Introduction

Plant architecture, including plant height, tiller number, tiller angle and leaf angle, is an important agronomic trait, which determines grain yield in crops (Wang et al. 2018). The plant architecture is regulated by internal developmental programs and external environmental factors (Wang et al. 2018). Phosphorus is one of the essential macronutrients in plant growth and development. Although the total abundance of phosphorus in soil is not deficient, the phosphorus that can be absorbed and utilized by plants is usually low, (Vance 2010). On the one hand, the phosphorus deficiency has become one of the important factors limiting crop yields and require the input of phosphorus fertilizer to sustain the crop yields. On the other hand, the overload phosphorus fertilizer has caused environmental concerns due to the low solubility and high chemical reactivity of phosphorus in soil. To improve the phosphate (Pi) using efficiency of crops is becoming more and more important for the sustainable agriculture development (Lopez-Arredondo et al. 2014).

To adapt to the Pi limited environment, plants have developed various mechanisms to increase effective absorption of phosphorus by roots from soil and to improve the using efficiency of phosphorus in plant during evolution (Lopez-Arredondo et al. 2014). The adaptation to the phosphorus deficiency mainly depends on the phosphorus starvation response (PSR) (Heuer et al. 2017), which can integrate the phosphorus homeostasis regulated by phosphorus level information of different parts of the plant and minimize the stress from the deficient phosphorus conditions (Bustos et al. 2010). MYB transcription factor family genes OsPHR1, OsPHR2 and OsPHR3 are core regulatory components of PSR regulation system (Guo et al. 2015). PHRs can bind to the P1BS (PHR1 binding site, GNATATNC) cis-elements of PSR genes to activate or repress their expression and regulate Pi homeostasis in plant (Wu and Wang 2008; Guo et al. 2015). The loss function of PHRs repressed the Pi starvation response and disrupted Pi homeostasis in rice (Wu and Wang 2008; Guo et al. 2015). On the contrary, when these transcription factors overexpressed in rice, for example OsPHR2, it would lead to Pi over-accumulation in the shoots under Pi-sufficient conditions (Wu and Wang 2008; Bustos et al. 2010). Recently, the SPX domain (SYG1/Pho81/XPR1) proteins were identified as negative regulators for PHRs and play critical roles in maintaining Pi homeostasis of plant (Lv et al. 2014; Shi et al. 2014; Wang et al. 2014). It has been found that the PSR is regulated by the delicate system of SPXs-PHRs (Wild et al. 2016). The SPXs interact with PHRs under Pi sufficient conditions, which prevents the nuclear localization of PHRs and their binding activity to P1BS element of the promoters of PSR genes under Pi deficiency conditions (Lv et al. 2014; Shi et al. 2014; Wang et al. 2014). When Pi is deficient, the SPXs cannot maintain their interaction with PHRs, which in turn leads to the nucleus enrichment of PHRs and the binding of PHRs to the P1BS element of downstream PSR genes (Lv et al. 2014; Shi et al. 2014; Wang et al. 2014). The expression of OsSPXs except OsSPX4 is in response to the change of Pi conditions to maintain Pi homeostasis (Lv et al. 2014; Shi et al. 2014; Wang et al. 2014). However, the stability of SPX4 is dependent on the Pi availability. Furthermore, promoters of most SPXs genes contain P1BS element. The expression levels of these genes are unregulated in Pi deficiency conditions, which in turn inhibits the function of PHRs. Recently, the SPX protein has been indicated to be a receptor of the endogenous Pi signaling molecule InsP8 (Wild et al. 2016; Dong et al. 2019). The endogenous level of InsPs responses to the change of Pi availability in environment and modulates the interaction between SPXs and PHRs (Wild et al. 2016), which mediates the association and disassociation between SPXs and PHRs.

Phytohormones were shown to be associated with the absorption and use of phosphorus (Lin et al. 2014). Phytohormones determine the use or utilization of phosphorus in plants by regulating the development of plant roots, modulating the symbiotic relationship between plants and microbiome in soil, and coordinating the above-ground part to under-ground parts of plant (Lopez-Arredondo et al. 2014). The developmental and biochemical mechanisms mediated by phytohormones action developed during evolution are able either to inhibit the unnecessary utilization and remobilize internal phosphorus, or promote absorption of external phosphorus, which enables plants to adapt to phosphorus limited environment (Chiou and Lin 2011). The changes in phosphorus availability can alter the synthesis, perception and transport of plant hormones while the phosphorus effectiveness can directly affect the metabolism and signal conduction of plant hormones in plants (Chiou and Lin 2011) Brassinosteroid (BR) plays an important role in plant development. The working model of BR biosynthesis and signaling has been well established in the past two decades (Nolan et al. 2019). BES1/BZR1 are key transcription factors in the BR signaling pathway in plants. When BR binds to its receptor BRI1, the glycogen synthesis kinase BIN2 would be dephosphorylated and inactivated, thereby inhibiting the phosphorylation of BZR1/2. The dephosphorylated BZR1/2 would accumulate in nucleus and regulate the downstream genes transcription (Nolan et al. 2019). The plant mainly improves the absorption and utilization of available phosphorus in the surface soil by inhibiting the growth of the main roots and promoting the formation of lateral roots and root hairs. The observations of phosphorus limited induced root system defects were attenuated in bes1/bzr2-D or bzr1-D indicating that BR is involved in the regulation of the plant adaptation to the phosphorus-limited environment (Singh et al. 2014).

