Ectopic expression of OsNF-YA8, an endosperm-specific nuclear factor Y transcription-factor gene, causes vegetative and reproductive development defects in rice

Bixio Niu, Jing Xu, Zhiguo E, Zhenyu Zhng, Xinming Lu, Chen Chen,b,c,*

a Jiangsu Key Laboratory of Crop Genomics and Molecular Breeding/Zhongshan Biological Breeding Laboratory, Agricultural College of Yangzhou University, Yangzhou 225009,Jiangsu, China

b Jiangsu Co-Innovation Center for Modern Production Technology of Grain Crops, Yangzhou University, Yangzhou 225009, Jiangsu, China

c Key Laboratory of Plant Functional Genomics of the Ministry of Education, Agricultural College of Yangzhou University, Yangzhou 225009, Jiangsu, China

d Key Laboratory of Rice Biology, China National Rice Research Institute, Hangzhou 311400, Zhejiang, China

Keywords:Rice OsNF-YA8 Plant height Flowering time Seed development

ABSTRACT Nuclear factor Y(NF-Y),a group of conserved transcription-factor complexes that consist of NF-YA,B,and C subunits, is essential for developmental regulation and for responses to environmental changes in eukaryotes.We previously found that some NF-Y genes, such as OsNF-YA8, were expressed specifically in the endosperm of rice.In the present study, overexpression of OsNF-YA8 in rice resulted in reduced plant height due to suppressed cell elongation in internodes.Gibberellin(GA)biosynthetic genes,including OsCPS1, OsGA20ox1, and OsGA20ox2, were down-regulated.OsNF-YA8 bound to the promoters of these genes to repress their expression.Endogenous GA content was decreased in OsNF-YA8 overexpressors, whose dwarf phenotype could be partially rescued by exogenous GA treatment.The findings suggested that ectopic expression of OsNF-YA8 causes defective GA biosynthesis in vegetative stage.Heading date in OsNF-YA8 overexpressors was delayed, especially under short-day conditions.OsNFYA8 bound to the promoter of Heading Date 3a (Hd3a), the florigen gene in rice, to negatively regulate flowering.Either ectopic activation or knockout of OsNF-YA8 impaired seed development, as indicated by reduced seed size and increased grain chalkiness.These results suggest that ectopic expression of the endosperm-specific OsNF-YA8 in rice disrupts both vegetative and reproductive development.

1.Introduction

Nuclear factor Y (NF-Y)transcription factor (TF), also known as the Heme activator protein (HAP), is highly conserved in eukaryotes [1].NF-Y is a heterotrimeric complex composed of three subunits:NF-YA,NF-YB,and NF-YC[2].Each NF-Y subunit usually has only one or two encoding genes in yeast and animals,but multiple encoding genes in plants [3–5].The rice (Oryza sativa) genome encodes 11 OsNF-YAs, 11 OsNF-YBs, and 13 OsNF-YCs [4].NF-Y TFs function in plant development, stress response, and flowering regulation [6–10].We previously [4] found that a group of NF-Y TFs were expressed predominantly in the endosperm of cereal species including rice, maize,barley,and sorghum.Many members of this group of genes in rice, including OsNF-YB1, OsNF-YB9, OsNFYC10, and OsNF-YC12, have been found to regulate storage compound accumulation in the endosperm [11–16].OsNF-YB7 and OsNF-YB9, which preferentially express in rice seed, are homologous to LEAFY COTYLEDON1 (LEC1), a master regulator of seed development in Arabidopsis [16].Overexpression of either OsNFYB7 or OsNF-YB9 caused dwarfing with aberrant spikelet development[16–19],suggesting that their ectopic expression may disrupt both vegetative and reproductive development of plants.But the underlying mechanisms await elucidation.

NF-Y TFs are essential for the vegetative–reproductive transition in many plant species.OsNF-YB11 (also known as DAYS TO HEADING 8,DTH8)accelerates heading under short-day(SD)conditions in rice by modulating the expression of Heading date 1(Hd1),an ortholog of Arabidopsis CONSTANS(CO).In contrast,DTH8 interacts with Hd1 for the transcriptional regulation of Hd3a,a florigenencoding gene of rice, under long-day (LD) conditions [20–22].Both OsNF-YB9 and OsNF-YC12 interact with rice GIGANTEA(OsGI), a key regulator of the expression of Hd1 and Early heading date 1 (Ehd1) [19].Ectopic expression of OsNF-YB10 and OsNF-YB8 repressed heading in rice [6], whereas OsNF-YC2 and OsNF-YC4 specifically affect flowering under LD by disturbing the flowering genes Ehd1, Hd3a, and Rice Flowering Locus T1 (RFT1) [23].

Plant height affects crop architecture, apical dominance, biomass, and mechanical harvesting [24].The Green Revolution increased the yield potential of rice, in part by breeding of semidwarf cultivars [25,26].It is desirable to characterize networks responsible for plant height regulation.Many genes controlling rice plant height are involved mainly in biosynthesis and signal transduction of phytohormones such as gibberellic acids (GAs), brassinosteroids, and strigolactones [24].GA intermediates are synthesized in several steps from geranylgeranyl diphosphate,which is converted to ent-kaurene by the ent-copalyl diphosphate synthase(CPS)and ent-kaurene synthase(KS).Ent-kaurene is converted to GA12by two cytochrome P450 enzymes,ent-kaurene oxidase (KO) and ent-kaurenoic acid oxidase (KAO).GA12is then converted to bioactive GA1or GA4by gibberellin 20-oxidase(GA20ox)and GA3ox[27,28].GA-deficient and insensitive mutants are usually much shorter than the wild type (WT) [29–31].The reduced plant height of semi-dwarf 1 (sd1) mutants is the result of impaired GA biosynthesis in the elongating stem, owing to defective OsGA20ox2 [26,32–34].

