Yanwei Wang,Chen Deng,Pengfei Ai,Xue’an Cui*,Zhiguo Zhang*

a Joint CAAS/IRRI Laboratory for Photosynthetic Enhancement,Biotechnology Research Institute/National Key Facility for Genetic Resources and Gene Improvement,Chinese Academy of Agricultural Sciences,Beijing 100081,China

b College of Bioscience and Bioengineering,Hebei University of Science and Technology,Shijiazhuang 050000,Hebei,China

Keywords:Rice Chloroplast biogenesis PEP complex Drought stress

ABSTRACT Chloroplasts are the center of plant life activities including photosynthesis,growth and development,and abiotic stress response.Chloroplast development and biogenesis in rice have been studied in detail,but how does abiotic stress affect chloroplasts is less studied.We obtained an albino mutant, alm1,whose chlorophyll content was greatly decreased.Transmission electron microscopy showed that chloroplast development in alm1 was blocked,especially in thylakoid-like structures,which could not form normally.The ALM1 gene encodes a chloroplast-localized superoxide dismutase.Full-length ALM1 successfully restored the non-albino phenotype,and in knockout lines,the albino phenotype reappeared.The ALM1 gene is expressed mainly in young leaves.alm1 plants died as a consequence of excessive reactive oxygen accumulation after the third-leaf stage.A series of biochemical assays verified that ALM1 interacted with the OsTrxz protein,which is one of the components of plastid-encoded RNA polymerase(PEP)complexes.A western blot experiment indicated that ALM1 played an important role in stabilizing OsTrxz in rice.An overexpression test of ALM1 revealed that ALM1 can increase drought resistance by removing excess reactive oxygen in rice seedlings.This study suggests that ALM1 not only participates in rice chloroplast biogenesis,but also increases rice stress resistance by scavenging excess reactive oxygen.

1.Introduction

Photosynthesis is the process by which green plants use solar energy to assimilate carbon dioxide and water to produce organic substances and release oxygen [1].Chloroplasts are not only important organelles in plants but the main site of photosynthesis[2].Chloroplasts are composed mainly of chlorophyll A and B,and are present in some cells of the mesophyll of higher plants and in algae cells [3].The mechanisms of chloroplast biogenesis inArabidopsisand rice are well understood [4–7].

Chloroplasts not only provide an energy source for most biological activities on earth via photosynthesis but also are an important biosensor by which plants respond to adversity [8,9].Chloroplasts are the most vigorous organelle in plant aerobic metabolism and the main source of active oxygenin vivo[10].The redox reaction is the basic metabolic pathway of life,which influences the aging and death of organisms,but at the same time produces reactive oxygen species(ROS)as a byproduct[11].The accumulation of ROS can damage cell macromolecules through oxidative stress,which leads to a series of harmful biochemical reactions and ultimately to a reduction in crop yield[12,13].To resist oxidative damage effectively,plants must produce a highly active defense system,including superoxide dismutase (SOD),peroxidase (POD),catalase (CAT),and ascorbate peroxidase (APX),which remove ROS [14].Chloroplasts play a key role in plant antioxidant processes [15].Under abiotic stress,the degree of chloroplast lipid peroxidation increases,malondialdehyde accumulates,and the activity of SOD increases [15].The change in active oxygen metabolism results in an increase in antioxidant enzyme activity and an increase in toxic substances including,H2O2,and salicylic acid [16,17].If these small molecules cannot be removed in time,they cause damage to chloroplasts.On the other hand,as a ubiquitous signal molecule,ROS can act as a sensing signal in life processes such as lignification [18].Thus,the rapid and accurate regulation ofconcentration in chloroplasts plays an important role in plant response to stress.

Some scavenging mechanisms have been reported inArabidopsis,including glutathione peroxidase (GPX),ascorbate peroxidase (APX),and SOD mechanisms.SOD is the first line of defense against oxidative damage and directly determines the scavenging efficiency of ROS under stress.Superexpression of the SOD gene has been used [19–21] to improve the resistance of plants to cold,heat,salt,drought and oxidative stress.However,the mechanism of fine modulation by SOD of chloroplast stress in rice awaits elucidation.

In this study,we conducted the mutants screening related to chloroplast development in rice T-DNA insertion mutant library.We expect some mutants to be associated with abiotic stress or reactive oxygen species scavenging.By analyzing the phenotypes and gene functions of these mutants,we can further study how abiotic stress affects chloroplasts.The study would provide guidance for plants how to improve photosynthesis under abiotic stress.

2.Materials and methods

2.1.Mutant screening

Wild-type and T2mutant seeds from a T-DNA enhancer mutant population [22] were germinated in a 28 °C incubator with a light intensity of 180 μmol m-2s-1.Thealm1mutant segregated from heterozygous plants.Segregation analysis indicated that thealm1was a single-gene recessive mutation (3:1 segregation,χ2=1.31,<).After germination and growth for 10–15 days,leaf color phenotypes were recorded.Thealm1mutant was selected and used for subsequent experiments.

