Seeds of wild-type rice (Oryza sativa L. “Nipponbare”)the osarf24-1 mutant, and transformants (described below) were sterilized in 1% NaClO for 30 min and sown on Murashige and Skoog agar medium. Seedlings were grown in a growth chamber at 28˚C under continuous light for 2 weeks. For morphological characterization, seedlings were transplanted and grown in the paddy field at the Experimental Farm of Ishikawa Prefectural University. For gene expression analyses, seedlings were selected for uniformity of growth and adapted to hydroponic culture for 2 days before treatment. IAA treatment (20 mM) was carried out by adding IAA to the culture medium.

Total RNA was extracted from whole seedlings of wildtype and mutant rice and from mature leaves of transgenic rice by using an RNeasy Plant Mini Kit (Qiagen, Venlo, Netherlands). Single-strand cDNAs were synthesized by using the Advantage RT-for-PCR Kit (Clontech, Palo Alto, CA, USA). Quantitative RT-PCR was performed with an iCycler iQ real-time PCR system (BioRad Laboratories, Hercules, CA, USA). The primer sequences were 5’-CAGGAAGCTGGTGT GTTGTC-3’ and 5’-CTTGATCAGGCGTGGCTGTG-3’ for OsARF23, 5’-AATGACGCCTGACATCACAC-3’ and 5’-GCTTG ATAAGACTCGATGAGG-3’ for OsARF24, 5’-ACCA AGAGCCGCTCAATGAG-3’ and 5’-ATCACACGTG GGCGAACATC-3’ for OsIAA1, 5’-GATGAACAGGC GGTCGCTGC-3’ and 5’-GGCTC CGGTAGTAGCTTG TG-3’ for OsGH3-1, and 5’-CGCC AGTTTGGTCGCT CTCGATTTCG-3’ and 5’-TCAGGA GCTCCGTGCTC TTCTGGTAC-3’ for Histone H3. These primers specifically amplified the target gene sequences. Expression levels were normalized against the values obtained for Histone H3, which was used as an internal reference gene. For the gene expression experiments, we performed 3 biological repeats.

Full-length OsARF24 cDNA was inserted in the sense orientation into the pET-32a expression vector (Novagen, Madison, WI, USA) to generate a thioredoxin fusion protein when expressed in BL21(DE3) E. coli cells (Stratagene, La Jolla, CA, USA). The recombinant protein was purified by using Talon Metal Affinity Resin (Clontech Co., Palo Alto, CA). OsIAA1 promoter fragments containing WT or MT AuxRE were amplified by PCR with rice genomic DNA. The primer sequences were 5’-GGTTGAAATTGGAACGATGTG-3’ and 5’-G GAACTTTCATCTACTACTAC-3’ for OsIAA1 AuxRE (WT), and 5’-TTTGGATTCTCCATTATGAGAAAATC AAAACATGGTTTTTT-3’ and 5’-TTAATAAAAAAC CATGTTTTGATTTTCTCATAATGGAGAATCC-3’ for generating the AuxRE mutation (MT). The amplified fragments were cloned into pBluescript II SK (Stratagene) and their identities were confirmed by sequence analysis. The PCR-amplified fragments were excised with restriction endonucleases, purified by 10% PAGE, and labeled with biotin using a Biotin 3’ End DNA Labeling Kit (Pierce, Rockford, IL, USA). The electrophoresis mobility shift assay was performed by using a LightShift Chemiluminescent EMSA Kit (Pierce).