Among the endosperm-specific NF-Y genes of rice that we previously identified [4], OsNF-YA8 has not been functionally characterized.NF-YA TFs contain two functional domains: the interaction domain for NF-YA and NF-YB/NF-YC interaction, and the DNA-binding domain for recognizing downstream targets [2].In Arabidopsis, overexpression of NF-YA2/4/7/10 led to dwarf phenotypes [35,36]; whereas overexpression of NF-YA1, 5, 6, and 9 affected male gametogenesis, embryogenesis, seed morphology,and seed germination [37].Null mutations of NF-YA1 and NF-YA2 were lethal [38].NF-YA8 bound to the CCAAT motifs of multiple MIR156 genes,thereby activating their transcription and inhibiting the juvenile-to-adult transition in Arabidopsis [39].In a rice OsMIR168a mutant, OsNF-YA8 was up-regulated and the juvenileto-adult transition was accelerated [40].In this study, we characterized the biological significance of OsNF-YA8 in rice by generating overexpression and knockout mutant lines.

2.Materials and methods

2.1.Plant materials,growth conditions,and phenotyping of agronomic traits

Kitaake(O.sativa ssp.geng)and the transgenic lines used in this study were grown in the experimental field of Yangzhou University(latitude 32°24′N, longitude 119°26′E) under natural conditions from May to October,and in the experimental field in Lingshui(latitude 18°22′N, longitude 109°45′E), Hainan province, China, from December to May of the following year.The osnf-ya8 mutants used in the study were generated in our previous study[4].In the greenhouse, plants were grown under well-controlled 12 h/12 h day/ni ght conditions with day/night temperatures of 32 °C/ 28 °C.

Plant height at heading stage was measured from the basal node to the flag leaf of the main tiller.Plant height at maturity was measured from the basal node to the top of the main panicle.Heading was measured in days from sowing to emergence of the first panicle.Three replicates were grown, with each replicate including at least 10 plants for phenotyping.Dehulled mature grains were used to measure grain length, width, thickness,1000-grain weight, and chalkiness rate, using a scanning system as described previously [16].

2.2.Histological analysis

Second internodes at heading stage were collected from WT(Kitaake) and oYA8OEfs (OsNF-YA8 overexpressors driven by the rice Ubiquitin promoter and tagged with 3 × FLAG (DYKDDDDK)),then fixed in FAA solution(60%(v/v)ethanol,5%(v/v)glacial acetic acid and 5% (v/v) formaldehyde) and vacuum evacuation for 40 min.Next,the internodes were dehydrated in an ethanol series as previously described [16], 15 min for each step, and then embedded into resin using the Technovit H7100 embedding kit following the manufacturer’s instruction.The samples were sliced in 3-μm thick-sections using a microtome (Leica RM2265), and stained with toluidine blue O for microscopic observation and photographing.The ImageJ package was used for cell length measurement.

2.3.Vector construction and plant transformation

The OsNF-YA8 coding sequence excluding the stop codon was amplified and recombined into the vector p1300U1-FLAG (L177),using the ClonExpress II One Step cloning kit (Vazyme), for generating oYA8OEf transgenic plants.The p1300U1-FLAG (L177) backbone contains a rice ubiquitin promoter to drive expression of the inserted gene.To generate mYA8OE transgenic plants, OsNF-YA8 coding sequence excluding the stop codon was inserted into pENTR/D-TOPO and then cloned into the pANIC6A vector with a Gateway cloning kit (Invitrogen, 12538120).The maize ubiquitin promoter embedded in the backbone of pANIC6A was used to drive OsNF-YA8.The constructs were introduced into Agrobacterium tumefaciens strain EHA105 for rice transformation as previously described [16].Primers used in the study are listed in Table S2.

2.4.RNA extraction and quantitative Real-time PCR(qRT-PCR)analysis

Total RNA from various tissues at seedling and heading stages was isolated using the RNA-easy Isolation Reagent kit (Vazyme).Then, 1 μg of total RNA was used for cDNA synthesis with a First Strand cDNA Synthesis kit (Vazyme).SYBR qPCR Master Mix(Vazyme) was used for qRT-PCR reactions on a CFX Connect Realtime System(BioRad).At least three biological replicates were used for each reaction.The rice Actin gene(Os03g0718100)was used as the internal control.Oligonucleotide primers are listed in Table S2.

2.5.RNA-sequencing and data analysis

Total RNA was extracted from the main stem of WT,oYA8OEf3,and oYA8OEf5 plants at heading stage using the plant RNAisolation kit (TIANGEN) following the manufacturer’s procedure.The RNA was submitted to Beijing Genomics Institute (BGI, Shenzhen, Guangdong, China) for library preparation and sequencing.CLC genomics Workbench 12.0 (Qiagen) was used for identifying differentially expressed gene(DEG),using the thresholds|log2fold change|>1 and P-value<0.05.The short reads were mapped to the Nipponbare genome (MSU 7.0).Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed using an online resource (https://www.omicshare.com/tools) with the default settings.

