Cytokinins regulate rice lamina joint development and leaf angle

Abstract

Leaf angle is determined by lamina joint inclination and is an important agronomic trait that determines plant architecture, photosynthetic efficiency, and crop yield. Cytokinins (CKs) are phytohormones involved in shaping rice (Oryza sativa L.) architecture, but their role in leaf angle remains unknown. Here, we report that CK accumulation mediated by rice CK OXIDASE/DEHYDROGENASE3 (OsCKX3) controls lamina joint development and negatively regulates leaf angle. Phenotypic analysis showed that rice osckx3 mutants had smaller leaf angles, while the overexpression lines (OsCKX3-OE) had larger leaf angles. Histological sections indicated that the leaf inclination changes in the osckx3 and OsCKX3-OE lines resulted from asymmetric proliferation of the cells and vascular bundles in the lamina joint. Reverse transcription quantitative PCR, promoter-fused β-glucuronidase expression, and subcellular localization assays indicated that OsCKX3 was highly expressed in the lamina joint, and OsCKX3-GFP fusion protein localized to the endoplasmic reticulum. The enzyme assays using recombinant protein OsCKX3 revealed that OsCKX3 prefers trans-zeatin (tZ) and isopentenyladenine (iP). Consistently, tZ and iP levels increased in the osckx3 mutants but decreased in the OsCKX3 overexpression lines. Interestingly, agronomic trait analysis of the rice grown in the paddy field indicated that osckx3 displayed a smaller leaf angle and enhanced primary branch number, grain size, 1,000-grain weight, and flag leaf size. Collectively, our results revealed that enhancing CK levels in the lamina joint by disrupting OsCKX3 negatively regulates leaf angle, highlighting that the CK pathway can be engineered to reduce leaf angle in rice and possibly in other cereals.

Introduction

Rice (Oryza sativa L.) is the major crop that provides the staple diet to more than half of the global population. Plant architecture is crucial for rice growth and yield (Sakamoto et al., 2006; Wang and Li, 2008). Leaf angle is an important agronomic trait that contributes to rice architecture and the effectiveness of leaf light interception. Erect leaves are suitable for rice dense planting, reducing mutual shading between plants to increase photosynthetic efficiency and crop yield (Sakamoto et al., 2006; Tong and Chu, 2018). Leaf angle is determined by the development of lamina joint; asymmetric cell division and elongation on the adaxial or abaxial sides of lamina joint cause leaf angle changes (Zhao et al., 2010; Feng et al., 2016).

Phytohormones play essential roles in regulating the lamina joint development and leaf angle. Phytohormones such as brassinosteroid (BR), gibberellic acid (GA), and auxin (indole-3-acetic acid [IAA]) are involved in regulating leaf angle (Luo et al., 2016; Tong and Chu, 2018; Huang et al., 2021). BRs regulate the asymmetric elongation of adaxial cells of the lamina joint to control rice leaf angle (Tong and Chu, 2018). The disruption of the components in BR biosynthesis and signaling in rice confers erect leaves but causes multiple growth defects (Yamamuro et al., 2000; Hong et al., 2003; Sakamoto et al., 2006). Rice BR-responsive module MicroRNA 159d-GAMYB-like 2 (OsmiR159d-OsGAMYBL2) regulates the expression of BRASSINOSTEROID UPREGULATED1 (BU1, a BR regulatory gene involved in BR signaling), ent-copalyl diphosphate synthase 1, and gibberellin 3β-hydroxylase 2 (CPS1 and GA3ox2, two GA biosynthesis genes), and results in a reduction in leaf angle, plant height, and grain size (Gao et al., 2018). Rice SPINDLY (OsSPY), a negative regulator of GA signaling, negatively regulates BR signaling by controlling SLENDER RICE1 (SLR1), suggesting that GA and BR interact in leaf angle regulation (Shimada et al., 2006; Bai et al., 2012; Li et al., 2012; Gao et al., 2018). Mutation of leaf inclination1 (lc1-D, a rice gain-of-function mutant) promotes cell elongation in the adaxial surface of lamina joint to increase leaf angle (Zhao et al., 2013), suggesting that IAA regulates leaf angle. As a factor involved in BR signaling, rice AUXIN RESPONSE FACTOR 19 (OsARF19) positively regulates GRETCHEN HAGEN 3-5 (OsGH3-5) and BRASSINOSTEROID INSENSITIVE 1 (OsBRI1), increasing adaxial cell division and leaf angle (Zhang et al., 2015). OsARF6 and OsARF17 are found to positively regulate leaf angle by regulating lignin biosynthesis genes and secondary cell wall formation (Huang et al., 2021). Additionally, abscisic acid (ABA) and jasmonic acid (JA) are also involved in the BR signaling pathway and regulate leaf angle (Gan et al., 2015; Gui et al., 2016; He et al., 2020; Li et al., 2021).