2.2.Chlorophyll content measurement and transmission electron microscopy

The chlorophyll contents of the wild type andalm1mutant were determined by acetone extraction with spectrophotometric assay as described [23].Briefly,0.5 g of plant leaves were placed in a round-bottom centrifuge tube and liquid nitrogen was used to freeze the sample.After the addition of steel beads,the sample was ground to powder with a tissue grinder at 200 S per minute.Next,80% acetone was added to the tube,and the sample was mixed by vortexing before overnight incubation in the dark at room temperature.After centrifugation for 10 min at 5000 r min-1,the supernatant was collected and transferred to a new centrifuge tube.The chlorophylla/bcontent was measured from its absorption spectrum by spectrophotometry according to the following equation,with 80% acetone used as a blank control:

Chla(mg g-1)=[12.7A663-2.69A645]×V/1000×W,Chlb(mg g-1)=[22.9A645-4.64A663] ×V/1000 ×W,

Chla+b(mg g-1)=[8.026A663+20.206A645] ×V/1000 ×W

WhereVis the volume (mL) of the extract andWis the weight(g) of the fresh leaves.

Fresh wild-type andalm1leaves were collected and cut into 5 × 5 mm pieces and immediately soaked in fixing solution(2% formaldehyde and 2% glutaraldehyde in 0.1 mol L-1Na cacodylate buffer).Fixed samples were infiltrated under a vacuum pump for 15 min,stored at 4 °C for 48 h,and immersed in acetone containing 2% osmium tetroxide and 0.1% uranyl acetate.The fixed samples were dehydrated by ethanol with different gradient concentration,then were embed using 67% resin for 48 h and polymerized using pure resin at 60 °C for 7 h.Finally,samples were embedded in Spurr’s resin medium,cut with a diamond knife,stuck on a copper mesh and bombarded with gold dust.They were photographed with a transmission electron microscope (Hitachi H-7500).

2.3.ALM1 gene cloning,complementation,overexpression,and knockout experiment

Thealm1mutant flanking sequence was obtained by the thermal asymmetric interlaced PCR (Tail-PCR) method following Wan[22].A co-segregation experiment was based on primer P3(180 bp from the T-DNA left border) and primers P1 and P2 flanking the insertion site.The amplified products were separated on a 2% agarose gel [22].

For a complementation experiment,we inserted the 768-bpALM1coding region into thepCAMBIA2300 vector between the restriction sites ofSmaI andXbaI,under the actin promoter.Thep-CAMBIA2300-ALM1 plasmid was introduced intoAgrobacterium tumefaciensAGL1 by electroporation,and positive recombinantAgrobacteriumclones were selected by PCR amplification of individual bacterial colonies.Thealm1mutants were unable to produce seeds.Homozygous calli were segregated fromALM1heterozygous line calli and identified by PCR detection.They were transformed with recombinantAgrobacteriumcontaining thep-CAMBIA2300-ALM1 plasmid [24].Geneticin (G418) was used as the selectable marker in selectable medium because the homozygous callus contained the hygromycin gene.Overexpression lines were created by transformation of wild-type calli with the same recombinantAgrobacterium.The regenerated plants were identified by PCR amplification.Transmission electron microscop (TEM observations of complementary lines were also made.

To obtain aALM1knockout mutant,a knockout experiment was performed using CRISPR/Cas9 technology [25].A specific target sequence was designed for the knockout ofALM1.Briefly,a 20-bp target site was selected from the geneLOC_Os06g05110exon sequence (http://cbi.hzau.edu.cn/cgi-bin/CRISPR#).The promoter of the gRNA was U6a.The targeted cassette was inserted into the CRISPR/Cas9 binary vector pYLCRISPR/Cas9Pubi-H by simultaneous digestion and ligation with a program of 37 °C for 2 min,10 °C for 3 min,20 °C for 5 min,15 cycles,and 37 °C for 2 min.The recombinant construct was introduced intoAgrobacteriumstrain AGL1 and transferred into wild-type calli byAgrobacterium-mediated transformation.Resistant calli were selected using the hygromycin selectable marker.Regenerated plants were selected by PCR amplification and sequencing.

2.4.Subcellular localization

For subcellular localization,theALM1open reading frame without its termination codon was first amplified by reverse transcription(RT)-PCR using the primer paire alm1-GFP-F/R.The full-length coding sequence for theALM1protein was integrated into the pAN580 vector to generate the p35S::ALM1:GFP translational fusion protein [26].Protoplasts cell from which the cell wall had been removed by enzymatic digestion were collected.The p35S::ALM1:GFP fusion or its negative control (PAN580-GFP) was transformed into rice protoplasts and incubated in the dark at 28°C for 16 h.Chloroplast autofluorescence was used as a marker in acquiring confocal laser scanning microscopy images [26].

2.5.Yeast two-hybrid,bimolecular fluorescence complementation,and Co-IP assay

For yeast two-hybrid screening,theALM1full coding sequences were amplified with the primers alm1-BD-f and alm1-BD-r.The amplified products were fused into the pGBKT7 vector,which was digested by the restriction enzymesEcoRI andBamHI.The recombination vector ALM1-BD was transferred into yeast strain AH109 by electric shock transformation following the GAL4 Two-Hybrid System 3 protocol (Clontech Laboratories,Mountain View,USA).TheOsTrxzandSOD2recombination vectors were fused separately into pGADT7 vectors for verifying interaction.