2.6.Hormone quantification and treatment

Content of GA4was measured in leaves of 2-week-old seedlings of WT, oYA8OEf3, and oYA8OEf5.Fresh seedlings were harvested and immediately ground into powder using liquid nitrogen.Three biological replicates were used for GA4quantification.Each replicate comprised at least five individuals with same genotype.The measuring procedures were as previously described [41].

For GA treatment,one-week-old seedlings of WT and oYA8OEf3 were sprayed with 100 μmol L-1GA3for 5 d.Three biological replicates were used,and at least 10 plants for each replicate were collected for recording plant height.

2.7.Dual-luciferase assay

The full coding sequence of OsNF-YA8 was cloned into the effector vector pGreenII 62-SK, using the ClonExpress II One Step cloning kit (Vazyme).The promoter sequences of OsCPS1, OsGA20ox1,and OsGA20ox2 were amplified from the rice genome and inserted separately into the reporter vector pGreenII 0800-LUC.Multiple reporter and effector combinations were transformed into rice protoplasts for transient expression.Rice protoplasts were prepared as previously described [42].Firefly luciferase (LUC) and renilla luciferase (REN) activities were measured using the Dual Luciferase Reporter Assay kit (Vazyme) with spectrofluorometer (Tecan Infinite M200 Pro).Primers used for the assay are listed in Table S2.

2.8.Chromatin immunoprecipitation coupled with quantitative PCR(ChIP-qPCR)

Dehulled seeds of oYA8OEf3 were sown on ½-strength Murashige and Skoog agar medium.Two-week-old seedlings were vacuum-infiltrated with 1% (v/v) formaldehyde for cross-linking.The plant materials were thoroughly ground in liquid nitrogen after quenching the cross-linking process.Chromatin complexes were isolated, sonicated,and incubated with anti-FLAG antibodies(Sigma, 0480), as described previously [43].The precipitated DNA was quantified by qRT-PCR.The values of qRT-PCR were normalized to the input sample to obtain the enrichment fold.Samples immunoprecipitated with Immunoglobulin G (IgG) were used as a negative control.Three biological replicates were used for each analysis.Primer pairs used for analysis are listed in Table S2.

2.9.Yeast two-hybrid assay (Y2H)

The coding sequences of OsNF-YC4 and OsNF-YC7 were cloned into the prey vector pGADT7 (Clontech), and the coding sequence of OsNF-YA8 was cloned into pGBKT7(Clontech),using the ClonExpress II One Step cloning kit(Vazyme).The prey and bait plasmids were cotransformed into the yeast strain Y2HGold and plated on SD/-Leu-Trp medium for 3 d at 30 °C.Interactions between bait and prey were further tested on SD/-Trp-His-Leu-Ade selective medium as previously described [4].

2.10.Co-immunoprecipitation (Co-IP) assay

The coding sequences of OsNF-YA8 and OsNF-YC4 were cloned into pCAMBIA1300-FLAG and pJIT163-GFP, respectively.The constructions were then co-transformed into rice protoplasts for transient expression.Total protein was extracted from the rice protoplasts using ice-cold NP-40 buffer(with 1%phenylmethylsulfonyl fluoride(PMSF))and incubated with GFP-Trap agarose beads(ABclonal)for 2 h at 4°C with agigation.After washing with NP-40 buffer(with 1%PMSF)for five times,the immunoprecipitated samples,together with the inputs,were boiled in 5×loading buffer for sodium dodecylsulfate polyacrylamide gel electrophoresis (SDSPAGE).Immunoblotting analyses were then performed using anti-FLAG (Sigma, F3165) and anti-GFP (Abclone, AE012) antibodies, respectively.

2.11.Bimolecular fluorescence complementation (BiFC) assay

The coding sequences of OsNF-YA8, OsNF-YC4, and OsNF-YC7 were separately cloned into pYCE or pYNE vectors[16].These constructs were introduced into Agrobacterium strain GV3101 and then used to infiltrate Nicotiana benthamiana leaves as described previously [16].Fluorescence signals were captured with a Zeiss LSM700 laser scanning confocal microscope.

