Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
•MicroRNA (miRNA)-mediated regulation of auxin signaling components plays a critical role in plant development. miRNA expression and functional diversity contribute to the complexity of regulatory networks of miRNA/target modules.
•This study functionally characterizes two members of the rice (Oryza sativa) miR393 family and their target genes, OsTIR1 and OsAFB2 (AUXIN SIGNALING F-BOX), the two closest homologs of Arabidopsis TRANSPORT INHIBITOR RESPONSE 1 (TIR1).
•We found that the miR393 family members possess distinctive expression patterns, with miR393a expressed mainly in the crown and lateral root primordia, as well as the coleoptile tip, and miR393b expressed in the shoot apical meristem. Transgenic plants overexpressing miR393a/b displayed a severe phenotype with hallmarks of altered auxin signaling, mainly including enlarged flag leaf inclination and altered primary and crown root growth. Furthermore, OsAFB2- and OsTIR1-suppressed lines exhibited increased inclination of flag leaves at the booting stage, resembling miR393-overexpressing plants. Moreover, yeast two-hybrid and bimolecular fluorescence complementation assays showed that OsTIR1 and OsAFB2 interact with OsIAA1.
•Expression diversification of miRNA393 implies the potential role of miRNA regulation during species evolution. The conserved mechanisms of the miR393/target module indicate the fundamental importance of the miR393-mediated regulation of auxin signal transduction in rice.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The plant hormone auxin influences virtually every aspect of growth and development in plants, including stem elongation, phototropic and gravitropic responses, apical dominance, and lateral and adventitious root formation (Vanneste & Friml, 2009). Auxin promotes the degradation of auxin/indole-3-acetic acid (Aux/IAA) transcriptional repressors, thereby allowing AUXIN RESPONSE FACTORS (ARFs) to activate the transcription of auxin-responsive genes (Worley et al., 2000; Gray et al., 2001; Tiwari et al., 2001). Degradation of Aux/IAA transcriptional repressors requires TRANSPORT INHIBITOR RESPONSE PROTEIN 1 (TIR1), an auxin receptor (Dharmasiri et al., 2005; Kepinski & Leyser, 2005). Auxin binding to TIR1, the F-box subunit of the ubiquitin ligase complex SCF (TIR1), stabilizes the interaction between TIR1 and the Aux/IAA proteins. This interaction results in Aux/IAA ubiquitination and subsequent degradation (Gray et al., 2001). Auxin signaling components have been conserved throughout land plant evolution, and have proliferated and specialized to control specific developmental processes (Chapman & Estelle, 2009). Research on monocots and eudicots has shown similar mechanisms for auxin signaling in these divergent species. However, monocots have a very differ ent anatomy from dicots (McSteen, 2010), and many of the characteristics that distinguish monocots and dicots involve structures whose development is controlled by auxin, such as cotyledons, the root system and the leaf vasculature. It is not yet clear whether auxin controls the differences in morphology seen in dicots vs monocots. So far, both conservation and diversification of mechanisms of auxin signal transduction have been discovered. There are 31 members in the Aux/IAA gene family in rice (Jain et al., 2006), 24 of which are regulated by auxin (Song et al., 2009a) and three of which have been functionally characterized. OsIAA1/3 transcription is induced by auxin and suppressed by light (Thakur et al., 2001), and the protein has a short half-life and is degraded by the proteosome (Thakur et al., 2005), similar to Aux/IAA proteins in Arabidopsis. Overexpression of OsIAA1 causes defects in root development and an additional defect of enlarged leaf angle (Song et al., 2009b). The F-box gene family, including homologs of the TIR1 gene family, has been catalogued in rice (Jain et al., 2007). However, it remains unclear whether the TIR1 and TIR1-like auxin receptor interaction with Aux/IAA proteins in Arabidopsis is conserved in other plants, particularly monocots.
Conserved microRNAs (miRNAs) target conserved homologous genes in diverse plant species (Jones-Rhoades et al., 2006). miR393 is a conserved miRNA family discovered in many plants. In Arabidopsis, four F-box genes, TIR1, AFB1, AFB2 and AFB3 (AUXIN SIGNALING F-BOX), have been validated as targets of miR393 (Jones-Rhoades & Bartel, 2004; Navarro et al., 2006; Parry et al., 2009). Loss of TIR1 has a modest effect on auxin responses and plant development; however, quadruple tir1/afb mutants exhibit a severe embryonic phenotype and a variety of growth defects, indicating that these genes have overlapping functions (Dharmasiri et al., 2005). Recent studies have demonstrated that the miR393/target regulatory module plays a role in root system architecture during nitrate responses (Vidal et al., 2010), leaf development (Si-Ammour et al., 2011), auxin response (Parry et al., 2009; Chen et al., 2011) and antibacterial resistance in response to pathogen attack (Navarro et al., 2006). In rice, the miR393 family is encoded by two loci, MIR393a and MIR393b (Jones-Rhoades & Bartel, 2004; Archak & Nagaraju, 2007). Recently, conserved miRNA targets OsTIR1 (Os05g05800) and OsAFB2 (Os04g32460), which encode putative TIR1-like proteins, were identified by degradome sequencing (Li et al., 2010; Zhou et al., 2010) and 5′-rapid amplification of cDNA ends (5′-RACE) (Xia et al., 2012). Genome-wide analyses have revealed that Oryza sativa and Arabidopsis thaliana possess apparently divergent patterns of miRNA regulatory systems (Zhang et al., 2011). To date, however, the expression patterns of miR393 in monocots remains virtually unknown. Therefore, it is necessary to examine whether diversified expression has occurred in the rice miR393 family and how it affects growth and development in monocots.
