Disruption of LEAF LESION MIMIC 4 affects ABA synthesis and ROS accumulation in rice

Hao Wu,Gaoxing Dai,Rao Yuhun,Kaixiong Wu,Junge Wang,Peng Hu,Yi Wen,Yueying Wang,Lixin Zhu,Bingze Chai,Jialong Liu,Guofu Deng,Qian Qiana,,,Jiang Hu,

a Department of Agronomy,College of Agriculture and Biotechnology,Zhejiang University,Hangzhou 310058,Zhejiang,China

b Guangxi Crop Genetic Improvement and Biotechnology Laboratory,Guangxi Academy of Agricultural Sciences,Nanning 530007,Guangxi,China

c State Key Laboratory of Rice Biology and Breeding,China National Rice Research Institute,Hangzhou 310006,Zhejiang,China

d College of Life Sciences,Zhejiang Normal University,Jinhua 321004,Zhejiang,China

Keywords: Rice Lesion mimic Reactive oxygen species Programmed cell death Zeaxanthin epoxidase Xanthophyll cycle Carotenoid Abscisic acid

ABSTRACT Lesion mimic mutants (LMMs) are advantageous materials for studying programmed cell death (PCD).Although some rice LMM genes have been cloned,the diversity of functions of these genes indicates that the mechanism of cell death regulation in LMMs needs further study.In this study,we identified a rice light-dependent leaf lesion mimic mutant 4 (llm4) that showed abnormal chloroplast structure,photoinhibition,reduced photosynthetic protein levels,massive accumulation of reactive oxygen species(ROS),and PCD.Map-based cloning and complementation testing revealed that LLM4 encodes zeaxanthin epoxidase(ZEP),an enzyme involved in the xanthophyll cycle,which functions in plant photoprotection,ROS scavenging,and carotenoid and abscisic acid (ABA) biosynthesis.The ABA content was decreased,and the contents of 24 carotenoids differed between the llm4 mutant and the wild type (WT).The llm4 mutant showed reduced dormancy and greater sensitive to ABA than the WT.We concluded that the mutation of LLM4 resulted in the failure of xanthophyll cycle,in turn causing ROS accumulation.The excessive ROS accumulation damaged chloroplast structure and induced PCD,leading eventually to the formation of lesion mimics.

1.Introduction

Lesion mimic mutants(LMMs)can spontaneously form necrotic lesions in local tissues without external injury or pathogen infection.In the 1920s,the first lesion mimic mutant was reported in maize[1].Since then,LMMs have been reported inArabidopsis,rice,wheat,and many other plants [2-4].Uncontrolled programmed cell death (PCD) influences the formation of lesion mimics [5-7].However,PCD is induced by multiple factors and its generation mechanism is complex [8-10].The underlying genetic,molecular,and physiological mechanisms involved in PCD can be revealed by studying uncontrolled cell death in LMMs [11].

Reactive oxygen species (ROS) are signaling molecules that mediate the occurrence of PCD and respond to a variety of abiotic and biological stresses in plants [12].ROS are produced in cytoplasm and various organelles,and low concentrations of ROS induce the activation of plant defense systems by participating in cell signal transduction [13].However,when plants are exposed to biotic or abiotic stresses,cellular ROS content increases,and high concentrations of ROS may affect cell membrane properties and cause oxidative damage to nucleic acids,lipids and proteins that may lead to cell death[14-17].In many LMMs,the formation of lesions is related to cell death,and excessive accumulation of ROS occurs during the formation of necrotic lesions.Excessive accumulation of superoxide anion radicals (O2.-) and ozone (O3)is directly responsible for the generation of necrotic lesions in theArabidopsismutantrcd1[18].RiceRF14eencodes a 14-3-3 protein,and RNA interference with the gene produces lesion mimics accompanied by excessive accumulation of ROS[19].ELL1is a cytochrome P450 monooxygenase gene involved in chloroplast development.Disruption ofELL1leads to destruction of chloroplast structure and accumulation of ROS,which eventually triggers PCD and the formation of lesions in rice[20].Collectively,excessive ROS accumulation acts in mediating cell death and lesions formation.