3.Results

3.1.Ectopic expression of OsNF-YA8 reduced plant height

Previously, we found that OsNF-YA8 was expressed preferentially in the endosperm [4].To investigate the gene’s function in rice, OsNF-YA8 overexpression lines driven by the maize Ubiquitin promoter (designated as mYA8OEs hereafter) were generated.Of four independent mYA8OE transgenic lines, all showed ectopic activation of OsNF-YA8 in leaves of transformants (Fig.S1A).In comparison to WT Kitaake, plant height of the mYA8OEs was significantly reduced, at either heading or maturing stages(Fig.S1B–D).We also overexpressed OsNF-YA8 tagged with 3 × FLAG, driven by the rice Ubiquitin promoter (designated as oYA8OEf hereafter).Of five independent oYA8OEf transformants obtained, only three (oYA8OEf3, 5, and 6) produced viable seeds for further analyses.The qRT-PCR analysis indicated that oYA8OEf3 and oYA8OEf5, but not oYA8OEf6, showed activation of OsNF-YA8 in leaves(Fig.1A),which was further confirmed by a Western blotting assay using an anti-FLAG antibody (Fig.1A).oYA8OEf3 and 5 displayed a more severe dwarf phenotype than the mYA8OE lines,whereas oYA8OEf6 showed no marked developmental defects(Fig.1B, C).Decreased plant height of oYA8OEfs was detectable as early as seedling stage(Fig.1D).oYA8OEf3 and 5 showed respectively 49.6%and 42.4%reduction of height at two weeks after germination (Fig.1E).In addition to reduced plant stature, oYA8OEf3 and 5 showed shortened panicles as well as reduced seed setting at maturation (Fig.S2).

To investigate why the oYA8OEfs had a more severe phenotypic abnormality than mYA8OEs, we compared the OsNF-YA8 expression levels in these lines.OsNF-YA8 transcripts were more abundant in oYA8OEf than in mYA8OE lines, indicating that OsNF-YA8 represses vegetative development, probably in a dosagedependent manner(Fig.S3).To confirm this inference,we investigated the progeny of the oYA8OEf3 T0plant.Three distinct plant height phenotypes were observed in the segregating population:severely reduced, moderately reduced, and normal.The severity of dwarfing was positively correlated with the abundance of OsNF-YA8 in leaves(Fig.1F).Seedlings with normal height showed no activation of OsNF-YA8 (Fig.1F).Seedlings showing the most severe dwarfing were not viable.We inferred that the failure of some of the oYA8OEf transformants to produce viable seeds was due to over-activation of OsNF-YA8.

3.2.Ectopic activation of OsNF-YA8 reduced internode number and repressed cell elongation

To identify the cause of reduced plant height of the OsNF-YA8 overexpressors, we measured total internode numbers and length of each internode of the main stem at maturation.The oYA8OEf lines presented fewer internodes than WT plants: >72% of the WT we observed had five internodes, whereas oYA8OEf3 plants had only four (Fig.2A).The mean length of the 5th internode of WT was 1.65 cm, accounting for ～18.9% of the length difference between oYA8OEf3 and the WT.The lengths of oYA8OEf3 internodes were consistently shorter than those of the corresponding internodes of the WT(Fig.2A,B).The shortened-internode phenotype was also observed in mYA8OEs, although their internode numbers were similar to that of the WT (Fig.S1E, F).

Fig.1.Overexpression of OsNF-YA8 caused dwarfing in rice.(A) Relative expression level (upper) and protein accumulation (lower) of OsNF-YA8 in oYA8OEfs and WT(Kitaake)leaves.The transformant oYA8OEf6 showed no overexpression of OsNF-YA8 and accordingly was used as a negative control.Error bars indicate SD(n=3).(B)Plant morphology of WT and oYA8OEfs at heading stage.Scale bar,10 cm.(C)Plant height of WT and oYA8OEfs at heading stage.Error bars indicate SD(n=30).Student’s t-test was used for statistical analysis.**, P < 0.01.(D) Plant morphology of 2-week-old WT and oYA8OEf seedlings.Scale bar, 5 cm.(E) Plant height of 2-week-old WT and oYA8OEf seedlings.Error bars indicate SD (n = 20).Student’s t-test was used for statistical analysis.**, P < 0.01.(F) Three distinct phenotypes of progeny of the oYA8OEf3 T0 plant(upper panel).The phenotypic abnormality of the segregating progeny showed correlation with the expression level of OsNF-YA8 in seedlings(lower panel).Scale bar,2 cm.Error bars indicate SD (n = 3).

To determine whether the dwarf phenotype was contributed by reduced cell size or cell number, we compared longitudinal sections of the second internode,which showed the most striking difference between the WT and oYA8OEfs.Cell length of oYA8OEfs was significantly reduced (Fig.2C, D), whereas cell number was not significantly changed (Fig.2E).These findings suggested that ectopic expression of OsNF-YA8 in rice represses cell elongation,thereby reducing plant height.

3.3.Impaired GA biosynthesis in OsNF-YA8 overexpression lines

We next analyzed transcriptome changes in the main stems of oYA8OEf3 and 5 at heading stage, employing a RNA sequencing(RNA-seq)approach.Respectively 3646 and 2812 DEGs were identified in oYA8OEf3 and 5,using the thresholds fold change>2 and P < 0.05 (Table S1).Among the 2010 common DEGs, 991 were upregulated and 1019 down-regulated in oYA8OEfs (Fig.3A;Table S1).Many biological GO terms, including transmembrane transport, response to abiotic stimulus, cell cycle, and metabolism were significantly enriched (Fig.S4B).As revealed by KEGG pathway analysis, DNA replication, biosynthesis of carotenoid, phenylpropanoid, monoterpenoid and diterpenoid, starch and sucrose metabolism, homologous recombination, and circadian rhythm pathways were enriched (Fig.3B), indicating that these pathways were disrupted by ectopic activation of OsNF-YA8.