In this study, we report the characterization of rice miR393 and knockdown of its two target genes, OsTIR1 and OsAFB2, the two closest homologs of Arabidopsis TIR1. Our results reveal conserved roles of the miR393/target module in auxin signal transduction between Arabidopsis and rice, but distinctive expression patterns and functions in growth and development in rice.
Materials and Methods
Plant materials and growth conditions
Rice (Oryza sativa L. ssp. japonica cv. Nipponbare) was used in this study. Plants were cultivated in an experimental field during natural growing seasons. For root measurements in seedlings, sterilized seeds were germinated in the dark for 2–3 d and then transferred to nutrient solution culture (Yoshida et al., 1976) at pH 5.0–6.0 in the glasshouse at 28°C with 16 h of light and 8 h of dark. For expression pattern analyses, various tissues were collected at different stages when the seedlings were grown in normal rice solution culture.
Gene constructs and rice transformation
To generate the 35S:miR393 constructs, 583- and 747-bp fragments surrounding the miRNA sequence including the fold-back structure were amplified from genomic DNA with the specific primers for 35S:miR393a and 35S:miR393b, respectively. The amplified fragments were sequenced and subcloned into KpnI and BamHI sites downstream of the Cauliflower mosaic virus (CaMV) 35S promoter in pCAMBIA13011 (a derivative of pCAMBIA1301 carrying the 35S promoter and the RBS terminator).
For the promoter:β-glucuronidase (promoter:GUS) constructs, 2.0- and 2.5-kb fragments upstream from the predicted fold-back of miR393 were amplified for ProMIR393a:GUS and ProMIR393b:GUS, respectively. The fragments were then transferred into the pENTR/D-TOPO vector (Invitrogen, Carlsbad, CA, USA) and subsequently into the destination vector pHGWFS7 containing an Egfp: uidA gene fusion by LR clonase reactions according to the manufacturer’s instructions.
For RNA interference (RNAi) constructs, 885- and 751-bp fragments were amplified from OsTIR1 and OsAFB2, respectively. The fragments were inserted into the pENTR/D-TOPO vector (Invitrogen) and subsequently cloned into pH7GWIWG2 (I) by LR clonase reactions.
MIM393 was generated as a target-mimic construct. The 24-nucleotide, miR399-complementary motif in IPS1 (Franco-Zorrilla et al., 2007) was exchanged for a sequence complementary to miR393. The construct was subcloned into KpnI and BamHI sites under the control of the 35S promoter in pCAMBIA13011.
Rice transformation was performed by the Agrobacterium tumefaciens-mediated co-cultivation approach. Transformed calli were selected on hygromycin-containing medium. Homozygous T3 seeds were used for further study.
All primers used in this work are listed in Supporting Information Table S1. The gene constructs used for rice transformation were verified by sequencing.
DNA gel blot analysis
For DNA gel blot analysis, total DNA was isolated from approximately 1 g of leaf tissues from transformed plants, as described previously. DNA (20 μg) was digested with EcoRI at 37°C overnight, separated on 1% (w/v) agarose gel and then transferred to a nylon membrane. Subsequent hybridization with a digoxigenin-labeled HPT-specific probe was performed using the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche) according to standard protocols.
Total RNA was isolated with the TRIzol RNA Extraction Kit (Invitrogen), and treated with RNAse-free DNAse I (TaKaRa, Dalian, China). Real-time PCR was performed by the Mastercycler ep realplex system (Eppendorf, Hamburg, Germany) with the SYBR PrimeScript RT-PCR Kit (Perfect Real Time; TaKaRa). The gene-specific primers are listed in Table S1. Primers for OsTIR1 and OsAFB2 were designed from each side of the miR393 cleavage site, and relative transcript levels were normalized using OsActin as a standard.
