Jiming Li,Minghui Zhng,Luomio Yng,Xinrui Mo,Jinjie Li,Lu Li,Jingguo Wng,Hulong Liu,Hongling Zheng,Ziho Li,Hongwei Zho,Xinwei Li,Lei Lei,Jin Sun,*,Detng Zou,*

a Key Laboratory of Germplasm Enhancement,Physiology and Ecology of Food Crops in Cold Region,Ministry of Education,Northeast Agricultural University,Harbin 150030,Heilongjiang,China

b College of Life Science,Northeast Agricultural University,Harbin 150030,Heilongjiang,China

c Key Laboratory of Crop Heterosis and Utilization of Ministry of Education and Beijing Key Lab of Crop Genetic Improvement,China Agricultural University,Beijing 100094,China

Keywords:Drought stress Rice ROS scavenging Glutathione peroxidase Abscisic acid

ABSTRACT The C (Cys) 2H (His) 2-type transcription factor is one of the most important transcription factors in plants and plays a regulatory role in the physiological responses of rice to abiotic stresses.A novel rice C2H2-type zinc finger protein,abscisic acid (ABA)-drought-reactive oxygen species (ROS) 3 (OsADR3),was found to confer drought stress tolerance by enhancing antioxidant defense and regulating OsGPX1.Overexpression of OsADR3 in rice increased tolerance to drought stress by increasing ROS scavenging ability and ABA sensitivity.In contrast,CRISPR/Cas9-mediated knockout of osadr3 increased the sensitivity of rice to drought and oxidative stress.An exogenous ROS-scavenging reagent restored the droughtstress tolerance of osadr3-CRISPR plants.Global transcriptome analysis suggested that OsADR3 increased the expression of OsGPX1 under drought stress.Electrophoretic mobility shift,yeast one-hybrid,and dualluciferase reporter assays revealed that OsADR3 modified the expression of OsGPX1 by directly binding to its promoter.Knockdown of OsGPX1 repressed ROS scavenging ability under drought stress in OsADR3-overexpression plants.These findings suggest that OsADR3 plays a positive regulatory role in droughtstress tolerance by inducing antioxidant defense and associated with the ABA signaling pathway in rice.

1.Introduction

Plants are subject to various abiotic stresses,with drought being one of the most important,as it influences plant growth and can lead to severe crop losses[1].Approximately 43%of the cultivated land area worldwide is subject to drought conditions,limiting the planting area[2].Understanding the mechanisms of drought tolerance in crops will allow maintaining or improving their yield and quality in areas subject to drought stress.

Reactive oxygen species (ROS),including superoxide (),hydrogen peroxide(H2O2),hydroxyl radical (HO),and singlet oxygen (1O2),play essential roles in signal transduction during plant drought stress [3].However,high levels of ROS may cause irreversible damage to cells,which is one of the most harmful effects of drought stress in rice,Oryza sativaL.[4].To balance ROS production and destruction,plants have evolved an antioxidant system that includes both antioxidant enzymes,such as ascorbate peroxidase (APX),glutathione reductase (GR),glutathione peroxidase(GPX),and catalase (CAT),and antioxidant compounds,such as ascorbate (ASC) and glutathione (GSH),to scavenge ROS [5].High ASC and GSH contents or ASC/dehydroascorbate (DHA) and GSH/oxidized glutathione(GSSG)ratios in tissues play a central sensing role in maintaining the ASC-GSH cycle under stress[6].Identifying genes regulating the activities of antioxidant enzymes and antioxidant compounds content will be essential for increasing plant stress tolerance.

The rice glutathione peroxidase (OsGPX) gene family contains five members distributed in various organelles:OsGPX1andOsGPX3in the mitochondria,OsGPX2andOsGPX5in the cytosol and chloroplast,andOsGPX4in the chloroplast[7].TheOsGPXfamily plays crucial roles in preventing H2O2accumulation and protecting cell membranes against ROS-induced damage [5].Previous study has investigated the role of GPX under abiotic and biotic stresses as well as in signal transduction and as redox sensor proteins [8].For instance,OsGPX1,encoding mitochondrial GPX,increases rice tolerance to salt stress by regulating seed germination and plays a role in ROS scavenging as well as in root and shoot development [9,10].Silencing ofOsGPX3impairs normal plant development and increases H2O2accumulation in young plants[9].However,the regulatory genes ofOsGPXand transcription factors (TFs) that bind theOsGPXpromoter are unknown.

C2H2-type TFs are a class of proteins defined by the QALGGH motif and are known to efficiently promote the stress tolerance in crops.For example,ZFP36,ZFP179,andZFP245are central regulators of ROS and abscisic acid (ABA) signaling transduction,and dramatically increase the activities of superoxide dismutase and peroxidase under drought and oxidative stresses [11–13].ZFP182andZFP252are transcription regulators that regulate the expression of stress-responsive genes and accumulation of osmoprotectants via the ABA pathway to improve drought-and oxidative-stress tolerance in rice [14,15].

In the present study,a novel rice C2H2-type zinc-finger protein,ABA-drought-ROS 3(OsADR3,LOC_Os03g55540),which is preferentially expressed in the leaf tissues of lowland rice (LR) in response to different abiotic stresses[16],was identified.The previous study has demonstrated that upland rice (UR) has evolved to be more drought-tolerant than LR,although previous studies suggested that LR might harbor more drought-tolerant genes than UR[16].To test this hypothesis,we transferred or knocked outOsADR3in the UR variety IRAT109 to determine whether it could promote drought tolerance.We evaluated the functions of the drought-tolerant geneOsADR3and rice genetic resources for breeding drought-tolerant cultivars,providing a theoretical basis for improving the molecular mechanisms of rice drought tolerance.

2.Materials and methods

2.1.Plant materials,growth conditions,and stress treatments

Rice cultivars Nipponbare and IRAT109(O.sativa japonica)were used for qRT-PCR analysis under various stresses and phytohormone treatments,and IRAT109 was used for all transgenic experiments.For qRT-PCR,rice seeds were first sterilized with 10%NaClO,germinated at 32°C for 3 days,and then grown in Hoagland nutrient solution with a 14-h light/10-h dark photoperiod,a 28 °C(light)/22 °C (dark) temperature range,200 μmol m-2s-2light intensity,and 80% relative humidity.Four-week-old plants were subjected to dehydration (20% PEG6000 (w/v)),salt (200 mmol L-1NaCl),cold (4 °C),H2O2(2% v/v),ABA (100 μmol L-1),and drought (no watering) treatments.Their leaves were harvested at 0,1,3,6,12,24,and 48 h after the beginning of treatments.All harvested leaf samples were rapidly frozen in liquid nitrogen and stored at -80 °C for RNA extraction.

For drought stress tolerance assays,uniformly germinated seeds of wild type (WT),OsADR3-OE,andosadr3-CR lines were sown in pots containing sterilized soil and grown in a greenhouse with 12-h light/12-h dark photoperiod,28 °C (light)/26 °C (dark) temperature conditions,200 μmol m-2s-2light intensity,and 80%relative humidity.Four-week-old plants had their watering withheld and were re-watered after 21 days of the drought stress treatment.

For exogenous ABA treatment,seeds of WT,OsADR3-OE,andosadr3-CR lines were sterilized with 10% NaClO and cultured on 1/2 MS medium with or without 10 μmol L-1ABA under a 14-h light/10-h dark photoperiod and 28°C(light)/22°C(dark)temperature conditions for 7 days.

