• Open Access

Rolling-leaf14 is a 2OG-Fe (II) oxygenase family protein that modulates rice leaf rolling by affecting secondary cell wall formation in leaves

Authors


(fax 86-23-68250158; email hegh1968@yahoo.com.cn)

Summary

As an important agronomic trait, leaf rolling in rice (Oryza sativa L.) has attracted much attention from plant biologists and breeders. Moderate leaf rolling increases the amount of photosynthesis in cultivars and hence raises grain yield. Here, we describe the map-based cloning of the gene RL14, which was found to encode a 2OG-Fe (II) oxygenase of unknown function. rl14 mutant plants had incurved leaves because of the shrinkage of bulliform cells on the adaxial side. In addition, rl14 mutant plants displayed smaller stomatal complexes and decreased transpiration rates, as compared with the wild type. Defective development could be rescued functionally by the expression of wild-type RL14. RL14 was transcribed in sclerenchymatous cells in leaves that remained wrapped inside the sheath. In mature leaves, RL14 accumulated mainly in the mesophyll cells that surround the vasculature. Expression of genes related to secondary cell wall formation was affected in rl14-1 mutants, and cellulose and lignin content were altered in rl14-1 leaves. These results reveal that the RL14 gene affects water transport in leaves by affecting the composition of the secondary cell wall. This change in water transport results in water deficiency, which is the major reason for the abnormal shape of the bulliform cells.

Introduction

The leaf is the major organ of plant photosynthesis. The shape of the leaf determines the efficiency of light capture and energy conversion in individual plants and also determines plant structure. For rice and other crops that are cultivated in high density populations, moderate rolling of the leaves helps to maintain the erectness of the leaves, which can improve light acceptance, delay leaf senescence, accelerate the accumulation of dry matter and increase the yield (Lang et al., 2004). Appropriate leaf shape is an important characteristic of the super rice idiotype (a hybrid of Oryza sativa L. with desirable characteristics) (Yuan, 1997). The identification of mutants with rolling leaves and isolation of genes that control leaf rolling will be beneficial for breeding crops with the desired architecture.

Analysis of the some leaf development mutants of Arabidopsis thaliana, maize has shown that some mutations for leaf rolling are related to the development of the leaf along the adaxial–abaxial axis. Development of the leaf along the adaxial–abaxial axis is controlled by both transcription factors and small RNAs (Moon and Hake, 2010). Genes that belong to the HD-ZIP III family, such as PHABULOSA (PHB), PHAVOLUTA (PHV) (McConnell et al., 2001), REVOLUTE (REV) (Otsuga et al., 2001) and ROLLED1 (RLD1) (Nelson et al., 2002), determine the development of adaxial cells in leaves. In contrast, members of the KANADI family, such as milk-weed pod1 MILK-WEED POD1 (MWP1) (Candela et al., 2008), and genes that belong to the YABBY (YAB) family, such as YAB2 and YAB3 (Eshed et al., 2004), determine the development of abaxial cells. The two small RNAs, trans-acting short interfering RNAs (tasiRNA)and miR165/166, show opposing polar distributions in leaf primordial tissue and are thought to establish the adaxial–abaxial axis during leaf development (Nogueira et al., 2007).

In rice, ROLLED LEAF9 (RL9)/SHALLOT-LIKE1 (SLL1) encodes a KANADI family protein and is involved in the development of leaf abaxial cells (Yan et al., 2008; Zhang et al., 2009a). Furthermore, overexpression of the rice OsAGO7 gene induces upward curling of the leaf blade, but suppression of YAB1 induces the formation of rolled leaves and abnormalities in flower development (Dai et al., 2007; Shi et al., 2007).

