3. RESULTS AND DISCUSSION
3.1. Sucrose content at different growth stages under different light conditions
The synthesis and transportation of sucrose were affected to some extent under low light conditions. The sucrose content in the grains gradually decreased during the grain development process. Under low light conditions, the content of sucrose in the rice grains was higher than the control before 10 d, it was lower than the control at 15 d, and higher than the control at maturity (Figure 1a). This result might indicate that, under low light conditions, the synthesis and transportation of sucrose, the primary photosynthetic product in plants, follows a dynamic process.
Sucrose enters the starch synthesis pathway after being transported to the rice grain through vascular bundles, and it regulates changes in sugar signals, inducing changes to various metabolic cycles in the grain. Studies have shown that sugar starvation may induce the expression of an amylase synthesis gene and improve its activity. Reductions in starch synthesis result in a decrease in starch accumulation and rice quality (L. Wang et al., 2016; Li Wang et al., 2013). Abscisic acid plays an essential role in the response of plants to various environmental stresses. In the developing grains in the control group, the ABA content first increased, then it reached a peak at 15 days, and then it gradually decreased (Figure 1b). After the shading treatment, the ABA concentration in the grain was significantly higher than the control at the beginning, it gradually decreased with time, and it was lower than the control after 15 days. These findings are similar to other research that found that the ABA concentration in firm and weak rice grains showed a decreasing trend during grain filling, but the ABA concentration decreased faster in the firm grains than in the weak grains at the same grain filling stage (Ang, Liang, & Lin, 2003; J. Yang, Zhang, Wang, Liu, & Wang, 2005). In a study on wheat, 9-14 days of spraying low concentrations of exogenous ABA was found to improve wheat grain sucrose synthetase (SuSase), soluble starch synthase, and ADP glucose focal phosphorylase activity (J. Yang, Zhang, Wang, Xu, & Zhu, 2004). Among these, SuSase activity is closely related to library strength and is considered an indicator of library strength. This finding infers that ABA promotes grain filling by increasing the reservoir strength by regulating the critical enzyme activity of sucrose-starch metabolism in grains (F. Wang, Sanz, Brenner, & Smith, 1993).
3.2. Overview of the circular RNAs in rice grains
Although there have been some reports on circRNAs in rice (Chen et al., 2018; Liu et al., 2017; Y. Wang et al., 2017; Z. Wang et al., 2017; Ye et al., 2015; Zuo et al., 2016), research on changes in circRNAs during the development of rice grains under different light treatments is limited. Circular RNAs are characterized by the occurrence of 3’-5’ connections in splicing reactions of individual RNA molecules. Based on the sequence reads, a total of 8015 circRNA candidates were identified by the Starchip, CIRCexplorer2, CIRI2, and CircRNA_finder software from a total of 18 samples from the two treatments of rice seeds (Figure 2a). This is 240.48% more than previously found in rice (Lu et al., 2015). After filtering and comparing the data from the two databases, Plantcircbase and Plantcircnet, 1661 known circRNAs were obtained (Supplemental Table 1). A total of 7270 of the sources of circRNA formation were concentrated in coding RNA, with 5507 in the exon region, making up 85.73% of all sources in the exon region, 373 sources were found in other regions, with 306 in the exon region making up 4.76% of all exon regions, 367 were from lincRNA, with 238 in the exon region making up 3.70% of all exon regions, and only five were from miRNA from other regions (Figure 2b).
3.3. Identification of differentially expressed circular RNAs
At different developmental stages, the number and expression of circRNAs were different. As fertility progressed, the amount of circRNA increased gradually under the control and low light conditions (Figure 2a). Overall, there was a small difference between the quantities detected under low light conditions (1455, 2032, 2360) and the control (1448, 2141, 2748) after 5, 10, and 15 days, respectively. There were 582 circRNAs detected at different times and under different light treatments, accounting for 7.26% of the total number of circRNAs. This result shows that a certain number of circRNAs show stable expression during the grain development process regardless of the environment, and this plays an important role in grain-filling development (Figure 2a).
