Further characterization of a rice AGL12-group MADS-box gene, OsMADS26

Plant MADS-box genes can be divided into 11 groups. Genetic analysis has revealed that most of them function in flowering-time control, reproductive organ development, and vegetative growth. Here, we elucidated the role of OsMADS26 , a member of the AGL12 group. Transcript levels of OsMADS26 were increased, in an age-dependent manner, in the shoots and roots. Transgenic plants of both rice and Arabidopsis over-expressing this gene manifested phenotypes related to stress responses, such as chlorosis, cell death, pigment accumulation, and defective root/shoot growth. In addition, apical hook development was significantly suppressed in Arabidopsis . Plants transformed with the OsMADS26 - glucocorticoid receptor ( GR ) fusion construct displayed those stress-related phenotypes when treated with dexamethasone (DEX). Microarray analyses using this inducible system showed that biosynthesis genes for jasmonate, ethylene, and reactive oxygen species as well as putative downstream targets involved in the stress-related process were up-regulated in OsMADS26 -overexpressing plants. These results suggest that OsMADS26 induces multiple responses that are related to various stresses. synthase


5
The MADS-box gene family encodes transcription factors with a conserved DNA-binding domain, called the MADS-box. These genes, ubiquitous in living organisms, have a wide range of functions.
Plant MADS-box genes can be grouped into two evolutionary lineages (Types I and II) (Alvarez-Buylla et al., 2000;Becker and Theissen, 2003). When restricted to the putative functional MADSbox genes, this list includes about 100 and 70 genes in Arabidopsis and rice, respectively (Nam et al., 2004). Approximately 40 Type II MADS-box genes each have been identified in Arabidopsis (Kofuji et al., 2003;Parenicova et al., 2003) and rice (Lee et al., 2003); these can be divided into 11 groups (Becker and Theissen, 2003;Lee et al., 2003;Arora et al., 2007).
Detailed genetic analyses have shown that, whereas some MADS-box genes are involved in reproductive organ development (being preferentially expressed in the floral organs), others are expressed in the vegetative organs, where they perform various roles in flowering-time control, vegetative growth, and root development (Becker and Theissen, 2003). Becker and Theissen (2003) have speculated that AGL12 group genes originated before the gymnosperm-angiosperm split about 300 million years ago. In northern blot analyses, AGL12, the sole MADS-box gene from the AGL12 group in Arabidopsis, shows root-specific expression (Rounsley et al., 1995). Recently, AGL12 overexpression analyses of suspension cells from Catharanthus roseus have demonstrated that this gene promotes the organization of those cells into globular parenchyma-like aggregates (Montiel et al., 2007). Loss-of-function analyses have elucidated that AGL12 regulates root meristem cell proliferation and flowering transition (Tapia-Lopez et al., 2008). In addition, in situ hybridization analyses have shown that this gene is also detected in leaves and floral meristems. In rice, the AGL12-group OsMADS26 is expressed not only in the roots but also in the shoots and panicles (Shinozuka et al., 1999). Pelucci et al. (2002) have also observed that OsMADS26 is highly expressed in the leaves and inflorescences. Furthermore, Arora et al.(2007) have shown the expression of this gene in panicles and seeds. These results imply that this gene functions in a broad 7 and 9 (Fig. 1B). In contrast, transcript levels continuously increased in leaves up to Day 70 (Fig.   1C). Within individual plants, expression was much stronger in the mature leaves than in young, still-developing leaves (Fig. 1D). Therefore, these results indicate that OsMADS26 is more active in older tissues.