For osmotic stress treatment,seeds of WT,OsADR3-OE,andosadr3-CR lines were sterilized with 10% NaClO and cultured on 1/2 MS medium,with or without 500 mmol L-1mannitol under a 14-h light/10-h dark photoperiod and 28 °C (light)/22 °C (dark)temperature conditions,for 7 days.

For oxidative stress treatment,the seeds of WT,OsADR3-OE,andosadr3-CR lines were germinated at 32 °C for 3 days,and then grown in Hoagland nutrient solution with or without 2% (v/v)H2O2under a 14-h light/10-h dark photoperiod,28°C(light)/22°C(dark) temperature conditions,200 μmol m-2s-2light intensity,and 80% relative humidity for 7 days.

Treatments were applied every day,and the phenotypes and physiological indices of the WT and transgenic plants under the experimental stresses were recorded.

2.2.RNA extraction and qRT-PCR analysis

Total RNA was extracted from rice tissues by the TRIzol method(Thermo Fisher Scientific,Waltham,MA,USA) and treated with DNase I to eliminate any DNA contamination.The quality of the total RNA was assessed using a NanoDrop 2000(Thermo Fisher Scientific,Shanghai,China).First-strand cDNA (10 μL) was synthesized according to the instructions for the PrimeScript RT Master Mix (Takara,Beijing,China).qRT-PCR primers were designed with Primer Premier 5.0 software (PREMIER,Palo Alto,CA,USA),based on theOsADR3transcript sequence (Table S1).qRT-PCR was performed as previously described [17].

2.3.Multiple sequence alignment and phylogenetic analysis

A multiple sequence alignment was performed for the amino acid sequences of the C2H2 conserved domain ofOsADR3and some C2H2-type zinc-finger proteins from multiple species using MEGA 7.0 software (https://www.megasoftware.net/) [18].Unrooted trees were constructed by the maximum-likelihood method with the following parameters:Poisson correction,pairwise deletion,1000 bootstrap replicates.

2.4.Plasmid construction and generation of OsADR3-OE and osadr3-CR plants

To construct theOsADR3-OE plasmid,the open read frame of theOsADR3was cloned in theAscI andPacI sites of the pMDC32 binary vector using a specific primer (Table S2).For the construction of theosadr3-CRplasmid,the target sites were identified with CRISPR-P 2.0(http://cbi.hzau.edu.cn/cgi-bin/CRISPR).The plasmids were introduced intoAgrobacterium tumefaciensstrain EHA105.A.tumefaciens-mediated transformation of IRAT109 was performed as described previously [19].Knockout lines were confirmed by PCR sequencing with primers 5′-ATGTCGAGCGCGTCGTCCAT-3′and 5′-TTACGCGGTGAGAAGCCGG-3′.

2.5.Transcriptional activation assay in yeast

The full-length ORF of theOsADR3was inserted into theBamH I/SalI site of the pGBKT7 vector to generate theOsADR3-GAL4 BD vector.The transcriptional activation activity ofOsADR3was determined as previously described [20].The trans-activation activities of each protein were evaluated according to growth status of yeast cells and activity of β-galactosidase.The primers used forOsADR3trans-activation analysis are presented in Table S2.

2.6.Subcellular localization

The coding sequences of theOsADR3were fused to green fluorescent protein(GFP)reporter coding sequences and were inserted into theHindIII/BamH I site of pJIT163-GFP plasmids.The plasmids were transformed into isolated rice protoplasts using polyethylene glycol (PEG)-mediated transformation methods [21].The nuclearlocalized protein fused with mCherry was used to stain nuclei.GFP and chlorophyll signals were detected using a Leica TCS-SP2 AOBS laser scanning confocal microscope (Leica,Heidelberg,Germany).

2.7.Osmotic stress tolerance assay with E.coli

The effect of 1000 μmol L-1sorbitol on the growth ofE.colistrain DH5α containing the empty pET32a plasmid as a control vector and the pET32a-OsADR3recombinant plasmid was investigated by shake cultivation in liquid cultures.E.coliwas inoculated into LB medium containing the same concentration of sorbitol,and then cultured under the same conditions to logarithmic phase(～4 h).The growth situation of the bacterial suspension was measured by flow cytometry using the Guava Easy Cyte HT system(Merck Millipore,Darmstadt,Germany).

2.8.Yeast one-hybrid assay

Y1H assays were performed to verify the physical interactions between promoters and TFs.The promoter fragment ofOsGPX1was cloned into the pHIS2 vector.The primers used are described in Table S2.The CDS ofOsADR3were inserted into the pGADT7 vector to generate recombinant pGAD-OsADR3constructs.The Y1H assay was performed according to the manufacturer’s instructions(Matchmaker Gold Y1H Library Screening System;Clontech Laboratories,Mountain View,CA,USA).pGAD-Rec2-53 and Phis2-OsGPX1-Promoter (pHIS2.1),pGAD-OsADR3and pHIS2.1,and pGAD-Rec2-53 and pHIS2.1 were used as negative controls.pGAD-Rec2-53 and p53HIS2 were provided in the kit as positive controls.The plasmids were co-transformed into yeast Y187 strains which were then plated on SD-Trp/-Leu/-His medium containing either 0 mmol L-13-AT (3-amino-1,2,4-triazole) (control)or 40 mmol L-13-AT as described previously [22].

2.9.Electrophoretic mobility shift assay (EMSA)

Histidine(His)-OsADR3fusion proteins were obtained byin vitroprokaryotic expression.The cDNAs encoding full-lengthOsADR3were cloned into pET28a to generate His-fusion recombinant vectors,which were then expressed inE.coliBL21 (DE3) (TransGen Biotech Co.,Ltd.,Beijing,China,CD601-02).0.2 mmol L-1isopropyl b-D-1-thiogalactopyranoside was applied to induce protein expression.His-fusion proteins were purified with a His-tagged Protein Purification kit (P2229S,Beyotime Biotechnology,Shanghai,China) according to the manufacturer’s instructions.

TheOsGPX1promoter fragment containing CACAAATAGTG motifs was synthesized by TSINGKE (Beijing,China).The EMSA Probe Biotin Labeling kit and Chemiluminescent EMSA kit (GS008 and GS009;Beyotime Biotechnology,Shanghai,China) were used for the EMSA,which were performed according to the manufacturer’s instructions.Unlabeled probes were used for probe competition,and His protein was used as a negative control.

2.10.Dual-luciferase reporter assay

TheOsGPX1promoter fragment containing the CACAAATAGTG motif was inserted into pGreenII 0800-LUC vectors to generate the reporter construct.35Spro:OsADR3effectors were generated by recombiningOsADR3into the pGreenII 62-SK vector.Transformation and dual-luciferase (LUC) activity determination was performed as described previously [23].The recombinant plasmids were introduced intoA.tumefaciensGV3101 which were then cultured to an OD at 600 nm of 0.15.The reporter and effector were combined in equal volumes,maintained at 20 °C without shaking for 3 h,transfected intoNicotiana benthamianaleaves,and incubated for 60 h.The LUC andRenillaluciferase (REN) activity levels were determined using a Dual-Luciferase Reporter Assay System(Promega,Madison,WI,USA).Transactivation was expressed as the ratio of LUC:REN.At least 15 replicates (five biological replicates × three technical replicates) were evaluated per experiment.

