Improving Rice Leaf Shape Using CRISPR/Cas9-Mediated Genome Editing of SRL1 and Characterizing Its Regulatory Network Involved in Leaf Rolling through Transcriptome Analysis

Leaf rolling is a crucial agronomic trait to consider in rice (Oryza sativa L.) breeding as it keeps the leaves upright, reducing interleaf shading and improving photosynthetic efficiency. The SEMI-ROLLED LEAF 1 (SRL1) gene plays a key role in regulating leaf rolling, as it encodes a glycosylphosphatidylinositol-anchored protein located on the plasma membrane. In this study, we used CRISPR/Cas9 to target the second and third exons of the SRL1 gene in the indica rice line GXU103, which resulted in the generation of 14 T0 transgenic plants with a double-target mutation rate of 21.4%. After screening 120 T1 generation plants, we identified 26 T-DNA-free homozygous double-target mutation plants. We designated the resulting SRL1 homozygous double-target knockout as srl1-103. This line exhibited defects in leaf development, leaf rolling in the mature upright leaves, and a compact nature of the fully grown plants. Compared with the wild type (WT), the T2 generation of srl1-103 varied in two key aspects: the width of flag leaf (12.6% reduction compared with WT) and the leaf rolling index (48.77% increase compared with WT). In order to gain a deeper understanding of the involvement of SRL1 in the regulatory network associated with rice leaf development, we performed a transcriptome analysis for the T2 generation of srl1-103. A comparison of srl1-103 with WT revealed 459 differentially expressed genes (DEGs), including 388 upregulated genes and 71 downregulated genes. In terms of the function of the DEGs, there seemed to be a significant enrichment of genes associated with cell wall synthesis (LOC_Os08g01670, LOC_Os05g46510, LOC_Os04g51450, LOC_Os10g28080, LOC_Os04g39814, LOC_Os01g71474, LOC_Os01g71350, and LOC_Os11g47600) and vacuole-related genes (LOC_Os09g23300), which may partially explain the increased leaf rolling in srl1-103. Furthermore, the significant downregulation of BAHD acyltransferase-like protein gene (LOC_Os08g44840) could be the main reason for the decreased leaf angle and the compact nature of the mutant plants. In summary, this study successfully elucidated the gene regulatory network in which SRL1 participates, providing theoretical support for targeting this gene in rice breeding programs to promote variety improvement.