Leaf inclination is an important agronomic trait in rice, providing the foundation of shoot architecture and affecting grain yields in the field. BR positively regulates rice leaf inclination (Sun et al. 2015). Both the BR biosynthesis mutant d2 and the signaling mutant d61 showed erect leaves (Yamamuro et al. 2000; Hong et al. 2003). Other phytohormones (such as ABA) can affect leaf inclination by modulating the BR biosynthesis or signaling in rice (Li et al. 2019). BR regulates leaf erectness by affecting the abaxial sclerenchyma cell number of rice lamina joints. The CYC U4;1 encoding U-type cyclin abundantly expresses in lamina joints. The BR regulated expression of CYC U4;1 has been proposed to be essential step in regulating lamina joints development through controlling the abaxial sclerenchyma cell proliferation in rice (Sun et al. 2015). BR can regulate the transcriptional expression of CYC U4;1 through BES1 to control the leaf erectness (Sun et al. 2015). Downstream of BR signaling, several transcription factors other than BZR1 has been found to be involved in the regulation cell division and elongation of lamina joint and affect the leaf inclination in rice. The HLH transcription factors BRASSINISTEROID UPREGULATED1 (BU1) (Tanaka et al. 2009), BU1-LIKE1 (BUL1) (Jang and Li 2017), and BUL1 COMPLEX1 (BC1) are able to enhance leaf inclination mainly by promoting cell elongation of lamina joint (Tanaka et al. 2009; Zhang et al. 2009a, b; Jang and Li 2017). Overexpressing INCREASED LAMININAR INCLINATION1 (ILI1) and ILI1 BINDING HLH1 (IBH1) in rice leads to leaf inclination (Zhang et al. 2009b). IBH1 can interact with BC1 and BUL1 and repress their function in regulating cell elongation (Jang and Li 2017). These regulatory networks of the transcription factors modulate the leaf inclination in rice. Under low-phosphorus conditions, the shoots of plants also reduce leaf size, tiller number and leaf inclination to reduce the amount of phosphorus requirement. Recently, it has been reported that SPX1, a central regulatory element in the phosphorus signaling system, is involved in the regulation of leaf inclination in low-phosphorus stress. Under low-phosphorus conditions, SPX1 can interact with transcription factor RLI1 (Regulator of Leaf Inclination 1), thus repressing the regulation of RLI1 on the transcriptional expression of BU1 and BC1 in lamina joint cells (Ruan et al. 2018). Increasing expressions of BU1 and BC1 promote cell elongation of lamina joint and increase the leaf angle in rice. RLI1 positively regulates leaf angle by binding to the promoters of BU1 and BC1, and activating their expressions. Pi deficiency can induce SPX1/2 expression resulting in the elevation of SPX1/2 proteins, which can interact with RLI1 and inhibit the expressions of the downstream BU1 and BC1 genes. The reduction of downstream genes, such as BU1 and BC1 can lead to the restriction of cell elongation of lamina joint and induced leaf inclination in rice (Ruan et al. 2018).

It has been showed that the cellular basis of leaf inclination induced by BR deficiency is different from the cellular basis induced by Pi-deficiency stress. Pi-deficiency stress can inhibit both abaxial and adaxial cell divisions (Ruan et al. 2018). BR deficiency can induce abaxial cell division. Furthermore, leaf erectness induced by Pi-deficiency was mainly caused by the inhibition of lamina joint cell elongation in which leads to restricted lamina joint size. Meanwhile, exogenous BR treatments can induce leaf inclination in the plants grown under Pi-deficient condition. Therefore, the similarity and diversity of the cellular basis for leaf inclination regulated by Pi-deficiency stress and BR deficiency indicated the regulation complexity of leaf erectness in rice (Ruan et al. 2018). Here, we investigated the possible molecular basis for the difference of Pi deficiency stress-induced and BR induced leaf inclinations in rice. Better understanding of the molecular mechanism can help us develop low Pi tolerance rice varieties.

