A LlWRKY33-LlHSFA4-LlCAT2 module confers resistance to Botrytis cinerea in lily

Abstract Gray mold caused by Botrytis cinerea is one of the major threats in lily production. However, limited information is available about the underlying defense mechanism against B. cinerea in lily. Here, we characterized a nuclear-localized class A heat stress transcription factor (HSF)-LlHSFA4 from lily (Lilium longiflorum), which positively regulated the response to B. cinerea infection. LlHSFA4 transcript and its promoter activity were increased by B. cinerea infection in lily, indicating its involvement in the response to B. cinerea. Virus-induced gene silencing (VIGS) of LlHSFA4 impaired the resistance of lily to B. cinerea. Consistent with its role in lily, overexpression of LlHSFA4 in Arabidopsis (Arabidopsis thaliana) enhanced the resistance of transgenic Arabidopsis to B. cinerea infection. Further analysis showed that LlWRKY33 directly activated LlHSFA4 expression. We also found that both LlHSFA4 and LlWRKY33 positively regulated plant response to B. cinerea through reducing cell death and H2O2 accumulation and activating the expression of the reactive oxygen species (ROS) scavenging enzyme gene LlCAT2 (Catalase 2) by binding its prompter, which might contribute to reducing H2O2 accumulation in the infected area. Taken together, our data suggested that there may be a LlWRKY33-LlHSFA4-LlCAT2 regulatory module which confers B. cinerea resistance via reducing cell death and the ROS accumulation.


Introduction
Necrotrophic fungi kill host cells and infect host plants by secreting keratase, cell wall degrading enzymes, phytotoxic metabolites, and other pathogenic factors in host plants and subsequently absorb nutrients from dead cells of the host and grow and expand rapidly [1].Botrytis cinerea is a typical necrotrophic fungal pathogen, which infects a broad range of plants including fruits, vegetables, f lowers, causing huge economic losses in the world every year [2].Through the co-evolution of plants and their potential necrotrophic pathogens, plants have evolved complex strategies to defend against these pathogens, including regulation of hormone-signaling pathways, production of antimicrobial metabolites, and transcription factor (TF)mediated transcriptional regulation to protect themselves from necrotrophic pathogens [3][4][5].TF-mediated transcriptional regulation is a critical step to activate plant immune response [6][7][8].
HSFs exist widely in plants and are classified into three classes, HSFA, HSFB, and HSFC, based on the structural characteristics [9].Typically, HSFs, especially HSFAs which are intensively studied in abiotic stress, have been found to play roles in abiotic stress by binding heat shock elements (HSEs) in the promoters of their targeted gene to regulate abiotic stresses, including heat [10][11][12], drought [13], salinity [14], and oxidative damage response [15].There is evidence that HSFBs could participate in the response to pathogen attacks.HSFB1 was reported to suppress the innate immune response of Arabidopsis and could also be involved in the defense initiation and systemic acquired resistance in Arabidopsis [16,17].Oryza sativa HSFB4d directly binds to the HSE in the promoter of OsHsp18.0-CI to defend against Xanthomonas oryzae [18].However, the function of HSFAs in the plant's response to pathogen remains poorly understood.
In different plants, many studies have shown that WRKY33 is a key positive regulator for host immunity to B. cinerea.In Arabidopsis, WRKY33 was directly phosphorylated by MAP kinases after B. cinerea infection and activated camalexin biosynthesis gene PAD3 expression [19,20].Arabidopsis WRKY33 was also required for the hormonal signaling and metabolic responses to B. cinerea infection [21].WRKY33 decreased levels of abscisic acid (ABA) by directly repressing the expression of NCED3 and NCED5 during B. cinerea infection [22].In tomato (Solanum lycopersicum), WRKY33 homologues (SlWRKY33A and SlWRKY33B) can restore the resistance of Arabidopsis wrky33 mutant plants to B. cinerea [23].However, it was still unclear whether HSFs are involved in the regulatory network of WRKY33.
In this study, we report the functional characterization of LlHSFA4 and LlWRKY33 following inoculation with B. cinerea spores.Both LlHSFA4 and LlWRKY33 contribute to defense against B. cinerea by reducing cell death and the ROS accumulation in the infected area.Further studies indicate that LlWRKY33 binds the W-box element on the promoters of LlHSFA4 and LlCAT2 and