Introduction
Rice, as an essential food crop, serves as a fundamental source of sustenance for millions of people worldwide [1]. The yield of rice is influenced by various factors, with the photosynthetic efficiency of leaves being one of the most significant factors [2]. Moderate leaf rolling enables the leaves to remain upright, reducing interleaf shadow, improving photosynthetic efficiency, and increasing rice yield. Therefore, moderate leaf rolling is recognized as a crucial agronomic characteristic in rice breeding [3][4][5]. In recent years, several genes responsible for controlling leaf rolling have been identified, providing a promising foundation for future breeding efforts to induce rice leaf rolling [6].
Although leaf rolling is a straightforward and intuitive agronomic trait, its regulatory mechanisms are complex. Mutations in genes related to leaf development and the impacts of diverse environmental conditions can both result in leaf rolling. Leaf development is an intricate process involving the formation of adaxial-abaxial axis, bulliform cells, sclerenchymatous cells, cuticle, and epidermal cells [7]. Mutations in genes involved in these processes can lead to abnormal leaf development and the occurrence of the rolled-leaf phenotype. Currently, more than 30 genes affecting leaf rolling have been cloned, and the majority of these genes operate by regulating the number, size, or arrangement of bulliform cells. SEMI-ROLLED LEAF 1 (SRL1) is a gene involved in the regulation of leaf rolling and encodes a putative glycosylphosphatidylinositol-anchored protein located on the plasma membrane. It is expressed in various tissues of rice and acts as a negative regulator of genes encoding vesicular H + -ATPase subunits and H + -pyrophosphatase, inhibiting the formation of bulliform cells and regulating leaf rolling [8]. CURLED LEAF AND DWARF 1 (CLD1) is allelic to SRL1, which is closely associated with cell wall formation, epidermal integrity, and the maintenance of water homeostasis. cld1 mutants exhibited a considerable decrease in the content of cellulose and lignin within the secondary cell wall of the leaf when compared with the wild type (WT). The loss of CLD1 function could lead to defective leaf epidermal development, reduced water retention capacity in leaves, and eventually rolled leaves [9]. RICE OUTER CELL SPECIFIC 8 (ROC8), a homologous structural domain leucine zipper class IV gene, is involved in the negative regulation of bulliform cell size. The overexpression of ROC8 reduced the number and size of bulliform cells, resulting in leaf adaxial rolling. Conversely, the knockdown of ROC8 increased the number and size of bulliform cells, leading to leaf abaxial rolling [10]. ROC5 and ROC8 exhibit similar biological functions in regulating leaf rolling and both have additive effects in function, and can form ROC8-ROC5 heterodimers to jointly regulate leaf rolling [11]. Additionally, ROLLED AND ERECT LEAF 1 (REL1) [12], ABAXIALLY CURLED LEAF 2 (ACL2) [13], ZINC FINGER HOMEODOMAIN 1 (OsZHD1) [14], and REL2 [15] have also been found to affect leaf rolling by regulating the number, size, or arrangement of bulliform cells.
In addition to leaf rolling caused by changes in bulliform cells, variations in cellulose and lignin content can also contribute to leaf rolling. NARROW AND ROLLED LEAF 2 (NRL2) [16], OsMYB103L [17], and ROLLING-LEAF 14 (RL14) [18] are involved in regulating the synthesis of cellulose and lignin, thereby affecting the morphology and mechanical strength of leaves. Moreover, changes in cellulose and lignin content also play a crucial role in the synthesis of secondary cell walls, which is essential for water transport in leaves. Alterations in cellulose and lignin synthesis can also lead to abnormal bulliform cell morphology that ultimately causes leaf rolling [18]. In addition, SRL10 [19], SHALLOT-LIKE 1 (SLL1) [20], CONSTITUTIVELY WILTED 1 (COW1) [21], ADAXIALIZED LEAF 1 (ADL1) [22], CURLY FLAG LEAF 1 (CFL1) [23], NRL1 [24] and ROLLED LEAF 17 (RL17) [25] also regulate the leaf rolling process through several different regulatory networks. In conclusion, the regulatory mechanism of leaf rolling is complex and influenced by several factors.
The CRISPR/Cas system is widely recognized as a mainstream gene editing tool for plant improvement. The guide RNAs effectively guide the Cas9 enzyme to create accurate cuts in the DNA double helix, which results in double-stranded breaks. These breaks can cause targeted gene mutations, such as gene knockout, when the DNA is repaired. This process ultimately achieves the desired modification of the genomic DNA [26]. With the continuous improvement of CRISPR/Cas9 editing technology, it is now possible to simultaneously target multiple sites, providing new methods for studying traits controlled by multiple genes and gene families [27]. The precise genome editing enabled by CRISPR/Cas9 has the potential to greatly enhance the quality and yield of agricultural production, while also facilitating the creation of new and improved varieties of rice with enhanced traits [28][29][30][31][32]. In recent years, the beneficial integration of CRISPR/Cas9 and high-throughput transcriptome sequencing technology (RNA-seq) has opened up new possibilities for investigating gene function [33,34]. Simultaneously, with the development of molecular biology technology, RNA-seq has been widely used in the study of gene expression levels and regulatory networks [35][36][37]. Combined with typical bioinformatics analysis methods such as differential expression analysis, gene network screening, and function enrichment, transcriptome data can comprehensively and systematically identify gene expression regulatory networks at the RNA level, providing a new direction for the further study of target genes [38].
Although numerous genes have been linked to leaf rolling, there are still gaps in our comprehension of the molecular mechanisms that regulate leaf morphology. The previous study employed microarray analysis to investigate the network controlled by SRL1. However, it only focused on the differential gene expression in the cells abutting the midrib at the adaxial surface of leaf blades, which develop into epidermal cells in wild-type plants and probably transform into bulliform cells in srl1-1. Due to the differences in sampling locations and limitations of microarray technology, this could potentially result in an incomplete understanding of the regulatory mechanisms of SRL1 in leaf development. In order to address the limitations of previous research and obtain a more comprehensive understanding of the regulatory network in leaf development involving SRL1, we used CRISPR/Cas9 to create a double-target knockout mutant of SRL1 in the indica rice background and compared the gene expression profiles between the mutant and control. The study primarily focused on the key genes influenced by SRL1 in vacuole development and cell wall synthesis, as well as the potential pathways through which it may affect changes in plant morphology.

Construction of CRISPR/Cas9 Knockout Vector
The SRL1 gene is composed of eight exons and seven introns, with its full-length complementary DNA (cDNA) sequence spanning 1871 bp and encoding a protein consisting of 441 amino acids. To completely disrupt the function of SRL1 and enhance the mutation efficiency, two knockout target sites were selected within the gene. Specifically, these target sites were located in the second exon (Target 1: 5 -GCAGGTGTGCCCAATTCATT-3 ) and the third exon (Target 2: 5 -GGCGGTTTGACAAGCTACGC-3 ) ( Figure 1A). Then, the U6a-sgRNA and U6b-sgRNA expression cassettes were obtained using overlapping PCR amplification ( Figure 1B), and the pY-CRISPR/Cas9 Pubi-H knockout vector was successfully constructed ( Figure 1C).

