YU Xiao-qi,XlE Wei,LlU He,LlU Wei,ZENG Da-li,QlAN Qian,REN De-yong

State Key Laboratory of Rice Biology,China National Rice Research Institute,Hangzhou 310006,P.R.China

Abstract Moderate leaf rolling can maintain leaf erectness,improve light transmittance in the population,and improve light energy utilization,thereby increasing rice yield. This study used ethyl methanesulfonate (EMS) to treat Yunjing 17 (YJ17) and obtained a semi-rolled leaf mutant that was named semi-rolled leaf 3 (srl3). We found that the rolled-leaf phenotype was due to the aberrant development of bulliform cells and the loss of sclerenchymatous cells. In addition,the shoot and root length of srl3 seedlings differed from the wild type. The srl3 mutant had significantly lower plant height and seed-setting rate but notably greater tiller number,panicle length,and primary branch number per panicle than the wild type. Genetic analysis showed that a single recessive nuclear gene defined the srl3 mutant,and it was precisely located in a 144-kb region between two insertion-deletion (InDel) markers,M8 and M19,on chromosome 2. In this region,no leaf-rollingrelated genes have been reported previously. Thus,the study indicated that SRL3 is a novel leaf-rolling-related gene,and the results laid the foundation for the cloning and functional analysis of the SRL3 gene.

Keywords: rice,srl3,semi-rolled leaf,bulliform cell,sclerenchymatous cell

1.lntroduction

Leaf is the main photosynthesis organ of plants,and an optimal leaf morphology is conducive to shaping ideal plant type and improving photosynthetic efficiency. In the model of super-high-yield hybrid rice,Yuan (1997)attributed the shape of rice functional leaves to long,erect,narrow,thick,and rolled. The moderate leaf rolling can help to keep the erectness of leaves,improve the light receiving conditions in the later stage of population growth,and improve the photosynthetic efficiency of the whole plant. In addition,moderate leaf rolling can reduce the radiation of the sun to the leaves and reduce the damage to the leaves,improve the resistance of the plants,and increase the accumulation of photosynthetic products,thereby increasing the yield of crops (Zhanget al.2009). Therefore,the study of leaf-rolling-related genes has always been the focus of rice breeders.

So far,a series of genes related to leaf rolling have been cloned in rice. These genes are mainly divided into four categories according to their causes of leaf rolling.Most of these genes take part in the regulation of leaf rolling by affecting the size or number of bulliform cells. In monocot,the bulliform cells usually cluster between two vascular bundles on the adaxial surface of the leaves,and they take part in the regulation of leaf rolling by controlling the gain and loss of intracellular water (Zhanget al.2021).The previous results showed that under drought-stressed conditions,bulliform cells lose expansion pressure and contract,resulting in the leaf-rolling phenotype. Once water is not deficient,the bulliform cells expand,and the leaves stretch again (Zouet al.2014). Moreover,the area and number of bulliform cells can control the direction of leaf rolling. When the area and number of bulliform cells increase,the leaves roll abaxially;while they decrease,the leaves roll adaxially. Overexpression ofACL1in wild type results in outward leaf rolling,but inhibition ofACL1expression does not lead to abnormality in leaves (Liet al.2010). The mutation of theLC2gene affects the cell division of leaf epidermis and increase the number of bulliform cells,so thelc2mutant showed increased leaf angle and abaxial-rolled leaf phenotype (Zhaoet al.2010).The overexpression of theRoc5gene leads to decreased size and number of bulliform cells and inward-rolled leaves;while the gene is inhibited,the leaves roll outward(Zouet al.2011). TheSRL1gene negatively regulates the expression of genes encoding H+-ATPase subunits and H+-pyrophosphatase,inhibits the formation of bulliform cells,and then regulates leaf rolling (Xianget al.2012).RL14changes the water transport by regulating the components of secondary cell walls,further leads to water loss of leaf,and finally causes the abnormal shape of bulliform cells(Fanget al.2012). In theoszhd1mutant,the arrangement of bulliform cells was abnormal,and the number increased,which lead to the rolling of leaves (Xuet al.2014).REL1andREL2function in leaf morphological regulation,and they may be involved in many biological processes,such as brassinolide (BR) signals coordination,auxin synthesis and transport,and bulliform cell development (Chenet al.2015;Yanget al.2016).