2.11.siRNA synthesis and transient expression assays in rice protoplasts

The siRNA fragment ofOsGPX1(F:5′-CAGUUGUAUGAGAA GUACA-3′;R:5′-UGUACUUCUCAUACAACUG-3′) was synthesized by TSINGKE.Rice plants were grown in the dark at 28 °C for 1–2 weeks.When plants were 4–8 in.tall,protoplasts from leaf and stem tissue were isolated as described previously [24].Introduction of siRNA into protoplasts was performed essentially as described [25].To evaluate osmotic stress in rice protoplasts,the protoplasts were cultured on the base of the cell culture plates and WI solution containing 0.5 mmol L-1mannitol,20 mmol L-1KCl2,4 mmol L-1MES,pH 5.65.The protoplasts were treated with a high concentration of mannitol;WI solution was replaced with buffer containing 1 mol L-1mannitol,20 mmol L-1KCl2,4 mmol L-1MES,pH 5.65.

2.12.H2O2 detection by confocal laser scanning microscopy

H2O2production in protoplasts was visualized using the H2O2-sensitive fluorescent probe 2′,7′-dichlorofluorescein diacetate(H2DCF-DA;Thermo Fisher Scientific) as described previously[26].The fluorescent signals were detected using Leica IMAGE software (Leica Microsystems,Wetzlar,Germany).

2.13.RNA sequencing (RNA-seq) analysis

Total RNA was extracted from leaves of four-week-old rice plants under drought and normal conditions at three time points.The RNA samples (3 μg) from the three biological replicates were mixed for each genotype and sent to Beijing Biomarker Technologies Co.,Ltd.(Beijing,China) for sequencing.Construction of the sequencing libraries was performed according to the manufacturer instructions of the Illumina HiSeq 2500 platform(Illumina Inc.,San Diego,CA,USA).DEGs responding to drought stress were defined by a ≥2-fold expression change and false discovery rate(FDR) <0.01,and genes which were down-regulated or upregulated between WT and OE plants were subjected to further gene ontology (GO) enrichment analysis.Heat maps were created with HemI 1.0 software and based on the expression data [27].The expression levels of selected DEGs were verified by qRT-PCR.

2.14.Physiological measurements

Fresh weight,shoot height,root length,germination rate,and survival rate under each treatment were measured as described previously [28].Six plants of each line were used per replicate,with three replicates for each line.Free Pro content in leaves was measured by the sulfosalicylic acid method [29] in the leaves of four-week-old seedlings for each line.Malondialdehyde (MDA)content was measured as described previously [30].Total protein extract from rice leaves was used for APX,monodehydroascorbate reductase (MDHAR),dehydroascorbate reductase (DHAR),GPX,CAT,and GR activity assays.Content of total soluble protein was measured by the Bradford method [31].For GPX activity,0.05 g fresh leaves were ground to powder in liquid nitrogen and 1 mL potassium phosphate buffer(pH 7.5)was added.GPX and GR activities were further measured as described[9,32].One unit of GR was defined as the amount of enzyme that reduced 1 optical density min-1at 340 nm and one unit of GPX was defined as the amount of enzyme that degraded 1 μmol of GSH min-1.For analysis of APX activity,2 mmol L-1vitamin C(AsA)was added to the extraction buffer and the homogenate was centrifuged at 12,000×g.APX activity was determined as previously described [33].One unit of APX activity was defined as the amount of enzyme that degraded 1 μmol of AsA min-1.The enzyme activity of CAT was determined as described previously [34].One unit of CAT was defined as the amount of enzyme that degraded 0.1 mol of H2O2min-1.The assay for MDHAR and DHAR activities was performed as described previously [35,36].One unit of MDHAR activity was defined as the amount of enzyme that oxidizes 1 nmol of NADH min-1at 25 °C and one unit of DHAR activity was defined as the amount of enzyme that produces 1 nmol of AsA min-1at 25 °C.

To determine the content of H2O2,1 g of leaf tissue was ground to powder in liquid nitrogen and 2 mL of 100 mmol L-1Kphosphate buffer (pH 6.8,containing 0.1 mmol L-1EDTA) was added.Total soluble protein content was measured by the Bradford method [31].The content of H2O2was measured using the Ampliflu Red (Sigma-Aldrich,St.Louis,MI,USA) method [37].ASC and DHA were measured by high-performance liquid chromatography as described previously [38].The contents of GSH and glutathione disulfide in 3%sulfosalicylic acid extract were determined as previously described [39,40].

For quantification of endogenous ABA content,leaves (0.3 g)were ground to powder in liquid nitrogen and then homogenized in 4 mL 80% methanol containing 1 mmol L-1butylated hydroxytoluene as an antioxidant.After overnight incubation at 4 °C,the mixture was centrifugated for 10 min at 5000×gand 4 °C.The supernatant was collected,dried under nitrogen gas,and washed with 0.2 mL 100 mmol L-1disodium hydrogen phosphate solution(pH 9.2).The resulting supernatant was extracted with 0.2 mL ethyl acetate and dried under nitrogen gas.The extract was dissolved in 0.5 mL 80% methanol.Quantification of ABA was performed using enzyme linked immunosorbent assays,as described previously [41].

2.15.Statistical analyses

Differences among treatments were evaluated using analysis of variance in SPSS 19.0 software (SPSS Inc.,IBM,https://www.ibm.-com/cn-zh/analytics/spss-statistics-software) and Microsoft Excel 2016 (Microsoft,Redmond,WS,USA).Statistically significant differences (P<0.05 orP<0.01) were identified based on Student’st-tests.

3.Results

3.1.Expression profiles of OsADR3 in UR and LR

In a previous cDNA microarray analysis,OsADR3was identified as a candidate gene preferentially expressed in LR [16].As shown in Fig.1A,the expression levels ofOsADR3were higher in the LR cultivar Nipponbare (Nip) than in the UR cultivar IRAT109 under the treatments,and for at least one time point in each treatment,expression was significantly higher in Nip than in IRAT109,indicating thatOsADR3is a gene preferentially expressed in LR.For each treatment,OsADR3showed higher expression levels under the drought,ABA,H2O2,and polyethylene glycol(PEG)treatments than under the salt and cold treatments,in both UR and LR.Interestingly,the expression level ofOsADR3differed significantly between the two cultivars at some time points under the treatments.For instance,after the drought treatment,OsADR3transcript levels rapidly increased in Nip,reaching a peak after 3 h,whereas in IRAT109OsADR3transcript levels slowly increased and peaked after 12 h.In the H2O2treatment,theOsADR3transcript levels peaked after 12 h in both cultivars but were significantly higher in Nip than in IRAT109.Thus,the expression ofOsADR3was induced in response to abiotic stresses or phytohormones,preferentially in LR compared with UR.

To investigate the tissue-specific expression ofOsADR3in the two rice cultivars,its transcript levels in multiple organs during the seedling and heading stages were measured by qRT-PCR.As shown in Fig.1B,OsADR3was expressed in different tissues at different stages,mainly in leaves at the seedling stage.There were no significant differences between the two cultivars at the seedling stage.However,the expression levels ofOsADR3were higher in Nip leaves and panicles than in IRAT109 at the heading stage.In addition,OsADR3expression was lower in the root than in other tissues at the seedling and the heading stages.Thus,OsADR3was expressed predominantly in leaf tissues at the seedling stage.

