Wenqing Li, Minjun Zhngb, Lei Qio Yunbo Chen Dpeng Zhng, Xiuqing JingPengfei Gn Yngbin Hung Junru Go Wenting Liu Chunhi Shi, Hongchng Cuie, Hifeng Li,Kunming Chen

a State Key Laboratory of Crop Stress Biology in Arid Areas, College of Life Sciences, Northwest A＆F University, Yangling 712100, Shaanxi, China

b The Sericultural and Silk Research Institute, College of Animal Science and Technology, Northwest A＆F University, Yangling 712100, Shaanxi, China

c College of Agronomy, Northwest A＆F University, Yangling 712100, Shaanxi, China

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

e Department of Biological Science, Florida State University, Tallahassee, FL 32306, USA

Keywords:Auxin transporter Root development Wavy root Gravitropism Abscisic acid (ABA)Drought tolerance

A B S T R A C T Root system architecture is influenced by gravity.How the root senses gravity and directs its orientation,so-called gravitropism, is not only a fundamental question in plant biology but also theoretically important for genetic improvement of crop root architecture.However,the mechanism has not been elucidated in most crops.We characterized a rice agravitropism allele,wavy root 1(war1),a loss-of-function allele in OsPIN2, which encodes an auxin efflux transporter. With loss of OsPIN2 function, war1 leads to altered root system architecture including wavy root, larger root distribution angle, and shallower root system due to the loss of gravitropic perception in root tips. In the war1 mutant, polar auxin transport was disrupted in the root tip,leading to abnormal auxin levels and disturbed auxin transport and distribution in columella cells. Amyloplast sedimentation, an important process in gravitropic sensing, was also decreased in root tip columella cells. The results indicated that OsPIN2 controls gravitropism by finely regulating auxin transport, distribution and levels, and amyloplast sedimentation in root tips. We identified a novel role of OsPIN2 in regulating ABA biosynthesis and response pathways. Loss of OsPIN2 function in the war1 resulted in increased sensitivity to ABA in seed germination,increased ABA level,changes in ABA-associated genes in roots,and decreased drought tolerance in the seedlings.These results suggest that the auxin transporter OsPIN2 not only modulates auxin transport to control root gravitropism, but also functions in ABA signaling to affect seed germination and root development,probably by mediating crosstalk between auxin and ABA pathways.

1. Introduction

Root system architecture is an important agronomic trait because of its influence on water and nutrient absorption [1,2].For plants, gravity represents a critical environmental signal and is a constant presence,constantly influencing root system architecture.How the root senses gravity and directs its orientation,the socalled gravitropism, is not only a fundamental question in plant biology [3] but also theoretically important for genetic improvement of root system architecture in crops [1,2]. However, the molecular mechanism of gravitropic perception remains unknown in most crops.

In the past two decades,the mechanism underlying gravitropic perception has been extensively studied in the model plantArabidopsis thaliana[4-7]. Increasing evidence has supported a critical role for polar auxin transport (PAT) in controlling gravitropism [8-10]. In higher plants, auxin (IAA) is synthesized mainly in leaf primordia and young leaves. It is directionally transported to the root through the central vascular tissue and then concentrated in the root cap to generate a basipetal auxin concentration gradient [11,12]. In the root cap, auxin is redirected to cortical and epidermal cells, leading to asymmetric auxin distribution [9].This is the so-called PAT,mediated by polarly localized auxin influx facilitators and auxin efflux carriers,guiding directional auxin flow in plants [11,12].AUX1,encoding an auxin influx facilitator,controls root gravitropism by facilitating auxin uptake in the root apical meristem [8].PIN-FORMED 1(AtPIN1) encodes a member of the auxin efflux carrier family [13], and its mutation reduces PAT in shoots [14].AtPIN2was identified among several alleles ofArabidopsisagravitropic root mutants [4,15-17]. As a homolog ofAtPIN1,AtPIN2was predominantly expressed in roots and responsible for auxin redistribution from root cap to cortical and epidermal cells inArabidopsisroots [16]. Accordingly,AtPIN2mutants show both reduced PAT and loss of gravitropic response in roots[4,17]. Recent investigation [18] further revealed thatAUX1andAtPIN2function in the same pathway during root gravitropic response andAUX1acts upstream ofAtPIN2.AtPIN2-mediated auxin transport and gravitropic perception may be modulated by protein phosphorylation [19], dephosphorylation [20], ubiquitination/proteasome [21,22], cytoskeleton-associated proteins [23],and other phytohormones [24]. PID kinase may control root gravitropism by modulation of AtPIN2-dependent basipetal auxin transport, which can be antagonistically regulated by RCN1 via dephosphorylation inArabidopsis[25]. A recent study [24] found that crosstalk between brassinolide signaling and endocytic AtPIN2 sorting is essential for determining the rate of gravity-induced root curvature.It has been suggested[26] that AtLAZY proteins control root gravitropism by coupling gravity sensing to the formation of auxin gradients. These studies suggest that the connection between polar auxin transport and root gravitropism are complex,involving many molecular components and their crosstalk.