To avoid excessive ROS accumulation,plants have evolved mechanisms to scavenge ROS,among which the xanthophyll cycle is among the most effective for protecting plants from oxidative stress [21].In this cycle,zeaxanthin is epoxidized by zeaxanthin epoxidase (ZEP) under low-light conditions to produce violaxanthin,which is reduced to zeaxanthin by violaxanthin deepoxidase under high-light conditions [22].Some carotenoids act as quenchers of chlorophyll singlet (1Chl*) formation,preventing ROS formation in this cycle[21,23].These findings suggest that disruption ofZEPmay lead to excessive ROS accumulation.InArabidopsis,the ZEP-deficient mutantaba1accumulates more singlet oxygen under high light conditions[24].But the first step in abscisic acid (ABA) biosynthesis pathway is the epoxidation of zeaxanthin to violaxanthin,catalyzed by ZEP [25].Thus,ZEP directly regulates ABA content and participates in physiological processes including biotic and abiotic stress response,seed dormancy,and growth and development [26-30].ZEP is also involved in carotenoid synthesis,and its null mutations can alter the carotenoid composition in plants.InArabidopsis,the total content of carotenoids in seeds of a ZEP-deficient mutant was increased sixfold compared to the wild type [31].Natural variation inZEPgene expression during seed development is identified as a potential mechanism for tuning carotenoid composition,stability,and content.In tomato,high-pigment 3,a mutant for theZEPgene,shows an increase of more than 30% carotenoid content and only 75% of the ABA content of the wild type [32].The splicing mutation ofZEPresults in higher zeaxanthin and total carotenoid content in mature pepper fruits [33].

Although the function ofZEPhas been studied in various plants,the relationship betweenZEPand PCD in rice remains unknown.To further elucidate the mechanism of action of PCD in rice,we isolated and characterized a rice LMM namedllm4.The lesion mimic phenotype of this mutant emerged at the seedling stage and was induced by light.Map-based cloning revealed thatLLM4encodes zeaxanthin epoxidase,an enzyme in the xanthophyll cycle.Mutation of theLLM4gene resulted in cell death as well as altered carotenoid composition and ABA sensitivity.These findings provide new insights into the function of riceZEPin PCD and growth and development.

2.Materials and methods

2.1.Plant materials and growth conditions

The lesion mimic mutantllm4was isolated from an ethyl methylsulfonate(EMS)-inducedjaponicarice(Oryza sativa)Yundao 32 (wild type,WT) mutant population.For morphological and genetic analyses,all plants were grown in paddy fields at the China National Rice Research Institute in Fuyang,Zhejiang province and Lingshui,Hainan province,China.For ABA treatment,seeds were incubated in a chamber (10 h dark/14 h light cycles with 70% relative humidity) at 28 and 30 °C.

2.2.Pigment measurement and transmission electron microscopy observations

Leaves (0.05 g) were collected from WT andllm4plants at the seedling,tillering and heading stages to measure pigment content using previously described methods[34].Absorbance values of the sample solution were determined with a microplate reader at 663 nm and 645 nm,and the pigment content was calculated following Tang et al.[35,36].Leaves from WT andllm4plants at the tilling stage were used for transmission electron microscopy(TEM) observations,performed following Hu et al.[37].

2.3.Histochemical assay and oxidation index determination

Leaves of WT andllm4plants at the seedling stage were harvested for histochemical assay.Trypan blue staining,3,3′-diaminobenzidine (DAB) staining,and nitroblue tetrazolium(NBT) staining were performed to detect non-viable cells,H2O2accumulation,and superoxide anion (O2.-) accumulation as previously described [38,39].2′,7′-Dichlorofluorescein diacetate (H2-DCFDA) staining was performed to monitor the production of total ROS as previously described[40].Tissues of transgenic plants were collected for β-glucoronidase (GUS) staining as previously described [41].

Leaves from the same sites of WT andllm4plants were collected to determine the oxidation index.Kits for measuring malonaldehyde (MDA) and hydrogen peroxide (H2O2),and for measuring the activities of superoxide dismutase (SOD),peroxidase (POD),and catalase(CAT),were purchased from Suzhou Grace Biotechnology Co.,Ltd.(Suzhou,Jiangsu,China).The assays were performed according to the manufacturer’s instructions.

2.4.Paraffin sectioning and terminal deoxynucleotidyl transferase dUTP nick-end labeling

Leaves were collected and fixed in 70%formalin-acetic acid-alcohol solution (FAA) overnight.Fixed samples were successively dehydrated in 85%,95%,and 100% alcohol,immersed in xylene,embedded in paraffin,and sliced with a rotary microtome to generate 8-μm thick sections.Slices were dewaxed in xylene,rehydrated in ethanol,stained with 1% safranin and 1% Fast Green,and observed with an ECLIPSE 90i microscope (Nikon,Tokyo,Japan).

For terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay,samples were processed,embedded,and sliced as described above.After rehydration with ethanol and treatment with proteinase K,the TUNEL assay was performed following the instructions of the TUNEL kit (Promega,Madison,WI,USA).Apoptosis signals were visualized with an LSM 700 confocal scanning microscope (Zeiss,Germany).

2.5.Map-based cloning and sequencing

A mapping population was derived from a cross between thellm4mutant and theindicacultivar Taichung Native 1(TN1).Plants displaying leaf lesions were selected from the F2generation for mapping.For fine mapping,insertion/deletion (InDel) markers were developed based on genome sequence differences between thejaponicarice Nipponbare andindicarice 9311.The primers are listed in Table S1.For whole-genome resequencing,WT andllm4samples were sequenced at Novogene Corporation (Beijing,China).Candidate genes were predicted using the Rice Genome Annotation Project database (https://rice.plantbiology.msu.edu).Genes with amino acid changes between WT andllm4plants were selected for further sequencing.

2.6.Vector construction

For complementation of thellm4mutants,a 9939-bp genomic fragment (including upstream and downstream sequences) was inserted into the pCAMBIA1300 vector.To construct the overexpression vector,a 1980-bp coding sequence (CDS) was inserted into the pCAMBIA1300s vector.To create knockout lines,the sgRNA targeting the first exon of theLLM4gene was ligated into the pYLCRISPRCas9Pubi-H vector as previously described [34].To construct the GUS promoter,the 2220-bp promoter fragment was cloned into the pCAMBIA1305.1 vector.The primers for vector construction are listed in Table S1.

2.7.Quantitative real-time polymerase chain reaction analysis

An AxyPrep total RNA miniprep kit (Axygen,New York,NY,USA) was used to extract total RNA from plant tissues.cDNA synthesis and quantitative real-time polymerase chain reaction (qRTPCR)were performed as previously described[17].The housekeeping geneUBQ5(Os01g0328400)was used as an internal control.The primers are listed in Table S1.

2.8.Phylogenetic analysis

The LLM4 protein sequence was selected as a query to identify similar sequences in the National Center for Biotechnology Information (NCBI,https://www.ncbi.nlm.nih.gov/).Sequence alignment was performed with DNAMAN v6.0.The phylogenetic tree was generated by the maximum likelihood method using MEGA v7.

2.9.Subcellular localization

Full-length cDNA ofLLM4without the stop codon was amplified by PCR using the primers LLM4-GFP-F/R(Table S1),and the resulting target fragment was ligated into the green fluorescent protein(GFP) pCA1301-35S-S65T-GFP(CK-GFP) vector to construct the pCA1301-35S-S65T-GFP-LLM4 (LLM4-GFP) vector.CK-GFP and LLM4-GFP vectors were co-introduced into rice protoplasts and transiently expressed as previously described [42].GFP signals were observed with the confocal scanning microscope.

2.10.Protein extraction and immunoblot analysis

Total protein was extracted and quantified with a BCA protein assay kit (Coolaber SK1070;Beijing Coolabo Technology Co.,Beijing,China)according to the manufacturer’s instructions.Antibodies against photosystem proteins were procured from Agrisera(V?nn?s,Sweden).Anti-actin and anti-histone H3 antibodies were used as controls.Immunoblot analysis was performed as previously described [43,44].

2.11.Measurement of abscisic acid and carotenoid contents

Leaves from the same sites on WT andllm4plants were collected to measure ABA or carotenoid content.Three biological replicates were measured for each sample.ABA content was detected by ultra-high performance liquid chromatography triple quadrupole mass spectrometry (UPLC-MS/MS) as previously described [45].Carotenoids were extracted following Fang et al.[46] with some modifications.Carotenoid composition was determined by Metware Biotechnology Co.,Ltd.(Wuhan,Hubei,China)by UPLC-MS/MS as previously described [47].

2.12.Fluridone treatment

Fluridone (Flu) is an inhibitor of phytoene desaturase (PDS),which can inhibit the synthesis of carotenoids.WT seedlings were grown under natural conditions in nutrient solution for three weeks and then treated with Flu solution at a concentration of 50 μmol L-1.Three days later,samples were taken and photographed,and the leaves were stained with DAB and NBT dye solution.

2.13.Abscisic acid treatment

To investigate the response of germinatingllm4mutant seeds to ABA,seeds were harvested at 35 d after flowering and dehulled.Fifty seeds were spread onto Petri dishes containing filter paper(three biological replicates for each sample),and 10 mL of water or 10 μmol L-1ABA solution was added to the dishes.Petri dishes were placed in a chamber under a 12 h light/12 h dark cycle at 30°C,and the percentage of germinating seeds was recorded every 12 h for 7 d.Germination completion was defined as the length of coleoptile or root equaling one half of the length of the seed [48].