The other three types of genes are in the minority. The second type of leaf rolling is caused by the imbalanced development between abaxial and adaxial surfaces.TheOsAGO7gene encodes a protein belonging to the big Argonaute (AGO) protein family,its transferred DNA(T-DNA) tagging mutant R05 has rolling leaves,and the upward rolling of leaves can cause the phenotype of erect leaves (Shiet al.2007). In theadl1mutant,the number of adaxial bulliform cells increases,and the abaxial surface produces similar bulliform cells,thus changing the polarity of leaves,leading to the outward leaf rolling(Hibaraet al.2009). TheSLL1gene modulates the shape of leaves by regulating the programmed cell death of sclerenchymatous cells on the abaxial side of leaves. Overexpression ofSLL1can promote the development of phloem on the abaxial surface and inhibit the development of adaxial epidermal cells and sclerenchymatous cells,leading to the leaves rolling in the adaxial direction (Zhanget al.2009). Inhibition ofOsAGO1agene expression results in slight rather than extreme abaxial rolling of leaves,and no significant negative pleiotropy has been found (Liet al.2013). The third type of gene regulating leaf rolling is caused by the aberrant development of sclerenchymatous,such asSRL2. TheSRL2gene encodes a function-unknown protein,and there are defective sclerenchymatous cells insrl2leaves,so the leaves do not have enough mechanical support and roll inward (Liuet al.2016). The last type of gene,such asCFL1,can inhibit the expression of the key genes related to cuticle development (BDGandFDH),resulting in abnormal cuticle development incfl1mutant,and the leaves showed rolling (Wuet al.2011).

In this study,we identified a mutant with semirolled leaves namedsrl3. The preliminary phenotypic identification and cytological analysis showed thatsrl3leaves semi-rolled to the adaxial surface gradually from the seedling stage to the mature stage,and the morphology of bulliform cells and sclerenchymatous cells both altered. Moreover,the shoot and root lengths ofsrl3seedlings also showed significant differences from the wild type. Fine mapping showedSRL3located at a physical distance of 144 kb on chromosome 2,and we speculated that it was a novel leaf-rolling-related gene.Our study provides a reference for further study on the molecular mechanism of leaf rolling in rice.

2.Materials and methods

2.1.Plant materials

Thesrl3mutant was identified from a mutagenized population of YJ17 (cv.japonica) treated with ethyl methanesulfonate (EMS). YJ17 represents the wild type for phenotypic analysis. Thesrl3mutant was crossed with TN1 (cv.indica),and the obtained F1seeds were sown and self-pollinated to produce the F2plants for constructing a mapping population. Plants were cultivated in the paddy fields in Hangzhou,Zhejiang Province and Lingshui,Hainan Province,during the natural growing season.

2.2.Measurement of the leaf-rolling index (LRl)

During 20 days after the flowering stage,we measured blade widths of the flag leaf,second upper leaf,and third upper leaf at natural state (Ln) and unfolded state (Lw) in thesrl3mutant and corresponding wild type,respectively.To determine the LRIs,we used the formula as follows:LRI (%)=(Lw-Ln)/Lw×100.

2.3.Measurement of leaf water content

To determine the leaf water content,we randomly selected some flag leaves,second upper leaves,and third upper leaves from the wild type andsrl3mutant. Firstly,we measured the weight of fresh leaf and the data were denoted by W1,and the leaf weight after soaking in water for 4 h at room temperature was recorded as W2. Then,we put the samples at 70°C for 24 h,and the dry weight of the leaf was recorded as W0. We calculated the leaf water content as the following formula: Leaf water content(%)=(W1-W0)/(W2-W0)×100.