Small RNA blot analysis
Total RNA was extracted from roots and leaves at various stages, flag leaves and panicles at the booting stage using TRIzol reagent (Invitrogen). Thirty micrograms of total RNA were fractionated on 17% polyacrylamide gels under denaturing conditions (7 M urea). Blots were hybridized using digoxigenin end-labeled LNA oligonucleotides (Exiqon, Vedbaek, Denmark) complementary to the miR393a sequence.
Histochemical detection of GUS activity
Histochemical localization of GUS staining was performed by incubating tissues and organs in a solution of 1 mg ml−1 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid (Sigma), 1 mM potassium ferricyanide, 0.1% Triton X-100, 0.1 M sodium phosphate buffer, pH 7.0, and 10 mM EDTA overnight at 37°C, followed by clearing with 70% ethanol.
Auxin resistance analysis
Rice seeds were dehusked and sterilized with 70% ethanol for 5 min and then with 20% sodium hypochlorite (v/v) for 30 min, rinsed five times with sterile water, and placed on N6 medium supplemented with or without 100 nM 2,4-dichlorophenoxyacetic acid (2,4-D; Sigma), and incubated at 28 ± 1°C in a growth chamber (12/12-h photoperiod and 250 μmol s−1 m−2 light intensity). After 2 wk, the seedlings were observed and measured. Photographs were taken with a Nikon D80 camera (Nikon, Ayutthaya, Thailand).
Alternatively, sterilized rice seeds were germinated in the dark for 2–3 d and then transferred to moistened filter paper. To evaluate auxin resistance, seedlings with a root length of approximately 3 cm were grown hydroponically, supplemented with various concentrations of 2,4-D (0, 50, 100 and 200 nM). The root length was measured at different time periods as indicated. More than 20 seedlings were used in each treatment. ANOVA followed by Tukey post-tests was performed to determine the significant differences.
The OsTIR1:GFP and OsAFB2:YFP fusions were obtained by Gateway cloning (Invitrogen). The full-length cDNAs of stop codon-less OsTIR1 and OsAFB2 were amplified using KOD FX DNA Polymerase enzyme (Toyobo, Kyoto, Japan). The cDNAs were transferred into pENTR/D-TOPO vector (Invitrogen) and subsequently cloned into pH7FWG2 and pB7YWG2 (Karimi et al., 2002), respectively, by LR clonase reactions. The fusion constructs were transformed into onion epidermal cells by particle bombardment. After transformation, onion epidermal cells were incubated at 22°C on Murashige and Skoog salt (MS) plates in the dark for 16 h before examination. Enhanced green fluorescent protein (EGFP), yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP) fluorescence and differential interference contrast (DIC) images were visualized by a Zeiss Leica TCS SP5 confocal microscope (Mannheim, Germany).
Yeast two-hybrid assay
The vectors and strains used for the yeast two-hybrid assay were provided in the Matchmaker GAL4 Two-Hybrid System 3 (Clontech, CA, USA). The yeast two-hybrid assay and α-galactosidase activity measurement were performed according to the Yeast Protocol Handbook (Clontech, CA, USA). The BD-OsTIR1, BD-OsAFB2 and AD-OsIAA1 plasmids were constructed by inserting PCR fragments of full-length cDNAs into the appropriate plasmids. The PCR fragment of OsTIR1 was amplified with the specific primers containing an EcoRΙ or SalI site. The resulting fragment was digested with EcoRΙ and SalI, and cloned into pGBKT7 to generate plasmid BD-OsTIR1. The PCR fragment of OsAFB2 was cloned into pGBKT7 to generate plasmid BD-OsAFB2. The PCR fragment of OsIAA1 was amplified with the primers containing an NdeI or XhoI site, and cloned into pGADT7 to generate plasmid AD-OsIAA1. The plasmids were transformed into yeast strain AH109. All primers used for yeast two-hybrid assays are listed in Table S1.
The cDNAs of OsTIR1/OsAFB2 and OsIAA1 were inserted into the pENTR/D-TOPO vector and subsequently cloned into pUGW2-nYFP and pUWG2-cYFP (Nakagawa et al., 2007), respectively, by LR clonase reactions (Invitrogen). As a result, OsTIR1/OsAFB2 were fused to the split N-terminal YFP fragment and OsIAA1 to the C-terminal YFP fragment. The fusion constructs were transformed into onion epidermal cells by particle bombardment. Fluorescence and DIC images were visualized by a Zeiss Leica TCS SP5 confocal microscope.
