The REL3-mediated TAS3 ta-siRNA pathway integrates auxin and ethylene signaling to regulate nodulation in Lotus japonicus

Authors

  • Xiaolin Li,

    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Mingjuan Lei,

    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Zhongyuan Yan,

    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Qi Wang,

    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Aimin Chen,

    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Jie Sun,

    1. The key Laboratory of Oasis Eco-agriculture, Agriculture College of Shihezi University, Shihezi, China
    Search for more papers by this author
  • Da Luo,

    1. School of Life Sciences, Sun Yat-Sen University, Guangzhou, China
    Search for more papers by this author
  • Yanzhang Wang

    Corresponding author
    1. National Key Laboratory of Plant Molecular Genetics, Institute of Plant Physiology & Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    • Author for correspondence:

      Yanzhang Wang

      Tel: +86 21 5492 4167

      Email: yzwang@sippe.ac.cn

    Search for more papers by this author

Summary

  • The ta-siRNA pathway is required for lateral organ development, including leaf patterning, flower differentiation and lateral root growth. Legumes can develop novel lateral root organs – nodules – resulting from symbiotic interactions with rhizobia. However, ta-siRNA regulation in nodule formation remains unknown. To explore ta-siRNA regulation in nodule formation, we investigated the roles of REL3, a key component of TAS3 ta-siRNA biogenesis, during nodulation in Lotus japonicus.
  • We characterized the symbiotic phenotypes of the TAS3 ta-siRNA defective rel3 mutant, and analyzed the responses of the rel3 mutant to auxin and ethylene in order to gain insight into TAS3 ta-siRNA regulation of nodulation.
  • The rel3 mutant produced fewer pink nitrogen-fixing nodules, with substantially decreased infection frequency and nodule initiation. Moreover, the rel3 mutant was more resistant than wild-type to 1-naphthaleneacetic acid (NAA) and N-1-naphthylphthalamic acid (NPA) in root growth, and exhibited insensitivity to auxins but greater sensitivity to auxin transport inhibitors during nodulation. Furthermore, the rel3 mutant has enhanced root-specific ethylene sensitivity and altered responses to ethylene during nodulation; the low-nodulating phenotype of the rel3 mutant can be restored by ethylene synthesis inhibitor L-α-(2-aminoethoxyvinyl)-glycine (AVG) or action inhibitor Ag+.
  • The REL3-mediated TAS3 ta-siRNA pathway regulates nodulation by integrating ethylene and auxin signaling.

Introduction

Trans-acting RNAs (ta-siRNAs) are a class of endogenous short-interfering RNAs (siRNAs) in plants that have a complex biogenesis involving an initial AGRONAUTE 1 (AGO1)- or AGO7-directed microRNA cleavage of their precursors (Allen et al., 2005; Yoshikawa et al., 2005; Adenot et al., 2006; Fahlgren et al., 2006). The cleaved transcript is converted into a long double-stranded RNA (dsRNA) by concerted action of SUPPRESSOR OF GENE SILENCING 3 (SGS3) and RNA-DEPENDENT RNA POLYMERASE 6 (RDR6; Allen et al., 2005; Nogueira et al., 2007). The long dsRNA is further recognized by DICER-LIKE 4 (DCL4) associated with DOUBLE–STRANDED RNA BINDING 4 (DRB4) and cut into phased 21-nucleotide (nt) small RNAs that undergo further methylation by HUA ENHANCER1 (HEN1, Gasciolli et al., 2005; Li et al., 2005; Xie et al., 2005; Howell et al., 2007). The mature ta-siRNAs are incorporated into RNA-induced silencing complexes (RISCs) containing AGO proteins to regulate their target genes that only show similarity at the ta-siRNA recognition site, and act in trans to repress expression noncell autonomously of specific target genes (Peragine et al., 2004; Vazquez et al., 2004; Chitwood & Timmermans, 2010). In Arabidopsis, four TAS gene families (TAS1 to 4) have been identified, which are themselves targets of specific microRNAs. TAS1 (a, b and c) and TAS2 are targets of miR173, which initiates formation of ta-siRNAs that target multiple mRNAs, including several pentatricopeptide repeat genes (Peragine et al., 2004; Vazquez et al., 2004; Allen et al., 2005; Axtell et al., 2006; Chen et al., 2007; Howell et al., 2007). TAS3 (a, b and c) is targeted by miR390, and the generated ta-siRNAs control the mRNA levels of AUXIN RESPONSE FACTORS (ARFs), including ARF2, ARF3 and ARF4 (Allen et al., 2005; Adenot et al., 2006; Fahlgren et al., 2006; Garcia et al., 2006; Hunter et al., 2006; Marin et al., 2010). TAS4 is targeted by miR828, and the produced ta-siRNAs regulate the expression of several MYB transcription factor genes (Rajagopalan et al., 2006). Among these ta-siRNAs, TAS3 ta-siRNA is relatively well defined for its biogenesis and regulatory roles. First, miR390 binds TAS3 precursor transcripts through its unique association with the effector AGO7 that sets the phase for future processing (Allen et al., 2005; Axtell et al., 2006; Montgomery et al., 2008). Next, SGS3-RDR6-DCL4 processes the TAS3 cleavage product into 21-nt ta-siRNAs, which are also termed tasiR-ARFs (Peragine et al., 2004; Allen et al., 2005; Montgomery et al., 2008). These tasiR-ARFs are bound by AGO1 and cause post-transcriptional cleavage of ARF3 and ARF4 transcripts, which regulates juvenile to adult phase transition in Arabidopsis, and the establishment of auxin-mediated polarity of leaf, flower and lateral root growth (Allen et al., 2005; Pekker et al., 2005; Fahlgren et al., 2006; Garcia et al., 2006; Hunter et al., 2006; Nagasaki et al., 2007; Nogueira et al., 2007; Douglas et al., 2010; Marin et al., 2010; Yoon et al., 2010). This TAS3 ta-siRNA pathway and its regulation of ARF3 and ARF4 are highly conserved throughout land plant evolution (Axtell et al., 2006; Nogueira et al., 2007; Allen & Howell, 2010; Douglas et al., 2010).

