WOX transcription factors play important roles in key developmental processes and in response to different abiotic stresses (Cheng et al., 2016; Jha et al., 2020; Wang L. Q. et al., 2021). Rice genome contains at least 13 WOX genes (Zhang et al., 2010). Among them, OsQHB encodes 200 amino acids including a homeodomain (HD) with 66 amino acids at the N-terminal (Supplementary Figure 1A). Phylogenetic analysis of rice and Arabidopsis WOX families indicated that OsQHB showed the highest similarity to AtWOX5 (Supplementary Figure 1B and Supplementary Table 1), which is specifically expression in the QC cells and is essential for stem cell maintenance in different meristems (Sarkar et al., 2007; Pi et al., 2015). β-glucuronidase (GUS) staining analysis using transgenic rice plants harboring OsQHB promoter-GUS construct showed that the transcription of OsQHB was detected in QC and stele of the root apexes, stem base, and crown root primordium (Figures 1A–D). Subsequently, the subcellular localization of OsQHB was investigated by fusing the OsQHB coding sequence with GFP and transiently expressed in rice protoplasts. The fluorescent signals of OsQHB-GFP fusion protein were found in the nucleus, as revealed by co-localization with the nuclear marker DAPI (4′,6-diamidino-2-phenylidone) (Yu et al., 2013; Figure 1E). These results suggest that OsQHB might be involved in the maintenance of the QC cells through a mechanism similar to that of AtWOX5.

To study the function of OsQHB, we generated knockout mutants in the Nipponbare (NIP) background via CRISPR-Cas9 mediated genome editing and isolated three alleles for further investigation. The osqhb-1 and osqhb-2 mutant plants contained a 5-bp and 8-bp deletion, whereas the osqhb-3 mutant plants contained a 1-bp insertion, in the coding region of the target gene, correspondingly, leading to a frame shift in the open reading frame and the generation of a premature stop codon (Supplementary Figures 2A,B). We also generated overexpression (OE) lines containing the coding region of OsQHB under the control of the CaMV35S promoter, the increased expression of the target gene was confirmed by qPCR (Supplementary Figure 2C). Although the expression of OsQHB was significantly increased in OsQHB-OE plants, whereas the OsQHB protein was increased slightly in OsQHB-OE plants (Supplementary Figure 2D), suggesting that OsQHB may be rapidly degraded by the ubiquitin system or the proteasome. To verify this hypothesis, we treated OsQHB-OE plants with MG132 (an inhibitor of the ubiquitin proteasome system) and then detected the OsQHB protein levels. The results showed that MG132 treatment significantly increased the accumulation of OsQHB protein (Supplementary Figure 2D), indicating that OsQHB is rapidly degraded by the ubiquitin proteasome system, and implying that OsQHB is essential for plant development and plants must maintain low levels of OsQHB protein to ensure normal growth.

Next, we investigated the agronomic traits of OsQHB overexpressed plants and osqhb mutants. At maturation stage, the osqhb mutants were shorter and the OsQHB-OE plants were slightly taller than wild-type plants (Figures 2A,B). The number of effective tiller and panicle length in wild-type and OsQHB-OE plants were identical, whereas osqhb mutants produced less effective tiller and smaller panicle than the wild type (Figures 2C–E). The total number of grains per panicle was significantly reduced in osqhb mutants, whereas slightly increased in OsQHB-OE plants, compared to that of the wild type (Figure 2F). All these analyses indicate that modulation of OsQHB expression affects yield-related traits in rice.

The grain size was also examined in osqhb mutants and OsQHB-OE plants using well-filled grains. The grain length and grain width was significantly reduced in osqhb mutants, whereas increased in OsQHB-OE plants, compared with those in the wild type (Figures 3A–C). The thousand-grain weight of various plants was further measured. The results showed that osqhb mutants had significant reductions, whereas the OsQHB-OE plants had significant increases in this parameter (Figure 3D). These results indicate that OsQHB positively regulates the grain size and thousand-grain weight in rice, implying that manipulation of OsQHB could be useful to improve rice yield.