Different types of circRNAs play different roles during different stages of seed development, and the number of circRNAs involved at different stages varies greatly. For example, 59 were detected in both the control and treatment after 5 d of shading, 36 after 10 d, and 295 after 15 d. Forty-three circRNAs were detected at both 10 d and 5 d, and 272 were detected at both 10 d and 15 d. However, in the control group, only 11 circRNAs were detected after each time period (Figure 2b). The number of upregulated expressions of circRNA at 10 d compared with 5 d under low light conditions was 96, which is significantly higher than that of 16 in the control group (Figure 2c, Supplymental Table 2). Such circRNAs may play different roles in grain development under low light conditions.
In this study, there was no differential expression of circRNA under different light treatment conditions, while in previous studies, there was more differential expression, which may be related to the different tissues that were examined. Previous studies mainly used leaves and roots (Lu et al., 2015; Ye et al., 2015), which express carbon and nitrogen metabolism during plant development, such as photosynthesis and nutrients. However, we only selected grains in rice, which are mainly concerned with the development and accumulation of embryos and endosperm. On the other hand, we also found that although there was no significant difference in the expression of circRNAs under different light treatments, the types of circRNA changed substantially.
3.4. Putative functions of the regulations of rice circRNAs acting as miRNA sponges
The RNAhybrid software predicted 612,874 targeted miRNA, while TargetFinder predicted 3,527, and 1,398 were predicted by both prediction software programs, making the combined number of predicted targeted miRNA 615003. The six circRNAs predicted to be miRNA sponges showed positive amplification from the expected corresponding circular template, while the six circRNAs showed expression differences consistent with the transcriptome sequencing results. In addition, of the 15 circRNA candidate genes that were tested, all were verified to be circular, and 13 showed expression patterns consistent with the RNA sequence results (Figure 3b).
Of the circRNAs that showed differences in light processing and development, 40 circRNAs have been reported as having known miRNA targets, including MiRNA164, miRNA398, miRNA167, and various isomers (Figure 3a). MiRNA164 and MiRNA167 can influence the formation of root caps, lateral root development, and adventitious root development through auxin response factors (Guo, Xie, Fei, & Chua, 2005; Meng et al., 2009), and they can also be induced by light conditions and participate in phyb-mediated signaling pathways (N. Suzuki et al., 2015). MiRNA167 can further influence the development of pollen mother cells to pollen grains by cutting the auxin response factorsLOC_Os10g33940, LOC_Os02g06910, LOC_Os04g57610, LOC_Os06g46410, and LOC_Os12g41950 (Peng et al., 2012). MiRNA164e was used as an OsDBH (DEAD Box Helicase) gene to encode helicase under simulated salt stress, and adapted to salt stress by regulating expression of helicase related genes (Macovei & Tuteja, 2012). The decreased expression level of miR398a and 398b under aluminum stress can cause the up-regulation of the target genes AtSOD1 andAtSOD2 (Superoxide dismutase1/2), while SOD is an important enzyme that promotes the conversion of superoxide free radicals into hydrogen peroxide and oxygen to reduce cell damage (Lima, Arenhart, Margis-Pinheiro, & Margis, 2011). In addition, many circRNAs in this study have not been discovered previously, which also indicates that more needs to be understood about the regulation of targeted miRNA by circRNA.
The results showed that the functions of these known targeted miRNAs are not directly related to the grain development and filling processes. There are two main reasons for this. Firstly, during the development of grains under different light conditions, circRNA might be regulated by a sponge mechanism targeting miRNA. In previous studies, it has also been suggested that during plant growth, circRNA might also regulate the entire biological process by regulating the expression level of the parent gene (Lu et al., 2015). Secondly, there is little evidence of miRNAs in plants playing a role in light regulation (Kumari, Rastoqi, Shukla, Srivastava, & Yaday, 2018), and there are a large number of undiscovered miRNAs in this study. These unknown miRNAs might act on grain filling and be targeted by circRNAs.
