Transcriptome analysis of knockout mutants of rice seed dormancy gene OsVP1 and Sdr4

OsVP1 and Sdr4 play an important role in regulating seed dormancy that involved in multiple metabolism and regulatory pathways. Seed dormancy and germination are critical agricultural traits influencing rice grain yield. Although there are some genes have identified previously, the comprehensive understanding based on transcriptome is still deficient. In this study, we generated mutants of two representative regulators of seed germination, Oryza sativa Viviparous1 (OsVP1) and Seed dormancy 4 (Sdr4), by CRISPR/Cas9 approach and named them cr-osvp1 and cr-sdr4. The weakened dormancy of mutants indicated that the functions of OsVP1 and Sdr4 are required for normal early seed dormancy. There were 4157 and 8285 differentially expressed genes (DEGs) were identified in cr-osvp1 vs. NIP and cr-sdr4 vs. NIP groups, respectively, with a large number of overlapped DEGs between two groups. The gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of common DEGs in two groups showed that genes related to carbohydrate metabolic, nucleoside metabolic, amylase activity and plant hormone signal transduction were involved in the dormancy regulation. These results suggest that OsVP1 and Sdr4 play an important role in regulating seed dormancy by multiple metabolism and regulatory pathways. The systematic analysis of the transcriptional level changes provides theoretical basis for the research of seed dormancy and germination in rice.


Introduction
Seed dormancy is a biological characteristic of plants adapting to the external environment, which has important ecological significance for the survival and development of plants.
In agricultural production, the germination and dormancy of seeds directly affect crop yields (El-Maarouf-Bouteau et al. 2015). Strong seed dormancy causes a decline in the germination rate, which will lead to a decline in the number of effective crop individuals, thus resulting in decreased yield (Shuai et al. 2016). But the weak dormancy leads to pre-harvest sprouting (PHS) in some areas where encounter high temperature and humidity during the harvest season (Hisano et al. 2022). Due to the long rainy harvest season in southern China, about 6% of the conventional rice and 20% of the hybrid rice planting area were damaged by PHS (Doebley et al. 2006). How to achieve the balance between dormancy and germination is the key problem in rice breeding (Hisano et al. 2022 In addition to the influence of external factors such as temperature and humidity, seed dormancy is also regulated by many genes and metabolites. Plant hormones are involved in the regulation of seed dormancy and development, abscisic acid (ABA) and gibberellin (GA) play antagonistic roles in the control of seed dormancy and germination by altering their synthesis and signaling. (Graeber et al. 2012). Mutations in many ABA synthesis and signaling genes cause weak dormancy (Fang et al. 2008). ABA-binding receptors REGULATORY COMPONENT OF ABSCISIC ACID RECEPTOR (RCAR)/PYRABACTIN RESIST-ANCE (PYR)/PYR1-LIKE (PYL) interact with PROTEIN PHOSPHATASE 2C (PP2C) and cause the formation of the ABA-RCAR/PYL/PYR-PP2C complex (Cutler et al. 2010), which helps the release of SNF1-related kinases 2 (SnRK2s) to phosphorylate the downstream transcription factors ABSCISIC ACID INSENSITIVE 3 (ABI3), ABI4, and ABI5 leading to activated ABA responses (Kobayashi et al., 2005;Furihata et al., 2006;Fujii et al., 2007;Fujii and Zhu, 2009;Nakashima et al., 2009). When GA is present in large quantities, GA binds to the response factor GID1 to form a complex, which promotes the ubiquitination and degradation of DELLA protein mediated by the F-box ubiquitin ligase SLEEPY1, which promotes the GA signaling and seed germination (Murase et al. 2008). Other plant hormones, such as auxin (AUX), ethylene (ETH), jasmonic acid (JA) and cytokinin (CK), coordinate with ABA and GA to regulate the dormancy and germination . Generally, AUX can promote seed dormancy by stimulating ABA signal transduction, and it also plays an important regulatory role in germination process (Liu et al. 2013). ETH counteracts ABA effects by regulating ABA metabolism and signaling pathway. Moreover, mutants insensitive to ethylene are more sensitive to ABA, and lead to enhanced dormancy in Arabidopsis (Corbineau et al. 2014). JA coordinate with ABA to delay seed germination (Varshney and Majee 2021). CK promotes seed germination by promoting GA-induced alpha-amylase activity in wheat (Eastwood et al. 1969).
The soluble sugar content can be used as an index for the selection of PHS resistance in rice. Low concentration of sugar can relieve the inhibitory effect of ABA and promote seed germination, while high concentration of sugar can activate ABA signal and inhibit seed germination (Garciarrubio et al. 1997;Finkelstein et al. 2000;Cheng et al. 2002;Rolland et al. 2006). Recent research found that mutations in PHS8 increased endosperm sugar content and induce preharvest sprouting in rice (Du et al. 2018). α-amylase hydrolyzes stored starch to produce dextrins, oligosaccharides and monosaccharides, which provide nutrients for seed germination. (Damaris et al. 2019). Proteomics analysis revealed that the main metabolic activities, such as repair of cell integrity and DNA damage, redox and pyruvate metabolism, are involved in the process of absorbing water and swelling (He et al. 2011;Yang et al. 2007;He et al. 2013).
Previous studies have shown that OsVP1 and Sdr4 are two key genes related to seed dormancy. OsVP1 encodes a transcription factor homologous to Zea mays Vp1 and Arabidopsis thaliana ABI3, and contains A1, B1, B2 and B3 domains (Hattori et al. 1994;Giraudat et al. 1992). OsVP1 was identified as an interaction factor of TRAB1, which binds to ABA-responsive elements (ABREs) to regulate ABAregulated transcription (Hobo et al. 1999). Sdr4, a major dormant QTL, can influence the seed dormancy difference between Nipponbare and Kasalath strongly. Sdr4 has two haplotypes Sdr4-k and Sdr4-n, of which Sdr4-k conferred deep dormancy and reduced pre-harvest sprouting of seeds (Sugimoto et al. 2010). In previous research, we proved that OsVP1 bounds to promoter of Sdr4 by the specific motif CAC CTG directly and regulates Sdr4 gene expression positively (Chen et al. 2021). Although the regulatory mechanism between these two genes is relatively clear, it is necessary to further explore how they regulate material changes and signal transduction in seeds, and then analyze the molecular mechanisms of seed dormancy and germination. In this study, we sequenced the transcriptome of cr-vp1, cr-sdr4 and their wild-type seeds, analyzed the data to explore the expression changes of genes, and revealed the regulatory mechanism of the two genes in seed germination. This work will provide important information for the further study of seed dormancy and germination in rice.