3.2.Sequence and phylogenetic analysis of OsADR3

TheOsADR3gene contains a complete open reading frame(ORF)of 810 bp.The predicted protein ofOsADR3contains 269 amino acids,with a calculated molecular mass of 28.3 kDa.Multiple sequence alignments among some of the reported C2H2-type zinc-finger proteins (Fig.S1) and homology searches in the Gen-Bank database (National Center for Biotechnology Information,U.S.National Library of Medicine),revealed that the OsADR3 protein comprised two C2H2-type zinc-finger conserved domains,both with a plant-specific QALGGH motif homologous to that of many C2H2-type zinc-finger proteins(Figs.S1A,S2).Phylogenetic analysis ofOsADR3and some known rice C2H2 TFs indicated thatOsADR3clustered withOsMSR15(Fig.S1B),a regulator of rice response to water stress [15,42].The phylogenetic relationships among 174 C2H2 proteins of rice,Arabidopsis,soybean,maize,wheat,and cotton are shown in Fig.S3.The phylogenetic tree displayed five clades (I to V),andOsADR3was clustered in clade II.

3.3.OsADR3 is a transcription factor and confers osmotic stress tolerance in protokaryon cells

To identify the transcriptional activation role ofOsADR3,a yeast one-hybrid assay was performed,in whichOsADR3was fused to the GAL4 DNA-binding domain and expressed in yeast cells.The yeast strain AH109 transformed with pGBKT7-OsADR3and the negative control pGBKT7 grew normally on the synthetic dextrose minimal medium (SD)without tryptophan(-Trp).Cells containing only the pGBKT7-OsADR3also grew normally on SD without tryptophan/histidine/adenine (-Trp/-His/-Ade),but negative control cells did not.Cells containing pGBKT7-OsADR3exhibited βgalactosidase activity (Fig.2A).Thus,OsADR3showed transcriptional activation activity in yeast cells.

OsADR3showed a B-box function as a putative nuclear localization signaling (NLS) motif,suggesting thatOsADR3plays a role in the nucleus (Fig.S1A).To investigate the subcellular localization of the OsADR3 protein in rice cells,the transient expression of OsADR3-green fluorescent protein (GFP) in rice protoplasts was detected using confocal laser scanning microscopy.As shown in Fig.2B,concurrent NLS-mCherry staining showed that the OsADR3-GFP fusion protein was localized specifically to the nucleus,evidence that OsADR3 is a nuclear protein.

To further investigate the role ofOsADR3under osmotic stress,the growth ofE.colicontaining the plasmid 2× 35S::OsADR3was analyzed quantitatively in liquid media,by flow cytometry to evaluate the defensive properties ofOsADR3to osmotic stress treatments.The results supported the idea that overexpression ofOsADR3in prokaryotes increases their ability to tolerate osmotic stress (Fig.2C and D).

3.4.Overexpression of OsADR3 conferred tolerance to drought stress in rice

To evaluate the biological role ofOsADR3,the full-length ORF ofOsADR3(Fig.3A) was overexpressed in the UR cultivar IRAT109.qRT-PCR analysis showed thatOsADR3was overexpressed in the two transgenic lines (OE-7 and OE-9) (Fig.3B),and there were no obvious morphological variations under normal growth conditions at the seedling stage (Fig.3C).The results of the later phenotypic and physiological analyses after drought stress,and followed by a 7-day recovery period (Fig.3C–E),revealed thatOsADR3overexpression increased plant fresh weight and survival rates after the drought treatment.As shown in Fig.3F–H,after drought stress,rice plants overexpressingOsADR3accumulated more proline and soluble sugar and less MDA.

Drought-stress treatments applied to WT and transgenic rice during the grain filling stage (rice water-sensitive period) showed that leaves and kernels of theOsADR3transgenic rice were not significantly affected,whereas most of the kernels of the WT rice had dried up (Fig.S4A–D).These results suggested thatOsADR3also increased drought tolerance during the rice filling stage.

3.5.Overexpression of OsADR3 increased expression of stressassociated genes under drought stress

To further identify the mechanisms and downstream targets ofOsADR3,RNA-seq analysis was performed using leaves from WT andOsADR3-OE (OE-9) rice lines under normal and drought stress conditions.In total,5398 and 4777 up-regulated DEGs were identified in the leaves of WT andOsADR3-OE rice lines after exposure to drought stress for 3 and 12 h,respectively (Fig.4A–C).As shown in Fig.4A–C,24 up-regulated genes were identified when the global transcriptome ofOsADR3-OE rice was compared with that of the WT after drought stress for 3 h,and 57 up-regulated genes were identified after drought stress for 12 h.Similar to upregulated DEGs,only six down-regulated genes were identified as present in the OE vs.OE-S,WT vs.WT-S,and OE-S vs.WT-S groups when the global transcriptome ofOsADR3-OE rice was compared with that of the WT after drought stress for 3 h,and 79 downregulated genes were identified in the same group comparisons after drought stress for 12 h (Fig.S5A and B).

Fig.2.Molecular characterization and osmotic stress tolerance assay in E. coli of OsADR3. (A) Transcriptional activation assay of OsADR3 in yeast cells.Vectors pGBKT7(negative control) and pGBKT7-53+pGADT7-T (positive control) were expressed in yeast.Plates were incubated for 3 days and subjected to β-galactosidase assay.(B) The OsADR3-GFP fusion gene was targeted to the nucleus in rice protoplasts(scale bar,10 μm).A nuclear-localized protein(red)was applied to mark the nucleus.The empty GFP vector was extensively localized in both the nucleus and the cytoplasm of the rice protoplast,and used as negative control.(C) Measurement of the apoptosis of E. coli transformed by empty pET32b plasmid versus E.coli transfected with the pET32a-OsADR3 by PI staining and a flow cytometry system under normal culture condition or 1 mol L-1 sorbitol stress.(D) Numbers of cells that survived or died under normal conditions,and under 1 mol L-1 sorbitol stress.Values are means ± SE.** indicates significant differences at P <0.01.

The transcript levels of those DEGs were further investigated.As shown in Fig.4B and D,LOC_Os02g44500andLOC_Os08g29910were identified as DEGs after both 3-and 12-h drought stress treatments.Interestingly,LOC_Os02g44500is aGPXgene,namedOsGPX1[43]orOsGPX3[9,10]in previous studies,and thus namedOsGPX1in the present study.GO analysis after 3 and 12 h revealed that the DEGs affected by the overexpression ofOsADR3were enriched mainly in detoxification,rhythmic processes,and antioxidant activities (Figs.S6,S7).These results indicated thatOsADR3was involved in the detoxification and ROS scavenging mechanisms of rice under drought stress.

Some previously identified rice stress-responsive genes involved in various pathways,includingPOD1,OsGSTU40,OsAMTR1,OsARF16,CYP76M5,SD25,CYP71D10,OsRMC,OsPM1,ATL36,RCE1,andOsRLCK318[44–52],were also up-regulated at 3 or 12 h of stress treatment.Expression levels of these genes under normal or drought conditions was further verified by qRT-PCR in the transgenic rice seedlings.In agreement with the analysis of DEGs,the expression of these genes was up-regulated and downregulated or unchanged in twoOsADR3-OE lines and twoosadr3-CR lines,respectively (Fig.4E,F).