LlHSFA4 is a B. cinerea-inducible HSFA in lily
To determine the expression patterns of LlHSFAs (LlHSFA1, LlHSFA2, LlHSFA3A, and LlHSFA4) under B. cinerea infection, the detached leaf discs of Lilium longif lorum 'White Heaven' were chosen to be inoculated with B. cinerea spores.The RT-qPCR results implicated that LlHSFA4 was the only inducible HSFA gene compared with control plants, while the expression of LlHSFA1, LlHSFA2, and LlHSFA3A was not induced by B. cinerea (Fig. 1A; Fig. S1, see online supplementary material).Notably, the necrotic lesions were obviously observed on infected leaves at 36 hours post inoculation (hpi) (Fig. 1A), which implicated cell death occurs between 24 hpi and 36 hpi in the infected leaves.Next, we isolated 1500-bp promoter region of LlHSFA4 and identified its activity after B. cinerea infection.With an online promoter analysis, several cis-acting elements were present in LlHSFA4 promoter, including W-box element, low temperature responsive element (LTRE), R response element (RRE), ABA response element (ABRE), and others, which implicated that LlHSFA4 might participate in response to biotic or abiotic stress (Table S1, see online supplementary material).Then, a proLlHSFA4-GUS vector was constructed and transformed into lily to determine the promoter activity of LlHSFA4 after B. cinerea inoculation.The inoculated lily petal discs were subjected to histochemical staining for GUS activity after 0 hpi and 24 hpi.The blue color observed in lily petal discs indicates the active expression of LlHSFA4 gene.It was found that the lily petal discs at 24 hpi appeared to have a deeper blue color than those at 0 hpi, indicating that the promoter activity of LlHSFA4 could be activated by B. cinerea (Fig. 1B).These data indicate LlHSFA4 is induced by B. cinerea.

Silencing of LlHSFA4 increases susceptibility of lily to B. cinerea
To further confirm the function of LlHSFA4 in the response to B. cinerea, we performed tobacco rattle virus (TRV)-mediated VIGS of LlHSFA4 in lily leaves.The 5-day infiltrated leaves were harvested to test the silencing efficiency of LlHSFA4.And the results indicated that the LlHSFA4 was effectively silenced in lily plants (Fig. 2A).After B. cinerea inoculation, the LlHSFA4-silenced lily plants showed more severe disease symptoms than VIGS-controls (Fig. 2B and C).Then, we determined the H 2 O 2 accumulation at 24 hpi in the infected leaves of LlHSFA4-silenced leaves.The results showed that the infected leaves of LlHSFA4-silenced plants had a high level of H 2 O 2 accumulation at 24 hpi than that of the VIGS-controls (Fig. 2D and E), suggesting that more ROS was accumulated in LlHSFA4-silencing leaves during B. cinerea infection.In addition, the results of microscopical observation of trypan blue staining showed that more dead cells were clearly visible at 24 hpi in LlHSFA4-silenced leaves than in VIGS-control leaves (Fig. 2F).All these results implicated that silencing of LlHSFA4 increased the susceptibility of lily to B. cinerea.

Overexpression of LlHSFA4 improves the resistance of transgenic Arabidopsis to B. cinerea
The three LlHSFA4-transgeniclines (OE-5, OE-6, and OE-7) were confirmed by RT-PCR (Fig. S2, see online supplementary material).The wild-type and three LlHSFA4-transgenic lines were inoculated with B. cinerea using droplet-inoculation or spray-inoculation to explore the role of LlHSFA4 under B. cinerea infection.At 48 hpi, the necrotic disease symptoms of wild-type were more severe than that of the transgenic lines (Fig. 2G and H; Fig. S3, see online supplementary material).Trypan blue staining implicated that overexpression of LlHSFA4 decelerated the cell death in transgenic Arabidopsis (Fig. 2I).

