An SPX-RLI1 Module Regulates Leaf Inclination in Response to Phosphate Availability in Rice

SPX and RLI1 regulate lamina joint cell expansion, thus increasing leaf inclination in response to Pi deficiency in rice.

Abstract

Leaf erectness is one of the key traits of plant architecture; in grains, plants with upright leaves can be planted close together, thus benefiting yield/unit area. Many factors, such as hormones, affect leaf inclination; however, how nutrition status, in particular phosphate (Pi) availability, affects leaf inclination remains largely unexplained. Here, we show that in rice (Oryza sativa), Pi deficiency stress inhibits lamina joint cell elongation, thus restricting lamina joint size and inducing leaf erectness in rice. The Pi starvation-induced proteins SPX1 (for Syg1/Pho81/XPR1) and SPX2 play a negative role in the regulation of leaf inclination. We further identified an SPX1-interacting protein, REGULATOR OF LEAF INCLINATION1 (RLI1), which positively regulates leaf inclination by affecting lamina joint cell elongation in rice. The rli1 mutants showed reduced leaf inclination and the RLI1 overexpressors showed increased leaf inclination. RLI1 directly activates the downstream genes BRASSINOSTEROID UPREGULATED1 (BU1) and BU1-LIKE 1 COMPLEX1 to control elongation of the lamina joint cells, therefore enhancing leaf inclination. We also found that Pi deficiency repressed the expression of RLI1. SPX1 protein interacts directly with RLI1, which could prevent RLI1 binding to the promoters of downstream genes. Therefore, SPX and RLI1 form a module to regulate leaf inclination in response to external Pi availability in rice.

INTRODUCTION

Leaf angle is an important agronomic trait in monocotyledonous plants, like rice (Oryza sativa), and contributes to both plant architecture and grain yields (Hoshikawa, 1989; Sinclair and Sheehy, 1999; Wang and Li, 2011). An erect-leaf trait in rice renders the plants more suitable for dense planting by enhancing the efficiency of sunlight capture for the seedling population (Sinclair and Sheehy, 1999). In rice, the leaf sheath and leaf blade are linked by the lamina joint (Hoshikawa, 1989). In the lamina joint, the parenchyma cells constitute the basic structure and abundant sclerenchyma tissues provide the mechanical support. Structurally, the parenchyma cells connect the abaxial large vascular bundles with adaxial side and the sclerenchyma cells are located at the fringe of vascular bundles (Zhou et al., 2017). During leaf development, after the complete elongation of the leaf blade and sheath, the blade diverges away from the vertical leaf sheath along with the development of the lamina joint to form the angle of leaf inclination (Hoshikawa, 1989). During this process, a number of factors, especially phytohormones, have been reported to affect the development of the lamina joint, thus modulating the extent of leaf inclination in rice.

Phytohormones are crucial players in the regulation of leaf inclination. For example, it has been reported that the hormone brassinosteroid (BR) regulates rice leaf inclination by inhibiting abaxial cell division and promoting adaxial cell elongation in the lamina joint (Sun et al., 2015). Consistent with this finding, the BR biosynthesis mutants in rice, dwarf4-1 (Sakamoto et al., 2006), ebisu dwarf (d2; Hong et al., 2003), and dwarf1 (Hong et al., 2005) and the signaling mutant d61-7 (Bai et al., 2007) display erect leaves. Gibberellin metabolism has been found to be modulated by BR, and thus is involved in the regulation of leaf inclination (Tong et al., 2014). Reducing the expression of the gibberellin negative regulator SPINDLY could increase the angle of leaf inclination (Shimada et al., 2006). The auxin indole-3-acetic acid also influences leaf inclination at high concentrations and has a synergistic effect with BR (Wada et al., 1981; Cao and Chen, 1995). Repressing the expression of auxin regulators OsAFB2 and OsTIR1 has been found to enhance leaf inclination (Bian et al., 2012). It has also been suggested that abscisic acid could modulate BR homeostasis to affect leaf inclination, since the gain-of-function epiallele of rice, RELATED TO ABSCISIC ACID INSENSITIVE3/VIVIPAROUS1 6, displayed greater leaf inclination with defects in BR homeostasis (Zhang et al., 2015). These indicate that phytohormones coordinate to regulate cell proliferation and cell elongation in the lamina joint and therefore affect leaf inclination in rice.

Downstream of phytohormone signaling, a number of transcription factors are active in directing lamina joint cell division and elongation and thus act to affect the extent of leaf inclination in rice. LC2, a vernalization insensitive 3-like protein, negatively regulates adaxial cell division in the lamina joint for leaf inclination (Zhao et al., 2010). The HLH transcription factors INCREASED LAMININAR INCLINATION1 (ILI1), BRASSINISTEROID UPREGULATED1 (BU1), BU1-LIKE1 (BUL1), and BUL1 COMPLEX1 (BC1) enhance leaf inclination by promoting elongation of lamina joint cells (Tanaka et al., 2009; Zhang et al., 2009; Jang et al., 2017). Overexpressing rice ILI1 BINDING HLH1 (IBH1) results in erect leaves (Zhang et al., 2009). IBH1 can interact with BC1 and BUL1 to inhibit their function in regulating cell elongation (Jang et al., 2017). These transcription factors form a regulatory network to integrate internal developmental cues to modulate leaf inclination in rice.

In addition to the regulation of leaf inclination by developmental programs, external environmental cues also modulate leaf erectness in rice. For example, silicon (Si) can deposit on the leaves to promote erect leaves in rice (Ma et al., 2001). As one of the essential macronutrients for plant growth, development, and reproduction, phosphate (Pi) availability also affects rice leaf inclination. The most distinctive characteristic of Pi-deficient rice is erect spindly leaves with minimal tillers on the plant; despite that, the contribution of Pi deficiency stress-induced leaf erectness to yield loss remains to be demonstrated (Mghase et al., 2011). However, the molecular mechanisms for determination of leaf inclination by environmental cues remain largely unknown.

Plants absorb phosphorus almost exclusively in the inorganic form (Pi). However, due to its propensity to form insoluble precipitates with metal cations in soil, the amount of Pi in the soil available to plants is often limiting. Pi deficiency stress elicits a number of developmental and biochemical adaptations in plants. Recently, an elegant Pi-signaling network based on the central regulators PHOSPHATE STARVATION RESPONSE1 (PHR1) subfamily members and SPX domain containing proteins (for Syg1/Pho81/XPR1) were identified in the regulation of Pi homeostasis and adaptation (reviewed in Chiou and Lin, 2011; Wu et al., 2013). In the network, AtPHR1 (in Arabidopsis thaliana) or OsPHR2 (in rice) activates the phosphate starvation-induced genes by binding to the P1BS (PHR1 Binding site, GNATATNC) cis-elements (Rubio et al., 2001; Zhou et al., 2008; Bustos et al., 2010; Guo et al., 2015). The malfunction of PHR in Arabidopsis or rice can result in the repression of Pi starvation signaling and the disruption of Pi homeostasis (Rubio et al., 2001; Bustos et al., 2010; Guo et al., 2015). Overexpression of AtPHR1 or OsPHR2 results in Pi overaccumulation in the shoots while inducing Pi starvation signaling even under Pi-sufficient conditions (Zhou et al., 2008; Bustos et al., 2010). In Arabidopsis, the Pi starvation-induced AtSPX1/2 can interact with AtPHR1 to inhibit the transcription activity of AtPHR1 in a Pi-dependent manner (Puga et al., 2014). In rice, a similar regulation mechanism has been observed under the control of OsSPX1/2 and OsPHR2 (Wang et al., 2014). However, whether these regulators participate in the regulation of leaf inclination in response to the Pi availability in rice is still unknown.