Phenotypes of ubi:OsMADS26S plants
To elucidate in vivo functioning, we regenerated transgenic rice plants that expressed either sense or anti-sense constructs of the full-length OsMADS26 cDNA. Plants that ectopically expressed the anti-sense OsMADS26 showed no visible phenotypic changes (data not shown). We previously identified an OsMADS26 knockout (KO) line (1A-16632) from a T-DNA tagging population via reverse-screening (Lee et al., 2003;Ryu et al., 2004). In that line, T-DNA is inserted into the first intron and the OsMADS26 transcript is not present. As observed with our antisense plants in the current study, the T-DNA insertional knockout plants exhibited no abnormality in their growth habit ( Supplementary Fig. 1).
In contrast, primary T1 transgenic plants expressing the sense OsMADS26 transcript (ubi:OsMADS26S plants) showed several abnormal phenotypes (Fig. 2). Among our 50 regenerated plants, 40 died at the young stage, after they manifested such traits as defective root/shoot growth ( Fig. 2A, B), chlorosis and cell death ( Fig. 2A, B), screw-like root curling (Fig. 2C, D), and pigment accumulation in their roots (Fig. 2B, D). The remaining 10 plants showed less severe phenotypes and survived to maturity, with the adults displaying semi-dwarfism (Fig. 2E), pale-green coloration ( Fig. 2E), spotted leaves (Fig. 2F), and shrunken seeds (Fig. 2G). Except for three lines, most of the plants were sterile. The T2 seedlings from those fertile lines had phenotypes similar to those observed from the primary transgenic plants, including retarded root/shoot growth, screw-like root 9 in various stress-related processes. However, some of those characteristics may have been due to indirect effects caused by ectopic overexpression at the regeneration stage. To observe the more direct effects, we generated transgenic plants carrying the OsMADS26-GR fusion construct (ubi:OsMADS26GR plants).
Among the 32 T1 primary transgenics, 11 independent lines were examined to test whether this inducible system would be successful when plants were treated with dexamethasone (DEX). Six lines clearly showed abnormal phenotypes ( Supplementary Fig. 4). For example, Line #33 developed curled and shorter roots while Line #17 had severe growth retardation. The six confirmed lines were followed through the next generations and genotyped to obtain homozygous (HO) plants from each line. For genotyping, at least 50 T2 seedlings were tested for hygromycin resistance. If all plants survived, the lines were regarded as HO; if all died, they were considered to be of the WT.
For further study, Line #33 was selected and its seedlings were treated with DEX in a dose-dependent manner to determine the effective concentration. In the WT segregants, DEX did with Line #17 (data not shown). To understand the nature of these shortened roots, we sectioned their maturation zones. Histological analysis showed that cell elongation was significantly inhibited in the DEX-treated plants (Fig. 3E, F).
The numbers of emerged roots and leaves were also reduced in DEX-treated transgenic plants in both line #33 and #17. For example, by Day 9, WT segregants of Line #33 had developed their 4 th leaves and had an average total of 8 to 10 roots (seminal plus nodal roots; Table I HO plants of Line #33 were treated with DEX, root numbers were reduced at the lowest concentration (10 nM) while leaf numbers were reduced in response to 100 nM DEX. However, by Day 12, both WT and HO plants grown at 100 nM DEX had developed a similar number of leaves (about 4) and roots (about 10). These results suggest that the production of fewer roots and leaves associated with OsMADS26 overexpression was caused by a slower growth rate rather than because of defective primordia development.
To see the direct effects of this overexpression, we applied 1 µ M DEX to six-day-old seedlings. When treated for three consecutive days, the transgenics manifested phenotypes of retarded growth, pigment accumulation by their roots, and wilting, chlorosis, and senescence in their shoots (Fig. 3C). To further understand the role of this gene in these processes, we measured

Identification of putative OsMADS26 downstream genes
To identify the OsMADS26 downstream genes, we compared genome-wide RNA expression levels between the ubi:OsMADS26GR plants and their WT segregants, using a 60 K-oligo chip. Total RNAs were prepared from the roots of 7-day-old seedlings treated with 1 µ M DEX for 3 or 9 h.

Supplementary
indicating that Line #17 generated more consistent results. When we applied a 2-fold difference as our cut-off criterion, 48 genes were identified, with respective Pearson correlation co-efficiencies of 0.806, 0.880, 0.937, and 0.861 for #33 (3 h), #33 (9 h), #17 (3 h), and #17 (9 h). Interestingly, this standard allowed us to identify only 13 down-regulated genes compared with the isolation of 35 upregulated genes, which implies that results fluctuated more with the former type. Our KMC clustering analyses showed global expression patterns for these 48 genes (Fig. 4). All were induced or suppressed more strongly at 9 h than at 3 h. Moreover, 5 were induced dramatically at both 3 and modification/protein turnover/chaperones, carbohydrate transport/metabolism, and secondary metabolite transport/metabolism. Genes belonging to three groups were changed more frequently, i.e., defense mechanisms (2.40%), inorganic ion transport and metabolism (1.02%), and secondary metabolite transport/metabolism (1.08%). Genes related to secondary metabolites were more abundant in the up-regulated group whereas those involved in defense were more abundant in the down-regulated group.