Cytokinins (CKs) are N⁶-substituted adenine derivatives essential for plant growth, development, and stress responses (Sakakibara, 2006). CKs play critical roles in stem cell division, panicle development, plant architecture, nitrogen fertilizer utilization, and environmental stress (Hwang et al., 2012; Wang et al., 2018; Yang et al., 2021). CK is mainly synthesized by three enzymes in plants, including the isopentenyltransferase (IPT), CK hydroxylase cytochrome P450 monooxygenases (CYP735A1/CYP735A2), and nucleoside 5′-monophosphate phosphoribohydrolases encoded by the LONELY GUY (LOG) gene (Kakimoto, 2001; Kuroha et al., 2009; Kiba et al., 2013). The CK catabolic enzymes mainly include glycosyltransferases, adenine phosphoribosyltransferase, and CK oxidase/dehydrogenase (CKX) (Sakakibara, 2006). Glycosyltransferase catalyzes the glycosylation of CK and converts it to an inactivated form (Mok et al., 2005; Sakakibara, 2006). Adenine phosphoribosyltransferase catalyzes the phosphorylation of CK adenine to reduce CK activity (Allen et al., 2002; Zhang et al., 2013). The glycosylated and phosphorylated CK can be converted to the active CK forms (Sakakibara, 2006). In contrast to these two enzymes, CKX irreversibly catalyzes CK to generate adenine or adenosine which lose their CK activity (Schmulling et al., 2003; Ashikari, 2005).

CKX genes are essential in determining crop agronomic traits (Chen et al., 2020). CK oxidase 2/Grain number 1a (OsCKX2/Gn1a) encodes a CKX that degrades CK in rice inflorescence meristem, stem, and roots (Ashikari, 2005). Allelic variations of the OsCKX2/Gn1a promote CK accumulation, which increases the number of spikelets and grains, and enhances lodging tolerance (Ashikari, 2005; Tu et al., 2022). OsCKX4 gene is regulated by OsARF25 and A-TYPE RESPONSE REGULATOR 2/3 (OsRR2/3) to control crown root development and by RICE LATERAL BRANCH (RLB)/Polycomb repressive complex 2 (PRC) complex to regulate lateral branches (Gao et al., 2014; Wang et al., 2022). Recent studies have shown that strigolactones (SLs) regulate CK catabolism by mediating the transcriptional expression of OsCKX9 during rice tillering (Duan et al., 2019). We previously showed that OsCKX11 determines grain number, and antagonizes CK and ABA during leaf senescence (Zhang et al., 2021). Although CKs were found to be decreased during rice lamina joint development (Zhou et al., 2017), the role of CKs in the development of lamina joint and leaf angle formation is largely unknown.

In this study, we found that OsCKX3 was highly expressed in lamina joint, and the fusion protein GFP-OsCKX3 was localized to the endoplasmic reticulum (ER). Enzymatic assays of the recombinant OsCKX3 protein indicated that it showed strong substrate specificities toward trans-zeatin (tZ) and isopentenyladenine (iP). In the osckx3 knockout mutants, levels of tZ and iP in lamina joint were increased, resulting in smaller leaf angles, while in OsCKX3 overexpressing lines, levels of these CKs were reduced, resulting in larger leaf angles. Moreover, field experiments showed that mutation of osckx3 reduced leaf angle and increased primary branch number, grain size, and leaf size. Together, these results reveal that CKs control leaf angle by mediating the CK homeostasis in lamina joint, providing insights into the hormonal regulation of leaf angle.

Results

CK oxidases mediate lamina joint inclination in rice

To study the functions of CK oxidases (OsCKXs) in rice, we generated 10 homozygous mutant lines of the 11 OsCKX genes by CRISPR/Cas9 technology (Zhang et al., 2021). By screening the osckx1 to 11 mutants at 7 days after germincation (DAG), osckx3-1 and osckx9-1 were found to show smaller leaf angles (Supplemental Figure S1, A and B), suggesting that CKs regulate lamina joint inclination. To further verify this finding, rice seedlings were treated with various CK forms exogenously. Compared with rice seedlings treated with mock, all CK forms negatively regulate leaf angles (Figure 1, A and B). The leaf angles of the seedlings treated with tZ and iP were most severely reduced up to 50% compared with the mock treatment, suggesting tZ and iP exhibit a stronger effect on lamina joint inclination than other CK forms (Figure 1, A and B).

Figure 1

CKs inhibited leaf angle. A, Leaf angle of 7-DAG Nipponbare seedlings treated with 6-BA, tZ, iP, cZ, and DHZ, respectively. Scale bars, 1 cm. B, Quantification of the leaf angle of the plants shown in (A). Data are means ± se (n = 25 plants). The data were analyzed by ANOVA one-way comparison by LSD test, different letters indicate a significant difference at P < 0.05.

OsCKX9 has been functionally characterized to be involved in tiller number and panicle number (Duan et al., 2019), but the function of OsCKX3 has not yet been characterized. Thus, the two allelic loss-of-function mutants osckx3-1 and osckx3-3 were selected to study the role of CK in leaf angle. DNA sequencing showed that osckx3-1 was mutated by a C insertion, while osckx3-3 was mutated by a 25-bp fragment deletion in the first extron of the OsCKX3 gene (Supplemental Figure S2, A and B). Meanwhile, OsCKX3-overexpression transgenic plants under the control of the cauliflower mosaic virus (CaMV) 35S promoter were generated. The transcript levels of OsCKX3 in these overexpression plants were quantified by reverse transcription quantitative PCR (RT-qPCR), and two representative lines with different expression levels of OsCKX3 were selected for studies (Supplemental Figure S2C). In contrast with the enhanced leaf erectness phenotype of the osckx3 mutants, the leaf angles of the OsCKX3 overexpression lines (OE1 and OE2) were enlarged (Figure 2, A and C). The loss-of-function osckx3 mutants and the OsCKX3 overexpression lines consistently demonstrated that OsCKX3 play an essential role in lamina joint inclination.