Here, we characterized the cellular basis of Pi-deficient stress-induced leaf erectness in rice. We further identified an SPX1 interacting protein RLI1 (REGULATOR OF LEAF INCLINATION1), encoding an HTH_MYB-like transcription factor, which positively regulates leaf inclination by affecting the elongation of lamina joint cells. RLI1 binds to the NNAKATNC cis-element to regulate the transcription of BU1 and BC1 to induce lamina joint cell elongation. The transcription activity of RLI1 is negatively regulated by SPX1. Therefore, Pi deficiency stress induced SPXs and repressed RLI1 form a regulatory module to regulate leaf inclination in response to Pi availability in rice.

RESULTS

Pi Deficiency Stress Inhibits the Cell Elongation of Lamina Joints to Reduce Leaf Inclination

The erect-leaf phenotype is a typical symptom of rice cultivated in Pi-deficient soil (Mghase et al., 2011). To understand the mechanism underlying the Pi-deficient stress-induced leaf erectness in rice, the leaf inclinations of wild-type plants grown under Pi-sufficient (+P) and Pi-deficient (–P) conditions were analyzed. It was found that the –P treatment leads to a 60% reduction in leaf inclination. Further observation revealed that the lengths of the marginal (d1) and middle sites (d2) of the lamina joint from plants grown under –P conditions were around 40 to ∼50% shorter than those grown under +P condition (Figures 1A to 1E). This suggests that Pi deficiency stress might inhibit the length of the lamina joint and thus decrease the leaf inclination.

Figure 1.

Pi Deficiency Stress Inhibits Lamina Joint Cell Elongation to Induce Leaf Erectness.

(A) Pi-deficient stress induces leaf erectness in rice. The 10-d-old wild-type seedlings cultured in Pi-sufficient (+P) conditions were transferred to +P and Pi deficient (–P) solutions and cultured for a further 20 d. The lamina joint of the third fully expanded leaf (as indicated by the red box) from the top of the main tiller was used for observation.

(B) The side views of the lamina joints indicated in (A).

(C) A 3D schematic of lamina joint. d1 indicates the length of adaxial side of lamina joint in vertical direction; d2 indicates the length of abaxial side of lamina joint in vertical direction; blue dotted line indicates the longitude section position; yellow dotted line indicates the transverse section position.

(D) Measurement of the length of the marginal and middle sites of the lamina joints (shown in Figure 1B). Error bars are sd (n = 20). The P values were determined by Student’s t test. The significant levels are as follows: **P < 0.01.

(E) Quantification of the leaf inclination of wild-type plants under +P and –P conditions. The third fully expanded leaf from the top of main stems was used for the quantification. Error bars are sd (n = 20). Student’s t test: **P < 0.01.

(F) Longitudinal and transverse section of the lamina joints under +P and –P conditions. The red box indicates the abaxial and adaxial sides of lamina joint respectively. m1 indicates the region between the adaxial epidermis and sclerenchyma in the middle area of the cross section; m2 indicates the region between the abaxial epidermis and the abaxial central vascular bundle (vb) of the cross section. Bars = 500 μm in 1 to 4 and 50 μm in 5 to 12.

(G) Measurement of lamina joint abaxial and adaxial cell lengths under +P and –P conditions (shown in 5, 6, 9, and 10 in [F]). Error bars are sd (n = 60 to 80). Student’s t test: **P < 0.01.

(H) Measurement of the lengths of m1 and m2 regions under +P and –P conditions (shown in 7 and 11 in [F]). Error bars are sd (n = 15). Student’s t test: **P < 0.01.

(I) The number of sclerenchyma cell layers in the m1 and m2 regions under +P and –P conditions (shown in 11 and 12 in [F]). Error bars are sd (n = 15).

BR deficiency induces the proliferation of lamina joint abaxial sclerenchyma cells and inhibits adaxial cell expansion resulting in leaf erectness (Sun et al., 2015). To test whether Pi deficiency stress modulates leaf inclination in a similar way, we set to determine the cellular mechanism. The cytological observation showed that, in contrast to the mechanism by which BR signaling results in leaf erectness, Pi deficiency stress inhibits both abaxial and adaxial cell proliferation and expansion in the lamina joint, which is reflected in the reduced lengths and cell layers of m1 (the region between the adaxial epidermis and sclerenchyma in the middle area of the cross section) and m2 (the region between the abaxial epidermis and the abaxial central vascular bundle of the cross section) (Figures 1F, 1H, and 1I). Longitudinal sections showed that the length of both the abaxial and adaxial sclerenchyma cells in the lamina joint were lower in plants grown under –P conditions, compared with +P conditions (Figures 1F and 1G). This indicates that in rice, Pi deficiency stress induces leaf erectness by inhibiting the elongation of lamina joint cells, resulting in reduced lamina joint length and restricting the ability of the leaf to bend.

SPX1 and SPX2 Function Redundantly in Regulating Leaf Inclination in Rice

To determine the possible molecular mechanism for Pi-deficiency-induced leaf erectness in rice, we first studied the leaf inclination of rice mutants defective in Pi signaling. PHR2, a MYB transcription factor, is the central regulator of both Pi signaling and Pi homeostasis in rice (Zhou et al., 2008). We found that the leaf inclination of the phr2 mutant is comparable to that of the wild type under both +P and –P conditions (Supplemental Figure 1), which indicates that the PHR2 might not be involved in the modulation of leaf inclination in rice. We further determined whether SPX1 and SPX2, key players in regulating Pi signaling and Pi homeostasis in rice (Wang et al., 2014), are involved in the regulation of leaf inclination. We found that the angles of leaf inclination of the spx1 and spx2 mutants were slightly greater than that in the wild type (Figure 2A). The angle of leaf inclination of the spx1 spx2 double mutant is twice that of the wild type (Figures 2A and 2C). These findings suggest that SPX1/2 might be involved in the regulation of leaf inclination in rice. To further characterize the role of SPX1 and SPX2, we also measured the leaf inclination of SPX overexpressors. Consistent with the negative role of SPX1/2 in regulating leaf inclination, the SPX1 and SPX2 overexpressors displayed a relatively erect leaf phenotype compared with the wild type (Figures 2A and 2C).

Figure 2.

SPX1 and SPX2 Negatively Regulate the Lamina Joint Length for Leaf Inclination in Rice.

(A) and (B) Lamina joint phenotypes of 21-d-old wild-type, spx1, spx2, spx1 spx2, SPX1-OE-1, SPX1-OE-2, SPX2-OE-1, and SPX2-OE-2 overexpression plants. Bars = 2 cm in (A) and 2 mm in (B).

(C) Quantification of the leaf inclination of wild-type, spx1, spx2, spx1 spx2, SPX1-OE-1, SPX1-OE-2, SPX2-OE-1, and SPX2-OE-2 plants (shown in [A]). The third fully expanded leaf from the top of the main stems was used for analysis.

(D) The lengths of lamina joint at middle site (d2) of wild-type, spx1, spx2, spx1 spx2, SPX1-OE-1, SPX1-OE-2, SPX2-OE-1, and SPX2-OE-2 plants (shown in [B]). The third fully expanded leaf from top of the main stems was used for the measurement. Data significantly different from the control are indicated. Error bars are sd (n = 20). P values were determined by Student’s t test. Mutants versus the wild type: *P < 0.05 and **P < 0.01. Overexpressors versus the wild type: #P < 0.05.