To detect the ABA sensitivity of thellm4mutant postgemination growth,the post-germination growth assay was performed under three concentrations of ABA solution (0,1,and 2 μmol L-1) for 7 d as previously described [49],and plant height was recorded.

3.Results

3.1.Phenotypic characterization of the llm4 mutant

In the field,the leaves of thellm4mutant were pale yellow,andllm4seedlings at four-leaf stage displayed small reddish-brown lesion spots on the basal leaves.With plant growth,the lesion spots gradually expanded,and the number of leaves with lesions increased (Fig.1A-F).In view of the differences in leaf color and lesion spots between thellm4mutant and the WT,the pigment content was measured in several growth periods.The Chlaand Chlbcontents inllm4leaves were markedly decreased relative to the WT (Fig.S1A-C).

A decrease in chlorophyll content is usually accompanied by abnormal chloroplast development[34].To investigate the chloroplast structure ofllm4,we performed ultrastructural observations by TEM.The chloroplasts of thellm4were smaller and fewer in number,and the thylakoid structure was vague and disordered(Fig.1G).Chloroplast dysplasia often affects plant photosynthesis[50].Thellm4plants showed significantly lower net photosynthetic rate,higher intercellular CO2concentration,and higher stomatal conductance than the WT.(Fig.S1D-F).The III,V,VI,and VII internode lengths,plant height,panicle length,seed setting rate,and flag-leaf length were markedly increased inllm4relative to the WT (Fig.1H-L,S1G),whereas effective panicle number,number of grains per spike,1000-grain weight,and number of secondary panicle branches were reduced relative to the WT(Fig.1MO,S1H).Thellm4mutant also showed clear preharvest sprouting(Fig.S1I).

Light is an inducer of the formation of lesion mimics,and a shading experiment is usually performed to determine whether the appearance of lesion mimics is induced by light [51-54].To determine whether the lesions of thellm4mutant were induced by light,a shading experiment was performed before lesions appeared.There were no lesions in covered regions or other parts of leaves.However,when the shading aluminum foil was removed for a week,the lesions appeared in covered regions ofllm4(Fig.S1J-L).Thellm4mutant showed a significantly lowerFv/Fmratio than did the WT(Fig.S1M).Fv/Fmis an indicator of the degree of photoinhibition,and a higher ratio represents a lower degree of photoinhibition[55].These results suggest that lesion mimics were induced by light inllm4.

3.2.Photosynthetic protein reductions and gene expression alterations in the llm4 mutant

In view of the destruction of chloroplast structure and the photoinhibition inllm4,we speculated that the photosynthetic proteins were damaged.To test this hypothesis,we performed Western blotting to examine the photosynthetic proteins inllm4and WT.The expression levels of photosystem I (PSI) chlorophylla/b-binding proteins (Lhca1,Lhca2),PSI core proteins (PsaB,PsaC,PsaE),PSII chlorophylla/b-binding proteins (Lhcb2,Lhcb3,Lhcb5),and PSII core protein(PsbD)were decreased inllm4relative to the WT (Fig.2A).Measuring the expression levels of chloroplast development-associated genes (RpoC1,RpoC2,Pps15,V1,V2,V3),photosynthesis-associated genes (Lhcb1,Lhcb4,PsaA,PsaC,PsbA,AtpE),chlorophyll degradation-associated genes(NYC3,SGR,RCCR1,RCCR2,PAO),and chlorophyll biosynthesis-associated genes (YGL1,YGL8,CAO1,PORA,PORB,CHLH,CHLD,CHLI,CHLM,CHEMA1,CHML,CHMB,GUN4,DVR),showed that the expressions of all genes were significantly down-regulated except for chlorophyll degradationassociated genes inllm4(Fig.2B-E).Thus,the damage to the chloroplast structure coincided with the changes in photosynthetic protein levels and the expression of genes associated with chloroplast development,photosynthesis,chlorophyll biosynthesis,and chlorophyll degradation.

