The broad host range pathogen Sclerotinia sclerotiorum produces multiple effector proteins that induce host cell death intracellularly

Abstract Sclerotinia sclerotiorum is a broad host range necrotrophic fungal pathogen, which causes disease on many economically important crop species. S. sclerotiorum has been shown to secrete small effector proteins to kill host cells and acquire nutrients. We set out to discover novel necrosis‐inducing effectors and characterize their activity using transient expression in Nicotiana benthamiana leaves. Five intracellular necrosis‐inducing effectors were identified with differing host subcellular localization patterns, which were named intracellular necrosis‐inducing effector 1–5 (SsINE1–5). We show for the first time a broad host range pathogen effector, SsINE1, that uses an RxLR‐like motif to enter host cells. Furthermore, we provide preliminary evidence that SsINE5 induces necrosis via an NLR protein. All five of the identified effectors are highly conserved in globally sourced S. sclerotiorum isolates. Taken together, these results advance our understanding of the virulence mechanisms employed by S. sclerotiorum and reveal potential avenues for enhancing genetic resistance to this damaging fungal pathogen.

. S. sclerotiorum also produces an array of cell wall-degrading enzymes, which macerate host tissue and facilitate penetration by fungal hyphae (Riou et al., 1991). Host tissue acidification by oxalic acid and other acids produced by S. sclerotiorum enhances both the expression and activity of cell wall-degrading enzymes, further promoting disease progression (Cotton et al., 2003;Favaron et al., 2004;Rollins & Dickman, 2001).
It is not well understood how fungal effectors enter host cells to localize to their intracellular virulence targets. SsSSVP1 was shown to translocate into host cells and move from cell to cell in the absence of S. sclerotiorum; however, the mechanism underlying this movement remains unknown (Lyu et al., 2016). Cell-to-cell movement has been shown previously in two Magnaporthe oryzae effectors (Khang et al., 2010). Intracellular oomycete effectors tend to possess two N-terminal motifs, RxLR (Arg-Xaa-Leu-Arg) and dEER (Asp-Glu-Glu-Arg), that mediate host cell entry (Bouwmeester et al., 2011). A common host cell entry motif has not been identified in intracellular fungal effectors; however, an RxLR-like motif has been found to be involved in host cell entry in some fungal plant pathogen effectors including Melampsora lini AvrL567 and AvrM, Fusarium oxysporum f. sp. lycopersici Avr2, and Leptosphaeria maculans AvrLm6 (Kale, 2012;Kale et al., 2010;Rafiqi et al., 2010). An RxLR-like motif has also been identified in the MiSSP7 protein secreted by the mutualistic ectomycorrhizal fungus Laccaria bicolor, which is required for entry into root cells and symbiosis development (Plett et al., 2011). Thus far, no functional RxLR-like motif has been described in a broad host range plant pathogen.
Intra-and extracellular immune receptors monitor host cells to detect pathogen invasion. Intracellular immune receptors are typically nucleotide-binding leucine-rich repeat (NLR) proteins that directly or indirectly recognize the presence of pathogen effectors in what is known as a gene-for-gene interaction (Flor, 1956;Jones & Dangl, 2006). Effector recognition results in activation of effectortriggered immunity, which often culminates in localized cell death termed the hypersensitive response (HR). Detection of a biotrophic or hemibiotrophic effector results in immunity. On the other hand, detection of a necrotrophic effector and activation of an HR results in susceptibility in an inverse gene-for-gene manner, as the pathogen derives nutrients from the dead tissue (Liu et al., 2006;Lorang et al., 2007). Inverse gene-for-gene interactions have been described in the interaction between narrow host range necrotrophic fungal pathogens and host species (Liu et al., 2017;Shao et al., 2021).
Whether S. sclerotiorum uses effectors to hijack NLR-mediated HR is currently unknown, although the finding that an Arabidopsis thaliana NLR contributes to S. sclerotiorum susceptibility suggests that this may occur during infection (Barbacci et al., 2020).
In a previous bioinformatic study, 70 putative effector genes were predicted in the S. sclerotiorum genome (Derbyshire et al., 2017).
In this study, we show that five of these putative effectors trigger necrosis in planta. We demonstrate that the effectors function intracellularly and localize to different subcellular compartments. For the first time, we show evidence that a broad host range pathogen effector, S. sclerotiorum intracellular necrosis-inducing effector 1 (SsINE1), enters host cells using an RxLR-like motif. A gene knockdown screen in Nicotiana benthamiana indicated that SsINE5 may induce host cell death via an NLR protein, a novel mechanism of cell death induction for a broad host range pathogen. Collectively, these results further our understanding of the molecular mechanisms underlying virulence of a broad host range fungal pathogen.