Given that –P treatment represses lamina joint elongation, resulting in an increase of the lamina inclination (Figure 1), we further determined whether the alteration of leaf inclination in the SPX mutants and overexpressors resulted from the effects on the lamina joint length. We found that the length of middle site of lamina joint (d2) in the spx mutants was significantly longer as compared with wild type, while those of the SPX1/2 overexpressors were shorter (Figures 2B and 2D). These results suggested that SPX1/2 also regulates the elongation of the lamina joint to affect leaf inclination. Further observation showed that both the adaxial and abaxial sclerenchyma cell lengths of the spx mutants were longer than those in the wild type, while those in the SPX1 overexpressors were shorter (Supplemental Figure 2). Together it was found that the negative regulators of Pi signaling, SPX1 and SPX2, play a negative role in regulating leaf inclination in rice.

RLI1 Physically Interacts with SPX1

Given that SPX proteins function as negative regulators by directly interacting with transcription factors (TFs) such as AtPHR1 and OsPHR2 to inhibit their transcriptional activity (Lv et al., 2014; Puga et al., 2014; Shi et al., 2014; Wang et al., 2014), we hypothesized that SPX1 and SPX2 may interact with some unknown TFs to regulate leaf inclination. To identify the possible TF that regulates leaf inclination, we employed the yeast two-hybrid (Y2H) system. The chimeric bait vector SPX1-BD was used to screen a cDNA library from rice shoot. Multiple independent cDNAs encoding a putative transcription factor (Os04g56990; designated as RLI1) were identified from the screening. RLI1 is a putative HTH-MYB transcription factor, showing similarity with the PHR family proteins. However, in contrast to the PHR proteins, RLI1 has no coil-coil domain (Supplemental Figure 3).

To verify the initial library screening, we used SPX1 as bait to test the interaction with RLI1. Y2H assays showed that SPX1 could interact with RLI1 and its MYB domain in yeast cells (Figure 3A; Supplemental Figure 4). Further analysis showed that SPX1 could form a homodimer in yeast cells, while RLI1 could not. To further validate the Y2H results, we conducted bimolecular fluorescence complementation assays in Nicotiana benthamiana leaves. The coexpression of SPX1-YFP^C with RLI1-YFP^N or SPX1-YFP^N led to a fluorescence signal in the nucleus, while coexpression of RLI1-YFP^C with RLI1-YFP^N did not produce a detectable fluorescence signal (Figure 3B). These results demonstrated that RLI1 interacts with SPX1 in the nucleus, but could not form a homodimer as SPX1 does. To further verify these results, we performed in vivo coimmunoprecipitation assays in N. benthamiana leaves cotransformed with RLI1-FLAG and SPX1-GFP, and found that RLI1-FLAG interacts with SPX1-GFP but not with the control GFP alone in planta (Figure 3C). Together, these results showed that RLI1 physically interacts with SPX1 in vitro and in vivo.

Figure 3.

RLI1 Interacts with SPX1 in Vitro and in Vivo.

(A) Y2h assay showing RLI1 interaction with SPX1. –Trp –Leu, medium without tryptophan (–Trp) and leucine (–Leu). –Trp –Leu –His –Ade, medium without tryptophan (–Trp), leucine (–Leu), histidine (–His), and adenine (–Ade).

(B) Bimolecular fluorescence complementation analysis showing that RLI1 interacts with SPX1 in the nucleus of N. benthamiana leaf epidermal cells. Nucleus stained with 4′,6-diamidino-2-phenylindole (DAPI). Bars = 50 μm.

(C) The coimmunoprecipitation assays showed that RLI1 interacts with SPX1 in planta. Protein extracts (Input) were immunoprecipitated with anti-FLAG magnetic beads (IP; Sigma-Aldrich). Immunoblots were developed with anti-FLAG antibody (Sigma-Aldrich) to detect RLI1 and with anti-GFP (Sigma-Aldrich) to detect SPX1. Molecular mass markers are shown (kD).

RLI1 Positively Regulates Leaf Inclination by Affecting Elongation of the Lamina Joint Cells

Given that SPX1 negatively regulates leaf inclination and RLI1 can interact with SPX1, we tested whether RLI1 could regulate leaf inclination in rice. First, we generated rli1 mutants using the transcription activator-like effector nucleases (TALENs) system. Two homozygous RLI1 mutants of rli1-1 and rli1-2 were obtained (Supplemental Figure 5). Both rli1-1 (deletion of 5 bp at 1464 from the start codon ATG) and rli1-2 (deletion of 2 bp at 1467 from the start codon ATG) were loss-of-function mutants due to the introduction of a premature stop codon. Compared with the wild type, the overall shoot appearance of the rli1-1 and rli1-2 mutants was more compact (Figure 4A). Leaf inclination measurement showed that the leaf angle of rli1 mutants was obviously smaller than the wild type (Figure 4B). These results indicated that RLI1 might positively regulate leaf inclination.

Figure 4.

RLI1 Positively Regulates Leaf Inclination.

(A) Phenotypic performance of 60-d-old wild-type (WT), rli1-1, and rli1-2 plants grown in pot experiments.

(B) The RLI1 mutants display smaller leaf inclination than that in the wild type. Three leaves from the top of the main stem were used for the leaf inclination quantification. 1st indicates the first fully expanded leaf from the top; 2nd indicates the second fully expanded leaf from the top; 3rd indicates the third fully expanded leaf from the top. Values represent means ± sd of 10 replicates. Data significantly different from the corresponding controls are indicated (**P < 0.01, Student’s t test).

(C) Phenotypic performance of 60-d-old wild type and the RLI1 overexpressors grown in pot experiments.

(D) The RLI1 overexpressors display larger leaf inclination than that in the wild type. Three leaves from the top of main stem were used for the leaf inclination quantification. 1st indicates the first fully expanded leaf from the top; 2nd indicates the second fully expanded leaf from the top; 3rd indicates the third fully expanded leaf from the top. Values represent means ± sd of 10 replicates. Data significantly different from the corresponding controls are indicated (**P < 0.01, Student’s t test).

To further verify the effect of RLI1 on leaf inclination, we generated RLI1 overexpressors. Three independent transgenic lines of 35:RLI1-FLAG (designated as RLI1-OE-3, RLI1-OE-7, and RLI1-OE-10), showing significantly higher RLI1 expression than the wild type (Supplemental Figure 5), were chosen for further analysis. In contrast with the loss-of-function mutants, the overall shoot appearance of the RLI1 overexpressors is obviously looser compared with that of the wild type (Figure 4C). The leaf inclination of the RLI1overexpressors was significantly greater than that of the wild type with the 2nd and 3rd leaf inclinations being twice those of the wild type (Figure 4D). These results demonstrated that RLI1 positively regulates leaf inclination in rice.

To further understand how RLI1 regulates leaf inclination, we observed the anatomical structure of the lamina joint in wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants (Figures 5A and 5B). We found that the lamina joint length of rli1-1 was obviously shorter, while those of the RLI1 overexpressors were significantly longer than those of the wild type (Figure 5F). Given that the lamina joint is a physical connection for the leaf sheath and leaf blade, its length can affect the ability of the leaf blade to bend. This indicates that the shorter lamina joint length in the rli1-1 mutants might restrict bending of the leaf blade and therefore lead to increased leaf erectness in the mutant (Figures 5A, 5B, and 5E). By contrast, the longer lamina joint in the RLI1 overexpressors gives more space for leaf blade bending, which leads to enlarged leaf inclination (Figures 5A, 5B, and 5E).

Figure 5.

RLI1 Regulates Leaf Inclination by Affecting Lamina Joint Cell Elongation.

(A) Lamina joint phenotype of 30-d-old wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants. Bar = 10 cm.

(B) The lamina joints of the third fully expanded leaf from top of main stems of 30-d-old wild type, rli1-1, RLI1-OE-3, and RLI1-OE-7. Bar = 2 mm.

(C) and (D) Longitudinal sections of the lamia joints of the third fully expand leaf from wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants. (C) is the abaxial longitudinal section of the lamia joint. (D) is the adaxial longitudinal section of the lamia joint. Bars = 50 μm.

