OsARF12, a transcription activator on auxin response gene, regulates root elongation and affects iron accumulation in rice (Oryza sativa)


Author for correspondence:
De An Jiang
Tel: +86 571 88206461
Email: dajiang@zju.edu.cn


  • Auxin has an important role in maintaining optimal root system architecture (RSA) that can cope with growth reductions of crops caused by water or nutrient shortages. However, the mechanism of controlling RSA remains largely unclear. Here, we found a limiting factor of RSA – OsARF12 – an auxin response factor whose knockout led to decreased primary root length in rice (Oryza sativa).
  • OsARF12 as a transcription activator can facilitate the expression of the auxin response element DR5::GFP, and OsARF12 was inhibited by osa-miRNA167d by transient expression in tobacco and rice callus.
  • The root elongation zones of osarf12 and osarf12/25, which had lower auxin concentrations, were distinctly shorter than for the wild-type, possibly as a result of decreased expression of auxin synthesis genes OsYUCCAs and auxin efflux carriers OsPINs and OsPGPs. The knockout of OsARF12 also altered the abundance of mitochondrial iron-regulated (OsMIR), iron (Fe)-regulated transporter1 (OsIRT1) and short postembryonic root1 (OsSPR1) in roots of rice, and resulted in lower Fe content.
  • The data provide evidence for the biological function of OsARF12, which is implicated in regulating root elongation. Our investigation contributes a novel insight for uncovering regulation of RSA and the relationship between auxin response and Fe acquisition.


The root is an important organ in mineral and water uptake. Optimizing root system architecture (RSA) can overcome yield limitations in crop plants caused by water or nutrient shortages (Werner et al., 2010). The dicotyledonous model plant Arabidopsis has a primary root (PR) and lateral roots (LRs), whereas the monocotyledonous crop rice has fibrous roots predominantly composed of adventitious roots (ARs) and LRs. Auxin acts as a versatile trigger in root developmental processes (Benkováet al., 2009), including the regulation of root growth (Ding & Friml, 2010), root patterning (Sabatini et al., 1999; Friml et al., 2002), and root cell division and elongation (Ullah et al., 2003; Campanoni & Nick, 2005; Mironova et al., 2010). These actions of auxins are mostly mediated by the auxin response factor (ARF) and auxin/indole acetic acid (AUX/IAA), an early response gene family. For instance, AUXIN RESPONSE FACTOR5 (ARF5)/MP, NONPHOTOTROPIC HYPOCOTYL4 (NPH4)/ARF7, BODENLOS/IAA12 and AUXIN RESISTANT3 (AXR3)/IAA17 have a regulating effect in PR formation (Hardtke & Berleth, 1998). ARFs specifically expressed in plants regulate embryonic root formation and critically depend on the ARF5/MP gene, which encodes a transcription factor that mediates auxin-responsive gene expression (Hardtke & Berleth, 1998; Ulmasov et al., 1999). The ARF5/MP gene controls embryonic root initiation by regulating a small mobile transcription factor, TARGET of MP 7 (TMO7) (Schlereth et al., 2010). In addition, MiARF2 inhibits root and hypocotyl growth of Arabidopsis (Wu et al., 2011).

A number of ARFs are implicated in LR development of Arabidopsis (Lau et al., 2008; Fukaki & Tasaka, 2009; Péret et al., 2009). Recent reports showed that a bimodular auxin response controls organogenesis in Arabidopsis: the BODENLOS/IAA12-MONOPTEROS/ARF5-mediated auxin response guarantees organized LR patterning downstream of the solitary root/IAA 14-ARF7/19 (De Smet et al., 2010). In rice, OsARF1 interacts with the auxin response elements (AuxRE) in the crown rootless1 (crl1) promoter to trigger its transcription in AR and LR initiation areas, resulting in root initiation (Inukai et al., 2005). In addition to these biological functions, ARFs also act in cell expansion. The interaction between ARF8 and BIGPETALp (BPEp) limits Arabidopsis petal growth by influencing cell expansion (Varaud et al., 2011).

Auxin response factors in plants are encoded by a multigene family. To date, 23 and 25 ARF members have been reported in Arabidopsis and rice (Guilfoyle & Hagen, 2007; Wang et al., 2007). Even though many ARF members have been identified and characterized in Arabidopsis as mentioned earlier, they have been less often found in rice – only OsARF1 was shown to be essential for growth in vegetative organs and seed development (Attia et al., 2009). To elucidate the functions of ARFs in rice, the structure of OsARF genes and interactions between OsARF and OsIAA were analyzed in our previous report (Shen et al., 2010). The present study focuses on gaining insights into the physiological and genetic features of knockout mutants of OsARF12 and the role of OsARF12 in auxin response and iron (Fe) accumulation for RSA.