LlWRKY33 directly activates the expression of LlHSFA4 by binding its promoter
It was found that six typical W-box elements were present in the promoter of LlHSFA4 (Fig. 3A).The WRKY transcription factor WRKY33 has been shown to act as a critical regulator in the response to B. cinerea infection [19][20][21][22][23][24][25].Therefore, we speculated that LlHSFA4 might be transcriptionally modulated by LlWRKY33.Yeast one-hybrid assay implicated that LlWRKY33 could directly bind the A4-P and only bind the truncated P6 fragment of LlHSFA4 promoter (Fig. 3B; Fig. S4, see online supplementary material).Promoter analysis of the A4-P6 fragment revealed that two W-box elements were present in the P6 fragment (Fig. 3A).Then, the A4-P6 fragment was further truncated into two fragments (A4-P6-1, A4-P6-2) for a Y1H assay (Fig. 3B).Further Y1H assay implicated that LlWRKY33 directly bound the W-box element located at −285 bp to −280 bp (GTCAA) on the A4-P6-1fragment, but not A4-P6-2 and muted A4-P6-1 fragment in which W-box elements (GTCAA) was mutated (TTTTT) (Fig. 3B).Based on the results of the Y1H assays, the promoter region containing W-box element of A4-P6-1fragment was synthesized into a 5 biotin-labeled probe for performing an electrophoretic mobility shift assay.The result of EMSA was consistent with those of Y1H (Fig. 3C).LUC activity of Nicotiana benthamiana leaves co-transformed with LlWRKY33 and proLlHSFA4-LUC was stronger than that of the SK-GFP and proLlHSFA4-LUC (Fig. 3E and F).Transient overexpression of LlWRKY33 also activated the transcription of LlHSFA4 in lily, while silencing of LlWRKY33 reduced its expression (Fig. 3G and H).Therefore, these results suggested that LlWRKY33 could serve as a direct regulator of LlHSFA4 and activate its expression.

Silencing of LlWRKY33 increases susceptibility of lily to B. cinerea
The LlWRKY33-GFP signal and the nuclear marker were colocalized in the nucleus in N. benthamiana leaves, indicating that LlWRKY33 was a nucleus-localized protein (Fig. S5A, see online supplementary material).The results of transactivation activity showed that LlWRKY33 had transcriptional activation in yeast cells (Fig. S5B, see online supplementary material).To verify the function of LlWRKY33 in the response to B. cinerea, we firstly performed RT-qPCR assay to examine the expression of LlWRKY33 in lily leaves after B. cinerea infection.The results showed that LlWRKY33's transcription was continuously elevated after B. cinerea infection and the highest expression level (26fold induction) was at 12 hpi (Fig. 4A).Next, a 521-bp promoter of LlWRKY33 was isolated and identified.Then, a proLlWRKY33-GUS was constructed and transformed into lily to determine the promoter activity of LlWRKY33 after B. cinerea inoculation.The proLlWRKY33-GUS activity was higher in lily at 24 hpi than its level at 0 hpi, indicating that the promoter activity of LlWRKY33 could be activated by B. cinerea (Fig. 4B).Then, we performed a VIGS experiment in lily.The results showed that LlWRKY33 expression was effectively silenced in LlWRKY33silenced lily plants compared with TRV2-controls (Fig. 4C).After VIGS-treatments, the detached leaves of 'White Heaven' were inoculated with B. cinerea spores to observe disease symptoms.The LlWRKY33-silenced lily plants showed more severe necrotic disease symptoms than VIGS-controls (Fig. 4D and E).The further DAB staining analysis revealed that the LlWRKY33-silenced leaves had a higher H 2 O 2 accumulation (Fig. 4F).The determination of H 2 O 2 content also confirmed this result (Fig. 4G).These data implicated that silencing of LlWRKY33 caused more B. cinereatriggered ROS generation in lily.The results of trypan blue staining show that cell death caused by B. cinerea infection was more clearly visible at 24 hpi in LlWRKY33-silenced leaves than in VIGScontrols (Fig. 4H).Based on these results, it could be concluded that LlWRKY33 confers B. cinerea resistance by alleviating cell death and the ROS accumulation.

Overexpression of LlWRKY33 enhances the resistance of transgenic Arabidopsis to B. cinerea
To further confirm the function of LlWRKY33, we transformed LlWRKY33 into Arabidopsis to further determine its function after B. cinerea infection.Three LlWRKY33-transgenic lines (OE-2, OE-3, and OE-4) confirmed by RT-PCR (Fig. S6, see online supplementary material) were selected for further studies.To examine the defensive effect of LlWRKY33 against B. cinerea, the transgenic Arabidopsis and wild-type were inoculated with B. cinerea  using droplet-inoculation or spray-inoculation.The LlWRKY33transgenic lines developed much milder symptoms than that of wild-type at 48 hpi (Fig. 4I and J; Fig. S7, see online supplementary material).Trypan blue staining assays also indicated that cell death was obviously inhibited in LlWRKY33-transgenic Arabidopsis compared with the wild-type (Fig. 4K), which suggested LlWRKY33-overexpression enhances the resistance of transgenic Arabidopsis to B. cinerea.