The entire OsARF24 coding region was inserted between the rice Actin promoter and the nopaline synthase polyadenylation signal of the hygromycin-resistant binary vector pAct-Hm2. This vector was modified from pBIH1 [15] to contain a rice Actin promoter. The resulting construct was introduced into Agrobacterium tumefaciens strain EHA105, and Agrobacterium-mediated transformation of rice (O. sativa L. “Nipponbare”) was performed as described [16]. Transgenic plants were selected on Murashige and Skoog agar medium containing 50 mg·L^-1 hygromycin, and then grown in a greenhouse at 28˚C under ambient light conditions.

osarf24-1 is a mutant of OsARF24 caused by insertion of the Tos17 retrotransposon. osarf24-1 showed a reduction in plant height (the height of osarf24-1 was 91% that of the wild-type, n = 10, P < 0.001; Figure 1(a) and Table 1) and a reduction in the leaf angle of flag leaves from the wild-type value of 22.3˚ to 13.0˚ (n = 10, P < 0.001). This mutant also showed the narrow leaf phenotype typical of auxin-deficient or auxin-insensitive rice mutants (Figure 1(b)) [17,18], and the ratio of blade width to blade length of the flag leaves was reduced (0.031, versus 0.050 in wild-type, n = 10, P < 0.001). Wild-type rice leaves form in a distichous alternate phyllotactic manner, and successive leaves develop on opposite sides of the shoot apical meristem with 180˚ of divergence (Figure 1(c), left plant). However, disordered phyllotaxis was found in the osarf24-1 mutant. In these seedlings, leaf divergence was not 180˚, but instead twisted gradually (Figure 1(c), right plant). Because similar twisted leaf development was also observed in 2,4-D-treated wild-type rice and constitutively active Aux/IAA repressor transgenic rice [19], an adequate auxin signal is important for normal rice leaf development. These phenotypes of osarf24-1 suggest that the osarf24-1 mutant has some defects in auxin response.

Sequence analysis revealed that osarf24-1 had an insertion of Tos17 in exon 13 (Figure 1(d)). In osarf24-1,

Table 1. Phenotype of the osarf24-1 mutant.

^aEach column represents mean ± s.d. of 10 independent plants.

Tos17 was inserted into the part of the gene encoding the C-terminal dimerization domain of the OsARF24 protein at the 773th leucine residue (Figure 1(d)). This insertion altered the amino acid sequence in the C-terminal dimerization domain of the OsARF24 protein; specifically, the C-terminal half of domain IV, which has a similar amino acid sequence to Aux/IAA proteins, was changed and truncated (Figure 1(e)). Therefore, osarf24- 1 is considered to be a truncation mutant of the C-terminal dimerization domain in OsARF24 protein.

The treatment of rice seedlings with IAA induced an increase in the expression of auxin-response genes OsIAA1 and OsGH3-1 [20,21]. After 10 min, IAA treatment of wild-type seedlings increased the expression of both OsIAA1 and OsGH3-1 to 1.5 times the levels in untreated seedlings (Figure 2). The expression levels gradually increased until 60 min after IAA treatment to 5.5 and 5.9 times those in untreated seedlings, respectively. In osarf24-1 seedlings, the steady-state levels of OsIAA1 and OsGH3-1 mRNA were 71% and 66%, respectively, of those in the wild-type seedlings (Figure 2). The kinetics of the increases in OsIAA1 and OsGH3-1 mRNA levels of osarf24-1 after IAA treatment were similar to those of wild-type rice. At 10 min after IAA treatment, the levels of OsIAA1 and OsGH3-1 mRNA in osarf24-1 seedlings were 1.0 and 0.9 times, respectively, those in the untreated wild-type control, and at 60 min after the treatment, had increased to 3.9 and 3.5 times, respectively. The levels of both OsIAA1 and OsGH3-1 mRNA in IAA-treated osarf24-1 were lower than those in the wild-type. However, the ratios of the relative expression levels of OsIAA1 and OsGH3-1 mRNA in IAA-treated osarf24-1 to untreated osarf24-1 were approximately 1.4 after 10 min and 5.3-5.4 after 60 min, demonstrating that osarf24-1 still has some ability to respond to IAA. Although multiple OsARF proteins are considered to regulate the expression of OsIAA1 and OsGH3-1, our results suggest that the osarf24-1 mutant is less sensitive to auxin than wild-type rice. Because the C-terminal dimerization domain in ARF proteins is considered to function in hetero-dimerization among ARF and Aux/IAA proteins [4,9,10], it is possible that the OsARF24 repressor protein with a C-terminal dimerization domain truncation in the osarf24-1 mutant cannot interact with Aux/ IAA proteins in rice, and therefore that this mutant

OsARF24 protein represses auxin-response genes constitutively.

