The barley powdery mildew effectors CSEP0139 and CSEP0182 suppress cell death and promote B. graminis fungal virulence in plants

The powdery mildew fungi secrete numerous Candidate Secreted Effector Proteins (CSEPs) to manipulate host immunity during infection of host plants. However, the function of most of these CSEPs in cell death suppression has not yet been established. Here, we identified several CSEPs from Blumeria graminis f. sp. hordei (Bgh) that have the potential to suppress BAX- and NtMEK2DD-triggered cell death in Nicotiana benthamiana. We further characterized two effector candidates, CSEP0139 and CSEP0182, from family six and thirty-two, respectively. CSEP0139 and CSEP0182 contain a functional signal peptide and are likely secreted effectors. Expression of either CSEP0139 or CSEP0182 suppressed cell death triggered by BAX and NtMEK2DD but not by the AVRa13/MLA13 pair in N. benthamiana. Transient overexpression of CSEP0139 or CSEP0182 also inhibited BAX-induced cell death and collapse of cytoplasm in barley cells. Furthermore, overexpression of either CSEPs significantly increased Bgh haustorial formation in barley, whereas host-induced gene silencing (HIGS) of the CSEP genes reduced haustorial formation, suggesting both CSEPs promote Bgh virulence in barley. In addition, expression of CSEP0139 and CSEP0182 reduced size of the lesions caused by the necrotrophic Botrytis cinerea in N. benthamiana. Our findings suggest that CSEP0139 and CSEP0182 may target cell death components in plants to promote fungal virulence, which extends the current understanding of the functions of Bgh CSEPs and provides an opportunity for further investigation of fungal virulence in relation to cell death pathways in host plants.

graminis species infecting crops of the grass family, normally colonizes barley as a host (Wyand and Brown 2003). Using integrated genomic, transcriptomic, and proteomic approaches, researchers have predicted and/ or identified~700 Candidates for Secreted Effector Proteins (CSEPs) from the Bgh genome, and investigated the functions of some of these CSEPs in fungal pathogenesis by host-induced gene silencing (HIGS) and overexpression analyses (Bindschedler et al. 2009;Godfrey et al. 2009;Nowara et al. 2010;Spanu et al. 2010;Pedersen et al. 2012;Pliego et al. 2013;Ahmed et al. 2015;Ahmed et al. 2016;Menardo et al. 2017;Frantzeskakis et al. 2018). It has been shown that~63% of the identified Bgh CSEPs are predicted to contain an N-terminal signal peptide (SP), and share an N-terminal Y/F/WxC motif that is speculated to have a function in effector translocation to host cells (Godfrey et al. 2010;Spanu et al. 2010;Pedersen et al. 2012). In addition, a large portion (c. 25%) of the Bgh CSEPs are RNase-like proteins associated with haustoria (RALPH effectors) that may act as pseudoenzymes with critical function Spanu 2017;Pennington et al. 2019). It is particularly interesting to note that most of the so far identified AVR A effectors, each recognized by a cognate MLA receptor, are RALPHs that share structural similarity with fungal RNases but lack the residues critical for catalytic activity Saur et al. 2019;Bauer et al. 2021). Furthermore, expression analyses have shown that many CSEP genes are differentially expressed during Bgh infection of barley, and many of them show predominant expression in haustoria (Spanu et al. 2010;Pedersen et al. 2012). Moreover, host targets that are involved in immunity and stress responses of some Bgh CSEPs have been identified, including pathogenesis-related (PR) proteins Pennington et al. 2016), small heat-shock proteins (HSPs) (Ahmed et al. 2015), and the ARF-GAP protein (Schmidt et al. 2014), among others (Pennington et al. 2016). However, the mechanisms by which Bgh CSEPs manipulate barley immune responses and enhance fungal virulence in host cells remain largely unknown.
Programmed cell death (PCD) plays an important role in a wide range of developmental processes and in responses to biotic and abiotic stresses in plants and animals (Dickman and Fluhr 2013). Although there are fundamental differences between plant and animal PCD, the basic morphological and biochemical features of PCDs are largely conserved (Das et al. 2009;Dickman and Fluhr 2013). BAX is a mouse pro-apoptotic protein with a conserved function of inducing cell death in plants. Expression of BAX in tobacco stimulates cell death, which closely resembles the hypersensitive responsive (HR) response induced by viral infection (Lacomme and Santa Cruz 1999). Cell death regulators have been identified in plants, using BAX-induced cell death. BAX inhibitor-1 (BI-1) homolog is one of the cell death suppressors isolated from different plant species, and has a conserved function in suppressing BAXinduced cell death (Huckelhoven 2004;Ishikawa et al. 2011;Xu et al. 2017). Similarly, using BAX and other types of cell death inducers, many pathogen effectors have been identified that can suppress cell death in plants (Lacomme and Santa Cruz 1999;Wang et al. 2011;Li et al. 2015;Xiang et al. 2016), such as Avr1b, Avr3b, Avh172, etc. from Phytophthora sojae (Dou et al. 2008;Dong et al. 2011;Wang et al. 2011), SNE1 from Phytophthora infestans, Six6 from Fusarium oxysporum, and Pst_8713 from Puccinia striiformis (Kelley et al. 2010;Gawehns et al. 2014;Zhao et al. 2018). Considering the existence of a large number of predicted effectors in Bgh and many other filamentous plant pathogens, it is envisaged that many of them should have the function in manipulating plant defense-related cell death (Panstruga 2003;Selin et al. 2016;Thordal-Christensen et al. 2018). However, so far only a few CSEPs, e.g. CSEP0264/ BEC1011 and CSEP0064/BEC1054, have been reported or implicated in suppressing cell death in host plant (Pliego et al. 2013;Pennington et al. 2019).
In the present study, we coexpressed CSEP candidates with a cell death inducer, BAX or NtMEK2 DD , and identified candidate Bgh effectors that may suppress HR-like cell death. NtMEK2 DD is a constitutively active mutant of NtMEK2 that activates defense responses and HR-like cell death in plants (Yang et al. 2001;Meng and Zhang 2013). We report here two CSEPs, CSEP0139 and CSEP0182 from family six and thirty-two respectively (Menardo et al. 2017), that could inhibit cell death responses in plants and enhance Bgh virulence during infection of barley. CSEP0139 and CSEP0182 suppressed BAX-induced cell death in both barley and N. benthamiana. Moreover, CSEP0139 and CSEP0182 suppressed cell death induced by NtMEK2 DD . Interestingly, they did not suppress cell death induced by the AVR a13 -MLA13 pair in N. benthamiana. We employed transient gene expression and HIGS assays, and demonstrated that CSEP0139 and CSEP0182 contribute to Bgh virulence in barley. Our data suggest that CSEP0139 and CSEP0182 are likely secreted effectors with cell death-suppressing activity, and able to promote Bgh virulence during infection of barley.