(E) Quantification of the leaf inclination in wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants. Values for leaf inclination are given as means ± sd (n = 15 to 20).

(F) The lengths of lamina joints at the middle site (d1) and marginal site (d2) of the wild type, rli1-1, RLI1-OE-3, and RLI1-OE-7. Values represent means ± sd (n = 15 to 20).

(G) Lamina joint adaxial and abaxial cell lengths. Values represent means ± sd (n = 30 to 50). Data significantly different from the corresponding controls are indicated (mutant versus the wild type, ##P < 0.01; overexpressors versus the wild type, ***P < 0.001; Student’s t test)

To determine whether the alteration of the leaf lamina joint length is due to the effect of cell elongation or cell division, we further observed the cytological composition of longitudinal and transverse sections of the lamina joint in wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants (Figures 5C and 5D). We found that the lengths of the adaxial and abaxial sclerenchyma cells in the rli1-1 and the RLI1-OE plants were significantly affected, with those of the RLI1 overexpressors and rli1-1 being larger and shorter, respectively, than those of the wild type (Figures 5F and 5G). However, the number of adaxial and abaxial sclerenchyma cell layers of both the rli1-1 and RLI1 overexpressors were comparable to those in the wild type (Supplemental Figure 6). No obvious difference for the shoot and root lengths and leaf epidermal cell size could be detected between the rli1 mutant and the wild type at seedling stage (Supplemental Figure 7). Together, these results show that RLI1 can positively regulate lamina joint cell elongation and lamina joint length to modulate leaf inclination in rice.

RLI1 Directly Promotes the Transcription of BU1 and BC1 to Regulate Leaf Inclination in Rice

Given the positive role of RLI1 in regulating leaf inclination in rice, we further sought to determine the possible molecular mechanism. First, we characterized the DNA binding property of RLI1. The RLI1 is a putative TF with an HTH_MYB-like domain, which showed high similarity to the PHR1 subfamily TFs in rice (Supplemental Figure 3). However, in contrast to the PHR1-subfamily proteins, there is no conserved coil-coil domain present in RLI1 (Supplemental Figure 3). This indicates that RLI1 might bind to a similar cis-element as that found in the PHR1 subfamily. To test this hypothesis, the interaction between RLI1 and P1BS cis-element, the typical binding sequences of PHR1 subfamily proteins (Rubio et al., 2001; Bustos et al., 2010; Guo et al., 2015; Ruan et al., 2017), was examined by electrophoretic mobility shift assay (EMSA). The EMSAs showed that RLI1 could bind to the P1BS motif (Figure 6A). To further determine the exact binding site of RLI1, the eight bases of P1BS motif (¹G ²N³A ⁴T ⁵A ⁶T ⁷N ⁸C) were mutated one by one to examine the preference of the RLI1. It showed that the 1st (G), 2nd (N), and 7th (N) bases are variable for RLI1 recognition and the 4th base could be A or G. However, the 3rd (A), 5th (A), 6th (T), and 8th (C) bases are essential for recognition by RLI1 (Figure 6B). Basing on these results, we suggest that RLI1 recognized the core bases of NNAKATNC (designed here as R1BS) (Figure 6C).

Figure 6.

RLI1 Directly Binds to the BC1 and BU1 Promoters.

(A) EMSA analysis showing that RLI1 binds to the P1BS cis-element. The DNA fragments containing P1BS were incubated with RLI1-His recombinant proteins as indicated. 6xHis proteins were used as negative controls.

(B) RLI1 can bind to the core DNA sequences of NNAKATNC. EMSA detects the interaction between RLI1 and different P1BS mutant versions; wild-type P1BS (GAATATGC) was used as positive control.

(C) Logo analysis of RLI1 recognition bases. Logo analysis was performed at http://weblogo.threeplusone.com.

(D) Relative expression levels of BU1 and BC1 in the lamina joints of wild-type, rli1-1, RLI1-OE-3, and RLI1-OE-7 plants. Lamina joint tissues from 21-d-old plants were harvested for RNA extraction and RT-qPCR analyses. Values represent means ± sd (n = 3). Data significantly different from the control are indicated. P values were determined by Student’s t test. Mutants versus the wild type: #P < 0.05; overexpressors versus wild the type: **P < 0.01.

(E) and (F) ChIP-qPCR assays showing that RLI1 associates with the promoters of BU1 and BC1 in vivo. Lamina joint tissues of 4-week-old wild type and 35S:RLI1-FLAG (RLI1-OE-3) were harvested for ChIP analysis. Enriched DNA fragments in the BU1 and BC1 promoters were quantified using RT-qPCR. The red lines indicate the candidates of RLI1 recognition sites in the promoters of BU1 and BC1. Values represent means ± sd (n = 3). Data significantly different from the control are indicated. P values were determined by Student’s t test. RLI1-FLAG versus No RLI1-FLAG: **P < 0.01 and *P < 0.05.

Our results showed that RLI1 regulates the leaf inclination by enhancing the lamina joint cell elongation. Thus, we sought to identify possible target genes that have R1BS in the promoter and are involved in lamina joint cell elongation in rice. BU1 and BC1 are key factors that are mainly expressed in the lamina joint and mediate the lamina joint cell elongating in rice (Tanaka et al., 2009; Jang et al., 2017). Two and four putative R1BS were detected in the promoters of BU1 and BC1, respectively (Figures 6E and 6F). We hypothesized that the RLI1 might bind to these R1BS to activate the BU1 and BC1 expression for lamina joint cell elongation, thereby modulating leaf inclination. Consistent with this hypothesis, the transcript levels of BU1 and BC1 were induced in the lamina joint of RLI1 overexpressors and repressed in those of the rli1-1 mutants (Figure 6D). These results indicate that BU1 and BC1 might be the direct target genes of RLI1. To test this hypothesis, chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) experiments were performed to determine whether the RLI1 could bind to the promoter of BU1 and BC1 in vivo. Chromatin from the RLI1 overexpressors was immunoprecipitated with an antibody to FLAG, and DNA was then amplified using primers surrounding putative R1BS in the promoter regions of BU1 and BC1. Clear enrichment of an S1 fragment and the S3 and S4 fragments can be detected with BU1 and BC1, respectively (Figures 6E and 6F). This indicates that RLI1 activates the transcription of BU1 and BC1 by directly binding to the R1BS on their promoters.

Given that BU1 and BC1 positively regulate lamina joint cell elongation and are directly activated by RLI1, we hypothesized that RLI1 regulates leaf inclination through BU1 and BC1. To test this hypothesis, we generated BU1 and BC1 loss-of-function mutants in both the wild-type and RLI1-OE-3 backgrounds by clustered regularly interspaced short palindromic repeat (CRISPR)-CRISPR associated 9 (Cas9) system technique (Supplemental Figure 8).

Consistent with the previous reports, both of the bu1 and bc1 mutants that we generated display erected leaves, with reduced lamina joint adaxial and abaxial sclerenchyma cell length (Tanaka et al., 2009; Jang et al., 2017). The enlarged leaf inclination in the RLI1 overexpressors was significantly suppressed by the introduction of a bu1 or bc1 mutation (Figures 7A to 7C). In agreement with the leaf inclination phenotype, the induced adaxial and abaxial sclerenchyma cell length in the lamina joint of RLI1 overexpressors was significantly reduced with the introduction of a bu1 or a bc1 mutation (Figures 7D and 7E).

Figure 7.

Knockout of BU1 or BC1 Suppresses the Leaf Inclination Phenotype of RLI1 Overexpressors.

(A) and (B) Phenotypic performance of the 60-d-old plants of the wild type, RLI1-OE-3, bu1-1, bc1-2, RLI1-OE-3/bu1-1, and RLI1-OE-3/bc1-2 under pot experiment conditions.

