A Role of the FUZZY ONIONS LIKE Gene in Regulating Cell Death and Defense in Arabidopsis

Programmed cell death (PCD) is critical for development and responses to environmental stimuli in many organisms. FUZZY ONIONS (FZO) proteins in yeast, flies, and mammals are known to affect mitochondrial fusion and function. Arabidopsis FZO-LIKE (FZL) was shown as a chloroplast protein that regulates chloroplast morphology and cell death. We cloned the FZL gene based on the lesion mimic phenotype conferred by an fzl mutation. Here we provide evidence to support that FZL has evolved new function different from its homologs from other organisms. We found that fzl mutants showed enhanced disease resistance to the bacterial pathogen Pseudomonas syringae and the oomycete pathogen Hyaloperonospora arabidopsidis. Besides altered chloroplast morphology and cell death, fzl showed the activation of reactive oxygen species (ROS) and autophagy pathways. FZL and the defense signaling molecule salicylic acid form a negative feedback loop in defense and cell death control. FZL did not complement the yeast strain lacking the FZO1 gene. Together these data suggest that the Arabidopsis FZL gene is a negative regulator of cell death and disease resistance, possibly through regulating ROS and autophagy pathways in the chloroplast.

dysfunction [12][13][14][15][16][17][18] . An Arabidopsis fzl mutation was previously shown to confer cell death and altered chloroplast morphology, associated with higher accumulation of some defense related molecules 19 . However, it is unknown how the Arabidopsis FZL gene affects cell death and whether it regulates mitochondrial function and plant resistance to pathogens.
We identified a lesion mimic mutant in the background of a transposon insertional mutant for the phosphate transporter gene PHT4;1 20 . Positional cloning revealed that the lesion mimic phenotype was not associated with the pht4;1 mutation, but was due to a mutation in the FZL gene. Besides its known role in affecting cell death and chloroplast morphology, we report here new function of the FZL gene in regulating plant defense and autophagy. We found that fzl mutations conferred enhanced disease resistance to the bacterial pathogen Pseudomonas syringae and the oomycete pathogen Hyaloperonospora arabidopsidis (Hpa). fzl-conferred cell death was associated with the activation of the autophagy and ROS pathways and was dependent on the key defense signaling mediated by SA. Our data also support a reciprocal negative regulation between FZL and SA signaling. Although being the only close homolog of Arabidopsis to the mitochondria-localized FZO proteins from yeast, flies, and mammals, a functional FZL-YFP chimeric protein was localized the chloroplast outer membrane but not the mitochondria of Arabidopsis cells. Thus, Arabidopsis FZL gene may have evolved new function different from its other homologs. Consistent with this idea, we found that neither FZL nor an FZL variant rescued the yeast strain lacking the FZO1 gene. Together, our study suggests that the FZL gene is a negative regulator of PCD and disease resistance to biotrophic pathogens in Arabidopsis, possibly acting through the autophagy pathway.

Results
FZL is a negative regulator of disease resistance to biotrophic pathogens in Arabidopsis. We isolated a lesion mimic mutant in the background of the phosphate transporter mutant called pht4;1-2 (GT_5_110509; a Landsberg erecta allele) 20 . The lesion mimic phenotype progressed more severely as leaves became more senescent (Fig. S1A). This phenotype was not co-segregated with the transposon insertion in PHT4;1 and kanamycin resistance. To identify the responsible gene, we crossed this mutant to Columbia-0 (Col-0) and conducted map-based cloning with 1108 recombinant F 2 lines to narrow down the mutation to a 50 kb region on chromosome 1. Further sequencing of the coding fragments in the region identified a one base pair substitution (G > A) in the FZL gene (At1g03160), which encodes a protein in the dynamin superfamily with demonstrated roles in regulating chloroplast morphology and cell death 19,21 . The mutation, previously called fzl-Ler, disrupts the 5′ end exon-intron junction of intron IV ( Fig. 1A and ref. 19).