For bimolecular fluorescence complementarity(BIFC),transient expression vectors pVYNE and pSCYCE were used for the transfer experiment.The full-length coding sequences ofALM1andOsTrxzwere amplified and cloned into the pSCYCE and pVYNE vectors to yield the recombinant vectors ALM1-YC and OsTrxz-YN.The four transformation combinations were then transferred into rice protoplasts(ALM1–YC and OsTrxz–YN;ALM1–YC andYN;YC and OsTrxz-YN;YN and YC) as described [27].After 16 h of incubation in the dark,images were acquired by confocal laser scanning microscopy.

The OsTrxZ-FLAG vector was constructed using full-length coding sequence(CDS)of OsTrxZ and pCambia1307.A 35S::ALM1:GFP plasmid was also generated in this experiment.For transient expression,each of the single constructs 35S::ALM1:GFP and OsTrxZ-1307 or 35S::ALM1:GFP and OsTrxZ-1307 were transformed into 4-week-old rice leaf protoplasts.After transformation,the protoplasts cells were completely lysed under non-denaturing conditions and the supernatants were reserved.The supernatants were separated by 8% SDS-PAGE and probed first with anti-Flag antibodies for the detection of OsTrxZ-Flag and then with anti-GFP antibodies for the detection of ALM1-GFP.

2.6.Quantitative real-time (qRT)-PCR and Western blot assay

To determine theALM1gene expression pattern,wild-typeRNAs from several stages and tissues were extracted with Trizol Reagent(Invitrogen,CA,USA).Each tissue RNA was used to synthesize first-strand cDNA with SuperScript III Reverse Transcriptase(Invitrogen),and the oligo (dT) was used as reverse-transcribed primer.The finished cDNA solutions were used for qRT-PCR in a 25 μL reaction solution system including SYBR Fluorescent reaction solution buffer and enzyme(Takara Co.Ltd.,Otsu,Japan).Reaction conditions were as follows:95°C for 30 s followed by 40 cycles of 95°C for 5 s,60°C for 30 s,and 72°C for 30 s.Actinwas selected as an internal reference gene for the calculation of relative quantitative abundances[28].For the expression analysis of genes involved in chloroplast development,whole RNAs in wild-type andalm1plants were extracted at the third-leaf stage.

Wild-type andalm1leaf protein was extracted at the third-leaf stage and Western blot assay was performed using the antibody Anti-OsTrxZ.The procedure employed the Pierce ECL Plus Western Blotting Detection Kit (Thermo Fisher Scientific,Waltham,USA).The detection imaging was visualized with an imaging system(ChemiDocTMXRS;Bio-Rad,Hercules,USA).The antibody anti-HSP90 was used as control.Antibodies were obtained from BGI(Shenzhen,China).

2.7.Polyethylene glycol(PEG)treatment for ALM1 overexpression lines

T1overexpressing transgenic plants were identified by PCR amplification.Three independent transgenic lines(OE1–OE3)with higher expression than the wild type were selected for PEG trial treatment.The OE1–OE3 lines and the wild type were all germinated for 3 days in a 28°C incubator,transferred into 96-well plastic plates with nutrient solution,and grown to the fourth-leaf stage.Next,the nutrient solution was replaced by 20% PEG 4000 solution.After all the plants wilted,the PEG solution was replaced with nutrient solution.After 10 days,images were acquired.Seedling survival rates were recorded.

2.8.Histochemical localization of and H2O2 and SOD activity measurement

Leaf localization of superoxide anion radicals was peformed inalm1and wild type using nitro blue tetrazolium (NBT) staining following Frahry and Schopfer[30].The general process was as follows.Leaves were excised and placed in 6 mmol L-1NBT in 10 mmol L-1Na citrate buffer(pH 6.0)at 25°C for 5 h under light.The treated leaves were then immersed in 95% ethanol for 1 h at 65 °C for decolorization.3,3′-diaminobenzidine (DAB) was used as an indicator of H2O2following Thordal-Christensen[31].Leaves were collected,immersed in a solution of 1 mg mL-1DAB(pH 3.8)and incubated for 8 h,then immersed in 95% ethanol for 1 h at 65°C.After cooling,NBT and DAB treatment leaves were examined under the light microscope and photographed.

SOD activities were measured according to Superoxide Dismutase(SOD)Assay Kit(Cat#:BC0170)(Solarbio,China).The general steps are as follows.Rice leaf samples were ground to powder.The samples were dissolved by the extract solution (10 times,Lot.No.20191016).The mixture was well mixed and centrifuged for 10 min (8000 r min-1) at 4 °C.The transmission at 560 nm of the supernatant was measured in a new tube with a spectrophotometer.

2.9.Sequence analyses and statistical analysis

To identify genes homologous toALM1,ALM1protein was used as a query at the National Center for Biotechnology Information database (NCBI;http://www.ncbi.nlm.nih.gov/).Rice andArabidopsisthalianaputative homologs were aligned with MEGA 4.0 neighbor-joining method [29] to generate unrooted phylogenetic trees.ALM1expression pattern analysis was conducted using Genevestigator (https://www.genevestigator.ethz.ch).The ChloroP server predicts the presence of chloroplast transit peptides (cTP)in protein sequences and the location of potential cTP cleavage sites (http://www.cbs.dtu.dk/services/ChloroP/).The primers are listed in Table S1.Each test was repeated at least three times.The mean ± SD is shown in the Figure.Significant differences are marked with asterisks based on Student’st-test:*,P<0.05 ;**,P<0.01.For multiple comparisons,means were compared by one-way analysis of variance and Duncan’s multiple range test with a 5% level of significance.

