![Leaf Rolling Controlled by the Homeodomain Leucine Zipper Class IV Gene Roc5 in Rice1[W]](http://data.docslib.org/img/f818916487f129bf2e44a96f44746d5e-1.webp)
Leaf Rolling Controlled by the Homeodomain Leucine Zipper Class IV Gene Roc5 in Rice1[W]
Liang-ping Zou, Xue-hui Sun, Zhi-guo Zhang, Peng Liu, Jin-xia Wu, Cai-juan Tian, Jin-long Qiu2, and Tie-gang Lu2* Biotechnology Research Institute/National Key Facility for Genetic Resources and Gene Improvement, Chinese Academy of Agricultural Sciences, Beijing 100081, People’s Republic of China (L.-p.Z., X.-h.S., Z.-g.Z., J.-x.W., T.-g.L.); and State Key Laboratory of Plant Genomics, Institute of Microbiology, Chinese Academy of Sciences, Beijing 100101, People’s Republic of China (P.L., C.-j.T., J.-l.Q.)
Leaf rolling is considered an important agronomic trait in rice (Oryza sativa) breeding. To understand the molecular mechanism controlling leaf rolling, we screened a rice T-DNA insertion population and isolated the outcurved leaf1 (oul1) mutant showing abaxial leaf rolling. The phenotypes were caused by knockout of Rice outermost cell-specific gene5 (Roc5), an ortholog of the Arabidopsis (Arabidopsis thaliana) homeodomain leucine zipper class IV gene GLABRA2. Interestingly, overexpression of Roc5 led to adaxially rolled leaves, whereas cosuppression of Roc5 resulted in abaxial leaf rolling. Bulliform cell number and size increased in oul1 and Roc5 cosuppression plants but were reduced in Roc5-overexpressing lines. The data indicate that Roc5 negatively regulates bulliform cell fate and development. Gene expression profiling, quantitative polymerase chain reaction, and RNA interference (RNAi) analyses revealed that Protodermal Factor Like (PFL) was probably down-regulated in oul1. The mRNA level of PFL was increased in Roc5-overexpressing lines, and PFL-RNAi transgenic plants exhibit reversely rolling leaves by reason of increases of bulliform cell number and size, indicating that Roc5 may have a conserved function. These are, to our knowledge, the first functional data for a gene encoding a homeodomain leucine zipper class IV transcriptional factor in rice that modulates leaf rolling.
Leaf functions, such as photosynthesis, respiration, through conventional genetic screening, and rl7 to and transpiration, are dependent on leaf shape or three- rl12 were mapped to chromosomes 5 (rl7 and rl8;Li dimensional architecture (Govaerts et al., 1996; Zhang et al., 2000), 7 (rl11; Shi et al., 2009), 9 (rl9 [Yan et al., et al., 2009). Leaf shape has long been considered an 2006] and rl10 [Luo et al., 2007]), and 10 (rl12; Luo et al., important agronomic trait in rice (Oryza sativa; Yuan, 2009). However, only a few mutant genes have been 1997). Moderate leaf rolling in rice leads to erect leaf cloned and characterized, and most of these leaf-rolling canopies and higher photosynthetic efficiency, improv- genes are associated with leaf adaxial-abaxial polarity ing stress responses by reducing transpirational water establishment. RL9 and SHALLOT-LIKE1 (SLL1) are the loss and radiant heat absorption (Lang et al., 2004; same gene (Yan et al., 2008; Zhang et al., 2009), and SLL1 Zhang et al., 2009), thereby increasing grain yield. encodes a SHAQKYF class MYB family transcription Therefore, moderate leaf rolling is an ideal trait for factor belonging to the KANADI family. Defective rice breeding (Price et al., 1997). To date, 12 rice mutants development of sclerenchymatous cells on the abaxial with rolled leaves (rl) have been isolated and reported side of sll1 mutant leaves leads to extreme leaf rolling. in several different studies, for which six genes (rl1–rl6) SLL1 deficiency also leads to defective programmed cell were mapped on corresponding rice chromosomes death of abaxial mesophyll cells and suppresses the development of abaxial features (Zhang et al., 2009). 1 This work was supported by the National High-Tech Research The rice adaxialized leaf1 (adl1) mutant has abaxially and Development Project “China Rice Functional Genomics” (proj- rolled leaves. ADL1 is an ortholog of maize DEFECTIVE ect nos. 2006AA10A101 and 2007AA10Z104), the National Science KERNEL1 (Becraft et al., 2002), which encodes a plant- Foundation of China (project no. 30971842), and awards for excellent specific calpain-like Cys protease and is required for researchers from the Chinese Academy of Agricultural Science (to establishment of the adaxial-abaxial axis in leaf primor- T.-g.L.) and the 100 Talents Program of the Chinese Academy of dia by promoting proper epidermal development, es- Sciences (to J.-l.Q.). 2 pecially in bulliform cells (Lid et al., 2002; Hibara et al., These authors contributed equally to the article. 2009). The leaves of adl1 mutants are covered with * Corresponding author; e-mail [email protected]. bulliform cells on both adaxial and abaxial sides The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy (Hibara et al., 2009). Bulliform cells are specialized described in the Instructions for Authors (www.plantphysiol.org) is: epidermal cells on the adaxial leaf blade surface in all Tie-gang Lu ([email protected]). monocotyledonous orders except Helobiae (Metcalfe, [W] The online version of this article contains Web-only data. 1960; Jane and Chiang, 1991). Shrinkage of bulliform www.plantphysiol.org/cgi/doi/10.1104/pp.111.176016 cells in the adaxial epidermis near the midrib has been
Ò Plant Physiology , July 2011, Vol. 156, pp. 1589–1602, www.plantphysiol.org Ó 2011 American Society of Plant Biologists 1589 Zou et al. linked to leaf rolling in rice and other grasses (O’Toole ference (RNAi) analyses revealed that Protodermal Factor et al., 1979; Kadioglu and Terzi, 2007). Overexpression Like (PFL) is probably down-regulated in oul1, indicating of Abaxially Curled Leaf1 (ACL1) and its homolog ACL2 that Roc5 may have a conserved function. Our findings in rice leads to increased bulliform cell number and provide, to our knowledge, the first functional data for a size and thus to leaf epidermal cell expansion, result- gene encoding an HD-Zip IV transcriptional factor in ing in abaxial leaf curling (Li et al., 2010). However, rice and its role in modulating rice leaf rolling. direct genetic and molecular evidence is lacking for the involvement of bulliform cells in leaf rolling. Homeodomain Leu zipper (HD-Zip) proteins are RESULTS plant-specific transcription factors grouped into HD- Isolation and Phenotypic Characterization of the Rice Zip families I to IV based on sequence (Elhiti and oul1 Stasolla, 2009). Sixteen genes belong to the Arabidopsis Mutant (Arabidopsis thaliana) HD-Zip IV gene family (Abe et al., To identify new genes modulating leaf rolling in rice, 2003). HD-Zip IVis also referred to as HD-GL2, because we screened more than 100,000 T-DNA insertion lines in GLABRA2 (GL2) was the first gene identified in this rice (var Nipponbare) from several populations (Yang family (Rerie et al., 1994). HD-Zip IV genes occur in et al., 2004; Peng et al., 2005; Wan et al., 2009). A total of species other than Arabidopsis, including maize (Zea 312 individual mutant lines showing leaf-rolling phe- mays; Ingram et al., 2000), rice (Ito et al., 2002), cotton notypes were isolated, identifying the oul1 mutant with (Gossypium hirsutum; Guan et al., 2008), and pine (Pinus leaves rolling toward the abaxial surface. In seedlings, spp.; Ingouff et al., 2001, 2003). The vast majority of HD- oul1 leaves gradually curved toward the abaxial surface, Zip IV genes are specifically expressed in the outer cell displaying a slight spiral pattern in the middle of the layer of plant organs (Vernoud et al., 2009). Identifica- leaf blade (Fig. 1, A and C). Leaf rolling became more tion of