LlWRKY33 directly activates the expression of LlCAT2 by binding its promoter
Because the accumulation of LlWRKY33 in lily affects the levels of H 2 O 2 , we proposed that the expression of H 2 O 2 scavenging enzyme genes may be activated by LlWRKY33 after B. cinerea infection.To test this proposal, we tested the expression of several H 2 O 2 scavenging enzyme genes such as LlAPX2 (ASCORBATE PEROXIDASE 2), LlCAT2, LlSOS1(SALT OVERLY SENSITIVE 1), LlGPX8(GLUTATHIONE PEROXIDASE 8), and LlSOD1(SUPEROXIDE DISMUTASE 1) in lily plants.Transient overexpression of LlWRKY33 activated the expression of LlAPX2 and LlCAT2 in leaves and transient silencing of LlWRKY33 decreased their expression compared with that of control plants (Fig. 5A and B; Fig. S8, see online supplementary material).To further investigate the relationship between LlWRKY33 and LlAPX2, LlCAT2, we isolated the promoters of LlAPX2 and LlCAT2 and found that there were two W-box elements (−403 bp to −398 bp GTCAA; −286 bp to −281 GTCAA) on LlAPX2 promoter and one W-box element (−223 bp to −218 bp TTGAC) on LlCAT2 promoter (Fig. 5C; Tables S2 and S3, see online supplementary material).
Further results of Y1H assay implicated that LlWRKY33 directly bound the LlCAT2 promoter but not the LlAPX2 promoter (Fig. 5D; Fig. S9, see online supplementary material).Next, we truncated the LlCAT2 promoter into three fragments for Y1H assay and found that LlWRKY33 could only bind the P3 fragment, but not P1 or P2 (Fig. 5D).LlWRKY33 could not bind the mutant P3 fragment in which the TTGAC was mutated as TTTTT (Fig. 5D).Based on these results, the P3 region containing the W-box element was synthesized into a 5 biotin-labeled probe for an EMSA.The result of EMSA was consistent with those Y1H assay (Fig. 5E).Additionally, the LUC activity of co-transformed with LlWRKY33 and proLlCAT2-LUC was stronger than that of the SK-GFP and proLlCAT2-LUC (Fig. 5G and H).All these results suggested that LlWRKY33 could directly activate the transcription of LlCAT2.

LlHSFA4 binds to the promoter of LlCAT2 and activates its expression
Because silencing of LlHSFA4 caused more H 2 O 2 accumulation after B. cinerea infection, we then tested the expression of several H 2 O 2 scavenging enzyme genes such as LlAPX2, LlCAT2, LlSOS1, LlGPX8, and LlSOD1 in lily plants.The expression of LlCAT2 was increased in transiently LlHSFA4-overexpression lily leaves, while silencing of LlHSFA4 led to an opposite effect (Fig. 6A and B; Fig. S10, see online supplementary material).Then we proposed that LlHSFA4 might act as a regulator of LlCAT2 because LlHSFA4 has the characteristic of a transcription factor which had transcriptional activation in yeast cells and localized in the nucleus (Fig. S11, see online supplementary material).We analysed the promoter of LlCAT2 and found a potential HSE in the promoter of LlCAT2 (Fig. 6C).The results of Y1H assay implicated LlHSFA4 bound the promoter of LlCAT2 and only bind the P3, but not P1 or P2 (Fig. 6D).We also found that LlHSFA4 could not bind the mutant P3 (Fig. 6D).In addition, the luciferase reporter assay implicated that the LUC activity co-transformed with LlHSFA4 and proLlCAT2-LUC was stronger than that of the SK-GFP (Fig. 6F and G).Therefore, these data suggested that LlCAT2 was directly regulated by LlHSFA4.

