The Lotus japonicus Ubiquitin Ligase SIE3 Interacts With the Transcription Factor SIP1 and Forms a Homodimer

The symbiosis receptor kinase SymRK plays an essential role in symbiotic signal transduction and nodule organogenesis. Several proteins bind to SymRK, but how the symbiosis signals are transduced from SymRK to downstream components remains elusive. We previously demonstrated that both SymRK interacting protein 1 (SIP1, an ARID-type DNA-binding protein) and SymRK interacting E3 ligase [SIE3, a RING (Really Interesting New Gene)-containing E3 ligase] interact with SymRK to regulate downstream cellular responses in Lotus japonicus during the legume-rhizobia symbiosis. Here, we show that SIE3 interacts with SIP1 in both yeast cells and Nicotiana benthamiana. SIE3 associated with itself and formed a homodimer. The cysteine 266 residue was found to be essential for SIE3 dimerization and for promoting nodulation in transgenic hairy roots of L. japonicus. Our findings provide a foundation for further investigating the regulatory mechanisms of the SymRK-mediated signaling pathway, as well as the biological function of E3 ligase dimerization in nodule organogenesis.


INTRODUCTION
The establishment of the rhizobia-legume symbiosis is a tightly regulated process that integrates bacterial infection steps with specialized organ development. Nodule formation is activated in response to rhizobia-derived nodulation factors (NFs), a group of lipochitooligosaccharides (LCOs) with different host-specific decorations (Long, 1989;Lerouge et al., 1990). NF action is accompanied by a series of signal transduction processes inside root cells. Intense research has focused on elucidating NF signaling, which plays an essential role in nodule organogenesis (Dénarié et al., 1996;Oldroyd and Downie, 2008;Masson-Boivin et al., 2009). Genetic dissection of nodulation in legumes such as Medicago truncatula and Lotus japonicus has led to the discovery of NF receptors and several key players in the NF signaling pathway Downie, 2004, 2006). Symbiosis receptor-like kinase (SymRK) is required for symbiotic signal transduction upon stimulation of root cells by microbial signaling molecules (Endre et al., 2002;Stracke et al., 2002). However, the exact biochemical functions and regulatory mechanisms of SymRK remain unknown.
In this study, we demonstrate that SIE3 is a novel plant dimeric E3 ubiquitin ligase whose disulfide linkage at Cys266 plays key roles in the formation and symbiosis function of SIE3 homodimers. Moreover, we demonstrate that the SIE3 E3 ligase interacts with the transcription factor SIP1.

SIE3 Interacts With SIP1 in Yeast Cells
We previously demonstrated that SymRK interacts with both SIP1, an ARID-type transcription factor and SIE3, a RING-type E3 ubiquitin ligase in L. japonicus (Zhu et al., 2008;Yuan et al., 2012). Here, we investigated whether SIP1 interacts with SIE3. To test this hypothesis, we conducted a yeast two-hybrid (Y2H) assay to examine the interaction between SIP1 and SIE3. As shown in Figure 1, yeast cells containing BD-SIE3/AD-SIP1 or AD-SIE3/BD-SIP1 grew on quadruple dropout SD medium and had higher β-galactosidase activity than the negative controls ( Figure 1A). The expression of recombinant proteins in yeast was confirmed by immunoblot analysis using anti-hemagglutinin (HA) or anti-Myc monoclonal antibodies ( Figure 1B). These results indicate that SIE3 associates with SIP1 in yeast cells.