Arabidopsis GL2, ANL2, ATML1, and PDF2 from evident during growth and finally became shallot like their corresponding mutants has highlighted the in- (Fig.1,B,D,andF).Notably,shallot-like1 leaves roll volvement of the HD-Zip IV transcription factors in the adaxially (Zhang et al., 2009). Accordingly, leaf rolling differentiation and maintenance of epidermal cell fate index values (LRIs) in the oul1 mutant were 0.77 to 0.93 (Rerie et al., 1994; Lu et al., 1996; Kubo et al., 1999; Soppe at later vegetative stages, whereas the LRIs in the flat et al., 2000; Ohashi et al., 2002; Abe et al., 2003). The wild-type leaves remained 0 (Fig. 1, E and H). functions of HD-Zip IV genes in monocots have only The leaf erection index values (LEIs) in oul1 were been shown for maize Outer Cell Layer1 (OCL1) and significantly higher than in the wild type and re- OCL4. ZmOCL1 may be involved in the specification of mained constant from the sixth leaf stage to the embryo protoderm identity, the organization of the flowering stage (Fig. 1I), suggesting that the abaxially primary root primordium, the maintenance of the L1 rolled leaves could enhance leaf erection. Leaf erection cell layer in the shoot apical meristem, and kernel can optimize canopy light transmission in the middle development (Ingram et al., 1999; Khaled et al., 2005). and later growing stages, thereby increasing photo- ZmOCL4 is expressed in the leaf blade epidermis and synthetic efficiency and consequently increasing grain inhibits trichome development in maize (Vernoud yields (Price et al., 1997; Lang et al., 2004; Zhang et al., et al., 2009). Therefore, it is intriguing to determine 2009). The stomatal conductance and photosynthetic whether the function of HD-Zip IV genes is evolution- rate of oul1 were significantly higher than in the wild arily conserved in other monocots. BLAST searches type (Table I). Interestingly, unlike most leaf-rolling indicated that there are nine GL2-type Rice outermost mutants (Lang et al., 2004), oul1 exhibited a higher cell-specific (Roc) genes in the rice genome (Ito et al., transpiration rate compared with the wild type (Table 2003). Full-length cDNAs for Roc1 to Roc5 have been I). oul1 had longer spikelets (Fig. 1G) and a higher cloned, and all five genes are specifically expressed in spikelet number per panicle (136) than in the wild type the rice epidermis with somewhat different temporal (106; Fig. 1K), although oul1 had a lower seed-setting patterns (Ito et al., 2002, 2003). However, the biological rate (64%) than in the wild type (88%; Fig. 1L). These functions of these Roc genes in rice morphogenesis and results indicated that OUL1 affected rice yields development remain to be determined. through photosynthesis and other factors. In a large-scale screen of a rice T-DNA insertion population, we isolated the leaf-rolling mutant outcurved Bulliform Cell Number and Size Are Increased in oul1 leaf1 (oul1). Here, we report that the phenotypes of oul1 are caused by T-DNA insertion in Roc5, a member of To investigate the formation of abaxial leaf rolling in class IV HD-Zip genes (Ito et al., 2003). Knockout of Roc5 oul1, cross sections from mature 10th leaves were led to increased bulliform cell number and size on the analyzed. In the wild type, bulliform cells occurred adaxial leaf blade surface. Interestingly, Roc5 overex- between two vascular bundle ridges in parallel with pression gave rise to opposite phenotypes of oul1,and the more adaxially localized veins. In cross sections, cosuppression lines displayed abaxially rolled leaves wild-type bulliform cells were typically arranged in like oul1. The data support a conserved function for HD- groups of 4 6 1 cells, with the middle cells larger than Zip IV family genes in dicots and monocots. Gene those on either side (Fig. 2, C and E). The oul1 mutant expression profiling, quantitative PCR, and RNA inter- had groups of 7 6 1 bulliform cells located between
1590 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Figure 1. Characterization of oul1 and wild-type (wt) morphology. A and C, At the five-leaf stage, oul1 leaves spiraled slightly in the middle of the leaf blade, whereas wild-type leaves were flat. B and D to F, In mature plants (B), the leaves rolled abaxially to form a cylinder-like shape in oul1 plants com- pared with the flat wild-type leaves (D– F). ab, Abaxial; ad, adaxial. Bars = 1 cm (A), 10 cm (B), 1 mm (C), and 5 mm (D–F). G and J, The spikelet in oul1 was longer than in the wild type. H and I, LRIs (H) and LEIs (I) of the wild type and oul1 are shown. K and L, The spikelet number per panicle (plump + empty) in oul1 (n = 136) was higher than in the wild type (n = 106; K), whereas oul1 had a lower (66%) seed- setting rate than the wild type (88%; L).
two vascular bundle ridges (Fig. 2, D and E). In cross (Zhang et al., 2009) or epidermal cells (Hibara et al., sections taken from similar positions, the average bulli- 2009). Chloroplast grana lamellae were disordered and form cell area was significantly larger in oul1 (5,128.21 6 irregular in rl9-1 mutant leaves (Yan et al., 2008), but 1,166.33 mm2) than in the wild type(2,298.55 6 447.26 cross sections of oul1 showed no significant difference mm2; Fig. 2F). Bulliform cells mainly consist of water- in organelles, including chloroplast grana lamellae filled vacuoles, and as expected, the water content (Supplemental Fig. S2). of the oul1 leaves was higher than in the wild type Toluidine blue O stains bulliform cells (Hernandez (Fig. 2G). et al., 1999). The adaxial mature 10th leaf surfaces of oul1 The bulliform cell phenotype was obvious through- and wild-type plants were peeled, and epidermal and out the whole oul1 mature blade, with no differences bulliform cells were stained blue and purple, respec- along the proximodistal and centrolateral axes (Sup- tively, with toluidine blue O (Fig. 2, A and B), again plemental Fig. S1). Other leaf cell types in oul1 mutants demonstrating increased bulliform cell size and number appeared normal. oul1 mutants do not have altered in oul1 compared with the wild type (Fig. 2, D–F). The adaxial-abaxial identity of the internal leaf structure positions and numbers of linear bulliform cell files on
Table I. Measurements of photosynthetic rate, transpiration rate, and stomatal conductance Samples were collected from the mid region of wild-type and oul1 flag leaves. Data show means 6 SD (n . 40) using the heteroscedastic t test to show significant differences (* P , 0.05, ** P , 0.01) compared with the wild type.
Leaf Transpiration Rate CO2 Stomatal Conductance Photosynthetic Rate 2 2 2 2 2 2 mmol m 2 s 1 mol m 2 s 1 mmol m 2 s 1 Wild type 2.90 6 0.52 0.189 6 0.022 11.47 6 0.73 oul1 3.26 6 0.33* 0.224 6 0.020** 13.03 6 0.77**
Plant Physiol. Vol. 156, 2011 1591 Zou et al.
unfolded stage (Fig. 3F). The average number of bulli- form cells was seven (nine maximum) and five (six maximum) in oul1 and the wild type, respectively (Fig. 3, E–L), suggesting that differences in oul1 and wild- type bulliform cell development occurred after folding and before unfolding of the first leaf. Furthermore, bulliform cell area in oul1 gradually increased and became higher than in the wild type during leaf development (Fig. 3M), which was consistent with the highest OUL1 expression level correlating with the appearance of bulliform cell differences between oul1 and the wild type (Fig. 3N). oul1 plants had no visible phenotypes up to the four-leaf stage. However, oul1 leaves spiraled slightly in the middle of the leaf blade at the beginning of the five-leaf stage and gradually curved toward the abaxial surface to become shallot like, whereas wild-type leaves remained flat through- out development.