Discussion
HSFAs are mainly implicated in plant thermotolerance by directly regulating their target genes [26][27][28].Previous studies have also shown that HSFAs (LlHSFA1, LlHSFA2, LlHSFA3, LlHSFA4) of lily are all involved in the heat stress response [29][30][31][32][33].In this study, we characterized LlHSFA4, which was induced by B. cinerea (Fig. 1; Fig. S1, see online supplementary material).In addition, our results indicated that LlHSFA4 was a positive regulator in the response to B. cinerea (Fig. 2).Studies in different species have shown that HSFA4 may be involved in a variety of stresses by regulating ROS homeostasis.Arabidopsis HSFA4A was reported to response to oxidative stress by regulating the expression of ROS scavenging enzyme genes [34].OsHSFA4a increased the Cd tolerance of rice by activating transcription of metallothionein gene [35].overexpression of Helianthus annuus HSFA9 enhanced the tolerance of transgenic tobacco to drought stress and oxidative stress [36].Additionally, HSFA4 enhanced the tolerance of Arabidopsis and chrysanthemum (Chrysanthemum morifolium) to salt stress through increasing regulating the ROS homeostasis [14,37].LlHSFA4 increased thermotolerance of transgenic plants through reducing H 2 O 2 accumulation and up-regulating the transcription of heat-related genes [33].These studies implicated that the HSFA4 could response to many stresses such as Cd, oxidative or salt stress by mediating ROS homeostasis.Consistent with previous studies, our results also indicate that LlHSFA4 acts as a positive regulator in the response to B. cinerea by directly binding the LlCAT2 promoter (Fig. 6).These results suggest that HSFA4 may play a role in the plant's response to biotic and abiotic stresses by regulating the ROS homeostasis.
Additionally, the several W-box elements were found in the promoter of LlHSFA4 (Table S1, see online supplementary material).It was found that, among the six typical W-boxes, only the W-box 6 appeared to be bound by LlWRKY33 (Fig. 3), which implicated that the binding of LlWRKY33 to the Wbox on the promoter of LlHSFA4 was selective.This result can be explained by a previous study on WRKY TF binding site preferences, which depended in part on the sequences f lanking W-box elements [38].Several studies have also found that the WRKY33 is an important regulator in the response to heat stress in addition to its involvement in the response to B. cinerea.In Arabidopsis, AtWRKY33 positively regulate the heat stress response by coordinately working with AtWRKY25, AtWRKY26 [39].Overexpression of SlWRKY33 can restore the tolerance of Arabidopsis wrky33 mutant plants to heat stress [23,40].Overexpression of Triticum aestivum WRKY33 enhances the tolerance of Arabidopsis to heat stress [41].Although WRKY33 and HSFA4 have been reported in the response to the heat stress and B. cinerea in plants, their regulatory mechanism between them is unclear.In this study, we demonstrated that LlHSFA4 is a direct target of LlWRKY33 (Fig. 3), which indicated that LlWRKY33 and LlHSFA4 were closely related in the response to these two stresses.However, whether the relationship between WRKY33 and HSFA4 is conserved in different plants requires further research.
In natural conditions, plants will encounter a variety of environmental stresses.Sometimes, plant may subject to one or combinations of two or more stress simultaneously because of the variability of the living environment [42,43].Pathogen attacks and temperature stress are one of the most frequently combinations of stress [44][45][46].Previous studies have also revealed that temperature affects plant-pathogen interactions by affecting disease resistance (R) genes effectiveness and durability or interfering with conidial germination and mycelial growth [47,48].B. cinerea occurs at low temperature and high humidity, which is most active at 23 • C and relative humidity above 80% in tomato [49].When the temperature rises above 30 • C, the virulence of B. cinerea will be weakened [50,51].This could be explained that high temperature increases the expression of defense genes, which induces the plant's resistance to B. cinerea [52].More reports have also shown that high temperature pretreatment has a direct positive effect on the host immunity, thereby improving their disease resistance [53][54][55].A recent study has shown that LlHSFA4 plays a positive role in the establishment of basal thermotolerance [33].In our study, LlHSFA4 confers resistance to B. cinerea (Fig. 2).These results suggest that LlHSFA4 may be potential linker in the response to heat stress and B. cinerea infection.However, the specific mechanism needs to be confirmed by further studies.
In this study, we identified the roles of LlHSFA4 and LlWRKY33 following inoculation with B. cinerea.Additionally, LlHSFA4, as a direct target of LlWRKY33, together with LlWRKY33 activates LlCAT2 to regulate ROS homeostasis (Fig. 7), thereby improving the resistance of lily to B. cinerea.