3.Results

3.1.An albino mutant was obtained in the rice T-DNA mutant population

The rice T-DNA enhancer-trapping mutant population was systematically screened and yielded～500 albino,pale,and yellow leaf mutants[32].One albino leaf mutant,namedalm1,attracted attention owing to its mutant phenotype cosegregation with GUS staining.Thealm1albino phenotype appeared 1–2 days after the seeds germinated (Fig.1A and B).With growth and development,thealm1albino phenotype became more pronounced,but there was no significant difference between root length and seedling length(Fig.1C–E).After approximately the third-leaf stage,thealm1mutant began to wilt and gradually died(Fig.1F).Chlorophyll content measurement showed that chlorophyllaandband total chlorophyll content in thealm1mutant decreased significantly compared with the wild type at the third-leaf stage(Fig.1G).These observations indicated that thealm1mutant phenotype was not due to a specific pigment but to a decrease in total chlorophyll content.Chlorophyll abnormalities in thealm1mutants may have resulted in the albino phenotype.Thealm1mutant was also more sensitive to low temperature than the wild type (Fig.S4).

The segregation ratio of the progeny of the heterozygous line was 75 (wild type):23 (mutant),indicating that theALM1gene was controlled by a single recessive nuclear gene.

Fig.1.Comparison of seed germination and seedling growth between wild type and alm1 mutant.Seed germinated for 1 day(A,bar=0.5 cm),2 days(B,bar=0.5 cm),3 days(C,bar=1 cm),6 days(D,bar=1 cm),10 days(E,bar=2 cm),15 days(F,bar=2 cm),Left,wild type;Right,alm1 mutant.(G)Pigment content comparison between wild type and alm1 mutant.

3.2.Transmission electron microscopy observation

To investigate the cytological origin of thealm1mutant,we selected wild-type andalm1leaves at several germination periods(3,6,10,and 15 days)for transmission electron microscopy(TEM).Compared with the wild type,on the third day,the chloroplast organization of the grana and lamellar structures ofalm1did not show marked changes(Fig.2A and B).On the sixth day,the lamellar structures ofalm1became degraded (Fig.2C and D).On the 10th day,lamellar structures were rarely seen and deformed as a whole in some fields of vision (Fig.2E and F).On the 15th day,chloroplast thylakoid-like structures were completely destroyed(Fig.2G and H).We concluded that chlorophyll development and synthesis were affected in thealm1mutant owing to the destruction of chloroplast cell structure.

3.3.Cloning the ALM1 gene by Tail-PCR

Given the cosegregation of the albino phenotype ofalm1with GUS staining [32],thealm1mutant may have been caused by TDNA tagging insertion.Tail-PCR amplification was used to obtain the flanking sequence in thealm1mutant.In total,three fragments were separated after the second round of PCR amplification.Sequence alignment in NCBI database showed that two of the three fragments aligned with the vector backbone,a frequent occurrence in rice transformants [32].One fragment sequence was aligned with the rice genome chromosome 6 BAC P0535G04,revealing that one T-DNA insertion was located 100 bp upstream ofLOC_Os06g05110(Fig.3A).The GUS gene expression orientation coincided with theLOC_Os06g05110native promoter orientation,suggesting thatLOC_Os06g05110was a candidate gene.

To determine whether the albino phenotype ofalm1was indeed caused by T-DNA insertion,a cosegregation test was performed.In 22 plants randomly selected from the segregating population,fragments could be amplified with primers P1 and P2 in all wild-type but not in all albino phenotypes.With primers P1 and P3,some expected fragments were amplified in all albino phenotypes and some wild-type phenotype plants,indicating that some wild-type plants were heterozygous.The progeny phenotypes of the two heterozygous plants were counted,and the separation ratio of the wild type to albino phenotype was 3:1 (Fig.3B).RT-PCR analysis showed that theLOC_Os06g05110gene was not expressed in thealm1mutant (Fig.3C).The cosegregation and genetic experiments were strong evidence thatLOC_Os06g05110was the candidate gene.

3.4.The ALM1 gene restores the alm1 phenotype

To further verify the correct assignment of the candidateALM1gene,two experiments were performed.First,a functional complementation test was performed.The full-length CDS ofALM1was amplified,fused withpCambia2300,and then transferred to the homozygousalm1callus (Fig.S1A).In the T0progeny,21 regenerated plants were obtained,all of which grew normally (Fig.S1B).PCR tests also showed that these were all positive plants(Fig.S1C).TEM showed that the chloroplasts of regenerated plants recovered normal structure,similar to that of the wild type(Fig.S1D–F).Second,using CRISPR/Cas9 technology,theALM1gene was knocked out,and one knockout line (crisp1-1) was identified by sequencing(Fig.S1G).The crisp1-1 line showed the albino phenotype and died at the fourth-leaf stage.Sequencings of crisp1-1 showed that a base insertion in the sixth exon caused a frameshift mutation (Fig.S1H).These experiments provided strong evidence that theALM1gene was responsible for thealm1mutant phenotype.

