An inhibitor of apoptosis (SfIAP) interacts with SQUAMOSA promoter‐binding protein (SBP) transcription factors that exhibit pro‐cell death characteristics

Abstract Despite the importance of proper cell death regulation across broad evolutionary distances, an understanding of the molecular machinery underpinning this fundamental process in plants remains largely elusive. This is despite its critical importance to development, homeostasis, and proper responses to stress. The identification of endogenous plant regulators of cell death has been hindered by the fact that many core regulators of cell death in animals are absent in plant genomes. Remarkably, numerous studies have shown that the ectopic expression of animal prosurvival genes in plants can suppress cell death imposed by many stresses. In this study, we capitalize on the ectopic expression of one of these animal prosurvival genes, an inhibitor of apoptosis from Spodoptera frugiperda (SfIAP), to identify novel cell death regulators in plants. A yeast two‐hybrid assay was conducted using SfIAP as bait to screen a tomato cDNA library. This screen identified several transcription factors of the SQUAMOSA promoter‐binding protein (SBP) family as potential SfIAP binding partners. We confirmed this interaction in vivo for our top two interactors, SlySBP8b and SlySBP12a, using coimmunoprecipitation. Interestingly, overexpression of SlySBP8b and SlySBP12a induced cell death in Nicotiana benthamiana leaves. Overexpression of these two transcription factors also induced the accumulation of reactive oxygen species and enhanced the growth of the necrotrophic pathogen Alternaria alternata. Fluorescence microscopy confirmed the nuclear localization of both SlySBP8b and SlySBP12a, while SlySBP12a was also localized to the ER membrane. These results suggest a prodeath role for SlySBP8b and SlySBP12a and implicate ER membrane tethering as a means of regulating SlySBP12a activity.