The ta-siRNA biogenesis pathway is required for proper development of plant lateral organs. Considerable studies demonstrate that mutations of core components implicated in ta-siRNA biosynthesis cause obvious developmental defects of leaves, flowers and lateral roots. In Arabidopsis, ta-siRNA biogenesis mutants, including dcl4, rdr6, sgs3 and ago7, generate an accelerated transition from juvenile to adult leaves (Peragine et al., 2004; Hunter et al., 2006). Also, dcl4 and rdr6 mutants produce shorter lateral roots (Marin et al., 2010). By contrast, rice and maize ta-siRNA biogenesis mutants produce more severe defects in organ patterning. Rice ta-siRNA defective mutants shootless 2 (shl2), shoot organization 1 (sho1) and sho2 – corresponding to Arabidopsis rdr6, dcl4 and ago7 – are defective in shoot meristem maintenance and leaf morphogenesis (Itoh et al., 2000; Nagasaki et al., 2007), and dcl4-1 mutants produces severe developmental defects in leaf and spikelet, and wavy leaf1 (waf1), mutation of Arabidopsis HEN1 ortholog gene, produces defects in leaf, flower and root development (Liu et al., 2007; Abe et al., 2010). Likewise, similar deficient phenotypes in leaf and flower development are produced in maize ta-siRNA biogenesis mutants (Timmermans et al., 1998; Nogueira et al., 2007; Douglas et al., 2010). Recently, leguminous plant Lotus japonicus TAS3 ta-siRNA biogenesis mutants, reduced leaflet1 (rel1) and rel3 mutants, were characterized by conspicuous defects in compound leaves with abaxialized leaflets, reduced leaflet numbers and infertile flowers (Yan et al., 2010). Given that ta-siRNA biogenesis is essential for plant lateral organogenesis, we ask whether ta-siRNA biogenesis is required for formation of nitrogen-fixing nodules, which are the unique lateral organs formed on legume roots. The identification of these legume ta-siRNA-defective mutants, rel1 and rel3, provides ideal materials to dissect the regulatory function of ta-siRNA pathway in nodule formation.

Nitrogen-fixing nodules result from the symbiotic interaction between legumes and rhizobia, which provide an ecological niche where soil rhizobia convert atmospheric N2 into ammonia for plant utilization. Unlike lateral roots, the formation of nitrogen-fixing nodules involves two distinct developmental processes: rhizobial infection and nodule organogenesis. Although these processes take place in different root cell layers, and can be separated genetically, they must be coordinated both spatially and temporally to ensure that infection threads are targeted into the developing primordium to form a nitrogen-fixing nodule (Madsen et al., 2010; Oldroyd et al., 2011). Despite differences from lateral roots in organogenesis and morphological structure, nodules develop postembryonically from pre-existing roots, and thus nodule organogenesis is integrated into the developmental program governing the root system. Moreover, the TAS3 ta-siRNA pathway was validated to control lateral root development in Arabidopsis (Marin et al., 2010; Yoon et al., 2010); we therefore hypothesized that this ta-siRNA pathway is likely involved in the developmental regulation of symbiotic nodules in legumes.

In order to validate this hypothesis, we used the TAS3 ta-siRNA biogenesis-deficient mutant rel3 to explore the role of the ta-siRNA regulation during nodulation. REL3, an ortholog of Arabidopisis AGO7, is a key component of TAS3 ta-siRNA biogenesis in L. japonicus (Yan et al., 2010). The loss-of-function rel3 mutant carries a retrotransposon insertion in its coding region that encodes PAZ and PIWI domains in AGO proteins (Yan et al., 2010). Our results showed that rel3 mutants developed fewer functional nitrogen-fixing nodules, with reduced nodulation zone ratio (the length of the nodulation zone divided by the main root length) and infection frequency. TAS3 ta-siRNA accumulation was compromised, and its target ARF3a, ARF3b and ARF4 were markedly upregulated in uninoculated and inoculated rel3 roots. Furthermore, rel3 mutants showed altered responses to auxin and ethylene during nodulation. Taken together, our data show that the REL3-mediated TAS3 ta-siRNA biosynthetic pathway acts as a key regulatory switch, integrating auxin and ethylene signaling to coordinate rhizobial infection and nodule initiation in L. japonicus.

Materials and Methods

Plant materials and growth conditions

The rel1 and rel3 mutants were isolated from EMS-mutagenized seeds of Lotus japonicus (Regel) K. Larsen ecotype Gifu B-129 (Yan et al., 2010). Seeds were surface-sterilized and rinsed completely with sterile water. The imbibed seeds were vernalized 4°C for 1 d on 0.8% (w/v) agar, and then incubated overnight at 28°C in the dark for germination. The germinated seeds were transferred on MS agar plates grown vertically for 5 d, and homozygous rel3 mutants with apparent abaxialized leaflets were selected (Yan et al., 2010). Five-day-old seedlings were transferred into pots filled with autoclaved vermiculite and perlite (3 : 1) or onto B&D agar plates, and inoculated with Mesorhizobium loti strain NZP2235 after 2 d. Plant growth conditions were as follows: photoperiod, 16 h light : 8 h dark, 200 μE m−2 s−1 light irradiance, 22°C and 75% hygrometry. Visible nodules, nodulation zone, root length and lateral root number of wild-type Gifu B-129 (WT) and rel3 were examined at 3 weeks post-inoculation (wpi).

Examination of infection events

Surface-sterilized seeds were germinated on 0.8% agar plates for 1 d in the dark, and then grown 3 d in the light. Homozygous rel3 mutants were selected, planted into pots and inoculated with M. loti NZP2235 harboring a hemA:lacZ construct after 2 d. Infection events were examined at 5 or 7 days post-inoculation (dpi). For examining β-galactosidase activity, whole roots were fixed for 1 h with fixative solution (1% (v/v) glutaraldehyde in 1× phosphate-buffered saline (PBS) buffer, pH 7.5). Fixed samples were subsequently washed twice with 1× PBS buffer and incubated a solution composed of 0.2× PBS, 2.5 mM K3[Fe(CN)6], 2.5 mM K4[Fe(CN)6], and 0.8 mg ml−1 of 5-bromo-4-chloro-3-indolyl-β-d-galactoside in N,N-dimethylformamide overnight at room temperature. Stained samples were rinsed three times with 1× PBS and cleared in different concentrations of ethanol solution. Infection events were observed under brightfield microscopy.

Hairy root transformation and GUS activity assay

Approximately 2.8 kb of promoter region of REL3 was amplified by PCR, and inserted between the KpnI and NcoI sites of the pCAMBIA1301p0G vector (Liu et al., 2010). Primers were used as follows: REL3pro-F, 5′-GGGGTACCAGAATGGGCTGAATGCGAAG-3′ and REL3pro-R, 5′-CATGCCATGGTTGTTTTTGTGGTGAGTTTCTGGGT-3′. The resulting REL3pro:GUS construct and the auxin-responsive reporter DR5:GUS (Casimiro et al., 2001) were introduced into Agrobacterium rhizogenes MSU440. The hairy root transformation was performed according to the Lotus japonicus handbook (Díaz et al., 2005). Two-week-old composite plants were planted into pots and inoculated M. loti after 2 d. GUS staining were performed as described in Liu et al. (2010). Roots and nodules were collected, incubated in GUS buffer overnight at 37°C, and then cleared for direct observation. GUS-stained roots and nodules for sectioning were fixed in FAA, dehydrated by ethanol, embedded in resin, sectioned and then observed under brightfield microscopy.