oul1 Phenotypes Are Due to T-DNA Insertion in Roc5 Figure 2. oul1 has increased bulliform cell number and size. A and B, The oul1 mutant was crossed to Nipponbare wild- Adaxial epidermal peels abutting the small veins of the wild type (wt; A) type plants. Genetic analyses of heterozygous F1 and oul1 (B). C to F, Cross sections of wild-type (C) and oul1 (D) mature progeny showed that the oul1 phenotype segregated leaf blades show significantly increased oul1 bulliform cell number (E) in a 3:1 ratio (x2 = 1.256 , x2 ) of wild-type (208) and area (F) between vascular bundle ridges. ab, Abaxial; ad, adaxial. 0.05,1 and mutant-like (81) plants, indicating that the oul1 Red lines (C and D) show the bulliform cells. Data show means and SD values of biological replicates (n . 23) and statistical analysis by leaf-rolling phenotype was caused by a single reces- heteroscedastic t test indicating significant differences (** P , 0.01). sive mutation. Bars = 20 mm (A–D). G, Relative water content of the 10th leaf of 120- Through PCR walking (Cottage et al., 2001; Peng d-old greenhouse plants grown in the soil. oul1 had higher water et al., 2005), a genomic DNA fragment flanking the content than the wild type. The data are means and SD (n . 5), with T-DNA insertion site in oul1 was isolated. BLAST statistical analysis using the heteroscedastic t test showing significant search with the flanking sequence (http://blast.ncbi. differences (** P , 0.01). nlm.nih.gov/Blast.cgi) identified the full sequence of the GL2-type homeobox gene Roc5 (P0657H12.28). Roc5, which spans approximately 7,049 bp on chro- the leaf blade were similar in the mutant, suggesting mosome 2, contains nine exons and eight introns. The that the outcurved oul1 leaf phenotype may be caused T-DNA was inserted into the sixth intron (Fig. 4A). by the increase in bulliform cell number and size. To confirm that the T-DNA insertion in Roc5 cose- gregated with the oul1 mutant phenotype, more than Increased Bulliform Cell Area Correlates with oul1 100 segregating individual plants were tested by PCR. Leaf Rolling Gene-specific PCR primers (AS39 and S39) flanking the insertion in Roc5 were used in combination with a Bulliform cells are large, thin walled, and highly primer to the T-DNA left border (primer LB2; Fig. 4A). vacuolated (Jane and Chiang, 1991). Bulliform cells Under these conditions, PCR amplification using wild- likely modulate leaf rolling (Li et al., 2010), with loss of type DNA or homozygous oul1-like DNA yielded only turgor pressure causing leaves to roll or curl (Moulia, a fragment with the gene-specific primers or AS39/LB2 1994; Moore et al., 1998) and cell shrinkage linked to primers, whereas heterozygous DNA amplified two rice leaf rolling (O’Toole et al., 1979; Kadioglu and fragments with the gene-specific primers (AS39 and Terzi, 2007). To investigate whether oul1 leaf rolling S39) and the combination primers (AS39 and LB2). PCR was due to changes in bulliform cells, we determined genotyping analysis confirmed that the oul1 mutant when the discrepancy of bulliform cell formation in phenotype was due to the presence of a homozygous different leaves at various developing stages occurred T-DNA insertion within Roc5. Furthermore, reverse between the wild type and oul1. Cytological analyses transcription (RT)-PCR of RNA from the homozygous of seedlings were performed on emerging (3 d), folded oul1 mutant did not detect full-length transcripts for (5 d), and unfolded (8 d) first complete leaves and also Roc5 (Fig. 4B). The RT-PCR primers spanned an intron on the second (14 d), third (22 d), and fourth (32 d) (Fig. 4A), and hence the genomic DNA should not leaves. Compared with the wild type, transverse sec- have been amplified and thus was used as a control for tions of oul1 first complete leaves revealed that bulli- RNA-specific amplification (Fig. 4B). form cell sizes were similar before the leaves folded Previous sequence analysis revealed that Roc5 en- (Fig. 3, A–D) and had increased number and size at the codes an ortholog of Arabidopsis GL2, a member of the
1592 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Figure 3. Leaf rolling in oul1 is due to increased number and size of bulliform cells. A to L, Cross sections of the first leaves of 3-d (A and B), 5-d (C and D; folded leaves), and 8-d (E and F; unfolded leaves) seedlings and second (14 d; G and H), third (22 d; I and J), and fourth (32 d; K and L) leaves of wild-type (wt) and oul1 plants. Red arrowheads (C and D) and lines (E–L) indicate bulliform cells. In the first leaf, bulliform cells were undifferentiated at 3 d (A and B) and were evident in the folded leaf of 5-d-old seedlings (C and D), but bulliform cell number increased in oul1 compared with the wild type in the 8-d-old unfolded leaf, although the leaves of oul1 were flat (E and F). Bulliform cell number also in- creased in first unfolded (E and F), second (G and H), third (I and J), and fourth (K and L) leaves of oul1 compared with the wild type. Bars = 20 mm. M, Bulliform cell area between vascular ridges from the first folded leaf to the fourth leaf stage shows a higher increasing trend in oul1 than in the wild type. Data show means and SD of bio- logical replicates (n . 23). N, qRT-PCR analysis of Roc5 expression in different developing leaves. Data show means and SD (n = 3).
HD-Zip IV family of proteins (Ito et al., 2003). GL2 is a open reading frame of Roc5 driven by its own pro- master regulator of epidermal trichome and non-root- moter. Of the numerous transgenic plants obtained, hair cell development in Arabidopsis (Rerie et al., 1994; most showed the wild-type level of Roc5 expression as Di Cristina et al., 1996; Masucci et al., 1996; Szymanski determined by quantitative real-time RT-PCR (qRT- et al., 1998). Roc5 contains the characterized Gly-rich, PCR; Fig. 5A). oul1 plants expressing Roc5 restored the homeobox, and START domains (Fig. 4A; http://www. wild-type leaf shape and bulliform cell morphology, uniprot.org/uniprot/Q6EPF0; Ito et al., 2003). However, number, and area (Fig. 5). These results confirmed that Roc5 does not contain a classical nuclear localization the outcurling leaf and abnormal bulliform cell devel- signal, although application of bioinformatics in the opment in oul1 were due to T-DNA insertion in Roc5. analysis found that Roc5 has two predicted nuclear lo- calization signals (GGRMLGGG and RKRKK; https:// www.predictprotein.org). To test whether Roc5 localized Enhanced and Suppressed Expression of Roc5 Leads to to the nucleus, the full-length Roc5 coding region was Adaxially and Abaxially Rolled Leaves, Respectively fused to the gene encoding GFP under the control of the cauliflower mosaic virus 35S promoter. Vectors to ex- To further elucidate the role of Roc5 in bulliform cell press the 35S::Roc5-GFP fusion protein and 35S::GFP (as development and leaf shape, the full-length Roc5 cod- a control) were introduced into onion (Allium cepa) ing region driven by the 35S promoter was introduced epidermal cells by particle bombardment transforma- into wild-type Nipponbare rice via A. tumefaciens- tion. Roc5-GFP localized only in the nucleus, whereas mediated transformation. Independent transgenic GFP was detected throughout cells, suggesting that lines were generated and confirmed by PCR detection Roc5 is a nuclear protein (Fig. 4, C–H). of the transgene. Roc5 expression was examined by qRT-PCR of total RNA extracted from leaves of each transgenic line. Roc5 mRNA levels were dramatically Complemented Expression of Roc5 Rescues the Mutant higher in overexpressing lines than in the wild type Phenotypes of oul1 (Fig. 6A). Furthermore, mature leaves of all these Roc5- overexpressing lines were adaxially rolled (Fig. 6B). To confirm that Roc5 disruption resulted in the oul1 Adaxial epidermis stained by toluidine blue O showed mutant phenotype, calli derived from the oul1 back- a decreased number of bulliform cells in ox1 and ox2 ground were transformed via Agrobacterium tumefa- plants (Fig. 6, E, F, K, and L). Cross sections of mature ciens with pRoc5::Roc5, a construct with the full-length leaves showed decreased bulliform cell number and
Plant Physiol. Vol. 156, 2011 1593 Zou et al.
Figure 4. Roc5 characterization and nuclear localization of Roc5 in onion epidermal cells. A, Schematic repre- sentation of the T-DNA insertion into the sixth intron of Roc5 (exons = black boxes; introns = white boxes) and Roc5 protein domain organization (http:// www.uniprot.org/uniprot/Q6EPF0). L and R represent the left and right T-DNA borders, respectively. Arrows indicate the primers used for genotyp- ing with AS39/LB2/S39 and with Q1f and Q1r. UTR, Untranslated region. B, RT- PCR analyses of Roc5 expression in the wild type (wt) and oul1, using ACTIN as a control. One RT-PCR primer for Roc5 spans the sixth intron, and ge- nomic DNA (gDNA) served as a neg- ative control. C to H, 35S::Roc5-GFP (C–E) and 35S::GFP (F–H) constructs were transiently expressed in onion epidermal cells, showing bright-field (C and F), GFP (D and G), and merged (E and H) signals.
size between two vascular bundle ridges in Roc5- negatively controls the development of bulliform cells, overexpressing lines compared with wild-type rice which in turn modulate leaf rolling. (Fig. 6, Q, R, U, and V). In some extreme cases, bulli- form cell sizes were similar to epidermal cells (data not PFL-RNAi Transgenic Plants Exhibit Reversely Rolling shown). Other leaf cell types in the Roc5-overexpressing Leaves, and PFL Is a Putative Target of Roc5 lines appeared normal, and there was no alteration in the adaxial-abaxial arrangement of the internal leaf To understand how Roc5 regulates downstream gene structure (Fig. 6, O, Q, and R). The transgenic lines with expression, expression profiling was performed using wild-type levels of Roc5 expression showed wild-type GeneChips (Affymetrix) to screen for putative target morphological phenotypes (data not shown), suggest- genes of Roc5. Triplicate mRNA samples of 8-d-old ing that the morphological changes in the overexpress- seedlings of oul1 mutants and the wild type were ing lines were not due to the transgenic procedure. harvested when the first complete leaf unfolded, be- Unexpectedly, some plants with abaxially rolled cause Roc5 expression was highest at this stage (Fig. leaves were observed in the transgenic lines (Fig. 3N). Microarray analysis indicated that 54 genes were 6B). qRT-PCR analysis of two independent lines ex- up-regulated and 64 genes were down-regulated in oul1 hibiting abaxially rolled leaves (cs1 and cs2) showed mutants, with differentially expressed genes having that Roc5 mRNA levels were decreased compared with more than 2-fold change and P , 0.01 detection values the wild type (Fig. 6A). The outcurved leaves of cs1 (Supplemental Table S1). and cs2 had an increased number and size of bulliform The L1 box is well conserved within promoter cells (Fig. 6, G, H, M, N, and S–V), as for ocu1 (Fig. 6P), regions of down-regulated target genes of all L1- suggesting that endogenous Roc5 expression was specific HD-Zip IV genes analyzed to date (Abe et al., probably suppressed by exogenous Roc5 cDNA, which 2001, 2003; Ohashi et al., 2003; Tominaga-Wada et al., has been referred to as a cosuppression (Napoli et al., 2009). Among the differentially expressed genes, 12 1990; van der Krol et al., 1990; Chen et al., 2007). These genes contained an L1 box in their promoter regions results and the phenotypes of oul1 suggested that Roc5 (Table II).