Fig.2.Transmission electron microscopic comparison between wild type and alm1 mutant.Seed germinated for 3 days(A,B),6 days(C,D),10 days(E,F),and 15 days(G,H)(Bar=100 nm).

Fig.3.Cosegregation analysis of T-DNA insertion with albino phenotype.(A)Schematic representation of the T-DNA insertion upstream of ALM1(exons,black boxes;introns,black lines) and the gene organization of ALM1 (http://rice.plantbiology.msu.edu/cgi-bin/ORF_infopage.cgi).ATG,start codon;TGA,termination codon;RB,right border;LB,left border.P1,P2,and P3 were the primers used for genotyping.(B)Cosegregation analysis of genotype and phenotype in a segregating population.All plants homozygous for the T-DNA insertion showed an albino phenotype,indicating that the recessive mutation was caused by T-DNA insertion.Numbers represent the different plants tested.P1/P2,PCR using primers P1 and P2;P1/P3,PCR using primers P1 and P3;He,hemizygous;Ho,homozygous;W,wild type.‘‘+” represents the wild-type phenotype;‘‘-”represents the mutant phenotype.Plants 4,9,and 11 were assigned as homozygous because only the 0.6-kb fragment was amplified using primers P1 and P3.Plants 3,7,8,10 were assigned as wild type because only the 0.5-kb fragment was amplified using primers P1 and P2.In contrast,both the 0.5-and 0.6-kb fragments were amplified from plants 1,2,5,6,12,15,and 16,indicating that they were heterozygous.(C)ALM1 expression analyses in wild-type and alm1 mutant plants using actin as a control.The RT-PCR primers P4 and P5 for ALM1 span introns 3 and 4.RT-PCR using RNA from homozygous alm1 plants did not detect full-length ALM1 transcripts.

3.5.ALM1 gene sequencing analysis

TheALM1gene encodesLOC_Os06g05110,a member of the iron superoxide dismutase family.The full length of theALM1gene is 4147 bp,and the full length of the coding frame sequence is 768 bp.ALM1encodes 255 amino acids.The gene contains nine exons and eight introns.There are two other superoxide dismutases in rice to target chloroplasts (LOC_Os06g02500.1andLOC_Os08g44770.1),which are located on chromosomes 6 and 8.MEGA analysis showed that theFSD2andFSD3genes ofArabidopsis thalianashowed the highest identity,with amino acid similarity reached 37.1% and 59.7%,respectively (Fig.S2).

3.6.Expression pattern and sub-cellular location

The development of rice leaves is the differentiation of leaf primordia into mature leaves.This process spans P0 (leaf initiation),P1 (youngest primordium),P2,P3,P4,P5,and P6 (Fig.4A).The expression patterns of theALM1gene in the leaf development stage(P0–P6) were recorded by quantitative real-time (qRT)-PCR.ALM1showed highest expression in the L4 stage of leaf development,which is also the early stage of chloroplast development.TheALM1expression level then decreased (Fig.4B).

Genevestigator microarray analysis showed thatALM1was more highly expressed in young leaves.For confirmation of these findings,RNAs were selected from several parts of wild-type plants for qRT-PCR analysis.TheALM1gene was expressed mainly in young leaves followed by old leaves and was weakly expressed in roots and panicles(Fig.4C).This expression pattern was consistent with thealm1mutant phenotype.

ChloroP analysis showed thatALM1contained a signal peptide located in chloroplasts,suggesting that theALM1gene is located in the chloroplast.For confirmation,the transiently transformed ALM1-GFP vector was constructed and introduced into rice protoplasts.The fluorescence ofALM1coincided with the spontaneous fluorescence of chloroplasts,further confirming the chloroplast localization of theALM1gene (Fig.4D–G).

3.7.Expression level of nuclear-encoded genes associated with photosynthesis decreased in alm1

Fig.4.Tissue expression and subcellular localization of ALM1.(A) Schematic illustration of a rice plant with a fully expanded third leaf.L1,L2,L3,and L4 indicate the first,second,third,and fourth leaves;developmental stages(P0–P6)are also indicated.SB(shoot base)corresponds to a 5-mm piece from the bottom of the shoot and contains asyet unmerged leaves at stages P0–P3.(B)qRT-PCR analysis of ALM1 in the SB,L1,L2,L3,and L4 tissues of wild-type seedlings.The SD was calculated from three independent experiments.(C)qRT-PCR analysis of ALM1 in the root,young leaf,old leaf,culm,and panicle tissues of wild-type seedlings.(D–G)Rice protoplasts expressing the ALM1-GFP fusion protein.(D) GFP fluorescence.(E) Chloroplast autofluorescence.(F) Bright field image of GFP.(G) Merged image of (D) and (E).Bar=5 μm.