Quantitative real-time reverse transcription PCR (qRT-PCR) analysis

Different organs and inoculated roots at various times post-inoculation were collected. All samples were immediately frozen in liquid nitrogen and stored at −80°C until use. Total RNA was extracted from frozen samples using the TransZol Plant kit (TransGen Biotech, Beijing, China) following the manufacturer's instructions. Samples containing 1 μg RNA were used for cDNA synthesis with PrimeScript RT reagent Kit with gDNA Eraser (TaKaRa, Dalian, China). The cDNA synthesis reaction mixture was used for qRT-PCR. qRT-PCR was conducted with Real-time PCR Master Mix (SYBR® Green I; TOYOBO, Osaka, Japan) and gene-specific primers. The LjUBIQUITIN gene was used as a reference gene to normalize expression.

qRT-PCR for TAS3 ta-siRNA was performed using One Step PrimeScript® miRNA cDNA Synthesis Kit (Takara) following the manufacturer's instruction. Briefly, 1 μg of RNA from each sample was polyadenylated, and then converted to cDNAs with a universal adaptor primer. qRT-PCR were performed using SYBR Premix Ex Tag II (TaKaRa) and specific TAS3 ta-siRNA primer in combination with the universal adaptor primer to examine TAS3 ta-siRNA expression. U6 was used as an internal reference. All primers for qRT-PCR analyses were listed in Supporting Information Table S1.

Examination of sensitivity to auxin and auxin transport inhibitors

Surface-sterilized seeds were germinated, transferred onto MS medium containing various amounts of 1-naphthaleneacetic acid (NAA) and auxin transport inhibitor N-1-naphthylphthalamicacid (NPA), and the lengths of primary roots were scored after 9 d in light conditions. For nodulation assay, 5-d-old rel3 or WT seedlings were transferred to B&D medium plates containing different concentrations of NAA, indole-3-acetic acid (IAA), NPA or 2,3,5-triiodobenzoic acid (TIBA), and inoculated with M. loti after 2 d. Plants were grown vertically in a growth chamber and visible nodules were counted at 3 wpi.

Ethylene sensitivity analysis

For triple response analysis, the imbibed seeds were transferred to 0.8% agar plates containing B&D medium supplemented with various concentrations of 1-aminocyclopropane-1-carboxylic acid (ACC), and kept in continuous dark conditions at 23°C. After 7 d incubation, seedlings were scanned and the length of roots and hypocotyls were measured by Image Tool Software. For the effects of ACC and L-α-(2-aminoethoxyvinyl)-glycine (AVG) on nodulation, 5-d-old-seedlings were transferred onto B&D agar plates containing different concentrations of ACC or AVG, and inoculated with M. loti after 2 d. The plants were grown vertically to ensure root elongation along agar surface. For the effect of Ag+ on nodulation, 5-d-old-seedlings were grown in pots filled with vermiculite and perlite, and watered regularly with B&D medium containing the indicated concentrations of Ag2SO4. Visible nodules were counted at 3 wpi.

Results

Symbiotic phenotypes of the rel3 mutant

In order to assess whether ta-siRNA regulation is involved in symbiosis, we investigated nodulation in the TAS3 ta-siRNA biogenesis-deficient mutant rel3. We examined the development of nodules and roots for 3 wpi. rel3 mutants developed fully infected nitrogen-fixing pink nodules (Fig. 1a–c), but developed only half as many nodules as WT (Fig. 1d). Cytological analysis showed that mature nodules of rel3 were normal, with fully differentiated symbionts, bacteroid-infected cells and uninfected cells that were indistinguishable from those in WT nodules (Fig. S1). However, the nodulation zone ratio was significantly reduced in rel3 (Fig. 1e), showing that root nodules were formed in a narrower nodulation zone on rel3 roots. Furthermore, the length of primary roots and lateral root number of rel3 were reduced (Fig. 1f,g).

Figure 1.

Nodulation and root growth phenotypes of the rel3 mutant. One-week-old Lotus japonicus wild-type (WT) and rel3 seedlings were inoculated with Mesorhizobium loti. Nodule formation and root growth were examined at 3 weeks post-inoculation. (a) Nodule formation of rel3 and WT. (b, c) Close-ups of mature nodules formed in WT and rel3. (d) Visible nodule numbers per plant. (e) The ratio of the nodulation zone per plant, calculated by dividing the length of the nodulation zone by the main root length. (f) The length of the primary roots. (g) Number of lateral roots. The data shown are means ± SE of three biological replicates are presented (n > 24). *, < 0.05; **, < 0.01, according to a two-tailed t-test. Bars, 1 cm.

We next investigated root development of rel3 grown on agar plates in nonsymbiotic conditions. We measured the length of primary roots and the number of lateral roots for 4 wk and found that the length of primary roots and the number of lateral roots in rel3 were decreased (Fig. S2). In addition, we measured the length of lateral roots of 9-d-old seedlings. The average length of lateral roots of WT was 5.1 ± 0.89 mm, whereas the average length of lateral roots of rel3 was 3.9 ± 0.97 mm, suggesting that the length of lateral roots in rel3 was reduced. The reduction in lateral root length in rel3 was consistent with the lateral root phenotype of the Arabidopsis TAS3a-1 mutant (Marin et al., 2010).

In order to verify whether the low-nodulating phenotype resulted from a mutation in REL3, we complemented rel3 mutants with a construct carrying REL3 cDNA driven by double 35S promoter (d35S:REL3) using hairy root transformation. The nodule number of transformed rel3 plants harboring d35S:REL3 was restored to the levels of WT plants transformed with an empty vector that only constitutively expressed GUS reporter (35S:GUS), while transformed rel3 plants harboring 35S:GUS formed reduced number of nodules compared with WT plants transformed with 35S:GUS (Table S2). Moreover, qRT-PCR analysis showed that REL3 expression in transgenic roots of rel3 plants harboring d35S:REL3 was similar to that in transgenic roots of WT plants harboring 35S:GUS, whereas REL3 expression in transgenic roots of rel3 plants harboring 35S:GUS was lower than that in transgenic roots of WT plants harboring 35S:GUS (Fig. S3). Therefore, REL3 expression in transgenic roots of tested plants was comparable to the nodule number formed on the roots. Additionally, the nodulation phenotype of rel3 was also complemented by introducing a construct harboring REL3 cDNA driven by its native promoter (Table S3). These data suggest that mutation of REL3 is responsible for the low-nodulating phenotype of rel3 mutants.