genetically regulated cellular suicide is referred to as programmed cell death (PCD). Programmed cell death has been studied extensively in animal systems, and the results of these research efforts have led to major treatment advances for many human diseases (Fuchs & Steller, 2011). In contrast, our understanding of the biochemical pathways underlying PCD in plants is severely lacking. This is largely due to the absence of obvious orthologs of core regulators of apoptosis, a well-studied form of PCD in animals (Kabbage, Kessens, Bartholomay, & Williams, 2017). While this has undoubtedly slowed progress on plant PCD research, it has also presented a unique opportunity for the discovery of novel cell death regulators in plant systems.
Apoptosis is a specific type of PCD characterized by distinct morphological and biochemical features (Kroemer et al., 2009). Apoptotic cell death in animals is executed through the activation of cysteinedependent aspartate-specific proteases termed caspases. Caspases exist as inactive proenzymes that can be activated by external or internal cellular cues. Once activated, caspases execute an orderly demise of the cell by targeting negative regulators of apoptosis, cytoskeletal components, and other caspases (Parrish, Freel, & Kornbluth, 2013). Due to the terminal nature of apoptosis, caspases must be kept under multiple layers of regulation. The inhibitor of apoptosis (IAP) family is an important group of proteins that negatively regulate caspase activity. The defining feature of all IAPs is the presence of one or more baculovirus IAP repeat (BIR) domains, which are used by IAP proteins for substrate binding (Verhagen, Coulson, & Vaux, 2001). Additionally, some IAPs contain a really interesting new gene (RING) domain that serves as a functional E3 ubiquitin ligase domain.
Inhibitor of apoptosis proteins can inhibit caspase activity by preventing procaspases from becoming active or by suppressing active caspases. This can be accomplished by simply blocking the active site pocket of a caspase or by utilizing the RING domain to ubiquitinate a caspase and mark it for proteasome-mediated degradation (Feltham, Khan, & Silke, 2012;Gyrd-Hansen & Meier, 2010).
Despite the fact that obvious orthologs of IAPs and caspases are absent in plant genomes, the ectopic expression of animal and viral apoptotic regulators in tobacco (Nicotiana spp.) and tomato (Solanum lycopersicum) modulate plant cell death. This was first reported nearly two decades ago when the expression of Bax, a mammalian proapoptotic gene absent in plant genomes, induced localized tissue collapse and cell death in Nicotiana benthamiana (Lacomme & Santa, 1999). Shortly thereafter, Dickman et al. (2001) demonstrated that expression of a viral IAP (OpIAP), as well as anti-apoptotic members of the Bcl-2 family, conferred resistance to a suite of necrotrophic fungal pathogens in Nicotiana tabacum. Pathogens with a necrotrophic lifestyle require dead host tissue for nutrient acquisition, and studies on Cochliobolus victoriae, Sclerotinia sclerotiorum, and Fusarium spp. revealed that these necrotrophic fungal pathogens hijack host cell death machinery to kill cells (Asai et al., 2000;Glenn et al., 2008;Kabbage, Williams, & Dickman, 2013;Lorang et al., 2012;Williams, Kabbage, Kim, Britt, & Dickman, 2011).
More recently, we showed that overexpression of an IAP from Spodoptera frugiperda (fall armyworm; SfIAP) in tobacco and tomato prevented cell death associated with a wide range of abiotic and biotic stresses (Kabbage, Li, Chen, & Dickman, 2010;. Tobacco and tomato lines expressing SfIAP had increased heat and salt stress tolerance, two abiotic stresses that induce cell death. These transgenic lines were also resistant to the fungal necrotroph Alternaria alternata and the mycotoxin fumonisin B1 (FB1) . Fumonisin B1 is produced by some species of Fusarium and is a potent inducer of apoptosis in animal cells and apoptotic-like PCD in plant cells (Gilchrist, 1997).
It has been over 15 years since it was first reported that overexpression of animal anti-apoptotic regulators in plants conferred enhanced resistance against a wide assortment of necrotrophic pathogens. During this time, numerous studies have confirmed the efficacy of animal apoptotic regulators in plants without identifying the means by which these regulators function. In this study, we used an unbiased approach to identify in planta binding partners of SfIAP in tomato to better understand how this insect IAP is able to inhibit cell death and confer stress tolerance in plants. Yeast two-hybrid and coimmunoprecipitation (CoIP) assays show that SfIAP interacts with members of the SQUAMOSA promoter-binding protein (abbreviated SBP in tomato or SQUAMOSA promoter-binding protein-like in some other species) transcription factor family. Overexpression of two tomato SBPs, SlySBP8b and SlySBP12a, induced cell death in tobacco leaves accompanied by enhanced production of reactive oxygen species (ROS). Overexpression of SlySBP8b and SlySBP12a also created an environment that was more conducive to the growth of the necrotrophic fungal pathogen A. alternata. In summary, our findings uncover SlySBP8b and SlySBP12a as novel SfIAP binding partners that exhibit prodeath attributes.