Using a chimeric construct FZL-YFP, created by fusing the FZL cDNA translationally at the 3′ end with the YFP reporter under the control of the CAMV 35S promoter, we were able to complement phenotypes conferred by fzl-Ler, including cell death, high SA accumulation, and high expression of SA marker gene PR1 (Fig. 1B,C and S1B-S1D). To test if FZL regulates disease resistance, we infected fzl-Ler with Pseudomonas syringae pv. tomato strain DC3000 (DC3000). The mutant showed more resistance than Ler, which was rescued by the FZL-YFP transgene (Fig. 1D).
We obtained two additional mutant alleles, fzl-2 (SALK_118335) and fzl-3 (SALK_152584) in Columbia-0 (Col-0) background (Fig. 1A). Compared with Col-0 leaves, leaves of fzl-2 and fzl-3 mutants appeared to be paler but did not show obvious cell death (Fig. 2). The fzl-2 and fzl-3 alleles showed either higher SA accumulation nor enhanced resistance to P. syringae infection, compared with Col-0 (data not shown). However, when infected with the oomycete pathogen Hyaloperonospora arabidopsidis (Hpa) isolate Noco2, these mutants demonstrated higher resistance than Col-0 ( Fig. 1E). Ler and fzl-Ler plants cannot be tested in this experiment because the Ler background confers hyper-resistance to Hpa Noco2. Thus these data suggest that FZL is a negative regulator of disease resistance to P. syringae and Hpa strains and cell death.
Consistent with the defense role of FZL, we found that expression of FZL was suppressed by P. syringae infection (Fig. 3A). To further test if FZL expression can be affected by SA, we treated plants with an SA analog, benzo (1, 2, 3) thiadiazol-7-carothioic acid (BTH), which was shown to activate similar defense responses in plants as SA but without the toxicity caused by SA [22][23][24] . Indeed, expression of FZL was lower in BTH treated plants in a dosage-dependent manner (Fig. 3B). We also observed a lower expression of FZL in the accelerated cell death 6-1 (acd6-1) mutant that shows constitutive defense and high SA levels ( Fig. 3B and refs 25,26). Thus these results suggest a negative feedback regulation of FZL by SA.
The cell death phenotype of fzl mutants can be enhanced by the Ler background. We were intrigued by the effects of different genetic backgrounds on cell death formation in the fzl mutants. We confirmed that the two Col-0 fzl mutants were allelic to fzl-Ler by crossing fzl-Ler with fzl-2 or fzl-3. The resulting F 1 plants demonstrated strong cell death in leaves as fzl-Ler (data not shown). Such phenotypic discrepancy in the fzl alleles suggests genetic background could affect cell death formation mediated by FZL in Arabidopsis. To test this possibility, we crossed fzl-2 and fzl-3 with Ler and found that fzl-2-Ler and fzl-3-Ler homozygous plants showed more severe cell death than the original mutants in Col-0 background (Fig. 2).
fzl-conferred cell death phenotype is regulated by SA. Some lesion mimic mutants showed SA-dependent cell death 9 . To test if fzl-conferred cell death phenotype is also SA-dependent, we conducted a mutant analysis. Given the complexity of genetic background on the manifestation of fzl-conferred cell death (Fig. 2), it is important to use the same genetic background in the mutant analysis. While most SA mutants are in Col-0 background 20,25,27,28 , the eds1-2 mutant is a Ler allele and is impaired in major SA accumulation 29 . Thus we crossed fzl-Ler with eds1-2 and found that the double mutant fzl-Ler eds1-2 had much suppressed but not completely abolished cell death and restored wt-levels of SA, compared with fzl-Ler (Fig. S2). Our data are largely consistent with the results from a previous study 19 , supporting a major role of SA in regulating cell death in fzl-Ler. To further test the role of SA in regulating cell death in fzl mutants, we used a single-cell system. We treated protoplasts of Ler, fzl-Ler and a complementation line (FZL-YFP + fzl-Ler #1) with BTH. Protoplast survival rate was recorded by using fluorescein diacetate (FDA) staining. We found that mock treatment did not induce much difference in cell death in the protoplasts. Upon BTH treatment, fzl-Ler protoplasts showed much faster and more cell death (Fig. 4). By 1 hr of BTH treatment, all fzl-Ler cells were dead while most Ler and FZL-YFP/fzl-Ler cells remained alive. Similarly, protoplasts of the two Col-0 FZL alleles, fzl-2 and fzl-3, also showed enhanced cell death with BTH treatment, compared with Col-0 control (Fig. S3). Since BTH treatment of wild type (wt) protoplasts did not activate cell death, these results suggest that SA signaling works together with FZL to control cell death.

