ABSTRACT
High temperature usually leads to the failure of grain filling during caryopsis development, resulting in the loss of yield, however the mechanism is not yet well elucidated. Here, we report that two rice caryopsis-specific NAM/ATAF/CUC domain transcription factors, ONAC127 and ONAC129, respond to heat stress and are involved in the caryopsis filling process. ONAC127 and ONAC129 are dominantly expressed in pericarp during grain filling and can form a heterodimer. To investigate the functions of these two ONACs, we obtained CRISPR/cas9 induced mutants and overexpression lines of them. Interestingly, we found that both knock-out and overexpression plants showed incompletely filling and shrunken phenotype of caryopses, which became more severe under heat stress. The shrunken caryopses of these transgenic lines are usually with ectopic accumulation of starch in the pericarp. Transcriptome analyses revealed that ONAC127 and ONAC129 mainly regulate stimulus response, cell wall construction and nutrient transport etc. ChIP-seq analyses identified the direct targets of ONAC127 and ONAC129 in developing caryopses, including monosaccharide transporter OsMST6, sugar transporter OsSWEET4, calmodulin-like protein OsMSR2 and Ethylene-Response AP2/ERF Factor OsEATB. The result suggested that ONAC127 and ONAC129 might regulate the caryopsis filling through sugar transportation and abiotic stress responses. Overall, this study demonstrates the transcriptional regulatory networks involving ONAC127 and ONAC129, which coordinates multiple pathways to modulate caryopsis development and heat stress response at rice filling stage.
INTRODUCTION
Rice caryopsis development starts from the fertilized ovary and finishes with a dehydrated, hard and transparent grain. This process can be divided into three stages by landmark events: cell division, organogenesis and maturation. (Agarwal et al., 2011). A rice caryopsis is mainly composed of endosperm filled with starch grains, and the biosynthesis of starch is closely related to carbohydrates transportation. Carbohydrates are synthesized in leaves and delivered to caryopses via phloem (Patrick, 1997), then unloaded to the pericarp as the energy and materials for starch biosynthesis (Zhang et al., 2007). The dorsal vascular bundle passes through the pericarp is the main nutrient transport tissue in caryopsis (Oparka and Gates, 1981). While the vascular bundles are not contiguous with endosperm tissue (Hoshikawa, 1984), the apoplasmic pathway is the only way for nutrient to reach the starchy endosperm. (Matsuda et al., 1979).
Carbohydrates may enter the nucellar epidermis directly through plasmodesmata, then be transported to the apoplasmic space with sugar transporter protein OsSWEETs (sugar will eventually be exported transporters) and partially hydrolyzed into monosaccharide by cell wall invertase OsCINs. The monosaccharide is transported into the aleurone layer mainly by monosaccharide transporter OsMSTs while the rest un-hydrolyzed sucrose is directly transported into the aleurone layer via sucrose transporter OsSUTs (Yang et al., 2018). The nutrient transportation is also regulated by a range of transcription factors such as OsNF-YB1 and OsNF-YC12, which activate the expression of OsSUTs to capture the leaked sucrose in apoplasmic space. Knockout of NF-YB1 or OsNF-YC12 leads to defective grains with chalky endosperms and similar phenotype has been observed in mutants of OsSUT1 (Bai et al., 2016; Xiong et al., 2019).
The nutrient transport processes in caryopsis are susceptible to changing environmental conditions, extreme external stimuli (especially the extreme temperature) during grain-filling stage will lead to a reduction of nearly 50% in rice (Hu and Xiong, 2014). Plants respond to the unexpected high temperature through some stress-specific signaling pathways. During heat stress signal transduction, the heat shock transcription factors (HSFs) will be activated to regulate the expression of downstream heat shock proteins (HSPs) and other stress-related genes. For example, AP2/EREBP transcription factor DREB2A can activate a heat shock transcription factor hsfA3 under heat stress to induce the expression of specific HSPs (Schramm et al., 2008). Similarly, the expression of OsbZIP60, which play important roles in moisture retention and heat-damage resistance, is induced under heat stress and endoplasmic reticulum stress (Oono et al., 2010). Although several plant heat stress response related proteins have been found and elucidated, the mechanisms of the response are still poorly understood.
NAC (NAM, ATAF1/2, CUC2) transcription factors are one of the largest family of plant specific transcription factors with 151 members in rice (Nuruzzaman et al., 2010). The typical structure of NAC transcription factors is a conserved NAC domain with about 150 amino acids in N-terminus, followed by a variable transcriptional regulation region (TRR) in C-terminus (Christianson et al., 2010). The NAC domain can be further subdivided into five subdomains [A-E], subdomains C and D are highly conserved and may mainly act in DNA-binding (Ooka et al., 2003). In addition, some subdomain D contains a hydrophobic NAC repression domain (NARD) which can suppress the transcriptional activity of NAC transcription factors by suppressing their DNA-binding ability or nuclear localization ability. The specific functional motif in NARD, LVFY can suppress not only the activity of NAC transcription factors, but also the other transcription factors by its hydrophobicity (Hao et al., 2010).