3.6.Loss of function of OsADR3 impairs rice tolerance to drought and osmotic stress tolerance

To further investigate the physiological functions ofOsADR3,the CRISPR/Cas9 system was used to editOsADR3.Two target sites within the CDS were designed and respectively integrated into the CRISPR/Cas9 editing vector(Fig.5A).Two homozygous mutants(CR-2 and CR-6)were identified by sequencing(Fig.5A).CR-2 contains a T insertion and CR-6 contains an A deletion in the CDS,which caused a frameshift mutation (Fig.5A).

In contrast toOsADR3-OE rice seedlings under drought stress,rice seedlings withoutOsADR3were more sensitive to drought stress(Fig.5B),and knockout ofOsADR3reduced plant fresh weight and survival rates after drought treatment for 7 days (Fig.5C and D).Further physiological analyses revealed that loss of function ofOsADR3reduced the content of free proline and soluble sugar and increased levels of MDA in leaves (Fig.5E–G).

Osmotic stress is one of the main factors causing damage to plants under drought stress [53].As shown in Fig.5H,under the 20%PEG6000 treatment,OsADR3-OE seedling tolerance to osmotic stress was higher than that of WT.In contrast,the growth ofosadr3-CR seedlings was inhibited by 20% PEG6000 compared to WT.The survival rate and relative water content ofOsADR3-OE seedlings were significantly higher than those of the WT,whereas lower survival rate and relative water content were observed inosadr3-CR seedlings (Fig.5I and J).Under the 500 mmol L-1mannitol treatment,the growth ofOsADR3-OE seedlings was less inhibited (Fig.S8A),exhibiting higher germination rate and shoot length,than the WT seedlings(Fig.S8B and C).Theosadr3-CR seedlings were more sensitive to mannitol treatment(Fig.S8A),and the germination rate and shoot length were markedly lower than those of the WT (Fig.S8B and C).

3.7.OsADR3 increased endogenous ABA levels,expression of ABAdependent stress-responsive genes,and sensitivity to exogenous ABA in rice

As shown in Fig.S9A,under the drought conditions,the endogenous ABA level was increased in all rice seedlings and significantly increased inOsADR3-OE seedlings,whereas the increase inosadr3-CR rice seedlings was minor.As shown in Fig.S9B,genes encoding 9-cis-epoxy-carotenoid dioxygenase,OsNCED4andOsNCED5increased ABA biosynthesis and showed higher expression levels in the twoOsADR3-OE rice lines than in the WT under the drought and ABA treatments.The transcript levels of genes encoding a dehydrin protein,OsRab16C,and an ABA-responsive factor,OsRAB21,OsLEA3,encoding a type of late embryogenesis abundant protein,andOsP5CS2,encoding key enzyme responsible for proline synthesis,were all increased in the twoOsADR3-OE rice lines compared with the WT after drought or ABA treatments (Fig.S9B).In contrast,knockout ofOsADR3reduced the expression levels of ABA-responsive genes under drought or ABA conditions(Fig.S9B).

As shown in Fig.S9C–E,under normal conditions,the different rice seedlings showed similar growth,root length,and germination rate.However,in the 1/2 MS medium containing 10 μmol L-1ABA,overexpression ofOsADR3severely inhibited root growth at the germination stage of the rice seedlings,whereas loss of function ofOsADR3promoted root growth and germination under ABA treatment.

3.8.OsADR3 regulated ROS accumulation under drought stress by inducing the antioxidant system

Fig.4.Analysis of genes up-regulated by OsADR3.(A)and(C)show Venn diagrams of up-regulated DEGs at 3 h(A)and 12 h(C)of the OE vs.OE-S,WT vs.WT-S,and OE-S vs.WT-S comparisons.OE-S and WT-S indicate the leaves from OsADR3-OE and WT rice seedlings after drought stress,respectively.(B) and (D) are the transcript levels of the DEGs found in the OE vs.OE-S,WT vs. WT-S,and OE-S vs.WT-S groups after drought stress for 3 and 12 h,respectively,identified by RNA-seq analysis.Genes in blue were differentially expressed under drought stress for 3 and 12 h.(E–F)Relative expression levels of some previously known stress-related genes up-regulated by overexpression of OsADR3 at 3 h (E) and 12 h (F) under normal or drought conditions in rice seedlings.Values are means ± SD (n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.

To investigate the role ofOsADR3in antioxidant defenses under drought stress,we measured the leaf H2O2content in the rice seedlings after drought stress.As shown in Fig.6A,overexpression ofOsADR3significantly lowered the content of H2O2in the rice seedlings,and knockout ofOsADR3led to accumulation of more H2O2in rice leaves under drought stress.Overexpression ofOsADR3significantly increased the content of ASC and GSH,and maintained high ASC/DHA (2.46) and GSH/GSSG (2.29) ratios with the increasing activities of GR,APX,DHAR,and MDHAR in theOsADR3-OE rice seedlings under drought stress (Fig.6B and C).In contrast,under drought stress,osadr3-CR showed lower ASC/DHA (1.06) and GSH/GSSG (1.11) ratios compared with WT andOsADR3-OE rice seedlings (Fig.6B,C).In agreement with the activities of antioxidant enzymes participating in the ASC/GSH cycle,overexpression ofOsADR3increased the activity of GPX,whereasosadr3-CR showed lower activity of GPX than WT andOsADR3-OE rice seedlings under drought stress.The activity of CAT and expression levels of the antioxidant genesCatB,APX1,andAPX2were higher in leaves ofOsADR3-OE seedlings than in those of the WT(Fig.S10).These results suggested thatOsADR3increased drought tolerance by inducing the antioxidant system in rice.

Given thatOsADR3reduced ROS accumulation under drought stress,we investigated whetherOsADR3plays a role in rice oxidative stress resistance.The overexpression and loss-of-function rice lines were subjected to 2%H2O2(v/v)treatment during the germination stage.Compared to the WT,theOsADR3-OE plants showed greater growth,germination rate (91.21%),and shoot length(3.37 cm)under the oxidative stress treatment(Fig.6D–F).In contrast,the growth ofosadr3-CR rice seeds was strongly inhibited by the H2O2(16.85% germination rate and 0.14 cm shoot length)(Fig.6D–F).Thus,OsADR3increased oxidative stress tolerance in rice.

3.9.OsADR3 regulated the expression of OsGPX1 by directly binding to its promoter

The results of RNA-seq and physiological assays in transgenic rice[43]suggested thatOsADR3is involved in rice ROS scavenging and regulates the expression ofOsGPX1to promote drought tolerance.To determine whetherOsGPX1showed higher mRNA levels inOsADR3-OErice seedlings than in the WT,qRT-PCR was performed for allGPXgenes.In agreement with the results of RNA-seq,the transcript levels ofOsGPX1were significantly higher inOsADR3-OE than in WT andosadr3-CR seedlings(Fig.7A).However,OsGPX4(LOC_Os06g08670) andOsGPX5(LOC_Os11g18170) were repressed after drought stress in all rice lines tested.AlthoughOsGPX2(LOC_Os03g24380) andOsGPX3(LOC_Os04g46960) were induced after drought stress,their expression levels in the WT were higher than in theOsADR3-OE andosadr3-CR rice seedlings.Thus,OsADR3may regulate the expression ofOsGPX1but has no effect on the otherOsGPXs.