| Plant material and growth conditions
Nicotiana benthamiana plants were grown on a 16-hr light cycle (~50 microeinsteins m −2 s −1 ) at 26°C and~60% humidity. Nicotiana glutinosa (PI 555510) and tomato (Solanum lycopersicum cv. Bonny Best) plants were grown on a 16-hr light cycle (~100 microeinsteins m −2 s −2 ) at 22°C and~60% humidity. The soil composition for all plants consisted of SunGro ® propagation mix and Sunshine ® coarse vermiculite in a 3:1 ratio. Plants were watered with deionized water supplemented with Miracle-Gro ® all-purpose fertilizer (1 g/L) as needed.
Amplicons were recombined into the entry vector pDONR ™ /Zeo using BP clonase II (Invitrogen). SlySBP8b(NLS mt ) and SlySBP12a (NLS mt ) constructs were generated using the Q5 ® Site-Directed Mutagenesis Kit (New England Biolabs). SlySBP12a(ΔTMD) and TMD SlySBP12a were amplified from SlySBP12a in pDONR ™ /Zeo using the primers indicated in Supporting information Table S1 and recombined into pDONR ™ /Zeo. For overexpression in N. benthamiana leaves and tomato protoplasts, entry vectors were mixed with the desired pEarleyGate destination vectors (Earley et al., 2006) and recombined using LR clonase II (Invitrogen). pEarleyGate vectors drive transgene expression using a cauliflower mosaic virus 35S (35S) promoter and were used to generate N-terminal yellow fluorescent protein (YFP; pEarleyGate104) or N-terminal influenza hemagglutinin (HA; pEarleyGate201) fusions. All constructs were verified using Sangar sequencing before being transformed into Agrobacterium tumefaciens GV3101.
Plasmids for the yeast two-hybrid screen were prepared as follows. SfIAP, SfIAP BIR1 , and luciferase cDNAs were cloned into the bait vector pGilda under control of the GAL1 promoter and in-frame with an N-terminal fusion of the E. coli LexA DNA binding protein (Takara Bio USA, Inc.). Luciferase (firefly luciferase from Photinus pyralis) was cut from an existing plasmid using a 5′-Nco1 restriction site in the START codon and a 3′-Not1 restriction site outside of the ORF and ligated into pGilda. Primers for SfIAP (GenBank: AF186378.1) and SfIAP BIR1 amplification were designed to place an EcoR1 site at the 5′ end and a BamH1 site at the 3′ end of the ORF. Primers used for amplification can be found in Supporting information Table S1.
Amplicons were cut using these restriction enzymes and ligated into pGilda. Tomato cDNAs were expressed from the GAL1 promoter with an N-terminal fusion of the B42 activation protein in the pB42AD plasmid (Takara Bio USA, Inc.). Bait and prey library were sequentially transformed into EGY48 yeast using standard protocols.

| Yeast two-hybrid screening
Yeast containing bait and plasmid were plated on SD galactose (-His/ -Trp/-Leu) to induce gene expression and select for bait-prey interactions. After incubating at 28°C for~5 days, colonies were pooled in 10 ml of sorbitol/phosphate buffer (1.2 M sorbitol, 0.1 M NaPO 4 , pH 7.5) per plate, pelleted, and resuspended in 2 ml of sorbitol/ phosphate buffer supplemented with 500 U of lyticase (Sigma: L2524-25KU) and 250 μg of RNase A. Yeast cells were incubated in the lyticase buffer for 3 hr at 37°C prior to plasmid recovery.
Plasmid DNA was extracted using a Wizard Plus SV Miniprep kit (Promega) and a modified protocol. Briefly, 2.5 ml of lysis solution and 80 μl of alkaline protease solution were added to yeast protoplasts and incubated at room temperature for 10 min. Next, 3.5 ml of neutralization solution was added and cellular debris was pelleted by centrifugation. Supernatant was run through the provided columns and plasmid DNA eluted according to the manufacturer's instructions. Low-cycle PCR was performed to amplify cDNA's from the prey library. Briefly, MyFi ™ proofreading DNA polymerase (Bioline) and pB42AD forward and reverse primers (flanking the cDNA insertion site of pB42AD) were used to amplify cDNA's (Supporting information Table S1). A QIAquick PCR purification kit (Qiagen) was used to clean PCR products before sequencing.