| Multiple S. sclerotiorum putative effectors act intracellularly to induce necrosis in planta
We aimed to discover novel S. sclerotiorum necrosis-inducing effectors. To this end, a set of 70 putative S. sclerotiorum effectors previously identified in the reference isolate 1980 were selected as initial candidates to screen for necrosis-inducing effectors (Derbyshire et al., 2017). All 70 putative effectors were found to be conserved in the aggressive Australian isolate CU8.24 (Denton-Giles et al., 2018).
The localization of the putative CU8.24 effectors was predicted using ApoplastP (Sperschneider et al., 2018). This revealed 46 putative apoplastic effectors and 24 putative cytoplasmic effectors, the latter of which are predicted to enter host cells during infection. One of the 24 putative cytoplasmic effectors was found to be a conserved GTP-binding protein, SarA, involved in membrane trafficking, and was therefore removed from further analysis. Several characterized S. sclerotiorum effectors have been shown to function inside host cells and we hypothesized that there are additional as yet uncharacterized effectors that function intracellularly to induce host cell death. Therefore, we set out to screen the 23 putative cytoplasmic effectors for necrosis-inducing activity in planta. Additionally, there were five putative apoplastic effectors that have not been screened for necrosis-inducing activity in other studies and were significantly up-regulated during infection of B. napus, which were also included in the assay.
Seventeen putative cytoplasmic and four putative apoplastic effectors were successfully cloned, tagged at the C-terminus with green fluorescent protein (GFP), and expressed in N. benthamiana leaves by Agrobacterium-mediated transient expression (agroinfiltration) with appropriate controls (Table 1). As expected, the negative control of GFP with a C-terminal 6×His-tag (GFP-his) induced no cell death and the positive control, Peyronellaea pinodes NLP2 expressed with the Medicago truncatula PR-1 signal peptide (SP) and a C-terminal 6×His-tag (SP-NLP2-his), induced strong necrosis (Debler et al., 2021). The putative cytoplasmic effectors were expressed without their predicted native SP to retain the mature effector proteins inside host cells. Four of these effectors consistently induced necrosis. Necrosis symptoms became visible from 3 days postinfiltration (DPI) and photographs were taken at 7 DPI.
These are hereafter referred to as S. sclerotiorum intracellular necrosis-inducing effector 1-4 (SsINE1-4) ( Figure 1, Table 1). When expressed with the MtPR-1 SP to export the effector proteins to the apoplast, SsINE1 and SsINE2 induced an attenuated cell death response, whereas SsINE3 and SsINE4 induced no macroscopic cell death symptoms, indicating that the effectors require cytoplasmic localization for full activity ( Figure S1). The mature sequences of the four putative apoplastic effectors were expressed with the MtPR-1 SP to export the effector proteins to the apoplast and assay for cell death symptoms. One of the effectors induced weak necrosis.
We expressed this effector without the MtPR-1 SP and surprisingly observed a stronger cell death response. This effector is hereafter referred to as SsINE5 (Figures 1 and S1, Table 1). Western blot analyses confirmed accumulation of the SsINE-GFP fusion proteins with and without an SP except for SP-SsINE4-GFP, which did not elicit necrosis ( Figure S3).
To validate and quantify the macroscopic cell death symptoms caused by the intracellular effector proteins, we conducted an ion leakage assay. As expected, PpNLP2 induced leakage of significantly more ions from the agroinfiltrated leaf section than GFP, as would be expected in necrotic tissue. The ion leakage results corroborated the visual cell death symptoms in that all SsINEs induced greater ion leakage than GFP. Furthermore, SsINE1 and SsINE4 were the most potent necrosis-inducing effectors, as their expression resulted in greater ion leakage than that caused by the other SsINEs ( Figure 1b).
In summary, five necrosis-inducing effectors were identified that function intracellularly.