To investigate howOsADR3regulated the expression ofOsGPX1,we performed prediction and found that putative C2H2 TFs may bind to theOsGPX1promoter via the C2H2 domain[54,55].The largest number of binding sites was found for the promoter fragment ofOsGPX1containing the CACAAATAGTG motifs.BecauseOsADR3contains two C2H2 domains with a conserved QALGGH motif(Fig.7B),we mutated two QALGGH motifs and performed a yeast one-hybrid assay and EMSA to determine which C2H2 domain binds the promoter ofOsGPX1(Fig.7C).

To determine whether OsADR3 bound to theOsGPX1promoter,a yeast one-hybrid (Y1H) assay was performed.The AD-OsADR3 and Phis2-OsGPX1-Pro vectors were constructed for the Y1H assay(Fig.7D),and 40 mmol L-1of 3-AT was added to prevent falsepositive results.The results suggested that the co-transformed AD-OsADR3 and Phis2-OsGPX1-pro yeast strains grew on synthetic-defined (SD) medium without tryptophan,leucine,and histidine (/-Trp/-Leu/-His) containing 40 mmol L-13-AT.In contrast,no yeast spot was detected in the negative control medium(Fig.7E).The yeast-transformed AD-OsADR3 with two QALGGH motifs substituted with EAMRHK did not grow on SD medium (/-Trp/-Leu/-His) containing 40 mmol L-13-AT (Fig.7E).Thus,OsADR3 bound to theOsGPX1promoter via two QALGGH motifs.

To confirm whether OsADR3 binds to theOsGPX1promoter,EMSA was performed.The 33-bpOsGPX1promoter fragments and mutant fragments were labeled with biotin as WT probe and mutant probe (Fig.7F),andOsGPX1promoter fragments without biotin label were used as competitors.The purified His-OsADR3 fusion proteins were used in the experiment.The EMSA displayed dark stripes of the DNA–protein complexes after the OsADR3 proteins and labeled WT probes were co-incubated (Fig.7G).In the presence of competitor probes with the same sequence,these complexes formed at a very low rate (Fig.7G).The dark stripe completely disappeared when theOsGPX1promoter fragments or QALGGH motifs of OsADR3 were mutated (Fig.7G).The Y1H and EMSA assays support the inference that OsADR3 directly binds to theOsGPX1promoters via two QALGGH motifs.

Fig.5.Knockout of osadr3 increased drought sensitivity in rice.(A) CRISPR/Cas9-edited osadr3 mutants.Three underlined letters indicate the protospacer adjacent motif(PAM)region.The red letter in CR-2 indicates a 1-bp insertion and the single line in CR-6 indicates a 1-bp deletion.(B)Phenotypes of WT,CR-2,and CR-6 rice seedlings under drought-stress for 7 days and recovery 7 days.(C–G)Fresh weight(C),survival rate(D),free proline level(E),soluble sugar level(F),and MDA content(G)of WT,CR-2,and CR-6 rice seedling leaves under drought stress.Error bars indicate ± SD (n=10).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.(H)Phenotypes of WT,OE-7,OE-9,CR-2,and CR-6 rice seedlings under 20%PEG6000-mediated osmotic stress for 3 days.(I,J)Survival rate(I)and relative water content (J) of WT,OE-7,OE-9,CR-2,and CR-6 rice seedlings under osmotic stress.Error bars indicate±SD(n=10).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.

To evaluate whether OsADR3 activates theOsGPX1promoterin vivo,a transient expression assay was constructed inN.ben-thamianaleaves.TheOsGPX1promoter fragment containing the CACAAATAGTG motif was introduced in the pGreenII 0800-LUC vector to generate a reporter plasmid.TheOsADR3coding sequence(CDS)was fused to a pGreenII 62-SK vector to generate an effector construct (Fig.7H).Dual-luciferase assay revealed that OsADR3 inducedOsGPX1expression under normal and drought conditions and that drought stress increased this induction (Fig.7I).Thus,OsADR3 activatedOsGPX1expression by directly binding to its promoter.

Fig.6.Analysis of ROS homeostasis in WT, OsADR3-OE,and osadr3-CR rice seedlings.(A) H2O2 content in leaves of WT, OsADR3-OE,and osadr3-CR four-week-old rice seedlings under normal conditions or exposure to drought stress for 12 h.Values are means ± SD (n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.(B)Levels of GSH,GSSG,ASC,and DHA,and ratios of GSH/GSSG and ASC/DHA in the leaves of WT,OsADR3-OE,and osadr3-CR four-week-old rice seedlings under normal condition or exposure to drought stress for 12 h.Values are means ± SD(n=5).Lowercase letters indicate significant differences at P<0.05.Statistical significance was determined by Student’s t-test.(C)Activities of GPX,APX,GR,DHAR,and MDHAR in leaves of WT,OsADR3-OE,and osadr3-CR rice seedlings under normal conditions or exposure to drought stress for 12 h.Values are means±SD(n=5).Lowercase letters indicate significant differences at P<0.05.Statistical significance was determined by Student’s t-test.(D) Phenotypes of WT, OsADR3-OE,and osadr3-CR four-week-old rice seedlings under normal conditions or exposure to oxidative stress(2%(v/v)H2O2 for 7 days).(E)Germination rates of WT,OsADR3-OE,and osadr3-CR seeds under normal conditions and oxidative stress for 7 days.Values are means±SD(n=5).Lowercase letters indicate significant differences at P<0.05.Statistical significance was determined by Student’s t-test.(F)Shoot lengths of WT,OsADR3-OE,and osadr3-CR rice seedlings under normal conditions or oxidative stress for 7 days.Values are means ± SD (n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.

3.10.Silenced OsGPX1 represses the ROS scavenging ability of OsADR3-OE plants

To investigate whetherOsADR3confers drought stress tolerance in rice by regulatingOsGPX1expression,we knocked downOsGPX1inOsADR3-OE plants by transforming siRNA-OsGPX1into rice protoplasts.The H2O2contents in the rice protoplasts were loaded with H2DCF-DA and visualized by confocal laser scanning microscopy,which showed that silencingOsGPX1in theOsADR3-OE plants accumulated more H2O2in the rice protoplasts (Fig.8A and B).The negative siRNA supplied by the manufacturer was used as a negative control.qRT-PCR and RT-PCR analyses revealed that RNAi-mediated silencing in rice protoplasts substantially reduced the expression ofOsGPX1(Fig.8C and D),demonstrating the effectiveness of the experiments.These results suggest that knockdown ofOsGPX1partly reduced the ROS scavenging ability ofOsADR3-OE plants under drought stress and thatOsADR3conferred droughtstress tolerance in rice by regulatingOsGPX1expression.

3.11.Exogenous GSH restored the drought tolerance of osadr3-CR plants.

To confirm that the deficiency of drought tolerance inosadr3-CR plants was caused by an impaired antioxidant system,we used exogenous glutathione to treat the leaves of WT andosadr3-CR plants exposed to PEG6000-simulated drought stress,to investigate whether increased ROS scavenging ability could restore the drought tolerance ofosadr3-CR plants.Under drought stress for 3 days,the leaves ofosadr3-CR plants were significantly more damaged than those of the WT (Fig.8E).However,when 5 mmol L-1GSH was added to the PEG6000 solution,the leaves of WT plants showed better drought tolerance and the phenotype of the leaves ofosadr3-CR plants returned to the same level as those of the WT(Fig.8F).Exogenous glutathione significantly lowered the content of H2O2and increased the activity of GPX in the leaves of WT andosadr3-CR plants (Fig.8G and H).These results suggested that exogenous glutathione restored the antioxidant system and drought stress tolerance ofosadr3-CR plants.