Results
Identification of CSEPs with cell death suppressing activity and expression analysis of CSEP0139 and CSEP0182 during Bgh infection of barley Previous studies have identified~500 candidate secreted effector proteins (CSEPs) in the Bgh genome, and many of these CSEP genes are highly induced and differentially expressed during Bgh infection of barley (Godfrey et al. 2010;Spanu et al. 2010;Pedersen et al. 2012;Hacquard et al. 2013). We selected some CSEPs for further study based on their relatively high expression in haustoria with respect to epiphytic tissues , and whether the CSEPs suppressed cell death in N. benthamiana triggered by BAX or NtMEK2 DD , a constitutively active mutant of NtMEK2 (Yang et al. 2001). We amplified~100 CSEP genes, which are distributed into 34 effector families, from isolate BghA6 by PCR, and subcloned them into the vector pGR107 for Agrobacterium tumefaciens-mediated transient expression in N. benthamiana . Fifteen out of these~100 CSEPs consistently suppressed BAX-and NtMEK2 DD -trigged cell death, and one of them varied in cell death suppression in different experiments (Additional file 1: Table S1).
Here, we further characterized CSEP0139 and CSEP0182, two effector genes from family six and thirty-two respectively (Menardo et al. 2017), which are highly expressed in haustoria but are not members of the RALPHs family, and expression of them suppressed cell death induced by BAX and NtMEK2 DD in N. benthamiana (Additional file 1: Table S1). The expression profiles of CSEP0139 and CSEP0182 were then confirmed in a time-course experiment in compatible interaction by RT-qPCR, using total RNAs collected from the entire infected leaves before haustorium formation at 0, 4, 8, and 12 h post inoculation (hpi), and RNAs separately collected from epiphytic Bgh tissues on leaves (E) and the remaining leaf tissues containing only the haustoria (H) after haustorium formation at 24 and 48 hpi. Transcript levels of CSEP0139 and CSEP0182 remained almost unchanged from 0 to 12 hpi, but significantly increased in both H and E samples at 24 and 48 hpi, with particularly more abundance in the H samples (Additional file 2: Figure S1). Moreover, the expression of CSEP0182 was significantly higher than that of CSEP0139, which is consistent with previously reported RNA-seq data . These results suggest that expression of both CSEP0139 and CSEP0182 genes is induced during haustorial formation and their transcripts are highly abundant in haustoria during Bgh infection of barley.

CSEP0139 and CSEP0182 contain a functional signal peptide
To examine if CSEP0139 and CSEP0182 are secreted proteins, we first analyzed if they contain an SP by using the Signal 4.0 software. Indeed, both CSEPs contain a putative N-terminal SP (Fig. 1a). In addition, a Y/FxC motif was also identified in both CSEPs, as previously indicated (Additional file 2: Figure S2) (Godfrey et al. 2010;Pedersen et al. 2012). We then examined if the SP was functional by using a yeast genetic assay, which is based on the requirement of invertase secretion for yeast cells to grow on a medium containing only raffinose as the carbon source (Gu et al. 2011). We constructed fusion plasmids in which the predicted SP sequences of CSEP0139 and CSEP0182 were fused in frame with a yeast invertase lacking its own SP in the pSUC2 vector. The first 25 amino acids from the M. oryzae protein Mg87 and the SP of the oomycete effector PsAvr1b were used as a negative and a positive control, respectively (Gu et al. 2011). The yeast cells of the YTK12 mutant were transformed separately with each of these plasmids to evaluate yeast growth on the YPRAA medium with raffinose as the carbon source. The invertase fusion plasmid with the SP from PsAvr1b enabled yeast to grow on the medium, while the plasmid with the 25 N-terminal amino acids of Mg87 did not, as expected. Importantly, plasmids with the SP from either CSEP0139 or CSEP0182 enabled the yeast to grow on the YPRAA medium (Fig. 1b). These results indicate that both CSEP0139 and CSEP0182 carry a functional SP and are likely secreted effectors.