Figure 2

Knockout mutants and overexpression lines of OsCKX3 showed altered leaf angle. A, Morphological phenotypes of the 7-DAG seedlings of osckx3-1, osckx3-3, and OsCKX3 overexpression lines OE1 and OE2. Scale bars, 1 cm. B, Morphological phenotypes of the adult plants of osckx3-1, osckx3-3, and the OsCKX3 overexpression lines OE1 and OE2 at 50 DAG. Scale bars, 20 cm. C, Quantification of the leaf angle of the plants shown in (A). Data are means ± se (n = 25 plants). D, Measurement of the leaf angle of the 50-DAG plants shown in (B). Data are means ± se (n = 25 plants). E, Phenotypes of the second lamina joint of the plants at 50 DAG. Ad, adaxial and Ab, abaxial. The dashed line indicates the shape of the lamina joint. Scale bars, 5 mm. F–H, Lengths of the adaxial (F), abaxial (G), and basal sides of the lamina joints from WT, osckx3, and OsCKX3-OE-1 lines at 50-DAG. Data are means ± se (n = 10 plants). *P < 0. 01, **P < 0.01, and ***P < 0.001 generated by Student’s t test.

To confirm whether other OsCKXs might have redundant roles with OsCKX3 in regulating leaf angle, we detected their expression in the lamina joint of osckx3. Of all the 11 OsCKXs, OsCKX1, 2, 3, 4, and 9 were significantly elevated in osckx3 mutant compared with wild type (WT) (Supplemental Figure S3), suggesting they play redundant roles in CK homeostasis in lamina joint. Interestingly, the individual mutants of osckx3 and osckx9 still showed a reduced leaf angle, suggesting OsCKX9 and OsCKX3 play partially redundant roles in regulating leaf angle. However, the leaf angles of osckx1, osckx2, and osckx4 were not obviously altered (Supplemental Figure S1, A and B), indicating their function in lamina joint may be redundant.

Disruption of OsCKX3 induces asymmetric growth and development of lamina joint

We further characterized the phenotypes of the loss-of-function mutants and the overexpression lines of OsCKX3 at the tillering stage. Accordingly, it was found that the leaf angles were increased in the OsCKX3 overexpression lines but were reduced in the osckx3 mutants (Figure 2, B and D). Since the shape and size of the lamina joint determine the leaf angle, the morphology of the lamina joints of the mutant and overexpression lines was imaged and measured (Figure 2, E–H). The results showed that the adaxial side length of the lamina joint was shorter in the osckx3-1 mutant but was longer in the OsCKX3 overexpression lines than in WT (Figure 2F). Adversely, the abaxial side length of the lamina joints was significantly longer in osckx3 but was shorter in the OsCKX3 overexpression lines than in WT (Figure 2G). The length of the basal sides of the lamina joints was longer in osckx3 but shorter in the OsCKX3 overexpression lines than in WT (Figure 2H). Therefore, the asymmetric growth and development of the lamina joints in the osckx3 mutant and the OsCKX3 overexpression lines lead to different leaf angles.

OsCKX3 affects the cell proliferation and vascular bundle number on the abaxial side of lamina joint

To further understand the lamina joint development, we performed the paraffin cross-sections of the lamina joint of the second leaf (from the top) in the osckx3 mutants and the OsCKX3 overexpression lines at the tillering stage. The region between the adaxial epidermis and the adaxial central vascular bundle was defined as D1, and the region between the abaxial epidermis and the sclerenchyma was defined as D2 (Guo et al., 2021). D1 and D2 were longer in the osckx3 mutant but were shorter in the OsCKX3-OE lines than in WT (Figure 3, A–J). Consistently, the osckx3 mutants had more cells, but OsCKX3-OE lines had fewer cells in the D1 and D2 regions of the lamina joint (Figure 3, A–K). Furthermore, the vascular bundle number on the abaxial side of the lamina joint was increased in the osckx3 mutants but was decreased in the overexpression lines; however, the vascular bundle number on the adaxial side of the lamina joints of all genotypes was all similar (Figure 3L). Thus, the asymmetric changes of the cell and vascular bundle number between the abaxial and adaxial sides of the lamina joints result in different lamina joint sizes and leaf angles.

Figure 3

OsCKX3 affected the cell proliferation and vascular bundle number in the lamina joint. A–C, Transverse section of the lamina joints of WT (A), osckx3-1 (B), and OsCKX3-OE-1 (C). The rectangle box represents the abaxial or the adaxial side of the lamina joint. Scale bars, 500 μm. D–F, High magnification images of the adaxial side of the lamina joints of WT (D), osckx3 (E), and OsCKX3-OE-1 (F). Scale bars, 100 μm. G–I, High magnification images of the abaxial side of the lamina joints of WT (G), osckx3 (H), and OsCKX3-OE-1 (I). Scale bars, 100 μm. J, Lengths of the D1 and D2 of the lamina joints shown in (D)–(I). K, Quantification of the parenchyma cell layers in D1 and sclerenchyma cell layers in D2 shown in (D)–(I). Data are means ± se (n = 20 plants). L, Quantification of vascular bundles on the abaxial and adaxial sides. Data are means ± se (n = 15 plants). *P < 0. 01, **P < 0.01, and ***P < 0.001 generated by Student’s t test. Without stars represent statistically not significant difference. M, Transcriptional analysis of CYC U4;1 expression in WT, osckx3-1, osckx3-3, OE1, and OE2 by RT-qPCR. Data are means ± se (n = 3 biological replications). **P < 0.01 and ***P < 0.001 generated by Student’s t test. WT was used as the reference set for significant test.