| The necrosis-inducing effectors localize to different subcellular compartments in host cells
Fungal pathogen effectors are known to function in various subcellular compartments. To provide insight into the cellular targets of the putative effectors, we predicted their subcellular localization using in silico methods and used confocal microscopy to determine the subcellular localization of the SsINEs in N. benthamiana epidermal cells. LOCALIZER did not predict any transit peptides or nuclear localization signals in the SsINEs. However, two of the nonnecrosis-inducing effectors were found to harbour putative mitochondrial transit peptides and two others were found to harbour putative nuclear localization signals ( mitochondria, although without mitochondrial marker colocalization we cannot draw any firm conclusions (Jaipargas et al., 2015). The visualization of chloroplast autofluorescence enabled the observation that SsINE3 appeared to cluster primarily around chloroplasts. SsINE4 primarily localized to punctate structures within the cells. SsINE3 and SsINE4 also partially localized to nuclei ( Figure 2). When expressed with the MtPR-1 SP, apoplastic localization of the effectors could be observed, demonstrating that the SP directed the proteins for export. Interestingly, GFP fluorescence could also be observed within nuclei of host cells when the SsINE-GFP variants were expressed with an SP ( Figure S2). Collectively, these data indicate that the SsINEs target different subcellular compartments to induce cell death.

| The necrosis-inducing effectors are highly conserved in globally sourced S. sclerotiorum isolates
As previously mentioned, the North American S. sclerotiorum reference isolate 1980 and Australian isolate CU8.24 both harbour all five SsINEs. We sought to investigate the conservation of the five SsINEs in S. sclerotiorum isolates found across the world and on diverse host species to provide some insight into their dispensability for S. sclerotiorum fitness. This analysis was conducted using genome sequences of 26 S. sclerotiorum isolates, including 1980 and CU8.24 (Derbyshire et al., 2017(Derbyshire et al., , 2019.
The SsINEs were present in all 26 isolates. Furthermore, the amino acid conservation was remarkably high. The amino acid sequences of SsINE2, SsINE4, and SsINE5 were identical in all isolates.
The amino acid sequence of SsINE1 was identical in 25 isolates; however, an isolate from South Africa (Sssaf) had one amino acid polymorphism in the predicted SP (V10I) and another amino acid polymorphism in the mature effector protein (D38G). The amino acid sequence of SsINE3 was identical in 19 isolates, with the other seven isolates harbouring only one amino acid polymorphism in the mature effector protein (Y145C). These seven isolates comprised the South African isolate Sssaf and six isolates sourced from Australia ( Figure 3a, Table S1). To summarize, the amino acid sequences of the SsINEs are highly conserved.