Distinctive expression patterns of miR393 family members
The rice genome contains two miRNA393 genes: OsmiR393a and OsmiR393b (Jones-Rhoades & Bartel, 2004; Navarro et al., 2006; Parry et al., 2009). The mature OsmiR393a and OsmiR393b are 21- and 22-nucleotide RNAs, respectively, and their sequences are identical, except that OsmiR393b has an extra U nucleotide at its 3′ end (Fig. S1). To characterize the expression pattern of the two miR393 family members, we analyzed miR393 expression via small RNA blots in various organs and tissues. Interestingly, the miRNA expression varied considerably between miR393a and miR393b in wild-type plants (Fig. 1a). Mature miR393a was detected in the roots of young rice seedlings at the one-leaf and three-leaf stages (1L-R and 3L-R), but was undetectable in leaves (1L and 3L). By contrast, miR393b expression was high in aerial organs, but undetectable in roots. miR393b accumulated in young leaves (1L and 3L), flag leaves (FL) and inflorescence tissues (BP), with the highest level in flag leaves. Thus, the accumulation patterns were different between miR393a and miR393b, especially in aerial and underground organs in rice.
To quantitatively investigate the transcription levels of the miR393 family members, we examined MIR393a and MIR393b precursors in various tissues by quantitative RT-PCR. Quantitative RT-PCR analysis showed that the transcripts of the MIR393a precursor were present at high levels in the roots of seedlings (Fig. 1b), whereas the MIR393b precursor was detected at a high level in young leaves (2L) of seedlings, and steadily decreased throughout plant growth (Fig. 1c). Therefore, the levels of MIR393a and MIR393b precursors were consistent with the accumulation of their mature miRNAs in various tissues. The transcription patterns demonstrate that MIR393a is actively expressed in roots, whereas MIR393b is expressed mainly in the aerial organs.
To examine the spatial expression patterns of the MIR393 genes in detail, we generated the reporter gene fusions ProMIR393a:GUS and ProMIR393b:GUS, in which 2–2.5-kb fragments upstream of the miRNA stem–loop regions were isolated and fused with GUS. These reporter constructs were introduced into rice by Agrobacterium-mediated transformation. GUS staining showed that miR393a was strongly expressed in emerging crown/adventitious roots, but not in the primary root (Fig. 2a,b). In emerging crown roots, the GUS expression domain stopped 0.5 mm from the root tip (Fig. 2c). In the initial stage of lateral root primordium formation, GUS staining was observed in pericycle cells, and gradually disappeared in the developing primordium (Fig. 2d). Interestingly, we found that GUS activity was very strong in the coleoptile tip of imbibed seed (Fig. 2e), and gradually appeared in the scutellar node of the embryo in germinating seed (Fig. 2f). GUS staining was also observed in the immature anther at early stages of floral development (Fig. 2g). By contrast, MIR393b was not detected in the roots, immature anthers or coleoptiles, but was expressed in the shoot apical meristem (Fig. 2h). This supports the notion that the MIR393a and MIR393b genes are transcriptionally regulated in a tissue-specific manner. Compared with the expression of MIR393b, MIR393a expression is restricted to more localized regions with high auxin concentrations, such as coleoptile tips and lateral root primordia. Together, all of the above results suggest that the miR393 family members possess distinctive, tissue-specific expression patterns in rice.
Overexpression of miR393 dramatically influences the growth and development of rice
The differential expression patterns of OsmiR393 family members prompted us to investigate whether functional diversification occurred in these two miR393 members. We generated transgenic plants expressing the stem–loop precursors of miR393a and miR393b under the control of the CaMV 35S promoter. Lines containing a single copy of the transgene were selected from independent T0 generation lines based on DNA gel blot analysis (Fig. S2). Homozygous T3 transgenic seeds were obtained for further investigation. Phenotypic observation of the transgenic plants showed that the overexpression of miR393 dramatically enlarged the lamina joint angle of flag leaves (Fig. 3a). The angle of the lamina joint of miR393-overexpressing plants was > 90°, whereas that of the wild-type was < 10° at the booting stage (Fig. 3b,c). In addition, overexpression of miR393 reduced plant height (Fig. 3c).
In addition to altered flag leaf angle and plant height, 35S:miR393a and 35S:miR393b seedlings had longer primary roots than the wild-type control after 2 wk of solution culture (Fig. 4a). The number of crown roots was reduced in 4-wk-old transgenic seedlings compared with the wild-type (Fig. 4b). These results demonstrate that both miR393a- and miR393b-overexpressing plants displayed typical phenotypes associated with altered auxin signaling.
In addition, the miR393-overexpressing plants displayed other changes throughout seed development. For example, the outer glume of the spikelet had an abnormal shape and failed to close (Fig. 5a). Furthermore, the size of the transgenic seeds was reduced compared with wild-type seeds (Fig. 5b). Small RNA blot analysis of miR393 expression showed that the transgenic plants with strong phenotypic aberrations had high expression levels of miR393 (Fig. 5c). A clear parallel between the severity of phenotypes and the level of miR393 expression suggests that the observed phenotypes were indeed caused by the ectopic expression of miR393. Taken together, our results demonstrate that overexpression of miR393 dramatically influences plant growth and development in rice.
