The NIN-Like Protein OsNLP2 Negatively Regulates Ferroptotic Cell Death and Immune Responses to Magnaporthe oryzae in Rice

Nodule inception (NIN)-like proteins (NLPs) have a central role in nitrate signaling to mediate plant growth and development. Here, we report that OsNLP2 negatively regulates ferroptotic cell death and immune responses in rice during Magnaporthe oryzae infection. OsNLP2 was localized to the plant cell nucleus, suggesting that it acts as a transcription factor. OsNLP2 expression was involved in susceptible disease development. ΔOsnlp2 knockout mutants exhibited reactive oxygen species (ROS) and iron-dependent ferroptotic hypersensitive response (HR) cell death in response to M. oryzae. Treatments with the iron chelator deferoxamine, lipid-ROS scavenger ferrostatin-1, actin polymerization inhibitor cytochalasin A, and NADPH oxidase inhibitor diphenyleneiodonium suppressed the accumulation of ROS and ferric ions, lipid peroxidation, and HR cell death, which ultimately led to successful M. oryzae colonization in ΔOsnlp2 mutants. The loss-of-function of OsNLP2 triggered the expression of defense-related genes including OsPBZ1, OsPIP-3A, OsWRKY104, and OsRbohB in ΔOsnlp2 mutants. ΔOsnlp2 mutants exhibited broad-spectrum, nonspecific resistance to diverse M. oryzae strains. These combined results suggest that OsNLP2 acts as a negative regulator of ferroptotic HR cell death and defense responses in rice, and may be a valuable gene source for molecular breeding of rice with broad-spectrum resistance to blast disease.


Introduction
Plants have evolved a sophisticated immune system to overcome attacks by microbial pathogens [1,2]. Microbial pathogens have evolved pathogen effectors to promote virulence and cause disease in host plants [3]. The interaction between rice (Oryza sativa L.) and Magnaporthe oryzae is a useful model system to investigate the molecular and biochemical mechanisms underlying plant immunity and disease. Avirulent M. oryzae infection causes highly resistant and immune responses in rice cells, including defense-related gene expression, reactive oxygen species (ROS) bursts, and ferroptotic cell death [4][5][6][7]. ROS act as cellular signaling molecules to trigger plant immune responses [3]. ROS accumulation suppresses M. oryzae effector BAS4 activity to prevent M. oryzae infection [8].
Ferroptosis, a nonapoptotic form of iron-dependent cell death [9], is triggered by the accumulation of ferric ion (Fe 3+ ) and toxic lipid ROS in mammalian cells [10]. Highly reactive Fe 2+ reacts with H 2 O 2 to produce Fe 3+ and highly toxic ·OH (Fenton reaction) [11,12], which results in iron-dependent accumulation of toxic lipid ROS [9]. The hydroxyl radicals produced by the Fenton reaction causes severe damage to lipids, proteins, DNA, and
Nicotiana benthamiana seeds were germinated in water for 7 days under continuous light at 25 • C. Germinated seedlings were planted in pots (8 cm diameter, 6 cm height) containing Baroker soil (Seoul Bio, Seoul, Korea). N. benthamiana plants were grown in a growth chamber under the following conditions: 25 • C, white fluorescent light (150 µmol photons m −2 s −1 ), and a 16 h day/8 h night photoperiod.

∆Osnlp2 Mutant Genotyping
T-DNA insertion ∆Osnlp2 mutant seeds were sterilized as follows: seed husks were removed, and seeds were treated with 100% ethanol for 1 min and 50% Clorox for 30 min. Sterilized seeds were cultured in one-half Murashige and Skoog (MS) (Sigma-Aldrich, St. Louis, MO, USA) medium at 25 • C under continuous light for 2 weeks. Fresh rice leaves were collected from each plant to extract genomic DNA for genotyping. The cetyltrimethylammonium bromide (CTAB) method was used to extract genomic DNA [30]. The primers were designed near the T-DNA insertion site based on information from the mutant database (http://signal.salk.edu./cgi-bin/RiceGE, accessed on 8 March 2017) to detect homozygous and heterozygous plants from the seed pool. The gene-specific primers and T-DNA rightborder primer (RB) were used to verify the T-DNA insertion in ∆Osnlp2 mutant plants. The primers were in Table S1. To germinate ∆Osnlp2 mutant seeds to grow to the whole plants, the peeled seeds of ∆Osnlp2 mutants were cultured in one-half MS media containing 30 mg·L −1 hygromycin.
Total RNA was extracted from homozygous plants using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and then used to synthesize cDNAs. OsNLP2 transcriptional levels in rice DJ and ∆Osnlp2 mutant lines were analyzed by performing RT-PCR and a real-time quantitative reverse transcription polymerase chain reaction (real-time qRT-PCR) using OsNLP2 RT primers. Rice Ubiquitin (OsUbi) transcript levels were used to normalize the transcript levels of OsNLP2. The data are presented as means ± SD of relative expression quantities of OsNLP2 in leaf sheaths from different rice plants (n = 4).

Fungal Inoculation and Infection Evaluation
To monitor disease symptoms on rice leaves, M. oryzae conidial suspensions (1.0 × 10 5 conidia mL −1 ) were uniformly sprayed on leaves of 3-week-old seedling plants.
Inoculated rice plants were incubated at 25 • C under dark and moist conditions for 24 h,  and then placed under normal conditions (16 h day/8 h night). Disease symptoms on rice leaves were monitored 5 days after inoculation.
For the rice leaf sheath tests, M. oryzae conidial suspensions (5.0 × 10 5 conidia mL −1 ) were inoculated on 6-week-old rice leaf sheaths (5 cm length). Inoculated rice sheaths were incubated in the dark for 2 days in a moistened chamber. Then, epidermal 3-4 cell thick layers were excised from the rice sheaths and monitored under a fluorescence microscope (Zeiss equipped with Axioplan 2; ZEISS, Campbell, CA, USA) as described previously [4]. Infected epidermal cells were categorized into three infection types: expanded infection (invasive hyphae colonized inside multiple cells), single-cell infection (limited hyphal growth inside a single cell), and HR cell death. Infected cells of each infection type were counted using four replicates from each of three independent experiments.

