2. MATERIALS AND METHODS
2.1. Plant materials and stress treatments
Rice seeds, Oryza sativa Linnaeus, of the Shuhui 498 variety were placed in barrels that were 27 cm high and 33 cm in diameter, and whole plants were placed in a Conviron A1000PG artificial climate box in 2018 and 2019 with light quantity of 700 mol.m-2s-1 for a photoperiod of 14 h at a temperature of 29℃ and a temperature of 20℃ during the 10 h night period (humidity: 75%) when rice was at heading stage. At the beginning of the flowering period, the rice plants were exposed to a low light treatment, of 233 mol.m-2 s-1 and the control remaining unchanged, with the other conditions remaining unchanged, and the glume was removed from the rice grains at the filling stage for RNA separation at 5, 10 and 15 days after treatment with low light.
2.2 Total RNA extraction, circRNA library construction, and sequencing
Total RNA was extracted from the control group and low light stress rice seeds using TRIzol reagent (Invitrogen Corporation, Carlsbad, CA, USA). The total RNA from each sample was used to prepare the circRNA sequencing library. After the RNA samples were qualified for detection, the rRNA was removed from the total RNA sample using the ribo-zero ™ kit. Some long noncoding RNAs (lncRNAs) have the same polyA-tailed structure as mRNA, so the removal of rRNA can maximize the retention of lncRNAs containing polyA-tailed RNA. A fragmentation buffer was added to the enriched RNA to break the RNA into small fragments. Then, a library was constructed using the segmented RNA as a template, and Qubit 2.0 was used for initial quantification and dilution of the library. After this, Agilent 2100 was used to detect the inserted fragment size of the library. After it was established that the inserted fragment size met expectations, real-time PCR was used to accurately quantify the effective concentration of the library to ensure quality of the library. After successfully passing library detection, different libraries were pooled into flow cells according to the requirements for effective concentration and target disembarkation data volume. After cBOT clustering, the Illumina HiSeqX high-throughput sequencing platform was used for sequencing.
2.3. Identification of circular RNAs
The filtered transcriptome sequencing data from each sample were combined, and four software programs, Starchip, CIRI2, CIRCexplorer2, and CircRNA_finder, were independently used to predict the back-spliced junction trans-shear site. The predicted circRNA genome sequence (osa_predict_seq.fa, as the query sequence) was compared with the known circRNA database genome sequence or transcriptome sequence to complete the preliminary verification of the circRNA. The Plantcircbase database, for genome sequences, was used with the Blastn ratio and a filtering condition of subject coverage greater than 90%, and the Plantcircnet database, for transcriptome sequences, was also used to identify circRNA that was located in the intergenic region or single exon region (i.e., circRNA spans multiple exons) with filtering conditions of more than 90% similarity and lengths greater than 100.
2.4. Prediction of miRNA targets of circRNAs, mRNA targets of miRNA, and annotation of functions
The sequence in the middle of the circRNA’s trans-shear site was extracted, and the reverse sequence was spliced from the middle and extended by default by 15 nt. The circRNA-miRNA regulatory relationship was predicted using two software programs, RNAhybrid and TargetFinder, using sequence-based and free energy-based calculations. The RNAhybrid result filter parameter that was used was an energy cutoff of -20 and significant P-value of 0.05. The TargetFinder result filter parameter that was used was a score of three. Regulatory networks were constructed, for the control group and low light treatment for the period 5-10 d after flowering and for the control group for the period 10-15 d after flowering, using differential expression (DE) circRNA-DE miRNA and DE miRNA-DE mRNA, and enrichment analysis was conducted of the mRNAs in the network. The filtering conditions of DE circRNA-DE miRNA were predicted using at least one predictive software. The filtering condition for the pairs of DE miRNA-DE mRNA effects was that the P-value of the expression correlation coefficient was less than 0.05 and was predicted by at least one predictive software program (Supplemental Figures 1-3). An analysis of targeted messenger RNA functions was performed using the Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and Wikipathway databases to predict the processes involved in rice grain development, such as photosynthesis, sugar and starch synthesis and metabolism, and the signaling of plant hormones, such as abscisic acid and auxin. Then the functional annotation results of the GO, KEGG, and Wikipathway databases were used to analyze the enrichment of differentially expressed mRNAs in the network.
2.5. Validation of differentially expressed circular RNAs
Quantitative real-time PCR and Sanger sequencing techniques were used to verify the circular structure and expression patterns of circRNAs identified using RNA sequences. During validation, 15 differentially expressed circRNAs were used, of which six were predicted to be sponges for more than two miRNA, and nine were selected randomly from differentially expressed circRNAs. Two micrograms of total RNA were used before real-time quantitative PCR with DNase I (2270 a, Takara, Japan).Primers were designed to ensure that the circular template was amplified (Ting, Miao, Gang, & Ting, 2015). The sequence of primers is shown in Table 1. The expression of circRNAs was quantified on an ABI StepOnePlus system (USA) using an SYBR green master mixture (Applied Biosystems, Foster City, CA, USA). The relative expression rate (△△Ct) of each circRNA was calculated using the 2 -△△Ctmethod and expressed as log2 of the value, where Ct is the periodic threshold value of the amplified target or reference gene (Kennet & Tomas, 2001). The expression of OsActin was used as a reference for data normalization. SPSS Statistics 19.0 software was used for statistical analysis. The Student t -test was used to compare the significant difference between the control and the light treatment group using a probability level of 0.05.
2.6. Quantification of sucrose and ABA
Approximately 100 mg of developing caryopses was used to quantify the ABA and sucrose content at the grain-filling stage 5, 10, and 15 days after treatment in both the light stress and the control groups, using liquid chromatography-tandem mass spectrometry as described previously (Mikiko et al., 2009).