To further determine that TAS3 ta-siRNA regulation is involved in nodulation, we conducted nodulation assay of the TAS3 ta-siRNA deficient mutant rel1 and double mutant rel1rel3 (Yan et al., 2010). REL1 encodes a homolog of Arabidopsis SGS3 and is required for TAS3 ta-siRNA production (Yan et al., 2010). The rel1 mutant exhibits a nucleotide substitution in its coding region, leading to a frame shift and premature stop codon. Both rel1 and rel1rel3 mutants formed functional nitrogen-fixing nodules, but both nodule number and nodulation zone ratio were significantly reduced (Fig. S4). These data indicated that REL1 and REL3 act in the same genetic pathway during nodulation. Given that REL1 and REL3 are two components involved in TAS3 ta-siRNA biogenesis and that similar phenotypes were observed in rel1 and rel3, we selected the rel3 mutant for subsequent study to specify nodulation regulation of TAS3 ta-siRNA.

Both rhizobial infection and nodule initiation are impaired in the rel3 mutant

In order to determine the stage at which nodulation in rel3 mutants is affected, we examined infection events and nodule primordium formation by inoculating rel3 and WT seedlings with M. loti harboring a hemA:lacZ construct. The infection thread formation in rel3 was normal: rhizobia were entrapped in an infection pocket at the curls' center of the root hairs, and then infection threads extended from the infection pockets towards the base of the root hair cells, subsequently entering and ramifying in nodule primordia (Fig. S5). Although normal infection threads developed in rel3, infection frequency and the extent of nodulation were reduced (Fig. 2). At 5 dpi, the total number of infection events, including infection threads and infection foci, was significantly reduced in rel3. At 7 dpi, infection events in rel3 were also greatly decreased (Fig. 2). Furthermore, the number of nodule primordia in rel3 was slightly reduced at 5 and 7 dpi compared with WT (Fig. 2), suggesting that the initiation of nodule primordia is weakly inhibited in rel3 at early stages of nodulation. We further assessed nodulation kinetics in rel3. The nodule number in rel3 was slightly less than that in WT at 1 wpi, but reduced significantly at 2 and 3 wpi (Table S4). These results demonstrated that the repression of nodule formation in rel3 is triggered at early stages of nodulation, but enhanced with emergence of mature nitrogen-fixing nodules.

Figure 2.

Analyses of infection events and nodule primordia in the rel3 mutant. Lotus japonicus wild-type (WT) and rel3 seedlings were inoculated with Mesorhizobium loti containing a hemA:lacZ construct. The inoculated roots were stained at 5 and 7 dpi post-inoculation (dpi) with X-Gal, respectively. Infection events include infection foci and infection threads. At least 10 plants were evaluated. The data shown are means ± SE. **, < 0.01, according to a two-tailed t-test.

Expression analyses of REL3 during nodulation

We examined REL3 expression during nodule development by measuring REL3 transcript levels in nodules and other organs by qRT-PCR. qRT-PCR analysis revealed that REL3 was expressed at varying levels throughout the plant; high levels of transcript were detected in roots and stems, medium levels in nodules, and lower levels in leaves, flowers and immature pods (Fig. 3a). Following inoculation, REL3 transcript levels were increased in the roots at 1 and 7 dpi compared with uninoculated roots (Fig. 3b). This expression pattern was consistent with REL3 transcript data from the Lotus japonicus Gene Expression Atlas (LjGEA; Table S5; Verdier et al., 2013). These data showed that REL3 expression is enhanced in response to rhizobial infection.

Figure 3.

Temporal and spatial expression of REL3 during nodulation. (a) Quantitative real-time reverse transcription PCR (qRT-PCR) analysis of REL3 expression in different organs: R, denodulated roots; S, stems; L, leaves; F, flowers; P, pods; N, mature nodules at 21 days post-inoculation (dpi). (b) qRT-PCR analysis of REL3 transcript levels in Lotus japonicus uninoculated roots (UN) and inoculated roots (IN) at 1 and 7 dpi. (c–e) Expression of the REL3pro:GUS in central vascular bundle of primary root (c), in the dividing pericycle cells of lateral root primordium (d) and in lateral root vascular bundle (e). (f, g) Expression of the REL3pro:GUS in dividing pericycle cells of nodule primordia. (h–j) Expression of the REL3pro:GUS in peripheral vascular bundles of immature nodule (h) and mature nodule (i, j). (k) Transverse section of a root showed GUS activity primarily in pericycle and phloem cells. (l) Transverse section of a root nodule showed GUS staining in nodule peripheral vascular bundles. (m) Strong GUS staining in pericycle and phloem cells of peripheral vascular bundle. Transcript levels of REL3 were estimated using the relative Ct value method and normalized by L. japonicus UBIQUITIN gene. Error bars indicate the range of possible values based on SE of replicate Ct values. Data presented are representative of at least three independent experiments. At least 100 transgenic roots harboring REL3pro:GUS were observed. Vascular bundle (VB), pericycle (pe), xylem (X) and phloem (ph) are indicated. Bars: (c–l) 1 mm; (m) 10 μm.

In order to determine the spatial expression of REL3 during nodule development, we examined expression pattern of REL3pro:GUS in roots and nodules. GUS staining showed that REL3pro:GUS expression was restricted to the central vascular bundle of the primary root, and not detectable in the meristematic region, root cap and root hairs (Fig. 3c). Also, REL3pro:GUS was expressed in the dividing pericycle cells of lateral root primordium and developing lateral root vascular bundle (Fig. 3d,e). GUS staining of nodules at different developmental stages showed that REL3pro::GUS was specifically expressed in dividing pericycle cells of nodule primordia (Figs 3f,g, S6), in peripheral vascular bundles of immature nodules (Fig. 3h) and mature nodule (Fig. 3i,j). The transverse root section exhibited GUS expression primarily in pericycle cells and phloem cells (Fig. 3k). Transverse sections of nodules showed GUS staining in nodule peripheral vascular bundles (Fig. 3l), particularly in pericycle cells and phloem cells (Fig. 3m). These data showed that REL3 was almost exclusively expressed in vascular system of roots and nodules, suggesting that REL3 plays an important role in the root development and nodulation.