(C) Leaf inclination angle analysis of wild-type, RLI1-OE-3, bu1-1, bc1-2, RLI1-OE-3/bu1-1, and RLI1-OE-3/bc1-2 plants. The third leaf (3rd) from the top of main stem was used for leaf inclination angle measurement and values for leaf angle are given as means ± sd (n = 10 to 15; **P < 0.01 and ***P < 0.001; Student’s t test).

(D) Longitudinal section of the lamia joint from wild-type, RLI1-OE-3, bu1-1, bc1-2, RLI1-OE-3/bu1-1, and RLI1-OE-3/bc1-2 plants. Lamina joint of the third leaf from the top of main stem was used for longitudinal section assay. Bar = 50μm.

(E) Lamina joint adaxial and abaxial cell lengths. Values represent means ± sd (n = 30 to 50). Data significantly different from the corresponding controls are indicated (RLI1-OE-3 versus RLI1-OE-3/bu1-1 and RLI1-OE-3/bc1-2; **P < 0.01; Student’s t test).

Taken together, these results suggested that RLI1 directly binds to the R1BS on the BU1 and BC1 promoters to activate their transcription for lamina joint cell elongation and therefore modulates leaf inclination in rice.

SPX1 Functions as a Negative Regulator of RLI1

Given that SPX1 interacts with RLI1 and negatively regulates leaf inclination, we hypothesized that SPX1 modulates leaf inclination by inhibiting RLI1 function in vivo. To test this, we first generated RLI1 and SPX1 double overexpressors and the spx1 spx2 rli1-1 triple mutant by crossing (Supplemental Figures 9 and 10). Our results showed that the enlarged leaf inclination phenotype of RLI1 overexpressors could be significantly suppressed (Figures 8A and 8B). Conversely, the enlarged leaf inclination phenotype in the spx1 spx2 double mutants was also suppressed by the introduction of an rli1-1 mutation (Figures 8C and 8D). We further analyzed the transcript levels of BU1 and BC1, the target genes of RLI1, in the wild type, rli1-1, spx1, spx2, spx1 spx2 rli1-1, RLI1-OE-3, SPX1-OE-1, and SPX1-OE-1/RLI1-OE-3 by RT-qPCR. We found that the induction of BU1 and BC1 by RLI1 overexpression was significantly suppressed by the addition of SPX1 overexpression, while the elevated expression of BU1 and BC1 in the spx1 spx2 double mutant was restored by the introduction of an rli1-1 mutation (Figure 8E). This finding indicated that SPX1 could inhibit the RLI1 from activating the expression of the downstream target genes in vivo.

Figure 8.

SPX1 and SPX2 Repress the Transcription Activity of RLI1 to Modulate the Leaf Inclination.

(A) and (C) Phenotypic performance of leaf inclination of wild-type, rli1-1, spx1 spx2, spx1 spx2 rli1-1, RLI1-OE-3, SPX1-OE-1, and RLI1-OE-3/SPX1-OE-1 plants. All plants were cultured under hydroponic conditions. Bar = 5 cm.

(B) and (D) Quantification of the leaf inclination in the plants of the wild type, rli1-1, spx1 spx2, spx1 spx2 rli1-1, RLI1-OE-3, SPX1-OE-1, and RLI1-OE-3/SPX1-OE-1. The third leaf (3th) from the top of the main stem was used for leaf inclination measurement, and values for leaf inclination are given as means ± sd (n = 10).

(E) Relative expression levels of BU1 and BC1 in wild-type, rli1-1, spx1 spx2, spx1 spx2 rli1-1, RLI1-OE-3, SPX1-OE-1, and RLI1-OE-3/SPX1-OE-1 plants. Lamina joint tissues of 21-d-old plants were harvested for RNA extraction and RT-qPCR analyses. Values represent means ± sd of three replicates.

(F) ChIP-quantitative PCR assays showed that SPX1 inhibits RLI1 from binding to the promoters of BU1 and BC1. Lamina joint tissues of 4-week-old wild type, 35S:RLI1-FLAG, and 35S:RLI1-FLAG×SPX1-OE-1 were harvested for ChIP experiments. Enriched DNA fragments in the BU1 and BC1 promoters were quantified by RT-qPCR. Data significantly different from the corresponding controls are indicated (*P < 0.05, **P < 0.01, and ***P < 0.001; Student’s t test).

To further determine the possible mechanism of SPX1 in the inhibition of RLI1 transcription activity, ChIP-qPCR analysis with the plants of wild type, RLI1-OE-3, and SPX1-OE-1/RLI1-OE-3 was used to test the effect of SPX1 on the DNA binding ability of RLI1 in vivo. The results showed that the enrichment of RLI1-FLAG on the BU1 and BC1 promoters were obviously reduced by SPX1 overexpression (Figure 8F).

Together, the results show that SPX1 can interact with RLI1 to prevent it from binding to the R1BS of the downstream target genes, therefore negatively regulating RLI1 function in modulating leaf inclination in rice.

The SPX-RLI1 Module Regulates Pi-Deficient Stress-Induced Leaf Erectness in Rice

Since SPX1 is a key negative regulator for Pi starvation signaling and Pi homeostasis and inhibits the activity of RLI1 to regulate the leaf inclination, we hypothesized that Pi-deficient stress-induced leaf erectness would also be regulated by the SPX-RLI1 module. To test this hypothesis, we first examined the expression patterns of RLI1, SPX1, and SPX2. Consistent with their role in the regulation of leaf inclination, RLI1, SPX1, and SPX2 are coexpressed in lamina joint as demonstrated by RT-qPCR (Supplemental Figure 11). Given that the expression of BU1 is induced by BR treatment, we also tested whether the expression of RLI1, SPX1, and SPX2 is affected by BR treatment. Similar to the previous report, 24-h 24-epibrassinolide treatment could induce BU1 expression (Tanaka et al., 2009). However, the expression of RLI1, SPX1, and SPX2 is not affected by the treatment (Supplemental Figure 12).

We further tested RLI1 expression under Pi-deficient conditions. The transcript levels of RLI1 were reduced by Pi deficiency stress. After growth for 20 d under Pi starvation conditions, the transcript level decreased to around half of that under Pi-sufficient conditions (Figure 9A). To verify the expression pattern of RLI1, especially in the lamina joint, we generated RLI1-pro:GUS transgenic lines. In agreement with the RT-qPCR results, GUS histochemical analysis showed that RLI1 was highly expressed in the adaxial and abaxial cells of lamina joint under Pi-sufficient conditions. The expression of RLI1 in these sclerenchyma cells was significantly repressed by Pi deficiency (Figure 9B). In contrast to RLI1, the expression of SPX1 and SPX2 in lamina joint was highly induced by the Pi-deficient stress (Supplemental Figure 13). Given that SPXs can interact with RLI1 to modulate its downstream gene expression, we also analyzed the expression of BU1 and BC1 under Pi-deficient stress. It showed that Pi-deficient stress could repress BU1 and BC1 expression in the wild-type lamina joint, while the BU1 and BC1 expression levels in rli1-1 lamina joint under both the +P and –P conditions were comparable to those of the wild type under –P condition (Supplemental Figure 13). Together, these findings indicated that Pi-deficiency might repress the expression of RLI1 and induce SPX to modulate the downstream genes' expression to restrict lamina joint cell elongation for leaf erectness.

Figure 9.

Pi Deficiency Stress Inhibits RLI1 to Promote the Leaf Erectness.

(A) RLI1 relative expression levels in plants of the wild type and RLI1-OE-3 under –P treatment for 10 d and 20 d. The seeds were sown on Pi-sufficient (200 μM Pi) hydroponic conditions for 10 d before +P (200 μM Pi) and –P (0 μM Pi) treatment. All plant shoots were harvest for RNA extraction and RT-qPCR analyses. Values represent means ± sd of three replicates.