2.4.Histological and microscopic observations

Fresh leaves of the wild type andsrl3mutant were collected at the maximum tillering stage. The materials were fixed in formalin-acetic acid-alcohol (FAA) solution at 4°C for about 24 h. Then,they were gradually dehydrated in a graded ethanol series,infiltrated with xylene,and embedded in paraffin. Cross sections (8 μm)were conducted with a rotary microtome and put at 42°C for at least 3 days. Some slices were chosen,stained with safranine and fast green,and dehydrated in a graded ethanol series. The stained slices were covered with neutral resins and observed with a LEICA microscope (DM4P,Leica,German) (Ruanet al.2020;Yuet al.2020).

2.5.Map-based cloning of SRL3

The F2population was constructed from the cross betweensrl3and TN1 for mapping of theSRL3gene.A total of 658 F2plants showed the semi-rolled leaf phenotypes the same as thesrl3mutant,and theSRL3gene was preliminarily located between two simple sequence repeat (SSR) markers,RM13633 and RM13840,on chromosome 2. All these SSR markers were downloaded from the GRAMENE database. The insertion-deletion (InDel) markers were developed by comparison of the sequences between theindica93-11 andjaponicaNipponbare (NCBI) for fine mapping. The primers we used are listed in Appendix A.

2.6.qRT-PCR analysis

To examine the expressions of leaf-rolling-related genes and candidate genes in the fine mapping region,total RNA from leaves 20 days after the seedling stage was extracted using the RNeasy Plant Mini Kit (Axygen,America). A total of 3 μg of total RNA was treated with RNase-free DNase,and the first-strand complementary DNA (cDNA) was synthesized by reverse transcription(TaKaRa,Japan). The qRT PCR analyses were performed on the Bio-Rad CFX96 Real-Time PCR System using the SYBR Green PCR Master Mix Kit(Toyobo,Japan). The primers used for the expression analysis of leaf rolling-related genes,cell cycle,expansion and proliferation-related genes were designed according to the previous reports. The riceubiquitin-1gene was used as a control to normalize the expression levels of the tested genes. Each sample had at least three biological replicates. The primers we used are listed in Appendix A.

3.Results

3.1.The phenotype of the srl3 mutant

Under the same natural conditions,thesrl3plants exhibited obvious differences from the wild type from the seedling stage. All leaves of thesrl3mutant showed adaxially and semi-rolled phenotypes and remained constant during the whole life span (Fig.1-A-D). During the tillering stage,the LRIs of thesrl3flag leaf,second upper leaf,and third upper leaf reached about 41,26,and 14%,respectively,while the corresponding last emerged three leaves in the wild type were almost flat (Fig.1-E). The leaf water contents of the flag leaf and third upper leaf were both significantly decreased,the water content of the wild-type flag leaf and third upper leaf were 92.7 and 92.1%,and they decreased to 89.3 and 90.0% in thesrl3leaves,respectively (Fig.1-F).The length and width of the top three leaves were also measured and analyzed. The length of the second upper leaf and third upper leaf of thesrl3mutant showed no difference from that of the wild type,while the length of the flag leaf was significantly increased (Appendix B). Also,thesrl3top three leaves were all significantly narrowed;the width of the flag leaf,second upper leaf,and third upper leaf was about 94,83,and 70% of the width of the wild type,respectively (Fig.1-G).

The agronomic traits were also investigated,and most of them had differences between the wild type and thesrl3mutant. Compared with the wild type,the plant height ofsrl3was significantly decreased,while the tiller number ofsrl3was notably increased (Table 1). The grain number per panicle showed no notable difference,but the panicle length of thesrl3mutant was increased by 2.07 cm (Table 1). The primary branch number was significantly increased by 3.3,but the number of the secondary branch showed no difference from the wild type (Table 1). The seed-setting rate of thesrl3mutant was sharply decreased,which was only 11% of the wild type rate (Table 1).