Subcellular localization of CSEP0139 and CSEP0182 in barley and N. benthamiana
To examine the subcellular localization of CSEP0139 and CSEP0182, we generated two sets of constructs to express YFP fusion of the mature form of the CSEPs (lacking SP) in barley and N. benthamiana, respectively. The CSEP0139-YFP and CSEP0182-YFP fusions were individually coexpressed with a free CFP marker in barley epidermal cells by particle bombardment. Confocal imaging revealed that CSEP0139-YFP and CSEP0182-YFP were localized to both the cytosol and the nucleus of barley cells (Additional file 2: Figure S3a). CSEP0139-YFP and CSEP0182-YFP were then individually expressed in N. benthamiana by agroinfiltrationmediated transient gene expression, and were also found to localize in the cytosol and the nucleus (Additional file 2: Figure S3b). These results suggest that CSEP0139 and CSEP0182 localize to both the cytosol and the nucleus upon overexpression in barley or N. benthamiana.
CSEP0139 and CSEP0182 suppress BAX-and NtMEK2 DDinduced cell death in N. benthamiana Our initial screen of Bgh CSEPs identified sixteen CSEPs, including CSEP0139 and CSEP0182 that suppressed cell death in N. benthamiana (Additional file 1: Table S1). To further confirm the activity of CSEP0139 and CSEP0182 in cell death suppression, we included CSEP0340 as a negative control that did not suppress cell death induced by BAX and NtMEK2 DD (this study), and the Oomycete effector Avh328 as a positive control that consistently suppressed BAX-induced cell death . The CSEPs and the effector controls were infiltrated and expressed 12 h prior to that of the cell-death inducers, with the effector and cell-death inducer infiltrated next to each other but with an overlapping area in N. benthamiana. Indeed, the expression of either CSEP0139 or CSEP0182, or the positive control Avh328 consistently suppressed BAX-induced cell death ( Fig. 2a) and NtMEK2 DD -induced cell death ( Fig. 2b) in the overlapping area. By contrast, the expression of CSEP0340 did not suppress BAX-induced or NtMEK2 DDinduced cell death in the overlapping area ( Fig. 2a, b, bottom panels). These results indicate that CSEP0139 and CSEP0182 can suppress BAX-and NtMEK2 DD -induced cell death in N. benthamiana.

CSEP0139 and CSEP0182 do not suppress the AVR a13 -MLA13 pair induced cell death
To further test the activity of the two CSEPs in suppressing HR-like cell death induced by plant proteins, we coexpressed the barley immune receptor MLA13 with its cognate AVR effector AVR a13 (also known as Bgh CSEP0372 from family 34, and recently shown to adopt a predict common RNase-like fold) to trigger HR cell death in N. benthamiana Saur et al. 2019). Similarly, CSEP0139 and CSEP0182 or the control effectors were expressed 12 h prior to the expression of the AVR a13 /MLA13 pair (Fig. 3). Unexpectedly, the expression of CSEP0139 or CSEP0182 did not suppress AVR a13 / MLA13-triggered cell death (Fig. 3, upper two panels), and similarly, the expression of Avh328 or CSEP0340 did not suppress AVR a13 /MLA13-triggered cell death in N. benthamiana, either (Fig. 3, bottom panel), although partial suppression of AVR a13 /MLA13-triggered cell death was sometimes observed for the CSEPs and the controls (Fig. 3). These results indicate that CSEP0139 and CSEP0182 most likely do not suppress the AVR a13 -MLA13 pair triggered cell death in N. benthamiana.

CSEP0139 and CSEP0182 inhibit BAX-induced cell death in barley
To further test whether these two CSEPs suppress cell death in host barley plant, we adopted a BAX-induced cell death assay in barley (Eichmann et al. 2006). In this assay, GFP expression in transformed barley epidermal cells displays normal cytoplasmic strands in the living cells, and coexpression of BAX induces cell death and causes cytoplasmic collapse of the transformed cells, whereas the expression of barley BAX inhibitor-1 (HvBI-1) can delay or block this cell-death response (Eichmann et al. 2006). In our experiments, coexpression of GFP and BAX and EV typically induced cell death and led to the disappearance of cytoplasmic strands and nucleus, Fig. 1 CSEP0139 and CSEP0182 contain a functional signal peptide. a Schematic representation of CSEP0139 and CSEP0182, the N-terminal signal peptide (SP) was predicted by Signal 4.0 (http://www.cbs.dtu.dk/services/SignalP), highlighted in red. Shades of blue indicate F/YxC motif. b Yeast invertase secretion assay for the verification of SP. The SP sequences of CSEP0139, CSEP0182, and P. sojae PsAvr1b and the N-terminal sequence of M. oryzae Mg87 were individually fused to yeast invertase lacking its own SP in the pSUC2 vector. CMD-W (minus Trp) media were used to select positive transformants of yeast mutant strain YTK12 carrying the pSUC2 vector. Yeast growth on YPRAA media indicates secretion of invertase. The pictures were taken at 3 days after plating and accumulation of some irregular patches in most of the transformed cells (Fig. 4a, 2nd cell), thus the percentage of transformed cells showing collapse of the cytoplasm was set to 100% (Fig. 4b, 2nd bar). By contrast, coexpression of GFP and EV did not typically result in cell death in the transformed cells (Fig. 4a, 1st cell), with a relative low rate of cells showing collapse of the cytoplasm (~32%) (Fig. 4b, 1st bar). Importantly, the coexpression of GFP and BAX plus CSEP0139 or CSEP0182 significantly reduced the rate of cells with collapse of cytoplasm by3 0%, as compared to the coexpression of GFP and BAX and EV (Fig. 4a, b, 3rd & 4th bar). These results indicate that expression of CSEP0139 and CSEP0182 can inhibit or alleviate BAX-induced cell death in barley cells. Oomycete effector Avh328 was used as a positive control, and CSEP0340 as a negative control. Expression of CSEPs, BAX and NtMEK2 DD was done by agroinfiltration in different but overlapping areas in leaves of N. benthamiana, with 12 h time lapse between the agroinfiltration of the effector and the cell death inducer. Cell death was visualized by trypan blue staining, and photographs were taken at 5 days after infiltration. Protein accumulation of CSEP fusions were detected at 2 days after infiltration by Western blotting