1594 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Figure 5. Complemented expression of Roc5 rescues the mutant phenotypes of oul1.A,qRT- PCR analyses of Roc5 expression in the wild type (wt), oul1, and independent complementation plant 1 (cp1) and cp2, showing the expression of cp1 and cp2 near the wild-type level. Ex- periments are biological replicates with SD.B, Comparison of the wild type, oul1, and the complementation lines shows that Roc5 could rescue the leaf-rolling phenotype in oul1. ab, Abaxial; ad, adaxial. Bars = 10 cm (top) and 5 mm (middle and bottom). C to J, Toluidine blue O-stained adaxial epidermal peels abutting large (C–F) or small (G–J) veins of the wild type (C and G), oul1 (D and H), and cp1 and cp2 (E, F, I, and J). K to N, Cross sections of the wild type, oul1, and complementation lines show similar bulli- form cell morphology in complementation plants (M and N) and the wild type (K). Line segments (red) highlight the bulliform cells. Bars = 20 mm (C–N). O to Q, LRIs of the second leaf from the top at flowering time (O), and the number (P) and area (Q) of bulliform cells in mature leaf blades for the wild type, cp1, and cp2.
To determine putative Roc5 targets in these 12 genes, in the wild type. This was confirmed by qRT-PCR (Fig. 7, RNAi constructs were made (Fig. 7A) and introduced B and C). Furthermore, all abaxially rolled leaves of PFL- into the wild-type rice calli through A. tumefaciens- RNAi transgenic plants exhibited decreased PFL expres- mediated transformation. The transgenic T0 plants sion compared with those without curling (Fig. 7D), weregrowninapaddyfieldoftheChineseAcademy which was consistent with microarray data. The bulli- of Agricultural Sciences in Beijing. In the paddy field, form cell number and area beside large veins of abaxially only PFL-RNAi transgenic plants displayed abaxially rolled leaves increased to varying degrees in transgenic rolled leaves (Fig. 7E). If Roc5 regulates PFL (GenBank plants (Fig. 7, F–H). All these results suggested that the accession no. AP003682.3), the expression of PFL should L1 box sequence [AACATTT(T)A; from approximately – be higher in overexpression lines and lower in oul1 than 1,906 to –1,898 bp and approximately –487 to –494 bp of
Plant Physiol. Vol. 156, 2011 1595 Zou et al.
Figure 6. Enhanced or suppressed expression of Roc5 leads to adaxially or abaxially rolled leaves, respectively. A, qRT-PCR analysis of Roc5 expres- sion shows that the two independent overexpres- sion lines (ox1 and ox2) had higher expression levels than the wild type (wt), whereas cosup- pression (cs1 and cs2) transcripts were lower than the wild type, with the expression of cs2 near the oul1 level. B, Morphological phenotypes of wild- type, oul1, ox1, ox2, cs1, and cs2 plants show that the leaves of overexpression lines were adaxially rolled and cosuppression lines were abaxially rolled, whereas those of the wild type were flat. ab, Abaxial; ad, adaxial. Bars = 10 cm (top) and 5 mm (middle and bottom). C to N, Toluidine blue O-stained adaxial epidermal peels abutting large (C–H) or small (I–N) veins of the wild type (C and I) and oul1, ox1, ox2, cs1, and cs2 (D–H and J–N) showing purple-stained bulli- form cells. O to T, Cross sections of leaves show that, compared with the wild type (O), bulliform cells in ox1 (Q) and ox2 (R) were smaller, whereas those in cs1 (S) and cs2 (T) plants were larger, with cs2 (T) similar to oul1 (P). Red lines highlight the bulliform cells. Bars = 20 mm (C–T). U and V, Measurement of numbers (U) and area (V) of bulliform cells in the wild type, ox1, ox2, cs1, and cs2. Data show means of biological replicates with SD (n . 30).
the transcription start site] located in the PFL promoter monocots, which have different epidermal cell types region was probably recognized by Roc5. and organization. In this study, we isolated a rice leaf- rolling mutant, oul1, and the corresponding cloned gene was Roc5 (Fig. 4A), a member of the HD-GL2 DISCUSSION (HD-Zip IV) gene family. Our work shows that Roc5 has essential roles in the formation and development The epidermis is the outermost cell layer covering of epidermal bulliform cells in rice. Therefore, HD-Zip the plant body, which prevents water loss and path- IV family genes seem to have a conserved function in ogen invasion (Martin and Glover, 2007). Epidermal regulating epidermal cell fate both in monocots and cells are specialized cells, which differentiate from the dicots. early undifferentiated epidermis in significant pat- The Roc5 T-DNA insertion knockout mutant oul1 terns and frequencies (Glover, 2000). Intensive genetic had significantly increased bulliform cell number and analyses have established the molecular mechanism size than the wild type (Fig. 2, C–F). However, the regulating epidermal cell fate in Arabidopsis (Ishida position and number of linear bulliform cell files on et al., 2008). However, it is unknown whether the same leaf blades did not change in oul1 (Fig. 2, A and B), mechanism works in other plant species, especially indicating that Roc5 plays a role in bulliform cell fate
1596 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Table II. Twelve genes that are significantly expressed in oul1, and their candidate L1 box-binding motif in the promoter region Italics indicated down-regulated genes in oul1 mutants; the others were up-regulated. RAP, Rice Annotation Project; TIGR, The Institute for Genomic Research; wt, wild type. Intensity Value Fold Putative Roc5 RAP or TIGR Change Binding Site Binding P Description Locus Identifier oul1-1 oul1-2 oul1-3 wt1 wt2 wt3 (oul1/Wild in the Promoter Motif Type) Region LOC_Os10g30670 234.9 217.7 183.1 10.22 7.37 9.65 23.3 5E-08 2910 to AACATTTA Putative transposon 2903 bp protein LOC_Os01g35330 901 804.1 736.9 91.88 116.8 94.96 8.04 7E-08 2976 to TAAATGTT Unknown protein 2969 bp LOC_Os03g19600 56.4 70.55 69.86 9.01 11.36 7.69 7.01 6E-06 2687 to TGCATTTA Unknown protein 2680 bp and TAAATGCA 2413 to 2406 bp Os01g0591000 481 466.2 454.5 68.86 64.15 68.59 6.95 7E-07 2754 to TAAATGTT Cytosolic aldehyde 2747 bp and dehydrogenase 2469 to 2462 bp Osllg0592100 547 670 536.4 85.12 87.7 80.52 6.92 5E-07 2879 to TAAATGTT Cell wall 2872 bp macromolecule catabolic process Os10g0504900 951 730.9 328.2 55.44 195.6 52.94 6.61 0.0004 2561 to AACATTTA Nonspecific 2554 bp lipid-transfer protein 2 Os07g0241600 123 159.4 130.5 25.08 33.03 25.83 4.92 2E-05 2973 to AACATTTA Transferring 2966 bp hexosyl groups Os05g0161500 3,474 3,168 2,435 784.41 550.2 917.1 4.03 1E-05 2894 to TAAATGCA Putative 2887 bp uncharacterized protein Os05g0225800 136 121 88.12 23.03 42.68 32.41 3.52 9E-06 2833 to AACATTTA Putative 2826 bp uncharacterized protein Os11g0226800 10.1 17.42 14.57 90.85 85.22 80.22 0.16 1E-06 2653 to TAAATGCA NBS-LRR-like 2646 bp protein (YR5) Os06g0553200 9.01 11.36 7.69 56.4 70.55 69.86 0.14 7E-07 2494 to AACATTTA Protodermal 2487 bp and factor-like 21,906 to protein 21,898 bp Os08g0473900 64.4 44.9 26.21 477.5 444.4 421.7 0.1 8E-06 2803 to TAAATGCT a-Amylase 2796 bp isozyme 3D but not its patterning. Consistent with this role, Roc5 is withRoc4,Roc6,ZmOCL1,ZmOCL3,AtANL2,and mainly expressed in the L1 layer of the meristem but AtHDG1 (Supplemental Fig. S3; Ito et al., 2003; Khaled not in mature leaves and other organs (Ito et al., 2003), et al., 2005). Roc5’s closest homolog is maize OCL1, with suggesting that Roc5 may be involved in the establish- 85.1% amino acid identity. Our study provides further ment of bulliform cells rather than their maintenance evidence that class IV HD-Zip transcription factors in the adaxial epidermis. Interestingly, Roc5 overex- have a conserved function in controlling epidermal cell pression led to the adaxially curved leaves and de- development in dicots and monocots, indicating that creased number/size of bulliform cells (Fig. 6, B, Q, R, some gene duplications took place before their diver- U, and V). In contrast, the Roc5 cosuppression lines gence. had abaxially rolled leaves, just like oul1 (Fig. 6, B, P, S, Clearly, Roc5 is not the only gene controlling bulli- and T). Overall, several lines of evidence indicate that form cell development. Bulliform cells generally occur Roc5 negatively regulates bulliform cell formation and intercostally as long strips several cells wide. Clonal development in rice. analysis of maize leaf development suggested that the Although Roc5 has no classical nuclear localization linear patterning of bulliform cells may be directed by signal, transient expression in onion cells showed that positional information, just like trichome pattern for- Roc5 is a nuclear protein (Fig. 4, C–H). Roc5 has no mation in Arabidopsis leaves (Hernandez et al., 1999). transcription activity in yeast (Ito et al., 2003) and was a In Arabidopsis, a network of interacting factors (MYB- negative regulator in bulliform cell formation, suggest- BHLH-TTG-GL2) promotes different epidermal cell ing that Roc5 may function as a transcriptional repressor. fates (Serna, 2004). Ectopic expression of the maize Phylogenetic analyses grouped Roc5 into a subfamily R gene, a basic helix-loop-helix family member, in
Plant Physiol. Vol. 156, 2011 1597 Zou et al.