The promoter sequence of OsQHB contains many putative hormone- and stress-response cis-elements, such as ABRE element (4 hits), AuxRR core (2 hits), P-box (1 hit), G-box (6 hits), and MYB recognition site (1 hit) (Figure 4A). ABRE element is associated with ABA response (Nakashima and Yamaguchi-Shinozaki, 2013), G-box and MYB recognition site with drought response (Ezer et al., 2017), AuxRR core with auxin response, and P-box with gibberellin response (Sakai et al., 1996; Hwang et al., 2010; Wang Y. et al., 2011; Ambawat et al., 2013; Wei et al., 2013; Zhang et al., 2020; Dutta et al., 2021). These elements are contained in the promoter of OsQHB suggests that OsQHB might be involved in modulating hormone response and stress tolerance.

Furthermore, we examined the responses of OsQHB to various abiotic stresses and hormone treatments by β-glucuronidase (GUS) staining. Considering that OsQHB is mainly expressed in root tips, we selected root tips for GUS staining. Our results showed that the GUS staining was significantly repressed by NaCl and weakly induced by osmotic stress (PEG6000) and ethylene, but apparently not affected by abscisic acid (ABA) treatment, respectively (Figures 4B–E). Numerous studies have shown that ABA and ethylene play important roles in regulating salt and drought tolerance (Pan et al., 2012; Liang et al., 2019; Liu C. Y. et al., 2019; Luo et al., 2020). Collectively, these results indicate that OsQHB might be associated with ethylene- and ABA-related environmental stimuli in rice.

Salt stress is a major environmental problem globally by affecting plant growth and causing crop production. Increasing numbers of studies have shown that the WOX gene family plays a role in salt tolerance (Wang L. Q. et al., 2021). As salt represses OsQHB transcription, we hypothesized that OsQHB is involved in regulating salt tolerance in rice. To verify this, we treated wild type, osqhb, and OsQHB-OE seedlings with NaCl. After 10 days of treatment with 120 mM NaCl, osqhb seedlings were obviously less affected than NIP seedlings by salt stress, whereas OsQHB-OE seedlings were all wilted to death (Figure 5A). After recovery for 10 days in non-salt conditions, the survival rate of osqhb seedlings was significantly higher, and that of OsQHB-OE seedlings was significantly lower, than that of the wild-type seedlings (Figures 5B,C). These results indicate that OsQHB negatively regulates the response of rice to salt stress at seedling stage.

To elucidate the molecular mechanisms regulated by OsQHB, we compared the transcriptomes of wild type, osqhb, and OsQHB-OE seedlings using transcriptome deep sequencing (RNA-seq). Totally 1079 differential expressed genes (DEGs) were identified in osqhb mutants, including 525 up-regulated DEGs and 554 down-regulated DEGs (Figure 6A and Supplementary Table 2). Gene Ontology (GO) enrichment analysis showed that these OsQHB-regulated DEGs were involved in stress response, regulation of metabolic and cellular biosynthetic process, ion transport, and kinase activity (Figure 6B), indicating that OsQHB is involved in a variety of biological processes and molecular functions, including those associated with abiotic stress responses.

Previous studies have shown that salt stress leads to over-production of ROS, and enhanced ROS scavenging capacity can improve salt tolerance in plants (Steffens, 2014; Xu et al., 2018; Yang and Guo, 2018). Thus we analyzed the expression of ROS scavenging-related genes in transcriptome data and found that multiple peroxidase (PRX) genes were regulated by OsQHB (Figure 6C and Supplementary Table 2), this regulation was further confirmed by qPCR (Figure 6D). PRX genes play diverse roles in plant physiology by scavenging ROS (Liu R. et al., 2021). These results indicate that OsQHB regulates the expression of ROS scavenging-related genes, which could further affect ROS accumulation.

To investigate whether ROS levels were altered in wild type, osqhb, and OsQHB-OE seedlings under salt treatment, we measured ROS levels in wild type, osqhb, and OsQHB-OE seedling leaf. Nitro blue tetrazolium (NBT) staining showed that the osqhb mutants accumulated less O^2– in the leaves compared with wild-type plants under salt stress (Figure 7A). On the contrary, the OsQHB-OE plants accumulated much more O^2– in the leaves compared with wild-type plants under salt stress (Figure 7A). Further determination of H₂O₂ content showed that the H₂O₂ content of osqhb mutants was lower, while that of OsQHB-OE plants was higher, than that of wild type after salt treatment (Figure 7B). Malondialdehyde (MDA) is an important product of lipid peroxidation, which represents the degree of oxidative damage to the plant cell (Qin et al., 2016). Salt stress significantly increased the MDA content in wild-type (Figure 7C). This tendency was weakened in osqhb mutants, but enhanced in OsQHB-OE plants (Figure 7C). These results indicate that knocking out OsQHB leads to the reduction of ROS accumulation under salt stress, thereby improving the salt tolerance of rice seedlings.