Chloroplast development depends on the integrated expression of both nuclear and chloroplast genes and is regulated by nuclearencoded genes.To determine whether the destruction of chloroplast cell structure in thealm1mutant influenced the expression level of nuclear-encoded genes associated with photosynthesis,qPCR analysis was performed inalm1and wild-type plants.The expression level of genes(OsPOLP1andFtsZ)influencing the generation and replication of plastid DNA did not vary(Fig.5A).Expression of the photosynthesis-associated nuclear genesRNRS(encoding the large subunit of RNR) [7],V2(encoding plastidial guanylate kinase),Cab1R(encoding photosystem II chlorophylla/bbinding protein) andRpoTp(encoding NEP core subunits) [1]were greatly decreased inalm1(Fig.5B and D).The reduced expression nuclear encoded genes associated with photosynthesis may impair photosynthesis and lead to thealm1albino phenotype.

3.8.The expression levels of plastid-encoded genes are changed in alm1

Plastid genes are transcribed by Nuclear -encoded RNA polymerase(NEPs)and/or PEPs.Plastid genes can be divided into three types:classes I–III.Class I genes are transcribed mainly by PEPs,class II genes are transcribed mainly by both NEPs and PEPs,and class III genes are transcribed exclusively by NEPs.The expression levels of plastid-encoded genes inalm1were measured.In class III genes(NEPs),expression levels of RNA polymerase(rpoA)increased greatly and the ribosomal protein (rps12) were slightly increased(Fig.5B).In contrast,the expression levels of class I genes includingpsaA,psbA,D1,andrbcLwere greatly decreased inalm1(Fig.5E).The expression of chlorophyll synthesis genes was greatly reduced,includingYGL1(encoding a chlorophyll synthetase),HEMA1(encoding glutamyl-tRNA reductase),CAO1(encoding chlorophyllide A oxygenase),andDVR(encoding divinyl reductase) (Fig.5C).The lower chlorophyll content ofalm1indicated thatALM1was involved in chloroplast development and indirectly influenced chlorophyll synthesis.The nuclear and plastid gene expression pattern inalm1was similar to that of PEP defective mutants such asptac2,clb19,andotp70,which show compromised plastid translation[33–35].The expression of nuclear-and plastid encoded genes suggested thatALM1influenced PEP complex activity.

3.9.ALM1 acts as one of the PEP complex members

Fig.5.Expression of nuclear and plastid genes associated with photosynthesis.Relative expression levels of plastid DNA replication genes (A),NEP-dependent genes (B),Chlorophyll biosynthesis genes(C),Nuclear-encoded RNA polymerase(NEP)associated with photosynthesis(D),plastid-encoded RNA polymerase(PEP)dependent genes(E).

To identify theALM1gene interaction partner,a rice yeast cDNA library was screened with the ALM1-BD construct.Twelve clones were verified.Three clones corresponds to the same gene (OsTrxz,LOC_Os08g29110).Yeast point-to-point hybridization confirmed the interaction of ALM1 and OsTrxz protein in SD (-Leu-Trp-His-Ade)culture medium(Fig.6A).The interaction was also confirmed by BIFC and Co-IP assay.No signal was generated by the OsTrxz-YN or ALM1-YC vectors alone,but the presence of both vectors produced a yellow signal,further confirming that the ALM1 and OsTrxz proteins were partners(Fig.6B).The Co-IP assay confirmed that ALM1 interacted with OsTrxz in rice.

Fig.6.ALM1 physically interacts with OsTrxz.(A)A yeast two-hybrid interaction assay was performed with ALM1 and OsTrxz.pGADT7 vector was used as a negative control.An ALM1-DNA binding protein was generated in the pGBKT7 vector(ALM1-BD).A OsTrxz-activation domain fusion was generated in the pGADT7 vector.(B)The full length ALM1 protein was fused to the C-terminus of YFP in pSCYCE (ALM1-YC).The full DUA1 protein was fused to the N-terminus of YFP in pVYNE (OsTrxz-YN).The BiFC assay showed that ALM1 interacts with OsTrxz in rice protoplasts.Scale bars,20 μm.(C)Co-IP assay for detection of interaction of ALM1 with OsTrxz.(D)ALM1 increases OsTrxz stability.Immunoblot assay was performed using an antibody to OsTrxz in wild type, alm1 mutant and overexpression line 1 (OE1) seedlings at the third-leaf stage.HSP90 was selected as a control.

The vectors containing the ALM1:GFP and/or OsTrxz:FLAG constructs were transformed into rice protoplasts.Anti-FLAG (recognizing OsTrxz:FLAG) pulled down ALM1:GFP from the resulting protoplasts (Fig.6C),showing that OsTrxz physically interacted with ALM1(Fig.6C).OsTrxz has been reported[36]to be a member of the PEP complex and to regulate chloroplast development in rice.Thus,ALM1 may also be one of the PEP-complex members.To investigate the interaction between the ALM1 and OsTrxz proteins,a Western blot assay was performed inalm1,the wild type,and overexpression line 1 (OE1) using anti-OsTrxz antibody at the third-leaf stage.The OsTrxz protein decreased inalm1butincreased in OE1 lines,compared with the wild type (Fig.6D).Thus,ALM1 played an important role in stabilizing OsTrxz protein in rice.