Selection of T-DNA-Free Double-Target Homozygous Mutant Rice Lines
Calluses of the indica rice line GXU103 were infested with Agrobacterium tumefaciens carrying a double-target CRISPR/Cas9 vector. Positive mutant plants were selected using specific primers, hyg-F/hyg-R (Table S1), which contained a hygromycin resistance marker, and a total of 14 positive plants were obtained ( Figure 2B). For each T0 individual plant, the targets were amplified by PCR and identified through sequencing. At the first

Selection of T-DNA-Free Double-Target Homozygous Mutant Rice Lines
Calluses of the indica rice line GXU103 were infested with Agrobacterium tumefaciens carrying a double-target CRISPR/Cas9 vector. Positive mutant plants were selected using specific primers, hyg-F/hyg-R (Table S1), which contained a hygromycin resistance marker, and a total of 14 positive plants were obtained ( Figure 2B). For each T 0 individual plant, the targets were amplified by PCR and identified through sequencing. At the first target position, there were five homozygotes, seven heterozygotes, and two WTs. Similarly, at the second target position, there were four homozygotes, five heterozygotes, and five WTs (Table 1). Three plants were identified as double-target mutants (GXU103-X represents different mutant types), resulting in a double-target mutation rate of 21.4%. GXU103-1 had a single-base insertion at the first target site and a 3 bp deletion at the second target site. GXU103-5 had a 4 bp deletion at the first target site and a single base insertion at the second target site, while GXU103-12 had a 5 bp deletion at the first target site and a single-base insertion at the second target site (Figure 2A). The plants exhibiting these three distinct types of double-target mutations, which were discovered in T 0 generation, were used in subsequent experiments.

Analysis of Leaf Physiological Morphology and Agronomic Traits
In order to further explore the variations in traits between the mutant line srl1-1 and the wild-type GXU103, the T2 generation of srl1-103 was utilized as the experiment materials for analysis. When investigating GXU103 and srl1-103 at maturity, it was o served that the leaves of GXU103 were naturally pendulous, while srl1-103 displayed overall erect and clustered growth status with inwardly rolled leaves ( Figure 3A,B Through the leaf cross-section, it was evident that the leaves of GXU103 are relatively fl with slightly upward leaf blade edges. In contrast, the leaf blades of srl1-103 exhibite significant inward rolling ( Figure 3C). Further experimental observations revealed th the bulliform cells on the leaves of GXU103 were neatly arranged, whereas those srl1-103 were larger and appeared irregularly arranged between two adjacent sm Self-pollinated seeds from all mutant lines in the T 0 generation were harvested and referred to as T 1 generation seeds. Using primers Cas9-F/Cas9-R (Table S1), 120 T 1 generation plants were detected, out of which 73 plants were identified as T-DNA-free. Among these 73 plants, a total of 26 were discovered to be homozygous for double-target mutations, and they were subsequently named srl1-103 ( Figure 2C). The sequencing results show that the double-target mutation types in the T 1 generation of srl1-103 were all derived from GXU103-X in the T 0 generation, confirming that the double-target mutation types detected in the T 0 generation were stably inherited in the T 1 generation of srl1-103. Three distinct double-target mutation types from the T 2 generation of srl1-103 were selected, and 100 individual plants were cultivated for each genotype. These 300 individual plants all exhibited defects in leaf development, with upright rolled leaves and a more compact plant type compared with WT. Finally, GXU103-1 mutant plants from the T 2 generation of srl1-103 were selected for subsequent experimental studies.

Analysis of Leaf Physiological Morphology and Agronomic Traits
In order to further explore the variations in traits between the mutant line srl1-103 and the wild-type GXU103, the T 2 generation of srl1-103 was utilized as the experimental materials for analysis. When investigating GXU103 and srl1-103 at maturity, it was observed that the leaves of GXU103 were naturally pendulous, while srl1-103 displayed an overall erect and clustered growth status with inwardly rolled leaves ( Figure 3A,B). Through the leaf cross-section, it was evident that the leaves of GXU103 are relatively flat, with slightly upward leaf blade edges. In contrast, the leaf blades of srl1-103 exhibited significant inward rolling ( Figure 3C). Further experimental observations revealed that the bulliform cells on the leaves of GXU103 were neatly arranged, whereas those of srl1-103 were larger and appeared irregularly arranged between two adjacent small vascular bundles. The total number of epidermal cells and bulliform cells was similar between GXU103 and srl1-103 ( Figure 3D).
To examine the impact of SRL1 gene knockout on rice agronomic traits, an investigation was conducted on mature GXU103 and srl1-103 plants. The results show that there were no significant differences in plant height, length of flag leaf, number of panicles, length of panicle, seed setting rate, and 1000-grain weight ( Figure 3E,F,I-L). However, compared with GXU103, the flag leaf width of srl1-103 was 12.6% narrower, showing a significant difference ( Figure 3G). Surprisingly, the leaf rolling index (LRI) of srl1-103 reached 57.12%, which was 48.77% higher than GXU103, indicating a highly significant difference ( Figure 3H). The above results indicate that the knockout of SRL1 altered the morphology of GXU103 leaves, resulting in increased leaf rolling and upright growth.