Salt tolerance is tightly related to the activities of ROS-scavenging enzymes. Therefore, we measured the activities of ROS-scavenging enzymes after exposed to NaCl. As shown in Figures 7D,E, after exposed to salinity stress, the activities of antioxidant enzymes in all the three genotypes were significantly increased, whereas the osqhb mutants showed a higher antioxidant enzymes activities and the OsQHB-OE lines showed a lower antioxidant enzymes activities, as compared with wild-type plants (Figures 7D,E), indicating that the OsQHB mutation leads to enhancement of salt tolerance in rice seedlings by enhancing the activities of antioxidant enzymes.

OsQHBp-GUS transgenic line was described previously (Ni et al., 2014). osqhb allelic mutants in the Nipponbare (NIP) background were generated by using CRISPR/Cas9 (Xu et al., 2017). Briefly, the target region 5′-GGAGCAGGTGAAGGTCCTGA-3′ was introduced into the pHUN4c12 vector backbone, and the recombinant vector was transformed into Agrobacterium strain EHA105-pSOUP for rice transformation. To generate the overexpression transgenic plants, the coding sequence of OsQHB was amplified by PCR and cloned in-frame with a Flag-tag into pCAMBIA1307 under the control of the CaMV 35S promoter (Wang J. et al., 2013). The recombinant plasmids were introduced separately into NIP or Zhonghua 11 (ZH11) via Agrobacterium-mediated transformation.

Rice seeds were imbibed in Petri dishes with sterile distill water at 37°C for 48 h. The germinated seeds were sown in a bottomless 96-well plate in a container of Yoshida’s culture solution (Cui et al., 2015) or in growing trays filled with soil. The plants were grown in a growth chamber under a 14 h light (30°C)/10 h dark (25°C) photoperiod, with a light intensity of ∼150 μmol m^–2s^–1 (white light) and 60% relative humidity. To assess salt-tolerance of rice seedlings, 3-week-old seedlings were treated with 120 mM NaCl solution for 10 days, and then transferred to water without NaCl for an additional 10 days. Subsequently, the survival rate of rice seedlings was counted (surviving seedlings indicated plants with green leaves).

For propagation and investigation of agronomic traits, the plants were cultivated at the Experimental Station of the Chinese Academy of Agricultural Sciences in Beijing during the natural growing seasons.

Arabidopsis WOX proteins sequences were available at the Arabidopsis genome sequence database¹ and rice WOX proteins sequences were downloaded from the National Center for Biotechnology Information (NCBI) database². Full-length protein sequences of all WOX proteins were aligned by multiple sequence alignment program of the Clustal X, and then phylogenetic tree was constructed by MEGA 6 software according to the following parameters: neighbor-joining method, 500 bootstraps (Tamura et al., 2007).

For promoter analysis, the promotor sequence (2,000 bp upstream of ATG) of OsQHB was uploaded to the PlantCARE database³ to identify stress related cis-elements as previously reported (Yang and Guo, 2018).

Total RNA was extracted from leaf tissues using an Ultrapure RNA Kit (CWBIO, China) according to the manufacturer’s instructions. ∼2 μg total RNA was reverse transcribed using HiScript II Q RT SuperMix (Vazyme, China) according to the manufacturer’s instructions. qPCR was performed on an optical 96-well plate with Bio-Rad iQ5 system using SYBR Green Real-time PCR Master Mix reagent (Vazyme, China) as previously described (Zhang et al., 2012). The rice Actin1 gene was used as an internal standard to normalize gene expression. The primers used for qPCR are listed in Supplementary Table 3.