Interaction of SIE3 With SIP1 in planta
We performed bimolecular fluorescence complementation (BiFC) assays to investigate whether the SIE3/SIP1 interaction occurs in planta using Nicotiana benthamiana leaf cells. SIP1 was fused to the split C-terminus of CFP (SIP1::SCC), while SIE3 was fused with the split N-terminus of this protein (SCN::SIE3). We analyzed N. benthamiana leaf epidermal cells 2-5 days after infiltration with Agrobacterium tumefaciens harboring these constructs. Strong fluorescent signals from leaves expressing both SIP1::SCC and SCN::SIE3 were observed in the plasma membrane and nucleus. This pattern was similar to that of leaves expressing the positive control proteins Arabidopsis CALCINEURIN B-LIKE (CBL) and CBL-INTERACTING PROTEIN KINASE 24 (CIPK24) (Figure 2A; Waadt et al., 2008). In the negative controls, where SIP1::SCC and SCN::SIE3 were expressed separately or SCN::NFR5 and SCC::SIP1, and SCN::SIE3 and SCC::NFR5 were expressed, no fluorescent signals were observed (Figure 2A). The expression levels of recombinant proteins in N. benthamiana leaves were confirmed by immunoblot analysis using anti-HA or anti-Flag monoclonal antibodies ( Figure 2B). These results indicate that SIP1 and SIE3 interact with each other in planta.

SIE3 Dimerization
When SIE3 was expressed in planta, we observed a high molecular mass ( Figure 2B). Based on its size, we reasoned that this band might correspond to SIE3 homodimer. To examine whether reduced thiol compounds would affect SIE3 dimerization, we tested whether SIE3 monomer could form homodimers with or without the reducing agent DTT (McGinnes and Morrison, 1998) in the reaction (Figure 3). In buffer lacking DTT, the majority of band signals were detected above the 130 kDa position, representing potential SIE3 dimers in the reaction ( . These results suggest that the upper band corresponds to the SIE3 homodimer and the lower band to the monomer, and that the presence of DTT in the reaction mixture helped to stabilize the SIE3 monomer.

Cysteine 266 of SIE3 Is Required for Homodimerization
To pinpoint the peptide domains required for homodimerization of SIE3, we generated deletions of SIE3 cDNA ( Figure 4A) and used the Y2H assay to test the interactions between them. The SIE3 deletion mutant containing the domains C-terminal region to LisH (CTLH) and CRA-RanBPM but lacking the C-terminal RING domain (SIE3-RING) interacted with itself, suggesting that the RING domain is not essential for homodimerization ( Figure 4B). When both the CRA-RanBPM and RING domains were removed (SIE3-CR), no self-interaction of SIE3 was observed ( Figure 4B). Taken together, these results suggest FIGURE 1 | SIE3 interacts with SIP1 in yeast cells. (A) Interaction between SIE3 and SIP1 in yeast cells. Proteins were fused with the Gal4 DNA binding domain (BD) in pGBKT7 or with its activation domain (AD) in pGADT7. Yeast cells harboring the constructs were maintained on SD/-Trp-Leu medium (SD-2) and selected for protein-protein interactions on SD/-Trp-Leu-His-Ade (SD-4) or SD-2/X-gal medium. The strength of the interaction was evaluated based on β-galactosidase activity (Miller units). At least three biological replicates were performed, and the data are presented as the mean ± SD. The combination p53/SV40 served as a positive control, and Lam/SV40, BD-SIE3/AD and BD-SIP1/AD served as negative controls. (B) Immunoblot analysis of protein levels in yeast cells. Anti-HA monoclonal antibody was used to detect the expression levels of HA-tagged proteins (AD-SV40, AD-SIE3, and AD-SIP1). Anti-Myc monoclonal antibody was used to detect the levels of Myc-tagged proteins (BD-53, BD-Lam, BD-SIE3, and BD-SIP1).
that the CRA-RanBPM domain is essential for the selfinteraction of SIE3.
Since DTT destroyed the dimerization of SIE3 (Figure 3), we hypothesized that the disulfide bonds whose formation is mediated by cysteine residues might be required for SIE3 homodimerization. Among the seven cysteine residues of SIE3, six are located in the RING-finger domain, which was not necessary for the SIE3-SIE3 interaction in yeast cells. The remaining cysteine reside (Cys266) is present in the CRA-RanBPM domain, which was essential for the self-interaction of SIE3. We therefore hypothesized that Cys266 might be the key residue controlling SIE3 homodimerization. In fact, the SIE3 form with Cys266 substituted to serine (SIE3-C266S) was unable to form homodimers in yeast or N. benthamiana leaf cells (Figures 4C, 5). This result, together with the observation that the SIE3 dimer was sensitive to DTT (Figure 3), indicates that the dimerization of SIE3 is mediated by the formation of a disulfide bond via Cys266.
The results of this work indicate that SIE3 forms a homodimer. We asked whether this homodimer is required for its function in the nodule symbiosis. We expressed SIE3 and SIE3-C266S under the control of the maize Ubiquitin promoter (SIE3-OX and SIE3-C266S-OX) in transgenic hairy roots of L. japonicus and scored the nodulation phenotypes 3 weeks after inoculation with Mesorhizobium loti MAFF303099 expressing β-galactosidase (lacZ), a constitutive marker used to observe rhizobial cells (Tansengco et al., 2003). Compared to control transgenic roots, significantly more nodules were produced in SIE3-OX hairy roots, whereas the number of nodules in SIE3-C266S-OX hairy roots was similar to that of the transgenic hairy roots expressing empty vector (control) (Figures 6A,B). Quantitative PCR (qPCR) analysis indicated that the expression level of the transgene in SIE3-OX or SIE3-C266S-OX hairy roots was 3-15-fold higher compared to the control ( Figure 6C). qPCR analysis of ENOD40-1, an early nodulin gene implicated in the processes of rhizobial infection, nodule initiation, and subsequent organogenesis (Kumagai et al., 2006) was performed. The results showed that ENOD40-1 was expressed at higher levels in SIE3-OX hairy roots than in the control ( Figure 6D). These data indicate that cysteine-266 of SIE3 is essential for its function in promoting nodulation in L. japonicus.