Characterization of target genes regulated by miR393 and their expression patterns
The rice miR393 family is predicted to target two TIR1-like genes, Os04g32460 and Os05g05800, which were identified by degradome sequencing (Li et al., 2010; Zhou et al., 2010). Each has a sequence that is partially complementary to that of miR393 with three mismatched nucleotides (Fig. S1). In Arabidopsis, miR393 targets the auxin receptors AFB1, AFB2, AFB3 and TIR1. Phylogenetic analysis of TIR1/AFB genes has shown that Os04g32460 and Os05g05800 are likely to be orthologous to Arabidopsis AFB2 and TIR1 (Jones-Rhoades & Bartel, 2004; Navarro et al., 2006; Parry et al., 2009). Recently, Os04g32460 and Os05g05800 were named as OsAFB2 and OsTIR1, respectively (Xia et al., 2012).
Target mimicry, in which an uncleavable target is expressed in the plant, has been used to sequester specific miRNAs in nonproductive interactions, thus increasing the levels of the endogenous target transcripts (Franco-Zorrilla et al., 2007). 35S:MIM393 was generated as such a target-mimic construct for miRNA393. We measured the transcripts of OsAFB2 and OsTIR1 in 35S:MIM393, 35S:miR393a and 35S:miR393b transgenic plants by quantitative RT-PCR analyses with primers flanking the miR393 target sites. We observed that the transcript levels of OsTIR1 and OsAFB2 increased in the flag leaves of 35S:MIM393 transgenic line compared with wild-type plants (Fig. 6a). The increased levels of miR393 caused a dramatic reduction in the transcript levels of OsAFB2 and OsTIR1 in miR393-overexpressing plants. Our results confirmed that OsTIR1 and OsAFB2, negatively regulated by miR393, are miR393 targets in rice.
To determine the expression pattern of miR393 target genes, real-time RT-PCR was performed with total RNA isolated from roots, leaves and shoots at various growth stages. The pattern of OsTIR1 expression is similar to that of OsAFB2, with the highest level of expression in leaf two (2L) from young seedlings (Fig. 6b,c). The expression of the two genes decreased dramatically in leaf three (3L), and then steadily increased from leaf three to leaf ten (10L) and flag leaves (FL). In roots, the expression levels of OsAFB2 and OsTIR1 were highest in young roots from one-leaf (1L-R) and two-leaf (2L-R) stage seedlings, respectively. Overall, the mRNA levels of OsTIR1 and OsAFB2 were largely similar in different tissues.
Overexpression of miR393 results in 2,4-D resistance
Considering that miR393 targets auxin receptor homologs, we examined whether the overexpression of miR393 affects the auxin response in rice. Wild-type and miR393-overexpressing transgenic seeds were germinated on N6 medium with or without the synthetic auxin 2,4-D. After 2 wk of growth, wild-type seedlings displayed dramatic inhibition of primary root elongation and increased numbers of crown roots in the 100 nM 2,4-D treatment (Fig. 7a). However, both 35S:miR393a and 35S:miR393b plants had longer primary root lengths and fewer crown roots than the wild-type. The shoot lengths were not obviously different between the overexpressing plants and the control. Interestingly, when cultured on callus-inducing medium with a high auxin concentration (10 μM 2,4-D), seeds of miR393-overexpressing lines still exhibited shoot and root hair growth instead of callus, whereas seeds of the wild-type displayed callus growth after 2 wk (Fig. S3). This indicates that the overexpression of miR393 results in 2,4-D resistance, implying that the miR393 targets function in the auxin response.
As root elongation is very sensitive to auxin, reduced inhibition of primary root elongation is a typical auxin-resistant phenotype (Chhun et al., 2003). To illustrate further the decreased auxin sensitivity of the transgenic plants, we examined primary root growth in different concentrations of auxin. After 2 wk of growth in solution medium containing 0, 50, 100 or 200 nM 2,4-D, the primary root lengths of the transgenic and wild-type plants were measured (Fig. 7b). The primary root length of wild-type plants treated with 200 nM 2,4-D was significantly shorter than the root lengths under 0, 50 and 100 nM treatments (ANOVA, P < 0.05). However, both 35S:miR393a and 35S:miR393b plants were able to maintain primary root elongation in the 200 nM 2,4-D treatment for 2 wk, showing an auxin-resistant phenotype. This result further demonstrates that miR393 overexpression results in an increase in auxin resistance.