In addition to defects in the establishment of leaf polarity, the effects of osmotic pressure or turgidity in bulliform cells are also important factors for the formation of rolled leaves (Alvarez et al., 2008; Hsiao et al., 1984). It is proposed that under water stress, bulliform cells lose turgor, which results in leaf rolling. When water stress is relieved, bulliform cells absorb water and swell up again (Price et al., 1997). Water and turgor potentials in mature leaves of the rice mutant CONSTITUTIVELY WILTED1 (COW1) are decreased significantly as compared with wild-type (WT) mature leaves (Woo et al., 2007). In NARROW AND ROLLED LEAF 1 (nrl1) mutants, the bulliform cells are smaller than those observed in the WT, which might account for the semirolled leaves that are seen in these mutants, rather than a defect in leaf polarity (Hu et al., 2010; Wu et al., 2010). However, the mechanisms that underlie leaf rolling in this type of mutant remain to be elucidated.

In the study reported herein, we report the isolation and functional characterization of a rice gene, RL14 that regulates leaf rolling. Protein analysis showed that RL14 coded for a protein of the 2OG-Fe (II) oxygenase family with unknown function, and we discuss the effect of this protein on the formation of the secondary cell wall and water transport.

Results

The RL14 mutant showed a phenotype of leaf rolling

Two allelic mutants were identified in a rice mutant population that had been generated by ethyl methanesulfonate mutagenesis. The mutants were designated as rl14-1 (Figure 1a) and rl14-2 on the basis of their incurved leaves (Figure 1b). Phenotypic observation showed that RL14 mutant plants had narrow and extremely rolled leaves, which appeared during the seedling stage. The rolling of leaves became more evident during plant growth. The leaf rolling index (LRI) of the top three leaves of the rl14-1 mutant ranged from 41% to 52%, at the heading stage (Figure 1c). The mutant rl14-2 had similar leaves to rl14-1.

Figure 1.

 The rl14 mutants have rolled leaves, with abnormal bulliform cells on the adaxial side. (a) Morphology of WT and rl14-1 plants. (b) Enlarged WT and rl14-1 leaves. (c) LRI of the top three leaves of rl14-1 at the heading stage. (d) Cross-section of a WT leaf; 1 and 2, close-up images of the insets in (d). (e) Cross-section of an rl14-1 mutant leaf; 3 and 4, close-up images of the insets in (e). Bulliform cells on the abaxial surfaces of rl14-1 leaves are samller and shrunken (black arrows). Scale bars = 500 μm in (d) and (e); and 50 μm in insets 1–4.

Cross-sections revealed no variation in the organization of the sclerenchyma and vascular tissues between leaves of the WT (Figure 1e) and those of rl14-1 (Figure 1f). However, in contrast to the WT, rl14-1 mutant leaves displayed dehydrated and atrophied bulliform cells in regions that were extremely rolled and smaller bulliform cells in regions that were slightly rolled (Figure 1f). In agreement with the histological analysis, scanning electron microscopy (SEM) did not reveal any change in the characteristics of leaf polarity in the abaxial–adaxial epidermis between the WT and rl14-1 (Figure 2a,b,c,d). However, we discovered that rl14-1 had significantly smaller stomata complexes than the WT had (Figure 2e). In view of the appearance of the bulliform cells and stomata, water status in rl14-1 leaves was analysed and compared with that in the WT. Stomata conductance and transpiration rate were significantly lower in the mutant than in the WT when plants were grown under full sunlight (Figure 3a,b). The rate of photosynthesis was slightly lower in the mutant than in the wild type, but the efficiency of water use was increased significantly (Figure 3c,d).

Figure 2.

 Scanning electron microscopy of WT and rl14 epidermis. rl14 had normal adaxial and abaxial characteristics, such as microhairs, but significantly smaller stomata complexes than the WT. (a) Abaxial epidermis of a WT leaf. (b) Adaxial epidermis of a WT leaf. (c) Abaxial epidermis of an rl14-1 leaf. (d) Adaxial epidermis of an rl14-1 leaf. (e) The area of the stomatal complexes (black box) was significantly smaller in the rl14-1 mutant than in the WT. Error bar represents SD (n > 30 in each measurement), and statistical analysis by a Student’s t-test indicated significant differences (*P < 0.05; **P < 0.01).

Figure 3.