CSEP0139 and CSEP0182 promote Bgh virulence in barley
To investigate the function of CSEP0139 and CSEP0182 in Bgh virulence, we transiently overexpressed CSEP0139 or CSEP0182 in barley epidermal cells by particle bombardment in both compatible and incompatible interactions (Shen et al. 2007;Bai et al. 2012). A GUS reporter was also coexpressed with the CSEPs to mark the transformed barley cells by GUS staining. The leaves of barley line P01 were used for bombardment and followed by inoculation with a virulent or an avirulent isolate, BghA6 or BghK1 respectively, and then fungal haustorium formation rate (haustorium index, HI%) was scored microscopically. Significantly, the expression of CSEP0139 or CSEP0182 increased Bgh haustorium index from~50% to~63% or~65% respectively in the compatible interaction, and from~16% to~24% or 26% respectively in the incompatible interaction ( Fig. 5a and Additional file 1: Table S2). These results suggest that both CSEP0139 and CSEP0182 can promote Bgh virulence during early stages of barley infection. We further took advantage of the HIGS approach in verifying the function of Bgh effectors (Nowara et al. 2010;Pliego et al. 2013;Ahmed et al. 2015). Silencing of CSEP0139 or CSEP0182 gene through HIGS was achieved by particle bombardments of an RNAi vector harboring sense and antisense fragments of the corresponding effector gene (Himmelbach et al. 2007). Silencing of either CSEP0139 or CSEP0182 significantly reduced the Bgh haustorium index by~50% or~40%, respectively, compared to the empty vector (EV) control ( Fig. 5b and Additional file 1: Table S3). Silencing of barley Mlo significantly reduced haustorium index by7 5%, similar as previously reported (Himmelbach et al. 2007). Thus, these HIGS data further support that CSEP0139 and CSEP0182 contribute to Bgh virulence in barley infection.

Expression of CSEP0139 and CSEP0182 reduces Botrytis virulence in N. benthamiana
Since necrotrophic fungi induce cell death for successful infection, we tested if the expression of CSEP0139 or CSEP0182 may affect the infection of a necrotrophic fungus, such as Botrytis cinerea. We individually expressed CSEP0139, CSEP0182 and GFP (used as a control) in N. benthamiana by agroinfiltration and inoculated B. cinerea spores in the middle of infiltrated area 24 h later, and scored disease symptoms at 2 and 3 days post inoculation (dpi) of B. cinerea (Fig. 6). Interestingly, pre-expression of CSEP0139 or CSEP0182 significantly reduced the size of disease lesions caused by B. cinerea by~22-29% at 3 dpi, as compared to the GFP control (Fig. 6). This result indicates that expression of CSEP0139 and CSEP0182 can reduce the virulence of B. cinerea in N. benthamiana, possibly by affecting cell death processes during B. cinerea infection.