In the rice genome, 12 PIN members have been identified and show highly tissue-specific expression patterns [27,28], implying their functional divergence in regulating rice growth and development.OsPIN2is the unique rice homolog ofAtPIN2and is expressed mainly in roots and stem base[27].Overexpression ofOsPIN2leads to abnormalities in tiller formation, plant height, and auxin transport [29], and also increased aluminum tolerance [30,31]. In two recent studies, [32,33]OsPIN2/LRA1mutation led to abnormalities in root elongation, lateral root formation, root gravitropism, and root architecture. However, the mechanism underlyingOsPIN2-mediated auxin transport and gravitropic regulation has yet to be elucidated and it remains unknown whetherOsPIN2is also involved in other developmental events.

The purpose of this study was to characterize an agravitropism allele,wavy root 1(war1), which showed abnormal root architecture including wavyroot,wider root angle,and shallower root system due to the loss of root gravitropism. A series of experiments including map-based cloning,CRISPR/Cas9-mediated genome editing, root gravitropic tests,DR5::GUSreporter and ABA assays, and RNA-sequencing (RNA-seq) and gene expression analyses were carried out to investigate the functional roles ofwar1in regulating rice growth and development.

2. Materials and methods

2.1. Plant materials and growth conditions

Thewar1mutant was isolated from the rice cultivar 93-11(Oryza sativaL.indica)by mutagenesis with60Co γ-rays.To remove unwanted genetic background after mutagenesis,thewar1mutant was crossed with the wild type 93-11 and the F1progeny were further backcrossed to the wild type.To observe the phenotype of root growth, germinated seeds were planted in hydroponic culture containing 1× Hoagland’s solution and 0.5 mmol L-12-(Nmorpholino)-ethanesulfonic acid monohydrate (MES). The hydroponic system was kept in a greenhouse with 12 h light (28 °C)and 12 h dark (25 °C) photoperiod at 180 μmol m-2s-1photon density and 70%-80% humidity. To test the response of roots to gravity, seeds were germinated and grown in vertical agar plates.Geotropic bending of roots was observed by inverting the agar plates for 24 h followed by 90° anticlockwise rotation for 24 h.

2.2. Map-based cloning

For genetic mapping,thewar1mutant was crossed with the rice cultivar 02428(Oryza sativaL.japonica).A total of 618war1plants were identified in the F2population. Preliminary genetic mapping was performed by genotyping of 94 F2mutant plants with SSR markers.For fine mapping ofwar1,526 F2mutant plants were further genotyped with SSR and InDel markers.To identify the mutation site, a candidate gene was amplified using genomic DNA extracted from thewar1mutant and 93-11 (wild type, WT) and sequenced.

2.3. Plasmid constructs and transformation

For CRISPR/Cas9-mediated genome editing ofOsPIN2, a single guide RNA (sgRNA) expression cassette with target sequence (5′-GTGGTGGGGGATATTCACGC-3′) located in theOsPIN2coding sequence (CDS) was constructed under the control of theOsU6apromoter. The sgRNA intermediate plasmid was assembled into the binary vector BGK03 to generate a CRISPR/Cas9 plasmid. The construct was introduced intoAgrobacterium tumefaciensstrain EHA105 and then transformed into thejaponicacultivar Nipponbare. Identification of theOsPIN2mutation in transgenic plants was performed by direct sequencing of the target region. To remove the CRISPR/Cas9 construct in the CRISPR/Cas9 mutants,the homozygous lines (war1-cr) were crossed with Nip and the F1progeny were again backcrossed to Nip.For subcellular location of OsPIN2 protein,the CDS was cloned into plasmid pTF486 to generate a 2×CaMV35S::OsPIN2-GFPtransient expression vector. The resulted vector and empty pTF486 were transiently expressed in rice protoplasts and GFP fluorescence was observed under Nikon A1R confocal microscope (Nikon Corporation, Minato-ku, Tokyo,Japan).