Transcriptome Analysis of the Wild-Type GXU103 and the Mutant Line srl1-103
To uncover the regulatory network involved in leaf rolling associated with SRL1, RNA-seq analysis was performed on mature leaves of the wild-type GXU103 and the T 2 generation of srl1-103. For each group, the samples were collected from three independent plants. After filtering out the ineligible reads from the raw data, six libraries were obtained. Each library consisted of over 18,852,288 reads, with more than 15,441,394 reads aligning to the reference genome at a unique comparison rate higher than 81.80% ( Table 2). The results indicate that high-quality sequencing data were obtained from RNA-seq. Additionally, the correlation between the samples was evaluated by the Pearson's correlation coefficient (PCC), revealing that the wild-type GXU103 exhibited a PCC value above 0.94, while the mutant line srl1-103 displayed a PCC value higher than 0.96. In general, all the results indicate that the high-quality sequencing data obtained from RNA-seq had a high level of confidence and could be used for subsequent transcriptome analysis ( Figure 4A).
were no significant differences in plant height, length of flag leaf, number of panicle length of panicle, seed setting rate, and 1000-grain weight ( Figure 3E,F,I-L). Howeve compared with GXU103, the flag leaf width of srl1-103 was 12.6% narrower, showing significant difference ( Figure 3G). Surprisingly, the leaf rolling index (LRI) of srl1-1 reached 57.12%, which was 48.77% higher than GXU103, indicating a highly significa difference ( Figure 3H). The above results indicate that the knockout of SRL1 altered th morphology of GXU103 leaves, resulting in increased leaf rolling and upright growth.

Transcriptome Analysis of the Wild-Type GXU103 and the Mutant Line srl1-103
To uncover the regulatory network involved in leaf rolling associated with SRL RNA-seq analysis was performed on mature leaves of the wild-type GXU103 and the generation of srl1-103. For each group, the samples were collected from three indepen ent plants. After filtering out the ineligible reads from the raw data, six libraries we obtained. Each library consisted of over 18,852,288 reads, with more than 15,441,3 reads aligning to the reference genome at a unique comparison rate higher than 81.80 ( Table 2). The results indicate that high-quality sequencing data were obtained fro  In srl1-103, a total of 459 differentially expressed genes (DEGs) were detected by RNAseq, of which 388 genes were significantly upregulated, while 71 genes were downregulated ( Figure S1, Table S2). Among them, the expression of SRL1 (LOC_Os07g01240) was notably decreased in srl1-103, suggesting the successful knockout of SRL1 through CRISPR/Cas9 ( Figure 4B). To investigate the involvement of DEGs in the development of the rolled leaf phenotype, a gene ontology (GO) analysis was performed on these genes. The analysis unveiled that DEGs were significantly enriched in various biological pathways, including cell structure, mitosis, phytohormones, epidermal development, stress response, cellular homeostasis, and cellular metabolism ( Figure 4C).

Impact of SRL1 Knockout on Cellular Structure
To further understand the mechanism of SRL1 regulation in leaf rolling, we focused on the genes related to the cell wall synthesis pathway (Table 3)