Plant materials
All rice lines used in this study are in NIP (Oryza sativa ssp. japonica) background. The stable knockout mutants cr-osvp1 and cr-sdr4 were obtained by CRISPR/Cas9 approach. For the experiment, rice plants were grown in the experimental station of China National Rice Research Institute (2020). Fertilizer and water management followed standard field production. In order to ensure the consistency of the growth period of experimental materials, the bloom time was marked in the seed hulls of these materials in the flowering period.

Seed germination assays
In the late stage of seed maturity, it encountered continuous rainy weather in Hangzhou (2020Hangzhou ( .09,10-2020. The air humidity is greater than 90%, and the average temperature is around 25 °C. Taking this advantage condition, we observed the pre-harvest sprouting phenotype of cr-osvp1, cr-sdr4 and NIP, and analyzed the germination rate on the 7th day, with three replications for each material.

RNA extraction, cDNA library preparation and sequencing
The NIP, cr-osvp1 and cr-sdr4 seeds that 25 days after flowering were taken, quickly frozen and stored at -80 °C for further sequencing with three replications for each material. RNA extraction, cDNA library preparation and sequencing were commissioned to Beijing Nuohe Zhiyuan Technology Co., Ltd. After extracted from each sample using TRIzol kit, the total RNAs were strictly controlled according to the following four aspects, including agarose gel electrophoresis, NanoPhotometer spectrophotometer, Qubit 2.0 Fluorometer and Agilent 2100 bioanalyzer. The library was constructed in accordance with NEB's normal library construction method, and the insert size of the library was detected by the Agilent 2100 bioanalyzer to ensure the quality of the library. The library was pooled according to the effective concentration and target offline data volume for Illumina sequencing. The raw data obtained from sequencing contained a small number of reads with sequencing adapters or low sequencing quality. In order to ensure the quality and reliability of data analysis, it is necessary to remove the reads with adapter contamination, containing N (N means base information cannot be determined) and low-quality (Q phred ≤ 20 bases accounted for more than 50% of the entire read length).