3.10.The ALM1 mutant accumulated excessive reactive oxygen in leaves

ALM1showed highest similarity to theFSD3gene ofArabidopsis,which has the function of scavenging reactive oxygen species.Given that SOD catalyzes the conversion of superoxide anion radicals (O2-) to H2O2and O2,NBT staining was used to measure the content of superoxide anion radicals (O2-) in plants.The blue regions inalm1were slightly larger than those in the wild type,indicating that more superoxide radicals accumulated in mutants(Fig.7A).In DAB staining assays,alm1showed slightly more spots than the wild type,indicating that more H2O2accumulated in mutants (Fig.7B).NBT staining and DAB staining results indicated the accumulation of excessive reactive oxygen in leaves inalm1,showing that thealm1mutant phenotype was caused not only by chloroplast development defects but also by excessive reactive oxygen in leaves.

3.11.Overexpression of the ALM1 gene confers stress tolerance at the seedling stage

Given that ALM1 is responsible specifically for eliminating excessive superoxide anion radicals (O2-) in plants,we attempted to determine whether overexpression of theALM1gene increased superoxide scavenging and thereby plant stress resistance.We selected three overexpression lines(OE1,OE2,and OE3)with moderate expression(～3 times,4 times,4.5 times)of drought response in seedlings (Fig.7C).When the plants reached the five-leaf stage in the greenhouse (～20 days old),20% PEG (drought simulation reagent) was used for the drought resistance test (Fig.8A).After 7 days of treatment,all plants showed a wilting phenotype,including the wild type.After rehydration for 10 days,almost all wildtype plants died,whereas most of the overexpression lines (OE1,OE2,and OE3)recovered(Fig.8B).The survival rates of the overexpression lines were 78%,60%,and 50% (Fig.8C).These results suggested that the stress resistance of overexpression lines improved at the seedling stage.Measurement of SOD enzyme activities suggested that OE1,OE2 and OE3 had higher SOD enzyme activities than the wild type under normal growth conditions (Fig.8D).Under PEG 4000 treatment(drought stress),SOD enzyme activities were induced and much higher in overexpression lines (Fig.8D).Expression levels ofAPX1,APX2,andCatBgenes increased significantly and CatC slightly increase (Fig.7D).Thus,overexpression ofALM1increased SOD enzyme activities,led to more scavenging of superoxide anion radical (O2-) and conferred drought stress in rice seedlings.

Fig.7.Reactive oxygen detection and expression in wild type,alm1 mutant,and overexpression lines.(A)NBT staining assay for detection of superoxide anion radicals(O2-)in wild type and mutant.(B) DAB staining assays for detection of H2O2 in wild type and mutant.(C) Relative expression of ALM1 gene in wild type, alm1 mutant and overexpression lines.(D) Reactive oxygen scavenging system marker genes ascorbate peroxidase (APX) and catalase (CAT) were quantified by qRT-PCR in wild type, alm1 mutant,and overexpression lines.

Fig.8.Tolerance of OE1-3 to PEG-induced drought stress(A)Wild type and OE1–3 seedling plants grew until the five leaf stage,and then were treated with PEG reagent(‘‘0 day”denotes before treatment)(B)After PEG treatment for 7 days,wild type and OE1–3 showed a wilting phenotype.Plants were then rehydrated for 10 days.Most wild type plants could not recover and died,whereas most of the overexpression lines (OE1,OE2,and OE3) recovered (B).(C) Survival rates of wild-type and OE1–3 plants were estimated after rehydration for 10 days.(D)SOD activities of wild-type and OE1–3 plants were measured under normal growth and drought treatment conditions.Values in C–D are presented as means±SE.Values are mean of three replicates and were compared by one-way analysis of variance and Duncan’s multiple range test.Different letters(a–c) indicate significant (P <0.05) differences between lines.

4.Discussion

An albino mutantalm1was obtained,and its function was investigated.Thealm1mutant was first found as a chloroplast developmental defective mutant,but follow-up experiments confirmed that thealm1mutant is more than a chloroplast developmental mutant.First,TEM observation verified that chloroplasts in thealm1mutant were gradually destroyed,unlike reported chloroplast development mutants.Second,although thealm1mutant phenotype first appeared after germination for 1–2 days,the chloroplast structure was destroyed in the later stage.The combination of chloroplast developmental and excessive active oxygen accumulation at the seedling stage leads to thealm1albino phenotype.Third,theALM1gene expression pattern indicated that theALM1gene was expressed mainly in the L4 stage,in which chloroplasts function normally in the wild type.These three points indicated that thealm1mutant was not only a chloroplast development defect mutant but also a promising material for studying stress.Thealm1mutant was also similar to theOscld1mutant[32].

Chloroplasts are the sites of the core redox reactions underpinning energy metabolism.Such reactions produce ROS.High levels of ROS cause damage to DNA,proteins,lipids,and carbohydrates,resulting in oxidative stress to plants.ROS signaling leads to responses by cells that enable them to adjust to changes in redox status.Recent study [37] inArabidopsisreveal that oxidation of the chloroplast NADH produces mitochondrial ROS which can activate signaling systems to modulate energy metabolism,and in certain cases can lead to programmed cell death.ROS in the chloroplast also rapidly activated GCN2 kinase.After activation,GCN2 kinase preferentially inhibited the ribosome loading of mRNAs for functions such as the chloroplast thylakoids,vesicle trafficking,and translation [38].Singlet oxygen (1O2) generated by photosystem II(PSII)can cause photo-oxidative damage of PSII.1O2was perceived by a chloroplast localized executer 1(EX1)protein and further activated chloroplast1O2-triggered posttranslational modification retrograde signaling[39].In the present study,based on the above studies,excessive ROS generation was coupled with a chloroplast development defect inalm1.