 Transpiration rates and other photosystem indexes measured at the heading stage. (a) Stomatal conductance. (b) Transpiration rate. (c) Photosynthetic rate. (d) Water use efficiency. Error bar represents SD (n > 10 in each measurement), and statistical analysis by a Student’s t-test indicated significant differences (*< 0.05; **< 0.01).

RL14 encoded a 2OG-Fe (II) oxygenase family protein

The sterile line Xinong1A was crossed with rl14-1 or rl14-2. All F1 plants displayed the WT phenotype, and their F2 progeny all showed a segregation ratio of 3 : 1 (rl14-1, 5800 wild-type plant/1892 rolling leaf plant X< X20.05 = 0.603; rl14-2, 3129 wild-type plant/980 rolling leaf plant X< X20.05 = 2.84). Therefore, the rolled leaf phenotype in the rl14-1 and rl14-2 mutants was controlled by a single recessive nuclear gene.

The RL14 gene was mapped first to an interval between the simple sequence repeat (SSR) markers RM3123 and RM1162 on the long arm of chromosome 10 (Figure 4a). To narrow down the search for the gene that was affected in rl14-1 and rl14-2 mutant plants, three SSR markers and two InDel polymorphic sequence markers were developed between markers RM3123 and RM1162. Finally, the RL14 locus was mapped to a 24.5-kb DNA region between the InDel marker SID2 and SSR marker SW24 on a single bacterial artificial chromosome, AC079874 (Figure 4b).

Figure 4.

 Map-based cloning of RL14. (a) RL14 was mapped first between the SSR markers RM3123 and RM1162, at genetic distances of 7.5 and 9.6 cM, respectively. (b) The RL14 locus was localized further to a 24.5-kb region between markers SID2 and SW24. (c) Amplification of relevant DNA fragments and sequence comparison revealed that rl14-1 and rl14-2 resulted from a single base deletion in the 517-bp upstream of the start codon and a single nucleotide mutation at codon 190 (T/C) in the second exon, respectively. (d) qRT-PCR analysis of RL14 (Os10g40960.1) expression in WT, rl14-1 mutant and transgenic rl14-1 plants leaves. Experiments are biological replicates. (e) WT and rl14-1 mutant plants and transgenic rl14-1 plants in which expression of RL14 has been complemented (c-RL14). (f) Leaf shape of plants at the six-leaf stage. (g) Recovered bulliform cells in rl14-1 plants in which expression of RL14 has been complemented. Bars = 50 mm. qRT-PCR analysis of expression of RL14. RL14 expression in WT or transgenic rl14-1 plants and rl14-1 mutants were analysed. Experiments are biological replicates.

Four annotated candidate genes, which encoded a polyol transporter (Os10g40950.1), a 2OG-Fe (II) oxygenase family protein (Os10g40960.1), a flavonol synthase (Os10g40990.1) and a retrotransposon protein (Os10g41000), were located in the 24.5-kb DNA region (Institute for Genomic Research; http://rice.plantbiology.msu.edu/). Further amplification of the relevant DNA fragments and sequence comparison revealed differences between the rl14-1 and rl14-2 alleles in the gene that encoded the 2OG-Fe (II) oxygenase family protein. The mutant rl14-1 carried a single base deletion (A) in the 517-bp upstream of the start codon, and rl14-2 had a single base substitution at codon 190 (T/C) in the second exon, which resulted in an amino acid change from Ile (I) to Thr (T) (Figure 4c). In addition, quantitative real-time PCR (qRT-PCR) showed that the level of expression of the 2OG-Fe (II) oxygenase gene was reduced significantly in rl14-1 (Figure 4d). Therefore, we tentatively designated the gene that encoded the 2OG-Fe (II) oxygenase family protein as the RL14 gene.