Fig.7.OsADR3 directly binds to the promoter of OsGPX1 and regulates OsGPX1 expression.(A)Expression analysis of all OsGPX genes in WT,OsADR3-OE,and osadr3-CR plants cultivated under normal conditions or drought stress.Measurements were performed after 12 h of drought.Values are means ± SD (n=5).Lowercase letters indicate significant differences at P<0.05.Statistical significance was determined by Student’s t-test.(B)Gene structure of OsADR3.Exons,untranslated regions,and C2H2 domains are marked by orange double-sided wedge,blue rounded-corner rectangles,and green rounded-corner rectangles,respectively.The scale bar at the bottom indicates the lengths of the exon,intron,and untranslated regions.(C)Amino acid sequence of the C2H2 domain used in Y1H and EMSA assays.The putative motifs(QALGGH)for DNA binding are in red.m1-OsADR3 and m2-OsADR3 represent the mutant OsADR3 proteins.Two putative motifs QALGGH were replaced with EAMRHK,respectively.(D) AD-OsADR3 and OsGPX1-promoter-PHis2 vectors used for yeast one-hybrid assay.Phis2,AD,AD-OsADR3,and OsGPX1-promoter-PHis2 represent the p53HIS2 vector and pGADT7-Rec2 vector,OsADR3 gene fused to the pGAD vector,and promoter of OsGPX1 fused to the Phis2 vector,respectively.(E) Yeast one-hybrid assay displayed OsADR3 directly binds to the promoter of OsGPX1 via its two QALGGH motifs.The AD and the Phis2-pro,AD-OsADR3 and Phis2,AD-m1-OsADR3 and Phis2-pro,AD-m2-OsADR3 and Phis2-pro,and AD and Phis2 to represent negative control,respectively.Yeast were grown on a SD/-Leu/-Trp/-His plate containing 0 mmol L-1 3-AT (3-amino-1,2,4-triazole) (as a control) or 40 mmol L-1 3-AT.(F)Diagram of the wild-type and mutated probes used for electrophoretic mobility shift assay(EMSA).The wild probe is a putative OsADR3 transcription factor binding site(CACAAATAGTG)on the OsGPX1 promoter.In the mutant probe,the putative binding site sequence CACAAATAGTG was replaced with CAGCGTACCTG.(G)EMSA assays displayed an interaction between His-OsADR3 protein and the OsGPX1 promoter via its two QALGGH motifs.(H)Schematic diagrams of the effector and reporter constructs used for a dual-luciferase assay.(I) Dual-luciferase assay.Transient expression assay of the promoter activities co-transformed with effector and reporter constructs in Nicotiana benthamiana leaves.Values are means±SD(n=6).Lowercase letters indicate significant differences at P<0.05.Statistical significance was determined by Student’s t-test.

Fig.8.Silenced OsGPX1 repressed the ROS scavenger ability of OsADR3-OE plants under drought and exogenous GSH restored the drought tolerance of osadr3-CR plants.(A,B)Production of H2O2 in protoplasts from WT plants, OsADR3-OE plants,osadr3-CR plants,OsGPX1-RNAi plants,and OsADR3-OE+OsGPX1-RNAi plants.Protoplasts were treated with osmotic stress medium (+mannitol) or normal incubation medium for 2 h and then loaded with H2DCF-DA (2′,7′-dichlorodihydrofluoresceindiacetate) for 10 min.Images were presented as mean pixel intensities.Fifty protoplasts per treatment were observed for three independent replicates?H2O2 was visualized by confocal microscopy(A),and fluorescence intensity was measured with Leica IMAGE software (B).Values are means ± SD (n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.(C) Expression analysis of OsGPX1 in rice protoplasts of WT plants, OsADR3-OE plants, osadr3-CR plants, OsGPX1-RNAi plants,and OsADR3-OE+OsGPX1-RNAi plants cultivated under normal conditions or osmotic stress.Values are means±SD(n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.(D) RT–PCR analysis of OsGPX1 expression in rice protoplasts of WT plants, OsADR3-OE plants,osadr3-CR plants, OsGPX1-RNAi plants,and OsADR3-OE+OsGPX1-RNAi plants cultivated under normal conditions or osmotic stress.GADPH was analyzed as a control.(E,F) Detached leaves from 4–week–old WT and osadr3-CR plants exposed to 15% PEG6000 with or without exogenous GSH (5 mmol L-1) for 3 days to indicate the stress tolerance.(G,H) Activities of GPX (G) and H2O2 content (H) in the leaves of WT and osadr3-CR rice seedlings under normal conditions or exposure to 15% PEG6000 with or without exogenous GSH for 3 days.Values are means ± SD (n=5).Lowercase letters indicate significant differences at P <0.05.Statistical significance was determined by Student’s t-test.

4.Discussion

Multiple sequence alignments and phylogenetic analysis indicated thatOsADR3belongs to the C2H2 zinc-finger family.Some zinc-finger proteins are transcriptional activators [13,56].Other zinc-finger proteins,also harboring a DLN-box/EAR-motif,have shown transcription-repressive activities and can repress the expression of osmotic stress-related and ABA-repressive genes in plants,such asZPT2-3in petunia [57] andSTZ/Zat10,AZF1,andAZF2inArabidopsis[58].In the present study,molecular characterization revealed thatOsADR3is a TF located in the nucleus and is a transcriptional activator (Fig.2).Thus,OsADR3is controlled by multiple signaling pathways and regulates unknown genes under plant drought stress to exercise various molecular functions in response to environmental stresses.

As the expression background ofOsADR3in the UR was lower than that in the LR (Fig.1),the results of transgenic experiments,particularly the analysis of differentially expressed genes,were precise and convincing.The present study overexpressed or knocked outOsADR3in WT plants,and the phenotype and physiological experiment indicated thatOsADR3increased drought and osmotic stress tolerance(Figs.3,5).The finding that overexpressed or knocked-outOsADR3did not affect plant growth under normal conditions suggested thatOsADR3offers practical benefits for molecular breeding to increase rice drought tolerance.

The synthesis of osmotic protectants is controlled by the corresponding functional genes,such as free Pro andLEAbiosynthesis genes [59].In our study,the content of osmotic protectants and the expression of the functional genes examined in theOsADR3-OE lines were higher than those of the WT line under drought stress.Moreover,the expression of many stress-related genes was up-regulated inOsADR3-OE plants (Fig.4).Stressed (S) and non-stressed seedlings of transgenic and WT lines displayed many identical DEGs after the drought stress treatment for 3 or 12 h(3883 and 3094,respectively),indicating that most of the DEGs were produced by the stress conditions and might not be associated withOsADR3overexpression.For this reason,in the present study,only DEGs present in the OE vs.OE-S,WT vs.WT-S,and OE-S vs.WT-S groups were further investigated.These DEGs have been shown to respond to drought or osmotic stress.For instance,OsPM1 isan ABA influx carrier and plays an important role in drought response [44].Overexpression ofOsGSTU40increased drought tolerance,as reflected in germination,root growth assay,and whole plant growth [60].These results suggest thatOsADR3regulates downstream genes or is an upstream modulator that control the content of osmo-protectants,thereby increasing rice drought stress tolerance.