Table 1 Agronomic trait of the wild type and the srl3 mutant

3.2.Histological morphology of the srl3 leaves

Leaf rolling is closely related to bulliform cells,which usually cluster between two vascular bundles on the adaxial surface of the leaves. The paraffin sectioning results showed that the number and area of bulliform cells between midrib,large vein,and small vein were all significantly decreased on the adaxial surface ofsrl3flag leaves compared with that of the wild type(Fig.2-A-F). According to our statistics,the number and area of bulliform cells between vascular bundles in thesrl3mutant were decreased to 77.7 and 64.8% of the wild type,respectively (Fig.2-I and J). In addition,we observed that there was no abaxial sclerenchymatous cell under the small veins of thesrl3leaves,but not all small vascular bundles lacked sclerenchymatous cells on the abaxial surface (Fig.2-C and F). We further performed the ultrastructural observation by SEM. In the wild type,dumb-bell-shaped silica cells were arranged orderly on both sides of vascular bundles and were parallel to the midrib. According to statistics,there is a row of dumbbell-shaped silica cells under every small vascular bundle.However,on the abaxial surface ofsrl3leaves,we could not observe these dumb-bell-shaped silica cells under some small veins (Fig.2-G and H). All these results indicate that the adaxial semi-rolled phenotype ofsrl3resulted not only from the decreased area and number of bulliform cells but also from the developmental defect of sclerenchymatous cells.

3.3.Germination of the srl3 mutant

During the germination stage of seeds,we observed that the germination ofsrl3mutant was different from that of the wild type. The statistical results showed that the germination rate ofsrl3seeds was obviously decreased compared with the wild type (Fig.3-A,B and D). Then we found that the shoot length of 5-d-oldsrl3seedlings was significantly increased compared with the wild-type seedlings in the same period,and the measurement results showed the shoots ofsrl3were about 21.0%longer than that of the wild type (Fig.3-C and E). In addition,we observed that there were fewer lateral roots in thesrl3mutant,so the total length ofsrl3roots was notably decreased to 58.0% of the wild type. The diameter ofsrl3roots was also significantly decreased to 82.2% of the wild type (Fig.3-C,F and G). We speculated that the difference in shoot length and root growth of thesrl3mutants might be because the mutation of theSRL3gene also affected the content of auxin or other hormones in the plant.

3.4.Genetic analysis of the srl3 mutant

For genetic analysis,a population was obtained by crossing thesrl3mutant with TN1. The leaves of all F1individuals exhibited normal phenotypes the same as those of wild-type individuals,and there were no mutated plants. In the F2populations,the abnormal phenotype of the semi-rolled leaf was segregated,and there was no intermediate phenotype. In the F2population derived fromsrl3/TN1,there were 944 wild-type plants and 324 semirolled-leaf plants. In the reciprocal crossing population derived from TN1/srl3,730 individuals showed normal phenotypes,and 257 individuals had semi-rolled leaves(Table 2). The results exhibited that the segregation ratio of normal plants to abnormal plants by crossing and reciprocal crossing both fit well with 3:1 ratio,indicating that the phenotype of adaxial semi-rolled leaves is probably controlled by a single recessive nuclear gene.

3.5.Fine mapping of SRL3

A total of 225 SSR markers distributed on 12 chromosomes were used for the screening of polymorphism between two parents (YJ17 and TN1).Among them,126 polymorphic markers were used for linkage analysis. Thesrl3mutants with semi-rolled leaves were randomly selected from the F2population and used to locate theSRL3gene,and it was initially localized between two SSR markers,RM13633 and RM13840,on chromosome 2,and their recombinants were 47 and 32,respectively (Fig.4). For fine mapping of theSRL3gene,another 18 InDel markers between the abovetwo SSR markers were developed. Among these 18 InDel markers,six showed polymorphism between two parents. These six markers were subsequently used to analyze another 327 semi-rolled leaf individuals. The results showed that there was only one recombinant with the InDel markers M8 and M19 (Fig.4). So we speculated that theSRL3gene was localized within a 144-kb region between the markers M8 and M19. Based on the Rice Genome Annotation Project Database,14 candidate genes were found in this region. Four of them are expressed proteins,and the others encode different proteins:LOC_Os02g45750encodes a kinase,LOC_Os02g45760encodes a chitin-inducible gibberellinresponsive protein,LOC_Os02g45770is a MADS-box protein gene,LOC_Os02g45780encodes a C3HC4 type domain protein,LOC_Os02g45810is a WD40 repeat gene,LOC_Os02g45820encodes a mitochondrial import inner membrane translocase subunit Tim,LOC_Os02g45850encodes a B3 DNA-binding domain protein,LOC_Os02g45870encodes a tobamovirus multiplication protein,LOC_Os02g45890encodes a sulfotransferase domain protein,andLOC_Os02g45900encodes a tRNAsplicing endonuclease positive effector-related (Table 3).