NAC transcription factors are involved in many biological processes in plants, such as organ development, secondary wall synthesis, stress response and so on. For instance, NAC29 and NAC31 regulate downstream cellulose synthase (CESA) by activating the downstream transcription factor MYB61 to control the synthesis of secondary wall (Huang et al., 2015). For stress response, especially in abiotic ways, the NAC transcription factors are master regulators in this process. RD26 gene is the first NAC transcription factor identified as a regulator in mediating abscisic acid (ABA) and jasmonate (JA) signaling during stress responses in Arabidopsis (Fujita et al., 2004). SNAC1 is a stress-responsive NAC transcription factor confers rice drought tolerance by closing stomata (Hu et al., 2006) and SNAC3 confers heat and drought tolerance through regulation of reactive oxygen species (Fang et al., 2015). OsNAC2 is also reported involving in drought stress response mediated by ABA (Shen et al., 2017). Meanwhile, it plays divergent roles in different biological processes, such as regulating shoot branching (Mao et al., 2007), affecting plant height and flowering time through mediating the gibberellic acid (GA) pathway (Chen et al., 2015). VNI2 in Arabidopsis is a bifunctional NAC transcription factor reported as a transcriptional repressor functioning in xylem vessel formation (Yamaguchi et al., 2010), but the transcriptional activator activity of VNI2 is induced by high salinity stress to regulate the leaf longevity (Yang et al., 2011).
Nine caryopsis-specific NAC transcription factors were found through transcriptome analyses in rice (Mathew et al., 2016).We selected four of these genes, ONAC025, ONAC127, ONAC128 and ONAC129 for further analyses, which located in the linked regions of chromosome 11 and were identified as a gene cluster (Fang et al., 2008). As recent study implied two NAC transcription factors specifically expressed in maize seeds, ZmNAC128 and ZmNAC130, playing critical roles in starch and protein accumulation in seeds filling (Zhang et al., 2019). We inferred that the four NAC transcriptional factors might also be involved in some important processes in rice caryopsis development. In this study, we selected ONAC127 and ONAC129 from the four genes, which could form a heterodimer and participate in apoplasmic transportation as well as heat stress response to regulate rice caryopsis filling. These findings broadened the understanding of the regulation network of stress response and grain filling in rice.
RESULTS
ONAC127 and ONAC129 are specifically expressed in caryopsis
To understand these four genes basically, we firstly checked the expression data of ONAC025, ONAC127, ONAC128 and ONAC129 in the microarray database CREP (http://crep.ncpgr.cn) (Wang et al., 2010). The data showed that the expression level of ONAC025 in caryopsis was significantly higher than that of the rest genes, while that of ONAC128 was the lowest among the four genes. For ONAC127 and ONAC129, they showed the highest expression level at 7 d after pollination (DAP) and then decreased gradually, such similar expression pattern attracted our further attention.
For validating the spatial and temporal expression patterns and further exploring the specific expression distribution of ONAC127 and ONAC129 in caryopsis, we mechanically isolated rice caryopses at 5-14 DAP to the mixture of pericarp and aleurone layer, starch endosperm and embryo for qRT-PCR analyses by the method described by Bai (Bai et al., 2016). The aleurone layer specifically expressed gene oleosin and the starch endosperm specifically expressed gene SDBE were used as markers (Ishimaru et al., 2015) (Supplemental Fig. S1). The result showed that ONAC127 and ONAC129 were mainly expressed in pericarp or aleurone layer while they were weakly expressed in starch endosperm, and the expression went to the summit at 5 DAP during the whole development process (Fig. 1A).
The mRNA in situ hybridization analyses using the unripe caryopses of ZH11 turned out that ONAC127 and ONAC129 were dominantly expressed in pericarp, when the expression level in aleurone layer and starch endosperm were quite weak (Fig. 1B). Histochemical GUS activity was also detected using the transgenic plants expressing pONAC127::GUS and pONAC129::GUS, the result was almost identical to that of in situ hybridization analyses, that ONAC127 and ONAC129 were preferentially expressed in pericarp of caryopsis (Fig. 1C). Considering that 5 DAP is a key time point of caryopsis development for the organogenesis is finished and maturation stage is just initiated (Agarwal et al., 2011). ONAC127 and ONAC129 might be involved in some substance exchange processes playing vital roles in caryopsis filling and maturation.
We then carried out the assay through transforming green fluorescent protein (GFP) fused ONAC127 or ONAC129 protein into rice protoplast transiently to investigate the subcellular localization pattern of the two genes. 35S::GFP was used as a positive control and 35S::Ghd7-CFP was used as a nuclear marker (Xue et al., 2008). The fluorescence generated by ONAC127-GFP and ONAC129-GFP was distributed in the nucleus and cytoplasm just like the way positive control did (Fig. 1D). For further confirming the subcellular localization, transgenic plants expressing pUbi::ONAC127-GFP and pUbi::ONAC129-GFP were generated in Zhonghua11 (ZH11) background. The subcellular localization pattern of ONAC127 and ONAC129 were validated by detecting the fluorescence from the roots of two-week-old rice seedlings of the plants, ONAC127 and ONAC129 proteins were indeed localized in the nucleus and cytoplasm (Supplemental Fig. S2).
ONAC127 and ONAC129 can form a heterodimer
It was noticed that the NAC transcription factors usually function as a dimer (Olsen et al., 2005), thus we investigated if there was any interaction between ONAC127 and ONAC129. The result indicated that ONAC127 could interact with ONAC129 in yeast (Fig. 2A). To confirm the interaction between ONAC127 and ONAC129, the in vitro GST pull-down assay was carried out. His-tagged ONAC127 and GST-tagged ONAC129 were expressed respectively and incubated together. The protein mixture of GST and ONAC127-His was used as a negative control. After purification by Glutathione Agarose, ONAC127-His was detected in the sample containing ONAC129-GST instead of the control with His-tag antibody (Fig. 2B). The in vivo bimolecular fluorescence complementation (BiFC) assay was also performed with rice protoplasts. Nuclear homodimer OsbZIP63 was chosen as positive control (Walter et al., 2004), and 35S::Ghd7-CFP was used as a nuclear marker. Yellow fluorescence generated from the interaction between ONAC127-YFPN and ONAC129-YFPC was detected, which confirmed the heterodimer ONAC127 and ONAC129 formed in nucleus (Fig. 2C). In view of previous studies, some of NAC transcription factors were located in the cytoplasm or plasma membrane first, and then were imported to nuclei under some specific conditions (Fang et al., 2014). Considering the subcellular localization pattern of ONAC127 and ONAC129 (Fig. 1D), we speculated that these two proteins might also exert their transcriptional regulation functions in this way.