Materials and Methods

Plant materials and growth conditions

Wild-type (WT) and mutant rice plants (Oryza sativa L.) were grown in normal culture solution combined with phytohormone treatments in a glasshouse with a light : dark cycle of 12 : 12 h at 30 : 24°C. Nicotiana benthamiana plants were grown in vermiculite containing Murashige and Skoog salt (MS) nutritional liquid in a growth chamber (light : dark of 12 : 12 h at 25 : 18°C). Six-week-old N. benthamiana plants were used for Agrobacterium tumefaciens-mediated transient expression. Phytohormone treatment was performed with 1 μM of 2,4-dichlorophenoxyacetic acid (2,4-D), 1-naphthylacetic acid (NAA) and 6-benzylaminopurine (6-BA); 10 μM of IAA, ethylene (ACC), ABA, brassinolide (BR) and methyl jasmonate (JA); and 1 mM acetyl salicylic acid (SA) for 3 h, respectively. The treatments of polar auxin transport inhibitors (PATIs) were carried out with 1 μM of 1-naphthylphthalamic acid (NPA), 2,3,5-triiodobenzoic acid (TIBA) and 1-naphthoxyacetic acids (1-NOA) for 3 h.

Identification of osarf mutants

Identification of T-DNA insertion sites in mutants of osarf12 (PFG_3A-14552), and osarf25 (PFG_ 2D-11520) and expression levels of OsARFs were determined according to the SIGnAL database, http://signal.salk.edu/cgi-bin/RiceGE. Right border primer Ngus-RB of T-DNA was used to confirm integration of T-DNA in Osarf12 and Osarf25, and gene-specific primers InARF12U/L or InARF25U/L were used to identify OsARF12 or OsARF25 genes. Identification of the TOS17 insertion mutant of osarf12T (RGRC-H0120) was based on the information in http://www.rgrc.dna.affrc.go.jp. The TOS17-tail16 was used to confirm integration of TOS17 into osarf12T, and gene-specific primers ARF12T17U/L were used to identify the WT bounds of OsARF12. These PCR products of various insertions were ligated with pUCm-T vector (Sangon, China), transformed into Escherichia coli DH5α, and the flanking sequences of T-DNA or TOS17 insertion sites were sequenced by Invitrogen. The expression of OsARF12 or OsARF25 in the related mutant was demonstrated by reverse transcription polymerase chain reaction (RT-PCR) using primers RTosarf12U/L or RTosarf25U/L located c. 330 or 230 bp from the start codon of various transcripts. The homozygous mutants of osarf12 and osarf25 were crossed to obtain the double mutant osarf12/25. The homozygous F4 generation of osarf12/25 was used for various analyses. For the complementation experiment on OsARF25, the open reading frame (ORF) of OsARF25 (AK121703) using primers OVARF25U/L was ligated to the binary vector pCAMBIA1300, which contained CaMV 35S promoter (35S). The 35S:OsARF25 were introduced into the A. tumefaciens strain EHA105 using electroporation. Agrobacterium-mediated transformation using mutant osarf25 and WT Dongjin of rice were performed as described by Hiei et al. (1994). To confirm the transcription level of each OsARF gene in the corresponding osarf mutant, RT-PCR of OsARF genes was performed. Total RNA extraction and reverse transcripts were carried out according to Wang et al. (2010). All primer sequences for the PCR and RT-PCR are listed in Table S2.

GUS staining

The construction of OsARF12 promoter-β-glucuronidase (OsARF12pro::GUS) was performed as described by Cheng et al. (2007a). Primers ProARF12U/L (Table S2) were used for amplification of the promoter region. The OsARF12pro::GUS was introduced into the A. tumefaciens strain EHA105 and transformed into rice WT, Nipponbare. GUS staining of seedlings was done using 100 mM sodium phosphate buffer (pH 7.0) containing 0.1% v/v Triton X-100 and 2 mM X-Gluc (Biobasic Inc., Sangon, Shanghai, China) at 37°C overnight. The various organs were observed with a Carl Zeiss laser scanning system, LSM510 (http://www.zeiss.com/).

Vector construction for 35S: OsARF12–GFP fusion protein

The ORF of OsARF12 (Os04g57610) was directly amplified from the full-length cDNA (AK071455) using the primers ARF12GFPU/L (Table S2), and then cloned into the binary vector pCAMBIA 1300 containing a CaMV 35S promoter::GFP cassette to create OsARF12–GFP fusion proteins. The 35S:OsARF12–GFP fusion constructs were transiently expressed in onion epidermal cell using the Bio-Rad biolistic PDS-1000/He system at 1100 psi (http://www.bio-rad.com/), carried out essentially as described previously (Varagona et al., 1992). GFP fluorescence was imaged using the LSM510 laser scanning system.

Agrobacterium-mediated transient expression in N. benthamiana leaves

The ORFs of OsARF12 (AK071455) and OsARF25 (AK121703) were ligated to the binary vector pCAMBIA1300, which contained CaMV 35S promoter (35S). Primers OVARF12U/L and OVARF25U/L for the vector construction are found in Table S2. The 35S:OsARF12 or 35S:OsARF25 were introduced into A. tumefaciens strain EHA105 using electroporation. Agrobacterium-mediated transformation in N. benthamiana was performed as described by Chen et al. (2008). Overexpression analysis of OsARF12 and OsARF25 genes was monitored by RT-PCR, and the primers RTARF12U/L and RTARF25U/L are listed in Table S2. DR5::sGFP (Ulmasov et al., 1997), 35S:OsARF12/DR5::sGFP vectors and 35S:OsARF25/DR5:: sGFP vectors were transformed into A. tumefaciens strain EHA105 for infecting leaves of N. benthamiana plants. After 3 d, the fluorescence was visualized under a Nikon microscope equipped with NIS-Elements Basic Research 3.0 software (http://www.nis-elements.com).