GAs are a large family of diterpenoid compounds[27].In agreement with the enrichment of the pathway of diterpenoid biosynthesis, we found that the GA biosynthetic genes OsCPS1,OsGA20ox1, OsGA2ox9, OsGA2ox5, OsGA2ox6, and OsGA2ox8 were down-regulated in oYA8OEf3 and 5 (Fig.S4B; Table S1).Consistently, qRT-PCR analysis showed that OsCPS1, OsGA20ox1,2, and OsGA2ox6,7 were significantly repressed in both transgenic lines;whereas OsGA2ox3 was upregulated (Fig.3C, D).These findings suggested that the dwarf phenotype of oYA8OEfs was induced by GA biosynthesis disruption.The endogenous GA4content was reduced in both oYA8OEf3 and 5 (Fig.4A).To further determine whether the dwarf phenotype of oYA8OEfs was caused by GA deficiency, we spray-treated 1-week-old WT and oYA8OEf3 seedlings with 100 μmol L-1bioactive GA3.In response to the treatment,oYA8OEf3 reached a height comparable to that of the WT at five days after treatment (Fig.4B, C), indicating that exogenous GA treatment alleviated the dwarf phenotype of the OsNF-YA8 overexpression plants.

3.4.OsNF-YA8 acts as a transcriptional repressor to hinder GA biosynthesis

Fig.2.Epitopic expression of OsNF-YA8 reduced internode cell length.(A) Images showing the main stems (left) and individual internodes of WT and oYA8OEf3.The internodes were counted from the uppermost to the lowest ones.Scale bar,2 cm.(B)Lengths of each internode and their sums of the WT and oYA8OEf3.Error bars indicate SD(n=10).Student’s t-test was used for statistical analysis.*,P<0.05;**,P<0.01.(C)Longitudinal sections of the second internodes of WT and oYA8OEfs.Arrows indicate the cells used for cell length measurement.Scale bar,100 μm.(D,E)Cell length(D)and cell number(E)of the second internodes of WT and oYA8OEfs as illustrated in(C).Error bars indicate SD (n = 3), at least 50 cells were measured for each biological replicate.Student’s t-test was used for statistical analysis.**, P < 0.01.

Cis-element analysis showed that there are several CCAATboxes,which are recognizable by NF-Y-TFs[1,44],in the promoter regions (2000 bp upstream of the start codon) of OsCPS1,OsGA20ox1,and OsGA20ox2.Because these genes were significantly down-regulated in oYA8OEfs (Fig.3D), we hypothesized that they were downstream targets of OsNF-YA8.To test this hypothesis, ChIP-qPCR assay was performed using an anti-FLAG antibody.In comparison with the IgG negative control, the binding of OsNF-YA8 was significantly enriched in the promoters of OsCPS1, OsGA2ox1, and OsGA2ox2 (Fig.5A).These findings suggested that OsNF-YA8 binds the promoters of these GA biosynthetic genes.

We next performed a dual luciferase reporter assay.The OsCPS1,OsGA20ox1, and OsGA20ox2 promoters were separately inserted into the reporter vector to drive expression of the LUC reporter;OsNF-YA8 was inserted into the effector vector (Fig.5B).We transiently co-expressed the effector with the reporters in rice protoplasts.OsNF-YA8 appeared to act as a transcriptional repressor to suppress the expression of OsCPS1, OsGA20ox1, and OsGA20ox2(Fig.5C).Together, the findings suggested that OsNF-YA8 directly targets multiple rice GA synthetic genes to repress their expression.

3.5.Overexpression of OsNF-YA8 delayed heading date

Previous studies showed that NF-YA, in complex with NF-YB/NF-YC proteins, may influence photoperiod-dependent flowering[6,45–47].The transcriptomic profiling revealed that the expression of Hd3a and Hd1 were significantly reduced in oYA8OEf3 and 5, whereas RFT1 and DTH8 were activated (Figs.6A, S4C).Consistently, under well-controlled SD conditions, the heading date of oYA8OEfs was delayed for ～12 days relative to the WT.Similar results were obtained when the plants were grown in Lingshui under natural SD conditions(Fig.6B, C).However, only slight difference could be observed under natural LD conditions in Yangzhou:the heading date of WT was 50.5 d and that of oYA8OEf3 was 54.7 d on average.

ChIP-qPCR assay showed that OsNF-YA8 was able to bind to the promoter of Hd3a (Fig.6D).A dual luciferase assay showed that,when OsNF-YA8 was used as the effector, expression of the LUC reporter was significantly reduced (Fig.6E, F), suggesting that OsNF-YA8 regulated photoperiodic flowering in rice by repressing Hd3a.

3.6.OsNF-YA8 overexpression led to reduced grain filling

OsNF-YA8 is expressed predominantly in developing seeds [4].Overexpression of OsNF-YA8 led to down-regulation of many starch and sucrose metabolism genes highly expressed in endosperm(Fig.S4D).In comparison with the WT, grain development of oYA8OEfs was significantly changed, with grain size and grain weight diminished(Fig.7A–D).oYA8OEfs displayed a higher chalky grain percentage and chalky degree of the endosperm than the WT control(Fig.7E,F).In case that the seed phenotype was caused by the severe dwarfing of oYA8OEf plants,we investigated seed development in weak overexpression lines driven by maize Ubiquitin promoter.Seeds of mYA8OE showed no width difference,but were significantly longer and thinner than WT seeds (Fig.S5A–C).mYA8OE2 and 4 showed significantly reduced grain weight, while those of mYA8OE1 and 3 were comparable to the WT.Grain of mYA8OEs displayed a higher rate of chalkiness than that of oYA8OEfs (Fig.S5B, E).

