CAPTIONS
Fig. 1. Location of and level of salt contamination at the study sites.
Fig. 2. Visible injury by salt stress within foliage and at a street site from Riga’s street greenery (Meierovica 1; September 16, 2014). For microscopy assessments, leaf center (light green) and rim (teal blue) disks were excised apart the main center vein and next to salt-triggered necrosis, within still green foliar tissues.
Fig. 3. Principal changes in the concentration of macro- and micronutrients, as a function of Na and Cl accumulation within foliage of trees at the study sites. The data points represent site averages (+ SE). Statistics refer to linear mixed effects models (LMEM). Solid lines and shaded areas indicate significant (P≤0.05) linear regressions and 95% confidence intervals.
Fig. 4. Tissue-(A -C ) and cell-level (H -K ) allocation of salt contaminants (Na, Cl) and some macro-nutrients (K, Ca) within leaf rim samples of Tilia x vulgaris leaves from moderately (B Meierovica 1) and strongly (C, H-K Aspazijas) contaminated versus control site (A NBG) in the street greenery of Riga. D -GElement spectra obtained for the element maps (D forA , E for B , F for C ,G for H -K ). The salt contaminants were found within all leaf blade tissues. At cell level, they primarily accumulated within vacuoles (v) and were missed within cell walls (cw). Other structures: uE – upper epidermis, pP – palisade parenchyma, sP – spongy parenchyma, lE – lower epidermis. The whitish globular structures represent ice contamination. Average foliar concentrations of Na/Cl: 86±8/760±121 mg kg-1 (NBG, A ,D ), 4020±739/7333±1746 mg kg-1 (Meierovica 1,B , E ), 9733±1964/13917±2022 mg kg-1(Aspazijas, C , F -H ). Technical specifications: High pressure frozen leaf samples planed by cryo-ultramicrotomy and examined using a cryo-FEG-SEM in high-vacuum cryo-mode (– 160 °C) at acceleration voltage of 20 kV, magnification of 1350x (A-F) and 4500x (G-K), 9 mm working distance, HDBSD detectors for imaging and X-ray Energy Spectrometer for micro-analysis; element spectra and maps obtained by scanning during 1500 s (200 µs dwell time); measurements cumulated over 30 frames and mapped at 512 x 384 ppi resolution (each point in elementary map representing the total number of counts for the element during mapping); element maps overlaid on HDBSD images; scale of spectra: 71517 cps.
Fig. 5. Nutrient content (mass percentages) of vacuolar medium within leaf rim samples of Tilia x vulgaris leaves from salt contaminated sites (average foliar concentrations of Na/Cl at Meierovica 1, Gertrudes, Aspazijas: 4020±739/7333±1746, 7067±677/7083±682, 9733±1964/13917±2022 mg kg-1. The values represent averages per tissue (+ SE) of measurements performed along two transect through the leaf blade, in one leaf sample per site (N = 3). The insert graph shows the vacuolar nutrient content at the asymptomatic NBG site (average foliar concentrations of Na/Cl: 86±8/760±121 mg kg-1). Abbreviations: uE – upper epidermis, pP – palisade parenchyma, sP – spongy parenchyma, lE – lower epidermis.Technical specifications: High-pressure-frozen leaf samples planed by cryo-ultramicrotomy and examined using a cryo-FEG-SEM in cryo-mode at acceleration voltage of 10 kV and magnification of 2200x (260 pA of current, 9.5 mm working distance). Collection of EDS spectra by each measurement point during 30 s. Half-quantitative nutrient mass percent composition (weight %) of vacuolar sap estimated on the basis of deconvoluted spectrum of each element (XPP deconvolution).
Fig. 6. Structural effects of salt contamination in symptomatic leaves of Tilia x vulgaris from moderately (B, E, H, L ) and strongly (C, F, I, J, M ) contaminated sites in the street greenery of Riga. Asymptomatic leaves from NBG (A, D, G, K ). Samples from the leaf center (A-F ) and leaf rim region (G-M ). In pP cells of symptomatic leaves, salt contamination caused an increase in the cell size driven by that of vacuome, with one large vacuole containing many vesicular inclusions (* in H-J ) filling most of cell volume finally (v; E, F, H-J ). The cytoplasm showed degenerative changes (# in M ) and an increase of autophagic vesicles (av). The chloroplasts (ch; L, M ) were also degenerated, with larger plastoglobules (pl; E, F,L, M ) protruding and being expelled (&) into the vacuole. Other structures: lE: lower epidermis, m: mitochondria, mvb: multivesicular body, n: nucleus, pP: palisade parenchyma, sP: spongy parenchyma, st: starch, ta: tannin body, uE: upper epidermis. Average foliar concentrations of Na/Cl: 86±8/760±121 mg kg-1(A, D, G, K ), 4020±739/7333±1746 mg kg-1(B, E, H, L ), 9733±1964/13917±2022 mg kg-1(C, F, I, J, M ). Technical specifications: A-F : 1.5 µm semi-thin cuttings stained with toluidine blue and acid fuchsine and observed in diascopic light microscopy; G-M: 75 nm ultra-thin sections stained with uranyl acetate and lead citrate, and observed in TEM.