TAS3 ta-siRNA accumulation is compromised in the rel3 roots

The rel3 mutant was shown to eliminate TAS3 ta-siRNA accumulation in vegetative shoot apices detected by small RNA filter hybridization (Yan et al., 2010), we therefore wanted to know whether TAS3 ta-siRNA biosynthesis is affected in the rel3 roots. We compared accumulation of mature TAS3 ta-siRNA, 5′D7(+), in the roots of rel3 and WT inoculated with M. loti at 3 dpi using qRT-PCR. qRT-PCR analysis demonstrated that TAS3 ta-siRNA accumulation in the rel3 roots was greatly reduced in the absence or presence of rhizobia compared with WT (Fig. 4a), suggesting that biogenesis of TAS3 ta-siRNA is impaired in the mutant roots. Furthermore, we detected TAS3 ta-siRNA accumulation in transgenic roots of rel3 mutants harboring d35S:REL3 fusion or empty vector for complementation assay. qRT-PCR analysis indicated that TAS3 ta-siRNA accumulation coincided with the levels of REL3 expression and nodule numbers of composite rel3 plants (Fig. S3). These data showed that REL3-dependent TAS3 ta-siRNA biogenesis is directly associated with nodulation.

Figure 4.

Quantitative real-time reverse transcription PCR (qRT-PCR) analyses of TAS3 ta-siRNA accumulation and expression of ARF3a, ARF3b and ARF4 in rel3 roots. (a) qRT-PCR showing expression of TAS3 5′D7(+) in the roots of Lotus japonicus wild-type (WT) and rel3 mutants that were uninoculated (UN) or inoculated (IN) with Mesorhizobium loti at 3 days post-inoculation (dpi). Transcript levels of TAS3 ta-siRNA were normalized to expression of U6. (b–d) Expression of ARF3a (b), ARF3b (c) and ARF4 (d)in uninoculated (UN) or inoculated (IN) roots of rel3 mutants and WT at 1 dpi. Transcript levels were estimated using the relative Ct value method and normalized to that of the UBIQUITIN gene. Error bars indicate the range of possible values based on SD of replicate Ct values. Data presented are representative of at least three independent experiments.

We further verified the involvement of TAS3 ta-siRNAs in nodulation by examining TAS3 ta-siRNA, 5′D7(+) in the WT roots at 1 and 3 dpi. qRT-PCR analysis indicated that TAS3 ta-siRNA accumulation was rapidly induced in the roots following inoculation (Fig. S7). This results can be supported by TAS3 transcript data from LjGEA (Table S5; Verdier et al., 2013). The increase of TAS3 ta-siRNA accumulation in response to rhizobial infection suggested the important role of TAS3 ta-siRNA during nodulation, and validated that REL3-dependent TAS3 ta-siRNA biogenesis is required for nodulation.

TAS3 ta-siRNA target genes ARF3 and ARF4 are upregulated in the rel3 roots in response to rhizobial infection

ARF3 and ARF4 transcripts are targets of TAS3 ta-siRNAs, and accumulate in mutants that block the production of this ta-siRNA (Peragine et al.,2004; Allen et al.,2005; Hunter et al., 2006). Previous study uncovered that ARF3a, ARF3b and ARF4 were all upregulated in seedlings, leaves and inflorescences of rel3, indicating that the REL3-mediated TAS3 ta-siRNA pathway negatively regulates the expression of these genes (Yan et al., 2010). To better understand expression pattern of ARF3a, ARF3b and ARF4 in response to rhizobial infection, we examined these gene expression in rel3 roots at 1 dpi by qRT-PCR. qRT-PCR analysis showed that ARF3a, ARF3b and ARF4 in inoculated roots of WT were slightly downregulated compared with WT uninoculated roots (Fig. 4b–d), and this expression pattern was consistent with the data from LjGEA (Table S5; Verdier et al., 2013). Moreover, this downregulated expression of the ARF genes and the increase of TAS3 ta-siRNA accumulation in the roots following rhizobial infection (Fig. S7, Table S5), suggest that TAS3 ta-siRNAs target the mRNA of ARF3a, ARF3b and ARF4 for degradation during nodulation. In uninoculated and inoculated roots of rel3, these ARF genes were remarkably upregulated compared with WT uninoculated or inoculated roots (Fig. 4b–d), showing that deficiency of TAS3 ta-siRNA biogenesis in rel3 results in de-repression of ARF3 and ARF4. These data indicated that the TAS3 tasiR-ARF pathway is operating during nodulation.

The rel3 mutant shows altered auxin response and transport

Because ARF3 and ARF4 are upregulated in rel3 roots in the presence or absence of rhizobia, it is possible that retardation of root growth and nodulation deficiency in rel3 might be caused by disturbed auxin homeostasis or an alteration in auxin transport or response. We therefore examined the response to the synthetic auxin NAA and auxin transport inhibitor NPA in rel3. At every concentration of exogenous NAA tested (0.1 and 1 μM), root growth of WT and rel3 was inhibited, but the inhibition of root growth in rel3 was less than in WT (Fig. 5a), suggesting that root growth of rel3 is more resistant to NAA than WT. Similar to NAA treatment, at all tested concentrations of NPA, root growth of WT and rel3 seedlings was inhibited, and the inhibition of root growth in rel3 was less than in WT (Fig. 5b). These results suggest that there are some differences in auxin response or transport between rel3 and WT roots.

Figure 5.

Effects of auxins and auxin transport inhibitors on root growth and nodulation in the rel3 mutant. (a, b) Relative root length of Lotus japonicus wild-type (WT) and rel3 seeds were grown on medium containing 1-naphthaleneacetic acid (NAA) (a) and N-1-naphthylphthalamicacid (NPA) (b) for 9 d. Root growth in the presence or absence of NAA or NPA was measured and expressed as a percentage of root growth compared with the control. (c–f) Nodule numbers formed in WT and rel3 mutants in response to different concentrations of NAA (c), indole-3-acetic acid (IAA) (d), 2,3,5-triiodobenzoic acid (TIBA) (e) and NPA (f). (g–j) Nodule formation in WT(g) and rel3 mutants (i), and root tip growth in WT (h) and rel3 mutants (j) treated with 10 μM NPA. The number of visible nodules that formed at 3 weeks post-inoculation was scored. Results are presented as means ± SE (n > 30). *, < 0.05; **, < 0.01, according to a two-tailed t-test. Bars, 1 cm.

