MYB Transcription Factor OsC1PLSr Involves the Regulation of Purple Leaf Sheath in Rice

Although several regulators associated with purple traits in rice have been identified, the genetic basis of the purple sheath remains unclear. In the present study, F2-1 and F2-2 populations were constructed using purple sheath (H93S) and green sheath (R1173 and YHSM), respectively. In order to identify QTL loci in purple sheaths, BSA analyses were performed on the two F2 populations. A crucial QTL for purple sheath was identified, tentatively named qPLSr6, and was located in the 4.61 Mb to 6.03 Mb region of chromosome 6. Combined with expression pattern analysis of candidate genes, LOC_Os06g10350 (OsC1PLSr) was suggested as a candidate gene. The homozygous mutant KO-1 and KO-2 created through CRISPR/Cas9 editing, lost their purple leaf sheath. The RT-PCR revealed that OsC1PLSr, anthocyanin synthase (ANS), diflavonol-4-reductase (DFR), flavanone-3-hydroxylase (F3H), and flavanone-3′-hydroxylase (F3′H) expression levels were dramatically down-regulated in the mutants. The yeast report system indicated that the 145–272 aa region at the C-terminal of OsC1PLSr is a positive transcriptional activation domain. The results indicated that OsC1PLSr synthesized anthocyanins by regulating the expression of ANS, DFR, F3H, and F3′H. This study provides new insights into the genetic basis of the purple sheath.

The anthocyanin accumulation induced the purple trait in distinct rice organs. OsPL6 locates on the short arm of chromosome 6, and upregulate-expression of OsPL6 led to purple accumulation in the leaf [12]. OsB2 is related to regulating anthocyanin accumulation. Ectopic expression of OsB2 generates the production of purple or black pericarp in rice [13]. The genome-wide association study and transcriptome analysis showed that OsC1, OsRb, and DFR were identified as the determinants of anthocyanin biosynthesis in rice leaves [14]. Through bulked segregant analysis with next-generation sequencing (BSA-Seq) and transcriptome sequencing (RNA-Seq) strategies, identifying a recessive gene plr4, which was regulating purple leaf in rice and was located near the 27.9-31.1 Mb interval of chromosome 4 [15]. The C-S-A gene system controls hull pigmentation in rice. C1 is a color-producing gene that produces brown when acted alone but purple when combined with A1. In addition, S1 interacts with C1 to activate the expression of the gene A1, then producing purple and brown hulls [16].
Despite multiple regulators responsible for the purple trait have been characterized, molecular detail of purple leaf sheath genes remains unclear. Previous research reports that the purple leaf sheath of rice was controlled by one or two pairs of genes through genetic analysis. Recently, researchers found Ra/Rb and OsC1, located on chromosomes 1 and 6 [17,18], which regulate anthocyanin accumulation in rice leaf sheaths. Through yeast two-hybrid analysis, OsC1 was shown to interact with Rb1/Rb2 [17]. Although OsC1 was mapped to produce rice purple leaf sheath, there has been little functional analysis of OsC1 due to a lack of genetic transformation validation.
Therefore, this study performed BSA-Seq analysis using two F2 populations; the qPLSr6 QTL was identified, and the candidate gene encoding MYB protein (OsC1 PLSr ) was functionally validated. Knockout homozygous mutants had a purple leaf sheath loss phenotype. We found that the OsC1 PLSr regulates anthocyanin synthesis through structural genes ANS, DFR, F3H, and F3 H. The results provided a theoretical basis for revealing the regulatory role of OsC1 in the purple leaf sheath.

Observation of Leaf Sheath Phenotype
According to phenotypic analysis, H93S, R1173, and YHSM have purple, green, and green leaf sheaths, respectively ( Figure 1a). Observation with a digital microscope showed that a large number of anthocyanins were distributed in the leaf sheaths of H93S, while the distribution of anthocyanins was rarely screened in R1173 and YHSM ( Figure 1b). Consistently, anthocyanin accumulation in the leaf sheaths of H93S was significantly higher than R1173 (Figure 1c). In the two F2 segregating populations, the color of the leaf sheath could be categorized into purple and green, and the expected Mendelian segregation ratios were both 3:1 (Table 1). These results indicated that the anthocyanin content is positively correlated with the leaf sheath color, implying a complete dominant gene controls the purple leaf sheath trait.  Note: χ 2 < χ 2 (0.05, 1) is considered as significant.