ONAC127 and ONAC129 are key players in starch accumulation during rice caryopsis filling
We obtained the specifically knockout ONAC127 or ONAC129 mutants (onac127 and onac129) in ZH11 background by CRISPR/Cas9 genome editing system (Ma et al., 2015). Considering that these two proteins could form a heterodimer, double knockout mutants (onac127;129) were also obtained. The sgRNA target sites were designed at the exons of ONAC127 and ONAC129 using the web-based tool CRISPR-P (Liu et al., 2017). There are two target sites of each gene, which were expected to generate different mutations in the coding region of the genes (Fig. 3A). The grains of three T0 homozygous plants of each mutant were selected for generating independent T1 transgenic lines. The sequencing data of the transgenic lines was decoded by the method described by Liu (Liu et al., 2015), and the genotypes of the homozygous lines were showed in Supplemental Fig. S3. On the other hand, we also generated the overexpression lines pUbi::ONAC127-FLAG and pUbi::ONAC129-FLAG (OX127 and OX129) (Fig. 3B). Relative expression level of the overexpression lines was analyzed by qRT-PCR (Supplemental Fig. S4).
During the rice reproductive stage, we found the development of caryopses in ONAC127 and ONAC129 transgenic lines arrested during 5-7 DAP (Supplemental Fig. S5), resulting in shrunken and incompletely filled grains after ripening (Fig. 3C, D). As we mentioned before, the expression level of ONAC127 and ONAC129 was very high during 5-7 DAP (Fig. 1A), the functions of ONAC127 and ONAC129 seem to be closely related to grain filling, especially at the early stage of maturation. Since the phenotype is illogical, we detected the expression level of the two ONACs in the transgenic lines immediately. It turned out that the expression of ONAC127 and ONAC129 was up-regulated in the overexpression lines and down-regulated in the mutants indeed (Supplemental Fig. S6). We speculated that ONAC127 and ONAC129 may regulate several different pathways contributing to caryopsis filling at the same time. The expression of the two ONACs may need to be maintained at a steady level, any fluctuations in their expression would interfere the normal running of different pathways and have the chance to bring caryopsis filling defection.
To investigate the reason of the shrunken phenotype further, histological analyses through resin embedded sections were prepared with 7 DAP incompletely filling caryopses. In ZH11, starch was stored transiently in the pericarp during the early stage, the degradation of starch in the pericarp correlated with starch accumulation in the endosperm (Wu et al., 2016). As we expected, the starch in the endosperm of ZH11 caryopses was arrayed neatly, densely and hardly found in pericarp (Fig. 3E). In contrast, many starch grains accumulated in the pericarp of the mutants and the overexpression lines of ONAC127 and ONAC129, especially in the double mutant onac127;129, in which substantial starch grains were hardly seen in endosperm (Fig. 3E). The finding above suggested that ONAC127 and ONAC129 might play critical roles in starch translocation and mobilization towards the developing endosperm.
ONAC127 and ONAC129 are involved in heat stress response
With the measurement of agronomic traits, the transgenic lines of ONAC127 and ONAC129 showed significant reduction in both seed set rate (percentage of fully filled seeds after harvest) and 1000-grains weight (Fig. 4A, B). The percentage of shrunken grains showed an obvious increase while proportion of blighted grains (glumes with no seed in it) did not change significantly in these transgenic lines (Fig. 4C). It proved that the reduction of seed set rate and yield was mainly caused by the increase of shrunken grains.
Notably, the transgenic plants during filling stages were suffered from severe heat stress in the summer of 2018 (Fig. 4D, E). Previous studies have shown that the ambient temperature higher than 35°C would have a serious impact on the yield at flowering and grain filling stages (Hakata et al., 2017). Therefore, we set 35°C as the heat damage temperature of rice (TB), and calculated the heat damage accumulated temperature per hour (THi), heat damage accumulated temperature during filling stage (TS) and heat damage hours during filling stage (HS) as described by Chen (Chen et al., 2019). Given that ONAC127 and ONAC129 function during early filling stage, we also calculated the heat damage accumulated temperature during 0-7 DAP (TS7) and heat damage hours during 0-7 DAP (HS7). Noting that we cultivated two batches of rice in the summer of 2018, the first batch of plants were flowering on about July 20 when the second batch goes on about August 20. It’s obvious that TS and HS of the first batch were much higher than that of the second batch, especially at the early development stage of caryopsis (Fig. 4D, E).Therefore, the caryopses of the plants cultivated in the first batch were defined as suffering from heat stress while those in the second batch were defined as planted under normal conditions with little or no heat stress. We found that heat stress caused sharp increase of shrunken grains compared to normal conditions, and such increase in onac127;129 was more dramatic than the other transgenic lines (Fig. 4C), which led to significant decrease of seed set rate and 1000-grains weight (Fig. 4A, B). Meanwhile, the expression level of ONAC127 and ONAC129 was highly increased under heat stress (Supplemental Fig. S6), which suggested that ONAC127 and ONAC129 might be involved in heat stress response during caryopsis filling stage.