Function annotation and enrichment analysis of differential expression genes
Using DESeq2 software, with padj < 0.05 as the threshold, the DEGs between the mutant and NIP were screened out. Cluster Profiler was used to perform enrichment analysis of the biological process (BP), molecular function (MF) and cell component (CC) of the DEGs, and perform GO gene function annotations to obtain functional annotations, biological functions, and related metabolic pathways of all DEGs, and P value ≤ 0.05 was set as the threshold to find significantly enriched molecular entries. At the same time, the KEGG database was used to locate and annotate the selected DEGs, and P value ≤ 0.05 was set as the threshold to find significantly enriched pathways for further interpret the function of genes.

Quantitative real-time PCR analysis
Quantitative real-time PCR (qRT-PCR) was performed using a Light Cycler 480 instrument (Roche) with the SYBR Green Real-time PCR Master Mix (Toyobo) in a 20 μL reaction volume with three biological replications. Twenty genes were randomly chosen and primers for qRT-PCR are listed in Table S7. The ubiquitin gene (Os03g0234200) was used as an internal control, and relative expression levels were calculated by the 2 −ΔΔCT method (Schmittgen and Livak 2008).

Germination analysis of cr-osvp1 and cr-sdr4 mutants
In previous studies, we created the mutation of two key germination regulation genes OsVP1 and Sdr4 by CRISPR/ Cas9 approach, and obtained their mutants named cr-osvp1 and cr-sdr4 (Chen et al. 2021). The vector of pVK005, that includes LB, Hyg, 35S, rU6, gRNA, mpCas9, Ubi and RB elements, was used in the CRISPR/Cas9 system (Fig. 1a). Sanger sequencing revealed that 1-bp insertion on the second exon of OsVP1, which lead to a frameshift and premature termination of the resulting protein, and Sdr4 also has 1-bp insertion on the only exon, which lead to a frameshift and termination failure of protein translation (Fig. 1b, c, Fig.  S1a, b). In order to confirmed that the mutants were reliable, seed germination was firstly analyzed. As shown in Fig. 2a, both mutants showed severe PHS after 7 consecutive days of high temperature and humidity (air humidity greater than 90%, average temperature is around 25 °C) after seed maturation compared with NIP. The mean seed germination rates of cr-osvp1 (77.7%) and cr-sdr4 (80.3%) were much higher than that of the NIP (3.0%)on the 7th rainy day (Fig. 2b).

Quality analysis of sequencing
In order to understand the molecular mechanism of OsVP1 and Sdr4 in regulating seed germination, RNA sequencing was performed to investigate gene expression using seeds of 25 days after flowering. There were 58,162,153, 57,420,762 and 55,070,625 raw reads were generated from wild-type NIP, cr-osvp1 and cr-sdr4 transcriptomes, and we obtained 57,652,107, 56,763,467 and 54,576,483 clean reads after removing the N base sequence, linker sequence and lowquality reads. The percentages of Q20 reached 97.58%, 97.93% and 98.02%, and the percentages of Q30 reached 93.35%, 94.07% and 94.25%, and the average percentages of GC were 55.81%, 54.45% and 52.26%, respectively (Table 1). The above data shows that the quality of this sequencing is good.