The expression ofOsADR3was strongly induced by exogenous ABA,indicating thatOsADR3is involved in the rice ABA signaling pathway.The phytohormone ABA plays a central role in plant responses to environmental stresses.Some TFs of rice,such asOsABI5[56] involved in ABA-dependent pathways increase rice seedling sensitivity to exogenous ABA.OverexpressingOsADR3increased rice seedling sensitivity to exogenous ABA and increased the endogenous ABA content in leaves(Fig.S9A–C),suggesting thatOsADR3participates in the ABA-dependent signaling pathway.OsLEA3[61];OsRab16C[62];OsRAB21[63],OsP5CS2[64],OsNCED4,andOsNCED5[65]showed higher expression levels in twoOsADR3-OE lines and lower expression levels in twoosadr3-CR lines than in WT rice seedlings under drought stress or exogenous ABA treatment (Fig.S9B),indicating thatOsADR3increased the synthesis of anti-permeable substances by regulating the expression of stressresistance genes to increase rice tolerance under drought stress conditions.These results suggest thatOsADR3plays a crucial role in protecting rice cells from drought stress via the ABA pathway.

Reactive oxygen species are key signal transduction molecules in plants,but their excessive accumulation can cause irreversible damage to cells [66].In previous studies,C2H2 TFs were reported to act in the antioxidant defense system by increasing the activity of antioxidant enzymes,which are the most effective mechanisms against oxidative stress [67,68].C2H2 zinc-finger proteins were shown to be essential for the expression and synthesis of SOD[13,69],POD [12],APX [70],CAT [71],and NADPH in rice under drought,low temperature,and high-salt-stress conditions,and conferred high tolerance to oxidative stress in rice seedlings.In the present study,overexpression ofOsADR3increased the activity of the ROS scavenging enzymes GR,GPX,DHAR,MDHAR,CAT,and APX in rice under drought stress and H2O2treatments(Figs.6,S10).Previous studies [72–75] indicated that increased APX,GR,MDHAR,and DHAR activities maintain the ASC-GSH cycle,thereby improving environmental-stress tolerance.Generally,GSH and ASC pools showed significant alterations in response to environmental stresses,and high ratios of ASC/DHA and GSH/GSSG are a general feature of increased oxidation processes,and contents of ASC and GSH are the sensor of the ASC-GSH cycle,the main mechanism preventing ROS accumulation in plants [5,76].Consistently,overexpression ofOsADR3increased the contents of ASC and GSH and maintained high ASC/DHA and GSH/GSSG ratios under drought stress (Fig.6B),suggesting thatOsADR3can maintain the ASCGSH cycle under drought stress.Furthermore,exogenous ROS reagent recovered theosadr3-CR phenotype and activity of antioxidant enzymes under drought stress(Fig.8E–H).These results indicate thatOsADR3can reduce ROS accumulation under drought and osmotic stress.

In plants,several TFs have been shown to bind to the promoter regions of antioxidant enzyme genes such asAPXs andCATs to activate or suppress their expression and enzyme activity involved in seed germination,disease resistance,and abiotic stresses[22,70,77].However,the TFs involved in regulatingGPXsmediated ROS scavenging are unknown.In previous studies[9,10],OsGPX1was found to be involved in H2O2homeostasis of rice and play an important role in osmotic and oxidative stress tolerance in rice.In the present study,the RNA-seq assay revealed that under both normal and drought conditions,OsADR3increased the expression ofOsGPX1and overexpression ofOsADR3increased the activity of GPX(Figs.4,6,7A).OsADR3maintained a high level of constitutive expression in all OE lines,suggesting that the genes expressed in the OE vs.OE-S,WT vs.WT-S,and OE-S vs.WT-S groups at both 3 and 12 h after treatment might be regulatory genes forOsADR3.The EMSA,Y1H,and dual-luciferase reporter assay indicated that OsADR3 bound to the promoter ofOsGPX1and regulateOsGPX1expression under normal and drought stress(Fig.7).OsADR3as a C2H2-type TF contains two plant-specific conserved amino acid sequences,QALGGH,in the DNA-recognition motif,which is important in phytohormone responses and tolerance to abiotic stresses [70].EMSA and Y1H assays showed that OsADR3 bound to the promoter ofOsGPX1only if both C2H2 domains were present (Fig.7).In tobacco,NtERF172confers drought resistance by regulatingNtCAT,and knockdown ofNtCATinNtERF172-OE plants reduced drought stress tolerance and ROS scavenging ability by more than inNtERF172-OE lines[22].Consistently,in the present study,transiently silencedOsGPX1repressed ROS scavenging ability inOsADR3-OE plants under osmotic stress(Fig.8A and B).Notably,the H2O2contents in the rice protoplasts of knocked-downOsGPX1alone were lower than those inosadr3-CR plants and higher than in siRNA-OsGPX1+OsADR3-OE,perhaps becauseOsADR3regulates other genes with redundant functions ofOsGPX1.Alternatively,OsADR3activity may not rely solely onOsGPX1.These results show strong evidence thatOsADR3increases drought and oxidative stress tolerance by modifying the expression ofOsGPX1.Thus,OsADR3increases drought stress tolerance by inducing antioxidant defense mechanisms and regulatingOsGPX1in rice,indicating its potential utility for breeding rice cultivars.

5.Conclusions

We identified a novel C2H2-type zinc finger protein,OsADR3,containing an NLS motif,L-box,DLN-box/EAR-motif,which was preferentially expressed in LR as a prominent regulator of the response to drought stress while causing rice to be sensitive to exogenous ABA.Overexpression ofOsADR3increased the activities of many antioxidant defense enzymes including CAT,GPX,and APX and maintained the ASC-GSH cycle by regulating ASC/DHA and GSH/GSSG levels to reduce damage caused by drought stress and accumulation of osmo-protectants under drought stress.OsADR3modifies the expression ofOsGPX1by directly binding to its promoter.OsADR3increases drought stress tolerance by inducing antioxidant defense and associated with the ABA signaling pathway in rice.In summary,OsADR3increases antioxidant defense mechanisms by regulating the expression ofOsGPX1and maintains the ASC-GSH cycle by regulating ASC/DHA and GSH/GSSG levels,thereby increasing the drought stress tolerance of rice.

Availability of data and materials

The RNA-seq data reported here have been submitted to the NCBI database (https://www.ncbi.nlm.nih.gov/) under the accession numbers PRJNA635251.

CRediT authorship contribution statement

Jiaming Li,Minghui Zhang,Jinjie Li,Zichao Li,and Detang Zou:designed the experiments.Jiaming Li,Luomiao Yang,Lei Lei,Xinrui Mao,Lu Li,Jingguo Wang,Hualong Liu,and Hongliang Zheng:performed the experiments.Jiaming Li,Jian Sun,Hongwei Zhao,and Xianwei Li:prepared the manuscript.

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

This work was supported by the National Key Research and Development Program of China (2017YFD0300501),National Science and Technology Major Project (2018ZX0800912B-002),National Natural Science Foundation of China (31701507),and China National Novel Transgenic Organisms Breeding Project(2016ZX08004002).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.12.005.