Figure 7. Construction of 35S::PFL-RNAi and characterization of transgenic plants. A, Sche- matic structure of 35S::PFL-RNAi vector includes the inverted repeat sequence of the 3# end regions of PFL DNA. LB, Left border; Hyg, hygromycin; NOS ter, A. tumefaciens nopaline synthase ter- minator; RB, right border. B to D, qRT-PCR anal- yses of PFL (B and D) and Roc5 (C) expression show that, compared with the wild type (wt), PFL expression was decreased in oul1 (B), increased in overexpression (ox) lines, similar to Roc5 (C), and decreased in nine independent T0 PFL-RNAi transgenic plants from which total leaf RNA was analyzed (D). E and F, Morphological phenotypes (E) and cross sections (F) of the flat wild-type and abaxially rolled PFL-RNAi leaves show varying increases in bulliform cell size (F; highlighted by red lines). Bars = 5 mm (E) and 20 mm (F). G and H, Measurement of number (G) and area (H) of bulliform cells in wild-type and PFL-RNAi lines. Graphs show means of biological replicates and SD (n . 30).
Arabidopsis induced trichome and root hair formation mation. Moreover, unlike Arabidopsis trichomes and (Lloyd et al., 1992), suggesting that some of these genes root hairs, which are single cells and never connected to have conserved functions in monocots and dicots. each other, rice bulliform cells occur as clusters of Initiation of bulliform cells correlates with procambium several cells, and the clusters are not adjacent to each formation (Jane and Chiang, 1991). Therefore, it would other. Expression of maize OCL4 under the control of be intriguing to identify rice orthologs corresponding to the GL2 promoter enhanced the glabrous phenotype Arabidopsis genes and investigate their functions in of the Arabidopsis gl2 mutant (Vernoud et al., 2009). epidermal cell specification. In Arabidopsis, GL2 pos- Therefore, identification of genetic and molecular in- itively regulates the formation of epidermal cells (tri- teracting partners of Roc5 would determine if Roc5 chomes) in aerial plant parts, but here we showed that has a different function in epidermal cell formation Roc5 negatively regulates rice leaf bulliform cell for- than GL2.
1598 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Roc5 interacts with Roc2 in yeast (Ito et al., 2003), probably regulates PFL by binding the L1 box motif in and hence it will be interesting to study the role of its promoter region (Fig. 7). PFL, which spans ap- Roc2 in bulliform cell development. Several genes proximately 1,392 bp on chromosome 6 and contains affect bulliform cell development in rice. Loss of three exons and two introns, encodes a protodermal function of BRD1, encoding a protein that catalyzes factor-like protein (protein identifier BAD53672.1). the C-6 oxidation step in brassinosteroid synthesis, A protein BLAST search (http://blast.ncbi.nlm.nih. increased the number of bulliform cells but did not gov/Blast.cgi) identified three hypothetical pro- cause the leaves to roll because the brd1 mutation also teins (SORBIDRAFT_10g021810 [Sorghum bicolor] and affected internal leaf cells, resulting in stiff leaf blades LOC100502555 and LOC100382257 [maize]) with 79%, (Hong et al., 2002). The YABBY family gene YAB1 is 78%, and 72%, respectively, amino acid sequence involved in feedback regulation of GA3 biosynthesis in identity to PFL, although none of these potentially rice. In YAB1 cosuppression plants, blades were abax- homologous proteins have putative conserved do- ially rolled to form a cylinder-like structure because of mains. the increased number of bulliform cells between the To date, several GL2-type homeobox genes have vascular bundles on the adaxial surface, whereas the been cloned. ATML1 is expressed in protodermal cells abaxial-adaxial identity in the internal leaf structure from an early stage of embryogenesis (Lu et al., 1996). was not affected. Hence, the rolled leaf phenotype may Similarly, ZmOCL1 is expressed in the protodermal be due to increases in bulliform cell number, and YAB1 cells from an early stage of embryogenesis and is also may have a function in bulliform cell division and expressed in the outer layer of the developing root differentiation (Dai et al., 2007). The rice adl1 mutant, apical meristem (Ingram et al., 1999). Although amino mutated in phytocalpain, showed defects in leaf po- acid sequence comparison and phylogenic analysis larity and abaxially rolled leaves (Hibara et al., 2009). showed that Roc5 is most similar to Arabidopsis ANL2 adl1 mutants have increased bulliform cell number (Supplemental Fig. S3; Ito et al., 2003), our study sup- present also on the abaxial blade surface, demonstrat- ports a Roc5 function that is similar to Arabidopsis ing that both differentiation and patterning of bulli- ATM, indicating that ANL2 function may have arisen form cells are affected. qRT-PCR analyses of the above after the divergence of monocots and dicots. However, three genes showed no significant difference in their delineating the involvement of cis-regulatory elements expression between wild-type and oul1 mutant plants (L1 box) in multiple DNA-protein interactions is a (data not shown), indicating that Roc5 regulates bulli- complex process. Whether PFL is a downstream target form cell differentiation either through a different of Roc5 remains to be determined. pathway or downstream of these genes. In the large, thin-walled, highly vacuolated bulli- form cells (Jane and Chiang, 1991), loss of turgor MATERIALS AND METHODS pressure (Moulia, 1994; Moore et al., 1998) and shrink- Plant Materials and Measurements age (O’Toole et al., 1979; Kadioglu and Terzi, 2007) have been linked to leaf rolling. However, several Rice (Oryza sativa spp. japonica) cv Nipponbare plants were grown in a other studies claimed that nonbulliform leaf cell types greenhouse at 30°C (16 h of light) and 22°C (8 h of dark) or in a paddy field of the Chinese Academy of Agricultural Sciences in Beijing from May to October control leaf rolling under water stress (Shields, 1951; of each year. LRI and LEI were measured as described by Shi et al. (2007). Metcalfe, 1960), and Linsbauer (1930) suggested that Wild-type and oul1 photosynthetic rate, transpiration rate, and stomatal bulliform cells are just for water storage. Both the oul1 conductance were measured at full heading stage using the portable photo- mutant (Fig. 2, A–F) and cosuppression lines (Fig. 6, B synthetic LCPRO+ instrument (ADC Bioscientific), with 500 mmol s–1 flow ° m –1 and S–V) had increased bulliform cell numbers, lead- velocity, 30 C leaf chamber, and 1,800 mol s light quantum flux density. Water content of wild-type and oul1 leaves (the 10th leaf of 120-d-old plants) ing to abaxial leaf rolling, whereas decreased bulliform was measured as described by Barrs and Weatherley (1962). Three leaf cell numbers in Roc5-overexpressing lines resulted in weights were taken: W1, immediately after leaf excision; W2, after wiping off adaxial leaf rolling (Fig. 6, B, Q, R, U, and V). In these water following a 4-h water saturation of excised leaves at room temperature; Roc5 mutant and transgenic lines, no other epidermal and W0, after subsequently drying leaves for 12 h at 70°C. Each measurement had at least five replicates. The formula for calculating the relative water cells or internal leaf structures were changed (Figs. 2, C content is as follows: relative water content (%) = (W1 – W0)/(W2 – W0) 3 100%. and D, and 6, O–T). Therefore, our study provides molecular genetic evidence supporting a direct role for bulliform cells in controlling leaf shape. Leaf rolling Histology and Microscopy Observation decreases the leaf surface area, which is believed to For paraffin section analysis, the basal half of each mature 10th leaf was reduce water loss during drought. The transpiration collected and fixed in Carnoy’s solution (60:30:10, ethanol:chloroform:glacial rate was higher in oul1 than in the wild type (Table I), acetic acid, v/v/v) for 12 h. Samples were washed with 60% (v/v) ethanol, immersed in 30% (v/v) hydrofluoric acid for approximately 7 to 10 d, and then indicating that, at least under some conditions, there is dehydrated with a graded ethanol series and embedded in paraffin. Sections no benefit of leaf rolling under water stress. On the (approximately 5–10 mm thick) were cut with a microtome (Leica RM2155), other hand, the photosynthetic rate of oul1 was signif- stained with 1% (w/v) safranin O (Amresco) and 1% (w/v) fast green FCF icantly higher than in the wild type (Table I), suggest- (Merck), examined with a fluorescence microscope (Zeiss AXIO Imager A1), and photographed. Bulliform cell numbers between two vascular bundle ing that altering leaf structure benefits photosynthesis. ridges from the midrib to the margin were counted, and their area was As a transcription regulator, Roc5, mainly expressed measured with AxioVision release 4.6 software using six separate leaves for in the L1 layer of the meristem (Ito et al., 2003), each sample.
