Plant materials
For the experiments, Arabidopsis thaliana ecotype Columbia-0 (Col-0) referred to as WT and salt overly sensitive 1 (sos1) mutant in the Col-0 gl–1 background were used; the sos1-1 allele was previously identified in an ethyl methane sulfonate screen17. The following published constructs were introduced either by transformation or crossing into sos1: SOS1-GFP (ref. 26); VHA-A1-RFP (AT2G28520: VACUOLAR PROTON ATPASE A1); mCherry-RABF2A (AT5G45130: RHA1 or ARABIDOPSIS RAB HOMOLOG F2A); mCherry-RABF2B (AT4G19640: ARA7 or ARABIDOPSIS RAB GTPASE HOMOLOG F2B); mCherry-VAMP711 (AT4G32150: VESICLE-ASSOCIATED MEMBRANE PROTEIN 711) (refs. 25,36). Rice (Oryza sativa) cv. Nipponbare (WT) and the mutant Ossos1 are lines previously described22.
Hydroponics plant growth
Seedlings grown under standard conditions on standard half-strength Murashige–Skoog (MS) media agar plates were exposed to low and variable levels of Na, which resulted in a high Na-background in the CryoNanoSIMS images. To ensure a controlled starting point for subsequent salt-stress experiments, a growth medium free of Na was required. This was achieved with a hydroponics system and using a medium formulation free of sodium (referred throughout the manuscript as Na-free medium) to allow fully controlled salt-stress assays for growth experiments.
Composition of liquid MS and Na-free medium for hydroponics
The Na-free medium was composed of MS basal salt macronutrient solution (formulation number M0654; Sigma-Aldrich) as the macroelement source. Each microelement solution was prepared separately and all solutions were prepared with MilliQ water. Essential micronutrients, free from Na-containing compounds, were used at the following concentrations: 50 µM boric acid (H3BO3), 0.05 µM copper sulfate pentahydrate (CuSO4.5H2O), 0.01 µM molybdic acid ammonium salt heptahydrate ((NH4)6 Mo7 O24.4H2O), 2.5 µM potassium iodide (KI), 0.05 µM cobalt(II) chloride hexahydrate (CoCl2.6H2O), 50 µM Fe(III)-EDTA (C10H12FeN2O8) (Chemie Brunschwig AG, CH), 50 µM manganese(II) sulfate monohydrate (MnSO4.2H20) and 50 µM zinc sulfate heptahydrate (ZnSO4.7H2O). 2.5 mM MES (2-morpholinoethanesulfonic acid) buffer was added, the pH was adjusted to 5.7 with KOH and the solution was sterilized. For comparison with a liquid half-strength MS medium, the seedlings were grown on a sterile liquid half-strength MS solution, pH adjusted to 5.7 with KOH.
Arabidopsis growth
Sterilized Arabidopsis seeds were sown on a wetted perforated disc placed in a sterile 12-well culture plate and stratified in the dark for 2 days. The pore size of the disc was approximately 120 µm, which prevented the seeds from falling through, but was sufficient for the roots to grow through. Post stratification, the wells were carefully topped with 4 ml of Na-free medium which enabled the discs with the seeds to float. The set-up was transferred to growth chambers at 21 °C with constant light and grown for 5 days. For NaCl treatments, stock solutions prepared in MilliQ water were used and the seedlings were directly treated with 2.5 mM, 25 mM and 100 mM NaCl for the indicated times in the hydroponics system under sterile condition and subsequently sampled for analysis with the cryogenic workflow described below (Fig. 1a). Root growth in the hydroponics system was compared with standard half-strength MS growth medium and growth on agar plates (Extended Data Fig. 1a). The concentrations of Na for growth stress were tested on agar plates and treatment conditions selected for hydroponics treatment (Extended Data Fig. 1d).
Rice growth
Sterilized rice seeds were germinated on wet filter paper for 3 days in the dark and then transferred to a hydroponics system on perforated sterile filter paper with the cotyledon in contact with the same Na-free medium described above. Growth was continued in a hydroponics settings in the growth chamber set to 28 °C with constant light and grown for 5 days. Salt treatments were given by adding NaCl directly to the hydroponic set-up and then sampled for high-pressure freezing (see below) at the set time point.
CryoSEM–CryoNanoSIMS workflow
A general description of the CryoNanoSIMS workflow can be found in ref. 1. Here we describe a specific workflow optimized for plants that focusing on Arabidopsis root meristems, as illustrated in Fig. 1 and Extended Data Fig. 1.