For GUS staining, 3-d-old OsQHBp-GUS seedlings were immersed in GUS staining buffer (50 mM sodium phosphate, pH 7.0; 10 mM EDTA; 0.5 mM K₃[Fe (CN)₆]; 0.5 mM K₄[Fe (CN)₆]; 0.1% [v/v] Triton X-100; and 1 mM 5-bromo-4-chloro-3-indolyl-β-D-glucuronic acid). The samples were vacuumed in a vacuum pump for 30 min, and then stained in 37°C for 2 d. After dehydrated in 75% (v/v) ethanol to remove the chlorophyll, the root tip was sectioned and photographed under a stereoscopic microscope.

To determine the subcellular localization of OsQHB, the full coding sequence of OsQHB was cloned into the pAN580-GFP vector under the control of the CaMV 35S promoter. Rice protoplasts prepared from etiolated shoots were transfected with OsQHB-GFP (Jiang et al., 2018). After incubated at 28°C for 16–20 h, equal volumes of DAPI (4′,6-diamidino-2-phenylidone) staining solution and protoplast suspension were mixed, and the fluorescence signal was observed with a confocal laser scanning microscope.

The leaves of 3-week-old seedlings were immersed in 20 mM phosphate buffer (pH 6.1) containing 2 mM NBT. After vacuumed in pump for 1 h, the samples were stained in 37°C for 12 h and then transferred to 75% (v/v) ethanol to stop the reaction. The material was photographed under a light stereomicroscope.

Ascorbate peroxidase and SOD activities were determined according to the instruction of the superoxide dismutase assays kits (Solarbio, China) and the ascorbate peroxidase assays kits (Solarbio, China). In brief, 0.1 g of normal or salt-stressed rice leaves from 3-week-old seedlings were harvested, ground in liquid nitrogen, and extracted by extraction buffer. After centrifuged at 8,000 g (SOD) for 10 min or 13,000 g (APX) for 20 min, the supernatant was used for SOD and APX activity measurement. The absorbance values were measured at 560 and 290 nm, respectively. SOD and APX activities were calculated from the formula provided with the SOD and APX assay kit (Solarbio, China).

H₂O₂ content was determined using the method described previously (Wang Y. et al., 2019). Briefly, 0.1 g rice leaves from 3-week-old seedlings with or without salt treatment were harvested and homogenized in 1 ml cold acetone. Then, H₂O₂ content was determined using a hydrogen peroxide assay kit (Solarbio, China) according to the manufacturer’s instructions.

To measure MDA, 0.1 g of normal or salt-stressed rice leaves were homogenized in 1 ml 0.1% (w/v) trichloroacetic acid (TCA) followed by centrifugation at 8,000 g for 10 min at 4°C. Four volumes of 0.5% (w/v) thiobarbituric acid (TBA) in 20% (w/v) TCA were added to one volume of supernatant; the mixture was incubated at 100°C for 1 h. The reaction was terminated by incubating the mixture on ice for 15 min, and the absorbance was measured by spectrophotometry at 450 nm, 532 and 600 nm. The content of MDA was calculated according to the formula provided in the MDA assay kit (Solarbio, China).

The primary roots of 3-d-old OsQHB overexpression line, osqhb mutant, and wild-type seedlings were collected and extracted using the TRIzol method (TIANGEN BIOTECH, China) and treated with RNase-free DNase I (TaKaRa, Japan). mRNA was purified from total RNA using poly-T oligo-attached magnetic beads and then subjected to RNA-seq library construction for the transcriptome experiments using NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (NEB, United States). Multiplex paired-end adapters were used for multiplex libraries. The library quality was assessed on the Agilent Bioanalyzer 2100 system. The library preparations were sequenced on an Illumina Hiseq 4000 platform by the Beijing Allwegene Technology Company Limited (Beijing, China) and paired-end 150 bp reads were generated. After removing adaptor and low-quality reads, clean reads were mapped to rice genome MSU7.0 using TopHat, and analyzed using Cufflinks according to Trapnell et al. (2012). Poisson-dispersion model of fragment was used to conduct statistical analysis (FDR < 0.05) and responsive genes were identified by fragments per kilobase per million reads (FPKM) requiring more than twofold change between two samples. Three biological replicates were used, and their repeatability and correlation were evaluated by the Pearson’s Correlation Coefficient (Schulze et al., 2012). The transcriptome data from this study can be found in the National Center for Biotechnology Information Sequence Read Archive (NCBI SRA) under the accession number PRJNA801123.

The primers used in this study are shown in Supplementary Table 3.