Results

BR deficiency attenuate the leaf inclination by comprimising Pi deficiency-induced BU1 expression

Given the role of BR in the regulation of leaf inclination in rice, we tested whether Pi-deficiency can repress leaf inclination of the loss-of-function mutant d2 and d61 (Supplementary Fig. 1A). The overall plant morphalogy did not obviously change between two tested conditions (Fig. 1a). But plants grow in Pi-defecient condition showed significantly increase of leaf inclination (Fig. 1a). The PSR genes IPS1 and PT6 were selected as marker genes to verify the effectiveness of treatment (Secco et al. 2013). It showed that the expression levels of IPS1 and PT6 under Pi-deficient condition were about 1000 times higher than those under Pi-sufficient condition in wild type (WT) (Fig. 1b, c), which is consistent with previous work and verified the effectiveness of treatment conditions. The leaf angles of d2 and d61 were less to that of the wild type plants under Pi-sufficient condition (Fig. 1d, e). Under Pi-deficient condition, the leaf angles of d2 and d61 were still less than that of the wild type, similar to those under Pi-sufficient condition (Fig. 1e). However, the leaf inclination in d2 and d61 induced by Pi-deficiency was less than those in WT (Fig. 1e), which suggested that the disruption of BR biosynthesis or signaling can attenuate the Pi-deficiency induced leaf inclination. The Pi-deficiency induced expressions of IPS1 and PT6 were somewhat compromised in d2 but not in d61 (Fig. 1b, c). The expression of OsDWARF4 was negatively regulated by BR pathway (Nakamura et al. 2009), while the expression of OsBZR1 seems not subject to the negative feed-back regulation of BR pathway (Bai et al. 2007). Consistently, the disruption of BR biosynthesis or BR signaling resulted in the upregulation of OsDWARF4 expression (Fig. 1f). The Pi-deficiency stress led to upregulation of OsDWARF4 and OsBZR1 in WT (Fig. 1f, g). On the contrary, the Pi-deficiency stress induced expressions of OsDWARF4 were compromised in d2 and d61 (Fig. 1f). These results indicate that the Pi-deficiency stress induced the expressions of genes OsDWARF4 which are regulated by BR require the intact BR biosynthesis or signaling. The Pi-deficiency stress induced expressions of OsBZR1 in WT whereas Pi-deficiency stress repressed expressions of OsBZR1 in d2 and d61 (Fig. 1g). Pi deficiency induced leaf inclination through repression of RLI1 expression and in turn induced BU1 expression (Ruan et al. 2018). The Pi deficiency repressed RLI1 in d2 and d61, which is similar to those in WT (Fig. 1h). However, the Pi deficiency induced BU1 expression in WT but did not induce BU1 expression in d2 and d61 (Fig. 1i). These results indicated that the less leaf inclination induced by Pi-deficiency in d2 and d61 may be partially through the disruption of the Pi deficiency-induced BU1 expression.

Fig. 1
figure 1

Pi-deficient induced leaf erectness was compromised in BR deficiency mutants. a The phenotype of WT, d2 and d61 plants after 7 days treatment in deficient (0 μM) or sufficient (300 μM) Pi hydroponic culture solutions, respectively. Bar = 2 cm. b, c The expression of PSR marker genes in d2 and d61 to verify the effectiveness of Pi deficiency treatment. d The leaf angles of mutants after 7 days Pi-deficiency stress treatment. e The statistical analysis of leaf inclination for mutants after 7 days Pi-deficiency stress treatment, n = 10. f–i The expression of genes related to leaf inclination which were regulated by BR or Pi deficiency. Error bars are standard error. Different letters indicate significant differences, statistic analysis by one-way analysis of variance (ANOVA) using Duncan’s multiple range test (P < 0.05)

Full size image

Constitutively activated or repressed BR signaling resulted in the insensitivity of Pi induced leaf inclination