Transcript analyses of putative OsMADS26 downstream genes
We chose eight genes to examine the reliability of our microarray data (Table III). Four iron/ascorbate family oxidoreductase genes were found in the up-regulated group, and could be divided into two groups: 1-aminocyclopropane-1-carboxylate (ACC) oxidase genes involved in ethylene biosynthesis (A09021902 and A05041211), and putative flavanone 3-hydroxylase genes (A05011009 and B10022103). From these, we selected one ACC oxidase gene (A09021902) and one flavanone 3-hydroxylase gene (A05011009) for further confirmation. The A05011009 protein showed high homology to gibberellin β -hydroxylase. We also identified a lipoxygenase gene (A09032318), an NADPH oxidase gene (B03011909), and the S-adenosylmethionine decarboxylase (SAMDC) gene (A10031622), which function in the biosyntheses of JA, reactive oxygen species (ROS), and polyamine, respectively. In addition, a MAP kinase gene (A05011217) involved in hormone signaling/biosynthesis and two harpin-induced protein genes (B05032110 and A03011404) were examined.
For semi-quantitative RT-PCR analyses of these eight selected genes, seven-day-old OsMADS26GR plants and their WT segregants were treated with 1 µ M DEX for up to 9 h, and RNAs were prepared from their roots. The OsMADS26 and actin genes were included as controls.
www.plantphysiol.org on August 19, 2017 -Published by Downloaded from Copyright © 2008 American Society of Plant Biologists. All rights reserved.

13
As expected, OsMADS26 transcript was highly expressed in the transgenic roots, with that level increasing after DEX treatment (Fig. 5). Transcripts of these eight genes were induced posttreatment, and showed expression patterns similar to those obtained from the microarray analyses.
Therefore, these data support the reliability of the microarray results.

The relationship between OsMADS26 and biosynthesis genes associated with stress responses
The phenotypes observed in our ubi:OsMADS26 and ubi:OsMADS26GR plants were broadly correlated with stress responses. Microarray analyses demonstrated the up-regulation of several genes for the biosynthesis of stress-inducing molecules such as ET, JA, ROS, and polyamine (Supplementary Table II).
In ET biosynthesis, ACC synthase and oxidase are the most important genes in mediating the final two steps. While none of ACC synthase genes was changed significantly, four ACC oxidase genes were up-regulated in three experimental sets. The JA biosynthesis genes include lipoxygenase, allene oxide synthase (AOS), allene oxide cyclase (AOC), oxo-phytodienoic acid reductase (OPR), and JA carboxyl methyltransferase (JMT) (Agrawal et al., 2004). Our microarray analyses showed that the following JA biosynthesis genes were up-regulated: OsLOX3, OsAOS1, OsAOS4, OsAOS5, OsOPR2, OsOPR12, and OsOPR13. Among the genes involved in ROS production, NADPH oxidase mRNA was clearly up-regulated, while a gene homologous to aldehyde oxidase was downregulated. Regarding polyamine biosynthesis, only SAM decarboxylase was up-regulated while genes encoding ornithine decarboxylase, arginine decarboxylase, and spermidine synthase were not changed. To further elucidate the role of OsMADS26, we utilized the Arabidopsis system, in which expression is under the control of the CaMV35S promoter. Of our 105 kanamycin-resistant T1 transgenic plants (35S:OsMADS26 plants), 11 developed severe dwarfism, chlorosis, and tilted leaves (Fig. 6A). Their growth was halted and they eventually died without developing reproductive organs. The rest of the T1 plants, which produced fertile seeds, were used for further analyses.

Phenotypes of
T2 segregants of these transgenic lines were analyzed phenotypically. Among the 11 independent lines examined, those from Lines #1, #8, and #11 displayed a wide range of abnormal phenotypes (Fig. 6B, C). Generally, their plants were smaller but had more lateral roots.
Furthermore, Line 1 accumulated red pigments while Lines 8 and 11 developed twisted leaves.
Plants from Line #11 were only about one-third as large as the WT, and they showed a delayed rate of leaf emergence. OsMADS26 transcripts were detected in all the plants with abnormal phenotypes.
To determine whether the abnormal phenotypes were induced by JA, we checked the expression levels of AtMyc2, VSP, and PDF1.2 (Bell and Mullet, 1993;Benedetti et al., 1995;Boter et al., 2004;Lorenzo et al., 2004;Penninckx et al., 1998) whose expressions generally are induced by that compound. Our analysis showed that transcript levels for VSP were increased in the transgenic plants with medium or strong phenotypes ( Supplementary Fig. 7). In contrast, PDF1.2 transcripts were down-regulated in the transgenic plants in proportion to their phenotypic severity, whereas AtMyc2 expression was unaffected. Therefore, these results suggest that OsMADS26 controls the subsets of JA-inducible genes.