Cloning and Plasmid Construction
To detect the subcellular localization of OsNLP2 and its functional domains, entry clones were recombined into the Gateway green fluorescent protein (GFP)-containing vector pGWB552 using the Gateway LR Clonase II enzyme (Invitrogen, Carlsbad, CA, USA) [31]. The coding sequences of OsNLP2 (LOC_Os04g41850) and its functional domains were amplified from the rice cDNA library using gene-specific primers containing Gateway attB1 and attB2 sites. The gene-specific primers were designed based on genome information provided by the Rice Genome Annotation Project (http://rice.plantbiology.msu.edu/, accessed on 5 June 2016). The primers were list in Table S1.
For the transient OsNLP2 expression assay in N. benthamiana leaves, the OsNLP2 coding region and its functional domains were cloned into the ligation-independent cloning (LIC) vector pCAM2300-LIC. These coding sequences were amplified from the rice cDNA library using primers with adaptors. Amplified PCR products were subcloned into the pCAM2300-LIC vector digested with a SnaB1 restriction enzyme using T4 DNA polymerase (New England Biolabs, Ipswich, MA, USA).

Agrobacterium Transformation, OsNLP2 Subcellular Localization, and Transient OsNLP2 Expression Assay
To determine the subcellular localization of OsNLP2, OsNLP2 full sequence and functional domains in pGWB552:GFP were transformed into Agrobacterium GV3101. The empty pGWB552:GFP vector was transformed as a negative control. Agrobacteria transformed with different constructs were cultured overnight at 28 • C in Luria Bertani (LB) liquid medium containing 100 µg mL −1 spectinomycin. The Agrobacterium cells were collected and then suspended in an infiltration buffer [97.5 mg 2-(N-morpholino)ethanesulfonic acid sodium salt (MES), 0.5 mL of 1 M MgCl 2 , and 0.1 mM acetosyringone in 50 mL water]. The Agrobacterium suspensions were infiltrated into N. benthamiana leaves, incubated for 2 days, and then the infiltrated leaves were stained with 4 ,6-diamidino-2-phenylindole (DAPI) solution (1 µg mL −1 ) for 10 min in the dark to visualize nuclei. The epidermal cells were observed using a fluorescence microscope (Zeiss equipped with Axioplan 2) with green fluorescence filters (Ex/Em: 488 nm/505-550 nm wavelengths) and DAPI filters (Ex/Em: 405/421-523 nm).
To transiently express OsNLP2 in N. benthamiana leaves, full-length OsNLP2 and its functional domains in the binary vector pCAM2300-LIC were transformed into Agrobacterium GV3101. The Agrobacterium suspensions in the infiltration buffer were adjusted to a final concentration of OD 600 = 1.0. The transformed Agrobacterium suspensions containing each of the three different constructs were infiltrated into individual sites on the same N. benthamiana leaves. The empty pCAM2300-LIC vector and Infestin 1 (INF1):pCAM2300-LIC were used as negative and positive controls, respectively. Cell death symptoms were observed on infiltrated N. benthamiana leaves after incubating for 2 days.

Complementation of OsNLP2 into ∆Osnlp2 Mutant
To restore the expression of OsNLP2 in ∆osnlp2 mutant plants, the OsNLP2 coding sequence inside the binary vector pB2GW7 (Bar gene as the selection marker) was transformed into Agrobacterium C58C1 [32]. The OsNLP2-complemented plants were created using the ∆Osnlp2 mutant line #6-4. Briefly, 4-week-old rice calli of ∆Osnlp2 mutant line #6-4 were used for the 3-day co-culture with Agrobacterium C58C1 containing OsNLP2:pB2GW7. The transformed calli were selected on the media containing increasing concentrations of phosphinothricin (PPT, 3 mg/L and 6 mg/L). After shooting and rooting, the calli were cultured in the regeneration media for 2 weeks in light at 25 • C, and transferred to soil after 2-day adaption in water. OsNLP2 primers and Bar gene primers were used to confirm the OsNLP2 expression in the transformants (Table S1).

RT-PCR and Real-Time qRT-PCR Analyses
Rice gene expression levels in M. oryzae-infected leaf sheath tissues were analyzed by reverse-transcription PCR (RT-PCR) and real-time quantitative PCR (qRT-PCR). Total RNAs were extracted from rice tissues using TRIzol reagent (Invitrogen), followed by first-strand cDNA syntheses using a cDNA synthesis kit (Invitrogen) according to the manufacturer's instructions. Equal amounts of cDNAs were used as templates for RT-PCR and real-time qRT-PCR. The expression of OsNLP2, OsPBZ1, OsWRKY104, OsRbohB, OsPIP-3A, and OsWRKY90 was analyzed with gene-specific primer sets by real-time qRT-PCR using TOPreal qPCR 2× PreMIX (SYBR Green with low ROX, Enzynomics). Relative expression levels of the tested rice genes were determined by normalizing them with respect to the expression levels of rice OsUbiquitin, 18S rRNA, and OsActin as internal controls. The relative expression values were calculated using the ∆∆Ct method [33]. The data are means ± SD of relative gene expression levels in leaf sheaths from three independent experiments. The list of gene-specific primers used in this study is provided in Table S1.