Protein localization and transcriptional activity analysis
The specific primers for genes cloning were shown in a list (Table S4, see online supplementary material).The pCAMBIA1300-GFP vector was used to construct LlWRKY33-GFP.The culture containing the pCAMBIA1300-LlWRKY33 plasmid was collected and resuspended in the buffer as previously described [56].The resuspended buffer was injected into N. benthamiana leaves.The GFP signal was observed by a laser scanning confocal microscope (LSM800, Zeiss, Germany).The pGBKT7 vector was used to construct pGBKT7-LlWRKY33 and pGBKT7-LlHSFA4.The vectors of pGBKT7 and constructed vector were introduced into AH109 yeast cells and screened on SD media (lacking Trp and His).

Pathogen inoculation
B. cinerea strain B05. 10 [57] was cultured on potato dextrose agar (PDA) medium for two weeks for sporulation.The B. cinerea spores were resuspended in sterile distilled water and collected by centrifugation and then adjusted to a concentration of 10 6 spores/ml with half-strength potato dextrose broth (PDB).For inoculation, the leaves of lily were chosen to be inoculated with 20 μL of the B. cinerea spores by droplet-inoculation.After inoculation of 60 h, phenotypes were photographed by a camera and lesion diameter of at least 20 infected leaves each time were determined using Image J software.The rosette leaves of one-month-old transgenic Arabidopsis and wild-type were chosen to be inoculated with 5 μL of the B. cinerea spores.After inoculation of 48 h, lesion diameter of at least 10 infected rosette leaves each time were determined using Image J software.All experiments were repeated three times.

Promoter isolation and GUS activity assay
All promoter sequences used in this study were obtained by the hiTAIL-PCR method [58].All promoter were analysed by the databases New PLACE (https://www.dna.affrc.go.jp/PLACE/?action=newplace).GUS staining was performed as previously described [59].

Gene expression assay
12-mm diameter samples detached from 'White Heaven' leaves with a puncher were chosen to be inoculated with 20 μL of the B. cinerea spores by droplet-inoculation.Then, the treated leaves were put on sterile filter paper and kept moist in a plastic dish.The treated leaves were collected at 0, 3, 6, 12, 24, 36, 48, and 60 hpi.RT-qPCR was performed with SYBR Premix Ex Taq II system (Vazyme, Nanjing, China) on the Quant Studio 6 Flex (ABI, Shanghai, China).Data was analyzed with the 2 -Ct method [60].The lily 18S rRNA was used as the endogenous gene.Three independent technical replicates were performed for each of three biological replicates.

Virus-induced gene silencing in lily
Agrobacterium-mediated infection of TRV-based VIGS assay [61] was used to silence gene expression in lily as described by our previous study [62].The 293-bp fragment of LlWRKY33 and 269bp fragment of LlHSFA4 were introduced into the pTRV2 vector.
The pTRV1, pTRV2, and the reconstructed vectors were transformed into Agrobacterium tumefaciens strain GV3101, individually.Agrobacterium culture containing TRV1 and TRV2 at a ratio of 1:1 was co-infiltrated into the leaves of lily.Then the infiltrated lily plants were put in dark conditions for 1 day and a culture room for 4 days.Then we collected 5-day leaves with a hole punch (12 mm in diameter) for inoculation to observe the disease symptoms or RT-qPCR analysis to detect the gene silencing efficiency.

Transient overexpression assay in lily
LlWRKY33 and LlHSFA4 were respectively introduced into the SK-II vector.Agrobacterium cultures containing SK-II, SK-LlWRKY33, or SK-LlHSFA4 were respectively injected into the leaves of lily.The 3-day infiltrated leaves were selected for RT-qPCR.

Histochemical staining
Inoculated leaves were stained for histological assay by DAB or trypan blue as previously described [63,64].The accumulation of H 2 O 2 was observed with DAB staining (1 mg/ml) and destained with 95% ethanol.The H 2 O 2 content was determined with a H 2 O 2 Content Assay Kit (Sangon Biotech, Shanghai, China).Trypan blue staining (0.04%) was used to detect cell death and destained with saturated chloral hydrate solution (2.5 g•mL −1 ).

Generation of transgenic Arabidopsis and phenotypic analysis
The bacterial solution of pCAMBIA1300-LlWRKY33 and pCAMBIA1300-LlHSFA4 were collected by centrifugation and resuspended in sucrose solution (5%).The resuspended bacterial solutions were infiltrated A. thaliana (Col-0) using the lowerdipping method [65] to generate transgenic Arabidopsis lines.The T3 generations of transformants were selected to perform the phenotypic analysis.To compare the resistance to pathogen of transgenic lines and wild type, the detached rosette leaves of transgenic lines and wild type were treated with B. cinerea spores.After 48 h, phenotypes were recorded.