Fig.3.Transcriptome analysis of OsNF-YA8-regulated genes.(A)Venn diagram showing differentially expressed genes(DEGs)identified in stems of oYA8OEf3 and oYA8OEf5 at heading stage.(B)KEGG enrichment of common DEGs identified in oYA8OEf3 and oYA8OEf5.The x-axis indicates the degree of enrichment.The y-axis indicates pathway terms.Dot size and color indicate the numbers of genes enriched in the pathways and the P-value,respectively.(C)An illustrative diagram showing the process of gibberellin(GA)biosynthesis in rice.(D)Relative expression of the indicated GA biosynthetic genes in internodes of oYA8OEf lines and WT.Error bars indicate SD(n=3).Student’s t-test was used for statistical analysis.**, P < 0.01.

Two null mutant lines with 2-and 4-bp deletions we generated previously [4] were used for the next investigation (Fig.S6A).The homozygous osnf-ya8 displayed no vegetative phenotype(Fig.S6B,C).However, osnf-ya8 plants showed increased grain length but reduced grain thickness (Fig.S6D, E).The grain weight of osnfya8 was significantly reduced (Fig.S6F).Seeds produced by osnfya8 mutants showed higher chalkiness (Fig.7G–J).The qRT-PCR analysis showed that several genes involved in sucrose and starch metabolism displayed misregulation in osnf-ya8 caryopses.Expression of Sucrose Phosphate Synthase1 (SPS1), Branching Enzyme 1(BE1),BEIIa,Soluble Starch Synthase IIIa(SSIIIa),and Rice Starch Regulator 1 (RSR1) was significantly up-regulated (Fig.S7).Together,these results suggested that OsNF-YA8 might function in accumulation of storage substances during grain-filling.

3.7.Ectopic expression of OsNF-YA8 affected the expression of other NF-Ys in rice

Fig.4.Application of GA alleviated the dwarf phenotype of oYA8OEf.(A)Quantification of endogenous GA4 contents in WT and oYA8OEfs at seedling stage.Error bars indicate SD (n = 3).Student’s t-test was used for statistical analysis.*, P < 0.05; **, P < 0.01.(B) Plant morphology of GA3-treated WT and oYA8OEf3 seedlings.Scale bar, 2 cm.(C)Heights of GA3-treated seedlings on 5 days after treatment (DAT).Error bars indicate SD (n = 10).

Fig.5.OsNF-YA8 bound to the promoters of OsCPS1,OsGA20ox1,and OsGA20ox2 for gene regulation.(A)ChIP-qPCR analysis showing the ability of OsNA-YA8 to bind to the promoters of OsGA20ox1, OsGA20ox2, and OsCPS1.Positions of the probes used for analysis are illustrated in the upper panels.The intron of OsCPS1 was used as a negative probe for ChIP-qPCR analysis.Error bars indicate SD(n= 3).Student’s t-test was used for statistical analysis.**, P <0.01.(B) Schematic diagram of the reporter and effector constructs for the transient dual-LUC assays.(C)Transient dual dual luciferase assay showing that OsNF-YA8 repressed OsGA20ox1,OsGA20ox2,and OsCPS1 in rice protoplasts.Error bars indicate SD (n = 3).Student’s t-test was used for statistical analysis.**, P < 0.01.

OsNF-YA8 is expressed exclusively in developing seed[4].Ectopic expression of OsNF-YA8 might form unusual NF-Y complexes in vegetative tissues and cause pleiotropic effects on plant development.Y2H and luciferase complementation imaging experiments showed that OsNF-YA8 interacted with OsNF-YC4 and OsNF-YC7,two NF-YC members expressed predominantly in vegetative tissues,either in yeast or in plant(Fig.8A,B).BiFC assay showed that interactions between OsNF-YA8 and OsNF-YC4/7 occurred predominantly in the nuclei of epidermal cells in N.benthamiana.We next performed Co-IP assays to test whether the interaction occurs in vivo.OsNF-YA8 tagging with 3 × FLAG (OsNF-YA8-FLAG) was co-expressed with OsNF-YC4 tagging with GFP (OsNFYC4-GFP) in rice protoplasts.OsNF-YA8-FLAG was coimmunoprecipitated by OsNF-YC4-GFP, but not by GFP (Fig.8C).We also examined the transcription level of other NF-Ys in OYA8OEf3 plants.The qRT-PCR results showed that OsNFYA1,3,4,5,7,10, OsNF-YB2,4,8,9 and OsNF-YC1,9 were strongly activated, whereas OsNF-YC3 and 5 were repressed (Fig.8D), suggesting that ectopic expression of OsNF-YA8 caused misregulation of NF-Ys in leaf.