Fig. 7. Structural responses within palisade cells of leaf rim samples, as a function of Na and Cl accumulation within foliage of trees at the study sites. The data points represent site averages (+SE). Statistics refer to linear mixed effects models (LMEM). Solid lines and shaded areas indicate significant (P≤0.05) linear regressions and 95% confidence intervals. Abbreviations: pP: palisade parenchyma, Pl size: percentage area of plastoglobules within chloroplasts. Vacuole size: percentage area of largest vacuole within pP cells.
Fig. 8. Multivariate structural responses to salt contamination in palisade parenchyma of leaf rim samples - RDA models. Correlation biplot based on a redundancy analysis of the structural data measured in mesophyll, showing the relationship between markers of salt injury (response variables, black arrows) and salt concentration in foliage (Na, Cl explanatory variables; blue arrows). The leaf injury (i.e. percentage area of leaf showing necrosis, light blue arrow) was passively projected, as a supplementary variable. Altogether, the first and second canonical axis explained 78.12% of total variance in the mesophyll structure dataset, whilst only the first axis was significant (P < 0.001). The color of each tree score shows the average Na-contamination within lime tree foliage at the five sampling sites (■ 86 + 7 mg kg-1, ■ 895+ 477 mg kg-1, ■ 4507 + 534 mg kg-1, ■ 8000 + 793 mg kg-1; N = 3 trees per site), with each site centroid indicated by a cross and label. Abbreviations for the descriptor variables: ChS chloroplast size, CS cell size, circ Cell circularity, PlD plastoglobule density, PlS plastoglobule size, VS vacuome size.
Fig. S1. Element mapping by means of compact FIB-Tof-SIMS performed during the microlocalisation trials, with a focus on Na and Cl ions within cross-sections of leaf rim samples collected in the street greenery of Riga. The Na and Cl concentration of leaf sample is indicated directly in the Fig. A, B Microlocalisation trial applying a freeze-substitution protocol. The Na and Cl contaminants have been dislodged from the vacuole to cytosol and cell wall compartments during the freeze-substitution. C, D Microlocalisation trial using the cryo-fixed sample block, after trimming by cryo-ultramicrotomy and controlled freeze-drying (as a consequence of of intercellular spaces, large vacuoles and lignified structures, the sections fell apart and could not be retrieved). Dislodging of salt contaminants was reduced but the tissue and cell structure was too much distorted for enabling the observation of salt distribution among the storage compartments. Sampling sites: Barona (A , B) ; Meierovica 1 (C , D ).
Technical specifications: Sample preparation: A, Bfreeze substitution of samples in LN2 using an EM AFS2 (Leica Microsystems) at ZMB (1% OsO4 in 100% acetone -90oC 8 hours, graded temperature increase to 0oC in 16 hours, rinsing in 100% acetone, infiltration by a series of graded acetone/Epon-Araldite mixture at +4oC and embedding in Epon-Araldite). Thin cuttings (100 nm) sectioned using a Reichert UltraCut S ultramicrotome, mounted on custom-made Si-wafers and coated with 4-6 nm Au. C, D stepwise freeze-drying of 400 µm-deep trimmed block sample (6.2 h in total, using a modified Bal-Tec BAF 060), prior to mounting on Al-stub. Element mapping: Visualization and mapping of 23Na+ and 35Cl- ions by means of Focused Ion Beam-Secondary Ion Mass Spectrometer microscope (compact FIB-SIMS) at the Swiss Research Institute for Applied Materials Sciences and Technology (EMPA) in Thun, Switzerland (Ga ion beam, 20 keV, 180 pA, 100x100 µm view field). 20keV 180-200 pA, 100x100 μm.
Fig. S2. Structural effects of salt contamination in the upper epidermis cells of Tilia x vulgaris foliage from moderately (B, E ) and strongly (C, F, G ) contaminated sites in the street greenery of Riga (leaf rim samples). A, BAsymptomatic material from the National Botanical Garden. Within epidermis cells, salt accumulation further hastened the ontological senescence by exacerbating the ongoing degenerative changes, such as the increase in size of plastoglobules (pl) within leucoplasts (l;B, E, F, G ), condensation of chromatin within nucleus (*n;B, E ), or higher frequency of autophagic vesicles (av) and multivesicular bodies (mvb; B, E ). Other structures: #, fibrillous mucilage material, m: mitochondria v: vacuole. Average foliar concentrations of Na/Cl: 86±8/760±121 mg kg-1(A, D ), 4020±739/7333±1746 mg kg-1(B, E ), 9733±1964/13917±2022 mg kg-1(C, F, G ). Technical specifications: 75 nm ultra-thin sections stained with uranyl acetate and lead citrate, and observed with TEM.