It has been showed that leaf inclination of rice grown under Pi-defecient condition still can be responsible to the exogenous BR application (Ruan et al. 2018). However, it is not sure whether leaf inclination in the plants with constitutively activated or repressed BR signaling be able to response to Pi-deficiency. The BR signaling specific activation of CYC U4 in lamina joint is essential for BR induced leaf erectness in rice (Sun et al. 2015). We tested whether Pi-deficiency can repress leaf inclination of the transgenic plants of gain-of-function pDWF4::bes1-D and pDWF4::bin2-1, which showed widder and narrower leaf angel than that in WT plants repectively (Fig. 2a). The Pi-defeciency can significantly induce leaf inclinationin in WT but neither in pDWF4::bes1-D nor in pDWF4::bin2-1 plants (Fig. 2b, c and Supplementary Fig. 1B). The Pi deficiency-induced expressions of IPS1 and PT6 in WT confirmed the effectiveness of low phosphorus treatment (Fig. 2d, e). The expression of CYC U4 in WT under Pi-suffecienct condition confirmed that the BR signaling is constitutively activated in pDWF4::bes1-D and constitutively repressed in pDWF4::bin2-1 respectively (Fig. 2f). Taken together, these results suggested that either constitutive activation or repression of BR signaling can disrupt the Pi deficiency-induced leaf inclination. The expression of CYC U4 in WT and in pDWF4::bin2-1 is not response to Pi-deficiency stress, however, Pi-deficiency stress enhanced the expression of CYC U4 in pDWF4::bes1-D, which is higher than that in WT under Pi-sufficient conditions (Fig. 2f). Moreover, the Pi-deficiency stress induced expression of OsDWARF4 was compromised in pDWF4::bin2-1 but not in pDWF4::bes1-D (Fig. 2g). Furthermore, we found that the Pi deficiency-induced upregualtion of IPS1 was hyposensitive in pDWF4::bes1-D and hypersensitive in pDWF4::bin2-1 compared to those in WT. However, the Pi deficiency-induced expression of PT6 showed little different responsiveness in pDWF4::bes1-D or in pDWF4::bin2-1 compared to those in WT (Fig. 2d, e). These results indicate that either constitutive activation or repression of BR signaling could affect part of the Pi deficiency-induced downstream genes.

Fig. 2
figure 2

Constitutive activate or repress BR signaling suppress the Pi-deficiency induced leaf erectness. a The phenotype of WT (Nipponbare), pDWF4::bes1-D and pDWF4::bin2-1 plants after 7 days treatment in deficient (0 μM)or sufficient (300 μM) Pi hydroponic culture solutions, respectively. Bar = 2 cm. b The leaf angles of mutants after 7 days Pi-deficiency stress treatment. c The statistical analysis of leaf inclination for mutants after 7 days Pi-deficiency stress treatment, n = 10. d, e The expression of PSR marker genes in pDWF4::bes1-D and pDWF4::bin2-1 to verify the effectiveness of Pi deficiency treatment. f–i The expression of genes related to leaf inclination which were regulated by BR or Pi deficiency. Error bars are standard error. Different letters indicate significant differences, statistic analysis by one-way analysis of variance (ANOVA) using Duncan’s multiple range test (P < 0.05)

Full size image

Under Pi-sufficient conditions, the expression of RLI1 significantly increased in pDWF4::bes1-D and dramatically increased in, pDWF4::bin2-1(Fig. 2h). The Pi deficiency was still able to repress RLI1 in pDWF4::bes1-D and in pDWF4::bin2-1, similar to those in WT (Fig. 2h), however, the fold change of Pi-deficiency induced down regulation of RLI1 expression was less in pDWF4::bes1-D than that in WT whereas more in pDWF4::bin2-1 than that in WT(Fig. 2h). In WT, the BU1 expression was downregulated in response to Pi-deficiency stress, which is consistent with the expression of RLI. The expression of BU1 was upregulated in pDWF4::bes1-D and down regulated in pDWF4::bin2-1 under Pi sufficient condition (Fig. 2i). When BR signaling is constitutive activated or repressed, the expression of BU1 is not response to Pi-deficiency stress, which may explain the insensitivity of leaf inclination in response to the Pi-deficiency stress. Although the cellular basis of Pi deficiency-induced leaf inclination is different from BR deficiency-induced leaf inclination, it seems that leaf inclination induced either by Pi-deficiency or by BR deficiency is dependent on their regulatory role on the expression of BU1.

Constitutively activated or repressed BR signaling altered expresion profile in lamina joint

To further analyze the effects of BR signaling on the expression profile of lamina joint, we used lamina joint from pDWF4::bes1-D and pDWF4::bin2-1 for RNA-seq under Pi sufficient condition. By Illumina sequencing of mRNA libraries, we generated about 10G data for each samples FPKM (for fragment per kilobase of exon per million fragments mapped) from RNAseq data which were used to compare transcript abundance among different samples. We used fold change of > 2 or < 0.5 as cutoffs in all the three biological replicates. We identified 576 up-regulated and 167 down-regulated genes in pDWF4::bes1-D compared with those in WT (Supplementary Fig. 2A, B). We also identified 117 up-regulated and 63 down-regulated genes in pDWF4::bin2-1 (Supplementary Fig. 2C-D). The Venn map showed that 3 identified up-regulated genes in pDWF4::bes1-D were down-regulated in pDWF4::bin2-1 compared to those in WT (Fig. 3a) and that 3 identified down-regulated genes in pDWF4::bes1-D were up-regulated in pDWF4::bin2-1 compared to those in WT (Fig. 3b). The 3 genes up-regulated in pDWF4::bes1-D but down-regulated in pDWF4::bin2-1 compared with WT include LOC_Os03g41480, LOC_Os07g01580 and LOC_Os11g24374. The 3 genes down-regulated in pDWF4::bes1-D but up-regulated in pDWF4::bin2-1 compared with WT are OsDWF4 (LOC_Os03g12660), BRD1 (LOC_Os03g40540) and a gene encoding a glycine rich protein (LOC_Os06g21250). We also found 9 genes were up-regulated (Fig. 3c) and 40 genes were down-regulated both in pDWF4::bes1-D and in pDWF4::bin2-1 (Fig. 3d). The Heatmap showed that most of these genes were insensitive to Pi-deficiency stress (Fig. 3e).