Discussion
Many filamentous fungi are predicted to encode numerous secreted effector proteins, some of which have been shown to contribute to fungal virulence (Franceschetti et al. 2017). More than 500 CSEPs have been identified via genome mining in powdery mildew fungi, such as Bgh and Bgt (Blumeria graminis f. sp. tritici), however, only a fraction of them have been demonstrated to contribute to fungal virulence on plant hosts (Pliego et al. 2013;Menardo et al. 2017;Bourras et al. 2018;Thordal-Christensen et al. 2018). As obligate biotrophs, Bgh and Bgt depend entirely on living host tissues for their survival, and therefore, it is essential for them to manipulate host defense responses that may trigger PCD. These fungi may utilize multiple CSEPs to target components of host cell death pathways, however, the functions of such CSEPs have rarely been reported so far. Previously, a study conducted using 50 Bgh candidate effectors identified one effector (BEC1011/CSEP0264) that interferes with host cell death in barley (Pliego et al. 2013). Here, we report that CSEP0139 and CSEP0182, two Bgh effectors from two different CSEP families that do not belong to the RALPHs superfamily, suppress BAX-induced cell death, as well as NtMEK2 DD -triggered cell death in plants. These two effectors also contribute to Bgh virulence during infection of barley. Based on these data, we Representative pictures indicate that GFP fluorescence shows normal cytoplasmic strands in cells expressing GFP + EV, or GFP + BAX + CSEP0139 or GFP + BAX + CSEP0182, but shows collapsed cytoplasm strands in cell expressing GFP + BAX + EV. Photographs were taken using Nikon A1 confocal laser scanning microscope at 10 h after bombardment. Scale bar = 50 μm. b Quantification of BAX-induced cell death suppressed by CSEP0139 or CSEP0182. The percentage of cells showing collapse of the cytoplasm was shown in various plasmid combinations. The combinations expressing GFP + EV and GFP + BAX + EV were set as controls. The coexpression of GFP and BAX and EV typically induced cell death, thus the percentage of transformed cells showing collapse of the cytoplasm in GFP + BAX + EV was set to 100%. Data shown are means of three replicates ±SE. Means with different letters are significantly different (P < 0.05). Duncan's multiple range test was used to compare all the means. At least 100 cells were counted per experiment envisage that there are more CSEPs from powdery mildew fungi having redundant functions in suppressing cell death and promoting fungal virulence in the host.
CSEP0139 was selected as one of the candidate Bgh genes for silencing in a previous study (Aguilar et al. 2016), and knock-down of CSEP0139 expression was found not to affect fungal aggressiveness during early stages of infection. However, in this study we observed the reduction of haustorial formation upon silencing of CSEP0139. One possible explanation for this might be the variation in silencing efficiency in different cells as well as in different experiments. Furthermore, overexpression of CSEP0139 significantly increased haustorial formation in barley cells, which was consistent with the silencing data, and expression analyses in both studies indicated that CSEP0139 expression was induced by Bgh infection and its transcripts were highly abundant in haustoria at 24-48 hpi ( Fig. 5 and Additional file 2: Figure S1) (Aguilar et al. 2016).
Although BAX-like homologs may not exist in plants, BAX can stimulate cell death in plants (Dickman and Fluhr 2013). In mammals, BAX-induced cell death is associated with mitochondria, and BAX translocation from cytosol to the mitochondrial membranes results in the release of cytochrome c and stimulates the formation of an apoptosome complex, subsequently triggering cell death (Ishikawa et al. 2011). In plants, it has been shown that BAX-induced cell death and certain types of PCD are associated with mitochondria and release of cytochrome c (Lam et al. 2001;Yao et al. 2004;Tateda et al. 2009;Li et al. 2017). Expression of BAX in barley epidermal cells induces apparent cell death responses that can be clearly visualized through microscope (Eichmann et al. 2006). Using this expression system, we successfully identified the cell death-suppressing function of effectors CSEP0139 and CSEP0182. Here, overexpression of CSEP0139 and CSEP0182 delayed or affected BAXinduced cell death and increased barley susceptibility to Bgh isolates (Figs. 4 and 5a). It is noteworthy that BI-1 proteins, which are conserved cell death suppressors in animals and plants, efficiently suppressed BAX-induced cell death (Ishikawaet al. 2011). Previously, overexpression of barley HvBI-1 had clarified different aspects of its cell death-related function, such as weakening a cell-wallassociated local hydrogen peroxide burst in a resistant mlo-mutant barley, delaying Mla12-mediated race-specific resistance responses to Bgh, and enhancing barley susceptibility to Bgh fungus (Eichmann et al. 2006;Babaeizad et al. 2009;Eichmann et al. 2010). However, the question of how to relate the function of the two CSEPs to that of HvBI-1 in the context of BAX-induced cell death still remains unanswered. Animal and plant BI-1 proteins are ER-resident transmembrane proteins that are believed to act downstream of the BAX-induced mitochondrial Overexpression of CSEP0139 and CSEP0182 increased Bgh haustorium formation in barley. One-week-old barley leaves (P01) were bombarded with GUS reporter along with pUBi-CSEP0139 or pUBi-CSEP0182 construct, respectively. Leaves were inoculated with spores of virulent isolate BghA6 or avirulent isolate BghK1 at 4 h after bombardment. Bgh haustorium index (%) was microscopically scored at 48 hpi. EV: empty vector. Error bar indicates SD, and similar experiments were repeated three times. The asterisk represents significant differences from EV at P < 0.01, determined by Student's t test. b Silencing of CSEP0139 and CSEP0182 reduced Bgh haustorium formation in barley. The experimental setting was the same as in (a), with pIPK007 as the empty vector; HIGS constructs, pIPK007-CSEP0139 and pIPK007-CSEP0139, were constructed and used for silencing of CSEP0139 and CSEP0182, respectively. Inoculation of virulent isolate BghA6 spores was carried out at 48 h after bombardment. Bgh haustorium index of each transformation was normalized to that of the empty vector pIPK007 (set to 100%). pIPK007-Mlo was used as a positive control. Data show the average values and SD, similar experiments were repeated three times. Means with different letters indicate significant difference (P < 0.05). Duncan's multiple range test was used to compare all the means membrane modification and to regulate ER-stress responses (Xu et al. 2008). As both CSEP0139 and CSEP0182 can localize to both cytosol and nucleus upon overexpression, whether or not these two CSEPs may directly interact with barley HvBI-1 protein remains to be an open question. Since HvBI-1 expression is also induced in epidermal tissues after Bgh infection (Eichmann et al. 2010), it will be of interest to examine if HvBI-1 and CSEP0139/CSEP0182 act independently in barley cells, and if they show any similarities in inhibiting BAXinduced cell death. Based on our data, we can only speculate that these two effectors act downstream of BAXinduced cell death. Future experiments may reveal if these two CSEPs act downstream of mitochondria and on the release of cytochrome c, as well as ER-stress responses in plant cells.
Recent studies have identified several barley powdery mildew AVRa effectors, which are recognized by the cognate MLA receptors, and the expression of the matching AVR a -MLA pair triggers cell death in barley and N. benthamiana Saur et al. 2019;Bauer et al. 2021). Here, we found that the AVR a13 -MLA13 pair induced visible cell death in N. benthamiana, however, pre-expression of CSEP0139 and CSEP0182 did not show clear suppression of this type of cell death (Fig. 3). This result may indicate that cell death triggered by the AVR a13 -MLA13 pair may be somewhat different from that induced by BAX or NtMEK2 DD , while BAX and NtMEK2 DD might induce cell death through a potentially shared cell death signaling and/or similar mechanism. Nevertheless, we cannot fully rule out the possibility that CSEP0139 and CSEP0182 may partially suppress the AVR a13 /MLA13 pair triggered cell death in N. benthamiana (Fig. 3). Therefore, examining the activity of CSEP0139 and CSEP0182 in suppressing ETI-related cell death in host barley remains to be an interesting topic in the future. Since host cell death signaling pathways are probably the primary targets of biotrophic fungi, these CSEPs might be used as useful tools in probing the components of cell death signaling pathways in barley. To further elucidate the underlying mechanisms of CSEP0139 and CSEP0182 in suppressing cell death in host plant, their host targets need to be identified using integrated genetic and biochemical approaches in the future. Further experiments are required to reveal if CSEP0139 and CSEP0182 target different or same components/pathways of cell death signaling in host cells.