In order to explore how the altered auxin response or transport affects nodulation in rel3, we examined nodule formation in rel3 treated with auxin. Inoculated rel3 and WT plants were grown on nitrogen-free B&D agar plates supplemented with various concentrations of NAA or IAA for nodulation. The low concentration of 0.1 μM NAA or 0.01 μM IAA increased the nodule number of WT (Fig. 5c,d), in agreement with the observation that low concentrations of auxin increase nodulation in Medicago truncatula (van Noorden et al., 2006). However, the low concentration of auxins failed to increase the nodule number of rel3 mutants (Fig. 5c,d), suggesting that the rel3 mutant is partially insensitive to exogenous auxins during nodulation. At the higher concentrations of 1 and 10 μM NAA or IAA, the nodule number of WT and rel3 was greatly reduced (Fig. 5c,d), but the relative inhibition in rel3 was less than in WT, and this finding further confirmed that the rel3 mutant is less sensitive to auxin than WT during nodulation.

We validated that alteration of auxin response or transport in rel3 influences nodulation by treating rel3 with the auxin transport inhibitors TIBA and NPA. Nodule number of WT and rel3 decreased with increasing concentrations of TIBA and NPA, but the relative inhibition of nodule number in rel3 was stronger than that in WT treated with 1 μM TIBA and NPA (Fig. 5e,f), suggesting that rel3 mutants are more sensitive to NPA and TIBA than WT during nodulation. Surprisingly, when treated with 1 or 10 μM NPA, rel3 roots appeared much more agravitropic than WT roots, even waving in the presence of rhizobia (Fig. 5g–j). These results suggest that rhizobial infection can enhance the agravitropic response to NPA more in rel3 than in WT, and that the altered auxin response or transport in rel3 roots is responsible for the low-nodulating phenotype of rel3 mutants.

Auxin distribution and accumulation is altered in the rel3 roots during nodulation

The alteration in response or polar auxin transport in rel3 roots likely causes change of auxin content and distribution. Research evidence indicates that activity of the auxin-responsive reporter DR5:GUS correlates well with endogenous auxin concentrations in the roots (Casimiro et al., 2001; Benkova et al., 2003). Therefore, we examined DR5:GUS activity in the root tips and elongation zone of rel3 and WT. In uninoculated transformed roots of WT, strong DR5:GUS activity was observed in central vascular tissues of the root elongation zone, and a maximum of GUS activity was observed in the columella initials, with faint activity in mature root cap cells (Fig. 6a). By contrast, in uninoculated transformed roots of rel3, relatively weak DR5:GUS activity was confined to root central vascular tissues, and GUS activity was expanded across the whole root tip, including columella initials, root cap cells, and proximal and distal lateral root cap cells (Fig. 6c). Following inoculation, DR5 activity shifted upward in the root tip domain of both rel3 and WT (Fig. 6b,d), suggesting that rhizobial infection can induce auxin redistribution in infected roots. However, DR5 activity in root central vascular tissue in rel3 was still lower than that in WT, as observed in uninoculated rel3 roots (Fig. 6). Studies on auxin-responsive reporters GH3:GUS or DR5:GUS expression indicate that local auxin accumulation in nodule primodium, pericycle and cortical cells of the infected sites is required for nodule initiation (Mathesius et al., 1998; Pacios-Bras et al., 2003; Wasson et al., 2006; Suzaki et al., 2012). Hence, DR5:GUS expression analysis reflected that auxin concentration and distribution in rel3 roots are altered, and this alteration is linked to the low-nodulating phenotype of rel3 mutants.

Figure 6.

Expression of the auxin-responsive reporter DR5:GUS in the rel3 roots. (a, c) Expression of DR5:GUS in uninoculated transgenic roots (UN) of Lotus japonicus wild-type (WT) (a) and rel3 (c). (b, d) Expression of DR5:GUS in inoculated roots (IN) of WT (b) and rel3 (d) at 1 dpi. Representative pictures of transgenic roots harboring DR5:GUS in rel3 and WT were shown. The number of transgenic roots for each construct and treatment were at least 100. Boxed regions indicate expression pattern of DR5:GUS. Bars, 100 μm.

The rel3 mutant exhibits different responses to ethylene

Ethylene negatively regulates nodulation, especially by inhibiting rhizobial infection (Penmetsa & Cook, 1997), and rel3 mutants had reduced infection events and nodules; therefore we wanted to know whether ethylene regulation is involved in the low-nodulating phenotype of rel3 mutants. We first examined the sensitivity of rel3 to ethylene by applying the ethylene immediate precursor ACC. In the presence of ACC, both rel3 and WT seedlings displayed the typical triple response, namely a shorter, thicker hypocotyl, inhibition of root elongation, and an exaggerated apical hook, and showed the triple response in a dose-dependent manner (Fig. 7a). However, the rel3 roots were shorter than WT roots at 10 or 100 μM ACC (Fig. 7b), suggesting that ACC inhibits root growth more in rel3 than in WT, and thus that the rel3 roots are more sensitive to ethylene. However, the hypocotyl elongation of rel3 was similar to WT at all concentrations of ACC tested (Fig. 7c). These results revealed that the rel3 dark-grown roots, but not hypocotyls, have an enhanced response or perception to the ethylene signal.

Figure 7.

Effects of ethylene on root growth and nodule formation of the rel3 mutant. (a) Ethylene triple response displayed by Lotus japonicus wild-type (WT) and rel3 seedlings. The imbibed seeds were grown on different concentrations of 1-aminocyclopropane-1-carboxylic acid (ACC) for 7 d in the dark. (b, c) Effects of ACC on elongation of roots (b) and hypocotyls (c). Relative root and hypocotyl lengths are shown as percentage of seedlings in the absence of ACC. (d) Effects of ACC on nodule formation. Nodule number in the presence or absence of ACC was measured and relative nodule number presented as the percentage of nodule number compared with the air control. (e) Nodules formed in different concentrations of L-α-(2-aminoethoxyvinyl)-glycine (AVG). (f) Nodules formed in different concentrations of Ag2SO4. One-week-old L. japonicus WT and rel3 seedlings were inoculated with Mesorhizobium loti, and then grown on B&D agar plates in the absence or presence of various concentrations of ACC and AVG, or planted into pots filled with vermiculite and perlite immersed with liquid B&D medium supplemented different concentrations of Ag2SO4, and watered regularly with B&D medium containing the indicated concentrations of Ag2SO4. Visible nodules were counted at 3 weeks post-inoculation. Mean values ± SE are presented. *, < 0.05; **, < 0.01, according to a two-tailed t- test.