Table 3 Function annotation of the candidate genes

Table 2 Genetic analysis of the semi-rolled-leaf trait in the mutant

3.6.The expression levels of leaf rolling-related and cell expansion-related genes

In order to determine whetherSRL3affects the expression levels of leaf rolling-related genes,17 genes were selected for qRT-PCR analysis. Firstly,we extracted total RNA from leaves of the wild type andsrl3mutant at the tillering stage. The results showed that the relative expression ofSLL1(a gene encoding a SHAQKYF class MYB transcription factor),SLL2,ADL1(a cysteine proteinase gene),SRL1(a gene encoding a glycosylphosphatidylinositol-anchored protein),andOsAGO7(a gene belongs to Argonaute family) were all notably up-regulated in thesrl3mutant (Shiet al.2007;Hibaraet al.2009;Zhanget al.2009;Xianget al.2012).On the contrary,the expression levels ofREL1(a gene coordinating BR signal to regulate leaf morphology),ACL2(a gene encoding a protein with unknown conserved domains),NAL7(a YUCCA gene),OsZHD1(a zincfinger transcription factor gene),OsZHD2(a zinc-finger transcription factor gene),andRL14(a gene encoding a functional unknown 2OG-Fe(II) oxygenase) were significantly decreased in thesrl3mutant (Fujinoet al.2008;Liet al.2010;Fanget al.2012;Xuet al.2014;Chenet al.2015) (Fig.5-A). The results indicated thatSRL3might be associated with these leaf rolling-related genes and regulate the morphological development of rice leaves.

We also found that the leaves were narrowed,so we investigated whether the expressions of some cell cycle,expansion and proliferation-related genes were influenced. The results exhibited that the cell cycle-related genes,CYCA2.3andMAPK,and the cell expansion-and proliferation-related genes,OsEXPA32andOsEXPB9,were up-regulated in thesrl3mutant. Moreover,the expression of cell cycle-related genes,CYCA2.1,CYCB2.2,CYCD4,CDKB,CDKA1,andCDKA2,and cell expansion-and proliferation-related geneOsEXPB10were all significantly decreased (Fig.5-B). These results suggested that the narrowing ofsrl3leaves may also be affected by the changes in the expression levels of these cell cycle,expansion-and proliferation-related genes.

Thesrl3mutant showed not only semi-rolled leaves but also narrowed leaves,and it elongated shoots and shortened roots in the seedling stage. The relationship between leaf rolling and indole acetic acid (IAA) content has been demonstrated in previous studies (Fujinoet al.2008),and auxin has been shown to act as a signal for cell elongation and cell division. Thus,the expression levels of some IAA-related genes were examined by qRTPCR. The total RNA was extracted from the shoots at the seedling stage and the leaves at the tillering stage. TheYUCCAgene family is involved in IAA biosynthesis,GH3.6andGH3.8are IAA inactivation-related genes,andARF25is an IAA signalling-related gene (Fujinoet al.2008;Maoet al.2020). The results exhibited that in thesrl3shoots at the seedling stage,the expression levels of several IAArelated genes,YUCCA1,YUCCA5,YUCCA6,andARF25,were notably increased,and the expression ofYUCCA2,YUCCA7,andGH3.6were significantly decreased compared with the wild type (Fig.6-A). In the leaves at the tillering stage,the expression levels ofYUCCA1,YUCCA2,YUCCA5,YUCCA6,YUCCA7,andARF25were significantly decreased,and the expression ofGH3.6andGH3.8were notably increased in thesrl3mutant (Fig.6-B).These results exhibited that both in the seedling and tillering stages,the expression levels of these IAA-related genes were altered,so we speculated that the alteration in shoot length and root growth of thesrl3seedlings and its leafrolling phenotype might also be related to the content of auxin in plants.