2.4. Auxin and DR5::GUS experiments

Quantitative analysis of endogenous level of free IAA followed Inahashi et al.[32].To inhibit auxin transport from shoots to roots,germinated seeds were grown in hydroponic solution and the shoot-root junctions were embedded in 1% agar containing 20 μmol L-1NPA (N-1-naphthylphthalamic acid). The seedlings were grown in hydroponic solution for one week and the agar was replaced every three days. The root phenotype of each plant was observed and photographed. To visualize auxin transport and distribution in root tip, thewar1mutant was crossed to aDR5::GUSline expressing the auxin-responsive GUS reporter in the wild-typejaponicacultivar Zhonghua 11. F1hybrids were confirmed by both GUS-histochemical staining and DNA sequencing of theOsPIN2locus. Wild-type andwar1plants in the F2population were selected by observation of root geotropic bending on vertical agar plates, root phenotype in hydroponic culture, and GUS staining of leaves. Root tips from 5-d-old, 2-week-old, and 8-week-old plants were examined for GUS-histochemical staining and microscopic observation following Zhou et al. [34]. At least 15 positive plants of each genotype were selected for theDR5::GUSassay.

2.5. ABA assays and drought stress treatment

Measurement of endogenous ABA levels of samples was performed by enzyme-linked immunosorbent assay following Chen et al. [35]. To test ABA sensitivity during seed germination,dehusked seeds were surface-sterilized and placed on 1/2 Murashige and Skoog (MS) medium plates supplemented with a range of concentrations of ABA.The plates were incubated at 28°C under continuous light and the seed germination rate was scored at a range of time points. Sixteen seeds of each genotype were used for each experiment and six independent replicates were tested.For a drought tolerance test, thewar1andwar1-crmutants and their wild-type controls were grown in hydroponic culture for two weeks. The seedlings were drought treated by withholding water (exposure of the whole plant to air) for 30 h, followed by recovery for 7 days in hydroponic solution. Numbers of surviving plants were recorded. At least three independent replicates were tested.

2.6. Microscopic observation

For anatomical analysis, cytological sections of young roots were prepared and observed under a Leica DM5000 B microscope(Leica,Wetzlar,Germany).To observe amyloplasts in root tips,germinated seeds were treated as described by Takahashi et al. [36].The root tip was observed under the microscope. The number of stained amyloplasts was determined with Image J software(https://imagej.net/Downloads).

2.7. Gene expression, RNA-seq, and sequence analyses

For expression analyses ofOsPIN2and other auxin-associated genes, total RNA was extracted from roots and leaves of 7-d-old seedlings using a RNeasy Plant Mini Kit (Qiagen 74904, Qiagen Inc., Hilden, Germany) according to the manufacturer’s protocol.The first strand of cDNA was synthesized from 2 μg of total RNA in a 25 μL reaction mixture using a M-MLV First Strand Kit (Invitrogen, Carlsbad, CA, USA). To amplify the full-length cDNA sequence ofOsPIN2, cDNA products equivalent to 100 ng of total RNA were used for RT-PCR with gene-specific primers.To measure the expression level ofOsPIN2and other auxin-associated genes,real-time qRT-PCR was performed with SYBR Es Mix (Roche Applied Science, Mannheim, Germany) on the Bio-Rad CFX96 Real-Time PCR system (Bio-Rad Laboratories Inc., Hercules, CA,USA). The relative expression level of target gene was obtained by normalization to the internal reference geneUBQ. The experimental procedures for RNA-seq and data analysis are detailed in Methods S1. Protein homology search of OsPIN2 against its orthologs was performed with the BLASTP program (http://blast.ncbi.nlm.nih.gov/).Multiple sequence alignments were conducted with PRALINE (http://www.ibi.vu.nl/programs/pralinewww/) using the default parameters.A phylogenetic tree of OsPIN2 and its orthologs was constructed with MEGA 7 (https://www.megasoftware.net/)by the neighbor-joining method with 1000 bootstrap replicates.

3. Results

3.1. Phenotypic characterization of war1

The mutant was identified in a60Co γ-ray-mutagenized population of the riceindicacultivar 93-11. Compared with 93-11, the mutant showed a wavyrooting phenotype in hydroponics and was accordingly namedwavy root 1(war1). Thewar1mutant showed wavyroots from the germination stage (Fig. 1A). At the seedling stage and tillering stage, thewar1still exhibited a wavy rooting phenotype and then showed a wider root distribution angle than 93-11 (Fig. 1B-E). In cytological sections, 93-11 showed regular rectangular cell shapes in the root cortex, butwar1showed abnormal shapes of cortical cells in root (Fig. 1F, G).