Inhibition of apical hook development in 35S:OsMADS26 Arabidopsis plants.
Ellis and Turner (2001)  dose-dependent manner while ethylene promotes such formations. Therefore, we employed this physiology to study any possible relationship between OsMADS26 and those hormones. As previously reported, MJ induced shorter roots and hypocotyls, and inhibited apical hooks, while ACC induced exaggerated development of the latter tissue (Fig. 7A). Proper hooks are defined as those where the angle between hypocotyl and cotyledon is <90°. When homozygous plants were grown in the dark, 86% of the transgenics did not have properly formed hooks (Fig. 7B). Moreover, when treated with 1 µ M MJ, all transgenic plants failed to develop apical hooks; ACC also did not induce drastic apical hook development. Therefore, these results suggest that some of the phenotypes observed in our OsMADS26-overproducing Arabidopsis plants are associated with MJ.

OsMADS26 transcript is more abundant in old tissues
OsMADS26 was the first of four rice genes identified in the AGL12 group. Its expression patterns have now been elucidated, with transcripts being detected in the roots, shoots, panicles, and inflorescences throughout all developmental stages (Shinozuka et al., 1999, Pelucchi et al., 2002.
In this study, we showed that the OsMADS26 transcript level was elevated in older leaves and roots, implying that this gene may be involved in senescence or maturation processes.  This suggests that the rice AGL12-group genes are functionally redundant.

Overexpression of OsMADS26 causes multiple stress responses in rice and Arabidopsis
To elucidate the role of OsMADS26, we regenerated transgenic rice plants over-expressing that gene.
Various phenotypes were displayed, such as defective growth, chlorosis, cell death, pigment accumulation, spotted leaves, and senescence. These were almost re-enacted in OsMADS26 overexpressing Arabidopsis plants, demonstrating the conserved role of this MADS-box gene in both model systems. We think that these phenotypes reflect the actual function of OsMADS26 because we employed an inducible system that showed the similar phenotypes to be independent of developmental stage. Therefore, the induced phenotypes are likely related to the action of OsMADS26. If the abnormalities had, instead, been artifacts due to disturbing the action of other proteins, we would have expected the influence to be linked with a particular growth stage. A number of overexpression analyses have been conducted previously to study gene function, especially when loss-of-function mutants do not provide any clues.
The phenotypes observed from the transgenics were similar to those previously reported for plants exposed to various stresses. In Arabidopsis, stresses mediated by heavy metals, nutrient deficiencies, and hypoxia induce the development of characteristic traits that include diminished leaf, shoot, and root elongation, as well as enhanced formation of lateral roots (see review by Potters et al., 2007). Similar abnormalities, e.g., chlorosis and cell death, are commonly observed in rice grown under extremely harsh conditions. Pigments are also accumulated in stressed plants (Harvaux and Kloppstech, 2001;Jordan et al., 1998). Although genes related to cell death and pigment accumulation were identified in our microarray analyses, no gene directly associated with chlorosis was detected. Therefore, the chlorosis phenotype seems to be more of an indirect effect compared with other abnormalities.