Identification of Candidate Regions for Controlling Leaf Sheath Color by BSA-Seq
A total of 60.72 + 60.42 Gbp of raw data were obtained and filtered into 58.47 Gbp and 57.67 Gbp of clean data from the F2-1, and F2-2 populations, respectively. Through the analysis of the data quantity and quality of seven samples (two PLSBs, two GLSBs, and three parents), it found that Q30s were all above 90.00%, and the GC content of the seven samples' clean data ranged from 44.33% to 44.77% ( Table 2). The average reading depth of seven samples ranged from 25× to 40×. According to sequencing data, 97.96-98.58% of the sequences can be successfully aligned to the reference genome R498. These results illustrated that all samples were of good sequencing quality and could be used for subsequent variation analysis. After filtering low-quality SNPs and InDels, finally, 675,512 SNPs and 140,374 InDels with credible and high quality were obtained from the F2-1 population, 466,796 SNPs and 96,738 InDels with credible and high quality from the F2-2 population (Table 3). Note: P, purple leaf sheath plants in the F2 population; G, green leaf sheath plants in the F2 population; Q30, the bases of quality value greater than or equal to 30 in the total base number; GC, the GC content of the samples; Mapped, the percentage of clean reads mapped to the reference genome R498 in the total clean reads; Ave depth, average coverage depth of samples; coverage, the ratio of the bases above given depth to the total bases in the reference genome. The ∆All-index (∆SNP-index and ∆Indel-index) and ED (Euclidean Distance) were calculated for variation analysis. The charts were plotted from the ∆All-index and ED ( Figure 2). In the F2-1 population, the interval for the candidate genes exceeding the threshold value was identified on the 3.33-8.72 Mb of chromosome 6 covering 1415 genes by SNP association analysis, while it was identified on the 3.37-8.55 Mb of chromosome 6 covering 1274 genes by InDel association analysis (Figure 2a,b,e,f; Table 4). Thus, the interval of the F2-1 population was identified on the 3.37-8.55 Mb of chromosome 6 and had 1274 genes. In the F2-2 population, the interval for the candidate gene exceeding the threshold value was identified on the 4.03-9.06 Mb of chromosome 6 and had 1257 genes by the SNP association analysis, and on the 4.61-6.03 Mb of chromosome 6 and had 415 genes by the InDel association analysis (Figure 2c,d,g,h; Table 4). Thus, the interval of the F2-2 population was identified on the 4.61-6.03 Mb of chromosome 6 and had 415 genes. Both the populations were located on the 4.61-6.03 Mb of chromosome 6. These results suggested that the candidate genes of the purple leaf sheath of H93S were located in the 1.42 Mb region of chromosome 6 from 4.61 Mb to 6.03 Mb.

Analysis of Candidate Genes
There were 415 genes with the 1.42 Mb region of chromosome 6. Among the 415 genes, 27 were expressed in rice leaf sheath by the public date expression spectrum CREP analysis (Table 5). Then, annotated by Rice Genome Annotation Project, three genes (LOC_Os06g10350, LOC_Os06g11270, and LOC_Os06g11330) were involved in the anthocyanin synthesis pathway and were considered as candidate genes (Table 5). qRTPCR results showed that the expression pattern of LOC_Os06g10350 in rice leaf sheath was "up-up-up-down" (Figure 3a) at four growth stages, which was consistent with the change of anthocyanin content in the four growth periods of H93S, and the expression level reached the peak at heading stage. While the expression patterns of LOC_Os06g11270 and LOC_Os06g11330 were "up-up-down-up" and "upup-down-down", respectively (Figure 3b,c). SO, LOC _Os06g10350 was considered to be a strong candidate for the OsC1 PLSr locus, which is co-located with OsC1 [18].