| Illumina sequencing and data analysis
Sequencing was performed by the Biotechnology Center at UW-Madison using Illumina next-generation sequencing with 100-bp paired-end reads. The sequencing data were uploaded to the Galaxy web platform, and we used the public server at usegalaxy.org to analyze the data (Afgan et al., 2016). Reads were groomed and trimmed to remove low-quality bases and adapter sequences before alignment (Bolger, Lohse, & Usadel, 2014). Bait (pGilda) and prey (pB42AD) plasmid sequences were concatenated with the Saccharomyces cerevisiae reference genome (S288C_reference_se-quence_R64-2-1_20150113) to create a FASTA file containing sources of plasmid and gDNA contamination. Reads were aligned to this file using Bowtie 2 (Langmead & Salzberg, 2012). Aligned reads (plasmid and gDNA) were discarded while unaligned reads were aligned to the tomato reference genome (Solgenomics: ITAG2.4) with Bowtie 2. Cufflinks (v2.2.1) was used to assemble transcripts from these aligned reads and calculate FPKM values for each locus (Trapnell et al., 2012). Enrichment scores for each locus were calculated using R Studio and scripts written in-house (RStudio Team, 2016). Details of Galaxy pipeline, in-house scripts, and complete dataset are available upon request.

N. glutinosa
Agrobacterium strain GV3101 was grown overnight in liquid LB supplemented with gentamycin and kanamycin (50 μg/ml) at 28°C with shaking. Cells were harvested by centrifugation, washed once with sterile deionized water, and resuspended in infiltration medium (10 mM MgSO 4 , 9 mM MES, 10 mM MgCl 2 , 300 μM acetosyringone, pH 5.7) to a final concentration of OD 600 = 0.9. Cultures were incubated at room temperature for 4 hr before infiltration.
Nicotiana benthamiana plants were infiltrated with a 1-ml needleless syringe at 4-5 weeks of age with the two youngest and easily infiltratable leaves being used. Nicotiana glutinosa plants were infiltrated at 5-6 weeks of age with a single leaf being used on each plant, typically corresponding to the 4th or 5th true leaf. Plants were transformed at different ages due to differences in rate of growth between the two species.
For total protein extraction, leaf tissue was frozen in liquid nitrogen and ground in 3× Laemmli buffer (10% β-mercaptoethanol). Samples were boiled for 10 min followed by centrifugation at 10,000 g for 5 min. Supernatants were removed and transferred to new tubes.  antibodies. The α-GFP antibody was detected using goat α-mouse IgG conjugated to horseradish peroxidase (HRP) (Cell Signaling 7076P2) while the α-HA antibody was detected using goat α-rabbit IgG conjugated to HRP (Cell Signaling 7074P2). Amersham ™ ECL ™ reagent (GE Life Sciences) was used to detect HRP-conjugated antibodies.

| Transient transfection of tomato protoplasts
Mesophyll protoplasts form tomato cotyledons were isolated from 10-day-old plants using the Tape Sandwich method (Wu et al., 2009). A total of 6 μg of plasmid was used for each transfection with an equal ratio used for cotransfections. Transfections were performed using polyethylene glycol (PEG) as described previously (Yoo, Cho, & Sheen, 2007). Protoplasts were used for imaging the day after transfection. Tris-HCl, 150 mM NaCl, 5 mM EDTA, 10% glycerol, 0.2% IGEPAL, and 1% plant protease inhibitor cocktail [Sigma]) was added at a concentration of 2 ml/g of leaf tissue. YFP-tagged proteins were immunoprecipitated by incubating the lysate with α-GFP magnetic agarose beads (GFP-Trap_MA; Chromotek) for 2 hr at 4°C. Beads were washed three times in extraction buffer (w/o IGEPAL) and boiled in 30 μl of 2× SDS loading buffer before loading on duplicate 12% Tris-Glycine-SDS polyacrylamide gels (Bio-Rad). Proteins were transferred to duplicate nitrocellulose membranes and probed with α-GFP (Cell Signaling 2955S) or α-HA (Cell Signaling 3724S) primary antibodies. The α-GFP antibody was detected using goat α-mouse IgG conjugated to HRP (Cell Signaling 7076P2) while the α-HA antibody was detected using goat α-rabbit IgG conjugated to HRP (Cell Signaling 7074P2). Amersham ™ ECL ™ reagent (GE Life Sciences) was used to detect HRP-conjugated antibodies.