Sample preparation and high-pressure freezing
This process is schematically summarized in Fig. 1a. For high-pressure freezing, the seedlings were gently transferred to a small petri dish with the media. The roots (radius approximately 120 µm) were cut approximately 3 mm from the tip and immediately placed in a 3-mm-diameter gold coated Type A Cu-carrier with a cavity of 0.1 mm (Article 1322, Wohlwend GmbH, Sennwald, Switzerland) filled with the cryoprotectant Dextran 40 (25% w/v in 50 mM MES, pH 5.7) and then covered with the flat side of an aluminium specimen carrier type B (Article 1313, Wohlwend GmbH, Sennwald, Switzerland) and directly high-pressure frozen using an EM ICE (Leica) device. From this point on in the workflow, the samples were maintained and handled under strictly controlled cryo-conditions.
Cryo-planing of vitrified sample block face
The Cu-carriers with the root samples were loaded into a sample holder inside a Leica instrument EM VCM with a stable cryo-environment (liquid N2 bath) and transferred to a Leica cryo-chamber EM FC7 on a Leica EM UC7 Ultramicrotome with the chamber and the sample holder maintained at −110 °C. A flat and smooth surface, which is essential for CryoNanoSIMS imaging, was created with a trimming diamond knife Cryo TRIM20 (Diatome, Biel), generating a surface block face containing a section of the root sample. A diamond cryo TRIM20 Diatome knife was used to further smooth the surface of the sample.
From this point on, a Leica EM VCT500 vacuum cryo-transfer (VCT) shuttle, holding the sample at about −150 °C in a vacuum of 8 × 10−4 mbar, was used to transfer the sample between different instruments.
Freeze etching
The samples were transferred to a Leica cryo-eBeam metal coater EM ACE600 under cryogenic conditions, pumped to a vacuum of 9 × 10−7 mbar. Samples were sublimated with a temperature cycle in which they were ramped up from −150 °C to −94 °C over 20 min, held at −94 °C for 90 s and ramped down to −150 °C over 20 min with a ramp rate of 3 °C min−1. These settings were empirically optimized to remove surface ice contamination and gently bring out ultrastructural detail of the root sections, which could be imaged in the cryoSEM.
Metal coating and cryoSEM imaging
Before cryoSEM imaging, the sample surface was coated with an approximately 3 nm layer of PtC using Leica cryo-eBeam metal coater (EM ACE600). The roots were then imaged using a Zeiss Gemini 500 SEM with a Leica cryostage operated at −140 °C with the vacuum at 2.7 × 10−8 mbar. Images were acquired with the Inlens detector and an accelerating voltage of 1.7 kV with a working distance of 3.5 mm and an aperture size of 10 µm. High-resolution images of 4,098 × 3,072 pixels were obtained both for overviews of the root samples and at higher magnification from areas of interest at the surface of the sample. These cryoSEM images served to select and correlate regions of the root sections with CryoNanoSIMS imaging.
Second metal coating
The samples were returned with the VCT under vacuum from the cryoSEM to the Leica cryo-eBeam metal coater EM ACE600. An additional approximately 17 nm layer of PtC was deposited on the sample surface and maintained at −150 °C. This was necessary to increase the conductivity and dissipate surface charging during CryoNanoSIMS analysis. The sample was then transferred with the VCT to the CryoNanoSIMS for imaging.
CryoNanoSIMS imaging
Elemental mapping was performed with the new CryoNanoSIMS instrument within the Laboratory for Biological Geochemistry at EPFL1. Images were systematically acquired on the root meristem with epidermal cells in the region where the cells transition from division to elongation (Extended Data Figs. 1c and 4a). Sample surface areas with ice or particle contamination (visible in cryoSEM images) were avoided. Areas selected for imaging were presputtered for 5 min with a defocused 16 keV O− ion beam to a total dose of about 1.5 × 1016 O− ions per cm2, to remove the metal coating. The sample surface was then bombarded with a primary ion beam of about 20 pA O− ions focused to a spot size of about 400 nm. Images of 45 × 45 μm2 with 256 × 256 pixels were obtained by rastering this primary beam across the sample surface with a pixel dwell-time of 5 ms. Secondary ions 23Na+, 24Mg+, 32S+, 39K+, 40Ca+, 55Mn+ and 56Fe+ were simultaneously collected in electron multipliers at a mass resolution of about 7,000 (Cameca definition), which was enough to resolve all mass interferences. A total of six acquisition cycles were recorded for a total analysis time of about 35 min per image (Extended Data Fig. 1b). Importantly, with these analytical settings, 23Na+ and 39K+, the two ions of direct importance for this study, were clearly detectable in all parts of the plant tissue from all experimental conditions (Extended Data Fig. 2), rendering a statistically robust direct comparison of the relative sodium and potassium distributions possible.