Fig. 3
figure 3

BR signaling regulated genes is insensitive to the Pi-deficiency in lamina joint. a Up-regulated genes in bes1-D and Down-regulated genes in bin2-1. b Down-regulated genes in bes1-D and Up-regulated genes in bin2-1. c Up-regulated genes in bes1-D and in bin2-1. d Down-regulated genes in bes1-D and in bin2-1. e Heatmap of genes identified in (a–d)

Full size image

Pi deficiency-induced leaf inclination mainly dependent on the BR mediating regulation expression of downstream genes

To explore the BR signaling-dependent and BR signaling-independent mechanisms that regulate Pi deficiency-induced genes, we further compared the genes expression profile of Pi-surfficienct condition and Pi-deficient condition with three biological replicates. Using fold change of > 2 or < 0.5 as cutoff in all the three biological replicates, we identified 1721 Pi deficiency-induced differentially expressed genes (DEG, 689 up- and 1032 down-regulated) and 14,539 no differentially expressed genes (NDEG) in WT (Supplementary Fig. 3a), 924 Pi deficiency-induced DEG (458 up- and 466 down-regulated) and 16,317 NDEG in pDWF4::bes1-D (Supplementary Fig. 3B), and 1212 Pi deficiency-induced DEG (847 up- and 1365 down-regulated) and 13,560 NDEG in pDWF4::bin2-1 respectively (Supplementary Fig. 3C). Among the up-regulated genes in response to Pi-deficiency stress, 40 and 1 genes were not in response to Pi-deficiency stress in pDWF4::bes1-D and in pDWF4::bin2-1, respectively (Fig. 4a). Among the down-regulated genes in response to Pi-deficiency stress, 46 and 7 genes were not response to Pi-deficiency stress in pDWF4::bes1-D and in pDWF4::bin2-1 respectively (Fig. 4b). Heatmap showed that most of Pi-deficiency up-regulated genes showed insensitive to Pi-deficiency stress in pDWF4::bes1-D (Fig. 4c). It may be due to the less fold change of gene expression in response to Pi-deficiency stress in pDWF4::bes1-D comapred to those in WT. However, most of Pi-deficiency down-regulated genes showed insensitive to Pi-deficiency stress in pDWF4::bes1-D may be caused by the lower expression levels Pi-deficiency stress in Pi-sufficient conditions. Among the Pi-deficiency induced genes, we note that Phosphate-Starvation Induced RING-Type E3 Ligase (OsPIE1, LOC_Os01g72480) was not in response to Pi-deficiency stress in pDWF4::bes1-D (Fig. 4c). OsPIE1 has been reported to regulate Pi homeostasis at least partially by repressing OsSPX2 in rice (Yang et al. 2018). These results suggested that constitutively activated BR signaling inhibited the expression of Pi-deficiency stress induced genes.

Fig. 4
figure 4

Constitutive activate BR attenuated the Pi-deficiency induced gene expression in lamina joint. a Pi-deficiency induced up-regulated genes in WT but their expression are not response to Pi-deficiency in bes1-D and/or in bin2-1. b Pi-deficiency induced down-regulated genes in WT but their expression are not response to Pi-deficiency in bes1-D and/or in bin2-1. c Heatmap of genes identified in (a–b)

Full size image

Pi deficiency-induced expression PSR genes were attenulated by constitutively activated or repressed BR signaling

The observation that OsPIE1 is not in response to Pi-deficiency stress in pDWF4::bes1-D promoted us to investigate that whether constitutively activated or repressed BR signaling alters the sensitivity of PSR genes. We further analyzed the genes in response to Pi-deficiency stress in pDWF4::bes1-D and in pDWF4::bin2-1 as well as in WT. We identified 187 common up-regulated genes (Fig. 5a) and 230 common down-regulated genes in response to Pi-deficiency in WT, pDWF4::bes1-D and in pDWF4::bin2-1 (Fig. 5b). Among the 187 genes, we find that the Pi-deficiency induced expression of several known PSR genes, such as OsSPX2 (LOC_Os02g10780), were significant compromised in pDWF4::bes1-D (Fig. 5a). We further selected 76 conserved PSR genes (Supplemental Table 1) and 27 BR related genes (Supplemental Table 2) to compare their Pi-deficiency induced expression in WT, pDWF4::bes1-D and pDWF4::bin2-1. The Venn map showed that 26 conserved PSR genes and 1 BR related genes overlap with the common Pi-deficiency stress induced genes in WT, pDWF4::bes1-D and in pDWF4::bin2-1 respectively (Fig. 5c). No Pi-deficiency stress down-regulated genes were found to overlap with the selected conserved PSR genes and BR related genes (Fig. 5d). However, the fold change of expression of these genes were significantly reduced in pDWF4::bes1-D but not in pDWF4::bin2-1 compared with those in WT (Fig. 5e). These results suggested that constitutively activated BR signaling attenuated the sensitivity of PSR genes in response to Pi-deficiency stress in lamina joint.