| Confocal laser scanning microscopy
Confocal laser scanning microscopy was performed on a Zeiss ELYRA LSM780 inverted confocal microscope using a 40×, 1.1numerical aperture, water objective. Protoplasts were stained with DHE at a concentration of 5 μM by adding DHE directly to protoplast solution and imaging approximately 10 min after DHE addition.
YFP fusions, chlorophyll autofluorescence, and DHE were excited with a 488 nm argon laser. YFP emission was detected between 502 and 542 nm, chlorophyll emission was detected between 657 and 724 nm, and DHE was detected between 606 and 659 nm. mCherry was excited with a 561 nm He-Ne laser, and emission was detected between 606 and 651 nm. Colocalization analysis was performed using the Coloc 2 package in ImageJ. A region of interest was selected on the image, and the analysis was performed using a PSF of 3.0 with 100 Costes randomizations.

| Electrolyte leakage analysis
Cell death progression in N. benthamiana leaves was assessed by measuring ion leakage. Approximately 24 hr post-agroinfiltration, eight leaf disks were collected from two leaves on the same plant and pooled into a single well of a 12-well plate, representing a single biological replicate. Leaf disks were washed for 30 min in 4 ml of deionized water by rotating plates at 50 rpm at room temperature.
Wash water was removed and replaced with 4 ml of fresh deionized water. Immediately after adding freshwater, the conductivity of the solution was recorded, representing the 24 hr post-agroinfiltration measurement. The conductivity of the water was measured using an ECTestr 11 + MultiRange conductivity meter (Oakton) at the indicated time points.

| DAB staining of N. benthamiana leaves
Staining solution was prepared by dissolving 3,3′-diaminobenzidine (DAB; Sigma) in HCl at pH 2. Once dissolved, this solution was added to Na 2 HPO 4 buffer (10 mM) for a final DAB concentration of 1 mg/ml. Tween 20 (0.05% v/v) was added, and the final pH was adjusted to 7.2. Whole leaves were collected, placed in petri dishes, submerged in DAB staining solution, and vacuum infiltrated. Plates were covered in aluminum foil and incubated at room temperature with shaking. After 4 hr, DAB staining solution was removed and replaced with clearing solution A (25% acetic acid, 75% ethanol).
Leaves were heated at 80°C for 10 min to remove chlorophyll.
Clearing solution A was removed and replaced with clearing solution B (15% acetic acid, 15% glycerol, 70% ethanol). Leaves were incubated in clearing solution B overnight at room temperature to remove residual chlorophyll.

| A. alternata inoculation of N. glutinosa leaves
Alternaria alternata isolated from potato was provided by Dr.
Amanda Gevens (University of Wisconsin, Madison, WI). For FB1 treatments, leaves were coinfiltrated with an Agrobacterium suspension containing the 35S:YFP construct and 5 μM FB1 (Cayman Chemicals). Leaves were harvested from N. glutinosa plants one day after agroinfiltration. Detached leaves were placed adaxial-side up in petri dishes (100 mm × 20 mm) containing three layers of wet filter paper. Five-mm-diameter agar plugs were collected from the edge of an actively growing fungal colony on potato dextrose agar. Leaves were wounded with a 1-ml pipette tip along the midrib, and agar plugs were placed fungal-side-down on top of the wound. Inoculated leaves were kept at room temperature (~23°C) for the duration of the experiment. Lesion areas were recorded 3 days after fungal inoculation.

| Image acquisition and analysis
All leaf images were taken using a Nikon D5500 camera with a Nikon AF-S NIKKOR 18-55 mm lens. Quantification of DAB staining intensity and fungal growth was performed using the Fiji package for ImageJ (Schindelin et al., 2012). For quantification of DAB staining intensity, the Colour Deconvolution package was used to isolate the DAB color channel for each DAB-stained leaf (Ruifrok & Johnston, 2001). Staining intensity caused by 35S:YFP expression on the left half of each leaf was subtracted from the staining intensity caused by 35S:HA-SlySBP8b or 35S:HA-SlySBP12a expression on the right half of the same leaf. Fungal lesions were quantified by tracing the periphery of the lesion and calculating the area within the periphery using ImageJ. Statistical analyses were performed using a one-way analysis of variance (ANOVA) with Tukey's honest significant difference (HSD) test in R Studio (RStudio Team, 2016).