3.2. The war1 is a loss-of-function allele of OsPIN2

In the F2mapping population,325 normal and 94 mutant plants were obtained, fitting a 3:1 Mendelian segregation (χ2= 1.76 ＜3.84),indicating thatwar1is a single recessive gene.Genetic mapping ofwar1revealed that the gene was linked to DNA markers on rice chromosome 6 (Fig. 2A). By genotyping of 526 F2mutant plants,war1was further confined between markers RM20508 and RM3430 (Fig.2B), an interval representing a physical distance of 410 kb. In this interval,LOC_Os06g44970encoding the putative auxin efflux carrier OsPIN2,was selected as a candidate.The genomic DNA and cDNA ofLOC_Os06g44970were amplified and sequenced, revealing a single base-pair deletion (343C) and two single base-pair substitutions in thewar1mutant (Fig. 2C). The deletion (343C) inwar1resulted in a frameshift mutation that replaces the 516 C-terminal residues with 130 new amino acids(Fig. 2C), suggesting thatwar1is a loss-of-function mutation inOsPIN2. CRISPR/Cas9-induced genome editing ofOsPIN2was performed in thejaponicacultivar Nipponbare. The homozygous transgenic lines,namedwar1-cr#1,#2,and#3,carried 1-bp insertions or deletions in the open reading frame(ORF)region ofOsPIN2(Fig. 2D), mimickingwar1(Fig. 2E, F). These results revealed that thewar1is a novel allele inOsPIN2.

3.3. OsPIN2 was preferentially expressed in root tips and the protein was membrane-localized

Temporal and spatial expression ofOsPIN2was examined in 18 tissue types at the seedling, tillering, booting, and heading stages.OsPIN2showed higher expression in roots and lower expression in stems and leaves, especially at the seedling and tillering stages(Fig.3A).Expression ofOsPIN2was observed at all examined developmental stages (Fig. 3A). Compared with expression at 0 days after germination (DAG) of coleoptiles, the expression ofOsPIN2was very low in callus(Fig.3B).After germination,OsPIN2showed higher expression in root tip, but lower expression in root base,middle of root, and root-shoot junction (Fig. 3B).OsPIN2showed highest expression in root tips at 4 DAG (Fig. 3B). Thus,OsPIN2was preferentially expressed in root tips.

Subcellular localization of OsPIN2 was investigated by transient expression ofCaMV35S::OsPIN2-GFPin rice protoplast. Fluorescence signals of GFP were distributed along the plasma membrane(Fig.3C),indicating that OsPIN2 was membrane-localizated.Multiple sequence alignment(Fig. S1) revealed that OsPIN2 was similar to its counterparts inArabidopsis thaliana, but that monocots and dicots showed divergence.OsPIN2 shared higher sequence similarity with monocots than dicots (Fig. 3D), suggesting a functional divergence between OsPIN2 and AtPIN2.

3.4. OsPIN2 controls gravitropic perception and amyloplast sedimentation in root tips

Fig.1. Comparison of phenotypes of 93-11(WT)and war1 mutant.Phenotypes of 1-d-old(A)and 3-d-old(B)seedlings of 93-11 and war1.93-11 produced straight radicle(seminal root), whereas war1 produced wavy radicle. (C) Enlarged view of wavy root phenotype in (B). (D) Phenotype of 12-d-old seedling roots. SR, seminal root (primary root);CR,crown root;LR,lateral root.(E)Phenotypes of root system for 8-week-old 93-11 and war1 plants.Green dotted line indicates root angle.(F)Longitudinal sections of seminal roots of 3-d-old seedlings. (G) Cross sections of seminal roots. S, stele; En, endodermis; C, cortex; Ex, exodermis; Ep, epidermis. Scale bars, 1 cm in (A), (B) and (D);1 mm in (C); 100 μm in (F) and (G).

The wavy root phenotype led us to further test gravity sensing inwar1andwar1-cr#1. When grown in vertical agar plates, the wild-type controls (93-11 and Nip) align their rooting direction to the gravity vector, whilewar1andwar1-cr#1lose the direction of root elongation(Fig.4A).When the seedlings were inverted and grown for 24 h, the wild-type controls realigned their rooting direction to the gravity vector, whereas the mutants showed no gravity-induced curvature of root tips(Fig.4B).When the seedling was further changed from inverted to horizontal orientation, the wild-type controls showed geotropic bending in root tips, but the mutants still produced no gravity-induced curvature of root tips(Fig.4C).Thus,loss of function ofOsPIN2resulted in no gravitropic perception in root tips,accounting for the wavy phenotype inwar1andwar1-cr#1.When grown in vertical agar plates for a week,the mutants and wild-type plants showed clear differences in root system architecture and distribution. The roots of the wild type elongated downward vertically, but the mutant roots did not elongate vertically, leading to shallower root systems (Fig. 4D).

Amyloplasts in columella cells of root cap have been proposed[37] to participate in determining gravity sensing. In comparison with 93-11 and Nip, the distribution and numbers of amyloplasts in root columella cells were significantly decreased inwar1andwar1-cr#1(Fig. 4E, F), demonstrating that loss of gravitropic perception was associated with decreased amyloplast sedimentation in columella cells.