| An RxLR-like motif facilitates translocation of the SsINE1 effector protein into host cells
To investigate the possible functions of the SsINEs, they were scanned for protein domains using InterProScan. Besides the To test if the reduction in cell death symptoms was caused by reduced expression or accumulation of the SP-SsINE1 RTLT-AAAA -GFP variant, we also measured GFP expression. There was no significant difference in GFP expression between the SsINE1 variants, indicating that the reduction in cell death symptoms was not a result of reduced accumulation of the protein (Figure 4c). This finding was confirmed by western blot analysis using an anti-GFP antibody. To clarify, a reduction in protein expression was not observed when SP-SsINE1 RTLT-AAAA -GFP was expressed relative to SP-SsINE1-GFP ( Figure S4). All induced variant proteins were detected by immunoblotting. Interestingly, an additional higher-molecular-weight band was detected when SsINE1 variants were expressed with an SP, which could be unprocessed effector proteins carrying the SP or posttranslationally modified proteins ( Figure S4). This was carried out in an independent laboratory to the initial study. SsINE1, SsINE2, SsINE3, and SsINE5 also induced robust cell death in assays conducted in the independent laboratory; however, SsINE4 did not ( Figure S6). Therefore, SsINE4 was excluded from the VIGS assay. In our preliminary screen, we tested the requirement of SsINE3-and SsINE5-induced cell death was attenuated in com5-2 VIGS plants (five out of eight biological replicates for both effectors; Figure 6a). The NLR targeted by com5-2 was, therefore, a strong candidate to be mediating cell death induced by SsINE3 and SsINE5.
To confirm these findings, VIGS assays were repeated in an independent laboratory. The N. benthamiana phytoene desaturase gene (NbPDS) was silenced as a positive silencing control in both assays; the bleaching phenotype indicated successful silencing of the target transcript ( Figure S7) (Liu & Page, 2008). In the first assay, com5-2 and com5-3 were expressed and then SsINE3 and SsINE5 were ex- was not significantly reduced in com5-2 or com5-3 VIGS plants ( Figure 6b). In a further experiment, SsINE3 and SsINE5 were expressed in com5-2 VIGS plants. Again, cell death induced by SsINE5 but not SsINE3 was reduced in com5-2 VIGS plants ( Figure S8).
The amino acid sequences encoded by the NLR genes targeted by the six com5 cassettes were aligned to 121 known NLR sequences with ClustalW using the Jukes-Cantor model. The neighbour-joining method was used to construct a phylogenetic tree, which showed that the NLRs targeted by com5-1, com5-4, and com5-5 cluster with Solanum demissum potato late blight resistance protein R1 and the NLRs targeted by com5-2, com5-3, and com5-6 cluster with Solanum lycopersicum bacterial speck resistance protein Prf (Figure 6c). The two NLR candidates that are targeted by com5-2 and com5-3 share 39% and 77% amino acid identity with SlPrf, respectively. Both NLR candidates also share amino acid identity with SdR1 (37% and 26%, respectively). The NLR gene silenced by com5-2 encodes NbNLR 061-1, which has an N-terminal coiled-coil domain (CNL-type NLR) (Ahn et al., 2022).
Notably, a BLASTP search revealed that NbNLR 061-1 has homologues in several plant families across the asterid clade of dicots. The following cut-offs were used: an E-value of 1e−5, a query cover of 50%, and 35% amino acid sequence identity (Choudhuri, (c) Section of a phylogenetic tree showing the relationship between the NLR proteins targeted by the six com5 cassettes and 121 known NLR proteins. Alignment of full-length amino acid sequences was performed with ClustalW using the Jukes-Cantor model. The neighbour-joining method was used to construct the phylogenetic tree. The NLR targeted by com5-2 is labelled as "NbNLR 061-1." 2014). There are a total of 57 species that harbour NbNLR 061-1 homologues, with around half of them (29) in the Solanaceae family (Table S3). These data suggest that SsINE5-induced cell death in N. benthamiana is at least partially dependent on the CNL 061-1, which has homologues in numerous species in the asterid clade of dicots. (83%), respectively, localized to the nucleus (Caillaud et al., 2012;Liu et al., 2018). In fact, nuclear targeting has been demonstrated for a large number of characterized effectors from all classes of pathogens and nuclear localization is required for the necrosis-inducing activity of many effector proteins (Deslandes et al., 2003;Rivas & Genin, 2011;Schornack et al., 2010;Yin et al., 2022). In agreement with the nucleus being an important virulence target of pathogen effectors, nuclear localization of host immune receptor NLR proteins is widespread to allow detection of effectors and rapid transcriptional reprogramming to fend off invading pathogens (Shen & Schulze-Lefert, 2007).