ROS Detection and Quantification
Cellular ROS (H 2 O 2 ) localization in rice leaf sheaths infected with M. oryzae was visualized using the following dyes: 5-(and 6-)chloromethyl-2 ,7 -dichlorodihydrofluorescein diacetate acetyl ester (CM-H 2 DCFDA) and 3,3 -diaminobenzidine (DAB). CM-H 2 DCFDA reacts with ROS in living cells and produces 2 ,7 -dichlorofluorescein (DCF), a highly fluorescent product that can be detected using a fluorescence microscope with GF filters [4,34]. The isolated epidermal layers of rice leaf sheaths were soaked in water for 5 min at 4 • C to wash away wound-induced ROS, followed by staining in 2 µM CM-H 2 DCFDA (Molecular Probes Life Technologies, Eugene, OH) for 30 min in the dark at room temperature. Samples were washed three times with 1× phosphate-buffered saline (PBS) buffer, and stained epidermal sheath cells were observed under a fluorescence microscope (Zeiss equipped with Axioplan 2) with green fluorescence (GF) filters (Ex/Em: 488 nm/505-550 nm wavelengths). The isolated epidermal layers from infected rice sheaths were incubated in 1 mg mL −1 DAB solution (Sigma-Aldrich, St. Louis, MO, USA) for 8 h at room temperature, and then destained with ethanol:acetic acid:glycerol (3:1:1, v/v/v) as described previously [4,34,35]. ROS localization inside epidermal cells was observed under a microscope.
ROS production in rice sheath cells after inoculation with M. oryzae was measured using the chemiluminescence assay [4,16]. The isolated epidermal layers were cut into small pieces (0.5 cm) and submerged in Milli-Q water for 5 min at 4 • C to reduce woundinduced ROS. A piece of the epidermal layer was transferred into each well of a 96-well plate containing a chemiluminescence solution [30 µL luminol (Bio-Rad, Hercules, CA, USA), 1 µL horseradish peroxidase (1 mg mL −1 ; Jackson Immunoresearch, West Grove, PA), and 69 µL Milli-Q water], and incubated in the dark for 5 min at room temperature. Chemiluminescence [relative luminescent units (RLU)] of the ROS signals was detected using a GloMax 96 Microplate Luminometer (Promega, Madison, WI, USA).

Prussian Blue Staining for Ferric Ion Detection
Ferric ion (Fe 3+ ) accumulation in rice leaf sheath cells was visualized by performing Prussian blue staining [5,36]. Epidermal layers of rice sheaths were isolated and stained with a Prussian blue solution (7% potassium ferrocyanide and 2% hydrochloric acid) for 15 h [37]. The stained epidermal cells were observed under a microscope (Zeiss equipped with Axioplan 2). Ferric ferrocyanides, which combine with Fe 3+ inside cells, appear as a bright blue color in sheath epidermal cells.

Lipid Peroxidation Assay
To detect lipid peroxidation levels in rice sheaths infected with M. oryzae, we quantified malondialdehyde (MDA), a product of unsaturated fatty acid peroxidation, by reacting it with thiobarbituric acid (TBA) and measuring with a spectrophotometer [4,38]. Rice leaf sheaths were ground with liquid nitrogen, and an equal amount of the powdered tissue was mixed with the TBA solution

Treatment with Deferoxamine, Ferrostatin-1, Cytochalasin A, and Diphenyleneiodonium
The iron chelator deferoxamine (DFO) [36], potent ferroptosis inhibitor ferrostatin-1 (Fer-1) [4][5][6][7][8][9], actin filament inhibitor cytochalasin A (Cyt A) [36], and oxidase inhibitor diphenyleneiodonium (DPI) [39] were used to investigate iron-and ROS-dependent ferroptotic cell death in leaf sheaths of ∆Osnlp2 mutant and complementation plants during M. oryzae infection. DFO, Fer-1, and DPI were purchased from Sigma-Aldrich, and Cyt A was purchased from Cayman Chemical Company (Ann Arbor, MI, USA). For the iron chelator assay, the epidermal layer of rice leaf sheaths was isolated at 42 h post-inoculation (hpi) and incubated in 3 mM DFO for 6 h at room temperature. For Fer-1 treatment, the epidermal layer of rice leaf sheaths was isolated at 24 hpi and incubated in 10 µM Fer-1 for 24 h in the dark at room temperature after vacuum infiltration. For Cyt A and DPI treatments, the rice leaf sheaths (5-7 cm length) were inoculated with conidial suspension (5 × 10 5 conidia mL −1 ) in water (mock), Cyt A (20 µg mL −1 ), or DPI (5 µM). The inoculated and treated rice leaf sheaths were then incubated in the dark at 25 • C for 48 h. The middle thin epidermal layers of the infected and treated rice leaf sheaths were stained and observed under a fluorescence microscope (Zeiss equipped with Axioplan 2).

Identification of OsNLP2, ∆Osnlp2 Mutant, and OsNLP2 Complementation Plants
A truncated RWP-RK domain-containing protein was screened from the rice cDNA library and identified as a NIN-like protein (named OsNLP2) by BLAST (Basic Local Alignment Search Tool) analysis of a DNA sequence in the National Center of Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/, accessed on 20 April 2016). NIN family proteins were identified as transcription factors (TFs) in nodulation and nitrate signaling [20,21]. We aligned the amino acid sequences of rice RWP-RK family proteins (NLP subfamily and RKD subfamily proteins) with other plant NLP proteins (Arabidopsis thaliana, Medicago truncatula, and Brachypodium distachyon) ( Figure S1). The OsNLP2 used in this study had the GAF-like domain, RWP-RK domain (N567 to V615), and C-terminal PB1 domain (L835 to D915) ( Figure S1). The GAF-like domain was detected at the N-terminal region of NLP subfamily proteins [21]. The function of the GAF-like domain in NLP proteins is not yet known. The PB1 domain contains a positively charged lysine (K) and a negatively charged OPCA motif (D-x-D/E-x-D/E) ( Figure S1) [40]. Most of the aligned NLP proteins shared the conserved RWP-RK domain and PB1 domain, except for OsNLP6 that lacked the standard RWP-RK motif. OsRKD proteins lacked the PB1 domain. We constructed a phylogenetic tree using the neighbor-joining method and MEGA7 software (Version 7.0, created by Sudhir Kumar, Glen Stecher, and Koichiro Tamura, Kent, UK) ( Figure S2). NLP subfamily proteins were distinctly separated from RKD subfamily proteins in the phylogenetic tree. The OsNLP2 used in this study was phylogenetically close to BdNLP4, OsNLP5, and BdNLP3 ( Figure S2). The amino acid sequences of plant RWP-RK family proteins were aligned based on information from the Rice Genome Annotation Project  Table S2.