Conclusions
We provide evidence for the key functions of two Bgh effectors CSEP0139 and CSEP0182 in promoting fungal virulence and suppressing cell death in host plant. Both CSEPs carry a functional signal peptide and are localized Fig. 6 Expression of CSEP0139 and CSEP0182 reduces the size of lesions caused by Botrytis cinerea in N. benthamiana. CSEP0139 (a) or CSEP0182 (b) was expressed in N. benthamiana by agroinfiltration, with GFP construct as control. B. cinerea was inoculated at 24 h after agroinfiltration with GFP or indicated CSEP in the middle of pre-infiltrated areas, and pictures were taken at 3 dpi. The lesion size caused by B. cinerea infection was measured as the diameter of infected area at 2 and 3 dpi, and shown as the average values and SE from three independent experiments, each with 10 leaves. The asterisk indicates significant difference at P < 0.05 by Student's t test in the cytosol and nucleus when overexpressed in plant cells. Both CSEPs have the potential to suppress cell death induced by BAX and/or NtMEK2 DD in barley and N. benthamiana. On the contrary, pre-expression of CSEP0139 or CSEP0182 reduces the virulence of the necrotrophic fungus B. cinerea in N. benthamiana. These findings provide bases for further understanding the virulent strategies of the biotrophic fungal pathogens and will facilitate the development of efficient strategies for combatting these pathogens in the field.

Plant and fungal materials
Barley (Hordeum vulgare L.) cultivars Golden Promise and 'P01' (isogenic line from cv Pallas containing Mla1) were grown in a growth chamber under the conditions of 20°C with light for 16 h and 18°C in darkness for 8 h. N. benthamiana plants were grown at 24 ± 1°C under a 16 h light/8 h darkness photoperiod in a greenhouse.

RNA isolation and RT-qPCR
Total RNA was isolated using Trizol solution (Invitrogen), and digested with RNase-free DNase I (Takara) to eliminate potential DNA contamination. The first-strand cDNA was synthesized with reverse transcriptase M-MLV (Invitrogen) and used as template. To analyze the expression patterns of CSEPs, one-week-old first barley leaves (P01) were inoculated with the virulent Bgh isolate A6, and total RNA was isolated from leaves at 0, 3, 6, 12, 24, and 48 hpi. In addition, at 24 and 48 hpi, Bgh epiphytic tissues from the leaf surface and leaf tissues containing Bgh haustoria were separately collected. The epiphytic tissues were collected by dipping Bgh-infected leaves in 10% cellulose acetate (Ahmed et al. 2015). qPCR was performed with specific primers (Additional file 1: Table S4), using an ABI step-one real time PCR system and Gotaq qPCR Master Mix (Promega). Relative expression was determined by comparing with expression at 0 hpi, arbitrarily set to 1. Expression of the Bgh glyceraldehyde 3-phosphate dehydrogenase gene (G3PDH) was used to normalize the CSEP expression of each sample, using the delta deltact method (Ahmed et al. 2015). Three biological repetitions were carried out at each observation time point, and two independent experiments were performed.