3.5. OsPIN2 finely regulates auxin levels, auxin transport, and auxin distribution in root tips

To investigate whether the loss ofOsPIN2function alters auxin levels, endogenous free IAA was quantified in 93-11 andwar1seedlings. In 5-day-old seedlings, there was no significant difference in auxin level between 93 and 11 andwar1(Fig. 5A). However, in 2-week-old seedlings,war1showed increased auxin levels in shoots, but no changes in root or top leaf, relative to 93-11 (Fig. 5B). Exogenous application of IAA did not change the wavy phenotype ofwar1, but exogenous NPA treatment of the root-shoot junction significantly reduced the wavy frequency inwar1(Fig. S2), indicating that the wavy-root phenotype inwar1was associated with abnormality of auxin transport.

Fig. 2. Map-based cloning of war1 and CRISPR/Cas9-induced mutation of OsPIN2. (A) Genetic mapping of war1 on rice chromosome 6. (B) war1 was further mapped to a chromosomal region between markers RM20508 and RM3430. (C) Map-based cloning revealed that war1 is an allelic mutation in OsPIN2. The mutation site of war1 is indicated. (D) A schematic of CRISPR/Cas9 target site in OsPIN2. Three types of mutation events in the target region are shown, and the resulting mutant lines were named war1-cr#1, #2, and #3. (E) Seedling phenotype of 2-day-old Nipponbare (Nip) and war1-cr mutants. Three independent war1-cr lines, #1, #2, and #3, are shown. (F) Root phenotype of 10-d-old war1-cr mutants. Scale bars, 1 cm in (E) and (F).

Auxin transport and distribution were further investigated using the auxin-responsive promoterDR5::GUSreporter. As indicated by GUS histochemical staining,war1and 93-11 showed completely different patterns of auxin distribution in root tips. In 5-day-old seedlings, locations of auxin were confined to the stele,quiescent center (QC), and columella (central root cap) in 93-11,but auxin was diffusedly present in the stele, zone of cell division,and root cap, as well as asymmetrically distributed in one side of the elongation zone at the root tip ofwar1(Fig.5C).In comparison with 93-11, less GUS signal was localized in the QC inwar1(Fig. 5C). On the whole, 5-day-oldwar1seedlings showed more GUS staining in root tips than the WT (Fig. 5C), indicating higher levels of auxin in the root tips ofwar1. In 2-week-old seedlings,auxin was distributed uniformly in the stele, quiescent center(QC), columella, and lateral root cap in 93-11, but was present in the stele, the columella, and one side of the root division zone inwar1(Fig. 5D). Compared with 93-11,war1showed less GUS staining in whole root tip (Fig. 5D), indicating lower auxin levels in root tips ofwar1. In 8-week-old plants, auxin was massively accumulated in the whole root cap,QC,and central vascular tissue in 93-11,but was confined to the stele,QC,columella and one side of the lateral root cap inwar1(Fig. 5E). The level of auxin was markedly lower inwar1than in 93-11 (Fig. 5E). As shown, in 5-d-old, 2-week-old and 8-week-old plants, auxin distribution and concentration in central vascular tissue and root cap were gradually increased in 93-11,but not increased inwar1(Fig.5C-E),indicating that directional auxin transport from central vascular tissue to root cap was inhibited in thewar1mutant.

Fig. 3. Spatiotemporal expression of OsPIN2 and subcellular localization of the protein. (A) Relative expression of OsPIN2 in different tissues and developmental stages. (B)Relative expression of OsPIN2 during seed germination and early seedling development. DAG, days after germination. (C) Subcellular localization of OsPIN2 by transient expression of the OsPIN2-GFP fusion in rice protoplasts.GFP fluorescence was observed under a confocal microscope.(D)A phylogenetic tree of OsPIN2 and its orthologous proteins.

3.6. Transcriptomic analysis reveals changes in both auxin and ABA pathways

To understand the molecular function ofOsPIN2, transcriptomic analysis via RNA-seq was performed in 5-d-old seedling roots. Of 28,182 genes expressed in 93-11 andwar1(Table S1), 233 differentially expressed genes (DEGs) including 149 down-regulated and 84 up-regulated genes, were identified inwar1(Fig. 6A; Table S2). Gene Ontology (GO, http://geneontology.org/) enrichment analysis revealed that the major biological processes included metabolic process, cellular process, and biological regulation and localization (Fig. 6B). In the category of cellular component, the DEGs were most enriched in membrane part (Fig. 6B). In the category of molecular function, the significant GO terms included catalytic activity, binding, nucleic acid binding transcription factor activity, and transporter activity (Fig. 6B). In a Kyoto Encyclopedia of Genes and Genomes(KEGG, https://www.kegg.jp/kegg/pathway.html) pathway enrichment analysis, the DEGs fell into several pathways including starch and sucrose metabolism, plant hormone signaling transduction, MAPK signaling pathway, and carotenoid biosynthesis (Fig. 6C).