OsNLP2 Expression Patterns in Rice-M. oryzae Interactions
To investigate whether OsNLP2 expression is involved in rice disease and immunity, we analyzed the OsNLP2 transcription levels in leaf sheaths of DJ rice during M. oryzae infection. Infected leaf sheaths of DJ (used as wild-type rice) were sampled at early time points after inoculation with M. oryzae PO6-6 (virulent) and 007 (avirulent) strains. Transcript levels of the internal control genes OsUbiquitin were used to normalize OsNLP2 transcript levels ( Figure 1A). Invariant expression of these reference genes normalized expression levels of OsNLP2 in rice leaf sheaths during infection. Real-time qRT-PCR data showed that infection with virulent M. oryzae PO6-6 distinctly induced OsNLP2 expression in DJ leaf sheaths at early infection stages (3-48 hpi). By contrast, infection with avirulent M. oryzae 007 did not induce OsNLP2 expression by 48 hpi during incompatible rice-M. oryzae interactions. These expression patterns ( Figure 1A) indicate that OsNLP2 is involved in blast disease (susceptibility), and may act as a negative regulator to suppress rice HR cell death and defense responses to M. oryzae infection.

Subcellular Localization of OsNLP2
The plant-specific RWP-RK family of TFs are, at least partly, localized in the nucleus [21]. OsNLP2 contained the characteristic RWP-RK domain and C-terminal PB1 domain ( Figures 1B and S1), which may function in DNA-binding and protein-protein interactions, respectively [40]. The OsNLP2 coding sequence and its N-terminal domain and RWP-RK or PB1 domain regions were constructed with a GFP tag, transiently expressed in N. benthamiana leaves using the agroinfiltration method; leaf cell nuclei were counterstained with DAPI; and cells were observed using a fluorescence microscope ( Figure 1C). The control GFP construct (00:GFP) was ubiquitous in the plasma membrane and cytoplasm of N. benthamiana cells. Full-sequence OsNLP2:GFPs were localized only in the nuclei, overlapping with the DAPI fluorescence ( Figure 1C). By contrast, the N-terminal domain:GFP was localized to both the plasma membrane and nuclei. RWP-RK + PB1:GFP and PB1:GFP were localized to both the cytoplasm and nuclei. This indicates that nuclear localization of OsNLP2 requires the full-length OsNLP2 coding sequence rather than a specific N-terminal domain, RWP-RK domain, or PB1 domain. These subcellular localization patterns collectively suggest that OsNLP2 is localized in the nucleus, where it may act as a TF and bind to a specific DNA sequence.

∆OsNLP2 Mutation Triggers Hypersensitive Resistance Reactions against Different M. oryzae Strains Infection
To investigate the roles of OsNLP2 in plant disease and immunity, ∆Osnlp2 knockout mutant lines were created from the wild-type rice cultivar DJ by performing T-DNA insertion mutagenesis. T-DNA pGA2715 was inserted at −385 bp in the 5 -untranslated region (UTR) of OsNLP2 genomic DNA ( Figure S3). Two pairs of primers (OsNLP2 forward and OsNLP2 reverse; OsNLP2 forward and T-DNA pGA2715 RB) were designed to select homozygous transgenic ∆Osnlp2 plants by performing PCR assays with OsNLP2 genomic DNA ( Figure S3, Table S1). The three ∆Osnlp2 mutant lines (#4-2, #5-1, and #6-4) were confirmed as homozygous mutants by electrophoretic analysis of the PCR products ( Figure S3B). The OsNLP2 transcription levels in wild-type rice DJ and ∆Osnlp2 mutant lines were analyzed by performing RT-PCR and real-time qRT-PCR assays with an RT primer set (LP and RP). OsNLP2 expression was significantly suppressed in the three ∆Osnlp2 mutant lines ( Figure S3C). However, all these ∆Osnlp2 mutant lines grew well normally, as wild-type rice DJ did, which could be used for further experiments. The ∆Osnlp2 mutant seeds germinated well in the 1/2 MS media to grow to the whole plants ( Figure S3D).
To verify that OsNLP2 is the causal gene for ∆Osnlp2, complementation plants were created by transferring the full-length OsNLP2 coding sequence into ∆Osnlp2 mutant line #6-4 using the OsNLP2:pB2GW7 binary vector ( Figure S3A). Levels of OsNLP2 expression were normalized by invariant expression of the internal control gene OsUbiquitin. OsNLP2 was not expressed in the ∆Osnlp2 mutant line #6-4. However, real-time qRT-PCR data showed significantly high levels of OsNLP2 expression in the complementation lines #3, #11, and #12 ( Figure S3C). This indicates that OsNLP2 expression was restored in ∆Osnlp2 mutant plants.
Leaves of 3-week-old wild-type rice DJ and ∆Osnlp2 knockout mutant lines were sprayinoculated with a conidial suspension (1.0 × 10 5 conidia mL −1 ) of M. oryzae PO6-6 (virulent) and 007 (avirulent). Whole-leaf disease phenotypes were observed and photographed at 5 days after inoculation with M. oryzae (Figure 2A These combined results indicate that ∆Osnlp2 mutant lines exhibit a broad-spectrum, racenonspecific resistance to different M. oryzae strains.