Impact of SRL1 Knockout on Cellular Structure
To further understand the mechanism of SRL1 regulation in leaf rolling, we focused on the genes related to the cell wall synthesis pathway (Table 3). Plant cell walls are mainly composed of cellulose, hemicellulose, lignin, and pectin. They provide mechanical support for plant tissues and play an important role in the development of plant morphology. In the transcriptome data, there were aberrant expressions of certain genes involved in cell wall synthesis. For example, LOC_Os08g01670 (RICE PECTIN METHYLESTERASE IN-HIBITOR 28, OsPMEI28), which exhibits pectin methyl esterase inhibitor activity and plays a key role in regulating the level of pectin methylation, was significantly upregulated in srl1-103. LOC_Os05g46510, which encodes a putative polygalacturonase (PG) with pectindegrading effects, was also similarly significantly upregulated. Glycosyl hydrolase (GH) is one of the most widely distributed protein hydrolases in plants, participating in multiple physiological processes such as plant cell wall remodeling, lignification, and phytohormone activation. It was found that LOC_Os04g51450, LOC_Os10g28080, LOC_Os04g39814, LOC_Os01g71474, LOC_Os01g71350, and LOC_Os11g47600, each of which encodes a putative glycosyl hydrolase, were likely to be involved in plant cell wall remodeling and were significantly upregulated in srl1-103. In addition to genes related to the cell wall synthesis pathway, significant changes were found in vacuole-related genes. Previous studies on SRL1 have suggested that it negatively regulates the expression of vacuole-related genes, which affect the formation of bulliform cells. The transcriptome data reveal that LOC_Os09g23300 (VACUOLAR IRON TRANSPORTER 2, OsVIT2), which encodes a vacuolar membrane transporter protein responsible for ion transport processes, was significantly upregulated in srl1-103. The differential expression of genes related to cell wall synthesis and vacuolar function might be involved in the regulation of leaf rolling by SRL1.

Impact of SRL1 Knockout on Plant Type and Leaf Angle
In comparison to WT, srl1-103 exhibited a more compact plant type characterized by smaller leaf angles and upright rolled leaves. To investigate the regulatory mechanism of plant type change in srl1-103, further analysis was conducted on differential genes related to the regulation of plant type (Table 3). It was shown that LOC_Os08g44840 (SLENDER GRAIN, SLG), which encodes a BAHD acyltransferase-like protein involved in the regulation of leaf angle, was significantly downregulated in srl1-103. A previous study showed that the downregulation of LOC_Os08g44840 led to a decrease in leaf angle and a more upright leaf in rice. Therefore, it could be inferred that the significant downregulation of LOC_Os08g44840 in srl1-103 was likely the main reason for the reduced leaf angle and compact plant architecture.