3.3.ROS accumulation in llm4 destroys cell structure and causes DNA damage

Accumulation of ROS during photosynthesis leads to chloroplast destruction,resulting in PCD and the lesion phenotype [3,56].To detect ROS accumulation inllm4,DAB and NBT staining were performed·H2O2,and O2.-were accumulated inllm4leaves(Fig.3A,B).Trypan blue stain also displayed a deeper color around lesions inllm4,indicating that much cell death occurred in the leaves(Fig.3C).The fluorescent probe H2DCFDA was used to detect ROS production.In a finding consistent with those of the lesions,oxidized-state fluorescence signals were observed inllm4but not in WT (Fig.3D).

Fig.3.ROS accumulation and cell death detection in WT and llm4.(A-C)DAB,NBT and trypan blue(TB)staining in leaves of WT and llm4.(D)Microscopic analysis of leaves of WT and llm4,incubated with H2DCFDA.Green indicates oxidized H2DCFDA and red indicates chlorophyll.(E,F)H2O2 and MDA content in leaves of WT and llm4.(G-I) CAT,SOD and POD activity in leaves of WT and llm4.(J)Relative expression of ROS scavenging-associated genes in WT and llm4.(K)TUNEL assay of WT and llm4.The red signal is propidium iodide (PI) staining;green fluorescence represents TUNEL-positive signals.*, P ＜0.05;**, P ＜0.01 (Student’s t-test);Scale bars,50 μm in (D,K);Error bars,± SD(n=3).

The contents of both H2O2and MDA were significantly higher inllm4than in the WT (Fig.3E,F).The accumulation of MDA indirectly reflects the degree of cellular damage [57].In paraffin sections,the cell morphological structure ofllm4was damaged,a finding consistent with the accumulation of MDA inllm4(Fig.S2A,B).Accumulation of H2O2may induce an increase in the activities of antioxidant system enzymes.CAT,SOD,and POD are the main enzymes in the plant antioxidant system and function in removing ROS and preventing physiological and biochemical changes caused by ROS[58].Therefore,we measured the activities of these three enzymes.Results revealed that CAT,SOD,and POD activities were significantly increased inllm4(Fig.3G-I).We further detected the expression of ROS scavenging-associated genes includingAPX1,APX2,SODA1,CATA,CATB,POD1,AOX1a,andAOX1b,and the expression levels of all these genes were significantly increased inllm4(Fig.3J).

ROS are major signaling molecules in PCD,and high concentrations of ROS can damage cells or induce PCD [59].DNA disintegrates in nonviable cells,and the TUNEL assay can be used to detect DNA fragmentation [6].Given that PCD is associated with DNA disintegration,we used the TUNEL assay to detect DNA fragmentation.Almost no signal was observed in WT leaves,whereas stronger signals were observed inllm4leaves (Fig.3K).DNA damage can cause changes in the expression of DNA replicationassociated genes and DNA repair-associated genes [17].We measured the expression levels of the DNA-damage marker genePAPRand the DNA-repair genesRAD51A2andRAD51C.Their expression was significantly down-regulated inllm4relative to the WT(Fig.S2C).These results suggest that excessive accumulation of ROS caused PCD inllm4.

3.4.Map-based cloning of LLM4

To isolate the mutant gene,an F2mapping population was constructed by crossingllm4with theindicacultivar TN1.Using bulked segregation analysis (BSA),thellm4locus was preliminarily mapped on chromosome 4 between markers B4-9 and B4-15 and finally confined to an interval between markers H35 and H9(Fig.4A).Next,the WT andllm4were re-sequenced and the base mutations in the mapped interval were analyzed.Finally,a single nucleotide substitution (G to A) at the junction of the 13th exon and the 13th intron ofLOC_Os04g37619was identified that resulted in a 78-bp alternative splicing insertion at the end of the 13th exon (Fig.3B-E).The mutation led to a frameshift at the 576th amino acid and a premature termination at the 577th amino acid(Fig.3B-E).According to the Rice Annotation Project database[60,61],LLM4encodes a zeaxanthin epoxidase and is acts in carotenoid and ABA biosynthesis.Accordingly,the ABA content ofllm4was reduced in comparison with the WT (Fig.4F).

Fig.4.Map-based cloning and identification of LLM4.(A)LLM4 was mapped on chromosome 4 between markers H35 and H9.(B)Structure of the LLM4 gene.Lines represent introns and blue boxes represent exons.(C)Mutation sites in llm4 at the genome and protein levels.Yellow sequences represent introns,and the green sequence represents the stop codon.(D)Sequencing comparison of mutation sites in WT and llm4;a G-to-A point mutation was identified(red frame).(E)DNA gel blot separation of cDNAs of WT and llm4.(F) ABA content in WT and llm4.**, P ＜0.01 (Student’s t-test);Error bars,± SD (n=3).