3.5. Functional categorization of predicted mRNAs
The differences in the biological processes in the control group from 5-10 d were mainly concentrated in carbohydrate metabolic processes (GO: 0005975), embryo development (GO: 00097990), lipid transport (GO: 0006869), and microtubule-based movement (GO: 0007018). These are essential metabolic and biological development processes in the development of grain embryos and endosperms. In terms of cell structure, monolayer-surrounded lipid storage bodies (GO: 0012511), cell walls (GO: 0005618), membranes (GO: 0016020), and extracellular regions (GO: 0005576) were identified. In terms of molecular function, 698 GO pathways were found to be involved. These were mainly related to hydrolase activity (GO: 000453, GO: 0016788) enzyme inhibitor activity (GO: 0004857), ion channel inhibitor activity (GO: 0008200), lipid binding (GO: 0008289), and carbohydrate metabolism and transport. The enrichment analysis that was conducted using the KEGG database showed significant differences in the expression of starch and sucrose metabolism (ma00500), phenylpropanoid biosynthesis (map00940), and glycerolipid metabolism (map00561). Analyses using the Wikipathway database found differences in seed development (WP2199), photosynthetic carbon reduction (WP1461), abscisic acid biosynthesis (WP626), sucrose metabolism (WP2623), and the ethylene signaling pathway (WP2851).
The difference between 10-15 d and 5-10 d after flowering in the control group was differences in electron carrier activity (GO: 0009055), protein heterodimerization activity (GO: 0046982), and the fructose 6-phosphate metabolic process (GP: 0005975). The main difference in enrichment found from the KEGG database was related to carbon metabolism (map01200), whereas the Wikipathway enrichment analysis found changes in photosynthetic carbon reduction (mapWP1461), glycolysis (mapWP2862), and the tricarboxylic acid cycle (mapWP2624).
When comparing the low light treatment with the control group at 5-10 d after flowering, there was specific differential expression of biological processes such as the abiotic stress response and photosynthesis, and when looking at cell structure, there were differences in nucleosomes (GO: 0000786) and cell wall structure development (GO: 0005618). The expression of the cell wall invertase gene decreases during the development of vulnerable granules with multiple consequences: inhibiting the development of vulnerable granules, delaying the sucrose conversion rate and synthesis rate of vulnerable granules, forming a stagnant period of filling of vulnerable granules, and playing a role in the unloading of assimilating effects (E. Wang et al., 2008). Under low light conditions, the grains of rice also showed similar weaknesses in development, which were similar to grout lag. The photosystem I reaction center (GO: 0009538, 7/2) also showed differential expression. Previous studies have shown that during rice grain filling, some photosynthesis still occurs in the grain. However, under low light conditions, photosynthetic products, such as sucrose, are reduced in the photosynthetically active tissues and thus the supply of these products to the grains is affected. Some photosynthetic cells in the rice grains appeared to be more active to make up for the inadequate synthesis of starch. In terms of molecular functions, protein heterodimerization activity (GO: 0046982), serine type endopeptidase inhibitor activity (GO: 0004867), and peroxiredoxin activity, which play important roles in the accumulation of nutrients, showed differences in activity. For the development of heterodimeric proteins, NAM, ATAF, and CUC (NAC) transcription factors are composed of two chains forming a stable heterogeneous complex. During the process of protein translation in the organization of new polypeptides, synthesis can lag behind and the wrong protein molecules bind (Wiedmann, Sakai, Davis, & Wiedmann, 1994). They are a dynamic component of the ribosome export channel, protect the formation of nascent polypeptides, and prevent inappropriate protein interactions (Rospert, Dubaquié, & Gautschi, 2002). After treatment with high-salt stress, the expression of the OsBTF3 gene in rice is significantly inhibited, and transgenic T2 plants that overexpress OsBTF3 have an increased resistance to high salt and cold stress, while resistance to RNAi transgenic lines is weakened, indicating that OsBFT3 may be present in rice (Li, Chen, Wu, & He, 2012). It plays a very important role in the regulation of high salt and low temperature stress. After the overexpression of Saβ-NAC in Arabidopsis thaliana , the transgenic Arabidopsis thaliana does not grow normally, but it has relatively strong resistance to coastal conditions and drought, and under conditions of abiotic stress and the overexpression of Saβ-NAC, the chlorophyll content and proline content of NAC transgenicArabidopsis plants is increased and the ion dynamic balance is also enhanced. When Spartina alterniflora are exposed to salt damage, drought, cold stress, or a specified concentration of ABA, the expression levels of Saβ-NAC in leaves and roots changes to varying degrees (Karan & Subudhi, 2012). Compared with the control at 5-10 d after flowering, the enrichment analysis conducted using the KEGG database showed differential expression of photosynthesis antenna-proteins (map00196), flavonoid biosynthesis (map00941, 48/8), and starch and sucrose metabolism (map00500).