Western blotting

Total proteins were extracted from N. benthamiana, which carried DR5::sGFP; 35S:OsARF12 and DR5::sGFP vectors; 35S:OsAR F25 and DR5::sGFP vectors. Of the proteins, 30 μg were resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane that was used for subsequent western blot analysis according to standard procedures. Mouse anti-cGFP-tag monoclonal first antibody (GenScript, A00185) and goat anti-mouse IgG(H&L)-AP conjugated secondary antibodies (CWBIO and CW0110) were used for the detection of DR5–sGFP protein.

osa-miRNA167d overexpression in N. benthamiana leaves and rice callus

The osa-miRNA167d (miR167d) precursor (MI0001109, 4163900–4165399 from http://www.mirbase.org/search.shtml) was ligated into pHB vector (Wang et al., 2005), transformed into A. tumefaciens strain EHA105, and infiltrated into rice callus (Hiei et al., 1994) or injected into tobacco (Chen et al., 2008). Primers OvmiR167U/L for constructs of miR167d overexpression, primers for reverse transcript of miR167d and expression detection are listed in Table S3 and the related methods were reported by Varkonyi-Gasic et al. (2007).

Microscopic analysis

Longitudinal sections of root tips of 3-d-old seedlings of WT, osarf12, osarf25 and osarf12/25 were observed. Microscopic analysis was performed by the previous distribution (Jiang et al., 2005).

IAA measurement

Indole-3-acetic acid concentration of leaves and roots in WT, osarf12, osarf25 and osarf12/25 was measured by gas chromatography-selected reaction monitoring mass spectrometry as described by Ljung et al. (2005). The 20 mg of each sample with five independent biological replicates were purified after addition of 250 pg of 13C6-IAA internal standard using ProElu C18 (http://www.dikma.com.cn).

Quantitative RT-PCR (qRT-PCR) analysis

Total RNA was isolated from leaves or roots of 7-d-old seedlings. The methods, including RNA extraction, reverse transcription and qRT-PCR, were performed according to Wang et al. (2010). The sequence of the related primers for qRT-PCR is listed in Tables S4–S8.

Measurements of metal ion concentrations and statistical test

Metal ion contents of 10-d-old seedlings of WT, osarf12, osarf25 and osarf12/25 mutants were analyzed with inductively coupled plasma mass spectrometry (Agilent 7500ce; Agilent Technologies, Palo Alto, CA, USA) according to Jia et al. (2011). Each sample in the experiment involving the five biological replicates were compared with the WT. < 5% by Student’s t-test was considered to indicate statistical significance.


Identification of mutants, osarf12T, osarf12, osarf25 and osarf12/25

Using the SALK database (http://signal.salk.edu/cgi-bin/RiceGE) and sequencing, T-DNA was integrated into intron 11 and exon 5 of the OsARF12 and OsARF25 genes, respectively; and TOS17 was integrated into intron 13 of OsARF12 (Fig. 1a). PCR analysis confirmed that the T-DNA or TOS17 fragment had been inserted into individual OsARF genes, and the homozygous lines were harvested (Fig. 1b–d). The RT-PCR results demonstrated that OsARF12 and OsARF25 genes were expressed in WT, but not in the corresponding mutants. Thus these genes were knocked out in mutants (Fig. 1e). OsARF25 expression in mutant osarf25 (Ov25/MT) and OsARF25 overexpression in WT (Ov25/WT) for complementation experiment and function research are shown in the Supporting Information, Fig. S1 and Table S1. OsARF25 was highly expressed in osarf25 and WT; however, there was no significantly different phenotype compared with WT. The primary root lengths (PRLs) in mutant osarf12T and osarf12 were c. 35% shorter than WT Nipponbare or Dongjin (Fig. 1f; Table S1). In particular, the PRL in double mutant osarf12/25 was c. 43% of that in WT. Thus the double mutation of OsARF12 (Os04g57610) and OsARF25 (Os12g41950) increased the effect of OsARF12 loss-of-function. However, the PRLs of both mutants osarf12 and osarf12T were all greater than WT under 2,4-D treatments, indicating that these mutants were insensitive to auxin (Fig. 1g); osarf25 was sensitive to axuin, while osarf12/25 was less sensitive to it (Fig. 1g).

Figure 1.

Identification of OsARF mutants. (a) TOS17 insertion site in osarf12T and T-DNA insertion sites in mutants osarf12 and osarf25. The black box represents the exon, and the black line represents the intron. (b) PCR analysis to confirm the integration of TOS17 in OsARF12; the upper bound indicates the OsARF12 gene fragment and the lower bound indicates the TOS17 insertion fragment. (c) PCR analysis to confirm the integration of T-DNA in OsARF12. The upper bound indicates the OsARF12 gene fragment and the lower bound indicates the T-DNA insertion fragment. (d) PCR analysis to confirm the integration of T-DNA in OsARF25. The upper bound indicates the OsARF25 gene fragment and the lower bound indicates the T-DNA insertion fragment. (e) Reverse transcription PCR analysis to confirm the knockout status of osarf. The upper and middle bounds show OsARF12 and OsARF25 expression (35 cycles) in various Oryza sativa wild-type (WT) and mutant lines, respectively. OsACTIN (26 cycles) is at the bottom as a control. (f) Phenotypic characterization of WT and mutants; normal nutritional solution. From left to right: the 7-d-old seedlings of Nipponbare (NIP), osarf12-Tos17 (osarf12T), DongJin (DJ), osarf12-TDNA (osarf12), osarf25, and osarf12/25. Bar, 2 cm. (g) Morphology of the above-mentioned lines under treatment of 1 μM 2,4-dichlorophenoxyacetic acid for 7 d. Bars, 2 cm.