In order to determine the ethylene sensitivity of nodulation in rel3, we examined nodulation on B&D agar plates supplemented with various concentrations of ACC. The nodule numbers in rel3 and WT were reduced in a dose-dependent manner at all concentrations of ACC (Fig. 7d). However, when treated with 1 or 10 μM ACC, the nodule number in rel3 was reduced less than in WT (Fig. 7d). This result demonstrated that the relative nodulation inhibition by ethylene in rel3 is weaker than in WT, and suggested that because there are already high ethylene responses in rel3 roots, ACC treatment could not cause ethylene response efficiently as WT did during nodulation, and thus give a differential nodulation response to ethylene. These data support the idea that root-specific enhanced response to ethylene signal is connected with the reduction in nodulation of rel3 mutants.

We further validated the root-specific enhanced sensitivity to ethylene of rel3 affects nodulation by comparing nodulation of rel3 and WT on either the ethylene biosynthesis inhibitor AVG, or the ethylene perception inhibitor Ag+. At low concentrations of AVG (1 μM), the nodule number of both rel3 and WT increased (Fig. 7e), consistent with the observations that AVG can increase nodulation in many tested legumes (Heidstra et al., 1997). At high concentrations of AVG (10 μM) or Ag2SO4 (5, 10 and 100 μM), nodulation in WT was decreased (Fig. 7e,f). Moreover, at 1 μM AVG and all tested concentrations of Ag2SO4, the nodule number of rel3 was similar to that of WT (Fig. 7e,f), suggesting that the appropriate amount of AVG and Ag+ can restore the low-nodulating phenotype of rel3. These data showed that the nodulation phenotype of rel3 may be caused by increased ethylene biosynthesis or perception.

Auxin regulation of nodulation is dependent on ethylene synthesis or response in the rel3 mutant

In light of the observation that rel3 has altered responses to ethylene and auxin during nodulation, we wanted to know how the interaction of ethylene and auxin affects nodulation. Therefore, we examined nodulation of rel3 and WT treated with 1 μM NPA and 1 μM AVG simultaneously. We observed that when treated with NPA only, the rel3 roots remained more agravitropic than the WT roots, but when treated with the combination of NPA and AVG, the rel3 roots showed no strong agravitropic response (Fig. S8), suggesting that the effect of NPA was alleviated markedly by AVG treatment in rel3. Only NPA treatment decreased the nodule number of both WT and rel3, but NPA inhibition of nodulation in rel3 is stronger than in WT (Fig. 8). Both NPA and AVG treatment can increase nodulation in rel3 and WT, and the extent of difference in nodule number between rel3 mutants and WT was reduced compared with only NPA treatment (Fig. 8). These results showed that effect of NPA on nodulation possibly acts downstream of the effect of AVG in rel3 roots.

Figure 8.

Nodule formation of the rel3 mutant in the presence of 1 μM NPA and 1 μM AVG. Nodule number of Lotus japonicus wild-type (WT) and rel3 plants in the absence (−) or presence (+) of 1 μM N-1-naphthylphthalamicacid (NPA) and 1 μM L-α-(2-aminoethoxyvinyl)-glycine (AVG). One-week-old L. japonicus WT and rel3 seedlings were inoculated with Mesorhizobium loti and grown on B&D agar plates supplemented with NPA or AVG. Visible nodules were counted at 3 weeks post-inoculation. Mean values ± SE are presented. * < 0.05; **, < 0.01, according to a two-tailed t-test.

Discussion

In this work, we found that the rel3 mutant formed fewer nodules, but the morphology and nitrogen-fixing function of these nodules are comparable to those of WT nodules (Fig. 1), suggesting that REL3 – a key component of TAS3 ta-siRNA biosynthesis in Lotus japonicus – is required for efficient nodulation. Furthermore, rel3 mutants showed reduced primary root length and lateral root number in symbiotic and nonsymbiotic conditions (Figs 1, S2). In addition, the rel1 and rel1rel3 mutants also displayed similar phenotypes (Fig. S4). These results revealed that root growth and nodule development are co-regulated by the TAS3 ta-siRNA pathway.

Four ta-siRNA precursors (TAS1-4) were identified in Arabidopsis, but cleavage of TAS3 is unique because it requires the specific action of the miR390/AGO7 complex for ta-siRNA production (Allen et al., 2005; Axtell et al., 2006; Montgomery et al., 2008). Numerous pieces of evidence indicate that: accumulation of all TAS1-, TAS2- and TAS3-derived ta-siRNAs depend on DCL4, RDR6 and SGS3; only TAS3-derived ta-siRNAs depend on AGO7; and AGO7 is required for the production and/or stability of tasiR-ARF, TAS3-derived ta-siRNA, which targets both ARF3 and ARF4 (Adenot et al., 2006; Fahlgren et al., 2006; Garcia et al., 2006; Hunter et al., 2006; Montgomery et al., 2008). This pathway and its target genes were also indentified in rice, moss and maize, indicating a highly conserved mechanism for the TAS3 ta-siRNA pathway (Timmermans et al., 1998; Itoh et al., 2000; Nagasaki et al., 2007; Nogueira et al., 2007; Douglas et al., 2010). REL3 is the ortholog of Arabidopsis AGO7, and mutation of REL3 results in elimination of TAS3 ta-siRNA production and upregulated expression of ARF3a, ARF3b and ARF4 in shoots (Yan et al., 2010). In this work, we showed that TAS3 ta-siRNA accumulation was greatly reduced, and ARF3a, ARF3b and ARF4 were strongly upregulated, in uninoculated or inoculated roots of rel3 (Fig. 4). Moreover, in response to rhizobial infection, these ARF genes were downregulated while TAS3 ta-siRNA accumulation was increased in WT roots. Together with research data on tasiR-ARF in other plants, these findings clearly showed TAS3 ta-siRNA cleavage of mRNAs of these ARF genes (Fig. S7, Table S5). Therefore, the low-nodulating phenotype of rel3 is due to de-repression of ARF3 and ARF4 by deficiency in TAS3 ta-siRNA biogenesis. The remarkable decrease in TAS3 ta-siRNA accumulation and evidently upregulated expression of ARF genes in rel3 during nodulation were consistent with their changes in aerial parts, suggesting that the conserved tasiR-ARF pathway for aerial organ development is recruited for nodule formation.

Because the formation of nitrogen-fixing nodules requires integration of rhizobial infection in epidermal cells and nodule primordium initiation in the cortex, there exist different features for tasiR-ARFs regulation in nodule formation and other organ development. We observed that both infection events and nodule initiation were reduced in rel3 (Fig. 2, Table S4), and these data showed that the REL3-mediated TAS3 ta-siRNA pathway modulates both rhizobial infection and nodule initiation concurrently. Although mutations of components involved in TAS3 ta-siRNA biosynthesis can cause more serious morphological and functional modification of aerial lateral organs (Itoh et al., 2000; Nagasaki et al., 2007; Abe et al., 2010; Yan et al., 2010), the morphology and function of rel3 nodules were not apparently affected, suggesting that the molecular mechanism and downstream components of TAS3-derived tasiR-ARF in controlling nodule formation are divergent.