ROS-and iron-dependent ferroptotic cell death is involved in the defense responses in rice cells infected with avirulent M. oryzae. To investigate the cell death triggered by
OsNLP2 knockout mutation, we analyzed ROS and ferric ion (Fe 3+ ) accumulation and lipid peroxidation in leaf sheath cells of rice DJ and ∆Osnlp2 mutants during M. oryzae PO6-6 (virulent) and race 007 (avirulent) infection. ROS accumulation was detected and visualized by CM-H 2 DCFDA and DAB staining. The ROS-sensitive CM-H 2 DCFDA dye was used to monitor ROS localization in living plant cells. DAB is oxidized by H 2 O 2 in the presence of peroxidase to generate a dark brown precipitate in plant cells [34,41]. CM-H 2 DCFDA (GF) and DAB (dark brown) staining showed that ROS (H 2 O 2 ) strongly accumulated inside and around IH in wild-type rice DJ cells infected with M. oryzae race 007 (avirulent), but not in cells infected with M. oryzae PO6-6 (virulent). However, ROS' strong accumulation could be detected in ∆Osnlp2 epidermal cells infected with both M. oryzae PO6-6 and race 007 ( Figures 3A and S5). CM-H 2 DCFDA-specific ROS-localized fluorescence was distinctly visible inside and around IH in ∆Osnlp2 mutant cells at 20-48 hpi ( Figure S6). DAB-stained epidermal cells with cellular aggregates (dark brown color) were observed at 48 hpi in the leaf sheaths of ∆Osnlp2 mutants ( Figures 3A and S5). The chemiluminescence assay showed that avirulent M. oryzae 007 infection distinctly induced ROS accumulation in both rice DJ and ∆Osnlp2 mutants to similar levels ( Figure S5B). ROS accumulation levels were significantly higher in ∆Osnlp2 mutant cells than those in rice DJ cells infected with different virulent M. oryzae strains at 48 hpi ( Figures 3B and S7).
Ferric ion (Fe 3+ ) accumulation and localization in rice epidermal cells were detected by Prussian blue (blue color) staining. Fe 3+ accumulation at 48 hpi was observed (blue color) inside and around IH in DJ cells in the incompatible interaction with M. oryzae race 007 (avirulent), but not in the compatible interaction with M. oryzae PO6-6 (virulent). However, Fe 3+ accumulation could be detected in ∆Osnlp2 epidermal cells infected with both M. oryzae race 007 and PO6-6. (Figures 3A and S5). These results indicate that ROS and ferric ion (Fe 3+ ) simultaneously accumulate in ∆Osnlp2 cells during M. oryzae infection. Lowmagnification images of M. oryzae-infected-rice leaf sheath cells showed all the infection phenotypes that were not specific to a particular single cell, but commonly detected in the individual constituent cells ( Figure S8).
The oxidation of lipids or lipid peroxidation is generated by ROS (hydroxyl radical and hydrogen peroxide) effects on polyunsaturated fatty acids in the membrane, resulting in significant tissue damage [42]. MDA, an indicator of lipid peroxidation, was quantified spectrophotometrically after reaction with TBA [4,43]. We analyzed lipid peroxidation in rice DJ and ∆Osnpl2 leaf sheaths infected with virulent and avirulent M. oryzae strains. Lipid peroxidation (MDA) levels were both induced in DJ and ∆Osnlp2 cells infected with M. oryzae race 007 ( Figure S5C). However, it was significantly higher in ∆Osnlp2 than in DJ cells infected with M. oryzae PO6-6 ( Figure 3C), indicating that higher ROS and Fe 3+ levels enhance oxidative stress levels in ∆Osnlp2 cells. ROS and ferric ion accumulation and the induced cell death indicated the ferroptosis was involved in the defense responses in ∆Osnlp2 mutant cells during the M. oryzae infection process.

Iron Chelator DFO and Ferroptosis Inhibitor Fer-1 Suppress ROS-and Iron-Dependent
Ferroptotic Cell Death in ∆Osnlp2 Mutants in Response to M. oryzae DFO, Fer-1, Cyt A, and DPI affected ferroptotic cell death in wild-type rice Dongjin and Hwayoung in response to M. oryzae infection but had no effect on non-infected rice [4]. Hence, these chemicals were used to investigate iron-and ROS-dependent ferroptotic cell death in leaf sheaths of ∆Osnlp2 mutant and complementation rice plants during M. oryzae infection in this study. DFO (3 mM) was applied to leaf sheaths of ∆Osnlp2 mutant #4-2 after inoculation with virulent M. oryzae PO6-6 and avirulent M. oryzae 007. ROS accumulation was detected by CM-H 2 DCFDA and DAB staining, and ferric ion accumulation was visualized by Prussian blue staining (Figures 4A and S9A). ROS and ferric ion accumulation and HR cell death were distinctly suppressed in infected ∆Osnlp2 leaf sheath cells treated with DFO. The chemiluminescence assay of ROS production showed that DFO treatment significantly suppressed ROS accumulation levels in ∆Osnlp2 leaf sheaths infected with both M. oryzae PO6-6 and race 007 (Figures 4B and S9B). We quantified the infected cell phenotypes (expanded infection, single-cell infection, and HR cell death) in rice leaf sheaths treated with DFO (3 mM) at 48 hpi ( Figures 4C and S9C). DFO treatment strongly inhibited the HR cell death response in ∆Osnlp2 leaf sheaths, which ultimately led to the compatible IH growth of M. oryzae. The ∆Osnlp2 mutant line #4-2 was treated with the ferroptosis inhibitor ferrostatin-1 (Fer-1) (10 µM). At 24 hpi, Fer-1 distinctly inhibited the accumulation of ROS and ferric ions in ∆Osnlp2 leaf sheath epidermal cells inside and around IH during M. oryzae PO6-6 (virulent) or 007 (avirulent) infection ( Figures 5A and S9A). Fer-1 treatment significantly inhibited HR cell death responses around IH with cellular aggregates (dark brown color) in ∆Osnlp2 leaf sheaths, which led to successful colonization of IH ( Figure 5A,D). ROS production was significantly suppressed in ∆Osnlp2 leaf sheath epidermal cells after Fer-1 treatment (Figures 5B and S9B). Lipid peroxidation levels in mock-and Fer-1-treated ∆Osnlp2 leaf sheath cells were tested at 48 hpi using MDA quantification ( Figure 5C). Avirulent and Virulent M. oryzae infection distinctly induced lipid peroxidation in ∆Osnlp2 leaf sheath cells, which displayed an HR cell death response. However, treatment with 10 µM Fer-1 suppressed lipid peroxidation in ∆Osnlp2 leaf sheaths infected with M. oryzae strains.