OsARF12 expression pattern and subcellular localization

Using GUS staining, the expressions of OsARF12 were constructively detected in various organs (Fig. 2a–i); however, OsARF12 expression had a tissue-specific pattern. GUS staining was mainly found in the stele and root tip (Fig. 2a and b) of PRs, ARs (same as for PRs, data no shown) and LRs (initiation to maturation) (Fig. 2f–j). In the stem, OsARF12 showed the highest expression in vascular tissue (Fig. 2c). OsARF12 expression was also observed in the stamen, ovary, glume and leaf vein (Fig. 2d,e). Semiquantitative RT-PCR (sqRT-PCR) confirmed that OsARF12 was expressed differently in various tissues (Fig. 2k). OsARFs contain a monopartite nuclear localization signal in the DNA-binding domain as found by Shen et al. (2010). The transient expression of OsARF12 in onion epidermal cells confirmed that it was localized in the cell nucleus (Fig. 2l,m).

Figure 2.

OsARF12 expression pattern and subcellular localization. (a–j) β-Glucuronidase (GUS) staining. Bars, 100 μm. (a) Region of maturation of Oryza sativa primary root (PR); (b) zone of elongation, meristematic region and root cap of PR; (c) stem (S); (d) flower; (e) leaf (L). (f–j) Lateral root (LR) initiation from pericycle; (k) semiquantitative reverse transcription PCR of OsARF12 expression in various organs or tissues: root (R), stem base (SB), stem (S), leaf sheath (LS), leaf (L), panicle(P); (l and m) subcellular localization of OsARF12 protein in onion epidermal cell: fusion protein was observed with green channel (l) 35S promoter::GFP as a control (m) 35S::OsARF12–sGFP.

Osarf12 or Osarf25 as a transcription activator on auxin response

Auxin response factor binds specifically to TGTCTC AuxREs in promoters to regulate expression of auxin response genes (Ulmasov et al., 1997). DR5::GFP constructs act as an auxin response reporting element containing AuxREs, which indirectly reflects auxin response (Ulmasov et al., 1997). To determine how OsARF12 and OsARF25 impact on the transcripts of auxin response gene, DR5::GFP and OsARF12 or OsARF25 were coexpressed in N. benthamiana. The fluorescence intensities of OsARF12/ DR5::GFP or OsARF25/DR5::GFP were significantly higher than control DR5::GFP (Fig. 3a–c). A subsequent experiment demonstrated that OsARF12 or OsARF25 were overexpressed in N. benthamiana carrying OsARF12/DR5::GFP or OsARF25/ DR5::GFP (Fig. 3d). Further, the western blotting of DR5::sGFP was also consistent with these results, with higher expression abundance under coexpression of OsARF12/ DR5::GFP or OsARF25/DR5::GFP in N. benthamiana (Fig. 3e). This evidence indicated that OsARF12 or OsARF25 was a transcription activator, which facilitated the transcription of the auxin response reporting gene.

Figure 3.

Coexpression of auxin-responsive DR5::GFP reporter gene and OsARF12 or OsARF25 in Nicotiana benthamiana. (a) Control, DR5::sGFP; (b) coexpression of 35S:OsARF12 and DR5::sGFP; (c) coexpression of 35S:OsARF25 and DR5::GFP. Bars (a–c), 100 μm. (d) Semiquantitative reverse transcription PCR analysis of OsARF12 and OsARF25 expression; (e) western blotting of DR5-sGFP expression.

OsARF12 expression was repressed by osa-miRNA167d in tobacco and rice callus

In Arabidopsis, the expression of ARF8 genes was regulated by miRNA167 (Yang et al., 2006). OsARF12 is highly homologous with ARF8 (Wang et al., 2007), and in the present study the abundance of 35S:OsARF12 expression in rice was not much higher than in the WT (data not shown). Hence, we hypothesized that OsARF12 could be cleaved by a small RNA. Additionally, the target gene of osa-miRNA167d (miR167d) was predicted by http://csrdb.ucdavis.edu/cgi-bin/rice_smrna (Fig. 4a). This implied that the 2467–2494 regions on the OsARF12 gene may be a target site of miR167d. To determine whether OsARF12 was controlled by miR167d, 35S:miR167d was transformed into tobacco and rice callus. The sqRT-PCR and qRT-PCR analyses showed that the OsARF12 expression level decreased fivefold in infected tobacco carrying 35S:OsARF12/35S:miR167d-1 and 35S:OsARF12/35S:miR167d-2 compared with 35S:OsARF12 (Fig. 4b,c). In addition, miR167d was overexpressed in rice callus (Fig. 4d). Both sqRT-PCR and qRT-PCR indicated that abundance of OsARF12 expression in rice callus ov-1 and ov-2 was also markedly decreased, as was found in tobacco (Fig. 4e,f). The results indicated that OsARF12 was regulated by miR167d, in a post-transcription event.