The physiological effect of RL14 in leaf rolling was confirmed by genetic complementation analysis. Phenotypic observation of transgenic rl14-1 plants confirmed that complementary expression of RL14 rescued leaf rolling with restoration of normal leaf shape at the sixth leaf stage (Figure 4e,f). In addition, observations of cross-sections of leaves from transgenic rl14-1 plants indicated that normal bulliform cells were formed at the adaxial epidermis (Figure 4g). Independent transgenic lines were identified through qRT-PCR analysis of RL14 expression using primers that matched sequences in the coding region. The results showed that the expression of RL14 was restored to the WT level in the transgenic plants (Figure 4d).

Comparison with the corresponding genomic sequence revealed that the RL14 gene consisted of three exons and two introns, and the coding sequence was 834-bp long and encoded a protein of 278 amino acids (Figure 5a). Protein sequence analysis revealed that the RL14 protein had a 2OG-Fe (II) oxygenase domain (amino acids 128–228) (Figure 5a; http://swissmodel.expasy.org/workspace/index) (Arnold et al., 2006). The mutation in rl14-2 was within this domain, which indicated the importance of the 2OG-Fe (II) oxygenase domain for the function of RL14. When a BLAST search with the full RL14 protein sequence was carried out on the Swiss Institute of Bioinformatics website (http://www.expasy.ch/tools/blast/), a large number of proteins that contained the 2OG-Fe (II) oxygenase domain were found. We then analysed the possible phylogenetic relationships between RL14 and its related proteins. RL14 was related most closely to members of the 2OG-Fe (II) oxygenase family from rice and sorghum (Figure 5b). Orthologous of RL14 was identified in other species, such as maize, Arabidopsis, Papaver somniferum, Vitis spp and Ricinus communis (Figure 5b).

Figure 5.

 Phylogenetic analysis and subcellular localization of RL14 protein. (a) Protein sequence analysis showed that RL14 contained a 2OG-Fe (II) oxygenase domain (residues 128–228). (b) Phylogenetic relationships among the members of the 2OG-Fe (II) oxygenase family. The numbers at each node represent the bootstrap support (percentage). The scale bar indicates the genetic distance based on branch length. (c) Subcellular location of the RL14GFP fusion protein in onion epidermal cells. (1) Onion epidermal cell expressing free GFP showed fluorescence in the nucleus and cytoplasm. (2) Onion epidermal cell expressing RL14::GFP showed fluorescence in the nucleus and cytoplasm. Scale bars = 50 μm.

To test whether RL14 localized, we generated a construct in which the full-length cDNA for RL14 was fused to the coding sequence of green fluorescent protein (GFP) and monitored the fluorescence of the transiently expressed RL14::GFP fusion protein and GFP alone in onion epidermal cells. As for GFP alone (Figure 5c, panel 1), RL14::GFP fluorescence was detected more strongly in the cytoplasm than in other parts of the cell (Figure 5c, panel 3).

Expression pattern of RL14

To study the expression pattern of RL14, qRT-PCR was performed with different WT tissues at the fourth leaf stage and heading stage. RL14 was expressed well in mature leaves (4L), leaf sheaths and roots. However, it was expressed at a very low level in leaves that were still wrapped in the sheath (5L), in only trace amounts in the stem and flowers, and not at all in shoot apical meristem (SAM) (Figure 6a).

Figure 6.

 Analysis of the expression pattern for the RL14 gene. (a) qRT-PCR analysis of RL14 expression in various tissues. Expression of RL14 in roots, SAM, 5th leaf (leaf still wrapped in leaf sheath), 4th leaf (middle part of 4th leaf), leaf sheath of 4th leaf, flag leaf, flowers and stems at the heading stage. For SAM samples, the precisely defined intact SAM and young leaf primordial (without the leaf sheath) were collected. Experiments were biological replicates. (b) Promoter–reporter gene (GUS) fusion studies on the expression of RL14. Expression of RL14 in the root (1), stem (2), leaf (3), leaf sheath (4) and glume (5) was analysed. Bars = 500 μm. (c) In situ hybridization analysis of the expression of RL14 in the underemerged leaf and mature leaf. The RL14 signal was detected in sclerenchymatous cells (black arrows) (1); the inset in (1) is enlarged in (2). In the mature leaf, the RL14 signal was detected in nearly all mesophyll cells (black arrows) (3). Bars = 50 μm.