(B) GUS staining of RLI1-pro:GUS transgenic plants under +P (200 µM) and –P (0 µM) conditions. The GUS staining is shown in lamina joint area containing parts of leaf blade and leaf sheath and cross sections of lamina joints (LJ). LJ-1 indicates the adaxial areas of lamina joint; LJ-2 indicates the abaxial areas of lamina joint. Bars = 100 μm.

(C) Phenotypic performance of lamina joints from the wild type, rli1-1, RLI1-OE-3, SPX1-OE-1, spx1 spx2, and spx1 spx2 rli1-1 grown under +P and –P conditions. The seeds were sown in +P (200 μM Pi) and –P (0 μM Pi) hydroponic solution for 15 d before phenotype performance. Bar = 5 cm.

(D) Longitudinal section of the lamia joints from wild-type, rli1-1, RLI1-OE-3, SPX1-OE-1, spx1 spx2, and spx1 spx2 rli1-1 plants grown under +P and –P conditions.

(E) Quantification of leaf inclination in the plants of wild-type, rli1-1, RLI1-OE-3, SPX1-OE-1, spx1 spx2, and spx1 spx2 rli1-1 plants grown under +P and –P conditions. Solid and hatched bars indicate +P and –P conditions, respectively. The first leaf was used for leaf inclination measurement, and values for leaf inclination are given as means ± sd (n = 10).

(F) and (G) Lamina joint adaxial and abaxial cell lengths. Values represent means ± sd (n = 30 to 50). Hatched bars indicate –P condition. Data significantly different from the corresponding controls are indicated (*P < 0.05 and **P < 0.01; Student’s t test).

To test this hypothesis, we first analyzed the lamina joint cell length and leaf inclination phenotype of RLI1 overexpressors under +P and –P conditions. Given that Pi-starvation-induced SPX proteins can regulate leaf inclination by inhibiting the function of RLI1, we hypothesized that both the enlarged lamina joint cell and leaf inclination phenotype of RLI1 overexpressors might be partially suppressed due to the induction of SPXs. Consistent with this, both the lamina joint cell length and leaf inclination of the RLI1 overexpressors were partially reduced by the Pi-deficient stress, despite the fact that the lamina joint cell length and leaf inclination of the RLI1 overexpressors are comparable to those of the wild type under +P conditions (Figures 9C to 9G). Conversely, the reduced leaf inclination and lamina joint cell length phenotypes of the SPX1 overexpressors were only slightly repressed by Pi deficiency (Figures 9C to 9G). Furthermore, due to the repression of RLI1 by Pi deficiency, both the enlarged leaf inclination and lamina joint cell length of the spx1 spx2 mutant were reduced by the –P treatment. However, because of the absence of SPX negative regulators, the leaf inclination angle and lamina joint cell length of the spx1 spx2 mutant under –P condition was similar to those of the wild type under +P conditions (Figures 9C to 9G).

Given the central role of RLI1 in the regulation of lamina joint cell elongation and leaf inclination in rice and that Pi-deficient stress can repress both the expression and function of RLI1, we hypothesized that both the lamina joint cell elongation and leaf inclination of the RLI1 loss-of-function mutant should be less sensitive to the Pi-deficient stress. To test this hypothesis, we analyzed the lamina joint cell length and leaf inclination phenotype of rli1-1 and spx1 spx2 rli1-1 under both the +P and –P conditions. We found that the lamina joint cell length of rli1-1 under both the conditions is similar to those of the wild type under –P conditions. Furthermore, in contrast to the wild type, Pi-deficient stress only slightly induced an increase in leaf erectness in the mutant (Figures 9C to 9G). Similar to rli1-1, the lamina joint cell length of the spx1 spx2 rli1-1 triple mutant was not responsible to the Pi-deficient stress and the magnitude of the leaf erectness response to the stress was decreased compared with that in the wild type (Figures 9C to 9G).

Taken together, our results showed that Pi-deficient stress could regulate leaf erectness by the cooperation of the SPX-RLI1 module.

DISCUSSION

Leaf inclination is an important agronomic trait in rice, providing the foundation of shoot architecture and affecting grain yields in the field. The erect-leaves phenotype allows greater light penetration to lower leaves, therefore optimizing canopy photosynthesis for dense planting in rice fields (Sinclair and Sheehy, 1999), while for the individual plant growing with sufficient space, the enlarged leaf inclination can benefit the plant with enhanced light capture. Despite the accumulation of knowledge in the regulation of rice leaf inclination, how the plant responds to the external environmental cues to modulate leaf inclination, therefore affecting photosynthesis and growth, is still obscure.

Due to low bioavailability, Pi deficiency limits crop production on more than 70% of cultivated land globally (López-Arredondo et al., 2014). To cope with Pi deficiency, plants activate a set of coordinated biochemical and developmental responses to increase Pi uptake and use and recycle P more efficiently in the plant (Nilsson et al., 2007). One of the typical developmental responses to Pi deficiency in rice is the induction of leaf erectness. This response might be a strategy to withstand Pi deficiency stress in rice. It is well known that Pi deficiency inhibits photosynthesis in plants (Dietz and Foyer, 1986). The stress can inhibit the export of triose-P from chloroplast stroma to cytosol by the Pi translocators (Natr, 1992), therefore leading to the conversion of photosynthates into starch in chloroplasts. This is because photosynthesis consumes a large amount of Pi for synthesis of ATP and triose-P and the conversion of triose-P to starch liberates Pi. Therefore, the induction of leaf erectness to reduce the light capture ability together with biochemical adaptations can inhibit photosynthesis to conserve Pi in rice. Further understanding the molecular basis of this developmental adaptation in rice may help us to develop crop varieties that use Pi more efficiently.

Here, we uncovered the cellular basis of the Pi deficiency stress-induced leaf erectness in rice. We noticed that the cellular basis of Pi-deficient stress-induced leaf erectness was slightly different from that induced by BR deficiency. It has been demonstrated that the BR deficiency can induce abaxial cell division and inhibit adaxial cell elongation for leaf erectness in rice (Sun et al., 2015). Since cell proliferation generates a demand for Pi that cannot be satisfied in Pi-deficient conditions, it is probable that a different strategy might be employed for the induction of leaf erectness under the stress. Consistent with this theory, we found that Pi-deficient stress inhibits both abaxial and adaxial cell divisions (Figure 1), which is different from the effect of inducing abaxial cell division by BR deficiency. Furthermore, Pi-deficiency-induced leaf erectness was mainly due to the inhibition of lamina joint cell elongation resulting in restricted lamina joint size (Figure 1). We also noticed that autofluorescence of the cell walls was enhanced in the lamina joints of plants grown under Pi-deficient conditions (Figure 1). It indicated that Pi deficiency not only inhibited lamina joint cell elongation but also induced cellular processes like cell wall lignification. However, exogenous BR treatments could still induce leaf inclination in the plants grown under –P condition (Supplemental Figure 14), which indicated that Pi deficiency stress-induced cell wall lignification might not restrict leaf bending ability in rice. Therefore, the similarity and diversity of the cellular basis for the regulation of leaf inclination by Pi-deficient stress and BR deficiency further emphasized the complexity of regulation of leaf erectness in rice.