Figure 4.

Overexpression of osa-miR167d in Nicotiana benthamiana leaves and rice (Oryza sativa) callus. (a) Bioinformatic prediction of target site of miR167d in the OsARF12 gene. (b) OsARF12 expression level in N. benthamiana leaves with semiquantitative reverse transcription PCR (sqRT-PCR). Upper panel, left to right: OsARF12 expression level in N. benthamiana (1) and N. benthamiana carried with 35S:OsARF12 (2); 35S:miR167d-1(3); 35S:miR167d-2 (4); 35S:OsARF12/35S:miR167d-1(5) and 35S:OsARF12/35S:miR167d-2 (6). Lower panel: NbACTIN expression in these six materials. (c) The relative expression level of OsARF12 in N. benthamiana leaves with quantitative reverse transcription PCR (qRT-PCR) (1–6 are as described in b). (d) Overexpression of miR167d in rice callus. NIP, Nipponbare; Ov-1, miR167d overexpression line 1; Ov-2, miR167d overexpression line 2. (e) OsARF12 expression level in transgenic lines NIP, Ov-1 and Ov-2 with sqRT-PCR; (f) relative expression level of OsARF12 in transgenic lines NIP, Ov-1 and Ov-2 with qRT-PCR. qRT-PCR used three independent biological replicates and was performed according to Wang et al. (2010). Error bars in the figure stand for standard deviation.

Mutant osarf12 was insensitive to auxin and OsARF12 expression was induced by auxin efflux transporter inhibitors and other phytohormones

To determine the relationship between OsARF12 and phytohormones in detail, PRL of osarf12 was measured for 1–7 d after germination (DAG); and OsARF12 expression was quantitatively studied under various phytohormones treatments. Primary root elongation (PRE) of osarf12 was significantly slower and PRL less than WT from 2 DAG, and the double knockout of OsARF12 and OsARF25 had a stronger effect on inhibition of PRE (Fig. 5a). However, there was an opposite result under 2,4-D treatment from day 5 to day 7 compared with control, which further revealed that mutant osarf12 was insensitive to auxin (Fig. 5b). In addition, OsARF12 was not induced by IAA or 2,4-D, which both need an auxin transporter to enter cells, but was induced twofold by NAA which can enter cells by diffusion. OsARF12 was also greatly induced by auxin efflux inhibitors NPA or TIBA, but suppressed by influx inhibitor 1-NOA (Fig. 5c). These results suggested that OsARF12 might be affected by auxin polar transport. OsARF12 was also increased by treatment with 6-BA, ACC and JA, but decreased by SA (Fig. 5c), and unchanged by BR and ABA. This indicates that these phytohormones in part regulated OsARF12 expression; however, OsARF25 did not show outstanding variability in the these treatments, with the exception of having nearly double the control values for the NAA treatment (Fig. 5c).

Figure 5.

Relationship between OsARF12 expression and polar auxin transporter inhibitors (PATIs) or phytohormones. Primary root elongation (PRE) of osarf12 under normal nutritional solution (CK) and 2,4-dichlorophenoxyacetic acid (2,4-D) treatment. (a) Primary root length (PRL) in CK for 1–7 d after germination (DAG); (b) PRLs in CK containing 1 μM 2,4-D for 1–7 DAG; (c) OsARF12 (open bars) and OsARF25 (closed bars) expression in roots of rice Oryza sativa wild-type (WT) treated with auxin, PATIs and various phytohormones. Quantitative reverse transcription PCR (qRT-PCR) used three independent biological replicates and was performed according to Wang et al. (2010). NAA, 1-naphthylacetic acid; 6-BA, 6-benzylaminopurine; ACC, aminocyclopropane carboxlic acid; BR, brassinolide; JA, methyl jasmonate; SA, acetyl salicylic acid; NPA, 1-naphthylphthalamic acid; TIBA, 2,3,5-triiodobenzoic acid; 1-NOA, 1-naphthoxyacetic acids. Error bars in the figure stand for standard deviation.

The decreased auxin contents and altered expression of genes related to short root, OsMIR and OsSPR1 in osarf12

The PRE of osarf12 was clearly inhibited (Figs 1a, 5a), leading us to question why the knockout of OsARF12 resulted in this change. Longitudinal PR sections of WT, osarf12, osarf25 and osarf12/25 were examined (Fig. 6a): the elongation zones of osarf12 and osarf12/25 were distinctly shorter than the WT. Additionally, the auxin concentration in roots or leaves of arf12 was also half that of the WT (Fig. 6b). These results suggested that decreased PRE might be caused by a shortened elongation zone of PR and reduction of auxin concentration in osarf12.

Figure 6.