Yeast one-hybrid assay(Y1H)
For Y1H assays, the ORF of LlWRKY33 and LlHSFA4 were respectively inserted into pB42AD plasmid (Clontech, CA, USA).The promoters of pLlHSFA4, pLlCAT2, and pLlAPX2 were respectively inserted into pLacZ vector (Clontech, CA, USA).Recombinant and empty vectors were introduced into EGY48a as described by a previous study [66], and cultured on the SD medium lacking Trp, Ura medium at 30 • C.

Luciferase reporter assay
The isolated promoter fragments of the LlHSFA4 and LlCAT2 were fused into pGreenII0800-LUC and the ORF of LlWRKY33 and LlHSFA4 were inserted into a pGreenII62-SK-GFP as described by our study [67].The resuspended bacterial solutions containing the effector plasmid and the reporter plasmid were injected into N. benthamiana leaves.The relative LUC activity was quantified using Image J v1.8.0 software.

Electrophoretic mobility shift assay (EMSA)
The pET32a vector was used to generate HIS-LlWRKY33 fusion protein.Target promoter region and mutant target region of LlHSFA4 and LlCAT2 were labeled with biotin by TSINGKE Biological Technology (Nanjing, China).Unlabeled DNA of the same sequence was used as a competitor.The EMSA assay was carried out with an EMSA kit (Thermo Fisher, New York, NY, USA).
The signal was detected by a CCD camera (Tanon 5200, Shanghai, China).

Figure 1 .
Figure 1.LlHSFA4 is induced by Botrytis cinerea.A The expression levels of LlHSFA4 in lily leaves after B. cinerea infection at various hours post inoculation.Disease lesion of leaf discs following inoculation with different hours was shown under the bar graph.Different letters indicate significant differences determined by Tukey's test with the standard deviation (P < 0.05).B. Analysis of the promoter activity of LlHSFA4 in lily petal discs at 0 hpi and 24 hpi.Scale bar = 1 cm.

Figure 2 .
Figure 2. LlHSFA4 positively regulates plant response to Botrytis cinerea.A The silencing efficiency of LlHSFA4 in TRV::LlHSFA4 plants detected by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).B Disease symptoms of TRV::00 and TRV::LlHSFA4 plants at 60 hpi by B. cinerea droplet-inoculation.C Lesion diameter of TRV::00 and TRV:: LlHSFA4 at 60 hpi by B. cinerea droplet-inoculation.The graph shows the average lesion size from three biological replicates (n = 60).Bars are determined by student's t test with the standard deviation ( * P < 0.05).D DAB staining of H 2 O 2 of TRV::00 and TRV::LlHSFA4 leaf discs at 0 hpi and 24 hpi.E The determination of H 2 O 2 content of TRV::00 and TRV::LlHSFA4 leaf discs at 0 hpi and 24 hpi.ns: not significant.F Trypan blue staining of TRV::00 and TRV::LlHSFA4 plants at 0 hpi and 24 hpi.Scale bar = 500 μm.G Disease symptoms of transgenic Arabidopsis (OE-5, OE-6, and OE-7) at 48 hpi by B. cinerea droplet-inoculation.Scale bar = 5 mm.H Lesion diameter of rosette leaves of the LlHSFA4-transgenic Arabidopsis at 48 hpi.Different letters indicate significant differences determined by Tukey's test with the standard deviation (P < 0.05).The graph shows the average lesion size from three biological replicates (n = 30).I Trypan blue staining of rosette leaves of the LlHSFA4-transgenic Arabidopsis at 48 hpi.Scale bar = 2 mm.

Figure 3 .
Figure 3. LlWRKY33 activates the expression of LlHSFA4.A LlHSFA4 promoter and W-box elements marked with blue triangles.B Y1H assay for LlWRKY33 and LlHSFA4 promoter.C EMSA of HIS-LlWRKY33 and the W-box element.D Schematic representation of the dual-luciferase reporter assay.E LUC signal in N. benthamiana leaves.Representative image was from one of three biological replicates.F Measurement of relative LUC activity.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).G The expression of LlWRKY33 and LlHSFA4 in LlWRKY33-overexpressing lily plants.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).H The expression of LlWRKY33 and LlHSFA4 in LlWRKY33-silencing lily plants.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05, * * P < 0.01).Three biological replicates were performed for each experiment.