We further found that RLI1 is a regulator of leaf inclination by modulating lamina joint cell elongation. RLI1 is an HTH_MYB-like domain contained protein, which shows high similarity with the MYB-CC (coiled coil) domain contained PHR1 subfamily members in rice (Supplemental Figure 3). However, there is no CC domain in the RLI1 protein. It has been reported that the CC domain of the PHR1 may be the dimerization domain and a deletion derivative of PHR1 lacking part of the CC domain showed impaired high-affinity sequence-specific DNA binding (Rubio et al., 2001). Consistent with the possible function of the CC domain in protein dimerization, we found that the RLI1 could not form a dimer in vivo (Figures 3A and 3B). However, we found that the RLI1 can also bind to the P1BS (GNATATNC) cis-element in the EMSA analysis (Figure 6A). Further analysis indicates that RLI1 binds to an even broader cis-element (R1BS; NNAKATNC). This discrepancy with the binding ability of the PHR1 lacking part of the CC domain might be that the residual CC domain in PHR1 interferes the MYB domain to access to DNA, while the complete lack of a CC domain in RLI1 allows it to function as monomer, conferring more flexibility to access DNA for binding. However, it has been shown that the PHR1 subfamily members of OsPHR1, OsPHR2, OsPHR3, and OsPHR4 have different DNA binding affinity for different P1BS sequences, even though they have similar conserved MYB-CC domains (Guo et al., 2015; Ruan et al., 2015, 2017). Therefore, we could not rule out the possibility that the divergence in the MYB domain of RLI1 enables its ability to bind to the R1BS cis-element.

We also demonstrated that RLI1 could bind to the candidate R1BS cis-elements in the BU1 and BC1 promoters in vivo. BU1 and BC1 are BR induced genes that are mainly expressed in the lamina joint. These genes positively regulate cell elongation to enhance leaf inclination in rice (Tanaka et al., 2009; Jang et al., 2017). It has been shown that BU1 is upregulated through both rice BR INSENSITIVE1 and RGA1 (rice G protein α-subunit 1) pathways by BR signaling, while the expression of BU1 is inhibited in the bri1 mutant and rga1 mutant backgrounds (Tanaka et al., 2009). Here, we showed that Pi-deficient stress could repress BU1 and BC1 expression through modulation of RLI1 activity (Figure 9; Supplemental Figure 13). Therefore, Pi-deficient stress and BR deficiency can both modulate the BU1 and BC1 expression to affect cell elongation in regulating rice leaf erectness.

It has been demonstrated that the SPX1/2 could inhibit PHR2 from binding to P1BS only in the presence of Pi in rice (Wang et al., 2014). In Arabidopsis, the competitive inhibition of PHR1 binding to P1BS by SPX1 is also highly dependent on the presence of Pi (Puga et al., 2014). Here, we found that SPX1 and SPX2 could bind to RLI1 and inhibit it from binding to the R1BS of downstream genes. Although we haven’t determined whether this inhibition is also in a Pi-dependent manner, the partial reduction of both the leaf inclination and lamina joint cell length of the RLI1 overexpressor by the Pi-deficient stress (Figures 9C to 9G) indicates that the SPX1 and SPX2 could still inhibit RLI1 activity under our Pi-deficient condition. Consistent with this, the leaf inclination and lamina joint cell length of the spx1 spx2 double mutant is significantly larger than those of the wild type under Pi-deficient stress (Figures 9C to 9G). However, whether the interaction between SPX1/2 and RLI1 might be attenuated under Pi deficiency stress needs to be further determined.

Basing on this acquired knowledge, we propose a working model that SPX and RLI1 form a module in lamina joint to regulate leaf erectness in response to Pi availability in rice (Figure 10). Under Pi-sufficient conditions, RLI1 can bind to the promoters of downstream BU1 and BC1 to activate their expression. The expression of BU1 and BC1 promotes lamina joint cell elongation and therefore enlarges the angle of leaf inclination in rice. Pi deficiency can repress RLI1 and induce SPX1/2 expression, respectively. The elevated SPX1/2 proteins can interact with the reduced RLI1 to inhibit its binding to the downstream BU1 and BC1 promoter, therefore repressing the expression of the downstream genes. The reduction of downstream gene expression restricts lamina joint cell elongation for leaf erectness in rice. Understanding this Pi-deficiency-induced mechanism regulating leaf erectness can enable us to breed new rice cultivars with different canopy forms that are sensitive or insensitive to Pi availability, which might help to increase Pi efficiency in field.

Figure 10.

A Working Model for Regulation of Rice Leaf Inclination Angle by the SPX-RLI1 Module in Response to Pi Availability.

Arrows represent positive effects, and lines ending with a short bar indicate negative effects. Blue arrows and lines represent transcript level regulation, and brown lines represent protein level regulation. Dotted lines indicate that this process does not work or is inhibited under related conditions. The black arrows indicate unknown mechanism.

METHODS

Plant Materials and Growth Conditions

All rice (Oryza sativa) plants used in this study were derived from the japonica variety Nipponbare. The mutants of RLI1 (rli1-1 and rli1-2) were obtained using the TALEN technique (Zhang et al., 2011). The right TALE recognition sequences are TCTGAAGCTGATGAATT, and the left TALE recognition sequences are GGCTCTTGATGTGGTAG. The sequencing identification primers for rli1-1 and rli1-2 are listed in Supplemental Table 1. The RLI1 overexpressors RLI1-OE-3, RLI1-OE-7, and RLI1-OE-10 were obtained by transforming 35S:RLI1-FLAG into wild-type Nipponbare.

The rli1-1 homozygous allele was crossed with the spx1 spx2 double mutant (Wang et al., 2014), and the rli1 spx1 spx2 triple mutant was identified from the F2 population. The RLI1-OE-3 line was crossed with SPX1-OE-1 (Wang et al., 2014) to generate SPX1-OE-1/RLI1-OE-3 double overexpressors.

The mutants of bu1 and bc1 (both in the wild type and RLI1-OE-3 backgrounds) were generated by the CRISPR-Cas9 technique (Mao et al., 2013; Feng et al., 2014). Both BU1 and BC1 were designed two CRISPR-Cas9 target sites (BU1 is GTCGTCGCGCTCCTCCGTGT and GTCGACGACGAAGCTGCTGA; BC1 is AATGGCTTCCTCTTCTGCCA and GGACAAGAAGAGGAAGCCCA). The identification primers for bu1 and bc1 are listed in Supplemental Table 1.

The Pi-sufficient (+P) and Pi-deficient (–P) treatments were performed using normal rice (O. sativa) culture solution (Yoshida et al., 1976) with 200 and 0 μM Pi, respectively. The nutrient solution was adjusted to pH 5.5 using 1 M NaOH and replaced every 2 d during treatment. The experiments were performed in a greenhouse with a 12-h-day (30°C)/12-h-night (22°C) photoperiod, with ∼200 μmol m⁻² s⁻¹ photon densities and ∼60% humidity.

Construction of Vectors and Generation of Transgenic Plant Lines

To construct the vector of 35S-RLI1-FLAG, the open reading frame of RLI1 (Os04g56990) was amplified from Nipponbare cDNA with Primer-Star-GXL (Takara) and cloned into a modified binary expression vector pF3PZPY122 after the cauliflower mosaic virus 35S promoter. To generate the vector of RLI1pro-GUS, a 2.6-kb fragment upstream of the start codon of RLI1 was amplified from genomic DNA of Nipponbare using Primer-Star-GXL (Takara) and then fused in frame to the 5′ terminus of GUS in the modified pBI101.3-GUS plasmid. The vectors of TALLEN-RLI1 for generating RLI1 mutation were constructed were generated as previously described (Zhang et al., 2011). The vectors of CRISPR-BU1 and CRISPR-BC1 were generated according to the instructions of Ma et al. (2015). In brief, two targets sequences of BU1 and BC1 were fused to the promoters of rice genomic RNA of U3 and U6a, respectively, and then constructed into the binary vectors of pYLCRISPR/Cas9-MT (I). All the constructs mentioned above were transformed into mature embryos from the seeds of Nipponbare or RLI1-OE-3 via Agrobacterium tumefaciens (strain EHA105) as described previously (Hiei et al., 1994).