To check the cell proliferation in the osckx3 mutant and OsCKX3 overexpression lines, we detected the transcription level of a rice U-type cyclin gene (CYC U4;1) which is involved in cell proliferation in lamina joint (Sun et al., 2015). The result showed that the CYC U4;1 gene was increased more than four times in osckx3 mutant but was decreased to ∼30% in the OsCKX3 overexpression lines compared with the WT level (Figure 3M), suggesting the OsCKX3 medicates cell proliferation via CYC U4;1 gene.

Expression pattern and subcellular localization of OsCKX3

The transcriptional expression pattern and subcellular location of osCKXs are essential to their biological functions. To detect the transcript levels of OsCKX3 in different tissues, we quantified its transcripts in the roots, stems, lamina joint, leaf blades, and young panicle of the rice plants at the heading stage by RT-qPCR (Figure 4A). It was found that OsCKX3 was expressed in the lamina joint, roots, stems, leaves, and small panicle, with the highest level detected in lamina joint (Figure 4A), which is consistent with the erect leaf phenotype of osckx3 mutants.

Figure 4

Spatiotemporal expression of OsCKX3 gene. A, Transcriptional analysis of OsCKX3 expression in various rice tissues in the Nipponbare background by RT-qPCR. Root, stems, lamina joints, leaf blades, and young panicle were harvested from rice plants. Data are means ± se (n = 3 biological replications). B, GUS staining of the seedlings of the pOsCKX3::GUS transgenic lines. Intense GUS staining was found in the lamina joint and root. Scale bars, 5 mm. C, Cross-sections of the lamina joint in (B) and the arrows indicate the vascular bundles. Scale bars, 100 μm. D–E, GUS staining in the flag leaf (D) and internode (E). Scale bars, 1 mm. F, Relative transcript levels of OsCKX3 in 7-DAG seedlings treated by IAA, BL, tZ, cZ, and iP for 3 h. OsUBQ5 was used as an internal control. Data are means ± se (n = 3 biological replications). *P < 0.01 generated by Student’s t test. Mock was used as the reference set for significant test. Without stars represent statistically not significant difference. G, Distribution of GFP-CKX3 fluorescence (left 1), AtWAK2-signal peptide-mCherry (ER marker) fluorescence (left 2) in N. benthamiana leaf epidermal cells, bright field image (left 3), and merged image (left 4). Scale bars = 20 μm.

To further confirm and clarify the tissue-specific expression of OsCKX3 in rice, we generated a construct in which the β-glucuronidase (GUS) gene was under the control of the OsCKX3 promoter. Consistently, the pOsCKX3::GUS transgenic plants showed the strongest GUS staining in lamina joints (Figure 4B). Using the GUS staining tissues of the pOsCKX3::GUS transgenic plants, we prepared the cross-section of the lamina joint to know the cellular expression of OsCKX3. The results showed that OsCKX3 was mainly expressed in the parenchyma cells and vascular bundles on the abaxial side of the lamina joint (Figure 4C). Therefore, the expression pattern of OsCKX3 in lamina joint is consistent with the erect leaf phenotype of the osckx3 mutants. In addition to the lamina joint, GUS staining was also detected in the root, flag leaf, and stem internode (Figure 4, D and E). Furthermore, we quantified the OsCKX3 gene transcripts in the seedlings treated with IAA, BR, and CK, respectively. It was found that the transcript levels of the OsCKX3 gene were increased by approximately five-, three-, and two-fold after treatments with iP, tZ, and BR, respectively (Figure 4F), indicating that OsCKX3 is possibly involved in maintaining CK homeostasis and the interaction between CK and BR.

To investigate the subcellular localization of OsCKX3, we transiently expressed a GFP-OsCKX3 fusion protein in Nicotiana benthamiana epidermal cells. The results showed that the green signal of GFP-OsCKX3 merged with the red signal of an ER marker AtWAK2-signal peptide-mCherry (Nelson et al., 2007), suggesting that GFP-OsCKX3 was sub-localized to the ER (Figure 4G). The localization of OsCKX3 is similar to that of AtCKX1 in Arabidopsis thaliana (Arabidopsis) (Niemann et al., 2018; Romanov and Schmulling, 2021).

OsCKX3 exhibits specific catalytic activities toward tZ and iP

The enzyme activity of OsCKX3 was investigated using purified recombinant maltose binding protein (MBP)-OsCKX3 fusion protein from Escherichia coli cells (Supplemental Figure S4). The catalytic activities of the recombinant MBP-OsCKX3 protein were measured using various CK forms including tZ, iP, isopentenyladenine riboside (iPR), tZ riboside (tZR), cis-zeatin (cZ), cZ riboside (cZR), and dihydrozeatin (DHZ). The substrate reduction was quantified by high-performance liquid chromatography (HPLC), and it was found that OsCKX3 had strong substrate specificities toward tZ and iP (Figure 5, A and B). The optimal conditions for the recombinant MBP-OsCKX3 protein toward tZ were 35°C and pH 7.0 (Supplemental Figure S5). Under the optimal conditions, the V_max of the recombinant MBP-OsCKX3 protein toward tZ and iP was 2.60 ± 0.15 and 0.54 ± 0.04 nmol/mg/min, respectively (Figure 5, C and D). The K_m of OsCKX3 toward tZ and iP was 1.74 ± 0.32 and 2.61 ± 0.52 μM, respectively (Figure 5, C and D). Compared with OsCKX9 and OsCKX11 (Duan et al., 2019; Zhang et al., 2021), OsCKX3 enzyme activity is less efficient but more specific to tZ and iP.