Fig. 5
figure 5

The Pi-deficiency induced expression of PSR genes is compromised when constitutive activate BR signaling. a Up-regulated genes in WT, bes1-D and in bin2-1. b Overlapped up-regulated genes in A overlapped compare to PSR and/or BR related genes. c Down-regulated genes in WT, bes1-D and in bin2-1. d Overlapped down-regulated genes in B overlapped compare to PSR and/or BR related genes. e Heatmap of 26 PSR genes identified in (b)

Full size image

The BR biosynthesis genes D11 (LOC_Os04g39430), BRD1 (LOC_Os03g40540) and OsDWF4 (LOC_Os03g12660) can still response to Pi-deficiency stress in WT and pDWF4::bin2-1 but not in pDWF4::bes1-D (Fig. 4c). We further noted that Pi-deficiency induced fold change of several BR related genes expression was altered when the BR signaling is constitutively activated or repressed. For example, the brassinolide (BL) unregulated gene OsBLE3 was in response to Pi-deficiency stress in WT, pDWF4::bes1-D and pDWF4::bin2-1(Fig. 5a, b). It also showed that higher expression of CYC U2;1 (LOC_Os04g46660) in pDWF4::bin2-1 were downregulated by Pi-deficiency stress (Supplemental Dataset). These results suggested that disruption of BR signaling could modulate the sensitivity of BR related genes response to Pi-deficiency stress in lamina joint and indicated the possibility to attenuate the downside effect of Pi-deficiency stress on rice plant architecture by manipulating the BR regulated development of lamina joint.

Validation of the RNA-Seq by RT-PCR

To validate RNA-seq results, we selected 12 genes to perform RT-qPCR analysis (Fig. 6). It showed that Pi deficiency induced the expression of OsGH3-2 and repressed the expression of OsEXP1, which is a member of α-Expansin (Lee and Kende 2002). The expression of OsGH3-2 and OsEXP1 in response to Pi-deficiency was more sensitive in pDWF4::bin2-1 and less sensitive in pDWF4::bes1-D. OsCesA4/BC7 (Zhang et al. 2009a) were unregulated in pDWF4::bin2-1 under the normal condition compared with those in WT. The expression of OsCesA4 in pDWF4::bin2-1 under Pi-deficiency condition was significantly down-regulated compared with those under normal condition. A putative cellulose synthase-like family F gene, CSLF8 (LOC_Os07g36630) also showed the similar expression pattern as OsCesA4. The expression of OsCesA9/BC6 (Kotake et al. 2011) was unregulated in pDWF4::bes1-D and in pDWF4::bin2-1. The Pi-deficiency stress repressed expression of OsCesA9 in pDWF4::bin2-1. These results indicated that modulation of cellulose synthesis are required for both Pi-deficiency stress induced and BR induced leaf inclination.

Fig. 6
figure 6

Validation expression of genes by RT-PCR. Relatively expression levels of selected genes, which are normalized to the expression of actin. aOsGH3-2 (LOC_Os01g55940), bOsEXP1 (LOC_Os04g15840), cOsCesA4 (LOC_Os01g54620), dCSFL8 (LOC_Os07g36630), eOsCesA9 (L OC_Os09g25490), fOsSPX2 (LOC_Os02g10780), gOsGRF8 (LOC_Os11g35030), hOSH71(LOC_Os05g03884), iOsBZR1 (LOC_Os07g39220), jBC1 (LOC_Os09g33580), kSLG (LOC_Os08g44840), L OsMYB91 (LOC_Os12g38400). Error bars are standard error, n = 3. Different letters indicate significant differences, statistic analysis by one-way analysis of variance (ANOVA) using Duncan’s multiple range test (P < 0.05)