Knockout of OsC1 PLSr Results in No Purple Leaf Sheath Phenotype
To verify the function of OsC1 PLSr in H93S, we used the CRISPR/Cas9 genomic editing system to knockout OsC1 PLSr in H93S to obtain homozygous lines KO-1 and KO-2 ( Figure 4a). The KO-1 and KO-2 harbored allelic homozygous insertions or deletions of base pairs, which caused premature termination of the OsC1 PLSr protein (Figure 4b,c). Anthocyanin contents were measured in transgenic plants at tillering stage. The anthocyanin contents in the leaf sheath of mutant lines were significantly lower than H93S (Figure 5a). Besides, the relative expression of structural genes DFR, ANS, F3H, and F3 H in mutant lines was down-regulated ( Figure 5b). Further phenotypic identification showed that the leaf sheath color was green of all these mutants (Figure 4d). In addition, the photosynthesis of mutant lines and H93S was not significantly different ( Figure 5c). These results demonstrated that the two domains of OsC1 PLSr play a critical role in inducing anthocyanin biosynthesis. In addition, the anthocyanin biosynthetic genes, DFR, ANS, F3H, and F3 H, were up−regulated by the activation of Myb transcription factor gene OsC1 PLSr , guided to the accumulation of anthocyanin in H93S.

Structure Analysis, Transcription Activity Analysis, and Toxicity Detection
H93S with purple leaf sheath possessed a full-length DNA sequence of OsC1 PLSr encoding 273 amino acids. Whereas R1173 and YHSM with green leaf sheath had a 10-bp deletion in exon 3 and only 207 amino acids. There was a 3-bp deletion in exon 2 in the green leaf sheath of Nipponbare (Figure 6a). OsC1 PLSr contains two conserved domains (R2-MYB and R3-MYB), and the two deletion sites were located in the R3-MYB domain (Figure 6a). The results revealed that an aberrant R3 MYB domain of OsC1 PLSr failed to induce anthocyanin synthesis in R1173, YHSM, and Nipponbare. (c), The full-length ORF of OsC1 PLSr and three truncated mutants (OsC1 PLSr -MYB1, OsC1 PLSr -MYB2, OsC1 PLSr -C) were fused with pGBKT7, and the transformed AH109 yeasts were selected from SD/-Trp + x-α-gal medium and SD/-Trp/-His/-Ade + x-α-gal medium, respectively. The empty BD vector was used as a negative control. "SANT" means "SANT SWI3, ADA2, N-CoR, and TFIIIB" DNA-binding domains.
To further study the function of OsC1 PLSr , by the toxicity detection assay, we found the growth curve of the yeast cells containing the negative control pGBKT7 (BD) was higher than pGBKT7-OsC1 PLSr (BD-OsC1 PLSr ). This result demonstrated that the protein encoded by OsC1 PLSr had a slightly toxic effect on the growth of yeast cells (Figure 6b). In addition, we constructed various regions of the OsC1 PLSr protein by using the yeast report system. These regions included the MYB1 domain, MYB2 domain, C-terminal region, and the full-length OsC1 PLSr . The regions were fused to the GAL4 DNA-binding domain in the pGBKT7 vector, respectively. Then, they were all transformed into AH109 yeast cells. These results indicated that the amino acids 145-272 region of the C-terminal in OsC1 PLSr were responsible for transcriptional activation activity; the MYB domain did not perform transcriptional activation function (Figure 6c).

Discussion
Anthocyanins are flavonoids, which are secondary metabolites produced by plants that generally increase with the stimulation of biotic or abiotic stress [19]. Target QTL can be located rapidly and precisely using BSA in conjunction with high-throughput genotyping technologies from next-generation sequencing. In this study, we first constructed two F2 populations and then used BSA-seq technology for preliminary mapping. Both F2 populations were used to map the purple leaf sheath gene. It is possible to swiftly and precisely identify the candidate interval by association analysis of the two populations. To accurately screen potential genes. There was no extensive population mapping done. As a result, labor is saved.
R2R3-MYB transcription factors have been clarified to participate in the determination of anthocyanin synthesis and regulation in plants. The gene TT2 of Arabidopsis encoding R2R3-MYB protein regulates proanthocyanidin biosynthesis by activating the expression of ANR, while the closely related MYB PAP4 protein (AtMYB114) specifically activates UFGT expression and regulates the anthocyanin biosynthesis [20]. However, few studies have reported how MYB domains regulated the synthesis of anthocyanin in the leaf sheath of rice. In this study, we found that two MYB domains of OsC1 PLSr were essential for anthocyanin synthesis in the leaf sheath. We found that only the amino acids 145-272 region of the C-terminal in OsC1 PLSr were responsible for transcriptional activation activity, while the MYB domain has no transcriptional activation effect, which indicated that MYB domain was only responsible for directly binding DNA sequence with target gene, but not for its own transcriptional activation effect. Therefore, it is speculated that the R2R3-MYB transcription factor OsC1 PLSr in H93S may also bind to the promoters of DFR, ANS, F3H, and F3 H genes to activate the expression of these four structural genes. However, through the analysis of 2 Kb promoters upstream of DFR and F3 H genes by the Plant CARE of promoter prediction analysis website, the results showed that the promoters of DFR and F3 H genes contained light responsive MYB binding site (AACCTAA). Therefore, OsC1 PLSr may directly regulate the expression of DFR and F3 H. The regulatory effect of OsC1 PLSr on DFR and F3 H gene expression needs to be further studied.
In addition, the purple leaf sheath is the most stable, early, and visible trait for rice tissues, which is an excellent visual marker-trait [21]. There was coupled CRISPR/Cas9 unit with the OsC1 PLSr unit, and the high-level expression of Cas9 was a good indication of anthocyanin accumulation of edited plants [22]. Thus, if the purple leaf sheath OsC1 PLSr was applied to rice breeding, the efficiency of rice hybrid breeding would be greatly improved, and the yield of hybrid rice would be increased. These results provided a basis for further understanding the genetic mechanism of leaf sheath color as well as its potential utilization in a two-line hybrid rice breeding system.