TheALM1gene was cloned by the Tail-PCR method,and cosegregation tests (GUS staining and PCR amplification) verified that thealm1mutant phenotype was caused by T-DNA insertion.Further complementary experiments and CRISPR/Cas9 tests strongly indicated that theALM1gene was responsible for the albino phenotype.TheALM1gene encodes a member of the SOD family.SOD is a key component of antioxidant enzymes in biological systems and is widely distributed in microorganisms,plants,and animals.Although superoxide dismutase has been well studied inArabidopsis,its function in monocotyledons lacks in-depth functional analysis.In the rice genome,there were two SOD family genes,which were predicted to be located in chloroplasts,and their functions were not assigned.Our study indicated that compared withArabidopsis FSD2orFSD3,theALM1gene has some similarities in function.First,ALM1,FSD2,andFSD3are located in chloroplast thylakoids.Second,lower H2O2was produced inalm1leaves andfsd2orfsd3.Reactive oxygen scavenging system marker genes,such as APX1,APX2,CatB were reduced significantly in mutants.The results indicated thatALM1,FSD2,andFSD3acted as protectors against active oxygen scavenging.Third,overexpression ofALM1andFSD2 or FSD3improved abiotic stress,although under different stress conditions(e.g.,drought,salinity,and cold).The comparison betweenALM1,FSD2,andFSD3indicated the conservative function ofALM1and homologous genes between monocotyledons and dicotyledons.

The SOD family inArabidopsishas been thought only to scavenge reactive oxygen species.FSD2/FSD3 was also assumed [40]to be a member of the PEP complex inArabidopsis,but there was a lack of detailed experimental evidence.In this study,we confirmed this hypothesis in rice.First,the quantitative PCR results showed that theALM1gene played a role mainly in the L4 stage,which is also the early stage of chloroplast development (plastid transcription and chloroplast protein synthesis stage).Second,quantitative PCR analysis of plastid-expressed genes showed that NEP genes were slightly upregulated and that PEP genes were downregulated,which is a feature of PEP complexes,similar to those in thewsp1anddua1mutants [4,41].Third,our yeast point-to-point hybrid test and BIFC experiment(ALM1 and OsTrxz interactionin vitro) verified that ALM1 interacted with the OsTrxz protein.OsTrxz has been reported to be a PEP-complex member.OsTrxz function perturbed the stability of the transcriptionally active chromosome(TAC)complex and PEP activity[33].We speculated that ALM1 functioned in TAC stability and PEP activity.TSV and WLP2 have been reported to be OsTrxz partners.WLP2encodes a putative pfkB-type carbohydrate kinase family protein and makes rice resistant to high-temperature stress [42].TSV encodes a putative plastidic oxidoreductase and protects rice from low temperature [43].Thus,OsTrxz is an important central protein that makes rice tolerant to cold or heat stress.Two cysteine residues in OsTrxz regulate the function of interacting proteins in a redox manner.Although ALM1 did not interact with TSV andWLP2,we speculated that an ALM1-OsTrxz-TSV or ALM1-OsTrxz-WLP regulatory module plays an important role in low-or hightemperature stress by regulating the redox balance in rice chloroplasts.The fine mechanism of ALM1 and OsTrxz awaits investigation.InArabidopsis,FSD2andFSD3genes can interact with each other[40].In contrast,the interaction of theALM1gene itself(ALM1-AD and ALM1-BD)was more tightly coupled than the interaction between ALM1 and SOD2 in a yeast assay,indicating thatALM1acted in the form of a homodimer rather than a heterodimer,in contrast toFSD2andFSD3(Fig.S2).The yeast assay indicated thatALM1is more important than other SOD members in chloroplast biogenesis.It is not known whether the function of other SOD genes is redundant.How the homodimer or heterodimer complex functions awaits further study.

The overexpression of SOD genes to improve plant resistance to cold,heat,salt,drought,and oxidative stress has been successful.All genes endow plants with drought tolerance and adaptability under high light intensity,and our results also suggest that theALM1gene can incease plant stress tolerance.This study demonstrates thatALM1not only participates in rice chloroplast biogenesis,but improves stress resistance by scavenging reactive oxygen.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

CRediT authorship contribution statement

Zhiguo Zhang and Pengfei Ai:conceived the original screen;Yanwei Wang:analyzed phenotypes,clonedALM1,performed the transformation test and conducted the yeast assays;Xuean Cui and Chen Deng:performed expression analysis and subcellular localization,Zhiguo Zhang:supervised and contributed to the writing.

Acknowledgments

This research was supported by Key Laboratory of Joint CAAS/IRRI Laboratory for Photosynthetic Enhancement,Biotechnology Research Institute/National Key Facility for Genetic Resources and Gene Improvement.This research was supported by the National Key Research and Development Program of China(2020YFA0907600),the Agricultural Science and Technology Innovation Program (CAAS-ZDXT2019003),and Fundamental Research Funds for Central Non-profit Scientific Institution.

Availability of data and materials

The materials used and/or analyzed in the present study are available from the corresponding author on request.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.09.013.