fzl-Ler shows increased autophagosomes and ROS accumulation besides altered chloroplast morphology.
Previous studies showed that fzl mutations caused abnormal chloroplast morphology 19,21 . We confirmed this phenotype with transmission electron microscopy (TEM), using the fourth to sixth leaves of 21-day old plants. The chloroplast of fzl-Ler cells showed reduced thylakoid stacks but more elongated thylakoid grana, compared with that of Ler cells ( Fig. 5 and S4). The average number of chloroplasts in fzl-Ler cells is 3.35 ± 0.16 (standard error of the mean; n = 78), which is significantly lower than that in Ler cells (4.76 ± 0.40; n = 60). No major difference was observed in the mitochondrial morphology.
Interestingly, we noticed the accumulation of vesicles in the cytoplasm of fzl-Ler cells, some of which were adjacent to the chloroplast in a cell ( Fig. 5D-F). These vesicles have a double membrane surrounding the vesicle content, resembling autophagosomes that represent the ancient vesicle mechanism to engulf and deliver cytoplasm content for degradation 3,30 . Leaf cells of Ler and fzl-Ler/FZL-YFP plants did not have obvious autophagosome accumulation (Fig. 5A,B and data not shown). To further confirm the activation of the autophagy pathway in fzl-Ler, we measured expression of several ATG genes (ATG5, 6, 7, 8c, 8f, 8i, 9, and 10) that are important for the formation of autophagosomes. We found that expression of ATG7, 8C, 8i, and 9 in fzl-Ler was at least two-fold higher than that in Ler and fzl-Ler/FZL-YFP plants (Fig. 6). Thus cell death formation in fzl-Ler likely involves the activation of autophagy.
Oxidative bursts lead to the production of reactive oxygen species (ROS), which are important for activation of defense and cell death [31][32][33] . Autophagy is tightly connected with ROS production by removing damaged proteins due to oxidation and thus affects cell survival 34,35 . To see how FZL could affect ROS accumulation and localization, we performed histochemical staining using cerium chloride (CeCl 3 ), which reacts with H 2 O 2 to produce electron-dense precipitates of ceriumperhydroxide that can be visualized using TEM 36,37 . The fourth to sixth leaves of fzl-Ler and Ler plants were fixed in the presence of cerium chloride followed by embedding and sectioning. Analysis of ultra-thin sections of the embedded samples showed abundant electron-dense cerium deposits, an indicator of H 2 O 2 accumulation, mainly on the cell wall of the fzl-Ler mutant ( Fig. 5C and S4). Minor cerium deposits were seen in the chloroplast, cytoplasm, and autophagosome membrane adjacent to the chloroplast ( Fig. 5C and F, arrows). However, Ler and fzl-Ler/FZL-YFP plants generally lack such dark deposits (Fig. 5A,B and data not shown). These results indicate that fzl-Ler-conferred cell death is associated with increased ROS production and autophagosome formation.