| DISCUSS ION
In addition to nuclear localization, SsINE2 and SsINE5 were also observed in small punctate structures. The size of these punctae is consistent with mitochondrial size in green plants, which is generally 0.2-1.5 μm, although without mitochondrial marker colocalization we cannot be certain (Jaipargas et al., 2015). Mitochondria play a key role in governing programmed cell death in plant and animal cells (Vianello et al., 2007). Salmonella and Shigella bacterial pathogens produce type III secreted effectors that target animal cell mitochondria, resulting in promotion or inhibition of cell death (Nandi et al., 2021). Many of these effectors lack canonical mitochondrial transit peptides, which is also the case for SsINE2 and SsINE5, highlighting the importance of experimental validation of protein subcellular localization. However, the WoLF PSORT prediction for SsINE2 and SsINE5 was apparently consistent with the experimental evidence. Harpin is an elicitor produced by bacterial pathogens Erwinia amylovora and Pseudomonas syringae that targets mitochondria to induce plant cell death through inhibition of ATP synthesis and rapid cytochrome c release (Krause & Durner, 2004;Xie & Chen, 2000).
Similarly, the P. syringae type III effector AvrRpt2 affects mitochondrial function to induce cell death (Yao et al., 2004). The targeting of S. sclerotiorum effectors to mitochondria may represent an effective mechanism to activate cell death in a broad range of host species.
Previously, the S. sclerotiorum effector SsSSVP1 was shown to interact with a highly conserved subunit of the mitochondrial respiratory chain, QCR8. SsSSVP1 did not directly target the mitochondria but interacted with QCR8 in the cytoplasm and disrupted its correct localization to mitochondria, thereby inducing host cell death (Lyu et al., 2016).
SsINE3 appeared to target chloroplast outer membranes. The

S. sclerotiorum effector SsITL localizes to host cell chloroplasts,
where it interacts with a chloroplast-localized calcium-sensing receptor, CAS. This interaction results in reduced salicylic acid accumulation and impaired S. sclerotiorum resistance (Tang et al., 2020).
The necrosis-inducing effector ToxA produced by the necrotrophic wheat pathogen Pyrenophora tritici-repentis localizes to chloroplasts in susceptible wheat lines and interacts with a chloroplast-localized protein named ToxA-binding protein 1 (ToxABP1) (Manning et al., 2007;Manning & Ciuffetti, 2005). The localization pattern of SsINE3 observed in this study is reminiscent of the PtrToxA localization reported by Manning and Ciuffetti (2005). Neither ToxA nor SsINE3 possesses a canonical chloroplast transit peptide; however, WoLF PSORT predicted chloroplast localization of the mature SsINE3 protein.
Export of the SsINEs to the apoplast reduced or abolished cell death activity. We hypothesize that during infection the conditions created by S. sclerotiorum enable host cell entry of effectors.
Oxalic acid has been shown to affect cell membrane integrity, which could allow for the translocation of effector proteins into host cells (Tu, 1989). f. sp. lycopersici, and L. maculans, but never in a broad host range fungal pathogen (Kale, 2012;Kale et al., 2010;Rafiqi et al., 2010).
This opens up the question as to whether SsINE1 requires the RxLRlike motif to induce cell death in multiple host species. Despite the evidence in the literature that RxLR motifs mediate host cell translocation, Wawra et al. (2017) showed that the RxLR motif of the P.
infestans effector AVR3a is in fact cleaved before it is secreted by the pathogen. The RxLR motif bears similarity to Plasmodium export element (PEXEL) and Toxoplasma export element (TEXEL) motifs in effectors of the apicomplexan parasites Plasmodium falciparum and Toxoplasma gondii, respectively (Hofmann, 2017). These are cleaved prior to secretion in a similar manner to AVR3a to direct the effectors for secretion (Coffey et al., 2016). Whether this is a general process for RxLR and RxLR-like effectors remains to be seen. Our experiments indicate that the S. sclerotiorum SsINE1 RxLR-like motif plays a role in entry into N. benthamiana epidermal cells; however, further experiments would be required to clarify whether the RxLRlike motif is cleaved during infection or whether it mediates host cell entry in colonized host tissue. Although we identified a dEER-like motif, we showed that this is not essential for translocation into host cells. A dEER-like motif has never been found to be required for RxLR-like fungal effector translocation into host cells. RxLR-like motifs were also identified in SsINE2, SsINE3, and several of the nonnecrosis-inducing effectors; however, experimental validation is necessary to determine whether these motifs play a role in effector secretion or translocation into host cells.
The hijacking of effector-triggered immunity by a necrotrophic effector has not been explicitly demonstrated in a broad host range necrotrophic fungal pathogen. We provide evidence that SsINE5 may induce necrosis dependent on the NbNLR 061-1 gene.  InterProScan was used to identify conserved protein domains in the amino acid sequences of SsINEs (Zdobnov & Apweiler, 2001).