OsMADS26 may generate various stress mediators
Because stress phenomena are connected with various factors, including phytohormones and ROS, the OsMADS26-mediated response described here might be related to hormonal activity. We speculated that jasmonate is the most probable candidate because the phenotypes observed from our transformants were similar to ones from plants that over-express JA-inducible genes. Using a genetics screening system to isolate mutants that constitutively express the thionin (Thi  Ibuprofen and acetylsalicylic acid work as lipoxygenase inhibitors (Doares et al., 1995;Nojiri et al., 1996). This might be because total oxylipin contents were increased. Recently, Vellosillo et al. induction among JA, SA, and ethylene treatments also has been reported (Cheong et al., 2003;Sasaki et al., 2004). For example, BWMK1, up-regulated in our microarray analyses, was increased in response to SA, JA, and ethephon. Therefore, OsMADS26 may act as a common positive regulator for a subset of genes that respond to these hormones.
Ethylene is another possible candidate because some of our phenotypes were similar to those from plants treated with ET, in which four ACC oxidase genes were up-regulated. However, ACC synthase transcript levels did not change here. Because ACC synthesis is the rate-limiting step in ET production, the effect of OsMADS26 in the ET-mediated response is restricted to the regions where ACC synthase activity is high. In the apical hook, ET-mediated signaling seemed not to be activated because we did not find any ET-induced exaggeration of a hook. However, the leaf-curling phenotype observed in the 35S:OsMADS26 seedlings was similar to that of WT plants treated with ET.
Finally, the third candidate is ROS -this may be possible based on our data showing that NADPH oxidase transcript levels were up-regulated in the OsMADS26-overexpressing plants. ROS induces morphogenic responses that include defective growth and a relatively large number of lateral roots (Olmos et al., 2006). Enhanced ROS production is associated with a broad range of biotic and abiotic stresses, e.g., heat, UV-radiation, heavy metal, anoxia, and pathogen attacks (Apel and Hirt, 2004). Unexpectedly, a gene with high homology to gibberellin β -hydroxylase was upregulated. However, we did not study a relationship between gibberellin and OsMADS26 because this hormone is rarely involved in stress-related responses observed in the OsMADS26overexpressing plants.
OsMADS26 may directly bind to the promoter regions of these biosynthesis genes.
Alternatively, it might regulate these genes via cross talk between stress-mediators or by positive feedback mechanisms (Sasaki et al., 2001;Zhong and Burns, 2003;Chung and Choi, 2007 example, three JA biosynthesis genes --OsAOS1, OsAOC1, and OsOPR1 -are induced not only by JA itself but also by treatment with ET, abscisic acid, salicylic acid (SA), or hydrogen peroxide (Agrawal et al., 2002(Agrawal et al., , 2003a. Likewise, OsACO2 transcript levels are elevated in IAA-treated etiolated rice seedlings whereas OsACO3 mRNA is greatly accumulated following ET exposure (Chae et al., 2000).

OsMADS26 regulates various stress-induced genes
Microarray analyses have produced a global spectrum for the genes regulated by JA, ET, and ROS.
MJ differentially controls the transcription of genes involved in oxidative bursts and programmed cell death, such as those for catalase, glutathione S-transferase, and cysteine protease (Schenk et al., 2000). Numerous genes associated with cell rescue, disease, and defense mechanisms have been identified as early ET-regulated genes (de Paepe et al., 2004). Extensive comparisons have demonstrated redundant and specific roles for ROS in connection with stresses (Gadjev et al., 2006). Furthermore, considerable cross talk occurs among these signaling pathways. Schenk et al. (2000) have reported that 50% of the genes induced by ET are also induced by MJ. Transcriptome analysis of Col;35S:ERF1 transgenic plants and ET/JA-treated WT plants has further revealed a large number of genes responsive to both ET and JA (Lorenzo et al., 2003). In the flu mutant, which overproduces 1 O 2 , the ethylene-responsive element-binding proteins are highly induced, indicating cross talk between 1 O 2 and ethylene-signaling (Gadjev et al., 2006).
Our microarray analyses showed that genes inducible by JA, ET, or ROS were upregulated in transgenic plants over-expressing OsMADS26. These include not only the biosynthesis genes already discussed here, but also many putative downstream genes, such as cysteine proteinase, S-adenosylmethionine decarboxylase, protease inhibitor, peroxidase, and MAP-kinase genes www.plantphysiol.org on August 19, 2017 -Published by Downloaded from Copyright © 2008 American Society of Plant Biologists. All rights reserved. (Biondi et al., 2001;De Paepe et al., 2004;Schenk et al., 2000;Zhao and Chye, 1999) (Table III).