4.Discussion

Leaf rolling is a common and easily observable phenotype,so it has attracted more attention from breeders. However,leaf rolling is caused by multiple factors in plants (Hibaraet al.2009). In rice,leaves are made up of various kinds of cells,including the parenchyma,sclerenchymatous,mesophyll,and bulliform cells. The leaf rolling phenotype may be caused by the changes in the morphology of one or even more cell types(Zouet al.2014). According to previous research,the main factors leading to leaf rolling inlcude the alteration of the area of bulliform cells,the imbalance development between adaxial and abaxial epidermis,the loss of sclerenchymatous cells on the abaxial or adaxial surface,and the abnormal development of cuticle (Hibaraet al.2009;Zhanget al.2009;Wuet al.2011). In this study,a semi-rolled leaf mutant,srl3,was identified from the EMS mutation population. According to our paraffin sectioning and SEM observation,the number and area of bulliform cells on the adaxial surface ofsrl3leaves were notably decreased,and some sclerenchymatous cells were lost under the small veins. These results suggested that the rolling-leaf phenotype of thesrl3mutant was caused by changes in both bulliform and sclerenchymatous cells.

In general,leaf rolling results mainly from the dehydration of bulliform cells on the adaxial surface of rice leaves (O’Toole and Cruz 1980). It suggests that the number,area,and even the morphology of bulliform cells might affect leaf rolling. Excluding the influence of some environmental factors,it is important for us to study the genes involved in regulating the development of bulliform cells and further understand the mechanism of leaf rolling.At present,several genes related to bulliform cells have been reported. TheYABBYgene,which mainly exists in plants and is named after its amino terminal,contains a highly conserved YABBY domain. The YABBY family determines the establishment of the polarity of the adaxial-abaxial axis during the development of plants’lateral organs. TheYAB1gene is cloned from rice,and it participates in the biosynthesis feedback regulation of gibberellin (GA). In the leaves ofYAB1inhibited plants,the number of bulliform cells is notably increased,which leads to abaxial rolling,indicating that theYAB1gene is involved in controlling the division and differentiation of bulliform cells in rice leaves (Daiet al.2007). Thebrd1mutant has more bulliform cells than the wild type,but its leaves do not roll outward. This is because theBRD1gene also affects the development of other leaf structures,so it exhibits the dwarfism and severe leaf distortion phenotype (Honget al.2002). In the abaxial rolled-leaf mutantadl1,the number of bulliform cells on the adaxial side is notably increased and even up to a dozen,while there are only five in the wild type. A gene,ADAXIALIZED LEAF1(ADL1),is isolated and cloned;it ensures the establishment of adaxial-abaxial axis polarity of leaves by affecting the normal growth and development of leaf epidermis (Hibaraet al.2009). There is no tissue-specific expression of theACL1gene in the apical meristem,leaf primordium,and young leaves. However,in theACL1overexpression plants,the size of epidermal cells and the number of bulliform cells are both increased,which results in abaxial leaf rolling. It indicates that theACL1gene plays a vital role in the morphological development of rice leaves (Liet al.2010). TheLC2gene,which encodes VIN3-like protein,controls the erectness of leaves and is induced by auxin and other several hormones in plants. The mutation of theLC2gene affects the division of leaf epidermal cells and increases the number of bulliform cells. Therefore,thelc2mutant has not only an increased leaf angle but also abaxial-rolled leaves (Zhaoet al.2010). In this study,the phenotype of thesrl3bulliform cells also changed. Both number and area of bulliform cells were significantly decreased in thesrl3mutant compared with the wild type. And our results showed that the expression levels of several bulliformcell-related genes were significantly altered. Therefore,we speculated thatSRL3might regulate the area and morphology of bulliform cells by affecting the expression of other genes related to the development of bulliform cells. This still needs to be further explored.