ONAC129 negatively regulates ONAC127 transcriptional activity
For deeper exploring the relationship of ONAC127 and ONAC129, we performed a dual luciferase assay in rice protoplasts as the transcriptional activity of the two genes was not detected in yeast assays (Fig. 2A). ONAC127 and ONAC129 were fused with the yeast GAL4 binding domain (GAL4BD) as effectors to be co-transformed into rice protoplasts with a firefly luciferase reporter gene driven by a CaMV35S promotor (35S-GAL4-fLUC; Fig. 5A). A significant increase in the transcriptional activation activity of ONAC127 and ONAC129 was detected, while the transcriptional activity of ONAC127 was significantly restrained when ONAC129 was co-transformed into rice protoplasts (Fig. 5B). These results hinted that ONAC129 might regulate the transcriptional activity of ONAC127 negatively and we wonder if ONAC129 could regulate the expression of ONAC127. However, we didn’t find significant fluctuation of the expression of ONAC127 in the transgenic lines of ONAC129 and vice versa. (Supplemental Fig. S6).
As we know, several NAC transcription factors contain both transcriptional activation domain and NAC repression domain (NARD), which means that the combinatorial effects of both transcriptional regulation activities determine downstream events (Hao et al., 2010). Hence, we decided to BLAST and analyze the protein sequences of ONAC127 and ONAC129 through the method described by Hao, and we found that there were NARD-like sequences in NAC subdomain D indeed in both genes (Fig. 5C), which suggested that ONAC127 and ONAC129 might be bifunctional transcription factors that could activate or repress downstream genes in different situation.
ONAC127 and ONAC129 regulate the transcription of genes related to sugar transportation and abiotic stimulus
RNA-Seq analyses worked in identifying differentially expressed genes. The 7 DAP caryopses of the overexpression lines and mutants of ONAC127 and ONAC129 under heat stress and normal conditions were used to generate RNA-seq libraries. There are 7056, 6485, 5765, and 1608 differentially expressed genes (DEGHs) identified in onac127 (onac127H), onac129 (onac129H), OX127 (OX127H), OX129 (OX129H) respectively under heat stress, compared with ZH11 (ZH11H), and 15334, 13284, 1776, 2041 differentially expressed genes (DEGNs) which were identified in onac127 (onac127N), onac129 (onac129N), OX127 (OX127N), OX129 (OX129N) respectively under normal conditions compared with ZH11 (ZH11N) (Fig. 6A, B; Supplemental Dataset S1). Gene Ontology (GO) enrichment analyses was performed to examine the biological roles of DEGs in caryopsis development. The significantly enriched GO terms were mainly associated with stimulus response, transcriptional activity regulation, signal transduction, cell wall construction and substance transportation (Fig. 6C).
Subsequently, the genes directly bound by ONAC127 and ONAC129 in caryopsis were identified using Chromatin Immunoprecipitation Sequencing assays (ChIP-seq). The ChIP assays were performed using the anti-FLAG antibody with 7 DAP caryopses of OX127 (OX127H), OX129 (OX129H) under heat stress, and OX127 (OX127N), OX129 (OX129N) under normal condition. The expression of the ONAC127-FLAG and ONAC129-FLAG fusion proteins was verified by western blot analyses to validate the effectiveness of the FLAG tags (Supplemental Fig. S7). After sequencing, 185, 6125, 220, 455 peaks were finally obtained respectively in OX127H, OX127N, OX129H and OX129N (Supplemental Dataset S2), and 54, 5092, 82, 212 putative ONAC127 and ONAC129 bound genes were identified in OX127H, OX127N, OX129H and OX129N respectively among the peaks (Fig. 6D). Then, we identified the binding motifs of the ONACs using MEME (Machanick and Bailey, 2011), and found that the significantly enriched motif was ‘CT(C)TTCT(C)TT’ (Fig. 6E, F), in line with ‘TT(A/C/G)CTT’, the specific motif of transmembrane NAC transcription factors NTL6 and NTL8 (Lindemose et al., 2014). Both the two genes were associated with stimulus response, implying that ONAC127 and ONAC129 were probably involved in stress response. Notably, the motif ‘CT(C)TTCT(C)TT’ was only found significantly in OX127H and OX129H. There was no significantly enriched motif in OX127N and OX129N, implying that the target genes of ONAC127 and ONAC129 might be more specific under heat stress.
For further exploring the target genes regulated by ONAC127 and ONAC129, we defined the genes which were either consistently bound by ONAC127/129 or simultaneously bound by ONAC127 and ONAC129 under heat stress or normal conditions as ONAC127/129 preferable-bound genes (Fig. 6D, red trapezoid). Based on the data from RNA-Seq analyses and phenotype of transgenic lines of ONAC127 and ONAC129, 8 genes with respect to sugar transportation or abiotic stress response were selected from 136 preferable-bound genes as the potential target genes (PTGs) of ONAC127 and ONAC129 for further investigation.