Arabidopsis genes
The online bioinformatic tool FindM was used to search the eukaryotic promoter database (EPD) for promoters containing the 5′-CCGTAC(A/G)-3′ cis-element bound by the SBP domain of SBP transcription factors. Promoter regions were defined as the genomic region 1,000-bp upstream of the transcription start site, and a bidirectional search mode was used. Genomic sequences used were from the TAIR 10 version of the Arabidopsis genome.

| Identification of SfIAP binding partners in tomato
To identify putative binding partners of SfIAP from tomato, we performed a yeast two-hybrid assay coupled with next-generation sequencing using a method developed by Lewis et al. (2012)  The cDNA library itself was also sequenced to account for biases in transcript abundance.
Enrichment scores were calculated for each locus using the equation in Supporting information Figure S1B. A total of 13 putative interactors with enrichment scores of 50 or higher were identified in our screen (Table 1). Interestingly, this list contained six members of the SQUAMOSA promoter-binding protein (SBP) family of transcription factors. Based on enrichment scores, the top interactor with full-length SfIAP was SlySBP8b (95.7) while the top interactor with SfIAP BIR1 was SlySBP12a (98.7). Also present at lower enrichments were SlySBP4, -6a, -6c, and an unannotated homolog referred to as "SlySBP-like" (Table 1).

| Induction of tissue death by SlySBP8b and SlySBP12a
SfIAP is known to inhibit apoptosis in S. frugiperda and suppress cell death when ectopically expressed in tomato and tobacco Muro, Hay, and Clem, 2002). Thus, we anticipated that SfIAPinteracting partners in plants may be positive regulators of cell death. To narrow our list of candidate genes, we transiently overex-   (Cerio, Vandergaast, & Friesen, 2010). This is particularly important as we show that cleavage at the N-terminus of the full-length protein occurs in N. benthamiana, thus removing N-terminal epitope tags (Supporting information Figure S3).
We noticed that SlySBP12a was not enriched in our yeast twohybrid when full-length SfIAP was used as bait but only appeared when the SfIAP BIR1 truncation was used (Table 1). To account for the possibility that SfIAP may interact transiently with SlySBP12a, an E3 ligase mutant of the truncated SfIAP protein was generated by mutating a conserved residue in the RING domain (Cerio et al., 2010). This construct, referred to as SfIAP M4 (I332A), is resistant to N-terminal cleavage in N. benthamiana (Supporting information Figure S3).

SlySBP12a abolishes cell death induction
In addition to its role in nuclear targeting, the NLS of SBP proteins forms a positively charged surface that is required for DNA binding (Birkenbihl, Jach, Saedler, & Huijser, 2005). Site-directed mutagenesis was used to substitute conserved lysine and arginine residues in the NLS with leucine to disrupt this positive charge (Supporting information Figure S4 Table S2).
Two of these genes, RPP4 and RRS1, are known lesion-mimic mutants in Arabidopsis (Huang, Li, Bao, Zhang, & Yang, 2010;Noutoshi et al., 2005). These data suggest a functional NLS within the SBP domain is