NanoSIMS image processing and analysis
Image analysis was performed using the L’IMAGE software v.10-15-2021 (developed by L. Nittler). Layers were drift corrected and accumulated. Regions of interest (ROIs) were drawn on ion images based on cryoSEM images to demarcate cells and the two primary subcellular areas of interest: vacuoles and cytosol (Extended Data Fig. 1e). For the 2.5 mM treatment, ROIs for the cell wall were conservatively selected with the calcium ion map as the basis for ROI assignment. Count-rate data were extracted from epidermal cells, without distinguishing between trichoblast (root hair) and atrichoblast (non-root hair) cells. Within individual cells, the vacuolar (or cytosolic) ROIs were combined into one ROI (that is, one ROI for each organelle type), because these regions are likely to be connected in three-dimensional space within the cells (Extended Data Fig. 1e). Only structures with smallest linear dimension of at least 5 pixels (that is, substantially larger than the size of the primary beam) were selected as ROIs. For each ROI, the average count rate for each elemental species (counts per second) was obtained from the accumulated images. In the absence of suitably matrix-matched standards, absolute elemental concentrations cannot be determined, but measured elemental ratios for specific subcellular compartments, obtained under identical analytical conditions, can meaningfully be compared within and between samples. To optimize comparability, WT and sos1 seedlings were grown on the same medium, under identical growth conditions, and went through identical cryo-workflow and analysis with identical instrument settings. Both WT and sos1 roots were included in each CryoNanoSIMS run. Line-profile data (5 pixels wide) were extracted from the regions highlighted on the image (Extended Data Figs. 2a–d and 3a,d).
Growth conditions for confocal microscopy experiments, complementation and root growth assays
In growth condition 1, seeds were surface-sterilized, sown on plates that contained half-strength MS with 0.8% agar (Roth), stratified at 4 °C in the dark for 2 days and grown vertically in growth chambers at 21 °C at constant light. In growth condition 2, seeds were sown in half-strength MS plates with 1% sucrose in a long-day chamber at 21 °C.
Complementation assay
After 4 days of growth on growth condition 1, seedlings were transferred to half-strength MS plates with 100 mM NaCl for 4 days and imaged. The primary root length was measured at transfer and after 4 days of growth and the increase in primary root growth was presented as delta primary root growth. For delta primary root growth recovery, the seedlings were transferred for 2 days to half-strength MS plates 8 days after salt stress and measured for root growth on day 10.
Confocal imaging
For the experiment with salt treatment in Fig. 4, 7 days after growth in growth condition 2, the seedlings were transferred to half-strength MS plates with 1% sucrose with 25 mM of NaCl or without NaCl, keeping the same growth conditions for another day. The short-term 25 mM and 100 mM NaCl treatments were done on 4-day-old roots grown in growth condition 1 (without sucrose). Brefeldin A (BFA) and Wortmannin treatments were performed on 7-day-old seedlings grown on growth condition 1 and treatments performed in liquid half-strength MS. BFA treatments were with 25 µM BFA and costained with 2 µM FM4-64 for approximately 1 h before imaging. For the BFA treatment on SOS1-GFP lines crossed with VAMP711-mCherry, the seedlings were not costained with FM4-64. Wortmannin treatments were with 33 μM Wortmannin for approximately 1 h before imaging.