To assess the expression pattern comprehensively, β-glucuronidase (GUS) activity was examined histochemically in transgenic plants that carried an RL14 promoter–GUS reporter gene. Histochemistry, as well as qRT-PCR analysis, showed that RL14 was highly transcribed in leaves, leaf sheaths and roots. It was also detected in stems and glumes, although its expression in these organs was not as intense as that in the mature leaves, leaf sheaths and roots (Figure 6b).

We further examined the spatial and temporal localization of RL14 during leaf development by in situ hybridization. Expression of RL14 mRNA was detected in the sclerenchymal cell layer of underemerged leaves and sheaths. In the mature leaf, RL14 mRNA accumulated mainly in the mesophyll cells that surrounded the vasculature (Figure 6c).

RL14 affected formation of the secondary cell wall

On the basis of the expression pattern of RL14, we conducted qRT-PCR analysis to determine the expression in rl14-1 of rice TF genes, lignin and cellulose biosynthesis genes for secondary cell wall formation. The expression of SND1 and VND4/5/6 (Ambavaram et al., 2011; Zhong et al., 2008), which are key TF genes for activating the developmental programme of secondary wall biosynthesis, was up-regulated in rl14-1 (Figure 7a). The MYB TF gene MYB58/63 and MYB20 (Ambavaram et al., 2011) targeting lignin and cellulose genes was significantly down-regulated and slightly up-regulated, respectively (Figure 7a). Accordingly, OSLAC4 and OSLAC17 the homology gene of Arabidopsis LAC4 and LAC17 which contribute to the constitutive lignification (Berthet et al., 2011) were down-regulated, but the OSCEA4 that plays a critical role in secondary cell wall cellulose biosynthesis (Xiong et al., 2010) was up-regulated in rl14-1 (Figure 7a). We also determined the abundance of the major constituents of secondary cell walls and showed that cellulose content was increased, but lignin content was decreased in rl14-1 (Figure 7b).

Figure 7.

 RL14 affected the formation of the secondary cell wall. (a) Relative expression of putative cell wall TF genes, cellulose and lignin biosynthesis genes in rl14-1 leaf compared with that in the wild type. Experiments were biological replicates. (b) Cellulose and lignin content in rl14-1 leaves. Error bar represents SD (n > 5 in each measurement), and statistical analysis by Student’s t-test indicated significant differences (*< 0.05; **< 0.01).

Discussion

Abnormal bulliform cells, rather than defects in leaf polarity, account for rolled leaves in the rl14 mutants

As a polymorphic crop, varieties and mutants of rice show several types of leaf rolling, including inward and outward rolling. According to previous research, two major factors affect leaf shape: one is the establishment of polarity and cell differentiation, and the other is physiological factors. Rice sll1 (Zhang et al., 2009a) and adaxialized leaf1 (adl1) (Hibara et al., 2009) are two leaf rolling mutants that are caused by alterations in leaf polarity. The sll1 mutants show altered leaf polarity in the abaxial epidermis and a lack of sclerenchymatous cells on the abaxial surface in regions of leaf curving. Leaf blades from adl1 show ectopic bulliform-like cells in the abaxial epidermis. In maize (Zea mays), the mwp1 mutant has adaxialized sectors in the sheath; the proximal part of the leaf. Ectopic leaf flaps develop in the mwp1 mutant where adaxial and abaxial cell types juxtapose (Candela et al., 2008). Using SEM and histological analysis of rl14 leaves, we did not find any obvious alteration in leaf polarity in these mutants (Figures 1d,e, and 2). However, we found that rl14-1 mutant leaves showed serious shrinkage and small bulliform cells. Some researchers think bulliform cells shape is related closely to the water content (Alvarez et al., 2008). Research on the rice cow1 mutant has shown that shrunken bulliform cells are related to a decreased rate of transpiration and turgor potential (Woo et al., 2007). In the abaxial leaf curling mutant by240, bulliform cells are exaggerated and increased, but under conditions of water deficiency, abaxially rolled leaves in by240 flatten gradually and curl adaxially (Li et al., 2010). Furthermore, genes related to the development of bulliform cells, such as ACL1 and ACL2, were down-regulated in rl14-1 mutants (Figure S1). Therefore, abnormal bulliform cells, rather than defects in leaf polarity, account for the rolled leaves in the rl14 mutants.