Knockdown of OsTIR1 or OsAFB2 alters flag leaf inclination
To investigate the role of OsTIR1/AFB2 genes in plant growth and development, we generated OsTIR1- and OsAFB2-RNAi transgenic lines in a wild-type background. We characterized the phenotype of 12 and 16 independent transgenic lines obtained for the knockdown of OsTIR1 or OsAFB2, respectively. Interestingly, droopy flag leaves appeared in AFB2-RNAi lines at the booting stage (Fig. 8a), resembling the trait observed in plants overexpressing miR393. RT-PCR analysis showed that the level of endogenous AFB2 expression was reduced in the AFB2-RNAi line (Fig. 8b). A similar flag leaf inclination phenotype was also observed in TIR1-RNAi lines, although the change in angle was relatively small in comparison with that in AFB2-RNAi lines (Fig. S4). However, except for the angle of the flag leaf, other morphological characteristics in OsTIR1- and OsAFB2-suppressing lines were similar in appearance to those of wild-type plants. These results indicate that the reduced expression of either OsAFB2 or OsTIR1 increases the inclination of flag leaves, but the reduced expression of either OsTIR1 or OsAFB2 alone has only a modest effect on plant growth and development in rice.
OsTIR1 and OsAFB2 interact with OsIAA1
Molecular genetic studies in Arabidopsis have illustrated that the TIR1/AFBs interact with Aux/IAAs, controlling auxin signaling. It is known that OsIAA1 protein is nuclear localized and plays an important role in auxin signaling pathways and plant morphogenesis in rice (Song et al., 2009b). To determine the subcellular localization of OsAFB2 and OsTIR1, the full-length cDNA of each gene was fused to the YFP and GFP reporter genes, respectively, and placed under the control of the 35S promoter. The IAA1:CFP fusion gene driven by the 35S promoter was used as a control. These constructs were co-transformed into onion epidermal cells, and the transformed cells were checked for the GFP/CFP signal. GFP and YFP were detected in the nuclei of the onion cells for OsTIR1:GFP and OsAFB2:YFP fusion constructs, respectively (Fig. 9a). Meanwhile, the CFP signal for the OsIAA1:CFP fusion protein was also located in the nuclei of the cells. Merged images demonstrated that the GFP/YFP and CFP signals overlapped (Fig. 9a). The results thus indicate that OsAFB2 and OsTIR1 are nuclear proteins.
To determine whether OsTIR1 and OsAFB2 interact with Aux/IAAs, we performed a yeast two-hybrid assay for interacting proteins using the full-length OsTIR1 and OsAFB2 proteins as baits and OsIAA1 as prey. Three strains of yeast containing each combination of bait (BD) and prey (AD) vectors were grown on medium with α-gal. The results showed that both OsTIR1 and OsAFB2 interact with OsIAA1 in yeast (Fig. 9b). To verify the interaction between OsTIR1/OsAFB2 and OsIAA1 in planta, BiFC assays were performed in onion epidermal cells (Fig. 9c). OsTIR1 or OsAFB2 was fused to the split N-terminal YFP fragment and OsIAA1 to the C-terminal YFP fragment. Bright YFP fluorescence was observed when OsTIR1-nYFP and OsIAA1-cYFP or OsAFB2-nYFP and OsIAA1-cYFP were co-bombarded into the onion cells. The cells transformed with IAA1-cYFP, TIR1-nYFP or AFB2- nYFP alone did not show any fluorescence. These results demonstrate that in vivo interactions occur between OsTIR1/OsAFB2 and OsIAA1 in plant cells, suggesting conserved interactions between TIR1 homologs and IAAs in both rice and Arabidopsis.
Expression diversification in the rice miRNA393 family
High-throughput sequencing analysis suggests that a minority of annotated miRNA gene families are conserved between plant families, whereas the majority are family or species specific (Griffiths-Jones et al., 2008; Sunkar et al., 2008). The evolution of miRNA processing and functional diversity has led to the identification of the complex regulatory networks of miRNA/target modules that control plant development and nutritional responses, together with other important processes (Cuperus et al., 2011). Therefore, it is necessary that the diversified functions of miRNAs be studied in a range of species. Based on the miRBase data (Release 17), the miR393 family is present in all angiosperm lineages (Kozomara & Griffiths-Jones, 2011). In Arabidopsis GUS-reporter lines, the expression patterns of miR393a and miR393b are similar, with both expressed ubiquitously in all tissues, such as roots, stems, rosette leaves, cauline leaves and inflorescences (Parry et al., 2009; Chen et al., 2011). This suggests that the expression levels of AtMIR393a and AtMIR393b are similar in the whole plants. As the sequences of mature AtmiR393a and AtmiR393b are identical, it is difficult to distinguish between miR393a and miR393b at the transcript levels of mature miR393. However, recent findings have shown that the high levels of miR393 accumulation in aerial organs are reduced dramatically in the miR393b-1 mutant, whereas the low levels in roots are not affected (Si-Ammour et al., 2011). This suggests that mature miR393 in aerial parts of the plant is produced predominantly by AtMIR393B, which helps to regulate the auxin-related development of leaves (Si-Ammour et al., 2011). Owing to the different size of mature OsmiR393a (21 nucleotides) and OsmiR393b (22 nucleotides, Fig. S1), the steady state levels of OsmiR393a and OsmiR393b can be detected in small RNA blots (Fig. 1a). In rice, we found that the two OsmiR393 family members exhibit differential expression in specific tissues and/or growth stages. Strikingly, OsmiR393a is expressed mainly in the roots, whereas OsmiR393b is expressed in aerial tissues. Spatio-temporal expression analysis indicated that OsmiR393a transcription is restricted to emerging crown root tips and lateral root primordia, but not in the primary root. However, both miR393a and miR393b of Arabidopsis are expressed in primary root tips (Parry et al., 2009; Chen et al., 2011). Based on a very different anatomy in root system between dicots and monocots, we speculate that the root-specific expression of OsmiR393a might play a partial role in the initiation of crown root and lateral root primordia in rice.