(a) Longitudinal sections of Oryza sativa primary root (PR) in wild-type (WT) plants and mutants. The thick (white) bars show the elongation zone of roots; the thinner bars are scale bars, 100 μm. (b) Contents of auxin in WT, osarf12, osarf25 and osarf12/25 in leaves (dark gray bars) and roots (light gray bars); (c) expression of OsYUCCA genes in WT (dark gray bars) and osarf12 (light gray bars); (d–f) expression of OsPIN, OsLAX and OsPGP genes in WT, osarf12, osarf25 and osarf12/25; (g) expression of genes related to short roots in WT and osarf12, osarf25 and osarf12/25. OsPLT1, Plethora1; OsPLT2, Plethora2; OsSHR, Short root; OsSCR1, Scarecrow1; OsSCR2, Scarecrow2; OsMIR, mitochondrial Fe-regulated; OsSPR1, Short postembryonic root1. Quantitative reverse transcription PCR (qRT-PCR) used three independent biological replicates and was performed according to Wang et al. (2010). Error bars in the figure stand for standard deviation.

Previous research has shown that auxin-related developmental changes occur when genes encoding conversion of tryptamine to N-hydroxytryptamine YUCCA for auxin biosynthesis are mutated (Cheng et al., 2006, 2007b; Stepanova et al., 2008; Zhao, 2008). Thus, we researched the expression of the OsYUCCA genes in WT and osarf12 (Fig. 6c). The OsYUCCA genes, except for OsYUCCA1, 2, and 5, were greatly down-regulated, which might result in the reduction of auxin biosynthesis and decreased auxin concentration. The result was similar to that of the TERMINAL FLOWER2 (tfl2) mutant, which had a lower rate of auxin biosynthesis and low concentrations of auxin (Rizzardi et al., 2011).

Plant development and physiology are widely determined by polar transport of the signaling molecule auxin. This process is controlled at the cellular efflux level catalyzed by members of the PIN (pin-formed) and ATP-binding cassette protein subfamily B/P-glycoprotein family that can function either independently or in concert (Bureau et al., 2010). Our results showed that the majority of OsPINs (except for OsPIN1, 1b, 1d, 2, and 8) and the majority of OsPGPs (except for OsPGP2, 3, 9, 10, 14 and 22) were significantly reduced in osarf12 (Fig. 6d–f). Experimental evidence suggested that OsARF12 may affect the expression of auxin transporter proteins and thereby modulate auxin distribution. However, it was inexplicable why the five auxin influx-like carriers (OsLAX) predicted by us (Shen et al., 2010) were all enhanced in osarf12 (Fig. 6e), and OsARF12 expression was decreased by the auxin influx inhibitor NOA (Fig. 5c). These results suggested that OsARF12-regulated auxin transport has a complexity that needs to be further investigated.

To uncover the reason for short roots of osarf12, expressions of seven genes related to the short-root phenotype were determined in WT and other mutants. The abundance of OsPLT1, OsPLT2, OsSHR and OsSCR1 altered little in osarf12, osarf25 and osarf12/25 compared with WT (Fig. 6g), indicating that these genes (correlated with regulators of roots) were not controlled by OsARF12. Despite this, OsSCR2 and OsMIR were up-regulated in osarf12 and osarf12/25. The knockout mutant of OsMIR accumulates more Fe than does the WT (Ishimaru et al., 2009). Conversely, short postembryonic root1 (OsSPR1) was inhibited in osarf12 or osarf12/25. The knockdown mutant of osspr1 accumulates less Fe compared with the WT (Jia et al., 2011). This all indicated that both mitochondrial proteins, OsMIR and OsSPR1, involved in root development and Fe homeostasis may be negatively/positively regulated by OsARF12. So, is OsARF12 also implicated in Fe accumulation? The related discoveries led us to study osarf12 in depth.

Knockout of OsARF12 represses Fe accumulation

To determine whether OsARF12 was also related to Fe accumulation, the Fe concentrations in WT and mutants were measured. The Fe concentrations in leaves and roots of osarf12 and osarf12/25 were distinctly lower than WT while the concentration in seed was slightly lower than WT (Fig. 7a–c). The results suggested that the knockout of OsARF12 might impair Fe accumulation. Similarly, the concentrations of the other divalent metal ions Zn, Mn, Mg, Cu and Ca also decreased in osarf12 and osarf12/25 (Fig. S2), which implied that OsARF12 not only functions in Fe signaling but was also related to homeostasis of the other divalent metal ions.

Figure 7.

Fe concentrations in Oryza sativa wild-type (WT), osarf12, osarf25 and osarf12/25, and expression of genes related to Fe uptake and transport in WT and osarf12. The five biological repeats in each line were used in measurement of Fe concentrations (μg g−1) of leaves (a), roots (b) and seeds (c). The bars in (a–c) show a significant difference (< 0.05) between the same lines. (d) Semiquantitative reverse transcription PCR (sqRT-PCR) analysis for genes related to Fe response in WT and osarf12; (e) quantitative reverse transcription (qRT-PCR) analysis for genes related to Fe response in leaves of WT (dark gray bars) and osarf12 (light gray bars); (f) qRT-PCR analysis for genes related to Fe response in roots of WT (dark gray bars) and osarf12 (light gray bars). qRT-PCR used three independent biological replicates and was performed according to Wang et al. (2010). Error bars in the figure stand for standard deviation.