The RL14 gene affects water transport in leaves indirectly by affecting the composition of leaf secondary cell walls

In WT rice plants, mesophyll cells are specified during the early stages of leaf development, when the young leaves remain wrapped inside the sheath in a crimped state. At plastochron 5, the differentiation of sclerenchymatous cells is visible; however, these cells have not yet developed mechanical strength, which requires the initiation of PCD and thickening of the secondary cell walls (Kawakatsu et al., 2006; Zhang et al., 2009a). In situ hybridization analysis indicated that expression of the RL14 gene correlated with the formation of the secondary cell wall in sclerenchymatous cells (Figure 6c). RL14 expression was only detected in mesophyll cells in mature leaves, because the secondary cell wall is only laid down in cells that have ceased growth (Brett and Waldron, 1990). Expression of TF genes related to secondary cell wall formation was altered, the expression of genes related to cellulose biosynthesis was increased, and the genes related to lignin biosynthesis were repressed in the rl14 mutants (Figure 7a). As a consequence, the cellulose and lignin content were changed significantly in rl14-1 mutant leaves (Figure 7b). These results show that RL14 is linked closely with the synthesis of the secondary cell wall in rice.

In plant leaves, water molecules are drawn from the non-living tracheids and vessels of the xylem through the living cells of the leaf mesophyll (middle layer) to the surface of the mesophyll cell walls in a process that is driven by transpiration. Some researchers have shown that water content and water loss in plant leaves might be affected by the composition and structure of the cell walls. For example, a mutant of Arabidopsis CesA7, which is involved in the synthesis of cellulose during the formation of the secondary cell wall, shows increased transpiration efficiency (TE), and there is evidence that this effect is because of the presence of smaller stomatal pores and reduced water transport because of collapsed xylem elements, as compared with the WT (Liang et al., 2010). Other Arabidopsis mutations that impair cell wall biosynthesis might also increase TE. The leaf wilting 3 mutant (lew3), which encodes a putative α-1,2-mannosyltransferase (ALG11), causes disruption of cellulose synthesis, a partially collapsed xylem, and reduced water loss from detached leaves (Zhang et al., 2009b). The rice NRL1 gene encodes a cellulose synthase-like protein D4 (OsCslD4). The phenotype of bulliform cells in nrl1 mutants supports the notion that OsCslD4 plays a crucial role in the development of vascular tissue, probably affecting cell wall biosynthesis (Hu et al., 2010). In addition to the changes in cellulose and lignin content (Figure 7b), rl14-1 showed smaller stomatal complexes (Figure 2e) and significantly decreased stomata conductance and transpiration rate relative to the WT (Figure 3a,b), as well as cow1 and AtCesA7irx3-5 mutants. Thus, we conclude that the RL14 gene affects water transport in leaves by affecting the composition and structure of leaf secondary cell walls and that water deficiency is the major reason for the abnormal shape of the bulliform cells.

Experimental procedures

Plant materials and growth conditions

The rl14 rice mutants were isolated from a population of the O. sativa L. ssp. Indicia restorer line Jinghui10 that had been treated with 1% ethyl methanesulfonate solution. Jinghui10 plants represent the WT. The mutants rl14-1 and rl14-2 were crossed with the sterile line Xinong1A (with flat leaves). The resultant F1 plants were self-crossed to produce the F2 seeds needed to construct the F2 mapping population. Rice plants were cultivated in an experimental field at the Southwest University Rice Research Institute under natural growing conditions. Field management adhered to normal agricultural practice. To grow transgenic rice, rice seeds were germinated in sterilized water and then grown in pots in a phytotron with a 12-h light and 12-h dark cycle.