It is known that auxin accumulates at locations of organ initiation (Heisler et al., 2005; Dubrovsky et al., 2008). Likewise, the expression of miR393 also actively occurs in the sites of organ initiation, implying its role in auxin accumulation and organ initiation. Interestingly, strong GUS activity was found in the coleoptile tips of germinated seeds in the ProMIR393a:GUS transgenic line (Fig. 2e). Monocot coleoptiles have been used to investigate IAA production, polar transport and tropisms since the early days of Darwin. IAA is synthesized at the tip of coleoptiles, mainly within the top 0–1-mm region (Nishimura et al., 2009). Our results indicate that the coleoptile tip, the region of IAA biosynthesis and accumulation, is also a region of active miR393a transcription. Accordingly, we speculate that miR393 expression might be involved in the local production and accumulation of auxin, especially in the region of a developing primordium.
Taken together, our data show that the OsmiR393 family is expressed in restricted regions of plant organs that are actively growing, whereas its two members display spatio-temporal expression patterns. Furthermore, although miR393 sequences are conserved between rice and Arabidopsis, expression patterns of the miR393 family differ in the two species. Similarly, miR827s also exhibit differences in tissue expression between the two species and regulate different classes of genes sharing a common SPX domain (Lin et al., 2010). The potential functional differentiation of the miRNA precursors is mirrored by subfunctionalization in expression patterns (Ru et al., 2006; Wu et al., 2006; Gutierrez et al., 2009). These data suggest that expression diversification has occurred in some miRNA families of rice.
OsmiR393 overexpression alters flag leaf angles and primary and crown root growth in rice
Increased flag leaf inclination was striking in miR393-overexpressing rice at the booting stage. It is known that an increased angle of the lamina joint is a typical brassinosteroid (BR)-related phenotype. Lamina joint inclination assays were thus recognized as an extremely sensitive system for the determination of the biological activity of natural or synthetic BRs and auxin (Nakamura et al., 2009). In addition to BR-induced effects on rice lamina inclination, auxin (IAA) also affects lamina joint inclination at high concentrations and has a synergistic interaction with BR (Zhao et al., 2010). OsIAA1-overexpressing transgenic rice exhibited an increased angle of the lamina joint and dwarfism through a cross-talk pathway with BR (Song et al., 2009b). Our results show that OsmiR393 overexpression causes dramatically increased inclination of the flag leaf, displaying an auxin-resistant phenotype. OsTIR1 and OsAFB2, the targets of miR393, exhibit high levels of expression in flag leaves. Combined with the enlarged inclination of flag leaves in OsTIR1 and OsAFB2 suppression lines, this suggests that OsTIR1 and OsAFB2, negatively regulated by OsmiR393, play an important role in the development of flag leaves.
In addition to flag leaf inclination, primary root elongation and crown root formation were clearly affected by miR393 overexpression in rice. In contrast, it has been reported that the overexpression of miR393 generates only a modest phenotype in Arabidopsis (Parry et al., 2009; Chen et al., 2011). This is perhaps a result of the differences in the formation of adventitious roots. It is well known that Arabidopsis rarely forms adventitious roots, whereas rice produces numerous adventitious or crown roots (Hochholdinger et al., 2004; Inukai et al., 2005). In agreement with the deficient crown root morphology of overexpressing plants, miR393a transcription is active in regions of emerging crown roots in rice. The localized expression of miR393a in roots corresponds well to the areas of crown and lateral root initiation, suggesting that miR393a might be involved in the initiation and emergence of these root systems. In rice, it is assumed that the first cell divisions yielding the new crown root primordium originate from the innermost ground meristem cells adjacent to the peripheral cylinder of vascular bundles in the stem (Itoh et al., 2005). Division of these cells eventually leads to a crown root meristem and could involve specific regulation of polar auxin transport via the GNOM1/PIN system. Crown rootless1 (Crl1) encodes a positive regulator for crown and lateral root formation, and its expression is directly regulated by an ARF in the auxin signaling pathway (Inukai et al. 2005). The auxin responsive AP2/ERF (APETALA2/ETHYLENE RESPONSE FACTOR) transcription factor CROWN ROOTLESS5 is involved in crown root initiation (Kitomi et al., 2011). Moreover, OsARF12 is implicated in the regulation of root elongation (Qi et al., 2012). In maize, Rum1 (an Aux/IAA protein) is also required for the control of embryonic seminal and post-embryonic lateral root initiation (von Behrens et al., 2011). It is known that local accumulation of auxin acts as a trigger for organogenesis (Dubrovsky et al., 2008). An auxin minimum zone restricts the occurrence of founder cell specification and lateral root initiation via the TIR1/AFB pathway (Dubrovsky et al., 2011). Considering that the miR393 targets are homologs of auxin receptors TIR1/AFB, we assume that miR393 affects crown root initiation and seminal root development through negative regulation of the TIR1/AFB-like genes in rice. Because root architecture is a key determinant of nutrient and water use efficiency in crops, the miR393-mediated regulation of adventitious root formation in rice is of considerable interest.