Plant Physiol. Vol. 156, 2011 1599 Zou et al.
To analyze bulliform cell arrangement, approximately 1-cm2 adaxial surfaces Overexpression Construct of Roc5 were peeled approximately midway along the mature rice blade (10th leaf), gently abraded with an aqueous silica mixture, and then soaked in 95% (v/v) The 2.5-kb Roc5 full-length coding region was amplified by PCR with ethanol for approximately 1 d, changing the solution every 4 h to remove the primers Roc5-5#SmaI(5#-CCCGGGGATCCAAGGAAGAGGACTTTC- chlorophyll. Leaves were then rinsed in water and soaked for 12 h in 1 N NaOH to TCG-3#) and Roc5-3#XbaI(5#-TCTAGACGTTCTTGACTCAGGCTCGTC-3#). extract cell contents, rinsed again in water, and then fixed in Carnoy’s solution for The PCR product was cloned into the pEASY-Blunt simple cloning vector 12 h. Samples were then stained overnight in 1% (w/v) toluidine blue O (Transgen Biotech) and sequenced, then excised from the vector by SmaI and (Hernandez et al., 1999). After washing with water, ribs were removed and leaves XbaI digestion and subcloned into pCAMBIA 23A. In pCAMBIA 23A, Roc5 were sealed for microscopy (Zeiss AXIO Imager A1) and photographed. was downstream of the 35S promoter. The 35S::Roc5 construct was introduced into wild-type rice calli by A. tumefaciens-mediated transformation.
Roc5 Isolation and Complementation Microarray and Data Analysis Genomic DNA flanking the T-DNA left border was cloned using PCR walking with nesting-specific primer pairs according to Cottage et al. (2001) Total RNA of both Nipponbare and oul1 was extracted using TRIzol and Peng et al. (2005). The primers used were LB1 (5#-CGATGGCTGTG- reagent, and GeneChip expression analysis was performed by CapitalBio. TAGAAGTACTCGC-3#), LB2 (5#-GTTCCTATAGGGTTTCGCTCATGTG- Each sample was assayed in triplicate. This array contains 22,582 probe sets # # # m TTG-3 ), AP1 (5 -GGATCCTAATACGAGTCACTATAGCGC-3 ), and AP2 representing 10,127 positive genes. In brief, total RNA (1–15 g) was first (5#-CTATAGCGCTCGAGCGGC-3#). PCR products were directly sequenced. reverse transcribed using a T7 oligo(dT) promoter primer in the first-strand The T-DNA insertion site in Roc5 was identified using National Center for cDNA synthesis reaction. Following RNase H-mediated second-strand cDNA Biotechnology Information BLAST searches of the rice genome database synthesis, the double-stranded cDNA was purified and served as a template (http://www.ncbi.nlm.nih.gov/Blast/) of the rescued flanking sequences. for in vitro transcription, which was performed with T7 RNA polymerase and The full-length cDNA of Roc5 was cloned into the vector carrying a 1.8-kb Roc5 a biotinylated nucleotide analog/ribonucleotide mix for copy RNA amplifi- promoter to generate pRoc5::Roc5. The vector was introduced into the Agro- cation and biotin labeling. The biotinylated copy RNA targets were then bacterium tumefaciens strain AGL1 using the heat shock method and then was cleaned up, fragmented, and hybridized to GeneChip expression arrays transformed into oul1 mutant calli. (GeneChip Expression Analysis Technical Manual; CapitalBio). After the arrays were washed and stained with streptavidin-PE in a GeneChip-Fluidics Station 450 (Affymetrix), the hybridized microarrays were scanned using the Genotyping of oul1 with PCR Affymetrix GeneChip Scanner 3000 and converted into DAT/CEL images for analysis (Zou et al., 2009). Genotyping of the oul1 segregating population was performed by PCR using Data normalization and comparison were performed for the GeneChip # # the following primers: AS39 (5 -CCACTACTTCTCCACTACCACTATCAC-3 ), expression experiment (Bolstad et al., 2003; Irizarry et al., 2003a, 2003b; Smyth, # # S39 (5 -TCAATCATTTCGATCAAGAGTGCAAC-3 ), and LB2 (T-DNA left bor- 2004). Genes with a greater than 2-fold change and a detection value of P , ° der). PCR was conducted with an initial step of 94 C incubation for 3 min and 0.01 were defined as differentially expressed genes. Sequences of 2 kb up- ° ° ° 30 cycles of 94 C for 30 s, 56 C for 30 s, and 72 C for 1 min. stream of the transcription start site (ATG) in all differentially expressed genes were identified from ftp://ftp.plantbiology.msu.edu/pub/data/Eukaryotic_ Transient Expression in Onion Epidermal Cells Projects/o_sativa/annotation_dbs/pseudomolecules/version_6.1/all.dir/all. 2kUpstream or annotation files of Affymetrix microarray and defined as pro- The 2.5-kb full-length coding sequence of Roc5 was amplified by PCR with moter regions of differentially expressed genes. Promoters containing an L1 box primers Roc5-GFPF (5#-CATGCCATGGAAGAGGACTTTCTCGAG-3#)and were selected, and their corresponding genes were termed as candidate targets of Roc5-GFPR (5#-GGAAGATCTTTGATGGTGCAGGAGATGAG-3#)fromRoc5 Roc5 (Abe et al., 2001, 2003; Ohashi et al., 2003; Tominaga-Wada et al., 2009). mRNA and cloned into pCAMBIA1302 to generate 35S::Roc5-GFP.Thefusion plasmid and pCAMBIA1302 (35S::GFP, as a control) were transformed into 35S PFL onion (Allium cepa) epidermal cells with the Bio-Rad PDS-1000/He device Generation of :: -RNAi Transgenic Rice Plants (Bio-Rad). Bombarded epidermal cells were incubated for 20 h at 25°Cinthe PFL (GenBank accession no. P0427B07.25) DNA encompassing 165 bp of dark. The cell layers were then examined with laser scanning confocal the 3# end (999–1,164 bp) was amplified with primers 5#-GGGGTACCAC- microscopy (Leica TCS SP2). GFP fluorescence was imaged using excitation TAGTATCTCTCCGGCCATCTCTC-3# and 5#-CGGGATCCGAGCTCAAAC- with the 488-nm line of the argon laser and a 505- to 530-nm band-pass GCACACAAGCCACA-3# and inserted into the pTCK309 vector using SacI, emission filter. SpeI, KpnI, and BamHI sites in inverted orientations. The PFL-RNAi construct was transferred to A. tumefaciens strain AGL1 by electroporation and then RT-PCR and qRT-PCR introduced into wild-type rice calli through A. tumefaciens-mediated transfor- mation. The regenerated T0 plants were grown in a paddy field at the Chinese Total RNA was extracted using TRIzol solution (Invitrogen) from leaves of Academy of Agricultural Sciences in Beijing. The abaxially rolling leaves of wild-type and oul1 plants. Total RNA (2 mg) from each sample was reverse PFL-RNAi transgenic plants were selected and analyzed. transcribed with oligo(dT) primer and PrimeScript RT Enzyme (TaKaRa) ac- cording to the manufacturer’s instructions. For RT-PCR, the PCR primers for Sequence data from this article can be found in the GenBank/EMBL data # # amplifying Roc5 were Q1f (5 -ATGGGCAGTAGTTGATGTGTC-3 )andQ1r libraries under accession number EU267976. (5#-CGAAGGAGTGGACGGTAGAG-3#), and primers for actin were actinf (5#-GACCTTGCTGGGCGTGATCTC-3#) and actinr (5#-GATGGGCCAGAC- TCGTCGTAC-3#). PCR conditions were as follows: preincubation at 94°C Supplemental