Figure 5

In vitro and in vivo enzyme activities of OsCKX3. A, Biochemical reaction catalyzed by OsCKX3 in vitro. B, Catalytic activities of recombinant protein OsCKX3 toward tZ, iP, tZR, iPR, cZ, cZR, and DHZ. Kinetics of recombinant OsCKX3 protein toward tZ (C) and iP (D). Data are means ± se (n = 3 experimental replications). E, Quantification of endogenous CK contents in the lamina joint of osckx3 mutant and OsCKX3 overexpression lines. Data are means ± se (n = 3 biological replications). *P < 0. 05 and **P < 0.01 generated by Student’s t test. WT was used as the reference set for significant test. Without stars represent statistically not significant difference. FW, fresh weight.

To further understand the biochemical functions of OsCKX3 in rice, we measured the CK levels in the lamina joint of the osckx3 mutants and OsCKX3 overexpression lines by ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). In osckx3-1 and osckx3-3, tZ and iP contents increased by ∼100% and 33%, respectively, but iPR, tZR, cZ, and cZR contents did not change significantly. Conversely, the contents of tZ and iP in the OsCKX3 overexpression were decreased by 68% and 46%, respectively, while the contents of iPR, tZR, cZ, and cZR did not differ from those of the WT lines (Figure 5E).

Disruption of osckx3 reduces leaf angle and improves other plant architecture parameters

To determine whether osckx3 disruption affects other agronomic traits, we measured these traits on a per plant basis (Figure 6, A–E). The agronomic traits of the WT, osckx3, and OsCKX3 overexpression lines grown in the paddy field were calculated at the mature stage. Compared with WT, the primary branch number, grain length, and 1,000-grain weight were significantly increased in the osckx3 mutants but decreased in the OsCKX3 overexpression lines (Figure 6, D and E). The osckx3 mutants showed no significant changes in plant height and secondary branch number (Figure 6E). The effective tiller number and grain width did not differ between WT, osckx3, and OsCKX3 overexpression lines (Figure 6E). In addition, the width of the flag leaves was mildly increased 25% in osckx3 and the leaf chlorophyll content and chlorophyll fluorescence parameter (F_v/F_m) was not changed (Supplemental Figure S6, A–F). Thus, disruption of osckx3 can improve leaf erectness, meanwhile, while also mildly improving primary branch number, grain length, 1,000-grain weight, and flag leaf size. Similar to the osckx11 mutant (Zhang et al., 2021), mutation of osckx3 reduced the fertility compared with WT (Figure 6E).

Figure 6

Agronomic traits of the mutants and the overexpression lines of OsCKX3 grown in the field. A, Morphology of osckx3 and the OsCKX3-OE lines at 110 DAG. Scale bars, 20 cm. B and C, Measurement of the leaf angle of the first flag leaf (B) and the second leaf (C) in the plants shown in (A). Data are means ± se (n = 25 plants). D, Grain shape of WT, osckx3, and OsCKX3-OE plants. Scale bars, 1 cm. E, Agronomic traits of osckx3 and the OsCKX3-OE lines. Data are means ± se (n = 15 plants for plant height, tiller number; n = 30 plants for primary branches per panicle, secondary branches per panicle, seed setting rate, and grain number per panicle; n = 10 groups for grain length, grain width, and 1,000-grain weight). WT was used as the reference set for significant test. *P < 0.01, **P < 0.01, and ***P < 0.001 generated by Student’s t test. Without stars represent statistically not significant difference.

Discussion

Leaf angle is essential in determining plant architecture and crop yield. Phytohormones play critical roles in the lamina joint development and regulate leaf angle (Chen et al., 2018; Tong and Chu, 2018). In this study, CKs were found to induce asymmetric proliferation of the cells and vascular bundles in lamina joint and negatively regulate leaf angle.

The cell number in D1 and D2 regions was both increased in the osckx3 mutants but decreased in the OsCKX3 overexpression lines (Figure 3, J and K). Meanwhile, the vascular bundle number on the abaxial side of the lamina joint was increased in the osckx3 mutant but reduced in the OsCKX3 overexpression line (Figure 3L). This observation is different from the lamina joint of osbhlh98-1, a mutant involved in lamina joint development in BR-signaling pathway, in which the cell number was increased in D1 region, but not in D2 region (Guo et al., 2021). Previously, BR was found to induce the asymmetric elongation and proliferation of the adaxial and abaxial cells and positively regulate leaf angle (Cao and Chen, 1995; Sun et al., 2015). Consistently, BR regulates cell elongation in rice lamina joint through the GA signaling pathway by mediating GA catabolism (Tong et al., 2014), and regulates cell proliferation in lamina joint through CYC U4;1 which is under the regulation of rice BRASSINAZOLE RESISTANT1 (OsBZR1) and glycogen synthase kinase 3 (GSK3)/SHAGGY-like kinase 2 (OsGSK2) at transcript and protein levels (Sun et al., 2015). In this study, CKs positively regulated CYC U4;1 gene expression in lamina joint since CYC U4;1 gene expression was increased in the osckx3 mutants but decreased in the OsCKX3 overexpression lines (Figure 3M), suggesting CK and BR act antagonistically to regulate the cell proliferation in lamina joint. On the other hand, OsCKX3 expression was found to be significantly induced by BR but not IAA (Figure 4F), suggesting that OsCKX3 is a key joint in the cross talk between CK and BR signaling in lamina joint. The molecular interactions of BR and CK in the cross talk remain to be elucidated in the future.