Discussion
Bulliform cells are a special type of epidermal cells that exist on the leaves of numerous plants. Typically, these cells are located in the upper epidermis or near the adaxial surface of the leaf, and display a large, empty, colorless, and bubble-shaped form, appearing between the two veins. Changes in the number, size, or arrangement of bulliform cells can result in leaf rolling. A previous report indicated that the leaf rolling phenotype of srl1-1 was caused by the enhanced formation of bulliform cells at the adaxial cell layers [9]. According to our data, the knockout mutant srl1-103 displayed larger and irregularly arranged bulliform cells in its leaves compared with WT ( Figure 3D), which may be one of the reasons for the rolling of leaves. This finding aligns with previous reports highlighting the significant role of SRL1 in the formation of bulliform cells. Statistical analysis was performed on several agronomic traits of srl1-103 ( Figure 3E-L). The results reveal that the knockout of SRL1 significantly influenced two agronomic traits, the width of the flag leaf and LRI. In the srl1-103 mutant, there was a noticeable reduction in the width of the flag leaf compared with WT, which could potentially affect its photosynthetic capacity and overall growth. Additionally, a significant alteration in LRI was observed in srl1-103, which reflects the degree of leaf rolling. The change in LRI suggests that leaf development has been disrupted, potentially affecting the plant's ability to efficiently absorb sunlight and exchange gases with the environment. Interestingly, while there were significant changes in the flag leaf width and LRI, other agronomic traits such as plant height, number of panicles, and 1000grain weight remained unaffected by the knockout of SRL1. This suggests that SRL1 plays a specific role in certain aspects of leaf development and morphology, while not exerting an influence on the overall plant structure or reproductive characteristics.
As an important agronomic trait that is easily observable during the growth of rice, leaf rolling has garnered considerable attention from breeders [5]. The mechanisms of leaf rolling are complex, typically involving the development of bulliform cells and cell wall formation [39]. In this study, it was found that genes related to cell wall synthesis and vacuolar functions were abnormally expressed in srl1-103, which may be the primary cause of the altered leaf morphology. To better understand the potential pathways associated with the SRL1 regulation of leaf rolling and plant type, the results of transcriptome data predictions were visualized ( Figure 5). Among the DEGs, LOC_Os08g01670 (OsPMEI28) exhibits pectin methylesterase inhibitor activity, and the overexpression of OsPMEI28 has been shown to decrease cell wall thickness and the content of cell wall components [40]. LOC_Os05g46510 encodes a putative polygalacturonase enzyme, which participates in cell wall biosynthesis by degrading the pectin within cellular components, consequently affecting the mechanical strength of the cell wall. LOC_Os04g51450, LOC_Os10g28080, LOC_Os04g39814, LOC_Os01g71474, LOC_Os01g71350, and LOC_Os11g47600 encode presumed glycosyl hydrolases with the ability to hydrolyze polysaccharides within cell walls. In general, the three types of genes related to cell wall synthesis were significantly upregulated in srl1-103, indicating that SRL1 may participate in cell wall formation by regulating the expression of these genes, thereby influencing leaf morphology. This finding is similar to previous studies on the regulatory mechanism of CLD1/SRL1 [9]. cell layers [9]. According to our data, the knockout mutant srl1-103 displayed larger and irregularly arranged bulliform cells in its leaves compared with WT ( Figure 3D), which may be one of the reasons for the rolling of leaves. This finding aligns with previous reports highlighting the significant role of SRL1 in the formation of bulliform cells. Statistical analysis was performed on several agronomic traits of srl1-103 ( Figure 3E-L). The results reveal that the knockout of SRL1 significantly influenced two agronomic traits, the width of the flag leaf and LRI. In the srl1-103 mutant, there was a noticeable reduction in the width of the flag leaf compared with WT, which could potentially affect its photosynthetic capacity and overall growth. Additionally, a significant alteration in LRI was observed in srl1-103, which reflects the degree of leaf rolling. The change in LRI suggests that leaf development has been disrupted, potentially affecting the plant's ability to efficiently absorb sunlight and exchange gases with the environment. Interestingly, while there were significant changes in the flag leaf width and LRI, other agronomic traits such as plant height, number of panicles, and 1000-grain weight remained unaffected by the knockout of SRL1. This suggests that SRL1 plays a specific role in certain aspects of leaf development and morphology, while not exerting an influence on the overall plant structure or reproductive characteristics.
As an important agronomic trait that is easily observable during the growth of rice, leaf rolling has garnered considerable attention from breeders [5]. The mechanisms of leaf rolling are complex, typically involving the development of bulliform cells and cell wall formation [39]. In this study, it was found that genes related to cell wall synthesis and vacuolar functions were abnormally expressed in srl1-103, which may be the primary cause of the altered leaf morphology. To better understand the potential pathways associated with the SRL1 regulation of leaf rolling and plant type, the results of transcriptome data predictions were visualized ( Figure 5). Among the DEGs, LOC_Os08g01670 (OsPMEI28) exhibits pectin methylesterase inhibitor activity, and the overexpression of OsPMEI28 has been shown to decrease cell wall thickness and the content of cell wall components [40]. LOC_Os05g46510 encodes a putative polygalacturonase enzyme, which participates in cell wall biosynthesis by degrading the pectin within cellular components, consequently affecting the mechanical strength of the cell wall. LOC_Os04g51450, LOC_Os10g28080, LOC_Os04g39814, LOC_Os01g71474, LOC_Os01g71350, and LOC_Os11g47600 encode presumed glycosyl hydrolases with the ability to hydrolyze polysaccharides within cell walls. In general, the three types of genes related to cell wall synthesis were significantly upregulated in srl1-103, indicating that SRL1 may participate in cell wall formation by regulating the expression of these genes, thereby influencing leaf morphology. This finding is similar to previous studies on the regulatory mechanism of CLD1/SRL1 [9]. A previous study on SRL1 indicated that some vacuole-related genes are significantly upregulated in srl1-1, suggesting that SRL1 negatively regulates the expression of vacuole H + -ATPase subunit and H + -pyrophosphatase genes to inhibit the formation of bulliform cells [8]. However, only one vacuole-related gene, LOC_Os09g23300 (OsVIT2) A previous study on SRL1 indicated that some vacuole-related genes are significantly upregulated in srl1-1, suggesting that SRL1 negatively regulates the expression of vacuole H + -ATPase subunit and H + -pyrophosphatase genes to inhibit the formation of bulliform cells [8]. However, only