Plant Materials and Genetic Population Construction
A purple leaf sheath two-line male sterile line Hang93S (H93S) was used as the female parent, and two green leaf sheath conventional lines (R1173, YHSM) were used as male parental lines. H93S hybrid with R1173 and YHSM to produce the F 1 generations, and two F 1 s self-pollinated to produce F2-1, and F2-2 populations, respectively. Both F2 populations, each of which consists of 1000 plants with no replication, were used to map the purple leaf sheath gene (OsC1 PLSr ). H93S and R1173 were used for leaf sheath histology analysis, anthocyanin content analysis, and UV-C irradiation treatment. All the plant materials were cultivated in the experimental field at South China Agricultural University (Guangzhou, China), followed by routine field management.

BSA Sequencing and Association Analysis
At the tillering stage, purple leaf sheath bulks (PLSBs) and green leaf sheath bulks (GLSBs) selected from the F2 population were constructed by pooling equivalent amounts of DNA from 50 purple and 50 green individuals. The healthy leaf sheaths of three parents (H93S, R1173, and YHSM) were collected to construct parental pools. The genomic DNA was extracted by CTAB for BSA sequencing. All DNA of two PLSBs, two GLSBs, and three parental pools were constructed into a library according to the manufacturer's recommended protocol. BSA sequencing was performed using whole genome sequencing on HiSeq 2000 platform (Illumina) by Biomarker Technologies Co., Ltd. (Beijing, China). Since all the parents (H93S, YHSM, and R1173) are indica lines, R498 was therefore selected as the reference genome (http://mbkbase.org/R498/, accessed on 10 June 2021). Sequencing data were subjected to quality control and mapped to R498. The raw reads of the fast format were first processed through a series of quality control procedures to remove low-quality paired reads. Before association analysis, SNPs and InDels must be filtered. Firstly, SNPs (InDels) with multiple genotypes were screened; Secondly, SNPs (InDels) with read support of less than 4 were filtered out. Thirdly, mixing pool genotype SNPs (InDels) were screened from the recessive mixing pool genotype SNPs (InDels) of non-recessive parents. Finally, high-quality and credible SNPs and InDels were obtained for subsequent analysis. Using All-index ( SNP-index and Indel-index) and ED (Euclidean Distance), the region linked to the target gene was mapped by comprehensive analysis.

Screening of Candidate Genes
The association regions obtained from two F2 populations were comprehensively analyzed, and the intersection of the association regions located by both populations was selected as the candidate region of the purple leaf sheath gene. The expression profile of the public database CREP (http://crep.ncpgr.cn/, accessed on 5 December 2022) was used to screen candidate genes. The Rice Genome Annotation Project database (RGAP) annotates the putative functions of candidate genes. In order to confirm the final candidate genes, qRT-PCR analysis was used to verify their expression at the heading stage.