ONAC127 and ONAC129 play a pivotal role in caryopsis filling through regulating key factors directly in substance accumulation and stress response
To validate the interactions between ONAC127/129 and promotors of their potential target genes (PTGs), the yeast one-hybrid assays were performed. It suggested that ONAC127 and ONAC129 could bind to the promoter sequences of OsEATB, OsHCI1, OsMSR2, OsSWEET4, bHLH144 and OsMST6 directly in yeast (Fig. 7A). The Dual-LUC assays in rice protoplasts were performed to figure out whether ONAC127 and ONAC129 affect the transcription of these genes directly. ONAC127 and ONAC129 were co-transformed into rice protoplasts with a firefly luciferase reporter gene driven by target gene promotors (Fig. 7B). It turned out that ONAC127 activated OsMSR2 and OsMST6 promotors significantly, and both ONAC127 and ONAC129 repressed the OsEATB and OsSWEET4 promotors strongly in vivo. It is noteworthy that ONAC129 could suppress the transcriptional activation activity of ONAC127 to OsMSR2 and OsMST6 promotors, while ONAC129 barely regulated the transcription of OsMSR2 and OsMST6 alone (Fig. 7C).
For testing whether the endogenous ONAC127 and ONAC129 bound to the target genes specifically, ChIP-qPCR was performed using the same rice materials of ChIP-Seq. The result indicated that ONAC127 and ONAC129 bound to the promoters of OsEATB, OsMSR2, OsSWEET4 and OsMST6 respectively under heat stress and normal conditions (Fig. 7D, E, F, G). The qRT-PCR for these genes was also performed in the 7 DAP caryopses of transgenic lines of ONAC127 and ONAC129, showing that the expression of OsEATB and OsSWEET4 was significantly up-regulated in the mutants and generally down-regulated in the overexpression lines. The expression of OsMSR2 and OsMST6 was both up-regulated in onac129 and down-regulated in onac127 (Supplemental Fig. S8). The results above indicated that OsEATB, OsMSR2, OsSWEET4 and OsMST6 were the direct targets of ONAC127 and ONAC129 in rice during caryopsis filling.
DISCUSSION
In this study, two rice caryopsis-specific NAC transcription factors ONAC127 and ONAC129 were identified, which can form a heterodimer and are predominantly expressed during the early and middle stage of rice caryopsis development (Fig. 1, 2). The caryopses of the transgenic plants were obviously unfilled (Fig. 3C, D, E; Supplemental Fig. S5), and the proportion of shrunken grains was higher under natural heat stress (Fig. 4C). By protein−DNA binding assays, we found that ONAC127 and ONAC129 may regulate the expression of the sugar transporters, OsMST6 and OsSWEET4 directly. Both the two genes are involved in the transportation of photosynthate from dorsal vascular bundles to endosperm during caryopsis filling (Wang et al., 2008; Sosso et al., 2015). ONAC127 and ONAC129 also bind to the promotors of OsMSR2 and OsEATB directly (Fig. 7), which participate in abiotic stress responses by responding to calcium ion (Ca2+) or plant hormones (Qi et al., 2011; Xu et al., 2011). These findings suggest that ONAC127 and ONAC129 may be involved in multiple pathways, therefore interfere the abiotic stress response and caryopsis filling process during rice reproductive stage.
ONAC127 and ONAC129 are involved in the apoplasmic transportation of photosynthates as a balancer
At the early stage of caryopsis development, some of the photosynthate from the dorsal vascular bundle was transported to endosperm through the apoplasmic space formed by degenerative nucellar cell wall, the others synthesized the starch grains in mesocarp cells (Hoshikawa, 1984). The starch accumulation in pericarp will reach a maximum at 5 DAP. After that, the starch in mesocarp cells will be disintegrated and the nutrient will be transported to the endosperm due to the continuous process of endosperm cells filling. The cell contents of pericarp cells will disappear, and the cells will eventually dehydrate and die, leaving only the cuticularized cell wall remnants (Wu et al., 2016).
Since the abnormal expression of ONAC127 and ONAC129, endosperm cells in abnormal caryopses may not obtain enough photosynthate for filling due to the functional defect of apoplasmic transport pathways. The photosynthates accumulates in pericarp as starch grains without been disintegrated, thus stop the endosperm cells from proliferation or filling by starch, eventually leading to incompletely filled and shrunken caryopsis (Fig. 3C, E).
It is noteworthy that the target transporters OsSWEET4 and OsMST6 are predominantly expressed at early stages of caryopsis development, and may act in downstream of the cell wall invertase OsCIN2 to import hexose into aleurone layer through apoplasmic space (Wang et al., 2008; Sosso et al., 2015; Yang et al., 2018). As previously mentioned, ONAC127 and ONAC129 suppress the expression of OsSWEET4 directly (Fig. 7C; Supplemental Fig. S8), and ossweet4-1, the mutant of OsSWEET4, has a distinct phenotype with shrunken grains similar with the abnormal caryopses in transgenic lines of ONAC127 and ONAC129 (Sosso et al., 2015). The fact confirmed our hypothesis that ONAC127 and ONAC129 are involved in the apoplasmic transport pathways directly by regulating the expression of transporters indeed.
Moreover, the shrunken caryopses we observed in mutants and overexpression lines of ONAC127 and ONAC129 are also similar with the incompletely filled caryopses of the double mutants ossweet11;15, in which two apoplasmic transporters OsSWEET11 and OsSWEET15 were dysfunction (Yang et al., 2018). For that OsSWEET15 has been excluded from the target genes by the yeast one-hybrid assays (Fig. 7A), we detected the expression of OsSWEET11 and OsSWEET15 in transgenic lines of ONAC127 and ONAC129 by qRT-PCR and found that both genes might be activated by ONAC127 and ONAC129 indirectly (Supplemental Fig. S9). The fact suggested that ONAC127 and ONAC129 might activate the expression of OsSWEET11/15 and suppress the expression of OsSWEET4 simultaneously. This may explain the fact that the phenotypes of mutants and overexpression lines of ONAC127 and ONAC129 are almost the same in incompletely filled caryopses (Fig. 3C, D, E; Supplemental Fig. S5). Hence, we speculate that ONAC127 and ONAC129 may act as a balancer in apoplasmic transportation by activating or suppressing the expression of transporters dynamically (Fig. 8), therefore keeping the amount of several different apoplasmic transporters at different stages of caryopsis development.