DISCUSSION
SymRK is an essential player in bacterial entry and is required for nodulation in legumes (Stracke et al., 2002;Radutoiu et al., 2003;Yoshida and Parniske, 2005). Various proteins that interact with SymRK have recently been identified. However, how NF signals are transduced downstream of SymRK remains unclear. To better understand the relationships between SymRK-interacting proteins, we investigated the interaction between SIP1 and SIE3, two SymRK-interacting proteins. Our results demonstrate that SIE3 and SIP1 interact with each other both in yeast and in planta.
Our results also demonstrate that SIE3 can form homodimers through a disulfide linkage at Cys266 and that a mutation at Cys266 of SIE3 abolished its role in symbiosis.
RING domain E3 ubiquitin ligases are modular proteins with RING and substrate-binding domains. Mammalian RING finger protein 4 (RNF4) is a dimeric RING E3 ligase. The RING domain of RNF4 is responsible for ubiquitin transfer and its dimerization, which are essential for E3 ligase activity (Liew et al., 2010). During the ubiquitination reaction, the activated ubiquitin tag is covalently linked to a cysteine (Cys) residue of E2 conjugase via a thioester bond. In the RNF4 dimer, one subunit is bound to E2 conjugase while the other subunit engages the thioester-linked ubiquitin (Plechanovová et al., 2011). Thus, the dimerization of RNF4 is required for the transfer of ubiquitin from E2 conjugase to the substrate. Dimerization and similar catalytic mechanisms have been observed for several other RING-type E3 ubiquitin ligases, including cIAP2 (Cellular inhibitor of apoptosis protein 2) (Mace et al., 2008), TRAF6 (TNF receptor-associated factor 6) (Vander Kooi et al., 2006;Yin et al., 2009), MDM2-MDMX19, BARD1 [breast cancer susceptibility gene 1 (BRCA1)-BRCA1-associated RING domain protein 1] (Brzovic et al., 2001) Ring1b-Bmi1 (B-cell-specific Moloney murine leukemia virus integration site 1) (Buchwald et al., 2006) and Arabidopsis SINA5 (Xie et al., 2002). However, little is known about the role of E3 dimerization in legume plants.
L. japonicus SIE3 interacted with itself in yeast cells and N. benthamiana leaves (Figures 2B, 4, 5). The SIE3 monomer might have become a homodimer during the reaction process (Figure 3). It appears that SIE3 primarily exists as a homodimer in plants. The ubiquitination of SymRK is mediated by SIE3 and the RING domain of SIE3 appears to be essential for its interactions with SymRK (Yuan et al., 2012). However, unlike the above-mentioned E3 proteins, the CRA-RanBPM domain of SIE3 rather than its RING domain is essential for SIE3 dimerization ( Figure 4B). Thus, SIE3 might represent a novel plant dimeric E3 ubiquitin ligase.
Protein dimerization can occur through the formation of disulfide bonds and is sensitive to DTT (McGinnes and Morrison, 1998). DTT has a strong effect on the formation of SIE3 homodimers (Figure 3). The CRA-RanBPM domain of SIE3 is essential for its homodimerization (Figure 4B). The CRA-RanBPM domain contains a single cysteine residue (C266). The replacement of this residue with serine (SIE3-C266S) blocked the formation of SIE3 homodimers in yeast and N. benthamiana leaves (Figures 4C, 5), further implying that SIE3 dimers form through a disulfide linkage at Cys266. In addition, significantly more nodules were produced in SIE3OX hairy roots than in the control (Yuan et al., 2012; Figures 6A,B), while there were no more or even fewer nodules in SIE3-C266SOX hairy roots than in the control (Figures 6A,B). In Arabidopsis, the expression of the dominantnegative SINA5-C49S affected the endogenous E3 ligase activity of SINA5, thereby downregulating the ubiquitination of the NAC1 (NUCLEUS ACCUMBENS-ASSOCIATED 1) transcription factor and auxin signaling in plant cells (Xie et al., 2002). Similarly, the presence of SIE3-C266S perhaps affected the ability of endogenous SIE3 to form homodimers, thereby downregulating the normal biological functioning of SIE3.
Previously, our laboratory has shown that SIP1 can interact with SymRK (Zhu et al., 2008). In addition, SIE3 can interact with SymRK and mediate ubiquitination of SymRK in L. japonicus (Yuan et al., 2012). In this report, SIE3 can interact with SIP1 in yeast and in planta (Figure 1 and Figure 2). It is possible that SIP1, SIE3, and SymRK form a trimeric complex, in which the E3 ligase SIE3 is bound to two protein substrates. In the gibberellic acid (GA) signaling pathway, the interaction between the GA receptor GID1 and its downstream component DELLA protein can help SCF SLY1/GID2 to gain its DELLA protein specificity (Schwechheimer, 2008). In a proposed model for CUL7-mediated TBC1D3 oncoprotein degradation in human, a third protein, Fbw8, has been shown to interact with CUL7 and also to be required for TBC1D3 degradation (Kong et al., 2012). However, whether a trimeric complex of SIE3, SymRK, and SIP1 exists in developing nodules during the establishment of the nodule symbiosis awaits further experimental confirmation.
In summary, SIE3 appears to represent a novel dimeric E3 ubiquitin ligase in plants. The cysteine 266 residue was found to be essential for SIE3 dimerization and for promoting nodulation in transgenic hairy roots of L. japonicus. SIE3, SIP1, and SymRK, have previously been shown to play important roles in the transduction of NF signals and during early nodule organogenesis. The current results shed new light on the complex relationships among SymRK, SIE3, and SIP1 and provide a foundation for further investigating the regulatory mechanisms of the SymRK-mediated signaling pathway, as well as the biological function of E3 ligase dimerization in nodule organogenesis.

Plant Materials and Growth Conditions
N. benthamiana plants were grown in a growth chamber at 22 • C and 40-70% relative humidity (RH) under a 16 h light/8 h dark photoperiod for about 4-6 weeks before infiltration with A. tumefaciens strain EHA105. After infiltration, plants were kept under the same growth conditions. Wild-type (WT) plants of L. japonicus (Handberg and Stougaard, 1992). "Miyakojima MG-20" were used for hairy root transformation. Seeds were scarified by immersion in concentrated H 2 SO 4 for 8 min before surface sterilization in 2% NaClO plus 0.1% Tween-20 for 20 min. Seeds were plated on 1/2 MS medium supplemented with 0.8% agar for germination at 28 • C in the dark for 2 days and then transferred to a growth chamber with a 16/8 h light/dark cycle at 23 • C.

Protein-Protein Interaction in Yeast Cells
Full-length cDNA of SIE3, SIE3-RING, SIE3-CR, and SIE3-C266S cDNAs were amplified by PCR and inserted into NdeI/EcoRI sites of pGBKT7 or pGADT7 (Yuan et al., 2012). AD-SIP1 and BD-SIP1 have been described in previous papers (Zhu et al., 2008;Wang et al., 2013). Small-scale mating tests were performed to examine protein-protein interactions. After mating, yeast cells harboring the constructs were maintained on SD/-Trp-Leu medium (SD-2) and selected for proteinprotein interaction on SD/-Trp-Leu-His-Ade (SD-4) or SD-2/Xgal. Colonies on SD/-Leu-Trp-His-Ade plates were transferred to the SD/-Leu-Trp/X-gal plate for further verification using β-galactosidase assay (Clontech). Proteins were extracted from yeast expressing recombinant proteins using yeast cracking buffer (40 mM Tris-HCl pH = 6.8), 8 M urea, 5% (w/v) SDS, 0.1 mM EDTA, 5% β-mercaptoethanol, 10 mM DTT, 0.4% (w/v) Bromophenol blue, Protease inhibitor solution (Roche), 1 mM phenylmethylsulfonyl fluoride. The cracking buffer was prewarmed to 60 • C, and cell pellets were thawn quickly by separately resuspending each sample in the prewarmed cracking buffer. Each sample was transferred to a 1.5 mL FIGURE 6 | Effect of SIE3-C266S overexpression on nodulation. (A) Nodulation phenotypes of WT, SIE3-OX, and SIE3-C266S-OX hairy roots. Transgenic hairy roots were produced using A. rhizogenes cells carrying the SIE3-OX or SIE3-C266S-OX construct. Transgenic hairy roots transformed with the empty vector (p1301U) served as the control. The hairy roots were inoculated with M. loti and grown in nitrogen fertilizer-free soil to induce nodulation. Three weeks after rhizobial inoculation, one representative independent transgenic hairy root (with nodules) per sample was photographed; yellow arrows indicate nodules. Bars, 5 mm. (B) Number of nodules per plant of transgenic hairy roots with altered SIE3 and SIE3-C266S transcript levels. Transgenic hairy roots expressing empty vector (pU1301) served as a control. Two large-scale experiments were conducted, and the total number of independent transgenic plants per sample is indicated in parentheses. Data represent the mean ± SE. (C) Analysis of SIE3 transcript levels in control (control), SIE3-OX (SIE3-OX-1, SIE3-OX-2, SIE3-OX-3), and SIE3-C266S-OX (SIE3-C266S-OX-1, SIE3-C266S-OX-2, SIE3-C266S-OX-3) hairy roots. Total RNA isolated from individual root systems was used for qPCR. Relative expression levels of SIE3 transcripts in SIE3-OX and SIE3-C266S-OX hairy roots were calculated with reference to that of control hairy roots. (D) qPCR analysis of Enod40-1 transcript levels in control and transgenic hairy roots. Total combined RNA isolated from the root systems (including small nodules) of multiple independent transgenic plants was used for qPCR analysis. Relative expression levels of Enod40-1 in transgenic hairy roots were calculated with reference to that of control hairy roots, at least three independent biological replicates and three technical replicates for each sample. The data are presented as the mean ± SD (C,D). "*"and "**" indicate significant differences between samples (t-test, p < 0.05 or p < 0.01, respectively). microcentrifuge tube containing glass beads, and heated at 70 • C for 10 min, vortexed vigorously for 2 min, and centrifuged at 14,000 rpm for 5 min. The supernatants were then transferred to 1.5 mL microcentrifuge tubes and kept as first supernatants. The pellets were boiled for 3-5 min, vortexed vigorously for 1 min, and centrifuged at 14,000 rpm for 5 min. These second supernatants were then combined with the first supernatants. The samples were briefly boiled and loaded on a SDS-PAGE gel. Anti-HA monoclonal antibody (EarthOx, 1:1000) was used to detect the expression levels of HA-tagged proteins AD-SIE3 (60 kD) and AD-SIP1 (63.2 kD); Anti-Myc monoclonal (EarthOx, 1:1000) antibody was used to detect the expression levels of Myc-tagged proteins BD-SIE3 (63.2 kD) and . Tests were performed with at least three biological replicates.

β-Galactosidase Assay
Yeast cells grown in liquid selection media were measured at OD 600 and centrifuged for 30 s at 13,000 rpm. Cells were washed twice with Z-buffer (60 mM Na 2 HPO 3 , 40 mM NaH 2 PO 3 , 10 mM KCl, 1.0 mM MgSO 4 , pH = 7.0), and resuspended in 100 µL of Z-buffer, followed by permeabilization by three freeze-thaw cycles in liquid nitrogen and 37 • C water bath. After centrifugation, cell extracts were added to 0.7 mL of Z-buffer containing 50 mM β-mercaptoethanol and 160 µL of ONPG (Onitrophenyl β-D-galactopyranoside, 4 mg/mL in Z-buffer). After incubation at 30 • C until the yellow color appeared, the reaction was terminated by the addition of 0.4 mL of 1.0 M Na 2 CO 3 . The reaction mixture was centrifuged for 10 min at 13,000 rpm to remove cell debris. β-Galactosidase activity in the supernatant was measured at OD 420 , and expressed in Miller units (Griffith and Wolf, 2002). All reactions were performed with at least three biological replicates.

BiFC Experiments
The full-length cDNA of SIP1, SIE3, and SIE3-C266S were cloned into the SpeI/KpnI site of pSCYCE-R, which contains a HAtag (Waadt et al., 2008) to obtain SIP1::SCC, SIE3::SCC and SIE3-C266S::SCC fusions. The full-length cDNA of SIE3 and SIE3-C266S were cloned into the BamHI/XhoI site of pSCYNE, which contains a FLAG-tag (Waadt et al., 2008) to obtain SIE3::SCN and SIE3-C266S::SCN. Plasmids were introduced into A. tumefaciens strain EHA105 cells by electroporation. Agrobacterium cells containing plasmids were grown in liquid medium (LB broth with 50 µM kanamycin, pH = 7.0, 10 mM MES, pH = 5.7, 40 µM acetosyringone) were pelleted and resuspended in the infiltration buffer (10 mM MgCl 2 , 200 µM acetosyringone). Agrobacterium strains containing different plasmids were mixed to a final OD 600 of 0.75 for each strain, and then mixed with the Agrobacterium strain containing the gene-silencing suppressor p19 at a final OD 600 of 1.0 in a volume ratio of 1:1. The mixture of Agrobacterium strains was incubated at room temperature for 4 h, and used for infiltration into the leaves of 4-6 week-old N. benthamiana plants using a 1 mL syringe. Cyan fluorescence was observed 3-5 days after infiltration of leaf cells expressing these proteins under the confocal microscope OLYMPUS BX61WI equipped with a CFP filter set (excitation/emission wavelengths of 405/477 nm). The BiFC analyses were performed with at least three independent biological replicates.

Protein Extraction and Immunoblot Analysis
Proteins were extracted from N. benthamiana leaves expressing recombinant proteins using native extraction buffer 1 (NB1) and denaturing buffer (DB) as described (Yuan et al., 2012). NB1 contained 50 mM Tris-MES, pH = 8.0, 0.5 M sucrose, 1 mM MgCl 2 , 10 mM EDTA, 5 mM DTT, and protease inhibitor cocktail CompleteMini tablets (Roche). DB contained 50 mM Tris-HCl, pH = 7.5, 150 mM NaCl, 0.1% NP-40, 4 M urea, and protease inhibitor cocktail CompleteMini tablets. The leaf areas surrounding the infiltrated sites were harvested, and ground in liquid nitrogen, then 1 mL of each leaf powder was filled in a 2 mL centrifuge tube. Leaf powders were resuspended in extraction buffer on ice. Total extracts were centrifuged at 13,000 rpm at 4 • C for 30 min. Supernatants were subjected to Western blot analysis. Same quantity of leaf powders were extracted (corresponding to same quantity of total leaf protein) and used to normalize loading the SDS-PAGE gels, and Ponceau S staining of the Rubisco band on the membrane was used to assess the amounts of samples on each lane on the SDS-PAGE gels.
The sources and dilutions of antibodies used in the experiments were as follows: anti-HA antibody (Sigma, 1:5000; EarthOx, 1:1000), anti-Flag antibody (EarthOx, 1:1000), anti-GFP antibody (Abmart, 1:1000). Proteins were separated by SDS-PAGE in a 10 or 12% acrylamide gel and semi-dry electroblotted to nitrocellulose membrane (Hybond-C, Amersham) 1 at 25 V for 40 min. The membrane was blocked with PBS containing 5% skimmed milk powder for 1 h at room temperature or overnight at 4 • C. The membrane was then incubated first with primary antibody in PBS containing 3% skimmed milk for more than 1 h at room temperature, and then with secondary antibody diluted in TBS containing 3% skimmed milk for 1 h at room temperature. Bands were visualized with the Millipore chemiluminescent HRP substrate kit. Anti-HA monoclonal antibody was used to detect the expression levels of HA-tagged proteins SIP1::SCC (58 kD), SIE3:: SCC (52.8 kD), and SIE3-C266S::SCC (52.8 kD). Anti-Flag monoclonal antibody was used to detect the expression levels of Flag-tagged proteins SCN::SIE3 (52.7 kD) and SCN::SIE3-C266S (52.7 kD). Anti-GFP monoclonal antibody was used to detect the expression levels of SIE3::GFP (68.5 kD). All experiments were performed at least three independent biological replicates.

Overexpression of SIE3 and SIE3-C266S by Hairy Root Transformation
The full-length CDS of SIE3 and SIE3-C266S were cloned into the KpnI/BamHI site of p1301U to obtain pMUb: SIE3 and pMUb: SIE3-C266S, respectively. A. rhizogenes strain LBA1334 cells carrying pMUb: SIE3 or pMUb: SIE3-C266S were used to induce hairy root formation in wild-type L. japonicus "MG-20" using a procedure as described previously (Yuan et al., 2012). Nodulation phenotypes of transgenic hairy roots were scored 3 weeks after inoculation with M. loti MAFF303099. Transgenic hairy roots expressing the empty vector (p1301U) were used as a control. The transgenic hairy roots (with nodules) were photographed; the mean values of nodule number and Student's t-tests were performed using software SPSS Statistics 17.0. The expression level of SIE3 and Enod40-1 in SIE3-OX or SIE3-C266S-OX hairy roots was determined by qPCR using the primers and the procedure as described previously (Yuan et al., 2012).

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.