| Gene cloning and construct generation
Effector coding sequences without predicted SPs were amplified using gene-specific primers with partial attB sites. Single-exon genes were amplified from CU8.24 genomic DNA, whereas multi-exon genes were amplified from cDNA derived from bulked in vitro and in  (Choi et al., 2021). All constructs and primers are listed in Table S4 and Table S5.

| Agrobacterium-mediated transient expression (agroinfiltration) in N. benthamiana
The generated expression vectors were transformed into Agrobacterium tumefaciens AGL1 by electroporation. The trans-

| Western blotting
A. tumefaciens strains carrying respective expression vectors were agroinfiltrated in N. benthamiana leaves. Leaves were sampled at 2 DPI and flash frozen in liquid nitrogen. Two leaves per strain were ground in liquid nitrogen and total proteins were extracted from 2 g of tissue in 2 mL of protein extraction buffer (20 mM Tris-HCl pH 7.5, 300 mM NaCl, 5 mM MgCl 2 ) supplemented with 0.5% (vol/ vol) IGEPAL CA-630, 5 mM dithiothreitol, and one tablet of cOmplete Protease Inhibitor Cocktail (Roche) per 50 mL. Samples were centrifuged at 3200 × g at 4°C for 15 min. Supernatants were filtered through two layers of Miracloth (Millipore) and used as total protein extract samples. Proteins were separated by SDS-PAGE and analysed by immunoblotting with an anti-GFP primary antibody (Thermo Fisher) and a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Agrisera). Proteins were detected using Pierce ECL western blotting substrate (Thermo Fisher). Polyvinylidene difluoride (PVDF) membranes were stained with Ponceau S (Sigma) to visualize protein loading.

| Cell death quantification assays
For the ion leakage assay, the protocol of Yu et al. (2012) was followed with some modifications. Three leaf discs were taken per agroinfiltrated zone at 5 DPI and floated on 2 mL water in a 12-well plate. This represented one biological replicate. Three biological replicates were taken per treatment. Conductivity was measured using a Horiba EC-11 LAQUAtwin compact conductivity meter (Horiba).
This experiment was conducted twice with similar results. Analysis of variance was conducted with a post hoc Tukey honestly significant difference test to determine significant differences between relative ion leakage values.
To quantify cell death using red light fluorescence, agroinfiltrated N. benthamiana leaves were sampled at 7 DPI and analysed using a ChemiDoc MP imaging system, model Universal Hood III

| Confocal microscopy
Agroinfiltrated N. benthamiana leaf tissue was sampled at 2-3 DPI, mounted on a microscope slide, and viewed on a Nikon A1+ point scanning confocal microscope (Nikon). To assess apoplastic localization, samples were incubated in 30% glycerol for at least 30 min to induce plasmolysis prior to mounting. GFP fluorescence and chloroplast autofluorescence were excited at 489 nm and emission was captured at 525 nm and 700 nm, respectively. Objective magnification of 10× or 20× was used.

| VIGS assay
N. benthamiana was grown for approximately 2 weeks until the plants had two to four true leaves. Two leaves were coinfiltrated with A. tumefaciens strains transformed with pTRV1 and a pTRV2 construct (OD 600 = 0.4 for each strain). Successful silencing was confirmed by silencing the NbPDS gene and observing the expected bleaching phenotype. After 2 or 3 further weeks of growth, plants were agroinfiltrated with effector expression vectors as described above. The development and use of the NbNLR VIGS library is described in Ahn et al. (2022). Confocal microscopy analysis was performed using the Curtin