ONAC129 suppresses the transcriptional activity of ONAC127 in some pathways, meanwhile, ONAC127 and ONAC129 may regulate downstream genes differently at dimer state and monomer state (Fig. 5B, 7C). We conjecture that the proportion of ONAC127/129 dimers may be vital in adjusting the balance of the amount of apoplasmic transporters at different stages of caryopsis development. Whether to knockout or to over-express ONAC127 or ONAC129, the proportion of the monomers and dimers of ONAC127/129 will be changed inevitably, which induces various changes in expression of the apoplasmic transporters, thus the apoplasmic transportation defects, leading to the incomplete filling caryopsis phenotype (Fig. 3C). In macroscopic view, the balance regulation of apoplasmic transportation may be a “rate-limiting step” of rice reproductive growth to ensure that the photosynthate is enough for both the filling of grains and the energy for plants to avoid premature aging, thus may be vital for plants to keep the balance of vegetative growth and reproductive growth.
ONAC127 and ONAC129 respond to heat stress during caryopsis filling
Heat stress has profound effects on plant growth, particularly in reproductive development. When plants encounter heat stress, the rise of cytosolic Ca2+ concentration is the most rapid response (Liu et al., 2003). With the increase of Ca2+ concentration, the expression of calmodulins (CaMs) and Calmodulin-like genes (CMLs) will be activated and modulate the phosphorylation state of HSFs to regulate their DNA binding ability, thereby regulating the expression of heat stress related genes including HSPs and some plant hormone responsive genes (Li et al., 2018).
The proportion of incompletely filled caryopses under heat stress is much higher than that of normal condition (Fig. 4C), and the expression of ONAC127 and ONAC129 was induced by heat stress (Supplemental Fig. S6), suggesting that heat stress might interfere the dimerization of ONAC127 and ONAC129 to affect the balance of apoplastic transportation. As previously mentioned, ONAC127 and ONAC129 regulate the expression of OsMSR2 and OsEATB directly (Fig. 6C; Supplemental Fig. S8). Previous study showed that OsMSR2, which can be strongly induced by heat, cold and drought stress, is a calmodulin-like protein working by binding Ca2+ and speculated having the function as a Ca2+ sensor in plant cells (Xu et al., 2011). OsEATB is a rice AP2/ERF gene which can suppress gibberellic acid (GA) synthesis while the GA signaling pathway involved in the regulation of cell elongation and morphogenesis under heat stress (Koini et al., 2009; Qi et al., 2011). We hereby speculated that ONAC127 and ONAC129 might participate in heat stress response by regulating the Ca2+ and GA signaling pathway.
The expression level of OsHSP101 and OsHCI1 was changed significantly in the transgenic lines of ONAC127 and ONAC129 (Fig. 7A, C; Supplemental Fig. S9). We believe that they are also the potential target genes of ONAC127 and ONAC129, though they may not be regulated directly. RING E3 ligase gene OsHCI1 is induced by heat stress and mediates nuclear–cytoplasmic trafficking of nuclear substrate proteins via mono-ubiquitination to improve the heat tolerance under heat shock (Lim et al., 2013). Meanwhile, OsHSP101 function as one of the important molecular chaperones that interact with OsHSA32 and OsHsfA2c, the latter of which is one of the central regulators of heat stress response (Singh et al., 2012; Lin et al., 2014). Accordingly, we believe that ONAC127 and ONAC129 are involved in some core reactions in the heat stress response during rice caryopsis development stage.
Ca2+ is one of the most important second messengers in heat stress response, cytosolic Ca2+ concentration would change rapidly in a short time when heat stress occurs (Li et al., 2018). The stress signal would be then transmitted by CaMs/CMLs like OsMSR2 to influence downstream gene expression consequently. When the downstream HSFs are activated, HSPs including OsHSP101 would be induced to express and function, and some nuclear-cytoplasmic trafficking reactions involving OsHCI1 would occur (Singh et al., 2012; Lim et al., 2013). Besides, a variety of plant hormones including GA would act in heat stress response to maintain plant growth, and OsEATB plays an important role in GA-mediated plant cell elongation (Qi et al., 2011; Li et al., 2018). Considering that all these direct or indirect target genes of ONAC127 and ONAC129 play vital roles in the key steps of the heat stress response regulatory network, we speculate that ONAC127 and ONAC129 may be the core transcription factors in heat stress response during rice caryopsis filling stage, and the stress and hormone response may be involved in the complex regulatory network of apoplasmic transportation.