one vacuole-related gene, LOC_Os09g23300 (OsVIT2) [41], was found to be significantly upregulated in srl1-103 in the transcriptome data. Leaf angle, which represents the inclination between the leaf blade and the vertical culm, plays a pivotal role in determining the overall plant architecture. Numerous studies have shown that the degree of the leaf angle is mainly associated with the biosynthesis and signal transduction pathways of brassinosteroid (BR), indole-3-acetic acid (IAA), and gibberellin (GA). Moreover, the interaction among these plant hormones also regulates the leaf angle to some extent [42][43][44]. The signaling pathway of BR is crucial for regulating the leaf angle in rice and involves the action of multiple genes. BRI1, encoded by the BR receptor kinase gene OsBRI1 in rice, shares high sequence similarity with Arabidopsis BRI1 and plays various roles in the growth and development of rice. Deficits in the function of OsBRI1 led to impaired BR signaling, resulting in reduced leaf angles and erect leaf traits in rice [45]. PRA2, encoded by the small G protein gene, can interact with BRI1 and inhibit its kinase activity, thereby negatively regulating BR signaling. The overexpression of OsPRA2 resulted in decreased leaf angle and reduced sensitivity to exogenous BR. In contrast, suppressing the expression of OsPRA2 through RNA interference led to an increase in leaf angle and enhanced sensitivity to exogenously applied BR [46]. In addition to the genes related to the BR signaling pathway, genes associated with BR homeostasis, such as RAV6 and SLG, also affect the magnitude of the leaf angle. RAV6 encodes a protein with a B3 DNA-binding domain, which affects rice leaf angle by mediating BR homeostasis, and the semidominant mutant Epi-rav6 showed an increase in leaf angle [47]. SLG encodes a BAHD acyltransferase-like protein that serves as a crucial regulatory factor in maintaining BR homeostasis. It functions as homomers and regulates grain size and leaf angle in rice [48]. The transcriptome data indicate that LOC_Os08g44840 (SLG) was significantly downregulated in srl1-103, possibly exerting a significant influence on plant architecture. This finding has not been previously emphasized in studies on SRL1. There may be potential interdependent regulatory mechanisms between SRL1 and SLG, which necessitate further investigation to fully understand their comprehensive effects on the structure of rice.
Currently, there are few reports on the involvement of leaf rolling genes in stress tolerance pathways. These stress tolerance phenotypes are typically regulated by multiple genes, making their study more difficult. In this study, we found that the knockout of SRL1 not only impacted leaf rolling and plant type, but it also affected the expression of some resistance genes (Table S3). This suggests that the SRL1 gene may be involved in certain stress tolerance regulatory networks. Chitinase is a disease course-related protein that holds significance in plant defenses against pathogens and insect pests. Moreover, it actively participates in the plant's response to abiotic stress [49][50][51]. The transcriptome data show that LOC_Os04g41680, which encodes a protein precursor of the chit3-chitinase family, was significantly upregulated in srl1-103 and may enhance disease resistance. In addition, LOC_Os02g14440 and LOC_Os07g48010, presumed precursors of peroxidase, were significantly upregulated in srl1-103, which may accelerate the peroxidase clearance process and improve the tolerance to oxidative stress. Furthermore, the knockout of SRL1 also affected the differential expression of some MYB transcription factors. MYB transcription factors, as one of the most abundant transcription factor families in plants, are widely involved in plant stress responses [52][53][54]. The overexpression of LOC_Os04g43680 (Osmyb4) can enhance cold tolerance of Arabidopsis thaliana and affect the expression of multiple genes in cold stress response pathways [55]. Interestingly, the transcriptome data reveal a significant upregulation of Osmyb4 in srl1-103, suggesting that it might enhance the cold tolerance of srl1-103. Similarly, LOC_Os11g45740 (JASMONIC ACID-INDUCIBLE RICE MYB GENE, JAmyb) encodes an R2R3-type MYB transcription factor. The overexpression of this gene has been shown to increase salt tolerance in transgenic lines [56], and its expression was significantly upregulated in srl1-103. Several genes encoding disease course-related proteins, whose expression is induced by jasmonic acid, play important roles in biological and abiotic stresses, and were also differentially expressed. LOC_Os01g28450 (PATHOGENESIS-RELATED CLASS 1B, OsPR1b) [57], LOC_Os03g18850 (JASMONATE INDUCIBLE PR10, JIOsPR10) [58], and LOC_Os12g36830 (ROOT SPECIFIC RICE PR10, RSOsPR10) [59] were found to be significantly upregulated in srl1-103, which may increase its corresponding resistance capabilities. Based on the transcriptome data, it is speculated that SRL1 is potentially involved in the regulation of the expression of these resistance genes, which may be a new direction in the study of leaf shape and stress tolerance. Additionally, the potential off-target effects of CRISPR/Cas9 gene editing technology may affect the analysis results. In this study, the predicted maximum off-target score for target 1 in indica rice was 0.23, while for target 2, it was 0.144. This indicates that further experiments are still needed to validate the results of the transcriptome analysis.
In this study, we conducted transcriptomic analysis to investigate the regulatory mechanism of SRL1 in rice, aiming to provide valuable insights for the further exploration of SRL1-related functions. The results indicate that SRL1 may affect leaf development through the regulation of genes associated with cell wall synthesis, vacuole development, and leaf angle. This fills a significant gap in previous research and enhances our understanding of the regulatory network in which SRL1 may be involved, making it clear and comprehensive. Furthermore, we also observed that the knockout of SRL1 led to a significant upregulation of certain resistance genes, suggesting that SRL1 may be involved in the regulatory network of stress tolerance. Under stressful conditions such as high temperature, high salinity, and drought, leaves tend to curl in order to reduce transpiration and minimize water loss, thus enhancing their tolerance. Whether leaf rolling under normal conditions contributes to resistance remains to be further investigated through methods such as resistance identification, gene expression analysis, and interaction analysis. With the continuous improvement of CRISPR/Cas9 editing technology, it is now possible to simultaneously target multiple sites, providing new methods for studying traits controlled by multiple genes and gene families in plants. In addition, CRISPR/Cas9 has demonstrated tremendous potential in the field of breeding. By modifying multiple genes in crops, it enables the simultaneous improvement of various traits, including resistance, quality, and yield, thus advancing the breeding process. The integration of traditional breeding techniques with gene editing technologies is emerging as a significant trend in breeding research, offering numerous opportunities and challenges for future agricultural production.