FZL is a chloroplast protein that functions differently from its yeast homolog. FZO1 proteins in yeasts, flies, and mammals are known to localize to the mitochondria and affect mitochondrial fusion 38 . Although being the only close homolog of FZO1 proteins, Arabidopsis FZL was localized to the chloroplast and affected chloroplast morphology. The predicted FZL protein has a potential mitochondrial targeting sequence. Whether FZL could also localize to the mitochondria and regulate mitochondrial function has not been explicitly ruled out. To address this question, we used a single cell system, utilization protoplasts from plants expressing a functional FZL-YFP under the control of the constitutive promoter CAMV 35 S ( Fig. 1 and S1B-1D). We first confirmed chloroplast-localization of FZL-YFP, which resided on the outer chloroplast membrane (Fig. S5A and B and 21 ). To test if the FZL-YFP protein also resides in the mitochondria in Arabidopsis, we stained protoplasts expressing FZL-YFP with the mitochondria-specific dye MitoTracker Red CMXRos. However, we did not observe a co-localization of YFP and MitoTracker red signals (Fig. S5C). Similarly, co-expression of a mitochondrial marker gene tagged with the red fluorescent reporter mCherry 39 in FZL-YFP protoplasts did not reveal a co-localization of mCherry with FZL-YFP (Fig. S6). Thus FZL-YFP is unlikely localized to the mitochondria.
Although not detected in the mitochondria of Arabidopsis cells, it is still possible that the Arabidopsis FZL gene shares conserved function with its yeast homolog FZO1 because the two proteins share significant similarity with 23% identity [16][17][18] . To test this, we expressed the full-length FZL cDNA and a truncated version with a deletion of the DNA fragment encoding the chloroplast transient peptide (FZL-ΔCTP) in yeast strains with or without the FZO1 gene. These yeast strains also expressed mitochondria-targeted GFP, mtGFP 40 . GFP fluorescence images showed typical wt-like tubular mitochondria in the strains expressing FZO1 (Fig. 7A left). However, the lack of FZO1 led to fragmented mitochondria (Fig. 7A right). Expression of Arabidopsis FZL or FZL-ΔCTP failed to rescue the mitochondrial defect in the FZO1-deletion yeast strain.
Yeast cells lacking FZO1 eventually lose mitochondrial DNA and can only grow on the fermentable YPD medium (containing glucose) but not on the non-fermentable YPG medium (containing glycerol) ( Fig. 7B and ref. 17). Consistent with results shown in Fig. 7A, only the strains expressing FZO1 but not the FZO1-lacking strains that expressed Arabidopsis FZL or FZL-ΔCTP, were able to grow on YPG (Fig. 7B). Together these data further support that the Arabidopsis FZL gene has evolved new function different from its yeast homolog FZO1.

Discussion
The FZL gene was previously reported to affect chloroplast morphology and cell death in Arabidopsis 19,21 . We identified an fzl mutant based on its lesion mimic phenotype. Further characterization of fzl mutants showed that besides altered chloroplast morphology and cell death, disruption in the FZL gene led to more resistance to both bacterial and oomycete pathogens, the activation of autophagy and ROS pathways. Cell death conferred by fzl is SA-dependent and can be further exacerbated by SA. While the lack of FZL led to increased SA accumulation, more SA could in turn suppress expression of FZL. Protein co-localization study and yeast complementation test suggest that the Arabidopsis FZL gene may have evolved new function different from its homologs in yeast, flies,  The yeast FZO1 is the founding member of FZO-like genes and its protein product was shown to localize to the mitochondria and affect mitochondrial function [16][17][18] . The mitochondria are dynamic organelles with constant movement inside of the cell along with frequent fusion and fission. Such dynamic behavior of the mitochondria is critical for mitochondrial function in many biochemical reactions, energy production, and cellular respiration 38 . The yeast fzo1 mutant showed highly fragmented mitochondria and instable mitochondrial DNA. Later studies with disrupted FZO homologs in flies, worms and mammals showed similar mitochondrial fragmentation [12][13][14][15]18 . These studies indicate that FZO genes regulate mitochondrial fusion, disruption of which could lead to disease in some organisms [41][42][43][44][45] . FZL is the only Arabidopsis gene that shares significant homology to the FZO genes in yeast, flies, and mammals. It was originally hypothesized to function similarly to these other FZO genes in regulating mitochondrial function. However two previous studies 19,21 and our data presented here support new function of the Arabidopsis FZL gene in the chloroplast. First, while mutations in FZL led to cell death in Arabidopsis, the yeast fzo1 mutant did not show a cell death phenotype (this study and refs 19,46). Second, it is the morphology of the chloroplast, not the mitochondria, that is altered in fzl-Ler when compared with Ler (Fig. 5, S4, and refs  19,21). Third, the FZL protein was localized to the chloroplast but not in the mitochondria of Arabidopsis cells (Figs S5 and S6, and ref. 21). Fourth, Arabidopsis FZL or an FZL variant did not rescue the FZO-lacking yeast strain (Fig. 7). Thus the Arabidopsis FZL gene likely functions differently from its homologs in yeast, flies, and mammals and Arabidopsis uses an FZL-independent mechanism in regulating mitochondrial fusion.
