Allocation of salt contaminants to leaf tissues
The high mobility of salt ions within plant tissues (Mengel & Kirkby, 2001; White & Broadley, 2001; Li et al., 2017) and lack of fresh material (Rokebul Anower et al., 2017) formed challenging constraints with regard to salt microlocalisation free of dislodging artefacts. Using a freeze substitution procedure (Wu & Becker, 2012) and with visualization and mapping of 23Na+ and 35Cl- ions using a Focused Ion Beam-Secondary Ion Mass Spectrometer (FIB-SIMS) microscope, dislodging of Na and Cl contaminants from the vacuole to cytosol and cell wall compartments was observed (Fig. S1A, B versus 4I, J). As a consequence, we sectioned the deeply frozen sample-carrier sandwiches directly, focusing on the leaf rim samples and using a Leica EM FC6 ultra-cryo-microtome at ScopeM, -150 °C cooling and LN2 flushing. Given the unrecoverable sections and distortion artefacts of stepwise freeze-dried block sample (Fig. S1C, D versus 6A-C), full cryo-conditions were also required with a view to visualizing and mapping the salt contaminants and leaf nutrients. Therefore and prior to observation, the planed samples transferred in LN2 to the UMR Silva-Silvatech Microscopy platform in Champenoux, France, were coated with 1.5 nm Pt (-120 °C, 2.5E-2 mbar Ar plasma) and the ice contamination freeze-etched (30 min, -85 °C, 2.7E-6 mbar), using an EM VCT100 vacuum cryo shutter (LEICA, UK) and an EM ACE600 double sputter coater. The samples were then transferred to a cryo- Focused Electron Gun Scanning Electron Microscope (cryo-FEGSEM; SIGMA HD VP, ZEISS, Germany) equipped with a High Definition Back Scattered Electron Detector (HDBSD) and an Energy Dispersive Spectrometer (X-MaxN EDS – SDD 80mm²; Oxford-Instruments, UK) interfaced to the SEM by the INCA software (Oxford–Instruments, UK). Element maps and line scans of salt contaminants and nutrients were obtained at – 160 °C, using magnifications of 1.35 or 4.5 kX and 20 keV acceleration voltages (30-35° take off angle, 9 mm work distance, 200 µs dwell time, 30 frames and 1500 s measurement time). The smooth and homogeneous material structure of samples allowed us the quantification of contaminants and main nutrients within the vacuolar compartment. At two randomly selected leaf blade transects per sample, point measurements were performed within the vacuole of one cell per tissue (10 kV HT, 30 s measurement time). The range of electron penetration in such conditions can be estimated to 2-3 µm (Huang & van Steveninck, 1989). The spectrometer was calibrated with reference to calcium standardization before micro-analysis. The half quantitative nutrient mass percent composition (weight %) of vacuolar sap was estimated on the basis of deconvoluted spectra for each element (INCA-XPP matrix deconvolution). The estimates per tissue type in the two investigated leaf blade transects were averaged.