The Actin Polymerization Inhibitor Cyt A and NADPH Oxidase Inhibitor Diphenyleneiodonium (DPI) Suppress ROS and Ferric Ion Accumulation and HR Cell Death in ∆Osnlp2 Leaf Sheaths in Response to M. oryzae
Conidial suspension (5 × 10 5 conidia mL −1 ) of virulent M. oryzae PO6-6 or avirulent M. oryzae 007 with Cyt A (20 µg mL −1 ) or DPI (5 µM) solution was inoculated on rice leaf sheaths. Cyt A treatment suppressed the M. oryzae-induced accumulation of ROS and Fe 3+ and HR cell death in ∆Osnlp2 leaf sheath cells, as shown in the images of CM-H 2 DCFDA, DAB, and Prussian blue staining (Figures 6A and S9A). The chemiluminescence assay showed that ROS production in ∆Osnlp2 leaf sheaths was significantly suppressed at 48 hpi after Cyt A treatment (Figures 6B and S9B). Both virulent and avirulent M. oryzae strains grew well with normal hyphal structures in ∆Osnlp2 leaf sheath cells treated with Cyt A (20 µg mL −1 ) ( Figure 6A,C). DPI (5 µM) distinctly inhibited the accumulation of ROS and ferric ions and HR cell death in ∆Osnlp2 leaf sheaths infected with M. oryzae PO6-6 or M. oryzae 007, leading to successful IH colonization (Figures 7 and S9). ROS and iron accumulation were hardly to be detected in ∆Osnlp2 cells after DPI treatment ( Figures 7A  and S9A). Chemiluminescence quantification indicated that DPI significantly suppressed ROS production in ∆Osnlp2 leaf sheaths at 48 hpi ( Figures 7B and S9B). DPI treatment distinctly promoted M. oryzae PO6-6 and M. oryzae 007 infection but inhibited HR cell death in ∆Osnlp2 leaf sheaths ( Figures 7C and S9C).

Time-Course Expression of Defense-Related Genes in ∆Osnlp2 Mutant Cells during M. oryzae Infection
To analyze the resistant responses, several defense-related genes expression levels in leaf sheaths of DJ and ∆Osnlp2 mutant line #4-2 were detected at different time points after inoculation with virulent M. oryzae PO6-6 and race 007 (Figures 8 and S10). OsPBZ1 was identified as a PBZ-inducible gene in rice [44]. In the incompatible interaction between DJ and M. oryzae race 007, OsPBZ1 was distinctly induced at 3-36 hpi in DJ cells, but not in the compatible interaction between DJ and M. oryzae PO6-6. However, OsPBZ1 was also induced at 3-36 hpi in ∆Osnlp2 mutant cells (Figures 8 and S10). Plasma membrane intrinsic proteins (PIPs), aquaporins, are membrane channels that facilitate the transport of water and small neutral molecules across biological membranes in living organisms, having a crucial role in the regulation of water transport and plant growth [45,46]. OsPIP-3A was highly induced at 6 and 24-48 hpi in ∆Osnlp2 mutant line #4-2 compared with DJ inoculated with M. oryzae PO6-6 ( Figure 8). WRKY proteins are a large family of TFs involved in various plant processes, especially in inducible defense responses to biotic and abiotic stresses [47]. OsWRKY104 induction was apparent in leaf sheaths of ∆Osnlp2 mutant line #4-2 at 3-24 hpi. By contrast, OsWRKY90 was only induced in DJ inoculated with avirulent M. oryzae race 007, but not in ∆Osnlp2 leaf sheaths inoculated with M. oryzae PO6-6, which was different with OsWRKY104 (Figures 8 and S10). Plant RBOHs produce ROS that are involved in plant immunity [3,48]. The NADPH oxidase OsRbohB was distinctly induced at 3-18 hpi in ∆Osnlp2 leaf sheaths ( Figure 8). These combined results indicate that OsNLP2 disruption triggers OsPBZ1, OsPIP-3A, OsWRKY104, and OsRbohB expression in ∆Osnlp2 leaf sheaths during the M. oryzae infection process.

Subcellular Localization and Functional Analysis of OsNLP2 Domains
Plant-specific RWP-RK family proteins are TFs that bind to specific DNA sequences (cis-elements) in the promoter regions adjacent to the genes that they regulate [49,50]. In this study, we screened a truncated RWP-RK domain-containing protein from the rice cDNA library. The NIN-like protein OsNLP2 was identified using DNA sequence BLAST in NCBI. OsNLP2 contains an N-terminal GAF domain, RWP-RK domain, and C-terminal PB1 domain. GAF domains are ubiquitous motifs present in cGMP-specific phosphodiesterases, adenylyl cyclases, and bacterial transcription factor FhlA [51]. The DNA-binding function of the RWP-RK domain determines its nuclear localization in plant cells. As expected, the subcellular localization study of OsNLP2:GFP showed that OsNLP2 was localized to the nucleus of N. benthamiana cells (Figure 1). The full-length OsNLP2, but not a specific N-terminal domain, RWP-RK domain, or PB1 domain, moved to the cellular nuclei. These results suggest that OsNLP acts as a TF that binds to a specific DNA sequence to regulate gene expression in rice. RWP-RK proteins have been proposed to be, at least partly, localized in the nucleus [21]. The RWP-RK motif (Arg-Trp-Pro-X-Arg-Lys, where X indicates any amino acid) binds to the nitrate-responsive cis-element (NRE) of nitrate-inducible genes [23,52,53]. The PB1 domain is responsible for protein-protein interactions [23,40]. The PB1 domain:GFP of OsNLP2 was localized to the cytoplasm and plasma membrane, which may mediate protein-protein interactions associated with nitrate-inducible gene expression in rice. Recently, Konishi and Yanagisawa [23] demonstrated that NLP-NLP interactions mediated by the PBI domain occur in a variety of different Arabidopsis NLP TFs.