The function of FZL is closely related to the chloroplast. Besides its critical role in photosynthesis, the chloroplast has been shown as the primary source of many important defense molecules, such as SA biosynthesis, production of ROS and some secondary compounds 47,48 . Like the mitochondria, the chloroplast is also an important player in PCD 49 . Mutations in several genes affect chloroplast-derived metabolites and confer the lesion mimic phenotype. Examples of such mutants include acd1 (impaired in pheophorbide an oxygenase) 50 , acd2 (impaired in red chlorophyll catabolite reductase) 44,51 , flu (impaired in a protein that is a part of a complex inhibiting tetrapyrrole synthesis) 52,53 , and the maize mutant les22 (impaired in uroporphyrinogen decarboxylase) 54 . FZL is an outer-membrane chloroplast protein (Figs S5 and S6, and ref. 21) and likely functions differently from these chloroplast metabolic proteins in affecting PCD. The lack of FZL could cause the change in chloroplast morphology and subsequently affect the function of the chloroplast. Such changes in the chloroplast can activate stress signals, leading to the production of ROS, which in turn could cause the accumulation of damaged proteins and result in cell death. Consistent with the role of autophagy in removing damaged proteins and regulating PCD, we observed increased autophagosomes using high-resolution TEM images and enhanced expression of autophagic genes in the fzl mutant exhibiting cell death (Figs 5 and 6). Thus our data support the regulation of autophay and ROS pathways by FZL. Further detailed morphological analysis of autophagosomes in fzl-Ler and genetic analysis of fzl mutants and other mutants impaired in autophagy and ROS pathways could contribute to a better understanding of FZL function.
Another factor affecting FZL function is SA as demonstrated by these supporting data: (1) fzl-Ler accumulates higher SA levels, associating with more cell death and disease resistance, compared to Ler ( Fig. 1 and S1); (2) The high SA level in fzl-Ler is suppressed by a normal FZL gene and by the eds1-2 mutation. Such suppression in SA accumulation is associated with reduced cell death and defense (Fig. 1, S1, and S2); (3) SA treatment activates cell death in fzl mutants but not in wt alone ( Fig. 4 and S3) and (4) expression of FZL is suppressed by high SA levels and defense activation (Fig. 3). Together these observations suggest a negative feedback loop formed between FZL and SA in regulating cell death and defense. The FZL-SA signaling loop likely also involves the ROS and autophagy pathways. To support this notion, ROS bursts are known to trigger SA production and signaling and in turn SA signal activation can induce more ROS bursts 33 . A similar interplay was found with autophagy and SA 3,11,55,56 . Thus FZL may normally function be to inhibit the activation of the signaling cascade involving ROS, SA and autophagy. In the absence of FZL, the activation of this signaling event leads to degradation of the chloroplast (the source of some pro-death signals) and eventually cell death. Indeed, we observed reduced numbers of chloroplasts in addition to cell death in fzl-Ler. While we know disrupting SA could suppress fzl-conferred phenotypes, it would be interesting to further investigate how disrupting ROS and autophagy pathways could interfere with fzl-conferred phenotypes in future studies.