3.5.Transgenic verification of LLM4 function

To investigate whether the mutation ofLOC_Os04g37619was responsible for the phenotype ofllm4,we performed a complementation test by introducing a 9939-bp tract of genomic DNA containing the upstream and downstream sequences of WT intollm4.Of the resulting 21 transgenic lines,all showed a rescued phenotype similar to WT(Fig.5A).The pigment content,LLM4expression level,and ABA content were also restored to those of WT or even higher (Fig.5B,C).The leaf phenotype and chloroplast structure of complementary transgenic plants also appeared normal and similar to those of the WT(Fig.S3A,B).When the coding sequence ofLLM4was overexpressed inllm4,the mutant phenotype was rescued in all T0positive overexpression plants and the pigment content and gene expression levels were higher than those ofllm4(Fig.S3C-E),indicating thatLOC_ Os04g37619is identical toLLM4.

Fig.5.Verification of gene function of LLM4 by functional complementation and gene knockout.(A) Phenotype of the complementation transgenic.(B) Pigment content of WT, llm4 and complementation (Com) plants at the tilling stage.Different lowercase letters indicate significant difference at P ＜0.05 (Tukey’s test).(C) Relative expression level of LLM4 in WT,llm4 and Com plants.(D)ABA content in WT and Com plants.(E,F)Phenotype of the WT,llm4,and knockout plants.(G)Leaf phenotype of the WT,llm4,and knockout plants.(H)Relative expression of LLM4 in WT,llm4,and knockout plants.(C,D,H)**,P ＜0.01(Student’s t-test);Scale bars,15 cm in(A),1 cm in(E,F),and 0.5 cm in (G);Error bars,± SD (n=3).

To further investigate the function ofLLM4,we performed a gene-editing experiment in the WT background using CRISPR/Cas9,obtaining five homozygous knockout lines with base insertions or deletions (Fig.S3F).The knockout plants displayed lesion mimic spots on leaves,and transcription ofLLM4was significantly down-regulated(Fig.5D-H).These findings indicate that the lesion mimic phenotype ofllm4mutant is caused by functional alterations ofLLM4.

3.6.Evolutionary tree and expression pattern of LLM4

Sequence comparison between genomic DNA and cDNA revealed thatLLM4is a ZEP gene containing 16 exons,and its protein consists of 659 amino acids with a molecular weight of 71.8 kD and an isoelectric point of 7.9.Comparison of the amino acid sequences indicated that LLM4 contains two FAD binding sites and one FHA domain highly similar to zeaxanthin epoxidases in other monocotyledon species (Fig.S4).To characterize the evolutionary relationships of LLM4 proteins,a phylogenetic analysis was performed.Homologs of LLM4 were found in many photosynthetic species including eudicots,monocots,and algae,with LLM4 showing the highest similarity with proteins from other monocotyledonous plants (Fig.6A).

Fig.6.Phylogenetic tree,expression pattern,and subcellular location of LLM4.(A)A phylogenetic tree of LLM4 putative homologs.LLM4 is highlighted in red.(B)Expression pattern of LLM4 in five tissues.(C) GUS signals in plants.(D) Subcellular localization of the LLM4 protein.Scale bars,1 cm in (C) and 50 μm in (D);Error bars,± SD (n=3).

To detect the expression pattern ofLLM4,qRT-PCR was performed to measure expression in root,stem,leaf,sheath,and panicle.LLM4was expressed in all tissues and its expression was higher in leaves than in other tissues(Fig.6B).InLLM4::GUStransgenic plants,GUS signals were observed in all organs,in agreement with the qRT-PCR analysis results(Fig.6C).To identify its subcellular localization,we constructed aLLM4-GFPvector and introduced it into rice protoplasts.In contrast to the control,the green fluorescence ofLLM4-GFPoverlapped with the spontaneous red fluorescence in chloroplasts (Fig.6D),revealing that the LLM4 protein is localized to the chloroplast.

3.7.Carotenoid content in the llm4 mutant

LLM4encodes a zeaxanthin epoxidase,which catalyzes the conversion of zeaxanthin (Zx) to violaxanthin (Vx) and functions in carotenoid and ABA synthesis (Fig.7A).To investigate the effects ofLLM4functional changes on carotenoid content,we performed carotenoid metabolomic analysis and found that the contents of 24 carotenoids were significantly altered inllm4compared with WT (Fig.7B;Table S2).Among them,Zx,Vx,and neoxanthin (Nx) are components of the ABA synthesis pathway that increased 66-fold,decreased 76-fold,and decreased 94-fold inllm4,respectively,revealing that mutation ofLLM4affected the synthesis of carotene and ABA.The expression levels of the carotenoid synthesis related genesPSY1,PSY2,PDS,ZDS,

Fig.7.Carotenoid content variation in WT and llm4.(A)Carotenoid and ABA synthesis pathways.(B)Heat map of the contents of 24 carotenoids with significant differences between WT and llm4.(C) Relative expression of carotenoid synthesis related genes.Error bars,± SD (n=3).