Increased tillers and early flowering are two phenotypes in artificial miR393-overexpressing rice (Xia et al., 2012). We also observed these phenotypes in miR393a- and miR393b-overexpressing lines (data not shown). However, it is surprising that the phenotypic changes in flag leaf angle and growth of primary and crown root observed by the overexpression of OsmiR393 in our study were not reported by Xia et al. (2012).
Conserved mechanism of miR393 and TIR1/AFB2 in auxin response
Previous phylogenetic studies have indicated that the TIR1/AFB proteins are conserved among all land plants. The TIR1/AFB2 clade divided into two distinctive clades (AFB1 and AFB2) before the separation of monocot and eudicot plants (Parry et al., 2009). In the rice genome, there are six TIR1/AFBs homologs, classified into TIR1 (Os05g05800), AFB2 (Os04g32460 and Os11g31620) and AFB4 (Os02g52230, Os11g27450 and Os03g08850; Parry et al., 2009) clades. Our results demonstrate that OsAFB2 and OsTIR1 are up-regulated in 35S: MIM393 plants, indicating that they are the targets of miR393 in rice (Li et al., 2010; Zhou et al., 2010; Xia et al., 2012).
Suppression of the expression of OsTIR1 or OsAFB2 alone increased the inclination of the flag leaf, as observed in the plants overexpressing OsmiR393. These results indicate that both OsTIR1 and OsAFB2 contribute to flag leaf inclination, suggesting an overlapping function. However, neither OsTIR1- nor OsAFB2-RNAi displayed any defects in root growth and seed development, which were observed in the miR393-overexpressing plants. Overexpression of miR393 causes a simultaneous dramatic decrease in its target mRNAs and leads to an increase in auxin resistance. This is similar to the mutation of multiple targets, such as in the quadruple tir1 afb1 afb2 afb3 mutant in Arabidopsis (Parry et al., 2009). Compared with miR393-overexpressing plants, TIR1- and AFB2-RNAi lines displayed only a weak phenotype. It has been predicted that most miRNAs have multiple targets. Therefore, functional redundancy may mask phenotypes if only one of the targets is suppressed. Genetic evidence in Arabidopsis indicates that TIR1 and three AFB F-box proteins act redundantly to mediate auxin responses throughout plant development (Dharmasiri et al., 2005; Kepinski & Leyser, 2005). Multiple miR156/7-targeted SPL genes, in concert with nontargeted SPL8, are essential for correct cell proliferation in early anther development (Xing et al., 2010). Thus, it is likely that OsTIR1 and OsAFB2 have redundant functions, or that other similar auxin receptors exist in rice.
Auxin homeostasis and related developmental processes depend on miRNA-mediated regulation of key components of auxin signaling (Rubio-Somoza et al., 2009). The highly conserved miR167–ARF6/ARF8 nodes modulate auxin early response regulators, and the overexpression of miR167a causes defects very similar to those seen in arf6 arf8 double mutants (Wu et al., 2006). Here, we found that the overexpression of either of the two miR393 family members caused a phenotype that was suggestive of severe defects in auxin response. Our results demonstrate that the regulation of TIR1 and AFB2 auxin receptors by miR393 links auxin action to transcriptional regulation.
In conclusion, the overexpression of miR393 negatively regulates mRNAs of OsTIR1 and OsAFB2. OsTIR1 and OsAFB2 interact with OsIAA proteins, probably releasing the activities of ARFs, and resulting in an increased resistance to auxin. The change in auxin response consequently affects diverse aspects of plant growth and development, such as flag leaf inclination, primary root growth, crown root initiation and seed development.
This work was supported by the National Science Foundation of China (Grant No. 30972016, No. 31171615, No. 30571197), and the Key Project of Science and Technology of Zhejiang Province (2009C12072).