Figure 4 .
Figure 4. LlWRKY33 positively regulates plant response to Botrytis cinerea.A The expression of LlWRKY33 in lily after B. cinerea infection at various hours post-inoculation.Disease lesion of leaf discs following inoculation with different hours was shown under the bar graph.Different letters indicate significant differences determined by Tukey's test with the standard deviation (P < 0.05).B Analysis of LlWRKY33 promoter activity in the lily petal discs.Scale bar = 1 cm.C The silencing efficiency of LlWRKY33 in TRV::LlWRKY33 plants detected by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).D Disease symptoms of TRV::00 and TRV::LlWRKY33 plants at 60 hpi by droplet-inoculation.E Lesion diameter of TRV::00 and TRV:: LlWRKY33 plants at 60 hpi.The graph shows the average lesion size from three biological replicates (n = 60).Significant differences are presented as means ± SD of three biological replicates (Student's t test, * P < 0.05).F DAB staining of H 2 O 2 accumulation of TRV::00 and TRV::LlWRKY33 plants at 0 hpi and 24 hpi.G Determination of H 2 O 2 content of TRV::00 and TRV::LlWRKY33 plants at 0 hpi and 24 hpi.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).ns: not significant.H Trypan blue staining of TRV::00 and TRV::LlWRKY33 plants at 0 hpi and 24 hpi.Scale bar = 500 μm.I Disease symptoms of rosette leaves from the LlHSFA4-transgenic Arabidopsis (OE-2, OE-3, and OE-4) at 48 hpi by droplet-inoculation.Scale bar = 5 mm.J Lesion diameter of rosette leaves of the LlHSFA4-transgenic Arabidopsis at 48 hpi.Different letters indicate significant differences determined by Tukey's test with the standard deviation (P < 0.05).The graph shows the average lesion size from three biological replicates (n = 30).K Trypan blue staining of rosette leaves from the LlHSFA4-transgenic Arabidopsis at 48 hpi.Scale bar = 2 mm.Three biological replicates were performed for each experiment.

Figure 5 .
Figure 5. LlWRKY33 activates the expression of LlCAT2.A The expression of LlWRKY33 and LlCAT2 in LlWRKY33-overexpressing lily plants confirmed by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).B The expression of LlWRKY33 and LlCAT2 in LlWRKY33-silencing lily plants confirmed by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05, * * P < 0.01).C The LlCAT2 promoter and the W-box element marked with blue triangle.D Y1H assay for LlWRKY33 and the LlCAT2 promoter.E EMSA of HIS-LlWRKY33 and the W-box element.F Schematic representation of the dual-luciferase reporter assay.G LUC signal in N. benthamiana leaves.Representative image was from one of three biological replicates.H Measurement of relative LUC activity.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).

Figure 6 .
Figure 6.LlHSFA4 activates the expression of LlCAT2.A The expression of LlHSFA4 and LlCAT2 in LlHSFA4-overexpressing lily plants confirmed by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).B The expression of LlHSFA4 and LlCAT2 in LlHSFA4-silencing lily plants confirmed by RT-qPCR.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).C LlCAT2 promoter and the HSE marked with purple triangle.D Y1H assay for LlHSFA4 and the LlCAT2 promoter.E Schematic representation of the dual-luciferase reporter assay.F The LUC signal in N. benthamiana leaves.Representative image was from one of three biological replicates.G Measurement of relative LUC activity.Significant differences are presented as means ± SD of three replicates (Student's t test, * P < 0.05).Three biological replicates were performed for each experiment.

Figure 7 .
Figure 7. Working model of LlWRKY33-LlHSFA4-LlCAT2 module under Botrytis cinerea attack.When lily is infected with B. cinerea, LlWRKY33 is rapidly induced and directly activate the expression of LlHSFA4 and LlCAT2 by binding to their promoters.Notably, LlHSFA4 could also directly bind to the HSE on the promoter of LlCAT2 and activate its expression.LlCAT2, a ROS scavenging enzyme, acts as a common target of LlWRKY33 and LlHSFA4 to prevent cell death caused by ROS accumulation to confer resistance to B. cinerea in lily.