Fig.4. Gravitropic perception in roots and amyloplast distribution in the root cap.(A)Germinated seeds were grown in vertical agar plates for three days.(B)Agar plates in(A)were inverted and then incubated for 12 h.(C)Agar plates in(B)were further rotated 90°counterclockwise and then incubated for 12 h.(D)Germinated seeds were grown in vertical agar plates for 7 days. (E) Amyloplast staining of root tip. (F) Numbers of amyloplasts in the root cap. Values are means ± SD (n = 12). Significance of differences between mutants and wild-type controls was determined by Student’s t-test. ***, P ＜0.001. Scale bars, 2 cm in (A-D) and 200 μm in (E).

More than 20 DEGs involved in the auxin and ABA pathways were identified by RNA-seq (Table 1). Two genes encoding auxininduced protein 5NG4 (efflux transporter) were down-regulated inwar1, and of six genes encoding indole-3-acetate betaglucosyltransferases (IAGLUs), three genes were up-regulated and three were down-regulated inwar1(Table 1). Twelve genes involved in ABA-associated processes were differentially expressed inwar1(Table 1). Two genes encoding the ABA receptors OsPYL/RCAR4 and OsPYL/RCAR6 were down-regulated;two genes encoding 9-cis-epoxycarotenoid dioxygenases OsNCED3 and OsNCED4,the key enzymes in ABA biosynthesis were up-regulated; and a gene encoding serine carboxypeptidaseOsSCP46,which is involved in ABA signaling and ABA-dependent seed germination,was downregulated inwar1(Table 1). Thus, loss ofOsPIN2function influenced ABA-associated processes.

Table 1Differentially expressed genes associated with auxin and ABA pathways.

3.7. OsPIN2 influenced ABA biosynthesis and response

To further investigate howOsPIN2functions in the ABA pathway,ABA inhibition of seed germination was investigated.In comparison with a control(no ABA treatment),exogenous application of 1 μmol L-1ABA showed little effect on the germination rate in 93-11 or Nipponbare,but caused 37.5%and 48.3%reductions of germination rate inwar1andwar1-cr#1,respectively(Fig.7A).Similarly,exogenous application of 1 μmol L-1ABA showed little effect on the formation of green shoots (coleoptiles) in 93-11 and Nipponbare (Nip), but resulted in 85.4%and 70.9%reductions of green shoot rates inwar1andwar1-cr#1(Fig. 7B, C). Thus,OsPIN2mutation resulted in an ABA hypersensitive phenotype during germination.In 5-d-old seedlings,war1showed significantly increased ABA levels in roots relative to the wild type (Fig. 7D), whilewar1-cr#1showed increased ABA levels in both roots and shoots(Fig.7E).Thus,OsPIN2mutation led to increased ABA levels in roots. In 5-d-old seedling roots,OsPYL6/RCAR6andOsPYL4/RCAR4, two genes encoding ABA receptors, were down-regulated;OsSCP46, a gene encoding serine carboxypeptidase and regulating ABA signaling and seed germination[40],was also decreased inwar1andwar1-cr#1(Fig.7F)relative to their expression in 93-11 and Nipponbare.OsNCED3,encoding the key enzyme in ABA biosynthesis, showed increased expression inwar1andwar1-cr#1(Fig. 7F). Thus, the loss of function ofOsPIN2affected ABA pathways by increasing ABA sensitivity during germination, increasing ABA level, and changing expression of ABAassociated genes.

Fig. 6. Transcriptomic analysis of war1 mutant. (A) Hierarchical cluster analysis of differentially expressed genes (DEGs) in war1. S1, S2 and S3 indicate three biological replicates. (B) KEGG pathway enrichment of DEGs in war1. (C) Gene Ontology (GO) enrichment analysis of DEGs in war1.

3.8. Loss of OsPIN2 function led to decreased tolerance to drought stress

We investigated the drought tolerance ofwar1andwar1-cr#1.Two-week-old hydroponic seedlings were subjected to drought stress (Fig. 8A, B, left). After water was withheld, mostwar1andwar1-cr#1plants were wilted or dry, whereas 93-11 and Nipponbare(Nip)plants showed only a few wilted and dry leaves(Fig.8A,B,middle).After re-watering for 7 days,most 93-11 and Nip plants survived and grew vigorously, whereas most ofwar1andwar1-cr#1plants were dry and dead (Fig. 8A, B, right). Compared with 93-11 and Nip, the survival rates were reduced by 68% and 36%inwar1andwar1-cr#1,respectively(Fig.8C,D).Thus,loss of function ofOsPIN2resulted in decreased drought tolerance in the mutants.