RNA Extraction, RT-PCR, and qPCR Assay

RNA was isolated from roots and shoots using Trizol reagent (Invitrogen) following the manufacturer’s instructions. Reverse transcription was performed using a Moloney Murine Leukemia Virus Reverse Transcriptase cDNA Synthesis Kit (Promega) according to the manufacturer’s instructions. RT-qPCR was performed as previously described (Zhou et al., 2008). The rice Actin gene was used as an internal control. Three biological replicates were performed per gene. Three technical replicates were performed per gene within an experiment.

Y2H Assays

The matchmaker GAL4 two-hybrid system (Clontech) was used for Y2H assays. Full-length SPX1 and RLI1 were cloned into the pGADT7 and pGBKT7vectors, generating AD-SPX1, BD-SPX1, AD-RLI1, and BD-RLI1 vectors. The 1 to 765 bp, 630 to 1053 bp, and 945 to 1053 bp of RLI1 fragments were cloned into the pGADT7 vector to generate AD-RLI1-N, AD-RLI1-C, and AD-RLI1-MYB vectors, respectively. Primers used are listed in Supplemental Table 1. Constructs were cotransformed into the yeast strain AH109. Medium without tryptophan (–Trp) and leucine (–Leu) was used for selecting transformed positive clones. Medium without tryptophan (–Trp), leucine (–Leu), histidine (–His), and adenine (–Ade) was used for selecting positive interact clones.

Bimolecular Fluorescence Complementation Assays

The coding sequences (CDSs) of RLI1 and SPX1 were cloned into either C-terminal or N-terminal fragments of YFP vectors (Shen et al., 2011). Primers used are listed in Supplemental Table 1. The constructs were transiently expressed in Nicotiana benthamiana leaves by Agrobacterium-mediated infiltration (stain EHA105) as described previously (Walter et al., 2004). The YFP fluorescence in N. benthamiana leaves was imaged at 3 d after infiltration using a Zeiss LSM710NLO confocal laser scanning microscope. Prior to fluorescence image observation, the infiltrated samples were treated for 5 min with 10 μM 4′,6-diamidino-2-phenylindole for observation of the nucleus.

Coimmunoprecipitation Assays

The CDS of SPX1 was constructed into the modified vector of pCABIA1300-GFP to generate 35-SPX1-GFP. The CDS of RLI1 was constructed into the modified vector of pF3PZPY122 to generate 35-RLI1-FLAG. The constructs were transiently expressed in N. benthamiana leaves by Agrobacterium. The coimmunoprecipitation assays were performed as described (Feng et al., 2008). Anti-FLAG magnetic beads (Sigma-Aldrich) were used to immunoprecipitate protein complexes and proteins were detected using an ECL reagent (Millipore) and the ChemDoc XRS system (Bio-Rad).

EMSA

To express the recombinant RLI1-HIS protein in Escherichia coli [strain BL21 (DE3); Novagen] the full-length CDS of RLI1 was amplified and cloned into the pET29b vector (Promega), resulting in the RLI1-HIS vectors. The recombinant proteins were extracted and purified according to the manufacturer’s instructions (Qiagen). The biotin-labeled P1BS probes were synthesized by HuaDa. The probe sequences and related mutation versions are listed in Supplemental Table 1. The EMSA was performed using a Light Shift Chemiluminescent EMSA Kit (Thermo Scientific) according to the manufacturer’s instructions.

ChIP-qPCR

One gram per plant of lamina joint tissues from 3-week-old wild type, 35S:RLI1-FLAG, and 35S:RLI1-FLAG×SPX1-OE-1 seedling were harvested for ChIP experiments. Nuclei were isolated using a Plant Nuclei Isolation/Extraction Kit (Sigma-Aldrich) following the procedure of Semi-Pure Preparation of Nuclei. Chromatin was isolated and sonicated to generate DNA fragments with an average size of ∼500 bp. The solubilized chromatin was immunoprecipitated by the anti-FLAG magnetic beads (Sigma-Aldrich). The coimmunoprecipitated DNA was recovered and analyzed by quantitative PCR with SYBGreen mix (Roche). Relative fold enrichment was calculated by normalizing the amount of a target DNA fragment against the respective input DNA samples and then against the wild-type promoter fragment. Three independent biological repeats were performed, yielding similar results. Representative data from one biological replicate is shown in the article. The primers used are listed in Supplemental Table 1.

Section and Cell Length Analysis

To produce sections of the lamina joint, lamina joint segments were embedded in 3% agar. Longitudinal and transverse sections (30 μm) of lamina joint were produced using a vibratome (Leica VT 1000 S). The images of lamina joint autofluorescence were taken under a microscope (Leica DM6000M). The lengths of adaxial and abaxial sclerenchyma cell in the lamina joint were measured using the microscope analysis software (Leica DM6000M).

GUS Histochemical Analysis

The lamina joints of 10-d-old seedlings of RLI1pro-GUS transgenic plants grown under +P and –P conditions for a further 20 d were used for histochemical GUS analysis. Histochemical GUS analysis was performed as previously described (Jefferson et al., 1987).

Measurement of Leaf Inclination

The measurement of leaf inclination was conducted as described previously (Yoshida et al., 1969). Briefly, seedlings were placed on a vertical board covered with paper. The culm is the vertical axis. With the leaves drooping normally from the axis, the positions of the leaf blade and lamina joint were marked on the paper. A line was drawn between the two points and the angle between the line and the vertical axis was measured with a protractor.

Statistics

To determine significant differences between two groups, a Student’s two-tailed t test was used for all experiments.

Accession Numbers

Sequence data from this article can be found in the database of the Rice Genome Annotation Project under the following accession numbers: RLI1 (Os04g56990), SPX1 (Os06g40120), SPX2 (Os02g10780), BU1 (Os06g12210), BC1 (Os09g33580), and ACTIN (Os03g50890).

Supplemental Data

• Supplemental Figure 1. The phr2-T mutant has no obvious defect in leaf inclination.

• Supplemental Figure 2. SPX1 and SPX2 negatively regulate lamina joint cell length.

• Supplemental Figure 3. RLI1 is an HTH_MYB-like transcription family protein.

• Supplemental Figure 4. SPX1 interacts with the MYB domain of RLI1 in yeast.

• Supplemental Figure 5. Identification of RLI1 mutants and overexpressors.

• Supplemental Figure 6. Malfunction of RLI1 has no obvious effect on the number of cell layers in the adaxial and abaxial sides of lamina joint.

• Supplemental Figure 7. The rli1-1 mutant has no obvious growth defects at seedling stage.

• Supplemental Figure 8. Identification of bu1, bc1, RLI1-OE-3/bu1, and RLI1-OE-3/bc1.

• Supplemental Figure 9. Identification of spx1, spx2, spx1 spx2, and spx1 spx2 rli1-1.

• Supplemental Figure 10. Characterization of the overexpressors SPX1-OEs, SPX2-OEs, RLI1-OE-3, and SPX1-OE-1/RLI1-OE-3.

• Supplemental Figure 11. Relative expression levels of RLI1, BU1, BC1, SPX1, and SPX2 in leaf sheath, lamina joint, leaf blade, and root

• Supplemental Figure 12. Effect of exogenous eBL-24 treatment on RLI1, SPX1, and SPX2 expression.

• Supplemental Figure 13. Relative expression levels of BU1 and BC1 in lamina joint tissues of the wild type and rli1-1 under +P and –P conditions

• Supplemental Figure 14. Exogenous eBL-24 treatments induce leaf bending in plants grown under both the Pi-sufficient and Pi-deficient conditions.

• Supplemental Table 1. Primers used in this study.

• Supplemental File 1. Text file of alignment used for phylogenetic analysis in Supplemental Figure 3.