Cloning and plasmid construction
For plasmid construct 1, SOS1-GFP26 (Extended Data Fig. 5b), a 4.8 kb DNA fragment containing the promoter and the first 13 exons and 13 introns (from −1,929 to +2,964) of the SOS1 locus, was amplified by polymerase chain reaction (PCR) using Arabidopsis genomic DNA as template and primers SOS1proHindIII-fwd (CGGGAAGCTTATACGTTTCGTAAGAAAC) and SOS1BamHI-rev (AAATTGGGTAGTGGATC). The PCR product was digested and subcloned in pBI321 using HindIII and BamHI sites. Next, pGPTV-II-GFP plasmid containing the SOS1 cDNA was digested at the unique BamHI and SacI sites. A 2.9 kb BamHI/SacI fragment containing the final part of the SOS1 cDNA fused to GFP was isolated and ligated in frame to the SOS1 genomic sequence of the pBI321 construct digested with the same restriction enzymes. For plasmid construct 2, SOS1-mCitrine (Extended Data Fig. 5c), a 8.357 kb DNA fragment containing the promoter and the full genomic sequence of SOS1 (from −2,250 to +6,073), was amplified by PCR using Arabidopsis genomic DNA as template and primers SOS1-fwd (ACAAAAAAGCAGGCTactaggatgatgtgtgttaatgtgagaggaagaagaa) and SOS1-rev (cccttgctcaccatTAGATCGTTCCTGAAAACGATTTTACTCGGAGA). For the mCitrine, a 7.47 kb DNA fragment was amplified by PCR from a pEN-R2-mCitrine-L3 plasmid (pEDO05) and primers mCitrine-fwd (TCAGGAACGATCTAatggtgagcaagggcgag) and mCitrine-rev (AACATCTCAATTCTCActtgtacagctcgtccatgc). For the terminator of SOS1, a 2.087 kb DNA fragment (from +6,077 to +8,127) was amplified by PCR using Arabidopsis genomic DNA as template and primers SOS1t-fwd (cgagctgtacaagTGAGAATTGAGATGTTTGTAACATAAGAAAACAAAATTGTTAGCT) and SOS1t-rev (CAAGAAAGCTGGGTTactagactgtgttatattggtaattagttggggct). Fragments were assembled by In-Fusion (Takara) into pENTR_L1L2 vector digested with BcuI (pRU335).
Confocal laser-scanning microscopy
Confocal images were acquired using a Leica Stellaris 5 WLL inverted confocal microscope with a ×63 objective with LASX (2022) software. For treatments, 7-day-old Arabidopsis seedlings were mounted with 30 µl of the corresponding medium (the same as used for the treatment) and imaged. For propidium iodide (PI) staining, PI was dissolved in liquid Na-free medium (10 μg ml−1 PI) and seedlings were immersed for 3 min, washed and imaged. GFP was excited with the 488 nm laser line of the MaiTai white laser and the emission was detected between 495 and 545 nm. PI and mCherry were excited with the 561 nm and emission was detected between 580 and 680 nm. One-shot images were taken with high-resolution 4,096 × 4,096, zoom ×3, in 8 bit depth, with ×4 line averaging.
Spinning disc confocal microscopy
We used a CSU-W1 Yokogawa spinning disk head fitted to a Nikon Eclipse Ti-E-inverted microscope, equipped with a CFI PlanApo ×100 numerical aperture, 1.40 oil immersion objective and with two EMCCD Andor iXon Ultra, operating with 1,024 × 1,024 pixels and 13 × 13 μm pixel size. GFP and RFP or mCherry were excited and detected simultaneously (excitation: GFP 488 nm, RFP/mCherry 561 nm; detection: GFP 525/50 nm; RFP/mCherry ET630/75 nm). Time-lapse images were processed using the in-built plugins of Fiji, that is, background subtraction (rolling ball radius, 30–50 pixels) and walking average projection (3–4 frames averaging) for the images for representative purposes.
Fluorescence signal and vesicle colocalization quantification
For intensity measurement of pixels, the ROIs were defined conservatively for the structures of interest: plasma membrane and internal area of the cell. The raw integrated density values were extracted for the ROIs. For the plasma membrane, the raw integrated density values were multiplied by a correction factor of 0.75 to account for the overlapping plasma membrane between cells and unaccounted cell faces. For the vesicle colocalization, background subtraction was performed using the ‘subtract background’ tool with a rolling ball radius of 50 pixels. The region encompassing a SOS1-GFP vesicle served as the ROI for each colocalization measurement for VHAa-1, RABF2a and RABF2b. A fixed size ROI with signal from mCherry-VAMP711 was used as a reference to quantify colocalization with SOS1-GFP for the vacuolar membranes. Pearson’s R values were determined using the Coloc2 plugin in FIJI.
Isolation of intact vacuoles
Protoplasts from leaves of 5-weeks-older Arabidopsis plants were isolated following the method described in ref. 37. The solution containing protoplasts was mixed (1:1) with lysis buffer, 200 mM sorbitol, 10% ficoll, 20 mM EDTA, 10 mM HEPES, pH 8.0, to release the vacuoles38.
Arabidopsis membrane isolation and purification
Total microsomes and tonoplast membranes used for western-blot analysis were isolated as described in ref. 39. Leaves were homogenized in a commercial blender and centrifuged at 10,000g for 20 min at 4 °C. The supernatant was centrifuged at 80,000g for 50 min at 4 °C, to obtain the microsomal fraction. Vacuolar membranes were isolated by loading the microsomal suspension onto a discontinuous sucrose gradient (0/20% w/v) that were centrifuged at 100,000g for 3 h at 4 °C.
Affinity purification of antibodies and western blotting
Anti-SOS1 antibodies were affinity purified according to ref. 40. A polypeptide comprising amino acids 998-1146 of Arabidopsis SOS1 was expressed and purified from E.coli as GST fusion. The protein was subjected to SDS-PAGE and transferred to nitrocellulose paper. After Ponceau S staining, the protein band corresponding to GST-SOS1 was cut out and incubated in TBS-T 5% non-fat milk for 1 h. The nitrocellulose strip was incubated overnight with anti-SOS1 serum41 diluted 1:1 in TBS. For elution of specifically bound antibodies, the nitrocellulose strip was incubated with 0.2 M glycine-HCl, pH 2.8 and 1 mM EGTA for 2 min. The eluate was neutralized immediately with 1 M Tris base. For western blots, protein samples were separated on 10% polyacrylamide SDS-PAGE gel and then electroblotted onto an Amersham Protran nitrocellulose membrane (Cytiva). The membrane was blocked for 1 h with TBS-T 5% non-fat milk. Then, the membrane was incubated overnight at 4 °C with the primary antibody (purified anti-SOS1 antibodies, 1:2,000 dilution; antiEpsilon subunit of tonoplast H+ATPase antibody AS07 213 from Agrisera, 1:10,000 dilution) in TBS-T 5% non-fat milk. Horseradish peroxidase-conjugated goat antimouse secondary antibody was used diluted 1:10,000 in TBS-T. The membrane was washed three times in TBS-T and incubated in ECL Select Western Blotting Detection Reagent for 5 min. The images were obtained using a GE Amersham Imager AI680 system. Refer to Supplementary Fig. 1 for full blots.
SEM image analysis and figure preparation
The SEM images were processed with levels setting in Photoshop (v.25.11.0). All NanoSIMS images were analysed using the L’IMAGE software v.10-15-2021 developed by L. Nittler. All confocal images were analysed using the ImageJ v.1.54 f (http://imagej.nih.gov/ij/)42 and Fiji programs. Figures were prepared using Adobe Illustrator (v.28.7.1) and Fiji. The only manipulation of images were histogram adjustments. Figures were then assembled in the Adobe Illustrator and Photoshop programs.
Statistics and reproducibility
We compared the effect of treatment on the two genotypes and considered each organelle separately. We performed two-way analysis of variance (ANOVA) and a paired comparison with treatment and genotype as the two factors, and significant interactions were determined by post hoc Tukey test for vacuole:cytosol ratio and Na:K ratio. Box plots for ratios show the mean with standard deviation and standard error. For Fig. 1b,c, the images are representative illustrations of a cryo-planed root meristem and elemental distribution performed on three different roots. The data for Fig. 2d,e were obtained using the following. Na-free medium: WT n = 15 cells; sos1 n = 13 cells, from one experiment with three different roots. 25 mM NaCl treatment: WT n = 49 cells; sos1 n = 33 cells, from two experiments with six roots in total. Pairwise comparison of vacuole:cytoplasm ratio was carried out with treatment and genotype as factors; Na-free medium: WT n = 15 cells; sos1 n = 13 cells, from one experiment with three different roots. 25 mM NaCl treatment: WT n = 22 cells; sos1 n = 23 cells, from two experiments with six roots in total. Significance: *** P < 0.001 analysed by two-way ANOVA with Tukey post hoc tests. The data for Fig. 3b,c were obtained using the following 100 mM NaCl treatment: WT n = 34 cells; sos1, n = 26 cells from two experiments with six roots in total. Box plots show the mean values (horizontal lines), 25% (boxes) and 75% (thin vertical bars) quartiles; outliers are indicated with solid circles. Pairwise comparison of ratios was carried out with treatment and genotype as factors. Significance: ***P < 0.001 analysed by two-way ANOVA with Tukey post hoc tests. For Fig. 3d, rice root meristems were treated with 100 mM NaCl, WT n = 18 cells. Statistical analyses were performed using OriginPro v.2023b and GraphPad software (Prism 10). Figure 4a is a representative picture of n = 7 roots with three biological replicates.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