Fig.6.Ectopic expression of OsNF-YA8 in rice altered heading date.(A)Relative expression of Hd3a,Hd1,DTH8,and RFT1 in WT and oYA8OEfs grown under controlled shortday conditions in greenhouse.Error bars indicate SD(n=3).Student’s t-test was used for statistical analysis.*,P<0.05;**,P<0.01.(B)Phenotypes of WT and oYA8OEfs grew under a controlled short-day condition in greenhouse.Arrows indicate emerging panicles.Scale bar, 10 cm.(C) Heading date of oYA8OEfs and WT under natural short-day conditions.Error bars indicate SD(n=10).Student’s t-test was used for statistical analysis.**,P<0.01.(D)ChIP-qPCR analysis showing the ability of OsNA-YA8 to bind to the promoter of OsHd3a.Positions of the probes used for analysis are shown in the upper panel.Error bars indicate SD(n=3).Student’s t-test was used for statistical analysis.**,P<0.01.(E)Schematic diagram of the reporter and effector constructs for the transient dual-LUC assay.(F)Transient dual luciferase assay showing that OsNF-YA8 repressed Hd3a in rice protoplasts.Error bars indicate SD (n = 3).Student’s t-test was used for statistical analysis.**, P < 0.01.

4.Discussion

4.1.OsNF-YA8 affects rice plant height via the gibberellic acid pathway

The control of rice height depends mostly on the biosynthesis and signaling of phytohormones [29,30,48–55].Accumulation of GA is well-controlled by complicated regulation networks in rice.For example, the TF Rice YABBY1 (OsYAB1) binds to the promoter of GA3ox2 to suppress its expression [56]; OsNAC2 negatively regulates plant height by repressing KO2 [53]; whereas OsWOX3A down-regulates GA biosynthesis by repressing KAO [49].Knockdown of the MADS-box TF family member, OsMADS57, reduced plant height in rice.OsMADS57 directly regulates the transcription of OsGA2ox3 and Elongated Upmost Internode1 to regulate GA metabolism [54].

In Arabidopsis, NF-YC3 and 9 interact with both CO and RGA, a DELLA in the gibberellin pathway, to directly regulate the transcription of SOC1 (Suppressor of Overexpression of Constans1) [46].In conifers,NF-YC1/4 and DPL(DELLA protein of Pinus tabuliformis)interacted with each other, suggesting that NF-YC1/4 might be involved in the GA signaling pathway [56,57].Overexpression of NF-YAs typically leads to severe growth retardation and developmental defects with hypersensitivity to abscisic acid (ABA) in Arabidopsis [37,58].The NF-YC–RGL2–ABI5 module integrates GA and ABA signaling pathways during seed germination[59].The findings suggested that GA and ABA have antagonistic effects in various developmental phases.

Fig.7.OsNF-YA8 regulated seed development.(A, B) Appearance of brown (A) and polished rice (B) of WT and oYA8OEfs.Scale bar, 1 cm.(C–F) Grain size (C), 1000-grain weight(D),Chalky grain percentage(E)and chalky degree(F)of WT and oYA8OEfs seed.Error bars indicate SD(n=3).Each biological replicate consisted of at least 100 seeds for phenotyping.**,P<0.01 by Student’s t-test.(G)Appearance of polished rice produced by the WT and osnf-ya8 mutants.Scale bar,1 cm.(H)Cross-sections of brwon seeds produced by the WT and osnf-ya8 mutants.(I,J) Chalky grain percentage (I) and chalky degree (J) of the WT and osnf-ya8 mutant seeds.Error bars indicate SD (n =3).Each biological replicate consisted of at least 100 seeds for phenotyping.**, P < 0.01 by Student’s t-test.

Our findings suggested that the OsNF-YA8-induced growth retardation in rice was due to GA biosynthetic defect.A series of GA biosynthetic genes were down-regulated in OsNF-YA8 overexpressors (Figs.3D, S4B).Reduced content of the bioactive GA4in oYA8OEfs and phenotypic recovery by the exogenous GA treatment further confirmed that OsNF-YA8 affects plant height by interfering with GA biosynthesis (Fig.4A–C).Likewise, a recent study [60]showed that quintuple mutant of the rice seed-specific OsNF-YC genes OsNF-YC8,9,10,11,12 displayed slower germination owing to GA deficiency.

4.2.Ectopic expression of OsNF-YA8 delayed rice heading date

In rice, Hd1 positively regulates Hd3a expression to promote flowering under SD [61], whereas RFT1 is required for flowering under LD [62].The NF-Y family member DTH8 plays a important role in rice flowering: DTH8 accelerates heading under SD by controlling the expression of Hd1; in contrast, it interact with Hd1 to mediate the transcriptional regulation of Hd3a under LD [20–22].NF-Y TFs also regulate flowering in other plant species.Systematic studies [9,63] in Arabidopsis have revealed that AtNF-YC3/4/9 are required for CO-mediated photoperiod-dependent flowering.In the present study, we found that ectopic expression of OsNF-YA8 delayed flowering under SD by repressing Hd3a (Fig.6).The expression of Hd1 was also suppressed, whereas that of DTH8 and RFT1 was up-regulated in oYA8OEfs (Fig.6A).