Plant Materials and Growing Environment
GXU103, an indica rice line developed by the Rice Research Institute of Guangxi University, possesses outstanding agronomic traits and has been utilized for conducting genetic transformation and gene editing experiments. All rice lines were cultivated in a mesh room with isolation conditions in Nanning, China, during the normal rice-growing seasons, and treated in accordance with the conventional planting management method. The materials used for phenotypic observation, agronomic trait statistics, and transcriptome analysis were from the T 2 generation of srl1-103, derived from the GXU103-1 mutant genotype. GXU103 was utilized as the wild-type control.

Genotype Determination of T 0 Generation and Screening for T-DNA-Free Plants
Genomic DNA was extracted from fresh leaves of each plant. T 0 -positive plants were detected using specific primers (hyg-F/hyg-R) with a hygromycin resistance marker. The target sequences were amplified using the specific primers SRL-F/SRL-R and identified by Sanger sequencing. T-DNA-free plants were screened by amplifying T-DNA isolates of the T 1 generation using Cas9-F/Cas9-R primers.

Statistical Analysis of Agronomic Traits
The agronomic traits of wild-type GXU103 and the T 2 generation of srl1-103 were investigated through field surveys. Five individual plants from each line were used for agronomic trait measurements. The main traits evaluated included plant height, length of flag leaf, number of panicles, length of panicle, seed setting rate, 1000-grain weight, width of flag leaf, and leaf rolling index. Statistical analysis was performed using the software GraphPad Prism 8. Significant differences were determined by the Student's t test.

Histology and Microscopy Observations
Flag leaves were fixed in 25% glutaraldehyde solution, dehydrated by graded ethanol solutions, and then embedded in low-temperature paraffin. Cross-sections (12 mm) were cut using a sectioning machine (Leice RM2235), spread and dried, dewaxed, rehydrated in graded ethanol solutions, stained with filtered 1% toluidine blue, and sealed with gum Arabic. Sections were microscopically examined and photographed (Zeiss Scope).

Leaf Rolling Index Analysis
Leaf rolling index (LRI) was determined for WT and mutant plants at the maturity stage using the natural state leaf edge distance (Ln) and the expanded state leaf edge distance (Lw). The LRI was calculated as: LRI (%) = (Lw − Ln)/Lw × 100.

Transcriptome Sequencing and Analysis
Transcriptional profiling was conducted using mature leaf tissues from wild-type and the T 2 generation of srl1-103. The samples for the RNA-seq experiment were obtained from three independent WT and mutant plants. The total RNA was extracted using an RNeasy Plant Mini Kit (No.74904, QIAGEN, Hilden, Germany). RNA sequencing libraries were constructed using the NEB Next Ultra TM RNA Library Preparation Kit Illumina (NEB, Omaha, NE, USA) and subsequently sequenced on a NovaSeq 6000 to generate double-ended 150 bp reads. The raw data were filtered and quality-checked using the default parameters of FASTP (an ultra-fast all-in-one FASTQ preprocessor). Then, the clean data were compared to the rice reference genome (ref, Minghui genome) using the default parameters of Hisat2 (a fast spliced aligner with low memory requirements). Stringtie (enables improved reconstruction of a transcriptome from RNA-seq reads) was used to calculate gene expression levels, and DESeq2 (moderates estimation of fold change and dispersion for RNA-seq data) was utilized to identify differentially expressed genes between samples, with the criteria of fold change > 2 and FDR (false discovery rate) < 0.05. Subsequently, GO functional enrichment analysis for GO terms with a p value < 0.05 was performed using GOATOOLs (a Python library for gene ontology analyses).

Data Availability Statement:
The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive [60] in the National Genomics Data Center [61] and China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA010568), which are publicly accessible at https://ngdc.cncb.ac.cn/gsa (accessed on 11 April 2023).