Full size image

The expressions of OsSPX1 and OsSPX2 are essential to maintain Pi-homeostasis and regulate the Pi-deficiency-induced leaf inclination (Lv et al. 2014; Shi et al. 2014; Ruan et al. 2018).The expression of OsSPX2 was in response to Pi-deficiency in pDWF4::bes1-D and in pDWF4::bin2-D as well as in WT. However, the fold change of OsSPX2 in pDWF4::bes1-D was much less than those in pDWF4::bin2-1 and in WT. The GRF can bind to the intron of KNOX families of transcription factors and regulate plant architecture. This modulation was found in both monocot and dicot plants (Kuijt et al. 2014). Similar scenarios as OsSPX2 have been observed for the expressions of OsGRF8 and OSH71/OsKn2. On the contrary, the Pi-deficiency stress slightly induced expression of OsBZR1 in pDWF4::bes1-D but not in pDWF4::bin2-1 as well as in WT. However, the Pi-deficiency stress-induced expression of BC1 was compromised in pDWF4::bin2-1 but not in pDWF4::bes1-D compared to those in WT. It has been reported that SLG can modulate BR homeostasis in rice and that overexpression of SLG lead to the increase of leaf angle (Feng et al. 2016). We found that the Pi-deficiency induced down-regulation of SLG in pDWF4::bes1-D and in pDWF4::bin2-1 as well in WT. Multiple abiotic stress stimuli and exogenous application of phytohormones can induce the expression of OsMYB91(Zhu et al. 2015). The Pi deficiency induced the expression of OsMYB91 in WT and pDWF4::bes1-D but not in pDWF4::bin2-1.

Discussions

BR deficiency induces abaxial cell division and inhibits adaxial cell elongation in rice. Pi-deficiency stress inhibits both abaxial and adaxial cell divisions (Sun et al. 2015). Introducing pDWF4::bes1-D can enlarge leaf angle through regulation of the transcriptional expression of CYC U4;1 which thus controls the leaf erectness. BR positively regulates rice leaf inclination (Sun et al. 2015). It had been showed that Auxin and ABA can affect leaf inclination by modulating the BR biosynthesis or signaling in rice (Li et al. 2019). We find that Pi-deficiency induced the expression of OsGH3-2 and repressed the expression of OsEXP1. OsGH3 can transform IAA to an IAA-amino acid conjugate. It has been found up-regulating OsGH3-1/LC1 and OsGH3.13/TLD1 transcription in the lamina joint leads to increasing of leaf angle. Enhanced expression of OsGH3-2 leads to the suppressed expression of expansin genes (Zhao et al. 2013). The sensetivity of expression of OsGH3-2 responsing to Pi-deficiency can be subject to the modulation of BR signaing. It has been revealed that auxin and BR play various similar roles in plant development by regualting a subset of common downstream genes. Pi-deficiency is able to mediate the regualtion of auxin biosynthesis and transport. How Pi-deficiency regualte leaf inclination by regualtion of auxin biosynthesis and transport in lamina joint needs to be further investigated.

Pi deficiency-induced leaf erectness was mainly caused by the inhibition of lamina joint cell elongation, which is restricted by cell wall. Cell wall thickening is a prominent feature during younger stages of lamina joint development, which declines over time. In response to Pi deficiency, cell wall-related genes were down-regulated which can lead to decrease of cell wall thickening, and thus decrease cell elongation. We find that expressions of OsCesA4 and CSLF8 in pDWF4::bin2-1 under Pi-deficiency condition were significantly down-regulated compared with those under normal condition. It indicated that Pi deficiency not only inhibited lamina joint cell elongation but also induced cellular processes like cell wall lignification. Plants responding to developmental and environmental cues require to reprogram cell wall structure. The Pi-deficiency stress and BR deficiency signals might convergence on modulating the expression of cell wall related genes to control cell elongation in lamina joint. Therefore, the similarity and diversity of the cellular basis for the regulation of leaf inclination by Pi-deficiency stress and BR deficiency further emphasized the complexity of regulation of leaf erectness in rice. Local sensing of Pi deficiency can promote cytoplasmic accumulation of BES1 and BZR1, thus modulate the transcriptional of downstream genes and drive shallower RSA for forage of Pi from soil in Arabidopsis (Singh et al. 2014). When the Pi is adequate, BES1 and BZR1 restore nucleus localized and reprogram the transcription of downstream genes. This mechamism provides an efficient modularity of root development in repsonse to the change of avalibility of Pi. The Pi deficiency stress induced expression of Pi starvation-induced genes in bzr1-D is similar to those acting in wild-type in Arabidopsis. It showed PSR genes expressed higher sensitivity of responsiveness to Pi deficiency stress in bzr1-D than that in WT (Singh et al. 2014). Similar, local sensing of Pi in lamina joint might reprogram the transcriptional of downstream genes and optimize the leaf angle in rice. The deficiency of Pi promotes the leaf inclination and reduces photosynthesis, which decreases the demand of Pi supply. When sensing adequate Pi concentration, the leaf inclination can be restored and promoted. However, we observed that although Pi deficiency stress can still induce the expression of conserved PSR genes in pDWF4::bes1-D and pDWF4::bin2-1. The responsiveness to Pi deficiency stress becomes more insensitive in pDWF4::bes1-D plant compared to those in WT, while the responsiveness to Pi deficiency stress in pDWF4::bin2-1 is similar as that in WT. It seems that BR signaling modulated PSR response in lamina joint of rice is different to those in root of Arabidopsis.