Data for 2.5 min, then 30 cycles at 94°C for 30 s, 52°C for 30 s, and 72°C for 1 min. qRT-PCR was performed on the iQ5 Muticolor Real-Time PCR Detection The following materials are available in the online version of this article. System (Bio-Rad) with real-time PCR Master Mixture (SYBR Green Mix). The Supplemental Figure S1. The bulliform cell phenotype in wild-type and # # primers for Roc5 were Roc5F (5 -CGCAAGAGGAAGAAGCGATAC-3 )and oul1 plants. Roc5R (5#-GCTCCAGTTGCGTCTTCATC-3#). The primers for PFL were PFLf (5#-ATCTCTCCGGCCATCTCTC-3#) and PFLr (5#-AAACGCACACAAGC- Supplemental Figure S2. Transmission electron microscopy analysis of CACA-3#). qRT-PCR was performed in triplicate for each individual line, wild-type (A) and oul1 (B) chloroplast grana lamellae (arrows). and threshold cycle values were quantified by qRT-PCR by calculating means Supplemental Figure S3. Phylogenetic tree showing the predicted rela- of normalized expression using the relative quantification method (Livak tionships between almost all plant members of class IV HD-Zip proteins and Schmittgen, 2001). The rice ACTIN gene, amplified with primers actinF analyzed to date. (5#-TGCTATGTACGTCGCCATCCAG-3#) and actinR (5#-AATGAGTAAC- CACGCTCCGTCA-3#), was selected as an internal standard to normalize Supplemental Table S1. Microarray data for the 118 differentially ex- the expression of Roc5. pressed genes.
1600 Plant Physiol. Vol. 156, 2011 Roc5 Controls Leaf Rolling
Supplemental Materials and Methods S1. Transmission electron micros- ZmOCL1, an HDGL2 family homeobox gene, is expressed in the outer copy analysis. cell layer throughout maize development. Plant Mol Biol 40: 343–354 Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP (2003a) Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31: e15 ACKNOWLEDGMENTS Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP (2003b) Exploration, normalization, and summaries of high We thank Dr. Chengcai Chu for providing the binary vector. density oligonucleotide array probe level data. Biostatistics 4: 249–264 Ishida T, Kurata T, Okada K, Wada T (2008) A genetic regulatory network Received March 9, 2011; accepted May 9, 2011; published May 19, 2011. in the development of trichomes and root hairs. Annu Rev Plant Biol 59: 365–386 Ito M, Sentoku N, Nishimura A, Hong SK, Sato Y, Matsuoka M (2002) LITERATURE CITED Position dependent expression of GL2-type homeobox gene, Roc1: significance for protoderm differentiation and radial pattern formation Abe M, Katsumata H, Komeda Y, Takahashi T (2003) Regulation of shoot in early rice embryogenesis. Plant J 29: 497–507 epidermal cell differentiation by a pair of homeodomain proteins in Ito M, Sentoku N, Nishimura A, Hong SK, Sato Y, Matsuoka M (2003) Arabidopsis. Development 130: 635–643 Roles of rice GL2-type homeobox genes in epidermis differentiation. Abe M, Takahashi T, Komeda Y (2001) Identification of a cis-regulatory Breed Sci 53: 245–253 element for L1 layer-specific gene expression, which is targeted by an Jane WN, Chiang SHT (1991) Morphology and development of bulliform L1-specific homeodomain protein. Plant J 26: 487–494 cells in Arundo formosana Hack. Taiwania Int J Life Sci 36: 85–97 Barrs HD, Weatherley PE (1962) A re-examination of the relative turgidity Kadioglu A, Terzi R (2007) A dehydration avoidance mechanism: leaf technique for estimating water deficits in leaves. J Biol Sci 15: 415–428 rolling. Bot Rev 73: 290–302 Becraft PW, Li K, Dey N, Asuncion-Crabb Y (2002) The maize dek1 gene Khaled AS, Vernoud V, Ingram GC, Perez P, Sarda X, Rogowsky PM functions in embryonic pattern formation and cell fate specification. (2005) Engrailed-ZmOCL1 fusions cause a transient reduction of kernel Development 129: 5217–5225 size in maize. Plant Mol Biol 58: 123–139 Bolstad BM, Irizarry RA, Astrand M, Speed TP (2003) A comparison of Kubo H, Peeters AJ, Aarts MG, Pereira A, Koornneef M (1999) ANTHO- normalization methods for high density oligonucleotide array data CYANINLESS2, a homeobox gene affecting anthocyanin distribution based on variance and bias. Bioinformatics 19: 185–193 and root development in Arabidopsis.PlantCell11: 1217–1226 Chen R, Zhao X, Shao Z, Wei Z, Wang Y, Zhu L, Zhao J, Sun M, He R, He G Lang YZ, Zhang ZJ, Gu XY, Yang JC, Zhu QS (2004) A physiological and (2007) Rice UDP-glucose pyrophosphorylase1 is essential for pollen ecological effect of crimpy leaf character in rice (Oryza sativa L.). II. callose deposition and its cosuppression results in a new type of Photosynthetic character, dry mass production and yield forming. Acta thermosensitive genic male sterility. Plant Cell 19: 847–861 Agron Sin 30: 883–887 Cottage A, Yang A, Maunders H, de Lacy RC, Ramsay NA (2001) Identi- Li L, Shi ZY, Li L, Shen GZ, Wang XQ, An LS, Zhang JL (2010) Over- fication of DNA sequences flanking T-DNA insertions by PCR-walking. expression of ACL1 (abaxially curled leaf 1) increased bulliform cells Plant Mol Biol Rep 19: 321–327 and induced abaxial curling of leaf blades in rice. Mol Plant 3: 807–817 DaiM,ZhaoY,MaQ,HuY,HeddenP,ZhangQ,ZhouDX(2007) The rice Li SG, He P, Wang YP, Li HY, Cheng Y, Zhou KD, Zhu LH (2000) Genetic YABBY1 gene is involved in the feedback regulation of gibberellin analysis and gene mapping of the leaf traits in rice (Oryza sativa L.). metabolism. Plant Physiol 144: 121–133 Acta Agron Sin 26: 261–265 Di Cristina M, Sessa G, Dolan L, Linstead P, Baima S, Ruberti I, Morelli G Lid SE, Gruis D, Jung R, Lorentzen JA, Ananiev E, Chamberlin M, Niu X, (1996) The Arabidopsis Athb-10 (GLABRA2) is an HD-Zip protein Meeley R, Nichols S, Olsen OA (2002) The defective kernel 1 (dek1) required for regulation of root hair development. Plant J 10: 393–402 gene required for aleurone cell development in the endosperm of maize Elhiti M, Stasolla C (2009) Structure and function of homodomain-leucine grains encodes a membrane protein of the calpain gene superfamily. zipper (HD-Zip) proteins. Plant Signal Behav 4: 86–88 Proc Natl Acad Sci USA 99: 5460–5465 Glover BJ (2000) Differentiation in plant epidermal cells. J Exp Bot 51: Linsbauer K (1930) Die epidermis. In K Linsbauer, ed, Encyclopedia of 497–505 PlantAnatomy,Vol.4(2).Gebru¨ der Borntraeger, Berlin Govaerts YM, Jacquemoud S, Verstraete MM, Ustin SL (1996) Three- Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data dimensional radiation transfer modeling in a dicotyledon leaf. Appl Opt using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. 