MATERIALS AND METHODS
Generation of transgenic plants and growth conditions
For generating CRISPR mutants, we used the web-based tool CRISPR-P v2.0 (http://cbi.hzau.edu.cn/CRISPR2/) designed by Liu (Liu et al., 2017) to get the specific gRNA cassettes targeting ONAC127 and ONAC129 and then cloned into a binary vector pYLCRISPR/Cas9-MH (Ma et al., 2015). For overexpression plants, we amplified the stop-code-less cDNA fragments of ONAC127 and ONAC129, fused 3×Flag or EGFP coding sequence at the 3’ end. The fused sequences were cloned into the binary vector pCAMBIA1301U (driven by a maize ubiquitin promotor). For tissue specific expression analyses, a 2000bp fragment of the 5’ upstream region of ONAC127, and a 2043bp fragment of the 5’ upstream region of ONAC129 were cloned into the binary vector pDX2181G (with GUS, β-glucuronidase).These recombinant constructs were introduced into rice Zhonghua11 (ZH11; Oryza sativa ssp. japonica) by Agrobacterium tumefaciens (EH105)-mediated transformation (Lin and Zhang, 2005). The plant materials were cultivated in paddy fields in Huazhong Agricultural University, Wuhan, China, under natural long-day conditions (approximately 12-14 h light/10-12 h dark) during May to October 2018. The temperature of growing areas during rice reproductive stage was showed in (Fig. 4D, E). We set 35°C as the heat damage temperature of rice (TB), and calculated the heat damage accumulated temperature per hour (THi), (Ti is the ambient temperature at i hour); heat damage hours during filling stage (HS), and heat damage accumulated temperature during filling stage (TS), using the method described by Chen (Chen et al., 2019). The heat damage accumulated temperature during 0-7 DAP (TS7) and heat damage hours during 0-7 DAP (HS7) was also calculated as the method above. The phenotypes were detected in homozygous T1 generation of transgenic plants. The primer sequences used in this study were listed in Supplemental Table S1 and would not be repeated in the following article.
Histochemical GUS staining
Unripe caryopses at 5 and 7 DAP from pONAC127::GUS and pONAC129::GUS transgenic plants were collected for the GUS staining assays following the methods described by Yang (Yang et al., 2018).
Microscopy analyses
For semithin section microscopy analyses, the unripe caryopses at 7 DAP were collected in 2.5% glutaraldehyde for fixation, vacuum infiltrated on ice for 30 min and incubated at 4°C for 24h. The plastic embedding and sectioning were performed as described by Wang (Wang et al., 2008). Slides were stained with toluidine blue and detected by a BX53 microscope (Olympus).
RNA in situ hybridization
Unripe caryopses (1-10 DAP) from rice ZH11 were collected for paraffin embedding. Paraffin section was performed according to the previous method (Xiong et al., 2019). Gene-specific fragments of ONAC127 and ONAC129 were amplified by the primers used in RT-PCR and cloned into the pGM-T vector. The probes were synthesized using the DIG RNA labeling kit (SP6/T7) (Roche) according to the manufacturer’s recommendations. RNA hybridization and immunologic detection of the hybridized probes were performed on sections as described by Kouchi (Kouchi and Hata, 1993). Slides were observed using a BX53 microscope (Olympus).
Transient expression assays
To investigate the subcellular localization of ONAC127 and ONAC129, the CDS of these genes were cloned into the pM999-35S vector, the nuclear located gene Ghd7 was used as a nuclear localization marker (Xue et al., 2008).For the BiFC assays, the CDS of ONAC127 and ONAC129 were cloned into the vectors pVYNE and pVYCE (Waadt et al., 2008) respectively. Plasmids were extracted and purified using the Plasmid Midi Kit (QIAGEN) and were then transformed into rice protoplasts according to the procedure described by Shen (Shen et al., 2017). The fluorescent signals were detected with a confocal laser scanning microscope (TCS SP8, Leica).
RNA isolation and transcript analyses
Total RNA was isolated using the TRIzol method (Invitrogen), and the first strand cDNA was synthesized using HiScript II Reverse Transcriptase (Vazyme). The real-time PCR was performed on QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems), with the 2-ΔΔCt method for relative quantification (Livak and Schmittgen, 2001). The significance of differences was estimated using Student’s t-test. Relevant primers were designed according to qPrimerDB (https://biodb.swu.edu.cn/qprimerdb) (Lu et al., 2018).
Yeast assays
For yeast two-hybrid assay, the CDS of ONAC127 and ONAC129 were cloned into the pGBKT7 and pGADT7 vector (Clontech). The fusion plasmids were transformed into yeast strain AH109 or Y187. The pGBKT7-53 was co-transformed with pGADT7-T as a positive control. For yeast one-hybrid assay, DNA fragments about 1000bp corresponding to the promoters of target genes were independently inserted into the pHisi-1 plasmid (Clontech). ONAC127 and ONAC129 were fused to GAL4 transcriptional activation domain (pGAD424). These constructs were transformed into the yeast strain YM4271. Yeast two-hybrid assay and one-hybrid assay were performed following the manufacturer’s instructions (Clontech).
Chromatin immunoprecipitation (ChIP)
Overexpression lines of ONAC127 and ONAC129 (fused with 3×FLAG) were used for ChIP-Seq analyses. We verified the expression of target proteins by western blot using ANTI-FLAG M2 Monoclonal Antibody (Sigma-Aldrich). ChIP assay was performed with ANTI-FLAG M2 Magnetic Beads (Sigma-Aldrich) as the method described previously (Xiong et al., 2019). For each library, three independent replicated samples were prepared. The immunoprecipitated DNA and input DNA were subjected to being sequenced on the HiSeq 2000 platform (Illumina), which was processed by the Novogene Corporation.
ChIP-Seq raw sequencing data was mapped to the rice reference genome (RGAP ver. 7.0, http://rice.plantbiology.msu.edu) (Kawahara et al., 2013) using BWA (Li and Durbin, 2009). MACS2 (Zhang et al., 2008) was used for peak calling and the peaks were identified as significantly enriched (corrected P-value <0.05) in the IP libraries compared with input DNA. Visual analyses were performed using IGV (Intergative Genomics Viewer, v2.3.26) (Robinson et al., 2011). Motif enrichment analyses were performed by MEME (Machanick and Bailey, 2011) with default parameters.