Validation of the predicted SP of CSEPs
The predicted SPs of CSEP0139 and CSEP0182 were validated using the invertase secretion-deficient yeast strain YTK12, as previously described (Gu et al. 2011). The predicted SP sequences of CSEPs and oomycete effector Avr1b and the first 25 amino acids of M. oryzae Mg87 (Gu et al. 2011) were fused in frame with the yeast invertase lacking its own SP in the pSUC2 vector. The pSUC2-derived constructs (Additional file 1: Table S5) were transformed into the invertase secretion-deficient yeast strain YTK12, which were then plated on CMD-W medium (0.67% yeast N base without amino acids, 0.075% tryptophan dropout supplement, 2% sucrose, 0.1% glucose, and 2% agar). Yeast positive clones were transferred onto YPRAA medium (1% yeast extract, 2% peptone, 2% raffinose, 2 μg/L antimycin, and 2% agar) for invertase secretion assay.

Subcellular localization analysis
For the subcellular localization assay in barley, CSEP0139 and CSEP0182 (without SP) were cloned into the destination vector, pUbi-GW-YFP, by gateway technology, and the constructs thus obtained, i.e., pUbi-CSEP0139-YFP and pUbi-CSEP0182-YFP, were co-transformed with the pUbi-CFP marker construct into barley epidermal cells by particle bombardment (Shen et al. 2007). Two days after bombardment, the fluorescent signal was monitored, and pictures were taken using a Nikon A1 confocal laserscanning microscope. GFP/YFP and CFP were excited at 488 nm and 405 nm, respectively.
For the subcellular localization assay in N. benthamiana, CSEP0139 and CSEP0182 (without SP) were cloned into the CaMV 35S promoter-driven destination vector, i.e., CTAPi-GW-YFP, using gateway technology (Bai et al. 2012), to obtain the constructs CTAPi-CSEP0139-YFP and CTAPi-CSEP0182-YFP. These constructs were then delivered into A. tumefaciens strain GV3101 for transient expression in N. benthamiana, as described in the following section. The leaves were used for fluorescence detection 2 days after infiltration.

Single-cell transient gene expression and HIGS in barley
Transient gene expression assay in individual epidermal cells of barley was performed by particle bombardment, as previously described (Shen et al. 2007). CSEP (without SP) sequence was cloned into the pUbi-GW vector to generate pUbi-CSEP constructs, which together with a ß-glucuronidase (GUS) reporter vector, were delivered into barley leaf epidermal cells. The leaves were inoculated with the virulent Bgh isolate A6 or the avirulent Bgh isolate K1 at 4 h after bombardment, and were stained with GUS staining solution (0.1 M Na 2 HPO 4 / NaH 2 PO 4 , 10 mM Na 2 EDTA, 5 mM K 4 Fe[CN] 6 , 5 mM K 3 Fe[CN] 6 , 0.1% Triton X-100, 20% methanol, 1 g/L X-gluc, pH 7.0) at 48 hpi. Fungal haustoria were observed and evaluated in the GUS-expressing cells after the leaves were stained with 0.6% Coomassie solution and cleared with water. Haustorium index was calculated using the number of GUS-expressing cells with developed haustorium divided by total number of GUS-expressing cells with germinated but aborted spores (incompatible) and with developed haustorium (compatible). Significant differences between the constructs and empty vector were determined using Student's t test.
For transient silencing of CSEP genes of Bgh by HIGS, the CSEP0139 and CSEP0182 fragments were cloned into the 35S promoter-driven hairpin destination vector pIPK007, as previously described (Himmelbach et al. 2007). The following experimental procedure was the same as that for the transient gene expression assay, except that the leaves were inoculated with Bgh spores at 48 h after bombardment (Nowara et al. 2010). Silencing of HvMlo was used as a positive control (Himmelbach et al. 2007). Relative haustorium index was calculated as the haustorium index of each construct divided by that of the EV (pIKP007) construct. Significant differences between constructs and EV (pIKP007) were determined using ANOVA and Duncan's multiple range test.

Inhibition of BAX-induced cell death in barley
Expression of mouse BAX can lead to cell death in yeast, tobacco, Arabidopsis and barley (Lacomme and Santa Cruz 1999;Kawai-Yamada et al. 2001). We tested the inhibition of BAX-induced cell death by CSEP0139 and CSEP0182 in barley using the transient transformation method with particle bombardment, as described in a previous study (Eichmann et al. 2006). In brief, 0.05 μg of pUbi-BAX plasmid DNA and 0.5 μg of pUbi-GFP plasmid DNA per shot were transferred into one-weekold barley epidermal cells, and GFP fluorescence signal was used to detect the intact scaffolds or collapsed cytoplasmic strands in living cells, using a Nikon A1 confocal laser-scanning microscope, 10-14 h after transformation. For assaying the inhibition of BAX-induced cell death in barley, 0.05 μg of pUbi-BAX plasmid DNA, 0.5 μg of pUbi-GFP plasmid DNA, and 1.6 μg of pUbi-CSEP0139, pUbi-CSEP0182 or pUbi-EV plasmid DNA were delivered into barley epidermal cells by particle bombardment. The percentage of cells showing collapse of the cytoplasm (%) was scored as the number of GFP-expressing cells with collapsed cytoplasm divided by the total number of GFPexpressing cells. For each experiment, at least 100 GFPexpressing cells were scored.