To better understand the mechanism of impaired accumulation of Fe, we analyzed the expression of genes involved in encoding Fe transport in WT and osarf12 using qRT-PCR. OsIRT1 and OsIRT2 were inhibited over threefold and twofold in osarf12 roots, respectively (Fig. 7d–f). The iron-regulated transporter (IRT) 1 gene encodes a probable Fe (II) transporter for Fe uptake. IRT1 also affected the uptake of other metals, for example, Mn and Zn (Eide et al., 1996). The low expression of OsIRT1 and OsIRT2 might result in decreased concentrations of Fe or other divalent metal ions in osarf12 roots. However, the expression of OsIRT1 and OsIRT2 was twofold and fourfold higher, respectively, in leaves of osarf12 compared with WT (Fig. 7d). We also observed impaired Fe concentrations (by a factor of 1.2) in leaves, while in roots of osarf12, concentrations were decreased by a factor of 2.4 (Fig. 7a,b). Thus, the enhanced abundance of OsIRT1 and OsIRT2 might result in a reduction of Fe concentrations in leaves of osarf12, but not to the same extent as in the roots. The Fe-responsive operator (OsIRO2) is expressed in a spatially and temporally similar manner to the genes nicotianamine synthase 1 (OsNAS1) and OsNAS2, which are involved in Fe absorption and translocation (Ogo et al., 2011). Our results showed similar trends to this report: OsNAS1, OsNAS2 and OsIRO2 had a higher expression in roots of osarf12 than in the WT (Fig. 7f). Yellow-stripe like 15 (OsYSL15) in rice is considered to be the dominant Fe (III)-deoxymugineic acid transporter responsible for Fe uptake and phloem transport (Inoue et al., 2009). The expression of OsYSL15 was similar in roots or leaves of osarf12 and in WT treated with normal solution containing divalent Fe (Fig. 7e,f). Natural resistance-associated macrophage protein 2 (OsNRAMP2), deoxymugineic acid synthase 1 (OsDMAS1) and nicotianamine aminotransferase 1 (OsNAAT1) had similar expressions in roots of osarf12 and WT, except that OsDMAS1 was expressed more in leaves of osarf12 than in the WT.


Root system architecture improvement has become a major focus of biotechnology (Ghanem et al., 2011) with the first task being to search and isolate genes implicated in regulating RSA. Auxin also sculpts RSA, although the mechanism of control of root development remains largely unclear. In the present study, OsARF12, an ARF, was researched in detail using the mutant osarf12 and the double mutant osarf12/25 – these mutants have altered RSA, suggesting that OsARF12 might be a limiting factor for RSA.

The knockout mutants of OsARF12, osarf12 with T-DNA insertion, and osarf12T with TOS17 insertion were identified and characterized. Both homozygous lines had shorter PRs and ARs and showed auxin insensitivity (Fig. 1f,g), implying that OsARF12 played a role in root development and auxin response. In Arabidopsis, the atarf7/19 or atarf10/16 double mutants generally have much stronger phenotypes than their corresponding single mutants (Remington et al., 2004; Okushima et al., 2005; Wang et al., 2005; Wilmoth et al., 2005). In the present study, the osarf12/25 double mutant had much short PRs than either single mutant, suggesting that both OsARF12 and OsARF25 also had somewhat redundant roles. Furthermore, ARF6 (Os02g06910) and ARF17 (Os06g46410) were also closely linked to OsARF12 and OsARF25 and showed similar anatomical expression patterns to OsARF12 (Os04g57610) and OsARF25 (Os12g41950) from the rice array database (http://www.ricearray.org/expression/meta_search.php). It will be interesting to know whether or not these four ARFs might have functional redundancy in root development, and whether loss-of-function approaches to all these genes cause rootless phenotypes.

OsARF12 was mainly expressed in the stele and root tip of PRs (Fig. 2a,b), ARs (data no shown) and LRs (Fig. 2f–k), which further indicated that it was implicated in root development. However, OsARF12 was also constructively expressed in the other tissues (Fig. 2c–e,k), suggesting it also functions in those tissues. When the gene expression patterns of OsARF12 were checked in a dynamic gene expression atlas covering the entire life cycle of rice (Wang et al., 2010), it was found to have higher expression levels in plumules or panicles than in radicles, indicating that OsARF12 also played important roles in early aerial tissue or panicle development. ARFs with a glutamine-rich middle region (MR) might function as activators of auxin-responsive gene expression in transiently infected protoplasts, while ARFs with proline- and/or serine-rich MRs (e.g. ARF1 or ARF2) may repress the transcription of reporter genes under the control of synthetic AuxREs (Tiwari et al., 2003). OsARF12 with a glutamine-rich MR was predicted by Shen et al., (2010). Here, we further confirmed OsARF12 as a transcription activator, which facilitated the expression of DR5::GFP (Fig. 3), consistent with the previous theory.

OsARF12 was predicted to be a target gene of miR167d (Fig. 4a), and our experiments using transient expression of miR167d in tobacco and rice callus confirmed this point (Fig. 4b–e). These results were similar to those suggesting that ARF8 genes were regulated by miRNA167 in Arabidopsis (Yang et al., 2006). In addition, it was difficult to get overexpression lines of OsARF12 in WT or osarf12 for its complementary (data no shown), also implying that excess OsARF12 might be cleaved by miR167d in tissues. The synonymous point-mutation in the target site of OsARF12 cleaved by miR167d would be required for construction of OsARF12 overexpression lines.