35: 6585–6598 Methods 25: 402–408 Guan XY, Li QJ, Shan CM, Wang S, Mao YB, Wang LJ, Chen XY (2008) The Lloyd AM, Walbot V, Davis RW (1992) Arabidopsis and Nicotiana antho- HD-Zip IV gene GaHOX1 from cotton is a functional homologue of the cyanin production activated by maize regulators R and C1. Science 258: Arabidopsis GLABRA2. Physiol Plant 134: 174–182 1773–1775 Hernandez ML, Passas HJ, Smith LG (1999) Clonal analysis of epidermal Lu P, Porat R, Nadeau JA, O’Neill SD (1996) Identification of a meristem L1 patterning during maize leaf development. Dev Biol 216: 646–658 layer-specific gene in Arabidopsis that is expressed during embryonic Hibara K, Obara M, Hayashida E, Abe M, Ishimaru T, Satoh H, Itoh J, pattern formation and defines a new class of homeobox genes. Plant Cell Nagato Y (2009) The ADAXIALIZED LEAF1 gene functions in leaf and 8: 2155–2168 embryonic pattern formation in rice. Dev Biol 334: 345–354 Luo YZ, Zhao FM, Sang XC, Ling YH, Yang ZL, He GH (2009) Genetic Hong Z, Ueguchi-Tanaka M, Shimizu-Sato S, Inukai Y, Fujioka S, analysis and gene mapping of a novel rolled leaf mutant rl12(t)inrice. Shimada Y, Takatsuto S, Agetsuma M, Yoshida S, Watanabe Y, et al Acta Agron Sin 35: 1967–1972 (2002) Loss-of-function of a rice brassinosteroid biosynthetic enzyme, Luo Z, Yang Z, Zhong B, Li Y, Xie R, Zhao F, Ling Y, He G (2007) Genetic C-6 oxidase, prevents the organized arrangement and polar elongation analysis and fine mapping of a dynamic rolled leaf gene, RL10(t), in rice of cells in the leaves and stem. Plant J 32: 495–508 (Oryza sativa L.). Genome 50: 811–817 Ingouff M, Farbos I, Lagercrantz U, von Arnold S (2001) PaHB1 is an Martin C, Glover BJ (2007) Functional aspects of cell patterning in aerial evolutionary conserved HD-GL2 homeobox gene expressed in the epidermis. Curr Opin Plant Biol 10: 70–82 protoderm during Norway spruce embryo development. Genesis 30: Masucci JD, Rerie WG, Foreman DR, Zhang M, Galway ME, Marks MD, 220–230 Schiefelbein JW (1996) The homeobox gene GLABRA2 is required for Ingouff M, Farbos I, Wiweger M, von Arnold S (2003) The molecular position-dependent cell differentiation in the root epidermis of Arabi- characterization of PaHB2, a homeobox gene of the HD-GL2 family dopsis thaliana. Development 122: 1253–1260 expressed during embryo development in Norway spruce. J Exp Bot 54: Metcalfe CR (1960) Anatomy of Monocotyledons. I. Gramineae. Claredon 1343–1350 Press, Oxford Ingram GC, Boisnard-Lorig C, Dumas C, Rogowsky PM (2000) Expression Moore R, Clark, WD (1998) Botany, Ed 2. WCB/McGraw Hill, Dubuque, patterns of genes encoding HD-ZipIV homeo domain proteins define Iowa specific domains in maize embryos and meristems. Plant J 22: 401–414 Moulia B (1994) The biomechanics of leaf rolling. Biomimetics 2: 267–281 Ingram GC, Magnard JL, Vergne P, Dumas C, Rogowsky PM (1999) Napoli C, Lemieux C, Jorgensen R (1990) Introduction of a chimeric
Plant Physiol. Vol. 156, 2011 1601 Zou et al.
chalcone synthase gene into petunia results in reversible co-suppression mutantsiscausedbygain-of-functionepigeneticallelesofahomeodo- of homologous genes in trans. Plant Cell 2: 279–289 main gene. Mol Cell 6: 791–802 Ohashi Y, Oka A, Rodrigues-Pousada R, Possenti M, Ruberti I, Morelli G, Szymanski DB, Jilk RA, Pollock SM, Marks MD (1998) Control of GL2 Aoyama T (2003) Modulation of phospholipid signaling by GLABRA2 expression in Arabidopsis leaves and trichomes. Development 125: 1161–1171 in root-hair pattern formation. Science 300: 1427–1430 Tominaga-Wada R, Iwata M, Sugiyama J, Kotake T, Ishida T, Yokoyama R, Ohashi Y, Oka A, Ruberti I, Morelli G, Aoyama T (2002) Entopically Nishitani K, Okada K, Wada T (2009) The GLABRA2 homeodomain additive expression of GLABRA2 alters the frequency and spacing of protein directly regulates CESA5 and XTH17 gene expression in trichome initiation. Plant J 29: 359–369 Arabidopsis roots. Plant J 60: 564–574 O’Toole JC, Cruz RT, Singh TN (1979) Leaf rolling and transpiration. Plant vanderKrolAR,MurLA,BeldM,MolJN,StuitjeAR(1990) Flavonoid Sci Lett 16: 111–114 genes in petunia: addition of a limited number of gene copies may lead Peng H, Huang H, Yang Y, Zhai Y, Wu J, Huang D, Lu T (2005) Functional to a suppression of gene expression. Plant Cell 2: 291–299 analysis of GUS expression patterns and T-DNA integration character- Vernoud V, Laigle G, Rozier F, Meeley RB, Perez P, Rogowsky PM (2009) istics in rice enhancer trap lines. Plant Sci 168: 1571–1579 The HD-ZIP IV transcription factor OCL4 is necessary for trichome Price AH, Young EM, Tomos AD (1997) Trait loci associated with stomatal patterning and anther development in maize. Plant J 59: 883–894 conductance, leaf rolling and heading date mapped in upland rice Wan S, Wu J, Zhang Z, Sun X, Lv Y, Gao C, Ning Y, Ma J, Guo Y, Zhang Q, (Oryza sativa). New Phytol 137: 83–91 et al (2009) Activation tagging, an efficient tool for functional analysis of Rerie WG, Feldmann KA, Marks MD (1994) The GLABRA2 gene encodes a the rice genome. Plant Mol Biol 69: 69–80 homeo domain protein required for normal trichome development in Yan CJ, Yan S, Zhang ZQ, Liang GH, Lu JF, Gu MH (2006) Genetic analysis Arabidopsis. Genes Dev 8: 1388–1399 and gene fine mapping for a rice novel mutant rl9(t) with rolling leaf Serna L (2004) A network of interacting factors triggering different cell character. Chin Sci Bull 51: 63–69 (in Chinese) fates. Plant Cell 16: 2258–2263 Yan S, Yan CJ, Zeng XH, Yang YC, Fang YW, Tian CY, Sun YW, Cheng ZK, Shi Y, Chen J, Liu W, Huang Q, Shen B, Leung H, Wu J (2009) Genetic Gu MH (2008) ROLLED LEAF 9, encoding a GARP protein, regulates analysis and gene mapping of a new rolled-leaf mutant in rice (Oryza the leaf abaxial cell fate in rice. Plant Mol Biol 68: 239–250 sativa L.). Sci China C Life Sci 52: 885–890 Yang Y, Peng H, Huang H, Wu J, Jia S, Huang D, Lu T (2004) Large-scale Shi Z, Wang J, Wan X, Shen G, Wang X, Zhang J (2007) Over-expression of production of enhancer trapping lines for rice functional genomics. rice OsAGO7 gene induces upward curling of the leaf blade that Plant Sci 167: 281–288 enhanced erect-leaf habit. Planta 226: 99–108 Yuan LP (1997) Super-high yield hybrid rice breeding. Hybrid Rice 12: 1–6 Shields LM (1951) The involution mechanism in leaf of certain xeric Zhang GH, Xu Q, Zhu XD, Qian Q, Xue HW (2009) SHALLOT-LIKE1 is a grasses. Pytomorphology 1: 225–241 KANADI transcription factor that modulates rice leaf rolling by regu- Smyth GK (2004) Linear models and empirical Bayes methods for assess- lating leaf abaxial cell development. Plant Cell 21: 719–735 ing differential expression in microarray experiments. Stat Appl Genet Zou X, Zou L, Foldager C, Bendtsen M, Feng W, Bu¨ nger CE (2009) Mol Biol 3: Article 3 Different mechanisms of spinal fusion using equine bone protein Soppe WJ, Jacobsen SE, Alonso-Blanco C, Jackson JP, Kakutani T, extract, rhBMP-2 and autograft during the process of anterior lumbar Koornneef M, Peeters AJ (2000) The late flowering phenotype of fwa interbody fusion. Biomaterials 30: 991–1004
1602 Plant Physiol. Vol. 156, 2011