Besides ROS, SA, and autophagy, fzl-conferred cell death can be affected by other factors. The fzl-2 and fzl-3 alleles in Col-0 background demonstrated minor defense and no cell death while the fzl-ler allele in Ler background showed strong cell death and defense phenotypes (Figs S1 and 2). We showed that the Ler background can enhance cell death in fzl-2 and fzl-3 mutants. Consistent with the influence of genetic background on fzl-conferred cell death, we observed residual cell death when the SA mutant eds1-2 in Ler background was crossed into fzl-ler while another study showed a complete suppression of cell death in fzl-Ler by a different eds1 mutant (Fig. S2 and  ref. 19). Such discrepancy could be due to genetic background and/or the growth environment in different laboratories. Developmental stages could also affect fzl-conferred cell death (Fig. S1A). Together these observations suggest that additional molecules existing under certain conditions could affect FZL function.
The roles of ROS and SA in plant defense against pathogens have been relatively well understood. Pathogen infection and activation of defense signaling are known to induce autophagy [5][6][7] . However how autophagy regulates plant innate immunity remains unclear. While they all had much reduced autophagosomes and increased cell death, atg mutants showed opposing effects on disease resistance. One atg mutant, atg6, was shown to be more susceptible to biotrophic pathogens 57,58 . Consistent with atg6-conferred disease susceptibility, studies on fzl and the mutant defective in glyceraldehyde-3-phosphate dehydrogenase (GAPDH) showed an association of constitutive autophagy with enhanced disease resistance against biotrophic pathogens. On the other hand a group of atg mutants (i.e. atg2, 5, 10, and 18a) was shown to be more resistant to biotrophic pathogens 56,59-61 . Plants usually respond to biotrophic and necrotrophic pathogens with opposing defense phenotypes 62 . Consistent with this idea, these atg mutants were more susceptible to necrotrophic pathogens. Interestingly another group of atg mutants (i.e. atg7 and 9) was shown to be more susceptible to both biotrophic and necrotrophic pathogens. It appears that there is a lack of consistent association between the number of autophagosome formation and plant defense. These seemly controversial data could suggest that different ATG genes are differentially required by pathogens of different lifestyles and the severity of cell death is not always coupled with the level of disease resistance in plants. Thus the mechanism of autophagy signaling in defense, in particular its connection with FZL function, still remains to be elucidated.
Taken together the fzl-ler mutant exhibits hallmarks of lesion mimic mutant phenotypes, including cell death, increased accumulation of defense-related molecules (i.e. ROS, SA and defense gene transcripts). We report that the activation of the autophagy pathway in this mutant is a possible mechanism leading to cell death in the plant. Data from this and other studies clearly support that the FZL gene of Arabidopsis has evolved new function different from its homologs in yeast, flies, and mammals. This new function lies in the regulation of chloroplast morphology and function, activation the signaling cascade involving ROS, SA and autophagy. The FZL-regulated processes ultimately affect plant growth, development, and response to pathogen attacks. Although we are still far from a complete understanding of the molecular mechanism by which FZL regulates innate immunity, FZL and its related genes can be used as excellent resources to uncover mechanisms of PCD and disease resistance. These genes can also be used as powerful tools to manipulate plant defense response in order to achieve a broad spectrum of disease resistance in plants. The fzl-Ler and eds1-2 mutants were previously described 19,29 . The fzl-2 (SALK_118335) and fzl-3 (SALK_152584) mutants were obtained from the Arabidopsis Biological Resource Center.

Methods
Protoplast Analyses. Arabidopsis protoplasts were isolated from leaves of 21-day old plants following the tape sandwich technique described previously 63 with few modifications. Briefly, the lower epidermal surface cell layer was peeled away from leaves using plastic tape. Fifteen peeled leaves were transferred to a Petri dish containing 10 ml of enzyme solution (20 mM MES (pH 5.7), 1% (w/v) cellulase R10, 0.25% (w/v) macerozyme R10, 0.4 M mannitol, 20 mM KCl, 10 mM CaCl 2 , 0.1% (w/v) BSA). Leaves were gently agitated in the dark for 60 to 120 min till the protoplasts were completely released into the solution. Protoplasts were centrifuged three minutes at 100 × g, washed twice with 10 ml of pre-chilled modified W5 solution (2 mM MES (pH 5.7)), 154 mM NaCl, 125 mM CaCl 2 , 5 mM KCl, 5 mM glucose, and incubated on ice for 30 min. Protoplasts were then centrifuged and resuspended to a final concentration of 5 × 10 5 cells/ml in modified MMG solution (4 mM MES (pH 5.7), 0.4 M mannitol, 15 mM MgCl 2 ) for further experiments.