Fig. 7. Loss of function of OsPIN2 affects ABA response, ABA level, and expression of ABA-associated genes. (A) Germination rate of seeds germinated on 1/2 MS plates containing no ABA (control) and 1.0 μmol L-1 ABA for two days. (B) Green coleoptile rates of seeds germinated on control and 1.0 μmol L-1 ABA plates at day 4. (C)Representative plots are shown for seeds germinated on 0, 0.5 and 1.0 μmol L-1 ABA at day 4.Three independent experiments were performed with similar pattern(n=25 seeds in each experiment).Scale bars,5 mm.(D,E)ABA contents in 5-day-old seedlings.Values in bar graphs represent three biological and three technical replicates for each genotype. Asterisks indicate significant differences between mutants and their corresponding wild-type controls determined by Student’s t-test. ***, P ＜0.001. (F) Relative expression of ABA-related genes in the mutants.

4. Discussion

In this study, functional characterization ofwar1revealed thatOsPIN2finely regulates auxin levels,auxin transport,and auxin distribution in root tips, leading to root gravitropism. This finding is consistent with those of previous investigations inArabidopsis[6,7] that PAT regulates root gravitropism. Loss of function ofOsPIN2led to disorder in auxin levels, auxin transport, and auxin distribution in root tips, although auxin content in whole roots was not changed inwar1.In fact, the previous studies have also suggested thatOsPIN2/LRA1mutation did not affect auxin level in the seedlings ofospin2/lra1mutants [32,33]. Our result ofDR5-GUSexperiment demonstrated the wild type has directional auxin transport in root tip, where auxin was confined in stele, QC, columella and lateral cap; however, thewar1mutant shows cleardefects in auxin transport and distribution in the root tip, where auxin distribution was disturbed especially in cap and columella cells. Considering that root columella cells are the site of gravity sensing [7,41], the abnormal transport and distribution of auxin at the columella will inevitably lead to gravitropic defects and wavy phenotype in root.These results show thatOsPIN2is required for auxin transport and auxin distribution in root tip, leading to gravitropic perception. In contrast, mutation ofOsPIN2disrupted auxin level, auxin transport, and auxin distribution in root tips,leading to loss of gravitropic perception.

InArabidopsis, amyloplast sedimentation in columella cells can trigger a re-localization of AtPIN2 to the lower side of the cell,facilitating asymmetric auxin redistribution in root tips and leading to root gravitropism [7].ARP3/DIS1controls root gravitropism by affecting amyloplast sedimentation and PIN-mediated polar auxin transport via regulation of AtPIN2 trafficking [23]. In the present study, amyloplast sedimentation in columella cells was significantly decreased inwar1andwar1-cr,implying thatOsPIN2control of gravitropic perception is associated with amyloplast sedimentation in root tips.In view of a similar mechanism previously shown inArabidopsis[23],we speculate thatOsPIN2disruption resulted in the absence ofOsPIN2in root tips,repressing both amyloplast sedimentation and polar auxin transport and leading to loss of root gravitropism.

Loss ofOsPIN2function led to increased sensitivity to ABA in seed germination, increased ABA levels, and changes of ABAassociated gene expression in roots, suggesting that auxin transporter OsPIN2 not only modulates auxin transport to control root gravitropism,but also functions in ABA signaling to affect seed germination and root development.These results suggest thatOsPIN2mediates crosstalk between auxin and ABA pathways.In fact,auxin and ABA interact extensively in the regulation of plant growth and development, especially in seed germination, root elongation, and lateral root formation [42]. BecauseOsPIN2plays a fundamental role in auxin transport, we speculate thatOsPIN2mediates crosstalk between auxin and ABA. This notion is supported by several investigations inArabidopsis. Auxin transport was required for the inhibitory effects of ABA on seed germination[43,44]and root elongation[45].Two previous studies[43,44]showed that an allele ofAtPIN2/EIR1confers resistance to ABA in seed germination.Thole et al. (2014) also identified a new allele of the auxin influx transporter AUX1.This allele had previously been identified as resistant to ABA in both root elongation [45] and seed germination [43].These investigations revealed that auxin transporters such as PIN2 and AUX1 regulate ABA-regulated processes. Similar situations have been reported inArabidopsis, where auxin receptors TIR1 and PP2A/RCN1 are also involved in crosstalk between auxin and ABA [44,46]. Besides being a molecular regulator of auxin transport and root gravitropic sensing[5],RCN1/PP2A can function as a general positive transducer of early ABA signaling [46]. However,in contrast torcn1andpin2/eir1, which showed ABA insensitivity in ABA inhibition of seed germination [43,44,46],war1showed increased sensitivity to ABA in seed germination and post-germination stages. This difference in ABA action was possibly due to species differences and to tissue-dependent variation of auxin-ABA crosstalk [42]. Our findings indicated thatOsPIN2mutation changed ABA perception in the seed germination and post-germination stages.This observation is consistent with changed expressions of genes encoding ABA receptors OsPYL6/RCAR6 and OsPYL4/RCAR4, and serine carboxypeptidase OsSCP46. In rice,OsSCP46controls seed germination by participating in ABA signaling [40] andOsPYL6/RCAR6may positively regulate ABA response during seed germination [47]. In the present study, loss ofOsPIN2function led to down-regulation ofOsPYL6/RCAR6,OsPYL4/RCAR4andOsSCP46, which could affect ABA response during seed germination. Loss ofOsPIN2function also caused an increase of ABA levels in root, consistent with up-regulation ofOsNCED3in thewar1mutants. These results suggest thatOsPIN2functions in regulating ABA signaling. Besides disrupting auxin transport, loss ofOsPIN2function changed expressions of auxin-associated genes,in particular six genes encoding indole-3-acetate betaglucosyltransferases (OsIAGLUs). Recent studies [38,39] have shown that OsIAGLUs not only participate in regulating auxin IAA homeostasis but mediate crosstalk between IAA and ABA. We suggest that loss ofOsPIN2function disrupted IAA homeostasis,leading to changes inOsIAGLUexpression, and changes in the expression ofOsIAGLUmay mediate the crosstalk between auxin and ABA in regulation of seed germination and early seedling development.