CRTISO,LCYB,LCYE,DSM2,BCH2,BCH3,CYP97A4,andVDEwere drastically reduced inllm4(Fig.7C).These findings suggest that the mutation ofLLM4affected the contents of carotenoids and the expression of related genes in the carotenoid biosynthetic pathway.

Carotenoids are required for quenching ROS in the chloroplast[62].Fluridone (Flu) is an inhibitor of phytoene desaturase (PDS)that can inhibit the synthesis of carotenoids.In response to Flu treatment,red-brown spots appeared on treated leaves.DAB and NBT staining also showed that much ROS accumulated in leaves,a finding consistent with the phenotype ofllm4(Fig.S5).These results suggest that disruption of carotenoid components influences the formation of lesion mimics inllm4.

3.8.llm4 mutant showed reduced dormancy and sensitivity to ABA

To determine the dormancy of thellm4seeds,and the response to exogenous ABA ofllm4plants,ABA treatment was performed.Following exogenous ABA application,the germination rate and growth kinetics ofllm4seeds were faster than those of WT under normal conditions but slower under 10 μmol L-1ABA treatment(Fig.8A-C).There was no difference in shoot length betweenllm4and WT under 2 μmol L-1ABA,although the shoot length ofllm4was significantly higher than that of WT in a control group(Fig.8D-F).The shoot length of WT decreased by respectively 15.4 % and 28.8 % and that of thellm4mutant by 22.7% and 43.8% under 1 μmol L-1and 2 μmol L-1ABA treatment(Fig.8G,H).Thus,llm4was more sensitive to ABA and the mutation ofLLM4could reduce dormancy.

Fig.8.ABA treatment of WT and llm4.(A) Germination phenotypes of WT and llm4 treated with 0 and 10 μmol L-1 ABA.Image was acquired on d 4.(B) Time-course germination percentage of seeds under normal condition.(C) Time-course germination percentage of seeds treated with 10 μmol L-1 ABA.(D,E) Germinated seeds were treated with ABA at 0,1,and 2 μmol L-1.The image was acquired after treatment for 7 d.(G)Shoot length after 7 d of ABA treatment.(H)Shoot length reduction percentage after 7 d of ABA treatment.**, P ＜0.01 (Student’s t-test);Error bars,± SD (n=3) in (B,C);± SD (n=10) in (G,H).

4.Discussion

LMMs are valuable genetic resources for studying plant hypersensitivity,PCD,and growth and development.In the present study,we cloned a lesion mimic mutant geneLLM4,which was previously reported to be an allele ofOsaba1[60].LLM4encodes a zeaxanthin epoxidase,and its mutantllm4showed changed carotenoid components,reduced ABA content,and accumulated ROS.

ZEP is a key enzyme in the ABA synthesis pathway (Fig.7A).Function-deficient mutants in ZEP or downstream enzymes are often used for ABA-deficient mutants in research.Given that ABA is a positive regulator of seed dormancy,a decrease in ABA content may lead to weak seed dormancy.Accordingly,mutation of genes involved in ABA synthesis causes a reduction in seed dormancy and an increase in germination rate [28].InArabidopsis,AtNCED6andAtNCED9are required for ABA biosynthesis during seed development,and their double mutantAtnced6 Atnced9shows reduced dormancy and increased seed germination rates [63].In rice,OsABA2functions in ABA biosynthesis,and its mutantlmm9150/osaba2displays decreased ABA content,reduced dormancy,and increased seed germination[64].In agreement with previous studies,the ABA content ofllm4was reduced.A germination test under normal conditions also showed thatllm4has a lower dormancy and higher germination rate than WT (Fig.8B).Decreases in seed dormancy and increases in germination rate can effectively avoid production problems of uneven emergence and low emergence rate.The growth rate ofllm4was higher than that of WT at the seedling stage,suggesting thatllm4may inhibit weed growth and survive when seedlings are flooded by rainwater.Thus,llm4has the potential to be an excellent direct-seeded rice variety.Althoughllm4displays the phenotype of pre-harvest sprouting,this response can be prevented by arranging the sowing date to avoid the rainy season and achieve a timely harvest.Although the yield traits ofllm4decreased slightly,the application ofZEPon production may be realized by gene editing to create weak allelic mutants.