In the low light treatment for the period 5-10 d after flowering, analysis of the Wikipathway database compared with the control group mainly showed changes related to sucrose metabolism (mapWP2623), ABA synthesis (mapWP626), and the chloroplast electron transport chain (mapWP2861). The content of endogenous ABA in grains under low light in this study was significantly higher than that of the control group at 5 and 10 d after flowering. Abscisic acid regulates sugar metabolism during stress, and α-amylase breaks down starch into fructose and glucose, which has a particular negative regulation effect on the decomposition of sucrose (Hakata et al., 2012). At present, it is believed that the mechanism of ABA regulation under adverse stress conditions, such as drought, high salt or low temperature, might start with adverse stress conditions promoting the accumulation of abscisic acid in plants, which induces the gene expression of ABA response elements and generates resistance to adverse stress. From the point of stimulation by stress to the plant response there are a series of complex information transmission processes, including three links: the cell or tissue must sense the original environmental stimulus and respond by producing an intercellular signal, the intercellular messenger must be transferred between cells or tissues and reach the active site of the receptor cells, and the receptor cells must accept, transduce and respond to the intracellular messenger, which makes the optimal combination of physiological, biochemical, and other functions in the receptor tissues, and finally reflects the adaptation or resistance of plants to environmental stimulation or adversity (Anderson, Ward, & Schroeder, 1994; Knetsch, Wang, Snaar-Jagalska, & Heimovaara-Dijkstra, 1996; Lee et al., 1996; Schwartz, Wu, Tucker, & Assmann, 1994; Stone & Walker, 1995).
The ABA biosynthesis-related gene expression of the low light treatment was significantly different from the control group at 10 days after flowering. Abscisic acid can participate in a series of processes in the adversity stress reaction of plants, it can improve the free proline and soluble sugar, sucrose osmotic regulation substances, and enhance the capacity of osmotic regulation. In this study, the ABA content increased at the same time as the sucrose content, indicating that grain under weak light can exhibit regulation of endogenous ABA to the detriment of resistance to abiotic stress. When the drought stress was postponed to reduce the water content of sesame leaves, the content of MDA increased significantly, and the content of soluble sugar increased first and then decreased (Yan et al., 2008). This result is consistent with our results (Figure 1).
Abscisic acid reduces the activity of ATPase, reduces the transporting force of H+ across the plasma membrane, and then affects the H+ / sucrose co-transport pathway (W., 1980). During the grain-filling period, the lower ABA levels in the early stage and the higher ABA level in the middle stage were not conducive to rice grain filling. The application of exogenous ABA at an appropriate concentration in the early stage of grain filling can promote the development of grain embryos, assimilate transport, promote grain filling, and increase the seed setting rate and yield of rice (Dewdney & Mcwha, 1979; Tang, Xie, Lu, & Liang, 2011). Zhang et al. (1996) used the method of tracer dynamics analysis to apply ABA to paddy rice ears. This can inhibit the formation of temporarily non-exportable substances in the flag leaves, promote the formation of transmissible substances, and increase the output rate of photosynthetic products. After this, it has an inhibitory effect on structural substances and respiratory consumption, and it has a promotion effect on the formation of injectable substances. Therefore, although photosynthesis was inhibited at the early stage of grain filling in this study, the endogenous ABA in the grains did not decrease, the sucrose content was still maintained at a certain level, and the dry weight of the entire ear was not significantly inhibited.
Yang et al. (2006) showed that the regulation effect of ABA on grain filling showed a dose effect: the concentration was promoted, but a high concentration was inhibited as ethylene could reduce the critical enzyme activity of the sucrose-starch metabolism pathway in the grain and inhibit grain filling. The interaction of ABA and ethylene regulates grain filling. Therefore, appropriately increasing the ABA level and the ratio of ABA to ethylene using methods such as moderate soil drought can promote grain filling in rice and wheat (Figure 4) (JiangChang Yang & Zhang, 2018).