21
Expression profiles for 22 rice peroxidase genes have revealed that many of them respond to disease, wounding, SA, JA, and ACC (Sasaki et al., 2004). A MAP kinase gene (A05011217), designated as BWMK1 (Genbank Accession Number AF177392), is induced not only by blast, wounding, and H 2 O 2 , but also by the phytohormones SA, JA, and ethephon (Cheong et al., 2003;He et al., 1999). The two putative flavanone 3-hydroxylase genes found from our microarrays may also cause pigment accumulation. In other species, MJ induces the accumulation of anthocyanin in soybean and Arabidopsis (Franceschi and Grimes, 1991;Jung, 2004). Furthermore, transcripts involved in anthocyanin production are co-regulated in response to O 2 ·¯, whereas H 2 O 2 negatively impinges on their expression (Gadjev et al., 2006).
Our analyses also showed the activity of HR-related genes that encode a harpin-induced protein or a cell death-associated protein. Harpin from Erwinia amylovora causes the HR response (Wei et al., 1992). A putative cell death-associated gene has close homology with hsr203J, which is expressed in the leaves of Nicotiana tabacum cv. Samsun NN infected with Ralstonia solanacearum 8107 (Kiba et al., 2003). An aldo/keto reductase family gene also was induced here. Members of the aldo/keto reductase superfamily can detoxify a major lipid peroxide degradation product, 4hydroxynon-2-enal (HNE) (Vander Jagt et al., 1995), and the rice aldo/keto reductase gene is induced in vegetative tissues in response to PEG-mediated water stress and salinity (Karuna Sree et al., 2000).
Three members belonging to the protease inhibitor family were down-regulated, suggesting their negative roles in stress-related responses. These proteins contain a domain commonly found in trypsin-alpha amylase inhibitors, seed storage proteins, and lipid transfer proteins (Rico et al., 1996). Some peroxidase genes were also down-regulated, perhaps causing the cell death signal to be amplified by reducing H 2 O 2 scavenging. Down-regulation of peroxidase genes by ethylene has previously been reported (de Paepe et al., 2004). Altogether, our results indicate that OsMADS26 controls various stress responses.

Plant materials and chemical treatments
Oryza sativa var. japonica cv. Dongjin and the Columbia ecotype of Arabidopsis thaliana were used.
Rice seeds were surface-sterilized and seedlings were grown at 28°C on gauze embedded in sterile Murashige and Skoog (MS) media containing 0.2% agar, 3% sucrose, and 0.01% myo-inositol.
Plants were grown to maturity in a greenhouse supplemented with artificial lighting during the winter period. DEXamethasone was dissolved in 95% alcohol at 1 mM and an appropriate amount was added to the growth media to arrive at the desired final concentration. Methyl jasmonate and aminocyclopropane-1-carboxylate (ACC) were dissolved in 95% alcohol and sterilized water, respectively, at 10 mM, before a suitable amount was added to MS solid media containing 0.2% agar, 3% sucrose, and 0.01% myo-inositol. For Arabidopsis, MJ and ACC were added to a 1/2 Gamborg B5 agar (0.8%) medium supplemented with 1% sucrose. The full-length cDNA clone of OsMADS26 (GenBank Accession Number AB003326) was isolated by nested PCR, using the following four primers: forward 1, 5'-atcaagcttggagctatcgatcatcaagc-3'; forward 2, 5'-atcaagcttgagacttatcttgatcgatgg-3'; reverse 1, 5'-ttgggtaccaaataaggtacatcagaatagc-3'; and reverse 2, 5'-ttgggtaccgttagaaggaatagcccatc-3'. These primers contained the Hind III and Asp718 restriction enzyme sites for subsequent cloning. The PCR product was first cloned into pBluescript SK-(Stratagene, La Jolla, CA). Afterward, the cDNA was sub-cloned into the pGA1611 binary vector between the maize ubi promoter and the nos terminator for the sense construct (Lee et al., 1999;Kim et al., 2003). For the anti-sense construct, we used the region between 404 and 900 of OsMADS26. For the DEX-inducible system, the OsMADS26 stop codon was changed to the Asp718 site by using the reverse primer (5'-ttgggtaccgaaggaatagcccatctcc-3'). The rat GR gene (AY066016) was inserted into that Asp718 site, generating an in-frame fusion between the two molecules. For Arabidopsis transformation, the pGA1535 binary vector with the CaMV35S promoter and a kanamycin selectable marker was used to sub-clone OsMADS26, with Hind III and Asp718.

Transformation
Rice transformation was performed according to the Agrobacterium-mediated methods described by Jeon et al. (1999) and Lee et al. (1999). All transgenic plants were grown in glass tubes, and then transferred to a confined paddy field. The Columbia ecotype was used for Arabidopsis transformation using floral dip method (Clough and Bent, 1998