Specific expression of REL3 in root and nodule vascular tissues (Fig. 3) was consistent with the expression pattern of Arabidopsis AGO7. AGO7 promoter is predominantly expressed in the vasculature of seedlinge and the pith region just below the shoot apical meristem (Montgomery et al., 2008; Chitwood et al., 2009). Additionally, AGO7 is also expressed within the adaxial-most cells of developing leaf primordia (Fahlgren et al., 2006; Garcia et al., 2006). This localized expression of AGO7 and TAS3A restricts the domain of tasiR-ARF biogenesis to the two upper-most cell layers of leaves which generates a tasiR-ARF gradient from adaxial surface to abaxial surface, suggesting that TAS3 ta-siRNA serves as a mobile silencing signal for leaf patterning (Chitwood et al., 2009). The vasculature-specific expression of REL3, together with research evidence on AGO7, leads us to assume that the apparent restricted region of REL3 expression may limit the domain of tasiR-ARF biogenesis in roots and nodule primordia, and thereby regulate nodulation. However, this assumption needs to be validated.

Substantial evidence has shown that auxin concentrations, localization and auxin polar transport are crucial for nodule development (Mathesius et al., 1998; Pacios-Bras et al., 2003; Wasson et al., 2006), while ethylene negatively regulates rhizobial infection and positions nodules opposite xylem poles (Ferguson et al., 2010; Desbrosses & Stougaard, 2011). However, how auxin and ethylene signaling interact during nodulation is poorly understood. In this work, the rel3 mutant showed an altered auxin and ethylene response during nodulation (Figs 5, 7). The rel3 mutant exhibited a decreased sensitivity to exogenous auxin during nodulation, but greater sensitivity to auxin transport inhibitors (Fig. 5). These altered auxin responses can be verified by the observation that the rel3 mutant showed a more agravitropic response in the presence of NPA and rhizobia than WT (Fig. 7). These data suggested that auxin response and transport in rel3 roots are different from those of WT during nodulation, supporting the conclusion that tasiR-ARFs mediate the effects of auxin (Williams et al., 2005; Fahlgren et al., 2006; Hunter et al., 2006). Furthermore, the rel3 mutant displayed an altered ethylene response due to the fact that ethylene inhibition of nodulation in rel3 is weaker than in WT, and because the low-nodulating phenotype of rel3 can be restored by chemicals that antagonize ethylene synthesis or action (Fig. 7). Additionally, ethylene-insensitive M. truncatula sickle (EIN2) mutants display hyperinfected and increased nodulation (Penmetsa et al., 2008), transgenic Lotus plants harboring the mutated melon ethylene receptor CmERS1/H70A or the dominant Arabidopsis ETR1-1 mutant gene show insensitivity to ACC, concomitant with an increase of infection threads and nodule primordia (Nukui et al., 2004; Lohar et al., 2009). These data support the inference that the enhanced root-specific ethylene sensitivity in rel3 correlates with its low infection events and reduced nodulation. Research on sickle mutants further indicates that ethylene signaling regulates nodulation by modulating auxin transport in nodule development (Prayitno et al., 2006). Consistent with this, the inhibition of nodulation by NPA in rel3 was weakened by AVG treatment (Fig. 8), showing that the effect of auxin transport inhibitor on nodulation in rel3 is dependent on modification of ethylene synthesis or response. Therefore, we suggest that the REL3-derived TAS3 tasiR-ARF pathway participates in ethylene-mediated regulation of auxin response and transport.

We examined auxin concentrations and distribution in uninoculated and inoculated roots of rel3 using the auxin-responsive reporter DR5:GUS, and found that auxin concentrations and distribution in rel3 roots were changed (Fig. 6). DR5:GUS expression in the root tip of rel3 was elevated compared with WT (Fig. 6a,c) in the absence of rhizobia. Following inoculation, DR5:GUS activity was extended and upshifted in root tips of rel3 and WT (Fig. 6); this result demonstrated that rhizobial infection can induce auxin redistribution in the root tips. Moreover, the more agravitropic phenotype of inoculated rel3 roots in the presence of NPA further confirmed that there are differences in auxin transport and distribution in rel3 and WT roots. Research evidence shows that localized auxin redistribution in the root tip is required for maintaining the root gravitropic response, and basipetal auxin transport that occurs in the peripheral layers of cells – primarily the epidermal and cortical cells – is also essential for gravitropism in Arabidopsis roots (Marchant et al., 1999; Rashotte et al., 2000; Casimiro et al., 2001). So it is reasonable to hypothesize that rhizobial infection can increase auxin accumulation in root tips but decrease basipetal auxin transport to root cortical cells in the elongation zone of rel3 roots treated with NPA, and thus produce a more agravitropic response. Moreover, The observation that DR5:GUS activity in central vascular tissues of the nodulation zone in rel3 is lower than that in WT with or without rhizobial inoculation also supports this hypothesis (Fig. 6). Studies on the supernodulating mutant sunn indicate that high concentrations of endogenous auxin correlate with increased nodule numbers (van Noorden et al., 2006). Therefore, low DR5:GUS expression in central vascular tissues of rel3 roots is associated with its low-nodulating phenotype.

Some ARFs connect auxin and ethylene signaling in Arabidopsis. ARF19 and ARF7 participate in auxin signaling and ethylene responses in roots (Li et al., 2006); ARF2 links auxin, ethylene and light-regulated apical hook formation (Li et al., 2004). This study revealed that the upregulated expression of ARF3 and ARF4 in rel3 resulted in altered ethylene and auxin responses during nodulation (Figs 4-8), and it is therefore likely that ARF3 and ARF4 may serve as an integration point for auxin and ethylene signaling mediated by the REL3-dependent tasiR-ARF pathway. Although our data reveal that the tasiR-ARF pathway regulates nodulation by integrating auxin and ethylene signaling, it remains unclear how ethylene affects auxin transport and concentrations in rel3 roots at the molecular level. Therefore, further work is needed to clarify the mechanisms by which ta-siRNAs govern ethylene-auxin crosstalk during nodulation in L. japonicus.

Acknowledgements

We thank Professor Jens Stougarrd and Fang Xie for their instructive suggestions on this manuscript. We thank Professor Ton Bisseling for providing bacterial strain MSU440. This work was supported by 973 Program of China (No. 2010CB126501) and NSFC Programs (No. 31270292 and 31071065).

Ancillary