Plant lines and growth conditions
The Lotus japonicus Gifu ecotype was used as the wild type. All plants were grown at 21 °C under 16 h light/8 h dark conditions. For germination, Lotus seeds were scarified with sandpaper and surface sterilized with 1% sodium hypochlorite for 10 min. Seedlings were washed with sterile water for 5 times and germinated on wet filter paper (AGF 651; Frisenette ApS) in sterile square Petri dishes at 21 °C for 2 days. Then, seedlings were transferred into the substrate mixture (leca:vermiculite=3:1). Three weeks post-inoculation, plants were treated with 10 mM KNO3 or KCl for 14 days (or as indicated). Subsequently, nodule number, nitrogenase activity (ARA), or leghaemoglobin content were recorded. For Zn treatment, 3 weeks post-inoculation, plants were watered with 500 µM MgCl2 (mock) or 500 µM ZnCl2 for 3 days followed by 10 days 10 mM KNO3 treatments. ARA and leghaemoglobin content were recorded. LORE1 insertion mutants were ordered through LotusBase (https://lotus.au.dk) and homozygotes were isolated for phenotyping and generation of higher order mutants as described43. Line numbers and genotyping primers are provided in Extended Data Fig. 2a. Mesorhizobium loti NZP2235 was used for nodulation assays.
Mutant screening and sequence analysis
A LORE1-mutant pool, in which there are random LORE1 insertions in the genome of each individual, were germinated in substrate mixture (leca:vermiculite 3:1) and inoculated with M.loti NZP2235. Four weeks post-inoculation, plants were watered with 10 mM KNO3 for three weeks. Most nodules became green or black, and we isolated plants with pink nodules for rescreening in subsequent generations. DNA from mutant plants was isolated and LORE1-flanking sequences sequenced to identify LORE1 insertion positions as previously described19. FUN protein sequences were identified by BLAST and SHOOT44 and aligned with MAFFT 7.490 and a tree constructed using FastTree 2.1.11. The tree was visualized using iTOL 6.7.345.
Hairy root transformation
For complementation assays, the Lotus ubiquitin promoter, FUN coding sequence, and 35S terminator were cloned into the pIV10 expression vector46. To study the expression pattern of FUN, native FUN promoter, glucuronidase (GUS) and the native FUN terminator sequence (tFUN) were cloned into the pIV10 expression vector. Constructs mentioned above were transformed into Agrobacterium rhizogenes AR1193. These agrobacteria were used to transform the hypocotyl of 6-day old seedlings. After three weeks, non-transformed roots were removed, and seedlings were transferred into the substrate mixture mentioned above or onto 0.25× Broughton and Dilworth 1971 medium plates. Subsequently, plants were inoculated with rhizobia and watered with nitrate as described above.
Acetylene reduction assay
ARAs were conducted essentially as described47. The nodulated root from single plants was placed in a 5 ml glass gas chromatography vial. A syringe was used to replace 500 µl air in the vial with 2% acetylene. Samples were incubated at room temperature for 30 min before ethylene quantification using a SensorSense (Nijmegen, NL) ETD-300 ethylene detector operating in sample mode with 2.5 l h−1 flow rate and 6 min detection time. The curve was integrated using the SensorSense valve controller software to calculate the total ethylene production per sample.
Leghaemoglobin content measurement
Leghaemoglobin content measurements were conducted using a spectrophotometric method as described previously41. Fresh nodules from each individual plant were first ground and homogenized in 16-fold volumes of 0.1 M precooled PBS (Na2HPO4:NaH2PO4 buffer at 5 °C, pH 6.8). The resulting slurry was then centrifuged at 12,000g for 15 min prior to assaying the supernatant by spectrophotometry at a wavelength of 540, 520 and 560 nm. The Leghaemoglobin content was calculated from a standard curve using bovine haemoglobin as a protein standard.
GUS staining
Three weeks post-inoculation, hairy roots were put into GUS staining buffer, which contains 0.5 mg ml−1 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid, 100 mM potassium phosphate buffer (pH 7.0), 10 mM EDTA (pH 8.0), 1 mM potassium ferricyanide, 1 mM potassium ferrocyanide and 0.1% Triton X-100. The roots were incubated at 37 °C overnight. Roots were washed with 70% ethanol twice before image acquisition. Quantitative GUS assays are described below for the trans-activation assays.
Gene expression
For RNA-seq, 3 weeks post-inoculation, plants were acclimatized prior to treatment by submerging in 0.25× Long Ashton liquid medium overnight, then treated with 0 or 10 mM KNO3 for 24 h. Mature nodules were collected. mRNA was isolated using the NucleoSpin RNA Plant kit (Macherey-Nagel) and RNA-seq (PE-150 bp Illumina sequencing) was conducted by Novogene. RNA-seq analysis was performed by mapping reads to the reference transcriptome using Salmon48 and quantification performed using DEseq249. A publicly available timeseries of nitrate-treated nodules17 was obtained from GEO using accession number GSE197362. GO enrichment was performed using GO_MWU with GO terms obtained from https://lotus.au.dk.
For the expression of target genes, RevertAid Reverse Transcriptase (Thermo) was used for the synthesis of first strand cDNA. LightCycler480 instrument and LightCycler480 SYBR Green I master (Roche Diagnostics) were used for quantitative PCR with reverse transcription. Ubiquitin-conjugating enzyme was used as a reference. The cDNA concentration of target genes was calculated using amplicon PCR efficiency calculations using LinRegPCR50. Target genes were compared to the reference for each of 5 biological repetitions (each consisting of 8 to 10 nodules). At least two technical repetitions were performed in each analysis. Primers used are listed in Extended Data Fig. 4b.
Electrophoretic mobility shift assay
The DNA probes with 6-FAM-label at the 5′ end were synthesized by Eurofins and are listed in Extended Data Fig. 5h. We incubated the purified FUN DNA-binding domain (residues 178–237) with the probes at 37 °C for 60 min in EMSA buffer (25 mM Tris-HCl pH 8.0, 80 mM NaCl, 35 mM KCl, 5 mM MgCl2). After incubation, the reaction mixture was electrophoresed in 6% native polyacrylamide gel and then labelled DNA was detected with the Typhoon scanner (Fujifilm). Probes without 6-FAM-label served as competitors, while probes with mutation in the core binding sites (TGACG) served as mutants.
Transient activation assay
Promoters of FUN candidate target genes (NRT2.1, HO1, NRT3.1 and AS1), the glucuronidase (GUS) coding sequence and 35S terminator were cloned into compatible Golden Gate vectors as reporters; while the 35S promoter, FUN coding sequence, eGFP and 35S terminator were cloned as the effector. The reporters and effector were cloned into the p50507 Golden Gate binary vector. These constructs were then transformed into Agrobacterium tumefaciens strain AGL1. These A. tumefaciens were diluted to OD600 = 0.2 and were infiltrated into N. benthamiana leaves. Three days after infiltration, samples of about 20 mg were collected for protein extraction. GUS activities were measured with 4-methylumbelliferyl-β-d-glucuronide as substrate (Sigma-Aldrich) using a Thermo Scientific Varioskan flash. For Zn treatment, 2 days after A. tumefaciens infiltration, N. benthamiana leaves were infiltrated with 500 µM MgCl2 (mock), 500 µM ZnCl2, or 2.5 mM EDTA. GUS activities were measured 1 day after treatments.
Protein production and purification
The FUN sensor domain (residues 244–480) with a 3C-cleavable N-terminal tag consisting of 10 histidines, 7 arginines and a SUMO tag was obtained from GenScript together with a construct of the FUN sensor with the zipper domain (residues 178–480) N-terminally tagged with 7 histidines and a GB1 tag. The plasmids were transformed into Escherichia coli LOBSTR cells51. The expression culture was grown to OD600 = 0.6 in LB medium with 0.1 mg ml−1 ampicillin and 0.034 mg ml−1 chloramphenicol at 37oC and 110 rpm. Cells were cold shocked on ice for 30 min before expression was induced with 0.4 mM IPTG at 18oC overnight. The cells were pelleted (4,400g, 4 °C, 10 min), resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 10% glycerol, 10 mM imidazole, 5 mM β-mercaptoethanol and 1 mM benzamidine) and lysed by sonication. The lysate was cleared by centrifugation (30,600g, 4 °C, 30 min), and the proteins were purified from the cleared lysate using a Protino Ni-NTA 5 ml column (Machery-Nagel). The protein was eluted with a high-imidazole buffer (50 mM Tris-Hcl pH 8.0, 250 mM NaCl, 5% glycerol, 500 mM imidazole, 5 mM β-mercaptoethanol). The FUN sensor with zipper was not purified further, while the FUN sensor was dialysed overnight against 50 mM Tris-HCl pH 8.0, 250 mM NaCl, 5% glycerol, 5 mM β-mercaptoethanol with 3C protease in a 1:50 molar ratio. The cleaved tag and the protease were subsequently removed by a second Ni-IMAC step. The FUN sensor was further purified by size-exclusion chromatography on a Superdex 200 Increase 10/300 GL (GE Healthcare) in minimal buffer (10 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM β-mercaptoethanol). For SAXS analysis, the FUN sensor was further purified on a ResourceQ 1 ml (GE Healthcare) and eluted with a linear gradient of 10–500 mM NaCl and 10 mM Tris-HCl pH 8.0 and 5 mM β-mercaptoethanol. Eluted fractions were pooled and dialysed against minimal buffer.
DLS and nanoDSF
The FUN protein was analysed on a Prometheus Panta instrument (NanoTemper Technologies) for alterations in thermal unfolding (nanoDSF) and size (DLS) upon addition of ligands. 0.8 mg ml−1 of the purified protein was incubated with 4 mM of different potential ligands or a 0–4 mM ZnCl2 series for 20 min whereupon 5 mM EDTA was added to samples analysed for reversible filamentation. Before addition, ZnCl2 was filtered using VivaSpin MWCO 5 kDa and immediately added to the protein samples. 10 consecutive DLS measurements were performed for each sample at 25 °C with 100% laser power and followed by a nanoDSF experiment measured at a temperature slope of 1 °C/min from 25–90 °C with 100% excitation power. All measurements were performed in triplicates.
SAXS
SAXS measurements were performed at the in-house NanoSTAR instrument at Aarhus University52,53 (Bruker AXS). The instrument uses a Cu rotating anode, has a scatterless pinhole in front of the sample47 and employs a two-dimensional position-sensitive gas detector (Vantec 500, Bruker AXS). The samples and buffer were measured in a homebuilt flow-through capillary. The intensity I(q) is displayed as a function of the modulus of the scattering vector. The buffer scattering was subtracted from the scattering from the samples and the intensities were converted to an absolute scale and corrected for variations in detector efficiency by normalizing to the scattering of pure water46. The data were plotted in Guinier of ln(I(q)) versus q2 to determine the radius of gyration Rg, and an indirect Fourier transformation54,55 was performed to obtain the pair distance-distribution function p(r), which is a histogram of distances between pair of points within the particles weighted by the excess scattering length density at the points. Note that the resolution of the SAXS data is about 400 Å and therefore the overall length of the fibrils induced by zinc is not resolved. The p(r) function is in this case related to the cross-section structure of the filaments.
Negative-stain electron microscopy
For electron microscopy, 0.1 mg ml−1 of the purified FUN sensor domain was incubated 20 min at room temperature with or without 100 µM ZnCl2 and with or without 5 mM EDTA. Samples for negative staining were prepared on 400 copper mesh grids that were manually covered with a collodion support film coated with carbon using a Leica EM SCD 500 High Vacuum Sputter Coater. Before staining, the grids were glow discharged with negative polarity, 25 mA for 45 s, using a PELCO easiGlow glow discharge system. 3 µl of the FUN sensor was deposited on the grid, incubated 30 s, and excess sample was removed from the grid using Whatman paper. After the blotting, the grid was floated 3 times on 2% uranyl formate solution for 15 s and then dried. Negative-staining micrographs were recorded using a Tecnai G2 Spirit microscope operating at 120 kV, equipped with a TemCam-F416 (4kx4k) TVIPS CMOS camera and a Veleta (2kx2k) CCD camera, at EMBION the Danish national cryo-EM facility in Aarhus, Denmark. Micrographs were recorded at a magnification of 42,000× and 52,000×.
Microscopy and confocal imaging
For the FUN expression pattern, the roots after GUS staining were observed by Leica M165FC Fluorescence stereomicroscope. Nodules were embedded in 3% agarose and sectioned in 100-µm slices using a vibratome. Nodule slices were observed by Zeiss Axioplan 2 light microscope. For FUN subcellular locations, Lotus hairy roots and N. benthamiana leaves expressing FUN–GFP were treated with 500 µM ZnCl2 (Zn) or MgCl2 (mock) for 3 days, and fluorescence were observed using a 491–535 nm filter on a Zeiss LSM 710 confocal microscope.
Zinpyr-1 imaging and quantification
Plants with pink nodules (3 weeks post-inoculation) were acclimatized prior to treatment by submerging in 0.25× Long Ashton liquid medium overnight, then treated with 0 or 10 mM KNO3 for 24 h. Mature nodules were embedded in 3% agarose and sectioned in 80-µm slices using a vibratome. Slides were stained with 5 µM Zinpyr-1 for 3 h and rinsed 3 times with water. Fluorescence was observed by Zeiss LSM 710 confocal microscope, using excitation at 488 nm and emission from 505 to 550 nm. Fluorescence densities were quantified by ImageJ.
Micro-XRF
XRF images were acquired at the ID21 beamline of the European Synchrotron Radiation Facility56. The scanning X-ray microscope at ID21 is equipped with a liquid nitrogen passively cooled cryogenic stage. Samples were prepared as described57. In brief, nodules were embedded in OCT medium and cryo-fixed by plunging them into liquid nitrogen-chilled isopentane. 20 mm sections of frozen samples were obtained using a Leica LN22 cryo-microtome and mounted in a liquid nitrogen-cooled sample holder between two Ultralene (Spex SamplePrep) foils. The beam was focused to 0.9 × 0.6 mm2 using Kirkpatrick–Baez mirror optics. The emitted fluorescence signal was detected with an energy-dispersive, large area (80 mm2) SDD detector equipped with a beryllium window (XFlash SGX, RaySpec). Images were acquired at a fixed energy of 9.8 keV by raster-scanning the sample with a step of 2 × 2 mm2 and a 220 ms dwell time. Elemental distribution was calculated with the PyMca software package58.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.