| SlySBP12a is anchored to the ER membrane by a C-terminal transmembrane domain
While YFP-SlySBP8b was found to be strictly nuclear localized, YFP-SlySBP12a was also localized to diffuse pockets outside of the nucleus (Figure 3). The presence of a putative C-terminal transmembrane domain (TMD) in SlySBP12a (Supporting information Figure S2 This ER marker consists of the fluorescent protein mCherry with a signal peptide at its N-terminus and an ER retention motif at its C-terminus (Nelson, Cai, & Nebenfuhr, 2007). We were also able to show colocalization between the full-length YFP-SlySBP12a construct and the ER marker in tomato protoplasts (Figure 8a). Intensity histograms were generated, and Pearson's R values and Costes

| ROS production and fungal growth in leaves overexpressing SlySBP8b and SlySBP12a
Reactive oxygen species (ROS) are important cell death intermediaries, and their accumulation is a key feature of cell death imposed by necrotrophic fungal pathogens and the death-inducing toxins they produce (Heller & Tudzynski, 2011;Kim, Min, & Dickman, 2008;Sakamoto, Tada, Nakayashiki, Tosa, & Mayama, 2005;Shi et al., 2007). Transgenic SfIAP plants were reported to accumulate lower levels of ROS under stress conditions compared to wild-type plants . Necrotrophic fungal pathogens are known to exploit host ROS production as means to kill host cells for their own benefit (Govrin & Levine, 2000;Heller & Tudzynski, 2011;Kabbage et al., 2013;Ranjan et al., 2017). In addition to reduced ROS accumulation, transgenic SfIAP plants are also resistant to the necrotrophic fungal pathogen A. alternata . Therefore, we reasoned that  (Figure 10a and b). This effect was more  Figure S6). Lesion areas were measured using ImageJ software (Schindelin et al., 2012). Overall, we show that these two transcription factors are able to increase ROS levels and promote A. alternata growth, phenotypes that are dampened in plants expressing SfIAP.

| DISCUSSION
Over 15 years ago, it was first reported that heterologous expression of a viral IAP (OpIAP) in tobacco suppressed cell death induced by the necrotrophic fungal pathogen S. sclerotiorum and the necrosisinducing viral pathogen tomato spotted wilt virus (Dickman et al., 2001). Subsequent studies revealed that an IAP from Spodoptera frugiperda (SfIAP) suppressed cell death imposed by numerous abiotic and biotic stresses (Hoang, Williams, Khanna, Dale, & Mundree, 2014;Kabbage et al., 2010;Li et al., 2010). However, the biochemical mechanism by which these IAPs suppress cell death in plant systems remains unknown. In this study, we utilize SfIAP as a in the cellular response to this mycotoxin (Stone, Liang, Nekl, & Stiers, 2005). Tolerance to FB1 is a phenotype that we have also observed in SfIAP-overexpressing tomato seedlings . Interestingly, AtSPL14 and SlySBP12a both reside in clade-II and display similar structural characteristics with large SBP proteins that contain a predicted C-terminal transmembrane domain (Preston & Hileman, 2013).
Another clade-II member, GmSPL12l from soybean, was shown to be a target of the Phakopsora pachyrhizi (Asian soybean rust) effector PpEC23 (Qi et al., 2016). This effector suppressed the hypersensitive response (HR) in soybean and tobacco and also interacted with other clade-II members from N. benthamiana and Arabidopsis: NbSPL1 and AtSPL1 (Qi et al., 2016). In another study, the N immune receptor of  . The results of four randomized and blind experiments clearly show that while the contribution of SlySBP8b or SlySBP12a overexpression to A. alternata lesion areas was small, it was significantly greater than leaves expressing the negative control 35S:YFP ( Figure 10). The small differences in growth could be explained by the fact that A. alternata is already an aggressive pathogen and the benefits of priming its host for death would be small. To test this, we also treated leaves with FB1, which is a structural analog of the AAL toxin produced by A. alternata f. sp. lycopersici that induces cell death in tomato (Mirocha et al., 1992). Pre-treatment of N. glutinosa leaves with FB1 led to enhanced growth of A. alternata comparable to SlySBP12a overexpression ( Figure 10). These results provide further evidence that SlySBP8b and SlySBP12a are positive regulators of cell death, which in this case, contribute to pathogenic development of A. alternata.