Differentially expressed genes analysis
To investigate the changes of gene expression in the PHS process, the DEGs of two groups were compared separately. In cr-osvp1 vs. NIP, there were 4157 genes differentially expressed, of which 2239 genes were up-regulated and 1918 genes were down-regulated; in cr-sdr4 vs. NIP, a total of 8252 DEGs were identified, and 4622 genes were up-regulated and 3663 genes were down-regulated (Fig. 3a). The number of differentially expressed genes in cr-sdr4 vs. NIP was significantly more than those in cr-osvp1 vs. NIP, and a total of 3030 genes were differentially expressed in the two groups (Fig. 3b). Consistent with previously studies that OsVP1 and Sdr4 regulated seed germination by the same pathway (Sugimoto et al. 2010;Chen et al. 2021), the DEGs in two groups have a large overlap, for example OsEm1, OsMFT2 and OsRab16A were regulated in the consistent direction among two mutants (Fig. 3c-e). OsEm1 encodes one of the late-embryogenesis abundant proteins, which was characterized to be regulated by abscisic acid (ABA) and the transcriptional activator OsVP1 at transcription level. OsMFT2 interacts with OsbZIP23/66/72 and participates in ABA signaling mediated regulation of rice seed germination (Song et al. 2020). OsRab16A contains an ABA-responsive element, and OsbZIP23/66/72 binds to its promoter to Fig. 2 Phenotypic characterization of NIP, cr-osvp1 and cr-sdr4. a PHS phenotypes of physiologically mature seeds of NIP, cr-osvp1 and cr-sdr4 after seven consecutive rainy days. Bar, 3 cm. b Comparison of seed germination rates of NIP, cr-osvp1 and cr-sdr4 calculated from a on the 7th day. **P < 0.01 by t test

Functional classification by gene ontology (GO)
In order to understand the main functional characteristics of DEGs in the two mutants, the DEGs of cr-osvp1 vs. NIP and cr-sdr4 vs. NIP were analyzed by GO significant enrichment analysis. A total of 90 GO molecular terms were enriched in cr-osvp1 vs. NIP, among them biological processes, cell components, and molecular functions were enriched to 45, 11 and 34 terms, respectively (Supplementary Table S1, 2, 3); in cr-sdr4 vs. NIP enriched a total of 123 GO molecular terms, with 85, 8 and 30 terms, respectively (Supplementary Table S4, 5, 6). The same DEGs in cr-osvp1 vs. NIP and cr-sdr4 vs. NIP groups were enriched into GO terms containing 34 biological process and 5 cellular component terms as well as 6 molecular function (Fig. 4a-c). The common enrichment of biological processes are carbohydrate metabolic process, DNA integration, nuclearide metabolic process, ribosome biogenesis and other terms. According to the statistics, there were 5 cellular component terms including intracellular ribonucleoprotein complex, ribonucleoprotein complex, ribosome, non-membrane-bounded organelle and intracellular non-membrane-bounded organelle. After analyzing the molecular function, 6 terms including amylase activity, cellulose synthase activity and glucosyltransferase activity were enriched.

Enrichment analysis of kyoto encyclopedia of genes and genomes (KEGG) pathway
In order to explore the metabolic pathway changes in the two mutants of rice seeds, the KEGG database was used to perform functional classification and pathway annotation analysis on DEGs. There were 15 KEGG pathways significantly enriched in cr-osvp1 vs. NIP, and 10 KEGG pathways significantly enriched in cr-sdr4 vs. NIP (Fig. 5a, b). Among them, ribosome biogenesis in eukaryotes, carbon fixation in photosynthetic organisms, carbon metabolism, starch and sucrose metabolism, taurine and hypotaurine metabolism, phosphatidylinositol signaling system and inositol

Analysis of gene expression patterns in plant hormone signal transduction pathway
A total of 80 DEGs related to auxin, cytokinin, gibberellin, abscisic acid, salicylic acid, ethylene and jasmonic acid were enriched in the plant hormone signal transduction pathway. The cr-osvp1 vs. NIP contained 42 DEGs, of which 26 were up-regulated and 16 down-regulated. The cr-sdr4 vs. NIP contained 67 DEGs, of which 39 were up-regulated and 28 down-regulated. Interestingly, a large number of genes had similar expression trend in the two mutants. Compared with wild-type NIP, auxin transport-related AUX1 genes, auxin response factors (ARFs), two Gretchen Hagen 3 (GH3)related genes with indole acetic acid amino acid synthase function and the small auxin up RNA genes (SAUR ) were down-regulated in two mutants. However, seven negative regulators of auxin signal transduction ( Aux/IAA) were upregulated. These results showed that the transcription of genes that play negative regulatory roles in the auxin signal transduction pathway were more active in the two mutants.
The expression of a JA receptor was down-regulated, and the transcription levels of genes encoding jasmonate ZIMdomain protein (JAZ) and jasmonyl-L-isoleucine synthase (JAR) were up-regulated, only GH3.3 (Os01g0221100) was down-regulated in cr-osvp1. We also performed the statistical analysis of the expression level of ABA receptors (PYR/ PYL), SnRK2, PP2C and ABA-responsive element binding factor (ABF), but all of them did not show a certain expression pattern. These results indicated that ABA signal transduction pathway may have a more complex function in participating in seed germination regulation. The expression of all of the 5 genes related to the CK signaling pathway were increased to varying degrees in the two mutants, it may be closely related to the rapid germination of seeds. GA signal pathway was only enriched in one gene SLRL1 (Os01g0646300), which is an inhibitor of GA signalling. The expression level of SLRL1 was down-regulated in the two mutants, indicating that the GA signal can be transmitted downstream more smoothly. In conclusion, seed germination is a complex process, and is involved in multiple hormone signal pathways.