Histological analysis

To prepare paraffin sections, samples were fixed in FAA (10% formaldehyde, 5% acetic acid, 50% ethanol) for 24 h at 4 °C, dehydrated in a graded ethanol series and embedded in Paraplast plus (Sigma-Aldrich, Co., St. Louis, MO). Microtome sections (8-μm thick) were stained with Safranin and Fast green.

SEM

Leaves from WT and rl14-1 plants were collected and fixed in formalin: glacial acetic acid: 50% ethanol (2 : 1 : 17, v : v : v) overnight. After dehydration in a graded ethanol series, the sections were critical point dried for 4 h (Hitachi), surface sprayed with gold powder and examined under a scanning electron microscope (Hitachi S-450). Measurements of guard cells were processed and analysed with the Image Tool software (http://ddsdx.uthscsa.edu/dig/itdesc.html).

Determination of transpiration rate and LRI

At the heading stage, stomatal conductance and the rate of transpiration were measured with a steady-state porometer (Li-6400; LICOR Biosciences, Lincoln, NE, USA). The widths of the top three leaves of rl14-1 plants at the heading stage were measured under the natural (Ln) or unfolding state (Lw). The LRI was calculated as LRI = (Lw−Ln)/Lw (Zhu et al., 2001).

Phylogenetic analysis

The BLAST search program (http://www.expasy.ch/tools/blast/) was used to identify sequences homologous to RL14. All the resultant sequences were aligned using the ClustalX1.83 multiple alignment mode, and a neighbour-joining phylogenetic tree was generated using MEGA4.0 (Tamura et al., 2007). The bootstrap values for nodes in the phylogenetic tree were from 1000 replications. The handling gap option was pairwise deletion, and the numbers at the branching points indicated the bootstrap values.

Map-based cloning of RL14

RL14 was mapped primarily with SSR markers (http://www.gramene.org/microsat/ssr.html) using 140 F2 mutant plants. To fine map the RL14 gene, we developed 39 SSR markers and two insertion/deletion (InDel) markers between the markers RM3123 and RM1162. SSR markers in the target region were obtained from the website http://www.gramene.org/ or identified by SSR Hunter, whereas InDel markers were designed using the Vector NTI software after the comparison of sequences from Xinong1A and the mutant. Among these (Table S1), five, three SSR markers and the two InDel markers, showed polymorphism between the mutant and Xinong1A. These five markers, together with RM3123 and RM1162, were used subsequently to analyse 1892 mutant-type F2 individuals from the population obtained by crossing Xinong1A with rl14-1 and 980 mutant-type F2 individuals from the population obtained by crossing Xinong1A with rl14-2. The RL14 locus was localized further within a 24.5-kb region between markers SID2 and SW24.

To define molecular lesions, a 24.5-kb region of genomic DNA was amplified by PCR from rl14 and the appropriate WT variety (Jinghui10). All PCR products were sequenced and the candidate gene was amplified from both rl14 and Jinghui10 genomic DNA using different primers (Table S1). The sequences obtained were analysed with the Vector NTI software.

Constructs and rice transformation

For complementation of the rl14 mutation, the full-length cDNA for RL14 was amplified by RT-PCR using the primers RL14exF and RL14exR (Table S1) from the WT. The primers were designed to incorporate a BamHI site at the N-terminal end and a SacI site at the C-terminal end of the open reading frame. PCR products were cloned into the pMD19-T vector (TaKaRa Biotechnology (Dalian) Co. Ltd., Dalian, Liangning, China). The WT RL14 cDNA fragment was digested with BamHI and SacI and ligated into the BamHI and SacI sites of the vector PTCK303. The resultant PTCK-RL14 plasmid, which contained the RL14 coding sequence driven by the ubiquitin promoter, was transformed into Agrobacterium tumefaciens strain LBA44O4 by electroporation and used to transform rl14-1 for complementation testing in accordance with a published method (Xiao et al., 2009).