CKs play essential roles in determining rice architecture, such as tiller number and panicle size (Ashikari, 2005; Koprna et al., 2016; Duan et al., 2019; Zhang et al., 2021). CKX gene family maintains CK homeostasis in the plant, and different members play roles in a unique physiology process based on its expression, localization, and biochemical activities (Zhang et al., 2021). The expression of OsCKX1, 2, 3, 4, 5, 8, 9, and 11 was all detected in lamina joint, of which OsCKX1, 2, 3, 4, and 9 were significantly increased in the osckx3 mutant (Supplemental Figure S3), indicating multiple OsCKX members coordinately regulate CK homeostasis in lamina joint. Consistently, pOsCKX3::GUS transgenic plants showed the highest expression in lamina joint (Figure 4, A–C), indicating it is the major player in maintaining CK homeostasis in rice lamina joint. The subcellular localization analysis showed that GFP-OsCKX3 fusion protein was predominantly localized to the ER (Figure 4G), consistent with the recent findings of CK transporters and receptors on the ER (Romanov et al., 2018; Romanov and Schmulling, 2021). CKX enzymes located on the ER were previously proposed to maintain CK homeostasis or reset the signal concentration in the ER (Niemann et al., 2018; Romanov and Schmulling, 2021); thus, our results support the concept of the ER as an essential compartment for CK signal perception and reset. At the biochemical level, recombinant OsCKX3 has high substrate specificities toward tZ and iP (Figure 5, A–D). Consistently, tZ and iP levels in lamina joint were increased in osckx3 mutants but decreased in OsCKX3-OE lines (Figure 5E). However, tZR, iPR, cZ, and cZR levels in the lamina joint of osckx3 or overexpression lines showed no change (Figure 5E), suggesting OsCKX3 catalyzes tZ and iP specifically. In conclusion, OsCKX3 regulates CK homeostasis on the ER and regulates rice lamina joint development and leaf angle.

The detailed phenotypic analysis demonstrated that disruption of osckx3 reduced leaf angle and improved other plant architecture parameters (Figure 6E). BRs are a kind of essential phytohormones that regulate many agronomic traits such as plant height, grain size, and leaf angle (Tong and Chu, 2018; Xiao et al., 2020). Rice BR-deficient mutants generally display smaller leaf angle but have multiple defects in plant architecture, such as high reduction of plant height and small grain size (Tong and Chu, 2018). Adversely, BR-sensitive rice increases grain size and grain number but increases leaf angles, causing loose plant architectures (Tong and Chu, 2018). These awkward characteristics have been challenging for the application of BRs in molecular breeding. In this study, we found knocking out some CKX genes can increase endogenous CK level appropriately and enhance leaf erectness. Interestingly, the primary branch number, grain size, 1,000-grain weight, and flag leaf size were all mildly improved in osckx3 mutants, although the fertility was reduced (Figure 6E). We attributed this observation to broad expression of OsCKX3 in the root, leaf, small panicle, and internode which may enable OsCKX3 play pleiotropic roles (Figure 4, A–F). Given the frequently observed trade-off between grain number and size in many grass species (Guo et al., 2018; Molero et al., 2019), OsCKX3 may have the potential to be used in overcoming this shortage.

In summary, we revealed that the OsCKX3 enzyme maintains CK homeostasis in lamina joint, which induces asymmetric proliferation of the cells and vascular bundles, and increases the size of lamina joint, thus resulting in enhanced leaf erectness and stronger mechanic strength to support leaf blade (Figure 7). Moreover, osckx3 disruption may improve the primary branch number, grain size, 1,000-grain weight, and flag leaf size. Hence, this study provides insight into hormonal regulation of lamina joint development and leaf angle and provides a target gene for improving plant architecture in molecular breeding of rice and possibly in other cereals.

Figure 7

A proposed schematic model that CK homeostasis mediated by OsCKX3 regulates lamina joint development and leaf angle. The OsCKX3 enzyme irreversibly degrades CKs and maintains CK homeostasis in lamina joints. As a result of OsCKX3 degrading CK in the WT and OsCKX3 overexpression lines, CYC U4;1 gene expression, and asymmetric proliferation of parenchyma cells and vascular bundle (VB) in lamina joints are suppressed, resulting in large leaf angles. Conversely, in the osckx3 mutant, accumulated CKs may enhance CYC U4;1 gene expression, promoting asymmetric proliferation of parenchyma cells and VB in lamina joints, resulting in reduced leaf angles.

Materials and methods

Plant materials and growth conditions

Rice (O. sativa L.) ssp. Japonica (cv. Nipponbare) was used as WT. Loss-of-function mutants osckx3 were generated via Agrobacterium (Agrobacterium tumefaciens)-mediated CRISPR/Cas9 technology in the WT background Nipponbare (Zhang et al., 2021). Plant seeds germinated in water at 37°C in the dark for 3 days and were grown in the growth chamber under a 12-h light (28°C)/12-h dark (22°C) photoperiod at 500–600 μmol m⁻²s⁻¹ light intensity and 50% humidity. After 1 week, the rice seedlings were transferred to Yoshida’s nutrient solution for hydroponic culture. In the field experiments, the WT, mutants, and overexpression lines were grown under a conventional cultivation environment in a paddy field, and the independent lines were arranged in each plot in the Botany Garden of Zhejiang Normal University in Jinhua (119°63′E, 29°13′N), China.