Histological, Anthocyanin Content, and Genetic Analysis
At the heading stage, longitudinal sections of rice leaf sheath were obtained by freehand sectioning, and the distribution of anthocyanins in leaf sheath cells was observed by using a digital microscope (OLYMPUS CX31). Anthocyanins were extracted from 0.5 g of finely ground leaf sheath tissue as described by Nakatskaa et al. [23]. Anthocyanin content (u.g −1 ) = OD/w, wherein "w" is the fresh weight of the leaf sheath. The leaf sheath color separation ratio in two F2 populations was investigated and calculated by the Chi-square test.

Sequence Analysis and Structure Analysis of OsC1 PLSr
Specific primers linked to the OsC1 PLSr gene were used to amplify gene sequences separately from the three parental lines (H93S, R1173, and YHSM) via using Phanta ® Max Super-Fidelity DNA Polymerase. Sequence alignment was performed with SnapGene. The CDS of OsC1 PLSr from the purple and green leaf sheath parental lines were cloned and subjected to alignment by SnapGene. Structure analysis of the OsC1 PLSr protein was performed by using the SMART database (www.smart.embl-heidelberg.de, accessed on 5 December 2022).

CRISPR/Cas9 Generated the OsC1 PLSr Mutant Plants
Based on the conversed structure of OsC1 PLSr protein, an independent site in the exon 3 was designed as CRISPR/Cas9 spliced sites by CRISPR-P [2]. The fragment containing the target was recombined with the pRGEB32 vector, which was digested by BsaI. Through the use of Agrobacterium tumefaciens, the constructs were transformed into H93S [24].

Transcription Activity Assay and Toxicity Detection in Yeast Report System
In order to determine the OsC1 PLSr transcriptional activity domain, various regions of the OsC1 PLSr protein coding sequence, including the R2-MYB domain (MYB1, 1-65aa), R3-MYB domain (MYB2, 66-144aa), C-terminal region (C,145-272aa) and the full-length OsC1 PLSr were amplified and inserted into EcoRI/BamHI-digested pGBKT7 vector. The recombination vectors and the negative control pGBKT7 (empty vector) were transformed into yeast strain AH109. The positive transformants were grown on SD/-Trp solid medium and SD/-Trp-His-Ade solid medium for three days at 30 • C. χ-α-gal was used to determine the transcription activation activity of the OsC1 PLSr protein regions in the yeast expression system. Besides, in order to determine whether the OsC1 PLSr protein produced a toxic effect on yeast growth, the yeast strains transformed with the recombination vectors pGBKT7-OsC1 PLSr (BD-OsC1 PLSr ) and others with the negative control pGBKT7 (empty vector, BD) were cultured with SD/-Trp liquid medium, respectively, which were shaken by shaker incubator (TENSUC). The OD 600 values were measured at five-time points (5 h, 10 h, 15 h, 20 h, and 25 h), and the growth curve was drawn.

RNA Extraction and Quantitative qRT-PCR
Total RNA was extracted from the leaf sheath of H93S across the entire growth period; first, cDNA was synthesized by using the GoScript™ (Fisher Scientific, Waltham, MA, USA) Reverse Transcription System (Promega). The qRT-PCR verification was performed with three techniques repeated. The AceQ ® qPCR SYBR ® Green Master Mix was used for the reaction on a StepOnePlus ® Real-Time PCR System 272007300 (applied biosystems ® life technologies TM , Carlsbad, CA, USA). Differences in gene expression were calculated using the 2 −∆∆Ct method. The actin gene was used as an internal reference control. The expression level of eight structural genes (ANS, DFR, F3H, F3 H, F3 5 H, PAL, CHS, and CHI) and candidate genes of the leaf sheath of H93S were analyzed by qRT-PCR.

Conclusions
The purple leaf sheath is a quality trait that can be recognized by the naked eye at the seedling stage. The application of purple leaf sheaths in breeding is of great significance. In this study, two F2 populations from the crossing of purple leaf sheath with green leaf sheath were constructed, BSA sequencing was performed, and candidate region qPLSr6 was located. OsC1 PLSr was found to be a potential candidate gene. The constructed knockout mutant showed a purple sheath deletion phenotype, indicating that OsC1 PLSr is closely related to the purple sheath phenotype. The results of the yeast report system showed that OsC1 PLSr had a transcriptional activation function. The discovery of this result laid the foundation for revealing the molecular regulation mechanism of the purple leaf sheath.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding authors.