OsNLP2 Expression Positively Regulates Susceptible Disease Development
A plentiful supply of nitrogen (N) is required for fungal pathogen growth, but increased nitrogen supply causes disease susceptibility in plants [18,54,55]. Rice NLP proteins, such as OsNLP1 and OsNLP4, enhance NUE and yield in rice [22,27]. OsNLP1, a key gene regulating nitrogen (N) utilization, was rapidly induced by nitrogen starvation in rice [27]. In rice, OsNLP3 is induced after germination and repressed by heat treatment [21]. OsNLP4 is repressed by several abiotic stresses and induced by low phosphate availability [21]. However, the question whether OsNLP genes function in plant disease and immunity has not been investigated. In our present study, virulent M. oryzae PO6-6 infection distinctly induced OsNLP2 expression and caused susceptible disease symptoms in rice DJ. By contrast, avirulent M. oryzae 007 infection did not trigger OsNLP2 induction in rice DJ, which subsequently led to HR-resistance reactions with necrotic and brownish spots. These infection phenotypes suggest that OsNLP2 induction is involved in susceptible blast disease and suppresses HR cell death and defense responses in rice during M. oryzae infection. OsNLP2 induction seems likely to create a beneficial nutrition condition for M. oryzae growth. Namely, OsNLP2 may act as a positive regulator of blast disease development during M. oryzae infection.

∆Osnlp2 Mutants Exhibit ROS-and Iron-Dependent Ferroptotic HR Cell Death Responses to M. oryzae Infection
We recently reported iron-and ROS-dependent ferroptosis in incompatible rice-M. oryzae interactions [4,5]. In the present study, we used ∆Osnlp2 mutant lines to investigate whether OsNLP2 was involved in ROS-and iron-dependent ferroptotic HR cell death responses to M. oryzae infection. ∆Osnlp2 mutants showed typical HR cell death and resistant phenotypes to M. oryzae PO6-6, which is virulent to wild-type rice DJ. ROS bursts during pathogen infection are required for HR cell death and immunity and disease-related cell death [3,56,57]. ROS and ferric ion accumulation and lipid peroxidation occurred in ∆Osnpl2 leaf sheaths. Ferric ion accumulation in M. oryzae-infected rice cells may stimulate ROS production, such as hydrogen peroxide (H 2 O 2 ) and hydroxyl radicals (·OH), which degrade DNA and other biomolecules [13]. Distinct focal ROS accumulation was detected around M. oryzae IH in ∆Osnlp2 mutant cells, which ultimately led to ROS-and iron-dependent ferroptotic cell death by Fe 3+ accumulation and lipid peroxidation. These results suggest that OsNLP2 mutation triggers ferroptotic cell death and HR-resistant reactions against M. oryzae infection. An iron-dependent ROS burst and lipid peroxidation may mediate ferroptotic cell death in rice [4,5]. Liu et al. [36] previously demonstrated that ferric iron deposition at the powdery mildew infection site mediates the ROS burst in epidermal cells of wheat (Triticum aestivum) leaves. The iron chelator deferoxamine (DFO) prevents iron-dependent ferroptotic cell death in mammalian cells [9,58]. During M. oryzae PO6-6 or 007 infection, DFO treatment inhibited the accumulation of ROS and ferric ions and HR cell death, in ∆Osnlp2 leaf sheaths, which may become susceptible to M. oryzae invasion. At that time, M. oryzae could not take up DFO-chelated iron from the host ∆Osnlp2 leaf sheath cells, as suggested by Dangol et al. (2019) [4]. The accumulation of lipid-based ROS such as lipid peroxides is involved in the iron-dependent ferroptotic cell death [14]. In this study, the lipid-based ROS scavenger ferrostatin-1 (Fer-1) distinctly inhibited lipid peroxidation and ferroptotic HR cell death in ∆Osnlp2 leaf sheaths. Fer-1 may inhibit the production of lipid hydroperoxides by blocking lipid peroxidation [9,59]. These combined results suggest that Osnlp2 mutation triggers iron-and ROS-dependent lipid peroxidation to induce ferroptotic HR cell death.
Cytochalasin (Cyt) binds to actin microfilaments and inhibits actin polymerization in plant cells [60][61][62]. Cyt A treatment distinctly suppressed the accumulation of ROS and Fe 3+ and HR cell death, leading to normal M. oryzae growth in ∆Osnlp2 leaf sheath cells. Cyt A may interfere with iron accumulation in the M. oryzae infection site in ∆Osnlp2 leaf sheath cells. Actin microfilament polymerization is required for the deployment of key early plant defense responses [62][63][64]. Diphenyleneiodonium (DPI) inhibits cellular ROS production mediated by NADPH oxidases in the plasma membrane [65][66][67]. NADPH oxidases, known as respiratory burst oxidase homologs (RBOHs), are responsible for ROS production in plants during pathogen infection [3,68,69]. DPI treatment distinctly suppressed ROS and ferric ion accumulation and HR cell death in ∆Osnpl2 mutant cells [4]. A possible inhibition of NADPH oxidase OsRbohB by DPI may suppress the iron-dependent accumulation of ROS and HR cell death in rice during M. oryzae infection. All DFO, Fer-1, Cyt A, and DPI treatments attenuated ferroptotic cell death and promoted M. oryzae infection in ∆Osnlp2 mutants. Collectively, these data suggest that OsNLP2 expression negatively regulates iron-dependent ROS accumulation, lipid peroxidation, and ferroptotic HR cell death during M. oryzae infection.