Among the 25 ARF genes in rice, the deduced amino acid sequence of OsARF24 is most closely related (71.4% identity) to OsARF23 (previously designated as OsARF1 [22]), and phylogenetic analysis grouped OsARF24 and OsARF23 with an Arabidopsis ARF, AtARF2 [13]. The predicted open reading frames (ORFs) of OsARF23 and OsARF24 encode proteins of 836 and 840 amino acids, respectively. The similarity of the deduced amino acid sequences of OsARF23 and OsARF24 is 71.4%, and the sequences are most closely related (55.1% and 52.7% similarity, respectively) to Arabidopsis AtARF2/HSS (At5g62010, 853 amino acids) [23]. The structures of OsARF23 and OsARF24 are similar to that of AtARF2/ HSS throughout their lengths: three domains found in repressor ARFs-a DNA-binding domain, an SPL-rich repression domain, and a C-terminal dimerization domain that is related in amino acid sequence to domains III and IV in Aux/IAA proteins-are highly conserved. Because the SPL-rich repression domain is a characteristic of repressor ARF proteins [24], both OsARF23 and OsARF24 are considered to function as transcriptional repressors [13,14].

Quantitative reverse-transcription PCR analysis revealed that OsARF24 and OsARF23 were expressed at different levels in all the organs of wild-type rice that we tested, including the vegetative shoot apices, leaf sheaths, leaf blades, elongating internodes, roots, inflorescences (immature panicles), and panicles at flowering time (Figure 3). Both genes were expressed at the highest level in inflorescences. OsARF24 was also preferentially expressed in panicles at flowering time, elongating internodes, leaf sheaths, and leaf blades, and at low levels in vegetative shoot apices and roots. OsARF23 was also preferentially expressed in elongating internodes, leaf sheaths, leaf blades, roots, and panicles at flowering time, and at a low level in vegetative shoot apices.

Previous observations indicate that the expression levels of some ARF genes were not affected by IAA treatment, and those of the others were either increased or decreased [10,13]. Thus, we examined the effect of exogenously applied auxin on the expression of OsARF23 and OsARF24. IAA treatment increased the expression of both OsARF23 and OsARF24 to 1.5 times that in untreated plants within 10 min (Figure 4). However, the kinetics of the increases in OsARF23 and OsARF24 mRNA levels after IAA treatment were different. The

expression level of OsARF23 gradually increased until 60 min after IAA treatment to 2.5 times that in untreated plants, whereas the expression level of OsARF24 reached a maximum within 10 min and was maintained at that level until 60 min after IAA treatment. These results suggest that the expression of OsARF23 and OsARF24 is regulated by different mechanisms.

The AuxRE has been identified in the promoters of some early auxin-response genes, and ARFs bind the AuxRE to regulate the transcription of these genes [24]. To examine whether the recombinant OsARF24 protein interacts with the AuxRE in the OsIAA1 promoter, we performed electrophoresis mobility shift assays (Figure 5). OsARF24 bound to a 377-bp OsIAA1 fragment containing the intact AuxRE (WT). The amount of retarded complex was reduced by the addition of increasing concentrations of unlabeled WT fragment as a competitor (“WT comp”, in Figure 5). The binding of OsARF24 with the WT fragment was not affected by addition of the unlabeled OsIAA1 fragment containing the mutated AuxRE sequence as a competitor (“MT comp”, in Figure 5). These results demonstrate that OsARF24 can bind to the promoter sequence of OsIAA1, and that this interaction depends only on the presence of an intact AuxRE sequence in the promoter fragment.