Agroinfiltration-mediated transient gene expression in N. benthamiana
CSEP0139, CSEP0182 and CSEP0340 sequences without SP were cloned into the PVX vector pGR107 and the sequence of Flag or HA tag was added to the genes in frame using suitable restriction enzyme cutting sites to obtain constructs pGR107-CSEP0139, pGR107-CSEP0182 and pGR107-CSEP0340 . All constructs were confirmed by sequencing, and then introduced into A. tumefaciens strain GV3101 (pJIC SA_Rep) for transient expression in N. benthamiana. For this, agrobacteria carrying the aforementioned constructs were cultured overnight, washed with 10 mM MgCl 2 three times, and then resuspended with 10 mM MgCl 2 to achieve a final OD 600 of 0.5. After incubation for 3 h at room temperature, the resuspended solution was infiltrated into 4-week-old N. benthamiana leaves, using a needleless syringe. For further expression of BAX, NtMEK2 DD , and AVR a13 /MLA13 (mixed agrobacteria transformed with AVR a13 or MLA13 expressing constructs), agrobacteria were infiltrated 12 h later into a different but with overlapping area expressing CSEPs on the leaves. pGR107-CSEP0340 was used as a negative control, and pGR107-Avh328 from P. sojae was used as a positive control suppressing BAX-induced cell death .

Trypan blue staining for assessing cell death in N. benthamiana
Agroinfiltration-mediated gene expression was performed following the same procedure as mentioned in the previous section. Cell death symptoms were monitored at 3-5 days after infiltration, and the photographs were taken at 5 days after infiltration. For trypan blue staining, the leaves were boiled with a 1:1 mixed solution of trypan blue staining solution (10 mL lactic acid, 10 mL glycerol, 10 g phenol, and 10 mg trypan blue, dissolved in 10 mL distilled water) and ethanol for 5 min, and maintained overnight at room temperature. The leaves were then de-stained in chloral hydrate solution (2.5 g/mL) until the color was completely faded.

Botrytis cinerea infection assay
Four to five-week-old N. benthamiana leaves were used for transient gene expression and inoculated with Botrytis cinerea. In brief, one half of a leaf was used for infiltration of GFP constructs and the other half for CSEP0139 or CSEP0182 constructs, by agroinfiltrationmediated transient expression. One day later, in the middle of infiltrated area, 10 μL conidial suspension (10 6 conidia/ mL) of B. cinerea was inoculated. Three experiments were conducted for each CSEP, and 10 to 15 leaves were used per experiment. Lesion diameters of B. cinerea were measured for~10 representative leaves at 2 and 3 dpi, and photographs were taken at 3 dpi.

Western blotting
Total proteins, expressing the indicated constructs, were extracted from N. benthamiana leaves at 2 days after infiltration. Leaf samples (0.1 g) were ground and resuspended in 250 μL of 2 × loading buffer (1 M Tris·HCl, 5% SDS, 25% glycerol and 0.25 mg/mL bromophenol blue). Suspensions were boiled in a water bath for 5 min and then centrifuged at 12,000 rpm for 10 min. Twenty microlitre of the supernatant was used for SDS-PAGE electrophoresis, following which the proteins were transferred onto a nitrocellulose membrane, for 90 min at 200 mA. The membrane was stained with Ponceau solution to give equal loading, and then incubated with anti-Flag or anti-HA antibody. Super pierce ECL western blotting substrate was used for detection under a cooled CCD camera (Tanon 5200).
Additional file 1: Table S1. CSEPs suppressing cell death triggered by BAX and MEK2 DD in N. benthamiana. Table S2. Scoring of haustorium index after overexpression of CSEP0139 or CSEP0182 in barley P01 in compatible (P01/BghA6) and incompatible (P01/BghK1) interaction. Table  S3. Scoring of haustorium index after silencing of CSEP0139 or CSEP0182 by HIGS approach. Table S4. List of primers used in this study. Table  S5. List of constructs used in this study.
Additional file 2: Figure S1. Expression of CSEP0139 (a) and CSEP0182 (b) is induced during Bgh infection of barley. The barley isogenic line P01 was inoculated with the compatible isolate BghA6 for the time course experiments. Total RNA was isolated from Bgh-infected barley leaves at 0, 3, 6, 12, 24, and 48 hpi. H and E denote haustorium containing leaf tissues and epiphytic Bgh tissues, respectively. Relative expression of CSEP0139 or CSEP0182 was determined by comparing with expression at 0 hpi, arbitrarily set to 1. Bgh glyceraldehyde 3-phosphate dehydrogenase was used as the reference gene. Error bars indicate SD of three biological repetitions. Means with different letters indicate significant difference (P < 0.05). Duncan's multiple range test was used to compare all the means. Two independent experiments were performed with similar results. Figure S2. Nucleotide and amino acid sequences of CSEP0139 and CSEP0182. Shaded sequence is the predicted signal peptide (SP). Red box represents the F/YxC motif. The sequences underlined are used in the HIGS constructs for gene silencing. Figure S3. CSEP0139 and CSEP018 are localized to the cytoplasm and nucleus. a One-week-old barley leaves were bombarded with pUBi-CSEP0139-YFP or pUBi-CSEP0182-YFP, together with pUBi-CFP construct. b CSEP0139-YFP or CSEP0182-YFP was expressed by agroinfiltration in N. benthamiana leaves. Confocal images were taken 48 h after bombardment in barley or agroinfiltration in N. benthamiana, with excitation at 405 nm (CFP) or 488 nm (YFP) channel, using Nikon A1 confocal microscope. Scale bar = 50 μm.