Verification with qRT-PCR
In order to verify the accuracy of RNA-seq data, 20 different types of DEGs were randomly selected for qRT-PCR analysis, such as glutelin (Os08g0012790), Embryonic abundant protein 1 (Os05g0349800), AWPM-19-like membrane family protein (Os10g0464300), AP2 domain containing protein (Os01g0131600) and expressed protein genes. The results of qRT-PCR showed that all the genes were consistent with RNA-seq results except for one cold protein gene (Os03g0794900) (Fig. 7).

Discussion
Rice seed dormancy is a very complicated phenomenon in the life cycle, which is affected by multi-gene regulations and environmental factors, so the research on the function and regulatory network of seed dormancy genes is of great significance. At present, CRISPR/Cas9 gene editing and transcriptome technology have been widely used in the research of rice gene function analysis. In this study, two dormancy genes OsVP1 and Sdr4 were edited by CRISPR/ Cas9 technology, and two stable mutants were obtained. Compared with the wild-type, the seeds of the two mutants had severely reduced dormancy and showed serious PHS. Seed maturation is usually accompanied by material accumulation, decreased water content, gradual cessation of internal metabolic activity, and seed dormancy. However, seed germination requires a lot of material and energy, so genes expression were very active. In this study, the Fig. 4 Analysis of GO significant molecular terms between cr-osvp1 vs. NIP and cr-sdr4 vs. NIP. a Biological process. b Cellular component. c Molecular function. Red means that GO term only exists in cr-osvp1 vs. NIP. Green means that GO term only exists in cr-sdr4 vs. NIP. Blue means that GO term both exists in cr-osvp1 vs. NIP and cr-sdr4 vs. NIP transcriptome analysis showed that cr-osvp1 and cr-sdr4 mutants had a lot of DEGs compared to NIP, and the number of genes with up-regulated expression was more than that of down-regulated. The amount of DEGs in cr-sdr4 vs. NIP was significantly more than cr-osvp1 vs. NIP group, which is consistent with the fact that cr-sdr4 have faster germination rate than cr-osvp1, proving the observation in the previous study (Chen et al. 2021). 72.9% of DEGs in cr-osvp1 vs. NIP was also differentially expressed compared with that in cr-sdr4 vs. NIP, indicating that OsVP1 and Sdr4 share similar regulatory networks when they participate in seed germination regulation, and which also consistent with the result that OsVP1 regulates the expression of Sdr4. GO enrichment found that biological processes such as carbohydrate metabolic process, polysaccharide catabolic, DNA integration process, and nucleic acid metabolic process were significantly enriched in both projects, indicating that DNA replication and material metabolism have undergone drastic changes, and these substances can promote seed germination. Cellular components such as ribonucleoprotein complex, ribosome, and non-membrane-bounded organelle are enriched. Seed germination involves cell division and differentiation, and these components are essential for seed germination. The genes participating in amylase activity regulation were significantly enriched. Amylase is the key enzyme that degrades starch in the early stage of seed germination (Damaris et al. 2019). The starch in the endosperm is converted into sugar under the action of amylase, and then transported to the growth part of the seed embryo (He et al. 2015). These results of KEGG enrichment showed that sugar, starch and sucrose metabolism pathways are very active in two groups, and the soluble sugars was metabolized from carbohydrates are substrates of respiration, providing sufficient energy for seed germination.