For co-localization studies, protoplasts expressing FZL-YFP were stained with the mitochondria-specific dye MitoTracker Red CMXRos (Molecular Probes) or transfected with a mitochondrial marker gene tagged with the red fluorescent reporter gene mCherry 39 . Images were captured using a confocal laser scanning microscope (Leica TCS SP2 AOBS) and analyzed using the Imaris software (version 7.0.0). Complementation Tests. The complementation construct for fzl-Ler was made by fusing FZL cDNA translationally at the 3′ end with the YPF reporter and the chimeric gene was expressed under the control of the CAMV 35 S promoter. This construct, named FZL-YFP, was used to transform fzl-Ler. At least 15 independently transformed lines were obtained and they all showed abolished cell death phenotype. Primers used for making the construct were listed in Table S1.
In order to complement the yeast FZO1-lacking mutant, we cloned the full-length FZL cDNA and a truncated version with a deletion of the sequence encoding the chloroplast transient peptide (FZL-ΔCTP) in the pYX122 vector under the control of the constitutive TPI promoter. These constructs and the empty vector were introduced into a yeast strain that has the chromosomal copy of the FZO1 gene deleted and carries a copy of the FZO1 gene on a plasmid with a URA3 marker. Upon growth on the 5-FOA medium, the URA3 plasmid was counter-selected against and thereby the FZO1 gene was eliminated in yeast cells, thus creating FZO1-lacking yeast mutant strains. Primers used for making yeast-complementation constructs were listed in Table S1.
Pathogen Infection. Bacterial culture and preparation of P. syringae strains were conducted as described 64 .
The fourth to sixth leaves of 21-day old plants were infiltrated with P. syringae-containing solution, using a 1 ml needleless syringe. Infected leaf discs were collected three days later for bacterial growth assay. The infection with Hyaloperonospora arabidopsidis isolate Noco 2 was conducted with 7-d old seedlings as previously described 64 . Cell Death Analyses. The fourth to sixth leaves from each genotype were stained with trypan blue for cell death visualization, according to a published protocol 27 . Photographs of the stained leaves were taken with a CCD camera (cool Snap HQ 2 , Photometrics, USA) connected to a dissecting microscope (Leica M205 FA, Leica Microsystems, Germany). At least three leaves from different plants of each genotype were stained and examined for cell death.
To assess cell death in protoplasts, 80 μ l of protoplasts at 5 × 10 5 /ml for each genotype were treated with 100 μ M of benzo (1,2,3)-thiadiazole-7-carbothioic acid (BTH, a kind gift from Robert Dietrich, Syngenta) or water (mock). Three replicates for each genotype were used in each experiment. Protoplasts were collected at the indicated times and were stained with fluorescein diacetate (FDA) (Sigma-Aldrich Co. LLC, St. Louis MO) for cell viability test, using 2 μ l of 5 mg/ml FDA per sample for 2 min in darkness. Living cells gave out green fluorescence that was detected with fluorescence microscopy. The cell survival rate was calculated based on the ratio between living cells and the total number of protoplasts in a sample. RNA Analysis. The fourth to sixth leaves of 21-day old plants were harvested for RNA extraction followed by qRT-PCR analysis as previously described 65 . Primers used for qRT-PCR were listed in Table S1. SA Measurement. Free and total SA (glucosylated SA) were extracted from 21-day old plants and quantified with a high-performance liquid chromatography (HPLC) instrument as previously described 25,28 .