RNA isolation and RT-qPCR

Total RNA was isolated from roots, leaves, lamina joints, culms, and panicles using the Trizol reagent (Invitrogen). The first-strand cDNA was synthesized from 1 μg of total RNA using HiScript qRT Supermix for RT-qPCR (+gDNA wiper) (R123-01; Vazyme Biotech, Nanjing, China). RT-qPCR was performed using SYBR Green (TaKaRa Bio Inc., Kusatsu, Shiga, Japan) on the qTOWER 2.2 Real-Time PCR Thermal Cycler (Analytik Jena AG, Germany) according to the manufacturer’s instructions. The primers used for RT-qPCR are P1F and P1R for OsUBQ5, and P2F and P2R for OsCKX3. All the primers used in this study are listed in Supplemental Table S1.

Plasmid construction

The target gRNA sequence of OsCKX3 was cloned into the CRISPR/Cas9 vector with the primers P3F and P3R to generate the osckx3 mutants. The primers P4F and P4R were used to identify the osckx3 mutants. OsCKX3 was amplified from cDNA using primers P5F and P5R. The coding sequence of OsCKX3 was cloned into the pCR8 vector (Invitrogen) using the sequence- and ligation-independent cloning method. The entry vector pCR8-OsCKX3 was cloned into the binary vector pMDC32 and pMDC43 to construct 35S::OsCKX3 and 35S::OsCKX3-GFP vectors by LR reaction using the Gateway LR Clonase II enzyme mix (11791-020, Invitrogen). In the same way, the OsCKX3 promoter sequence was amplified by primers P7F and P7R and cloned into the pMDC163 vector to generate the pOsCKX3::GUS construct. OsCKX3 full-length fragment was amplified using primers P6F and P6R from the pCR8-OsCKX3, and was cloned into pMAL-c2X vector to generate pMAL-OsCKX3 expression vector using Clonexpress II One Step Cloning Kit (Vazyme, Nanjing, China).

GUS staining and histologic sectioning

Histochemical GUS assays and sectioning were performed using a previously described method (Jefferson et al., 1987). Samples were infiltrated with 90% acetone for 30 min on ice and then washed with ultrapure water three times, and then were infiltrated under vacuum for 30 min in the GUS staining buffer for various times. After removing the staining buffer and clearing with 70% EtOH (v/v). The cross-section was performed following a method previously described (Zhao et al., 2019). The sample was fixed using formalin–acetic acid–alcohol (1.8% formalin [v/v], 5% acetic acid [v/v], and 90% methanol [v/v]) and embedded in paraffin. Sections with a thickness of 7 μm were cut for observation.

Subcellular localization

The subcellular localization of OsCKX3 was performed following the method previously described (Zhang et al., 2021). 35S::eGFP and 35S::eGFP-OsCKX3 plasmids were transformed into N. benthamiana leaves mediated by GV3101. The fluorescence images were observed by the Zeiss LSM 880 Confocal Microscope system (Carl Zeiss Microscopy GmbH, Jena, Germany) using an excitation 488-nm laser with an emission wavelength of 505–550 nm for GFP and 561 nm laser with an emission wavelength of 600–660 nm for mCherry.

Enzymatic assays of the recombinant protein OsCKX3

The pMAL-OsCKX3 expression vector was transformed into BL21 (DE3 Invitrogen), and the recombinant protein and MBP were purified following the previously described method (Zhang and Gan, 2012). The protein concentration was quantified according to the Bradford method. The electrophoretic separation of MBP-OsCKX3 fusion protein was performed on a 12% (w/v) polyacrylamide gel. The enzyme activities were assayed in the 200 μL reaction mixtures containing 40 µg MBP-OsCKX3 fusion protein, 5 µM CK substrate, and 0.5 mM 2,6-dichlorophentolindophenol in a 75-mM Tris/HCL buffer with different pH (ranging from 3.5 to 10) at different temperatures (ranging from 4°C to 50°C) for various time. For the assays of enzyme substrate specificities, the reactions mixture was incubated at optimized temperature and pH for 20 min and the MBP protein or boiled MBP-OsCKX3 were used as a control. For enzyme kinetics assays, different concentrations of tZ (1, 2, 4, 5, and 10 µM) or iP (1, 2, 3, 4, 5, and 10 µM) were incubated with 40 µg MBP-OsCKX3 enzyme under optimized conditions. Trichloroacetic acid was added to stop the reaction and centrifuged at 15,000 g for 20 min. The supernatant was quantified by HPLC with ultraviolet absorption at 254, 280, and 310 nm following a previously described method (Zhang et al., 2021).

Quantification of endogenous CKs

Approximately 0.02 g lamina joint tissue from 7 days after germination (DAG) rice seedlings were frozen in liquid nitrogen and powdered by a tissue lyser. Then 1.5-mL extraction buffer (80% methanol with internal standards including 45 pg [²H₅]tZ, [²H₅]tZR, [²H₆]iP, and [²H₆]iPR) was added to the powder and mixed well, then the mixture was incubated at 4°C for at least 2 h. After centrifugation (10 min, 15,000 g, 4°C), the supernatant was dried under nitrogen and redissolved in 300 μL of 30% methanol, and was filtrated though 0.22 μm membrane. The filtered extracts were quantified by a UPLC/MS/MS system (AB SCIEX/QTRAP 5500) following a previously described method (Zhao et al., 2019). Briefly, the CKs were separated by the Exion LC (AB SCIEX) equipped with the Acquity UPLC BEH C18 column (2.1 × 100 mm, particle size of 1.7 µm). The column was maintained at 40°C and the mobile phases were composed of water (A) and methanol (B) for CK using a multi-step linear gradient elution: 5% B at 0–2.5 min, 5%–0% B at 2.5–3 min, 20%–50% B at 3–12.5 min, 50%–100% B at 12.5–13 min, 100% B at 13–15 min, 100%–5% B at 15–15.2 min, and 5% B at 15.2–18 min. The CKs were analyzed by the triple quadruple mass spectrometer QTRAP 5500 system. The optimized conditions were as follows: curtain gas, 40 psi; ion spray voltage, 5,000 V for positive ion mode; turbo heater temperature, 600°C; nebulizing gas (Gas 1), 60 psi; heated gas (Gas 2), 60 psi. Analyst software (version1.6.3, AB SCIEX) was used to control the instrument and to acquire the MS data. The data analysis was processed by MultiQuant software (version3.0.2, AB SCIEX).