In plants,ZEP is responsible for the epoxidation of zeaxanthin,and its products are both components of carotenoids and precursors of ABA.The common feature of ZEP-deficient mutants is the accumulation of zeaxanthin and the deficiency of violaxanthin and neoxanthin.In tobacco,theZEPmutantaba2-slaccumulates 7% of zeaxanthin,whereas violaxanthin and neoxanthin were not detected in mutant leaves,and the ABA content of theaba2-slranges from 23%to 48%relative to the WT[65].InArabidopsis,violaxanthin and neoxanthin were not detected,and ABA content decreased,and zeaxanthin increased in ZEP-deficient mutantsaba1-101andaba1-102[66,67].In pepper,mutations inZEPresulted in zeaxanthin accounting for 60.05% of the total carotenoids in the variety SO with orange fruits,whereas no zeaxanthin was detected in the variety SY with yellow fruits.Compared with SY,SO showed reduced ABA content and only minor amounts of antheraxanthin and violaxanthin were detected[33].In this study,the contents of 24 carotenoids were significantly different betweenllm4and WT(Fig.7B).Among them,zeaxanthin,violaxanthin,and neoxanthin are essential components in the ABA biosynthesis pathway and were dramatic change inllm4(Table S2).And the ABA content was lower than that of WT (Fig.4F),these results as observed for ZEP-deficiency mutation reported in other plants[65-67].

Zeaxanthin and violaxanthin are components of the xanthophyll cycle,which functions in photoprotective mechanisms and ROS scavenging in plants[21,68].Thus,an imbalance of zeaxanthin and violaxanthin content may impair the photoprotection and ROS-scavenging in the xanthophyll cycle.InArabidopsis,because thenpq2mutant cannot perform xanthophyll cycling owing to the lack of violaxanthin,theFv/Fmratio is lower than that of WT under both low-and medium-light conditions[69].TheArabidopsisdouble mutantnpq1 lut2is sensitive to strong light and accumulates a large amount of ROS owing to a failure to synthesize xanthophyll-cycle components [70].The thylakoids of the ZEPdeficient mutantaba1generate twice as much1O2as WT under strong red light,leading to chloroplast rupture and cell death[23].Consistent with previous studies,the content of zeaxanthin and violaxanthin was imbalance between WT andllm4,and the maximum photochemical efficiency of PSⅡ(Fv/Fm) decreased inllm4,indicating that photoprotective mechanisms and the photosystem were also impaired (Fig.S1M).DAB staining,NBT staining,H2DCHDA probe detection,and H2O2and MDA content measurement showed that excessive ROS was accumulated inllm4(Fig.3A,B,D-F),in agreement with the behavior of theArabidopsismutantaba1with impaired xanthophyll cycling [24].ROS are factors in triggering PCD,and a correlation between excessive ROS accumulation and PCD has been demonstrated[2-4].PCD occurred inllm4(Fig.3C,K).Previous studies have shown that excessive ROS accumulation causes cell membranes to be highly oxidized,affecting cell permeability and ultimately leading to cell death [8,37].We speculate that the mutation inLLM4results in the failure of the xanthophyll cycle,leading to ROS accumulation and eventually the formation of PCD and lesion mimics.

CRediT authorship contribution statement

Hao Wu:Project administration,Investigation,Visualization,Writing -original draft,Writing -review &editing.Gaoxing Dai:Investigation,Visualization,Writing -original draft.Rao Yuchun:Formal analysis,Writing -review &editing.Kaixiong Wu:Investigation,Methodology.Junge Wang:Investigation,Data curation.Peng Hu:Investigation,Resources.Yi Wen:Investigation.Lixin Zhu:Investigation.Bingze Chai:Investigation.Jialong Liu:Investigation.Guofu Deng:Supervision,Resources.Qian Qian:Supervision,Resources,Funding acquisition.Jiang Hu:Conceptualization,Supervision,Resources,Writing -review &editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors acknowledge the financial support of the National Natural Science Foundation of China (32060454,32272109),Hainan Yazhou Bay Seed Laboratory (B21HJ0215),National Natural Science Foundation of China (32072048,U2004204),and Specific Research Fund of The Innovation Platform for Academicians of Hainan Province.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.04.001.