Promoter–reporter gene fusion studies

For promoter–GUS fusion studies, a 2.1-kb genomic DNA fragment that contained the promoter region of the RL14 gene was amplified by PCR using primers RL14PF and RL14PR (Table S1) and then subcloned into vector pCAMBIA1301 (Cambia), which resulted in a fusion of the RL14 promoter and the GUS reporter gene. The construct was transformed into WT plants as described earlier. About nine independent transgenic lines were obtained after screening, and GUS activity was detected histochemically at the heading stage as described previously (Jefferson et al., 1987).

In situ hybridization analysis

Leaves were collected from fourth leaf stage of wild type and fixed in FAA (10% formaldehyde, 5% acetic acid, 47.5% ethanol) overnight at 4 °C. Samples were then dehydrated through a butanol series and embedded in Paraplast Plus (Sigma–Aldrich, http://www.sigmaaldrich.com/). Sections that were 8 μm thick were obtained using a Leica RM2135 microtome (Leica Biosystems, http://www.leica.com). To prepare the RL14 probe, a 352-bp fragment of RL14 cDNA was amplified using primers RL14 T7 and RL14 sp6 (Table S1). The probe was synthesized using a DIG RNA labeling kit (SP6/T7; Roche Diagnostics Ltd, http://www.roche.com) in accordance with the manufacturer’s recommendations. Pretreatment of sections, hybridization and immunological detection was performed as described previously (Xiao et al., 2009).

Subcellular localization of the RL14 protein

The subcellular localization of the RL14 protein was verified. The full-length RL14 coding sequence was amplified using primers RL14gF and RL14gR2 (Table S1), and the full-length coding sequence for GFP was amplified using primers gfpF and gfpR (Table S1). The two DNA fragments were cloned into vector PTCK303 to create PTCK-RL14GFP. PTCK-RL14GFP or CaMV35S-sGFP was transformed into onion epidermis by Agrobacterium-mediated transformation. After 48 h, fluorescence was visualized under a fluorescence microscope.

Real-time qRT-PCR analysis

Real-time qRT-PCR analysis was performed to examine the expression pattern of RL14 in various tissues and at different stages of leaf growth, to identify the transgenic lines in which RL14 expression was enhanced and complemented. Transcripts from genes related to secondary cell wall and the development of bulliform cells that were altered in rl14-1 plants as compared with the WT were also identified. Total RNA was extracted using TRIzol solution and reverse transcribed using the Superscript Preamplification System in accordance with the manufacturer’s instructions and then analysed quantitatively on a Bio-Rad CYF96 using the real-time PCR Master Mix (TaKaRa Biotechnology (Dalian) Co. Ltd.). The rice ACTIN gene was amplified and used as an internal standard to normalize the expression of RL14 and the other genes tested. The primers used to test the expression of RL14 and the other genes are listed in Table S2.

Measurement of carbohydrate and lignin

Carbohydrate was assayed as described previously (Updegraff, 1969). The leaves were ground into a fine powder in liquid nitrogen. The powder was washed in phosphate buffer (50 mm, pH 7.2) three times, extracted twice with 70% ethanol at 70 °C for 1 h and dried under vacuum. The dried cell wall material was assayed for cellulose content using the anthrone reagent, with Whatman 3 MM paper as the standard. After vacuum drying, lignin content was quantified in accordance with the method described by (Zwiazek, 1991).

Acknowledgements

This work is supported by the National Natural Science Foundation of China: Cloning and functional analysis of gene URL2(t) for unilateral rolling leaf in rice (31071480) and Genetically modified organisms breeding major projects of china: Cloning and functional analysis of three genes for high chlorophyll content, unilateral rolling leaf and sheated panicles in rice (2009ZX08009-109B), by Project supported by the Found for Distinguished Young Scholars of Chongqing: gene cloning and function analysis of the important trait in rice (Grant No. CSTC, 2008BA1033), by Chongqing research projects: the cultivation of stress resistance transgenic novel material of Rice (Oryza sativa L.) and cucumber (Cucumis sativus L.) (CSTC, 2010AA1019).

Ancillary