Sclerenchymatous cells play a significant role in maintaining the normal morphology of leaves in rice. Sclerenchymatous cells are lignified dead cells differentiated from mesophyll cells,which are characterized by thickened secondary cell walls. The deficiency of sclerenchymatous cells on the abaxial side of the leaf leads to adaxial leaf rolling due to the lack of force provided by these cells,while the abnormal formation of adaxial sclerenchymatous cells is related to abaxially rolled leaves.SHALLOT-LIKE1(SLL1) is a typical sclerenchymatous cells-related gene,which controls the morphology of rice leaves by regulating the programmed cell death of abaxial sclerenchymatous cells. Thesll1mutant produces extremely adaxial rolled leaves because of the absence of sclerenchymatous cells on the abaxial surface of leaves (Zhanget al.2009). Thesrl2mutant is a semi-rolled leaf mutant,and its phenotype is also caused by the defect of abaxial sclerenchymatous cells.SRL2andSLL1play roles in different regulatory pathways of rice abaxial leaf development (Liuet al.2016). According to the SEM and paraffin sectioning analysis ofsrl3leaves,we found that the abaxial sclerenchymatous cells were absent under some small vascular bundles. It was because the sclerenchymatous cells provide mechanical strength to maintain the proper leaf morphology thatsrl3leaves showed adaxial rolling phenotype. The expression analysis showed that the expression of theSLL1gene was notably increased insrl3leaves compared with that of wild type. Our results and previous studies indicated that sclerenchymatous cells play a vital role in regulating leaf rolling in rice,and theSRL3gene participates in the formation of sclerenchymatous cells.

The rolling-leaf phenotype is also related to the IAA content in plants. InPetuniaandArabidopsis,FLOOZY and YUCCA genes encoding flavin monooxygenase-like proteins function in auxin biosynthesis;the mutations of these genes lead to the change of IAA level and the rolling-leaf phenotype (Zhaoet al.2001;Tobeña-Santamariaet al.2002). In rice,there are at least sevenYUCCA-likegenes encoding proteins similar to YUCCA proteins inArabidopsis(Yamamotoet al.2007).OsYUCCA1predominantly functions in various organs,and otherOsYUCCAgenes also take part in IAA biosynthesis,but their functions are very likely to be redundant (Yamamotoet al.2007). The role of GH3 proteins is to transform active IAA into its inactive form by binding IAA with amino acids (Staswicket al.2005).GH3.8encodes a protein that prevents the accumulation of free IAA. Overexpression ofGH3.8inhibits expansion expression and auxin signaling and causes abnormal plant morphology and stagnant growth and development(Dinget al.2008).OsARF25is an IAA response gene(Maoet al.2020). In our qRT-PCR analysis results,the expression levels of these IAA-related genes in seedlings and mature leaves were both significantly altered. In this study,thesrl3mutant showed abnormal phenotypes such as semi-rolled leaf,low seed-setting,elongated shoots,and shortened roots at the seedling stage,indicating that these abnormalities may also be associated with the transportation and distribution of IAA.

5.Conclusion

We characterized a rolling-leaf mutant,semi-rolled leaf3derived from YJ17 treated with EMS. In thesrl3mutant,the leaves were adaxially rolled and semirolled,the number and area of bulliform cells were significantly decreased,and the sclerenchymatous cells were absent on the abaxial surface under some small veins. During the seedling stage,thesrl3shoots were elongated,becoming short and thin. TheSRL3gene was finally mapped in a 144-kb region on chromosome 2,with 14 candidate genes in this region. Taken together,characterization and fine mapping of theSRL3gene help gain insight into the molecular mechanisms involved in leaf rolling in rice and may provide a new opportunity to improve light energy utilization. In addition,thesrl3mutant is perfect material for further studies on ideal plant type breeding.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (32071993 and 91735304).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm