A recently evolved Nodulin recruited by NIN for root hair rhizobial infection in Medicago truncatula

Summary

• In root nodule symbiosis, rhizobia enter into legume nodules either by intracellular infection via root hairs or by intercellular infection such as crack entry. The formation of the infection thread (IT) is key to root hair infection, but the underlying mechanisms are unclear. Although transcriptome studies have revealed gene regulatory networks of some central regulators such as NODULE INCEPTION (NIN), the specific roles of many downstream genes in IT development and their integration mechanism into the symbiotic pathway remain unknown.

• Through genetics, cell biological approaches and evolutionary analysis, we characterized the role and evolutionary origin of Medicago truncatula Nodulin 6 (N6) in rhizobial infection.

• MtN6 is specifically induced by rhizobia in root hairs. Mtn6 mutant shows rhizobial infection defects with abnormal ITs in root hairs and nodules. MtN6 encodes a cell wall‐localized protein which is secreted into infection chamber and ITs. NIN regulates rhizobia‐induced MtN6 expression by directly binding its promoter. N6 is only present in legumes adopting root hair rhizobial infection mode and is likely to have originated from Nodulin/glutamine synthase‐like protein by gene duplication.

• We conclude that N6 is a newly formed gene recruited by NIN for root hair rhizobial infection during the evolution of root nodule symbiosis.

Expression pattern of MtN6 in root hairs and nodules.

Introduction

The symbiotic nitrogen‐fixing system between leguminous plants and rhizobia constitutes a fundamental component of the nitrogen cycle in nature (Canfield et al., ²⁰¹⁰). Legumes establish this symbiotic relationship with soil‐dwelling rhizobia through their root systems, leading to the formation of specialized structures known as root nodules (Downie, 2014). In root nodules, rhizobia convert atmospheric nitrogen into ammonia, which is accessible to plants, thereby supplying essential nitrogen nutrition to the host (Roy et al., ²⁰²⁰). The formation of indeterminate root nodules, such as those in Medicago truncatula, originates from cell division of the pericycle and cortical tissues in the root (Xiao et al., ²⁰¹⁴; Kohlen et al., ²⁰¹⁸). Nodule development is tightly coordinated with ‘rhizobial infection’, the process of soil rhizobia entry into host roots and subsequent colonization of developing nodules (Oldroyd & Downie, 2008; Schiessl & Jhu, ²⁰²⁵). Both processes are dependent on precise molecular dialogue between plants and rhizobia (Liu & Murray, 2016; D. Wang et al., ²⁰²²). When the Nodulation (Nod) factors secreted by rhizobia are recognized by specific receptors on the surface of host root hair cells, the symbiotic signaling pathway is activated, which involves nuclear calcium signaling (Ehrhardt et al., ¹⁹⁹⁶; Lévy et al., ²⁰⁰⁴; Charpentier et al., ²⁰¹⁶; Kelly et al., ²⁰¹⁷). This activation triggers a cascade of gene expressions and cellular reprogramming, initiating the processes of rhizobial infection and nodule organogenesis (Oldroyd, 2013; Mergaert et al., ²⁰²⁰; Luo et al., ²⁰²³).

The infection of legumes by rhizobia primarily occurs through two pathways: intercellular infection associated with lateral root emergence (‘crack entry’) or between root epidermal cells and intracellular infection starting from root hairs (Sprent, 2007; Sharma et al., ²⁰²⁰). The former is common in legumes such as Aeschynomene and Arachis spp., whereas the latter strategy is predominantly utilized in legumes such as pea, soybean, Medicago truncatula and Lotus japonicus (Quilbé et al., ²⁰²²). Distinct signaling pathways may mediate intercellular and intracellular rhizobial infection, but the conservation and divergence of genetic networks controlling these two infection modes is still unclear (Montiel et al., ²⁰²¹; Bhattacharjee et al., ²⁰²²). Partly due to the model systems used for legume rhizobial symbiosis, where intracellular infection mode is predominant, our current understanding of mechanisms underlying intercellular rhizobial infection is rather limited (Horta Araújo et al., ²⁰²⁴). Intracellular infection via root hairs involves a specialized tubular structure known as the ‘infection thread’ (IT). In most legumes, this process begins with the perception and attachment of rhizobia to root hairs, which subsequently curl to encapsulate the bacteria, forming an infection chamber (IC) (Fournier et al., ²⁰¹⁵). Within this chamber, rhizobia proliferate to form microcolonies (Gage et al., ¹⁹⁹⁶). Then, from the IC, the cell wall and membrane of the root hair grow inwardly to initiate the tubular IT, which is also deemed as a newly formed apoplastic compartment (de Carvalho‐Niebel et al., ²⁰²⁴). The IT extends downward through the root hair via polar tip growth, with its growth direction aligning with the movement trajectory of the nucleus (Gage, 2004). The leading nucleus and the growing IT remain connected through a cytoplasmic bridge rich in endoplasmic reticulum (Fournier et al., ²⁰⁰⁸). The IT provides a temporary habitat for rhizobia to proliferate, and the residing rhizobia population moves forward along with the progression of ITs. This process reiterates in the underlying cortical cells, leading to the continuous branching and extension of the IT in the developing nodule (Gage, 2004; Fournier et al., ²⁰⁰⁸; Murray, ²⁰¹¹). In indeterminate nodules such as that of M. truncatula, rhizobia continuously ‘infect’ the host via ramifying ITs in the infection zone of nodules. Eventually, the rhizobia are released from the IT into nodule cells, enveloped by the host membrane as symbiosomes and differentiate into bacteroids capable of nitrogen fixation (Ivanov et al., ²⁰¹²; Zhao et al., ²⁰²⁵).

Rhizobial infection necessitates extensive cell wall remodeling (Brewin, 2004; Parniske, ²⁰¹⁸; Monroy‐Morales et al., ²⁰²⁶). The legume host secretes cell wall enzymes such as Nodulation Pectate Lyase (NPL), symbiosis‐specific pectin methyl esterases (SyPME1), Glycoside Hydrolase 9C (GH9C) and L. japonicus α‐expansin EXPA1 for IC formation, IT progression and/or rhizobial release (Xie et al., ²⁰¹²; Liu et al., ^2019a; Su et al., ²⁰²³; Zhang et al., ^2024a; Montiel et al., ²⁰²⁵; Zhao et al., ²⁰²⁵). Rearrangement of the cytoskeleton is also required at different stages of rhizobial infection, that is, from initial root hair deformation to the polar growth of IT, and a few actin‐ and microtubule‐related proteins have been shown to be required for these processes (Yokota et al., ²⁰⁰⁹; Hossain et al., ²⁰¹²; Qiu et al., ²⁰¹⁵; Gavrin et al., ²⁰²⁰; Su et al., ²⁰²⁰; Liang et al., ²⁰²¹; Gao et al., ²⁰²²). Genetic screens have revealed several host components that regulate the polar growth of IT, which includes VAPYRIN (VPY), LUMPY INFECTIONS (LIN) and RHIZOBIUM‐DIRECTED POLAR GROWTH (RPG) (Kuppusamy et al., ²⁰⁰⁴; Arrighi et al., ²⁰⁰⁸; Kiss et al., ²⁰⁰⁹; Yano et al., ²⁰⁰⁹; Murray, ²⁰¹¹; Liu et al., ²⁰²¹). These proteins form the so‐called ‘infectosome’ protein complex, together with an exocyst subunit EXO70H4 at the very tip of IT, and around the nucleus (Liu et al., ^2019b; Lace et al., ²⁰²³; Li et al., ²⁰²³; Deng et al., ²⁰²⁴). Although the exact function of the infectosome is far from clear, it is thought to be required for the polar delivery of cell wall and membrane materials for building the IT (Liu et al., ^2019b; Yang et al., ²⁰²³).

Transcriptional reprogramming is a prerequisite for the establishment of legume‐rhizobial symbiosis and genes involved in rhizobial infection are under the tight control of transcription factors of the legume host (Mergaert et al., ²⁰²⁰). One of the central transcription regulators is NODULE INCEPTION (NIN), which is also prominent for the evolution of root nodule symbiosis (Schauser et al., ¹⁹⁹⁹; Marsh et al., ²⁰⁰⁷; Liu & Bisseling, ²⁰²⁰; Shen & Feng, ²⁰²³; Geurts & Huisman, ²⁰²⁴; Doyle et al., ²⁰²⁵; Liu et al., ^{2025, 2026}; Bhardwaj et al., ²⁰²⁶; Cathebras et al., ²⁰²⁶; Nosaki et al., ²⁰²⁶). NIN directly regulates the expression of genes required for IT development, such as NPL, NF‐YA1, SCARN and RPG (Xie et al., ²⁰¹²; Soyano et al., ²⁰¹³; Qiu et al., ²⁰¹⁵; Li et al., ²⁰²³). A previous comparative transcriptome study reveals that the expression of a large set of rhizobium‐induced genes is affected in the M. truncatula nin‐1 mutant, many of which are potential transcriptional targets of NIN (Liu et al., ^2019a). However, whether and how these genes regulate rhizobial infection are still unknown.

Here, we characterize M. truncatula Nodulin 6 (N6) as a direct transcriptional target of NIN. MtN6 is localized to the IT wall, and Mtn6 mutants exhibit defective rhizobial infection in root hairs and nodules. Systematic analysis suggests that N6 has recently evolved in certain legume lineages, likely recruited by NIN for intracellular rhizobial infection via root hairs.

Materials and Methods

Plant and rhizobia materials

Medicago truncatula Gaertn. ecotypes R108 and Jemalong A17 were used as wild‐type (WT) for genotyping and generating composite transgenic plants, respectively. Mtn6‐1 (NF16674) and Mtn6‐2 (NF16505) were insertion mutants in the R108 genetic background with tobacco (Nicotiana tabacum) Tnt1 retrotransposon, and both mutant alleles were identified in this research. nin‐1 is a fast‐neutron mutagenesis mutant in the A17 background (Marsh et al., ²⁰⁰⁷).

The following rhizobial strains were used in this study: Sinorhizobium meliloti strain Rm1021 (Rm1021) harboring pXLGD4 (hemA:lacZ) (Rm1021‐LacZ), S. meliloti strain Rm1021 harboring a pHC60‐derived plasmid conferring constitutive Cerulean CFP expression (Rm1021‐CFP) and S. meliloti strain 2011 with mCherry tagged (Sm2011‐mCherry) (Liu et al., ^2019b).

Rhizobial infection and NOD phenotyping

Medicago truncatula seeds were treated with sulfuric acid for 5 min, washed with sterilized ddH₂O for five times, treated with bleach for 5 min and washed again with sterilized ddH₂O for five times. The seeds were immersed in sterilized ddH₂O for 3 h and transferred onto water agar (0.4%) plates at 4°C for 2 d. The seeds were transferred to the dark at 23°C overnight before sowing.

For symbiotic phenotyping, germinated seedlings of R108, Mtn6‐1 and Mtn6‐2 were sown in vermiculite/perlite (1 : 1) mix and cultivated in a controlled environment chamber with a 16 h : 8 h, light : dark photoperiod. The seedlings were inoculated with S. meliloti 1021‐LacZ (cultured overnight at 28°C, OD₆₀₀ = 0.05). Plant roots were harvested at 10 dpi, and the roots were fixed in a 2.5% glutaraldehyde solution for 1 h, followed by staining in 0.8 mg ml⁻¹ X‐GAL (5‐Bromo‐4‐chloro‐3‐indolyl‐β d‐galactoside; Sangon Biotech, Shanghai, China) solution (100 mM sodium phosphate, 10 mM KCl, 1 mM MgSO₄, pH = 7.4) at 28°C in the dark overnight. Rhizobial infection events were then quantified.

For Nod phenotyping, roots were harvested at 10, 14, 21 and 30 dpi, and nodule numbers were quantified. For infection phenotyping in nodule cells, nodules were treated with an ethanol gradient for dehydration, embedded in paraffin (Biosharp, Beijing, China), sectioned into slices at a thickness of 5 μm using a Leica RM 2235 rotary microtome and stained with 0.5% toluidine blue for 12 min. Images of ITs and nodule sections were captured using a digital camera mounted on either a Zeiss Axio Scope A1 microscope or a Leica DM2500 LED microscope.

Promoter‐GUS assay

A 2195‐bp MtN6 promoter was amplified using HiFi Platinum Polymerase (Invitrogen) and ligated into pDONR207 using Gateway BP clonase (Invitrogen). The BP product pDONR207‐pMtN6 was cloned into pKGWFS7 using Gateway LR clonase (Invitrogen) to make pMtN6:GUS. For truncation and deletion analysis of the MtN6 promoter, additional 1443, 700 and 1973 bp (with S1 region deleted) promoters of MtN6 were used as level 0 modules, cloned into level 1 vectors and then cloned into level 2 backbone vector EC50507 (Patron et al., ²⁰¹⁵).

Transgenic hairy root plants harboring corresponding promoter‐GUS vectors were cultured in a controlled environment chamber for 4 wk before inoculation with S. meliloti 1021‐LacZ (OD₆₀₀ = 0.05). Plants were harvested at 5, 7 and 21 dpi, and subsequently stained with 1 mg ml⁻¹ X‐Gluc solution (100 mM potassium phosphate, 1 mM potassium ferricyanide, 1 mM potassium ferrocyanide, 10 mM EDTA, pH = 7.0) for several hours at 37°C in the dark. To investigate GUS staining within the nodules, nodules collected at 21 dpi were sectioned into 70‐μm‐thick slices using a vibratome.

Confocal laser scanning microscopy for rhizobial infection

The DNA fragment encoding the N6 sequence was synthesized by Sangon Biotech (Shanghai), and the MtN6 promoter was amplified by PCR from genomic DNA utilizing Phanta Max Super‐Fidelity DNA Polymerase (Vazyme, Nanjing, China). Both components served as level 0 modules, and the level 1 construct pN6:N6‐GFP was generated by assembling level 0 modules. A level 2 vector with the backbone EC50507 was made by integrating another level 1 construct, pAtUBQ10:DsRed, which served as a transgenic marker and a control for protein subcellular localization. pVPY:MtN6‐GFP was constructed using the same method. For the colocalization construct, the secondary vector pMtN6:MtN6‐GFP pGH9C2:GH9C2‐mCherry was generated by combining the level 1 vectors pMtN6:MtN6‐GFP and pGH9C2:GH9C2‐mCherry.

A live‐cell imaging system based on confocal laser scanning microscopy was employed to observe rhizobial infection in M. truncatula root hairs (Fournier et al., ²⁰⁰⁸). Transgenic hairy roots harboring different constructs were cultivated on plates containing modified Fahraeus medium (with 3 mM MgSO₄, 0.5% Phytagel and 100 nM 2‐aminoethoxyvinyl glycine) and vertically cultured in a growth chamber. The roots were covered with Lumox film (Sarstedt, Leicester, Leicestershire, UK) and inoculated with S. meliloti 1021‐CFP or S. meliloti 2011‐mCherry (OD₆₀₀ = 0.001). Root systems were imaged at different infection stages using the Leica STELLARIS 5 or Olympus FV3000 confocal laser scanning microscopy systems.

Nodules were collected from composite transgenic plants from 10 dpi for protein localization and subcellular dynamics during rhizobial infection. Manually made nodule sections were placed on glass slides containing PBS buffer for further imaging using the Leica STELLARIS 5 or Olympus FV3000 confocal laser scanning microscopy systems.

The wavelength settings for fluorophore excitation (ex) and signal collection (em) were as follows: GFP 488 nm (ex)/500–530 nm (em), DsRed 561 nm (ex)/570–610 nm (em), mCherry 561 nm (ex)/600–630 nm (em), CFP 457 nm (ex)/465–485 nm (em). Confocal image analysis and Z‐axis projection (maximum) were performed using FV31S‐SW. GFP, mCherry and CFP were pseudo‐colored in green, red and magenta, respectively.

Confocal microscopy of nodules stained with fluorescent dyes

Nodules of different genotypes were harvested at 10 and 21 dpi, fixed in 2.5% (v/v) glutaraldehyde for 1 h and sectioned at a thickness of 70 μm with a vibratome before staining.

To analyze the rhizobial infection phenotype, nodule sections were double‐stained with Calcofluor White Stain (Sigma‐Aldrich) and SYTO13 (Invitrogen). Nodule sections were stained with 5 μM SYTO13 in the dark for 20 min and washed three times with distilled water. Then, the nodule sections were transferred onto a microscope slide with a drop of Calcofluor White solution (composed of 0.1% (w/v) Calcofluor White M2R and 0.05% (w/v) Evans blue) and a drop of 10% (w/v) KOH. The sample was imaged after 1 min. FM4‐64 (Invitrogen) was utilized to stain the IT membrane. Nodules from composite transgenic plants at 10 dpi were manually sectioned, placed on glass slides containing FM4‐64 at a final concentration of 20 μg ml⁻¹ and observed under the microscope immediately.

The wavelength settings for fluorophore ex and signal collection (em) were as follows: SYTO13, 488 nm (ex)/500–530 nm (em); Calcofluor White, 405 nm (ex)/420–470 nm (em); and FM4‐64, 561 nm (ex)/700–780 nm (em). The images were captured using the Olympus FV3000 or Leica STELLARIS 5 confocal laser scanning microscopy system. Confocal image analysis was performed using the FV31S‐SW software. SYTO13, Calcofluor White and FM4‐64 were pseudo‐colored in green, magenta and red, respectively.

Transmission electron microscopy

Nodules from WT R108, Mtn6‐1 and Mtn6‐2 were collected at 10 dpi for transmission electron microscopy (TEM) observation. The nodules were longitudinally sectioned and immediately fixed in 0.1 M PBS buffer containing 2.5% glutaraldehyde and 4% paraformaldehyde under vacuum at room temperature for 3 h, stored overnight at 4°C, washed three times with 0.1 M PBS buffer and fixed with 1% (w/v) OsO₄ overnight at room temperature. The samples were dehydrated at room temperature using a gradient series of ethanol: 30%, 50%, 70% and 95% EtOH, each for 20 min, followed by two treatments with 100% EtOH, each lasting 30 min.

The nodules were then sequentially infiltrated with a mixture of SpurrResin (SPI Supplies) and acetone in the ratio of 1 : 3, 1 : 1 and 3 : 1, each step for 24 h. The nodules were then incubated in 100% resin twice, each for 24 h and placed in an oven at 70°C for 2 d for polymerization. Sectioning was made at a thickness of 70 nm using a Leica EM UC7 ultramicrotome, and sections were collected on 200 mesh copper grids and stained with 2% uranyl acetate for 5 min followed by Reynolds' lead citrate solution for 5 min. The sections were observed using a Tecnai T12 120 kV TEM equipped with an Eagle CCD and a CryoBox.

Transient dual‐luciferase reporter assay

The S1 fragment of the MtN6 promoter was amplified from genomic DNA and cloned into the pGreenII 0800‐LUC vector using the HindIII and KpnI restriction sites to make the reporter construct. The full‐length sequence of NIN was cloned into the pRI101 vector using KpnI and EcoRI sites to generate the effector construct. Four‐week‐old Nicotiana benthamiana plants were co‐infiltrated with GV3101 strains carrying different combinations of constructs with the target reporter gene, effector or empty vector. The activities of firefly luciferase (LUC) and Renilla luciferase (REN LUC) were quantitatively measured using the dual‐LUC Reporter Assay system (Promega). The reporter gene LUC activity was normalized by co‐expressed REN LUC activity.

Protein expression and purification

The constructed recombinant plasmid was transformed into Escherichia coli BL21 competent cells and plated onto LB solid medium containing ampicillin antibiotics. A single colony was inoculated into an LB liquid medium containing ampicillin, followed by overnight incubation at 37°C. The cultures were inoculated at a 1 : 100 ratio into 200 ml antibiotic‐containing LB medium and incubated at 37°C for c. 2–3 h until OD₆₀₀ reached 0.5–0.6. The culture was then kept at 10°C for 20 min, followed by the addition of isopropyl β‐d‐1‐thiogalactopyranoside (IPTG) with a final concentration of 0.25 mM. A culture without IPTG was used as the control. The bacteria were incubated at 16°C with shaking for 16 h, kept at 10°C for 10–30 min and centrifuged at 4°C at 3434  g for 10 min to obtain the pellet. The bacteria were resuspended using 20 ml of prechilled HIS‐extract buffer (containing 50 mM NaH₂PO₄, 300 mM NaCl and 10 mM imidazole, pH = 7.4, with the final concentrations of 1 mM PMSF and 1 mg ml⁻¹ lysozyme), transferred to a 50‐ml centrifuge tube and incubated at 10°C for 30 min to ensure complete cell lysis.

The resuspended solution was ultrasonicated until it became clear and centrifuged at 30 000  g for 20 min at 4°C. Then, the supernatant was filtered through a 0.45‐μm membrane to a new centrifuge tube. The protein purification column was washed with 5 ml of HIS‐extract buffer and 20–50 μl of Ni‐NTA agarose beads was then added. The column containing the beads was washed twice with HIS‐extract buffer. The protein filtrate was allowed to flow through the column naturally. The HIS‐wash buffer (containing 50 mM NaH₂PO₄, 300 mM NaCl and 40 mM imidazole, pH = 7.4) was applied to the column one to three times to elute impurity proteins. Subsequently, HIS‐elution buffer (containing 50 mM NaH₂PO₄, 300 mM NaCl and 300 mM imidazole, pH = 7.4) was applied to the column. Finally, imidazole was removed from the eluate through ultrafiltration to obtain the purified protein.

Electrophoretic mobility‐shift assays

To create constructs for electrophoretic mobility‐shift assays (EMSAs), the partial‐length coding DNA sequence (CDS) of NIN (encoding C‐terminal residues 566–933 aa) was cloned into pDONR207 using the KpnI and EcoRI restriction sites, recombined into pCOLD‐TF to create pCOLD‐NIN, which was subsequently transformed into E. coli BL21. The pN6 probes were synthesized (Sangon Biotech, Shanghai, China) and labeled with 6‐FAM at the 5′ end. Probes were converted from single‐stranded to double‐stranded by annealing at 95°C for 10 min.

The purified NIN protein was then subsequently incubated with the 5′ 6‐FAM labeled probe at 37°C for 30 min in EMSA buffer (20 mM Tris, pH = 7.9, 5% glycerol, 40 ng μl⁻¹ bovine serum albumin (BSA), 50 mM MgCl₂ and 5 mM dithiothreitol). The reaction mixture was then electrophoresed on a 6% native polyacrylamide gel with running buffer (25 mM Tris, 192 mM EDTA, Boric acid) and imaged with a Tanon MINI Space 3000 (Shanghai Tanon Science & Technology Co. Ltd, Shanghai, China).

The MEME suite was utilized to analyze the potential NBS sites contained within the 2195 bp region upstream of the N6 promoter. For the N6 promoter fragment being used in EMSA, the mutant variant N6m had two base pair mutations in the conserved NBS (from 5′‐AATATATGTCCTTATGCATATTTGAACAAGCATGAGAA‐3′ to 5′‐AATATATGTCCTTATGCATATTTG‐CAAGCATGAGAA‐3′).

RNA extraction and real‐time polymerase chain reaction analysis

The root segment spanning the susceptible zone was harvested under stereomicroscope. Total RNA was extracted from samples using an Ultrapure RNA kit (Tiangen, Beijing, China). The cDNA was synthesized using the HiScript III RT SuperMix for quantitative polymerase chain reaction (+gDNA wiper) according to the manufacturer's protocols. The quantitative reverse transcription polymerase chain reaction (qRT‐PCR) was carried out using StepOne Plus (Applied Biosystems, Singapore). Medicago truncatula Elongation Factor 1α (EF1α) was used as the reference gene. The sequences of primers are listed in Supporting Information Table S1. Relative expression was quantified using the 2^−ΔΔCt method from three biological replicates.

N6 evolutionary analysis

A total of 39 N6 homologs from 20 legume species were used for phylogenetic reconstruction. A total of 190 homologs of MtN6 from 88 species were used for phylogenetic tree. Aeschynomene evenia C. Wright, Arachis ipaensis Krapov. & W. C. Greg, Faidherbia albida (Delile) A. Chev., Lupinus angustifolius L. and Pisum sativum L. from the Legume Information System (Berendzen et al., ²⁰²¹; Table S2; Notes S1) and others from Phytozome (https://phytozome‐next.jgi.doe.gov/) databases. Phylogenetic analysis was performed using OneClick workflow on NGphylogeny.fr with default parameters (Lemoine et al., ²⁰¹⁹). The phylogenetic trees were visualized using Interactive Tree of Life (iTOL) v.6 (Letunic & Bork, 2024). Synteny analysis utilized genome assemblies from M. truncatula Gaertn. 4.0, Trifolium pratense L. v.2, L. japonicus (Regel) K. Larsen Lj1.0v.1, Phaseolus vulgaris L. v.2.1 and Vigna unguiculata (L.) Walp. in Phytozome (Goodstein et al., ²⁰¹¹) and A. ipaensis and A. evenia v.1.2 from the Legume Information System (Berendzen et al., ²⁰²¹). Syntenic relationships among the legume genomes (Aeschynomene, Arachis, Phaseolus, Vigna, Lotus, Medicago and Trifolium) were identified using the JCVI comparative genomics toolkit (Tang et al., ²⁰²⁴). CDSs were processed to retain only the primary transcript isoforms for each gene. Gene coordinates were extracted from GFF annotations and converted into BED format using jcvi.formats.gff bed. Pairwise orthologous gene relationships between each species and M. truncatula were inferred using the jcvi.compara.catalog ortholog module with a C‐score cutoff of 0.7 to retain high‐confidence matches. These orthologs were used to construct initial synteny anchors. Syntenic blocks were generated from anchor pairs using the MCScan algorithm implemented in JCVI, followed by a single round of synteny screening and refinement to remove weak or ambiguous alignments. The resulting high‐confidence anchor blocks were retained for downstream analyses. For multispecies comparisons, pairwise synteny outputs were consolidated using JCVI base utilities to generate a unified dataset of syntenic blocks across all species. To examine local collinearity around MtN6, syntenic regions were further refined by extracting genes located within 10 genes upstream and five genes downstream of the MtN6 locus across all genomes. Final syntenic relationships were visualized using jcvi.graphics.synteny, based on curated synteny block files and genome layout definition. The evolutionary timetree was generated using TimeTree 5 (TimeTree.org) by inputting the selected species list (Table S2) and retrieving divergence time estimates from the TimeTree database.

Results

MtN6 expression is associated with rhizobial infection in root hairs and nodules

Early gene expression investigations in M. truncatula nodules showed that MtN6 encodes an early Nodulin, which is homologous to the protein domain at the N terminus of Nodulin/glutamine synthase‐like protein (NodGS) (Gamas et al., ¹⁹⁹⁶; Mathis et al., ¹⁹⁹⁹; D'Souza et al., ²⁰⁰¹). In our previous Medicago root hair ‘infectome’ studies (Breakspear et al., ²⁰¹⁴), MtN6 (Medtr1g062710, Mt4.0v.1) was found to be expressed in root hairs in the early stages upon rhizobial inoculation (Fig. 1a). Compared with that in WT, MtN6 expression is greatly reduced in nin‐1 root hairs at 5 d post inoculation (dpi) with S. meliloti, (Fig. 1b) (Liu et al., ^2019a). In the root hairs of M. truncatula, the expression of N6 is also induced by Nod factors at 4 hpi (Fig. S1) (Jardinaud et al., ²⁰¹⁶). These results prompted us to study MtN6's function during rhizobial infection.

Fig. 1.

Expression pattern of MtN6 in root hairs and nodules. (a) MtN6 expression is induced in Medicago truncatula root hairs at 3 and 5 d post inoculation (dpi) with Sinorhizobium meliloti 1021 (Sm1021). Data analyzed from Breakspear et al. (2014)). (b) Rhizobia‐induced root hair expression of MtN6 is dependent on NODULE INCEPTION (NIN). Data analyzed from Liu et al. (2019a). WT‐C, wild‐type (WT) inoculated with S. meliloti 1021 nodΔD1ABC. All other labels (WT, ern1‐1, nf‐ya1‐1, nin‐1) indicate plants with corresponding genotypes inoculated S. meliloti 1021. (c–g) Expression pattern of MtN6 as shown promoter‐GUS assay. MtN6 is expressed in a root hair harboring an infection thread (IT) (c) and the cortical cells beneath the infected root hair at 5 dpi (d, e). Expression of MtN6 in a nodule primordium at 7 dpi (f) and a mature nodule at 21 dpi (g). The nodule section was cut at a thickness of 70 μm (f, g). Red arrowhead indicates ITs in root hairs (c, d) or in the cortex (e) and white arrowhead indicates cortical cells underlying an infected root hair (d). Promoter GUS analysis was performed using at least 12 independent transgenic roots in each experiment and three independent experiments were performed. Two‐tailed Student's t‐test: **, P < 0.01; ***, P < 0.001. Error bars represent SD. Bars, 25 μm (c, d), 50 μm (e, g), 100 μm (f).

To further investigate the spatiotemporal expression pattern of MtN6, we utilized a construct with fusion of the MtN6 promoter to the GUS gene (pMtN6:GUS) in composite plants generated via Agrobacterium rhizogenes‐mediated hairy root transformation (Boisson‐Dernier et al., ²⁰⁰¹). X‐Gluc staining showed that MtN6 was primarily expressed in the root tip, lateral root primordia and root stele (Fig. S2a). After rhizobial inoculation, MtN6 was highly expressed in root hair cells, especially those harboring ITs (Figs 1c, S2b). As rhizobial infection proceeds, MtN6 was expressed in the outer cortical cells underlying the infected root hair and dividing inner cortical cells (Figs 1d,e, S2c). During nodule development, MtN6 was consistently expressed in nodule primordia, young nodules and mature nodules (Fig 1e–g). In mature nodules with zonation, promoter GUS activity was mostly observed in the apical region of the nodule including the infection zone (Fig. 1g). Taken together, the expression of MtN6 is closely associated with rhizobial infection, both in root hairs and in nodules.

MtN6 is required for IT growth in M. truncatula

The expression pattern of N6 led us to further investigate its role in rhizobial infection. To this end, we isolated two M. truncatula Tnt1 insertional lines of MtN6 (NF16674 and NF16505) which were designated as Mtn6‐1 and Mtn6‐2, respectively (Fig. S3). Time‐course phenotyping showed that both Mtn6‐1 and Mtn6‐2 mutants had fewer nodules than WT (Fig. 2a). This phenotype was most obvious at a later stage of 30 dpi, and the nodule number difference between WT and Mtn6 mutant alleles was mainly due to significantly decreased pink nodules in the mutants (Fig. 2a,b).

Fig. 2.

Rhizobial infection and nodulation (Nod) phenotype of Mtn6. (a) Time course quantification of nodule numbers in wild‐type (WT), Mtn6‐1 and Mtn6‐2. The number of nodules for each genotype was counted at 10, 14, 21 and 30 d post inoculation (dpi) with Sinorhizobium meliloti 1021‐LacZ. n ≥ 15 for each genotype. Nod phenotyping was repeated three times. (b) Quantification of pink nodules and white nodules in WT, Mtn6‐1 and Mtn6‐2 at 30 dpi with S. meliloti 1021‐LacZ. n ≥ 15 for each genotype. Nod phenotyping was repeated three times. The data in (a) and (b) at 30 dpi are the same set of data. (c) Phenotypes of WT, Mtn6‐1 and Mtn6‐2 at 10 dpi with S. meliloti 1021‐LacZ. IT, infection thread; MC, microcolony; N, nodule; NP, nodule primordia. (d) Quantification of abnormal infection events in WT, Mtn6‐1 and Mtn6‐2 at 10 dpi with S. meliloti 1021‐LacZ. (e–k) Typical infection events in WT (e), Mtn6‐1 (f–h) and Mtn6‐2 (i–k). Red arrowheads indicate blocked ITs showing sac‐like structures, red arrows indicate branched ITs. n ≥ 15 for each genotype. Infection phenotyping was repeated two times. Two‐tailed Student's t‐test: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Error bars represent SEM. Bars, 25 μm.

We further quantified rhizobial infection events for WT and Mtn6 mutants at 10 dpi and the results revealed a significant decrease in the total number of infection events in both Mtn6 mutants, in comparison with WT (Fig. 2c). These decreased infection events included both infection chambers and ITs (Fig. 2c). In addition, there were more abnormal ITs in both Mtn6 mutants (Fig. 2d). In contrast to the smooth elongating ITs in WT (Fig. 2e), these defective infections in Mtn6 mutants manifested as enlarged microcolonies, blocked and/or branched ITs (Fig. 2f–k).

To examine the rhizobial infection process in Mtn6 nodules, we first made confocal laser scanning microscopy analysis using nodule sections with double staining of SYTO13 for rhizobia and Calcofluor White for cell wall (Fig. 3a–c). In some Mtn6 nodules, ITs with thicker walls were often observed, and there were more intercellular ‘blobs’ of rhizobial accumulation in both Mtn6 alleles (Figs 3b,c, S4). The more frequent intercellular blockage of rhizobial infection in Mtn6 nodules was further confirmed by paraffin sections and TEM imaging (Fig. 3d–f). Quantification analysis of TEM pictures further showed that the IT walls in Mtn6 mutants were significantly thicker than those in WT (Fig. S5). These results indicate that in addition to its role in IT development in root hairs, MtN6 also regulates rhizobial infection in nodules.

Fig. 3.

Defective rhizobial infection in Medicago truncatula Mtn6 nodules. (a–c) Abnormal infection threads (ITs) with thicker wall (white arrow) in Mtn6 mutants (b, c), compared with that in wild‐type (WT) (a) (white arrowhead). Magenta, Calcofluor white‐stained cell wall and IT wall. Green, SYTO13‐stained rhizobia. Number of nodules showing phenotypes as shown in the figures out of total number of nodules are 59/61, 17/68 and 18/60 for WT, Mtn6‐1 and Mtn6‐2, respectively. Nodules are from at least 30 plants for each genotype. (d, e) Paraffin nodule sections of WT and Mtn6 mutants. Zoomed‐in images are shown in (e) for the rectangle‐marked region in (d). Red arrowheads indicate ITs, and red arrows indicate intercellular ‘blobs’ (e). (f) Transmission electron microscopy (TEM) images of intercellular rhizobial accumulation (asterisks) in Mtn6 mutants. Paraffin sections were performed using at least 10 nodules from 10 plants in two independent experiments. TEM was performed using at least 10 nodules from 8 plants in two independent experiments. cw, cell wall; r, rhizobia; rrf, rhizobial release foci; s, symbiosome; IT, infection thread; v, vacuole. Bars, 20 μm (a–c), 200 μm (d), 40 μm (e), 5 μm (f).

We further examined whether MtN6 could promote rhizobial infection by overexpressing MtN6 using the LjUBQ1 promoter. The results showed that compared with control transgenic plants (EV), there were more ITs and nodules in pLjUBQ1:MtN6 transgenic roots (MtN6‐OE) (Fig. S6).

MtN6 is localized to IT

The subcellular localization of MtN6 was investigated in A17 roots expressing MtN6‐GFP fusion driven by the native MtN6 promoter (pMtN6: MtN6‐GFP). This construct could complement the rhizobial infection phenotypes in Mtn6‐1 and Mtn6‐2, suggesting that pMtN6: MtN6‐GFP was functional (Fig. S7). Under nonsymbiotic conditions, no GFP signal was found in root hairs or roots (Fig. S8a,b). Following rhizobial inoculation, MtN6‐GFP was first found at the tip of curled root hairs entrapping rhizobia (Fig. S8e), and later encircling the newly formed infection chamber (Figs 4a, S8c,d). In root hairs harboring ITs, MtN6‐GFP was present around the IT, especially at the growing tip region of the ITs (Fig. 4c). As a negative control, we could not observe comparable fluorescent signals like that of MtN6‐GFP in infected root hairs of nontransgenic roots (Fig. S9). A similar localization pattern was observed in both infection chambers and ITs in transgenic roots expressing MtN6‐GFP under the control of the infection‐specific VPY promoter (Fig. 4b,d).

Fig. 4.

Subcellular localization of MtN6 in infected root hairs in Medicago truncatula. MtN6‐GFP localization in infection chambers (ICs) (a, b) and infection threads (ITs) (c, d) within root hairs. MtN6‐GFP was driven by native MtN6 promoter (a, c) or VPY promoter (b, d). MtN6‐GFP localization in root hairs was observed in three independent experiments from 5 dpi with Sm1021‐CFP or Sm2011‐mCherry. White arrows indicate MtN6‐GFP localization at the IT tips. Each experiment examined at least 60 root hair infection events (including both ‘IC’ and ‘IT’ types) from 15 transgenic plants. Bars, 10 μm (a, b), 50 μm (c), 20 μm (d).

We next examined MtN6 localization in nodules. In the early developmental stages before the establishment of histological nodule zonation, MtN6‐GFP, driven by its native promoter, was broadly localized throughout the IT‐filled regions of young nodules (Fig. S10a). Later, MtN6‐GFP was located in the nodule infection zone and was prominent in the region containing ITs (Figs 5a, S10b). MtN6‐GFP was observed in the IT, with the strongest signal at the IT tip (Fig. 5b). It seemed that MtN6‐GFP localized not only in the IT cell wall but also in the matrix, especially at the tip region of ITs (Fig. S10c). MtN6‐GFP signal was also found in intracellular rhizobial release sites and sometimes in the apoplast (Figs 5c,d, S11). Furthermore, when rhizobia were released from the IT, scattered localization of N6‐GFP was also found around the released rhizobia (Figs 5d, S12).

Fig. 5.

Subcellular localization of MtN6 in Medicago truncatula nodules. (a) Localization of MtN6‐GFP in the infection zone of nodules. (b) Localization patterns of MtN6‐GFP during infection thread (IT) development in nodule cells. White arrowheads indicate the developing IT tip. (c) Localization of MtN6‐GFP in the apoplast (arrows) and in ITs. (d) Punctate localization of MtN6‐GFP in rhizobial release foci (rrf, arrowhead) and with newly released rhizobia (arrow). IT, infection thread (longitudinal sections in (b) and cross sections in (c)). The localization of MtN6‐GFP in nodules was observed in three independent experiments. For each experiment, more than 20 nodules from 15 transgenic plants were selected. The cells in (b), (c), (d) were from the distal, middle and proximal region of the infection zone, respectively. Bars, 200 μm (a), 20 μm (b–d).

MtN6 and GH9C2 colocalize during rhizobial infection

To verify that the localization of MtN6‐GFP was actually to the IT wall, we employed FM4‐64 dye to stain the IT membrane. This process revealed that MtN6‐GFP and FM4‐64 exhibited adjacent but nonoverlapping localization (Fig. 6a–c). This finding confirmed that MtN6 was not localized on the cell membrane but rather on the cell wall of the IT. The IT wall localization of MtN6 is reminiscent of GH9C2 localization during rhizobial infection (Zhao et al., ²⁰²⁵). Next, we performed a colocalization experiment to further compare their localization pattern.

Fig. 6.

Colocalization of MtN6 and GH9C2 during rhizobial infection in Medicago truncatula. (a–c) MtN6‐GFP localization imaged with FM4‐64 staining during rhizobial infection. Zoomed‐in view is shown in (b) for the red rectangular region indicated in (a). White arrows indicate the infection thread (IT) membrane labeled by FM4‐64, whereas white arrowheads indicate the localization of MtN6‐GFP. The white dashed line in the ‘Merge’ panel (a) indicates the signals for colocalization analysis of MtN6‐GFP with FM4‐64 in (c). Red arrows indicate FM4‐64 fluorescent signal peaks, and green arrows indicate MtN6‐GFP fluorescent signal peaks in (c). (d, e) Colocalization of MtN6‐GFP and GH9C2‐mCherry in infected root hairs. White arrowhead indicates the localization of MtN6‐GFP at the tip of a curling root hair entrapping rhizobia. (f) Colocalization of MtN6‐GFP and GH9C2‐mCherry in ITs in the nodule infection zone. White arrowheads indicate tips of the ITs. For FM4‐64 staining, colocalization analysis was performed from two independent experiments with at least 15 nodules from 10 transgenic plants in each experiment. Colocalization of MtN6‐GFP and GH9C2‐mCherry was performed in root hairs and nodules in two independent experiments, with at least 50 root hair infection events (including both ‘IC’ and ‘IT’ types) from 15 transgenic plants for each experiment. Bars, 10 μm (a, d, e), 5 μm (b), 20 μm (f).

Composite plants of M. truncatula were generated using a construct containing both pMtN6:MtN6‐GFP and pGH9C2:GH9C2‐mCherry. Live‐cell imaging of roots inoculated with S. meliloti revealed colocalization of MtN6‐GFP with GH9C2‐mCherry to the IT wall in root hairs (Figs 6d, S13). Although both GH9C2 and MtN6 localized to the infection chamber, during the process of IC formation only MtN6 was found at the tip of a curling root hair (Fig. 6e). In nodules, GH9C2 and MtN6 colocalized to transcellular ITs and IT branches (Fig. 6f).

MtN6 is a direct transcriptional target of NIN

Root hair transcriptome analysis suggests that MtN6 is under the control of NIN (Fig. 1b). To further investigate the transcriptional regulation of MtN6, we first examined MtN6 expression in nin‐1 mutants using qRT‐PCR. The results showed that S. meliloti‐induced MtN6 expression is abolished in the nin‐1 mutant at both 5 and 14 dpi, confirming its dependence on NIN (Fig. 7a).

Fig. 7.

NODULE INCEPTION (NIN) regulates MtN6 expression in Medicago truncatula. (a) Quantitative reverse transcription polymerase chain reaction analysis of MtN6 transcript levels in WT and nin‐1 at 5 or 14 d post inoculation (dpi) with Sinorhizobium meliloti 1021. Relative expression level is normalized to Medicago truncatula EF1α. Two‐tailed Student's t‐test: *, P < 0.05. Error bars represent SEM. (b) Diagram of the truncated MtN6 promoter region and assays of P1:GUS, P2:GUS and P3:GUS expression in transgenic M. truncatula roots. S1, 222 bp sequence spanning NBS. (c) Assays of P3:GUS and P3∆S1:GUS expression in transgenic M. truncatula roots. Number of roots showing typical GUS staining out of total number of roots is labeled as x/x in b, c. Hairy roots for GUS assays were collected at 5 dpi with S. meliloti 1021. Three independent experiments were repeated and one typical result is shown. (d) Dual‐luciferase (LUC) reporter assay shows the effect of NIN on MtN6 promoter activity. The pMtN6‐S1:LUC‐35S:REN reporter was co‐infiltrated with either the 35S:NIN effector (NIN) or the empty vector (EV) into Nicotiana benthamiana leaf epidermal cells. Co‐infiltration of 35S:REN with the empty vector or 35S:NIN was used as negative control for the system. Left panel, fluorescence imaging in N. benthamiana. Right panel, the ratio of firefly LUC to REN reflecting MtN6 promoter activity (n = 3). Two‐tailed Student's t‐test, *, P < 0.05. Error bars represent SEM. (e) Electrophoretic mobility‐shift assays (EMSA) analysis of NIN binding to the MtN6 promoter. Sequence alignment showing the putative NBS of MtN6 aligned with previously identified NBS in Lotus japonicus NIN (Nishida et al., ²⁰²¹). The image was created using the MEME suite. In the EMSA experiment, unlabeled DNA fragments (38 bp) were added at a 10‐fold (10 μM) and 100‐fold (10 μM) excess of labeled DNA fragments (38 bp, 1 μM) as binding competitors. M, mutated probe (Supporting Information Table S1). Bars, 1 cm.

Next, we performed truncation analysis of the MtN6 promoter by generating promoter fragments of 2195, 1443 and 700 bp upstream of the start codon of MtN6. These promoter fragments were subsequently fused with the GUS gene and used for promoter‐GUS expression experiments upon rhizobial inoculation (Fig. 7b). Histochemical GUS staining revealed no significant differences in staining intensity between fragments 2195 and 1443 bp (Fig. 7b). Notably, the 700‐bp fragment exhibited the weakest staining, indicating that critical cis‐regulatory elements may reside within the region between 1443 and 700 bp in the MtN6 promoter (Fig. 7b).

Based on the LjNIN‐specific binding sites (Nishida et al., ²⁰²¹), we used the MEME website (https://meme‐suite.org/meme/tools/meme) to identify potential NIN binding sites (NBS) in the MtN6 promoter. One potential NBS was found between 1146 and 1368 bp in the MtN6 promoter upstream of the start codon of MtN6 (Fig. 7b,e). We performed an NBS site analysis within the 2 kb promoter sequences of legume N6 homologs (Table S1) and found that the NBS sites of legume N6 (N6‐NBS) showed high similarity to reported NBSs and NREs (Fig. S14) (Soyano et al., ²⁰¹⁵; Jiang et al., ²⁰²¹). Additionally, the N6‐NBS exhibits the highest similarity with Nod‐associated cis‐regulatory elements 2 (NACE2) among the three NACE elements (Yu et al., ²⁰²⁴). We conducted another promoter‐GUS analysis using a 222‐bp (S1 sequence) deletion spanning the NBS from the 2195 bp full‐length promoter (P3∆S1:GUS) (Fig. 7c). GUS staining in hairy roots showed that the staining intensity in P3∆S1:GUS roots became much weaker compared with the 2195 bp full‐length MtN6 promoter GUS fusion (Fig. 7c). This finding was further validated by the dual‐LUC reporter assay showing that NIN could activate MtN6 promoter expression in N benthamiana leaves (Fig. 7d). We further introduced P3:GUS into WT and nin‐1 mutant roots, and the results showed that the expression of P3:GUS is much compromised in nin‐1 (Fig. S15). These results were consistent with the presence of NBS in the N6 promoter.

To determine whether NIN could bind to the predicted NBS, an EMSA was performed. A mobility shift was observed when the synthetic NBS oligonucleotides of the MtN6 promoter were incubated with the carboxyl‐terminal portion (566–933 aa) of the recombinant NIN protein, and this binding was attenuated by the addition of a competitive probe, whereas the mutated competitive probe failed to produce the same effect (Fig. 7e). These results demonstrate that NIN positively regulates MtN6 expression by directly binding to an NBS in the MtN6 promoter.

N6 is a new gene evolved in legumes, likely recruited for root hair infection in root nodule symbiosis

A phylogenetic analysis was conducted to investigate the evolutionary relationships between N6 and nodulin/glutamine synthetase–like protein (NodGS) homologs across plant lineages. NodGS proteins are characterized by a C‐terminal glutamine synthetase–like domain fused to an N‐terminal nodulin‐like domain (Fig. S16). NodGS homologs are widely distributed throughout the plant kingdom. However, their biological function remains largely unresolved, with functional evidence limited to Arabidopsis thaliana, where NodGS has been implicated in root morphogenesis and microbial elicitation (Doskočilová et al., ²⁰¹¹). In contrast to the broad distribution of NodGS, N6 is restricted to the legume family (Fig. S17; Table S2). We analyzed legume species which have genomes sequenced and annotated. Among these legumes, 13 species, including Cicer arietinum, Glycine max, Lens culinaris, L. ervoides, L. japonicus, Lupinus albus, L. angustifolius, M. truncatula, Phaseolus vulgaris, Pisum sativum, T. pratense, Vicia faba and V. unguiculata, possess both N6 and NodGS homologs. By contrast, several early diverging legume lineages, including Chamaecrista fasciculata and Faidherbia albida (Caesalpinioideae) as well as Arachis spp. and A. evenia (Papilionoideae, tribe Dalbergieae) have only NodGS homologs and no N6 homologs (Fig. 8a). These basal legume lineages diverged c. 55–60 million years ago, earlier than the diversification of the 13 N6‐containing species (Zhao et al., ²⁰²¹). Cercis canadensis, belonging to the Cercidoideae subfamily (c. 34 million years ago), which does not form nitrogen‐fixing nodules, also harbors only a single NodGS gene but without N6 homologs. These observations indicate that N6 likely arose after the divergence of basal legume lineages and became associated with later‐diverging, nodulating legumes (Fig. S18). Notably, from the above N6 containing legume species, all but Lupinus exploit IT‐mediated rhizobial entry, indicating a likely functional association between N6 and this infection mode. For Lupinus, although rare short IT‐like structures are formed in nodules, rhizobia enter intercellularly at the junction of the root hair base and neighboring epidermal cells (Tang et al., ^{1992, 1993}; González‐Sama et al., ²⁰⁰⁴; Zhao et al., ²⁰²¹). Interestingly, L. angustifolius retains an intact N6, whereas L. albus contains only four fragmented remnants of N6, including two N‐terminal and two C‐terminal fragments that can pair to form intact N6 (Fig. S19). This lineage‐specific pattern suggests that N6 may be undergoing degeneration or functional divergence in Lupinus.

Fig. 8.

Phylogenetic and microsynteny analysis of N6 across representative legume species. (a) Phylogenetic tree of legume N6s. The phylogenetic tree was made using NGphylogeny.fr and was viewed by Interactive Tree of Life (iTOL). A total of 39 N6 homologs from 20 legume species were used. N6 homologs are highlighted in blue, NodGS are highlighted in purple, and N6 homologs in Lupinus spp. are in orange. MtN6 homologs were retrieved from Phytozome 13 (https://phytozome.next.jgi.doe.gov/blast‐search) by Blast of proteome data using Medicago truncatula N6 sequence. (b) Syntenic relationships of the N6 region among several legume species. A collinear block encompassing the 10 genes upstream and five genes downstream of MtN6 is shown. The gray rectangle represents a gene on chromosome, N6 homologs are in blue, and NodGS homologs are in purple. Gray lines represent conserved syntenic gene pairs.

To further investigate the evolutionary origin of N6, we performed synteny analyses of genomic regions surrounding NodGS and N6. The results revealed a high degree of collinearity among NodGS‐containing regions across the examined species, with N6 embedded within conserved syntenic blocks (Fig. 8b). Notably, N6 is conserved only in IT‐forming legumes and is absent from crack‐entry legumes, such as Arachis spp. and Aeschynomene spp. (Fig. 8b). Comparative analysis of gene structures further demonstrated high exon identity (77%–100%) between N6 and NodGS genes (Fig. S20), highly supporting that N6 originated via duplication of an ancestral NodGS gene. Collectively, these data indicate that N6 evolved within legume lineages predominantly employing IT‐mediated rhizobial entry, suggesting that it is a recent symbiotic adaptation.

Discussion

In many legume rhizobial symbioses, the IT serves as a critical pathway for the delivery of rhizobia into host cells. Years of genetic study has discovered many host genes regulating IT formation and development, but the underlying cellular mechanisms are still unclear (de Carvalho‐Niebel et al., ²⁰²⁴; Gao et al., ²⁰²⁴; Zhang & Ott, ²⁰²⁴; Zhang et al., ^2024b). High‐resolution transcriptomic approaches include single cell type or single cells using rhizobial‐infected or Nod factor–inoculated roots, root hairs and nodules have also revealed many potential regulators for rhizobial infection in different legumes (Libault et al., ²⁰¹⁰; Breakspear et al., ²⁰¹⁴; Roux et al., ²⁰¹⁴; Jardinaud et al., ²⁰¹⁶; Cervantes‐Pérez et al., ^{2022, 2024}; L. Wang et al., ²⁰²²; Ye et al., ^{2022, 2024}; Frank et al., ²⁰²³; Sun et al., ²⁰²³; Liu et al., ^2023a,b; Pereira et al., ²⁰²⁴). However, functional studies on the role of these genes in the intricate cytological process of IT development are quite limited. Further investigations are also needed to elucidate the evolutionary origins of key players and their integration mechanisms within the symbiotic network. In this study, we find that M. truncatula N6 encoding an IT‐wall protein has recently evolved in certain legumes and has been recruited by NIN for IT development in intracellular root hair rhizobial infection.

Medicago truncatula n6 mutants have decreased rhizobial infection events and show typical defective rhizobial infection phenotypes at IT initiation and elongation stages, such as enlarged microcolonies and blocked ITs, suggesting a role for MtN6 in IT polar growth (Fig. 2). Besides rhizobial infection defects in root hairs, abnormal ITs with thicker walls and increased intercellular rhizobia accumulation are often found in nodules (Fig. 3). This intercellular rhizobia accumulation in nodules was also present in other mutants defective in rhizobial infection such as vpy and gh9c2 (Liu et al., ^2019b; Zhao et al., ²⁰²⁵), and we speculate that when the intracellular infection pathway is obstructed, plants may activate alternative infection strategies (Karas et al., ²⁰⁰⁵).

The abnormal IT phenotype of Mtn6 mutants is in line with MtN6 localization in IT wall. Similar to GH9C2 and NPL, N6 also localizes to the wall of the whole IT but shows more focused localization in the very IT tip region, where N6, GH9C2 and NPL may function together to secure the cell wall remodeling tailored for IT polar growth (Liu et al., ^2019b; Su et al., ²⁰²³; Zhao et al., ²⁰²⁵). Although all these cell wall proteins are also secreted into the IC, only MtN6 is found at the root hair tip during the root hair curling process leading to the formation of IC, implying a distinct requirement for MtN6 at this early stage. This tip localization of MtN6 is reminiscent of VAMP721, suggesting a possible relevant secretory pathway (Fournier et al., ²⁰¹⁵). N6 encodes a putative amidohydrolase protein, which may modify cell wall component per se for IC formation and IT development. Another possibility is that N6 may generate a molecule or certain metabolite which exerts signaling or nutritional functions for rhizobia to propagate and move in the IT (Guo et al., ²⁰²²; Chen et al., ²⁰²³). Further explorations are needed to elucidate the biochemical activity of N6 to test above hypotheses.

It is thought that formation of new genes during evolution is an important driver for functional innovation and newly evolved genes have a fundamental role in phenotypic complexity (Innan & Kondrashov, 2010; Chen et al., ²⁰¹³; Xia et al., ²⁰²⁵). Through phylogenetic analysis, we find that N6 is a new gene, which is not universally present across all plant species but is predominantly distributed among legume lineages that adopt root hair infection. This distribution pattern strongly suggests that its emergence may be closely related to the evolution of the key symbiotic strategy known as root hair infection in certain legumes. N6 is homologous to the N terminal domain of NodGS, which is a bi‐domain protein present in plants and fungi (Doskočilová et al., ²⁰¹¹; Guo et al., ²⁰²²; Chen et al., ²⁰²³). NodGS may have its C‐terminal GS domain and N‐terminal domain functioning as independent functional modules. For instance, tea NodGS homolog CsTSI has been shown to possess glutamine synthetase activity in its C‐terminal GS domain (She et al., ²⁰²²). It was also shown that in A. nidulans, the C terminus of fluG (NodGS homolog) is required for asexual development and the N terminus of fluG contributes to conidial production (Iradi‐Serrano et al., ²⁰¹⁹). Our microsynteny analysis suggests that N6 was likely generated by duplication from the NodGS gene. This duplication might occur after the divergence of Dalbergioids, for example Arachis and Aeschynomene, and before the divergence of Lupinus during the evolution of legumes (Zhao et al., ²⁰²¹). What's more, N6 acquired a secretory signal peptide at the N‐terminus which is not present in NodGS, and this critical adaptive change drove the functional transition from intracellular metabolism to localization at the symbiotic interface of the IT wall. We propose an evolutionary scenario that in legumes where intracellular rhizobial infection mode is utilized, N6 was recruited by NIN into the genetic network regulating IT development, possibly to specifically optimize intracellular infection efficiency via the root hair pathway. Supporting this hypothesis is that in Lupinus species, where intercellular rhizobial infection strategy is used, L. angustifolius N6‐like protein exhibits lower sequence conservation than other legume N6 homologs, and L. albus N6 homologs are fragmented (Fig. S19). In addition, NIN is found to be inessential for epidermal intercellular rhizobial infection in A. hypogaea (Bhattacharjee et al., ²⁰²²).

Our finding provides new molecular insights into the mechanisms underlying NIN regulation of rhizobial infection and reveals an evolutionary strategy of leguminous plants in optimizing their symbiotic capacity by recruiting new genetic elements. More examples are expected in future to show how new genes were formed and integrated into the existing signaling network to drive phenotypic evolution in legume‐rhizobial symbiosis.

Competing interests

None declared.

Author contributions

RW and C‐WL designed the research. RW performed most of the experiments with help from YC and H‐QW for sectioning. WW did RT‐PCR experiments and made phylogenetic and evolutionary analyses. LZ made the constructs for EMSA. RW, WW, H‐QZ, Z‐BL, J‐LD, LZ, C‐YJ and C‐WL collected and analyzed the data. WW assembled all the figures. RW wrote the first draft. RW, WW and C‐WL revised the manuscript with input from all authors. RW and WW contributed equally to this work.

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Supporting information

Fig S1 Induction of MtN6 by Nod factor in Medicago truncatula root epidermis.

Fig. S2 Expression pattern of MtN6 in roots.

Fig. S3 Characterization of the Mtn6 mutants.

Fig. S4 Intercellular blockage of rhizobial infection in Mtn6‐1 and Mtn6‐2 nodules.

Fig. S5 The infection thread walls are affected in Mtn6 mutants.

Fig. S6 Overexpression of MtN6 promotes rhizobial infection.

Fig. S7 Complementation of rhizobial infection phenotypes in Mtn6 mutants by pMtN6:MtN6‐GFP.

Fig. S8 Subcellular localization of MtN6‐GFP in root hairs.

Fig. S9 Imaging of a nontransgenic infected root hair.

Fig. S10 Subcellular localization of MtN6‐GFP in nodules.

Fig. S11 MtN6‐GFP localization in the apoplast.

Fig. S12 Colocalization analysis of MtN6‐GFP with rhizobia.

Fig. S13 Colocalization analysis of MtN6‐GFP and GH9C2‐mCherry in infected root hairs.

Fig. S14 Comparison of N6‐NBS with known NRE and NBS motifs.

Fig. S15 Assays of P3:GUS expression in transgenic hairy roots of WT and nin‐1.

Fig. S16 Protein domain structures of N6 and NodGS.

Fig. S17 Phylogenetic analysis of MtN6 and NodGS.

Fig. S18 Gene age analysis of N6.

Fig. S19 Amino acid alignment of legume N6 and Lupinus N6‐like proteins.

Fig. S20 Comparison of the gene structures of MtN6 and two MtNodGSs.

Notes S1 Protein sequences of N6 and NodGS for phylogenetic analysis.

Table S1 List of primers used in this study and sequences for N6‐NBS motif discovery.

Table S2 Species and protein information used for phylogenetic and microsynteny analyses.

Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Data availability

All data generated in this study are included in the main text and Figs S1–S20, Tables S1, S2 and Notes S1 of this article.