Further investigation of osarf12 PRE and auxin insensitivity (Fig. 5a, b) triggered more study on linkages of OsARF12 with PATIs or other phytohormones (Fig. 5c). Our experiments also demonstrated that OsARF12 might be implicated in auxin polar transport and other phytohormone signaling, playing a role in cross-talk between auxin and other phytohormones.

Furthermore, the events that caused the short roots of osarf12 were revealed by the shortened elongation zone of PRs and reductions in auxin concentration (Fig. 6a, b). The down-regulated abundance of most auxin synthesis genes OsYUCCAs (Fig. 6c) and auxin transporter genes OsPINs and OsPGPs (Fig. 6d–f) confirmed these results, although the detailed function of these genes in rice remains unclear. Both OsPIN1c and OsARF12 are expressed in roots, with the latter being expressed in the stele and the root tip (Wang et al., 2009; this study). The OsPIN1C abundance was impaired in osarf12, suggesting that OsARF12 might directly regulate OsPIN1C to affect auxin transport. The temporal and spatial distribution of auxin mainly depends on the dynamic expression and subcellular localization of auxin efflux proteins, PIN (Bureau et al., 2010), and the local auxin concentrations were determined by biosynthesis and intercellular transport in Arabidopsis roots (Ding & Friml, 2010). Root length is determined by cell division in the root meristem and cell elongation in the root elongation zone. These results illustrated that decreased auxin concentrations in the root might influence cell elongation. Moreover, in the mutant osarf12, the OsMIR and OsSPR1 genes implicated in root development and Fe homeostasis were also up-regulated or down-regulated, leading us to explore the Fe signal in osarf12.

Rhizosphere availability of microelements and Fe regulate root system branching as a strategy to adjust nutrient uptake and soil availability (López-Bucio et al., 2003). Fe, Zn, Mn, Mg, Cu and Ca are essential micronutrients for plant growth and development. In osarf12, the contents of these micronutrients were lower than in the WT (Fig. 7a–c; Fig. S2); osarf12 showed lower expression of the two genes, OsIRT1 and OsIRT2, related to Fe acquisition and deficiency response in roots compared with WT. This result was consistent with a previous report that OsIRT1 decreased in roots of mutant osspr1, whose Fe content was also lower than in the WT (Jia et al., 2011). Conversely, osmir has higher Fe content and OsIRT1 abundance than the WT (Ishimaru et al., 2009). Together, OsARF12 may regulate positively with OsSPR1 and negatively with OsMIR. Despite this, the change in these genes did not perfectly explain the reduction of Fe concentrations in osarf12. Higher plants generally have two Fe-affinity systems – low and high. So far the molecular mechanism of the low Fe-affinity system (Fe-sufficient condition) is unclear; hence, we had to use the genes related to the high Fe-affinity system (Fe-deficient condition) to determine Fe uptake and transport. However, the abundance of these genes in osarf12 changed little compared with WT under Fe deficiency (Fig. S3), suggesting that osarf12 experienced less influence of the high Fe-affinity system. It is worth mentioning that the mutant osarf12, with lower Fe content under Fe-sufficient conditions (Fig. 7a–c), could become a novel experimental material to reveal the action mechanism of the low Fe-affinity system; for example, using osrf12 to make a microarray to analyze the more down-regulated genes implicated in Fe uptake or transport.

In conclusion, several lines of evidence indicate an important role of OsARF12 in regulating RSA and Fe accumulation in rice. To our knowledge, this is the first report to define the regulation of Fe accumulation by an ARF in higher plants. Based on our studies, the biological functions of OsARF12 as a transcription factor are twofold: in the first case,OsARF12 mediates auxin synthesis and transport via OsYUCCAs, OsPINs and OsPGPs – the down-regulated auxin synthesis or polar auxin transport in osarf12 might result in decreased auxin concentration and shortened PR; in the second case OsARF12 regulates Fe concentrations through adjusting expression of OsMIR or OsSPR1 to affect abundance of OsIRT1. Also, the two biological functions appear to be closely linked; that is, the short PRs in the osarf12 mutant may affect Fe acquisition in roots. These findings reveal the cross-talk between auxin response and Fe signaling and will form the basis for further exploration of the relationship between both signaling pathways.


This research was supported by the National Natural Science Foundation of China (grant no. 31071392 and 30971703), the Genetically Modified Organisms Breeding Major Projects (2009ZX08009-123B) and the Natural Science Foundation of Zhejiang province, China (grant no. Y3080111). We gratefully acknowledge Professor Akio Miyao of the RGRC (Rice Genome Resource Center) in Japan for providing the full-length cDNA clones of OsARF gene families and osarf12-TOS17 mutant, and Professor Gynheung An of the PFG (Plant Functional Genomics Laboratory) in Korea for contributing T-DNA insertion mutants of osarf12 and osarf25. We would like to thank the laboratory of XiaoYa Chen (Shanghai Institute for Biological Sciences, Chinese Academy of Science) for providing the pHB vector and the laboratory of Ping Wu (State Key Laboratory of Plant Physiology and Biochemistry, College of Life Sciences, Zhejiang University) for providing certain experimental instruments and apparatus.