Plant hormone signal transduction pathways play important role in seed dormancy and germination. Among them, ABA, AUX and JA inhibit seed germination, but GA and CK play opposite effects on seed germination through specific regulation mechanisms (Holdsworth et al. 2008). In this study, genes related to hormone signal transduction were significantly enriched in cr-osvp1 vs. NIP. Although, in cr-sdr4 vs. NIP, enriched genes related to hormone signal transduction was few, the DEGs of the two groups were highly repetitive, indicating that the two gene-regulated phytohormone signal transduction pathways are consistent. Several genes related to seed dormancy in ABA signaling pathway were identified, and the proteins encoded by these genes including transcription factors, phosphatases and protein kinases (Née et al. 2017;Merlot et al. 2001;Vlad et al. 2009;Xiang et al. 2014). In this study, four important gene families, ABF, PYR/PYL, PP2C and SnRK2, involved in ABA signaling pathway were partially up-regulated and down-regulated, indicating that mutation of OsVP1 and Sdr4 have effects on ABA signal transduction, leading to the expression of related genes changed. Auxin receptor TIR1/ AFB, transcriptional repressor AUX/IAA and transcription factor ARF are important proteins in auxin signal transduction pathway (Salehin et al. 2015;Winkler et al. 2017;Powers et al. 2019). In Arabidopsis, the down-regulated of the ARF genes will lead to reduced seed dormancy and accelerated germination rate (Liu et al. 2013). In this study, the expression level of AUX1 decreased in two mutants, but the other AUX/IAA genes expression were up-regulated, and five ARF genes expression were down-regulated, and the expression of the enriched early response genes decreased, implying that auxin signal transduction may be inhibited. The JA receptor CORONATINE INSENSITIVE 1 (COI1) plays a key role in JA signaling by interaction with JAZs, leading to the ubiquitin-dependent degradation of JAZ proteins through the 26S proteasome pathway (Pan et al. 2020;Thines et al. 2007). JAZ proteins interact with ABI5 and suppress its transcriptional activity, resulting in seed germination (Ju et al. 2019). In this study, the expression of COI1 was down-regulated and JAZs were up-regulated, indicating that JA signaling was suppressed. GA can stimulate the induction of hydrolase, weaken barrier tissues (endosperm or seed coat), and stimulate embryo growth to promote germination (Itoh et al. 2005). There is one negative regulator SLRL1 (Os01g0646300) was identified in the GA signal transduction pathway, and the expression of this gene was significantly reduced, suggesting that GA signaling may be activated. The cytokinin signal transduction pathway is also called two-component signaling system, which mainly includes histidine kinase (HK), histidine phosphotransfer protein (HP) and response regulator (RR) (Hwang et al. 2012;Kieber et al. 2018;Zubo et al. 2020). In this study, the expression of ARRs and AHP were both up-regulated, indicating that the cytokinin signaling was activated to promote germination.
The phenomenon of rice seed dormancy is the result of the interaction of multiple metabolic pathways and signaling pathways, which require the participation of multiple genes, and the function of a single gene cannot be completed independently. Sdr4 has three haplotypes, Hap2 has strongest dormancy, and Hap3 has stronger dormancy than Hap1 (Magwa et al. 2016). Even though Hap2 conferred strong dormancy, not all varieties with Hap2 have lower germination rate than varieties with Hap1. The analysis of the changes in the genome-wide expression profiles of the two mutants and WT will help to understand the interaction of various metabolic pathways during the dormancy process, and provide a theoretical basis for the screening of rice dormancy major genes. At the same time, it provides more possibilities for PHS resistant breeding. Fig. 7 Verification of expression of 20 selected DEGs in rice. Relative expression of 20 genes were tested using qRT-PCR. Three independent replicates were performed for each gene. Fold changes of these genes detected in RNA-Seq are also shown for reference

Conclusions
OsVP1 and Sdr4 are key regulators in seed dormancy, two severe PHS mutants cr-osvp1 and cr-sdr4 were obtained by CRISPR/Cas9. OsVP1 and Sdr4 are responsible for the same regulatory pathway, because transcriptome data show that they have a large number of identical DEGs. GO and KEGG enrichment analysis revealed a variety of biosynthetic and metabolic pathways, indicating that seed dormancy and germination is a complex biological process in which nucleotide and glucose metabolic pathways play key roles. The signal transduction pathway of plant hormones has undergone drastic changes, and the signals of various hormones influence each other and jointly regulate seed germination. The results provided important information for perfecting the regulation mechanism of seed dormancy.