Broad-Spectrum, Nonspecific Resistance of ∆Osnlp2 Mutants to Different M. oryzae Strains
In this study, OsNLP2 mutation induced OsPBZ1, OsPIP-3A, OsWRKY104, and Os-RbohB in ∆Osnlp2 leaf sheaths during M. oryzae infection. Probenazole (PBZ) induces race-nonspecific resistance in rice plants against M. oryzae. OsPBZ1 is known as a PBZinducible gene in rice [44]. Plasma membrane intrinsic proteins (PIPs), aquaporins, function as membrane channels to mediate H 2 O 2 transport inside the plant cell across the plasma membrane [3,45]. OsPIP2;1 has a crucial role in the regulation of water transport and plant growth [46]. The rice genome contains more than 100 WRKY genes. The OsWRKY71 TF is involved in rice defense response [70]. The NADPH oxidase OsRbohB exhibits apoplastic ROS-producing activity in rice cells [17]. The distinct induction of defense-related genes, such as OsPBZ1, OsPIP-3A, OsWRKY104, and OsRbohB, by the OsNLP2 mutation may trigger iron-and ROS-dependent ferroptotic cell death and broad-spectrum resistance in ∆Osnlp2 mutants during M. oryzae infection.
∆Osnlp2 mutant lines were highly resistant to different M. oryzae strains KJ401, Y34, RO1-1, and 70-15, which are virulent to wild-type rice DJ. A complementation test verified that OsNLP2 is the causal gene for ∆Osnlp2 mutant phenotypes. The complementation lines exhibited susceptible infection types, i.e., 40-60% cells with expanded infection hyphal growth of different M. oryzae strains. These data support the hypothesis that loss-offunction of OsNLP2 may trigger broad-spectrum, non-specific resistance to diverse M. oryzae strains. To investigate whether the OsNLP2 mutation affects plant growth, we measured parameters of ∆Osnlp2 and wild-type rice DJ plants. There was no significant difference in the number of spikes, the number of kernels per spikes, and the weight of one thousand seeds (data not shown). These results demonstrated that the OsNLP2 mutation induced enhanced resistance without compromising growth and yield. In this study, we have screened OsNLP2 from rice cDNA library by yeast two-hybrid screening using a M. oryzae effector as bait. A partial, but not entire, OsNLP2 protein was found to interact with the M. oryzae effector Avr-Pikm (data not shown). This suggests the possibility that M. oryzae effectors interact with OsNLP proteins inside rice cells. Furthermore, different M. oryzae strains may secrete non-specific effectors to induce OsNLP2 expression, which ultimately leads to create a beneficial cellular condition for M. oryzae infection and suppress HR cell death and defense responses in rice.
OsNLP2 mutation triggered the expression of some defense-related genes, such as OsPBZ1, OsPIP-3A, OsWRKY104, and OsRbohB, in ∆Osnlp2 mutants. Rice defense-related genes, which function in pattern-triggered immunity (PTI) signaling, have been proposed to mediate broad-spectrum resistance to two or more pathogen species independently [71]. A natural allele of the C 2 H 2 -type TF BSR-D1 in rice confers broad-spectrum resistance to M. oryzae [72]. Zhou et al. [73] recently demonstrated that the rice bsr-k1 (broad-spectrum resistance Kitaake-1) mutant confers broad-spectrum resistance against M. oryzae and Xanthomonas oryzae pv. oryzae. Loss-of-function of the Bsr-k1 gene induced OsPAL1 expression in the bsr-k1 mutant. OsPAL1 overexpression in wild-type rice conferred resistance to M. oryzae. Collectively, our study suggests that OsNLP2 negatively regulates ferroptotic cell death and immune responses in rice, and provides a valuable gene source for molecular breeding of rice with broad-spectrum resistance to diverse M. oryzae strains.

Conclusions
In this study, we identified that OsNLP2 was highly induced during the virulent M. oryzae infection process but suppressed during avirulent M. oryzae infection. OsNLP2 was localized in the nucleus, suggesting that OsNLP2 function as a transcriptional factor regulating the rice defense response. The ∆Osnlp2 lines exhibited an immune response with hypersensitive cell death against the virulent M. oryzae strains. However, the complementation lines of OsNLP2 restored the susceptible phenotype to the virulent M. oryzae strains. Deletion of OsNLP2 induced non-race specific resistance to virulent M. oryzae strains. ROS and ferric ions are highly accumulated in ∆Osnlp2 cells to ensure the strong resistance to virulent M. oryzae strains. Small molecules of DFO, Fer-1, Cyt A, and DPI suppressed the ferroptotic cell death in ∆Osnlp2 cells during M. oryzae infection. The defense-related genes OsPBZ1, OsPIP-3A, OsRbohB and OsWRKY104 were highly induced to ensure broadspectrum resistance in ∆Osnlp2 cells. In summary, OsNLP2 acts as a negative regulator of ferroptotic cell death and defense responses in rice with broad-spectrum resistance to rice blast disease.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox11091795/s1. Figure S1: Amino acid sequence alignment of GAF-like domain, RWP-RK domain, and PB1 domain of rice NLP family proteins with other plant NLP proteins using MEGA7 software. Figure S2: Phylogenetic tree of rice NLP family proteins with other plant NLP proteins. Figure S3: ∆Osnlp2 mutant and complementary lines genotyping. Figure S4: Agrobacterium-mediated transient expression of OsNLP2 and its PB1 and RWP-RK-PB1 domains does not trigger cell death in Nicotiana benthamiana cells. Figure S5: ROS and ferric ion accumulation and lipid peroxidation in leaf sheath cells of rice DJ and ∆Osnlp2 mutants during avirulent Magnaporthe oryzae 007 infection. Figure S6: ROS accumulation in rice DJ and ∆Osnlp2 mutant cells at different time points after inoculation with Magnaporthe oryzae PO6-6. Figure S7: ROS quantification in leaf sheaths of rice DJ and ∆Osnlp2 mutant lines infected with different Magnaporthe oryzae strains. Figure S8: Low-magnification images of ROS and ferric ion accumulation and cell death response in leaf sheath cells of rice DJ and ∆Osnlp2 mutant during virulent Magnaporthe oryzae PO6-6 infection. Figure S9: Suppression of HR cell death in ∆Osnlp2 leaf sheaths infected with avirulent Magnaporthe oryzae 007 after treatment with deferoxamine (DFO), ferrostatin-1 (Fer-1), cytochalasin A (Cyt A), and diphenyleneiodonium (DPI). Figure S10: Real-time qRT-PCR analysis of time-course expression of defense-related genes in leaf sheaths of rice DJ and ∆Osnlp2 mutant plants infected with Magnaporthe oryzae PO6-6 (virulent) and 007 (avirulent). Table S1: Primers used in this study. Table S2: Accession numbers of RWP-PK proteins used in this study.