Hormone treatment and leaf angle measurement

To determine the relative transcript levels of OsCKX3 in response to different phytohormone treatments, the plants at 7 DAG were treated by 1 µM IAA, 1 µM 24-epibrassinolide (BL), 1 µM tZ, 1 µM cZ, and 1 µM iP in nutrient solution at 30°C for 3 h. The whole seedlings were harvested and immediately transferred into liquid nitrogen for RNA extraction and RT-qPCR analysis. OsUBQ5 was used as an internal control.

Leaf angles of the 7 DAG seedlings were photographed and measured with Image J. For leaf lamina inclination assays, 6-benzylaminopurine (6-BA), tZ, iP, cZ, and DHZ were added to the Yoshida’s nutrient solution at the final concentration of 1 µM. Seedlings at 7-DAG were treated for 4 days and leaf angles were measured. A protractor was used to measure the leaf angles of the plants grown in the field.

Chlorophyll content index and chlorophyll fluorescence assays

The leaf chlorophyll content index was measured with a SPAD meter (SPAD-502 plus, Japan). The chlorophyll fluorescence parameter (F_v/F_m) was measured with a portable fluorescence system (model OS1p, Opti-Sciences, Hudson, New Hampshire, USA) using the fully expanded rice flag leaves at 120 DAG.

Statistical analyses

The Student’s t test and one-way analysis of variance by LSD test were used to analyze differences between plants and/or treatments by SPSS v. 13.0 (IBM Corp., Armonk, New York, USA). In the figures, the different letters or stars above each column indicate the significant differences according to the test. Without stars represent statistically not significant difference.

Plant architecture analysis

The rice architecture parameters including plant height, tiller number, primary branches per panicle, secondary branches per panicle, seed setting rate, grain number per panicle, grain length, grain width and 1,000-grain weight were measured on a single-plant basis. To measure the grain size, ∼150 seeds per plant and 10 plants of each line were photographed using high-speed photographic equipment (Eloam, Shenzhen, China). The images were processed and analyzed using SC-G software (Hangzhou Wanshen Detection Technology Co., Ltd., Hangzhou, China, www.wseen.com). To image the grain size, 10 seeds were spread on the glass plate of a flatbed scanner (Microtek i600, Shanghai, China). The scanner was set in the reflection imaging mode, RGB color, and 600 dpi.

Accession numbers

Sequence data from this article are found in the rice genome annotation project by the following accession numbers: OsCKX1 (Os01g0187600), OsCKX2 (Os01g0197700), OsCKX3 (Os10g0483500), OsCKX4 (Os01g0940000), OsCKX5 (Os01g0940000), OsCKX6 (Os02g0220000), OsCKX7 (Os02g0220100), OsCKX8 (Os04g0523500), OsCKX9 (Os05g0374200), OsCKX10 (Os06g0572300), OsCKX11 (Os08g0460600), CYC U4;1 (Os10g0563900), and OsUBQ5 (Os01g0328400).

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Screening for mutants with altered leaf angle from the rice seedlings of OsCKX loss-of-function mutants osckx1–11.

Supplemental Figure S2. Characterization of the loss-of-function mutants and overexpression lines of OsCKX3.

Supplemental Figure S3. The relative expression levels of OsCKX1–11 in the lamina joint of osckx3-1.

Supplemental Figure S4. Purification of the recombinant MBP-OsCKX3 protein.

Supplemental Figure S5. Optimization of the temperature and pH for the recombinant OsCKX3 enzyme in vitro.

Supplemental Figure S6. Growth of the flag leaves of the WT, osckx3 mutants, and the OsCKX3-OE lines.

Supplemental Table S1. Primers used in this study.

Acknowledgments

We thank Prof. Xuelu Wang in Henan University for sharing the BR-related materials with us. We also thank Ms. Xiaoxian Zhu and the other members in Zhang laboratory for their critical reading of the manuscript or the assistance in the experiments.

Funding

This study was supported by the National Natural Science Foundation of China. (Grant 32100270 to J.Z.), the Natural Science Foundation of Zhejiang Province (Grant LQ20C130004 to P.H. and LY22C020003 to J.Z.), and the National Key R&D Program of China (2016YFD0100901 to K.Z.).

Conflict of interest statement. The authors have no conflicts of interest to declare.

K.Z. conceived and designed the experiments. P.H., J.Z., J.H., B.Z., S.X., E.Z., and P.F.H. performed the experiments. K.Z., P.H., J.Z., J.H., B.Z., S.X., E.Z., and P.F.H. analyzed the data. P.H., J.Z., and K.Z. wrote and revised the manuscript. All the authors discussed the results and collectively revised the manucript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Kewei Zhang ([email protected]).

Author notes

These authors contributed equally (P.H. and J.Z.)