To validate the specific targets genes, the immunoprecipitated DNA and input DNA was applied for ChIP-qPCR analyses. The enrichment value was normalized to the input sample. The significance of differences was estimated using Student’s t-test.
RNA-Seq
Total RNA was extracted from unripe caryopses at 7 DAP. For each library, three independent replicated RNA samples were prepared. The RNA samples were then sequenced on the HiSeq 2000 platform (Illumina), which was processed by the Novogene Corporation.
The raw reads were filtered for adaptors and low-quality reads. Clean reads were mapped to the reference genome of rice (RGAP v. 7.0) using HISAT2 (v.2.0.5) (Kim et al., 2015). The gene expression level was calculated by FPKM method (Trapnell et al., 2010), and the differentially expressed genes (|Fold Change| ≥ 2 and p-value<0.05) were selected by DESeq2 software (Love et al., 2014). GOseq (Young et al., 2010) was used for GO enrichment analyses, and p-value was converted to −log10(p-value) for display.
In vitro GST pull-down assay
The CDS of ONAC127 and ONAC129 were cloned into pET28a and pGEX-4-1 vectors respectively for His-tagged ONAC127 and GST-tagged ONAC129 protein expression in vitro. The fusion plasmids were transformed into Escherichia coli BL21 strain. The GST pull-down assays were performed according to the previous methods (Xiong et al., 2019). The protein was separated on a 10% SDS-PAGE gel and further analyzed by immunoblotting using anti-His and anti-GST antibody (Sigma-Aldrich).
Dual Luciferase Transcriptional Activity Assay
Full-length cDNAs of ONAC127 and ONAC129 were cloned to yeast GAL4 binding domain vectors (GAL4BD) and “None” as effectors. The 35S-GAL4-fLUC and 190-fLUC were used as reporters, and AtUbi::rLUC was used as an internal control. The constructed plasmids were purified and transformed into rice protoplasts with the procedure mentioned previously. Dual Luciferase Reporter Assay System (Promega) was used to measure the luciferase activity by Tecan Infinite M200 (Tecan).
Agronomic traits analyses
Harvested ripe rice grains were air dried and stored at room temperature for at least 2 months before measuring. Only full grains were used for measuring 1000-grain weight which was calculated based on 100 grains. The seed set rate/composition was calculated based on the grains from 3-5 panicles (include the full, shrunken and blighted grains). All measurements of the positive transgenic plants were undertaken using three independent lines.
Accession numbers
The sequence data from this paper can be found in the RGAP database (http://rice.plantbiology.msu.edu) under the following accession numbers: ONAC025 (LOC_Os11g31330), ONAC127 (LOC_Os11g31340), ONAC128 (LOC_Os11g31360), ONAC129 (LOC_Os11g31380), OsMST6 (LOC_Os07g37320), OsSWEET4 (LOC_Os02g19820), OsEATB (LOC_Os09g28440), OsMSR2 (LOC_Os01g72530), bHLH144 (LOC_Os04g35010), OsHCI1 (LOC_Os10g30850), HSP101 (LOC_Os05g44340), OsSWEET11 (LOC_Os08g42350), OsSWEET15 (LOC_Os02g30910), OsbZIP63 (LOC_Os07g48820). The RNA-seq and ChIP-seq data are deposited in the NCBI Gene Expression Omnibus (Edgar et al., 2002) with accession number GSE140167.
Supplemental Material
Supplemental Fig. S1 Expression level of mechanical isolation marker in isolated tissue.
Supplemental Fig. S2 Subcellular localization of the ONAC127 and ONAC129 in the roots of two-week-old rice seedlings of pUbi::ONAC127-GFP and pUbi::ONAC129-GFP transgenic plants.
Supplemental Fig. S3 Mutation sites in onac127, onac129 and onac127;129 lines, as compared with wild-type (ZH11) sequences.
Supplemental Fig. S4 The relative expression level of the overexpression lines of ONAC127 and ONAC129.
Supplemental Fig. S5 Different stages of caryopses development.
Supplemental Fig. S6 Expression level of ONAC127 and ONAC129 in 7 DAP caryopses of transgenic lines compared with that of ZH11.
Supplemental Fig. S7 Detection of FLAG fusion proteins in the ZH11 and overexpression lines.
Supplemental Fig. S8 Expression level of the target genes of ONAC127 and ONAC129 in 7 DAP caryopses of transgenic lines compared with that of ZH11.
Supplemental Fig. S9 Expression level of the indirect target genes of ONAC127 and ONAC129 in 7 DAP caryopses of transgenic lines compared with that of ZH11.
Supplemental Table S1 Primers used in this study.
Supplemental Dataset S1 Differentially expressed genes.
Supplemental Dataset S2 Binding sites identified by ChIP-seq.
Author contributions
Y.R. and J.Y. conceived the original research plans; Y.R. carried out most of the experiments; Z.H., Z.W., F.W. and Y.X. provided technical assistance to Y.R.; Y.R. and H.J. analyzed the data; Y.R. conceived the project and wrote the article with contributions of all the authors; J.Y. supervised and complemented the writing.
ACKNOWLEDGEMENTS
We thank Prof. Min Chen, Prof. Honghong Hu, Prof. Yidan Ouyang and Ms. Wenjing Guo for helping revise the manuscript. This research was supported by grants from the National Natural Science Foundation of China (no. 31570321). The funders had no role in the study design, data collection and analysis, the decision to publish, or in the preparation of the manuscript.