Fig.8. Comparison of drought tolerance between wild type and mutant seedlings.(A)Drought treatment of 93-11 and war1 seedlings.(B)Drought treatment of Nipponbare(Nip)and war1-cr#1 seedlings.Phenotypes of mutants and wild-type controls before drought treatment(left),after withholding water for 30 h(middle),and after recovery for 1 week (right) are shown. (C, D) Survival rates of war1 (C) and war1-cr#1 (D). Values are means ± SDs from three independent experiments (n = 40 seedlings in each experiment).Asterisks indicate significant difference between mutants and the corresponding wild-type controls determined by Student’s t-test.*,P ＜0.05;**,P ＜0.01.Scale bars, 2 cm in (A, B).

It is usually accepted that ABA is the major stress hormone promoting plant drought tolerance, and the ABA hypersensitive phenotype is associated with drought resistance [48-50]. Thewar1mutants were hypersensitive to ABA during the seed germination and post-germination stages (Fig. 7), but their drought tolerance was decreased (Fig. 8). An explanation of this inconsistency might be thatOsPIN2plays a more important role in regulating auxinrelated processes than in affecting the ABA pathway. The loss of function ofOsPIN2disrupted auxin transport and distribution,possibly accounting for the drought-sensitive phenotype ofwar1.ABA hypersensitivity resulting fromOsPIN2mutation, may not necessarily lead to increased drought tolerance because of the fundamental role ofOsPIN2in regulating auxin transport and distribution. This notion is consistent with the findings of some previous studies [51-54] that auxin pathways function in plant response to environmental stresses. Overexpression ofOsPIN3timproved drought tolerance in rice [51] and activation ofAtPIN2via PINOID kinase increased salt tolerance inArabidopsis[53],suggesting that PIN proteins act in regulating tolerance to drought and salt stresses. A more recent study [54] suggested that auxinsensitive Aux/IAA proteins mediate drought tolerance inArabidopsisby regulating glucosinolate levels.In fact,the response of plants to drought stress may involve other processes including the growth and development of root and leaf stomata. Thewar1mutants had defects in the root system that might confer the decreased tolerance to drought stress. The auxin transporter OsPIN2 modulates auxin transport and distribution, functions in the ABA signaling pathway,and affects drought tolerance,suggesting that crosstalk between auxin and ABA signals mediated byOsPIN2is involved in drought response.

CRediT authorship contribution statement

Wenqiang Li:Conceptualization, Visualization, Investigation,Formal analysis,Writing-original draft,Funding acquisition.Minjuan Zhang:Formal analysis, Methodology, Investigation, Data curation.Lei Qiao:Resources, Methodology, Investigation.Yunbo Chen:Methodology, Validation.Dapeng Zhang:Resources.Xiuqing Jing:Software.Pengfei Gan:Methodology.Yangbin Huang:Validation.Junru Gao:Validation.Wenting Liu:Software.Chunhai Shi:Resources.Hongchang Cui:Writing - review＆editing.Haifeng Li:Supervision, Visualization,Writing - review＆editing.Kunming Chen:Supervision, Project administration, Writing -review＆editing.

Declaration of competing interest

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

Acknowledgments

This work was supported by the National Natural Science Foundation of China(32070197,31570181 and 31200148)and Chinese Universities Scientific Fund(2452018149).We thank Rice Genome Resource Center (Tsukuba, Japan) for providing full-length cDNA clone ofOsPIN2.