In comparison with WT, OsNF-YA8 ectopic expression showed limited effects on flowering under LD.This differs from the OsNF-YC2,4,6- and OsNF-YB8,10,11-mediated flowering regulations, which function specifically under LD, via modulating Ehd1,Hd3a, and RFT1, with assistance of the OsNF-YB8/10/11 proteins[6].Because OsNF-YA8 is barely expressed in tissues other than rice seeds,we believe that the pleiotropic effects induced by the ectopic expression of OsNF-YA8, such as delayed flowering, cannot reflect the true function of the gene in rice.Instead, the phenotype was most likely induced by abnormal NF-Y complexes that formed in oYA8OEfs.Ectopically expressed OsNF-YA8 can compete with the other NF-Y members to form unusual NF-Y complexes that are not present in WT to disrupt the heading regulation mediated by normal NF-Y TFs.In agreement, we found that OsNF-YA8 is able to interact with OsNF-YC4 and 7 in vitro and in vivo (Fig.8A–C).In addition to OsNF-YA8, ectopic expression of other seedpreferential NF-Y members, such as OsNF-YB7 and OsNF-YB9, may also cause similar developmental defects [17–19].Moreover, activation of OsNF-YA8 in vegetative tissues interrupted the expression of other OsNF-Y genes (Fig.8C), which may also contribute to the delayed flowering in oYA8OEfs.

4.3.OsNF-YA8 functions in grain development of rice

We previously [4] identified a group of NF-Y genes, including OsNF-YA8, OsNF-YB1,9 and OsNF-YC8,9,10,11,12, preferentially expressed in rice endosperm.Recent studies[11–15,19,60,64]have revealed that most of these genes are necessary for rice endosperm development and grain filling.Either overexpression or knockout of OsNF-YA8 reduced grain thickness and chalkiness (Figs.7, S5,S6), resulting in a phenotype like that of some other endospermspecific NF-Y mutants.

Fig.8.Ectopic expression of OsNF-YA8 impaired expression of other NF-Ys in rice.(A)Yeast two-hybrid assay showing interactions between OsNF-YA8 and OsNF-YC4/7.The indicated construct combinations were cotransformed into yeast.The transformants with gradient dilution were grown on double (SD/-T-L) and quadruple (SD/-T-L-H-A)selective medium.(B) Luciferase complementation imaging assay showing interactions between OsNF-YA8 and OsNF-YC4/7 in Nicotiana benthamiana.nLUC and cLUC indicate respectively N and C termini of luciferase; NF-YA8-nLUC indicates OsNF-YA8 recombinant protein tagging with nLUC; NF-YC4/7-cLUC indicated OsNF-YC4/7 recombinant protein tagging with cLUC.The indicated construct combinations were transiently coexpressed in epidermal cells of N.benthamiana.(C) Validation of the interaction between OsNF-YA8 and OsNF-YC4 using co-immunoprecipitation assay.OsNF-YC4-GFP and OsNF-YA8-3 × FLAG were coexpressed in rice protoplasts.Proteins before (input) and after IP were detected by anti-GFP and anti-FLAG antibodies, respectively.IB, immunoblotting.(D) Relative expression of OsNF-Ys in WT and oYA8OEf3 leaves.Error bars indicate SD (n = 3).*, P < 0.05; **, P < 0.01 by Student’s t-test.

GA has been recently [65] reported to be an intrinsic signal for endosperm development.Given the GA deficiency in vegetative tissues of OsNF-YA8 overexpressors, we hypothesized that OsNF-YA8 probably regulate seed development by modulating GA biosynthesis.A recent study[60]showed that the quintuple mutant of OsNFYC8,9,10,11,12(pnfyc)germinated more slowly than the WT due to GA deficiency.The expression of most GA biosynthetic genes was down-regulated in pnfyc, indicating that NF-YCs transcriptionally regulate them [60].The seeds of OsNF-YA8 overexpression lines also displayed germination defects (Fig.1F).However, it is still unclear whether NF-Y is directly involved in the regulation of seed germination, because seed developmental defects may also cause germination difficulties.

5.Conclusions

Overexpression of OsNF-YA8,an NF-Y family member expressed preferentially in endosperm, disrupted both vegetative and reproductive development in rice.Ectopically expressed OsNF-YA8 bound to the promoters of some GA biosynthetic genes, including OsCPS1, OsGA20ox1, and OsGA20ox2, to repress their expression,eventually causing a defect in GA biosynthesis,resulting in reduced plant height.In addition, heading of OsNF-YA8 overexpressors was delayed.The expression of some key regulators of heading,including Hd3a,Hd1,RFT1,and DTH8,were disturbed.Either ectopic activation or knockout of OsNF-YA8 impaired seed development.These findings suggest that OsNF-YA8 is necessary for developmental regulation in rice.

CRediT authorship contribution statement

Baixiao Niu:Conceptualization, Investigation, Validation,Visualization, Writing – original draft, Funding acquisition.Jing Xu:Investigation, Visualization, Writing – original draft.Zhiguo E.:Software.Zhenyu Zhang:Investigation.Xinming Lu:Investigation.Chen Chen:Supervision,Conceptualization,Writing–review& editing, Funding acquisition.

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

We thank Dr.Zongju Yang for helpful suggestions and discussions.We thank Mr.Xuefeng Wei for management of plant materials.This work was supported by the National Natural Science Foundation of China (31701392 and 32170344), the Six Talent Peaks Project in Jiangsu Province (NY-142), the Jiangsu Province Government (JBGS[2021]001), and the Independent Scientific Research Project Funds of the Jiangsu Key Laboratory of Crop Genomics and Molecular Breeding (PLR202101).

Appendix A.Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cj.2023.07.001.