To assess the activity of the OsARF24 gene product in vivo, we fused the full-length OsARF24 cDNA to the rice Actin promoter in the sense orientation (Act::OsARF24) and introduced the construct into wild-type rice by Agrobacterium-mediated gene transfer (Figure 6). In this experiment, we selected 3 transgenic lines in which the OsARF24 expression level was doubled relative to wild-type (Figure 7). The Act::OsARF24 transformants showed a slight reduction in plant height (the average plant height of 5 plants from each of 3 different transgenic lines was 85% that of wild-type; Figure 6(a)). These transformants also showed the narrow leaf phenotype typical of auxin-deficient or auxin-insensitive rice mutants (Figure 6(b)), and the ratio of blade width to blade length of the flag leaves was reduced to 0.034 (the average of 5 flag leaves from each of 3 different transgenic lines), versus 0.050 in wild-type. Disordered phyllotaxis was also found in Act::OsARF24 transgenic seedlings (Figure 6(c)). These results suggest that the Act:: OsARF24 transformants have some defects in auxin response, as was the case for the osarf24-1 mutant.

We also examined the expression of auxin-response genes in the Act::OsARF24 transformants. The steadystate levels of OsIAA1 and OsGH3-1 mRNA in the leaves of Act::OsARF24 transformants decreased to about 50% and 20%, respectively, of those in the leaves of wild-type plants (Figure 7), confirming that OsARF24, which has an SPL-rich middle region, acts as a transcriptional repressor of auxin signaling in rice.

As previously mentioned, OsARF24 protein, which contains the SPL-rich repression domain, is considered to be

an ARF that functions as a transcriptional repressor [13, 14]. In our experiments, transgenic rice overexpressing the OsARF24 cDNA showed auxin-deficient phenotypes including dwarf stature, narrow leaf, and aberrant phyllotaxis. In addition, the expression levels of auxin-response genes OsIAA1 and OsGH3-1 were decreased in these transformants. These results strongly support the hypothesis that OsARF24 acts as a repressor ARF. However, the Tos17 retrotransposon insertion mutant of OsARF24, osarf24-1, also showed both auxin-deficient

phenotypes and decreased levels and auxin responses of OsIAA1 and OsGH3-1 expression. Although another rice ARF, OsARF23, shows high amino acid sequence similarity with OsARF24, the expression of OsARF23 and OsARF24 is regulated by different mechanisms in various organs of wild-type rice plants and in response to IAA treatment. Based on these results, we consider that OsARF23 and OsARF24 do not function redundantly in rice.

Most ARF proteins contain a C-terminal dimerization domain related to domains III and IV in Aux/IAA proteins [4,7,10]. The C-terminal dimerization domains in both ARF and Aux/IAA proteins are protein-protein interaction domains that allow homoand heterodimerization of ARF proteins and hetero-dimerization among ARF and Aux/IAA proteins [4,9,10]. Although ARF repressors can dimerize via their C-terminal dimerization domains, ARF repressor-Aux/IAA and ARF represssorARF activator interactions are much weaker than ARF activator-Aux/IAA and ARF activator-ARF activator interacttions, and it remains unclear whether Aux/IAAs interact with ARF repressors, or whether ARF repressors interact with ARF activators, to regulate target gene expression in plants [3,10,12,14].

In the osarf24-1 mutant, a Tos17 insertion in OsARF24 altered the amino acid sequence in the C-terminal dimerization domain of the OsARF24 protein, and the C-terminal half of domain IV was truncated. This mutation may reduce the formation of repressor ARF dimers; however, this does not explain why the osarf24-1 mutant showed reduced sensitivity to auxin because both monomers and dimers of ARF repressors can target and repress the expression of auxin-response genes [3]. Another hypothesis is that the C-terminal dimerization domain, especially the C-terminal half of domain IV, affects the stability or the repressor activity of OsARF24. The Arabidopsis ARF1 repressor is targeted for proteasomal degradation via a different set of machinery than that used for Aux/IAA degradation [25]. In addition, we cannot exclude the possibility that the weak interactions found between ARF repressors and Aux/IAA proteins in vitro are enough to function in vivo, but that the truncated OsARF24 repressor protein in the osarf24-1 mutant is unable to interact with rice Aux/IAA proteins and thus constitutively represses auxin-response gene expression. We envisage that detailed analyses of the osarf24-1 mutant and OsARF24 protein will help to reveal the function of the C-terminal dimerization domain in ARF repressors.