In order to improve grain yield by increasing the plant density of crop, it needs to optimise the coordination of photosynthesis and the absorbtion of nutrient from soil. To dissect the molecualr mechanism underlying the interaction of Pi-deficiency stress and BR deficiency-induced leaf inclination will provide an efficient modularization strategy to increase the plant density without yield penalty. It might due to BR regulated expression of OsPIE1, which acts partially through repression of OsSPX2 and desensitizes Pi deficiency stress induced expression of PSR genes in lamina joint. Furthermore, leaf erectness induced by Pi-deficiency was mainly caused by the inhibition of lamina joint cell elongation resulting in restricted lamina joint size. Meanwhile, exogenous BR treatments can induce leaf inclination in the plants grown under Pi-deficient condition as well. Therefore, the similarity and diversity of the cellular basis for leaf inclination regulated by Pi-deficiency stress and BR deficiency indicated the regulation complexity of leaf erectness in rice.

Methods

Plant materials and growth conditions

The wild type rice cultivar used in this study is TP309 and Nipponbare (Oryza sativa Japonica). The BR biosynthesis mutant d2 and the BR signaling mutant d61 were in the TP309 background. The d2 used in this study is d2-1, which contains a premature stop codon at residue 83 (Hong et al. 2003). The d61 used in this study is d61-1, which contains change of threonine to isoleucine at residue 989 (Yamamuro et al. 2000). The pDWF4::bin2-1 and pDWF4::bes1-D transgenic plants were in Nipponbare background. Surface-sterilized rice seeds were incubated in sterile water at 30 °C for 2 days in the dark. The germinated seeds were transferred to hydroponic culture medium in the 96-well plates and cultured at 30 °C with 16 h light/8 h dark photoperiod for 7 days. The seedlings were transferred to plastic container with 20 L hydroponic culture solution (300 μM Pi/0 μM Pi) and grown at 30 °C for 7 days. Roots and lamina joints of the plant materials were collected respectively on the seventh day. The hydroponic solution was refreshed every 3 days. The hydroponic culture solution contained 1.25 mM NH4NO3, 0.35 mM K2SO4, 1.0 mM CaCl2, 1.0 mM MgSO4, 0.39 µM (NH4)6Mo7O24, 9.0 uM MnCl2 20 µM H3BO3, 0.32 µM CuSO4, 0.77 µM ZnSO4, and 20 uM EDTA-Fe, 0.5 mM NaSiO3, with or without 0.3 mM KH2PO4, resulting in the plus Pi (300 µM) and minus Pi (0 µM) conditions. The pH of the solution was adjusted to 5.5.

The measurement of leaf inclination

According to the described measurement methods in previous references (Sun et al. 2015), the angle between the second leaf lamina and sheath was measured with a protractor on the seventh day of low Pi treatment. The second leaf inclination was 180° minus the measured angle.

RNA extraction and sequencing

There are three biological replicates for every treatment condition. Total RNA was extracted from lamina joints using Trizol reagent (Invitrogen) according to the manufacture’s protocol. The quality of the total RNA was detected using NanoDrop2000 and Agilent 2100 RNA 6000 Nano kit. The X-ten (Illumina) platform was used for RNA-Seq, and cDNA library construction and sequencing were done by Berry Genomics, Beijing, China.

RNAseq data analysis

The raw data size is at least 10 Gb for every sample, and the reads length is 150 bp. The raw RNAseq reads were processed by the fastQC using Trimmomatic tool to remove low quality reads and adapters. All analyzed reads were aligned to the rice reference genome and gene model annotation file (https://rice.plantbiology.msu.edu/pub/data/Eukaryotic_Projects/o_sativa/annotation_dbs/pseudomolecules/version_7.0/all.dir/) using bowtie2 and tophat2 tools. And then the data was processed with cufflinks, cuffmerge and cuffdiff tools respectively, and based the cuffdiff results we collected genes whose |log2(fold_change)| ≥ 1 in different treatments and mutants as differential expression genes (DEGs).

qRT-PCR

cDNA was synthesized with the ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOYOBO). qRT-PCR was performed on a real-time PCR instrument (Bio-Rad CFX Connect Real-Time PCR Detection System) using ChamQ Universal SYBR qPCR Master Mix (Vazyme).

Statistical analysis

The data were analyzed using the online SAS software with one-way analysis of variance (ANOVA) (https://www.sas.com/en_us/software/university-edition/download-software.html#windows-setup). And the significant differences (P < 0.05) were observed using the Duncan’s multiple range test.