| SlySBP8b and SlySBP12a require a functional SBP domain for cell death induction
As members of a transcription factor family, we hypothesize that SlySBP8b and SlySBP12a exert their prodeath activity through the regulation of genes involved in cell death. The NLS of SBP transcription factors serves a dual role in nuclear import and DNA binding, making this an essential motif for SBP function (Birkenbihl et al., 2005). We show that these transcription factors are clearly localized To identify genes involved in cell death that may be regulated by SBP transcription factors, we searched the Arabidopsis genome for promoters that contain the SBP cis-element (Supporting information Figure S5). We identified 523 genes involved in a diverse array of biological processes, which is expected of a large transcription factor family known to be involved in diverse developmental and stressrelated processes. Further investigation of these genes revealed a subset with known roles in stress responses (Supporting information  Hsu et al., 2013;Lai, Vinod, Zheng, Fan, & Chen, 2008). Perhaps the most interesting finding is that two genes, RPP4 and RRS1, are known lesion-mimic mutants (Huang et al., 2010;Noutoshi et al., 2005). Due to the large number of genes with predicted SBP-binding sites in their promoters, future studies will need to utilize chromatin immunoprecipitation sequencing (ChIP-Seq) to determine genes regulated by SlySBP8b and SlySBP12a in vivo. These studies will provide more concrete information on the downstream components responsible for cell death execution in plants.

| SlySBP12a localizes to the ER
Unlike SlySBP8b, which we found to be strictly nuclear localized, SlySBP12a was also present outside of the nucleus (Figures 3 and   6). By fusing the putative C-terminal TMD of SlySBP12a to YFP, we were able to show that the TMD of SlySBP12a localized YFP around the nucleus and at the periphery of N. benthamiana epidermal cells ( Figure 6). We hypothesized that this pattern was due to ER localization. This was confirmed in tomato protoplasts, where both YFP-SlySBP12a and YFP-TMD SlySBP12a colocalize with the ER marker SP-mCherry-HDEL (Figure 8a and b).
In response to environmental stress, plant cells increase production of secreted proteins, which in turn can cause ER stress due to the sudden influx of proteins that must be properly folded before  (Figures 3, 6, and 7).
With our data and previous studies of ER-MTTFs, we can speculate that SlySBP12a is cleaved from the ER membrane upon stress perception and translocates to the nucleus where it regulates genes involved in cell death. However, we must keep in mind that our experiments were performed with a cDNA copy of SlySBP12a, preventing the detection of splice isoforms that could lack the TMD. This is important to consider as bZIP60 was originally thought to be proteolytically cleaved from the ER membrane upon stress induced by tunicamycin treatment (Iwata, Fedoroff, & Koizumi, 2008). A follow-up study by the same group showed that in addition to being proteolytically cleaved, bZIP60 is also alternatively spliced in response to tunicamycin treatment, resulting in a truncated protein lacking the C-terminal TMD (Naga- confers enhanced stress tolerance, the animal-derived nature of these genes will likely prevent their broad commercial use. Thus, the identification of endogenous plant cell death regulators, such as SBP transcription factors, that can be targeted to ameliorate stress tolerance is appealing. This is exemplified by recent interest in exploiting SBP genes for crop improvement due to the many developmental traits they regulate (Liu, Harberd Nicholas, & Fu, 2016;Wang & Wang, 2015). Efforts are underway in our laboratory to determine whether the disruption of these transcription factors impact tolerance to a range of abiotic and biotic insults.

ACKNOWLEDGMENTS
We would like to thank Andrew Bent and members of his laboratory for their valuable insight and shared laboratory resources. The Alternaria alternata isolate used was provided by Shunping Ding of performed and designed experiments, analyzed data, and wrote the manuscript. All authors reviewed the manuscript.