Arabidopsis thaliana ecotype Columbia-0 were grown hydroponically using an Araponics growth system (Araponics, Liège, Belgium), as previously described in Podgórska et al. (2013). The plants were grown for 7 weeks continuously on nutrient medium containing Ca(NO₃)₂ (2.5 mM) or (NH₄)₂SO₄ (2.5 mM) as the sole nitrogen source. NO₃^–-treated plants were considered as controls. Experiments were carried out on plant leaves.

The growth parameters of leaves were monitored by measuring the area and the thickness of the ninth fully developed leaf. The area of the leaf blade was calculated using WinFOLIA leaf analysis software (Image Analysis for Plant Science, Regent Instruments Inc.). The thickness of the leaf blade samples (the mean for three locations) was determined using Nikon SMZ1000 (Japan) stereo microscope. The measurements were made using ImageJ (v.1.51f) software.

To gain insights into the mechanical properties of growing leaves we performed tensile tests on the leaf lamina in full turgor and rehydrated frozen-thawed samples. Leaf sections were cut (5 mm × 2 mm) from the front part of the leaves next to the central nerve, with a pair of razor blades mounted between a plastic separator. The leaf specimens were glued at the ends to plastic sticks with small amount cyanoacrylate adhesive. A half of the samples were examined for mechanical properties at full turgor and the other half was tested after freezing at -20°C for 1 h to destroy cell membrane. This procedure led to a complete reduction in turgor pressure of tissues so that the results depended only on the cell wall mechanical properties. Unidirectional tensile tests were performed on the computer-driven Instron universal machine (model 5542, High Wycombe, United Kingdom) at a deformation rate of 1 mm/min. The plastic sticks were firmly connected by means of grips to the load cell (10 N capacity) and to the immovable part of the instrument. Sample strains were measured locally at high accuracy (1/100 pixel) from video recordings analyzed using Video gauge system (Imetrum, United Kingdom). The displacement of selected points on the sample surface could be tracked at an accuracy level of approximately 10 μm for a gauge length of approximately 0.5 mm. The use of this non-contact video extensometer allowed us to avoid the interfering effects due to the sample slippage within clamps, and non-linearity in strain distribution close to the grids. The stress-strain curve was the basis for the calculation of mechanical parameters.

The relative strain (%) was calculated directly from the displacement of tracked points during tensile sample deformation. The strain at break was defined as (x_max – x₀) ⋅ 100%/ x₀ where x₀ was the distance between the selected points before loading (the gauge length) and x_max was the distance between the same points at the break. The tensile stiffness EA of leaf structure was determined as the ratio of the load increment to the corresponding relative strain at the linear phase of sample deformation. The Young’s modulus (E) of the leaf blade was calculated as follows: E = EA/A, where EA is the tensile stiffness and A is the cross-sectional area of the sample. The tensile strength was defined as the ratio of the load at break to the value of A.

Cell walls were prepared according to the method described by Solecka et al. (2008). Isolated cell walls were utilized for total polysaccharide determination, cell wall fractionation and protein extraction for enzymatic measurements. Cell walls were extracted from around 5 g of frozen leaf tissue in 0.05 M HEPES-KOH, pH 6.5 and centrifuged at 5,000 × g for 10 min. The pellet was washed three times with cold water. Proteins from the crude cell wall preparations were extracted using 1 mL 0.05 M HEPES, 1 M NaCl, and 1 dose of protease inhibitor cocktail (Complete Ultra Mini EDTA-free Easy pack, Roche, Penzberg, Germany) under continuous agitation at 4°C for 12 h. The homogenates were centrifuged at 13,000 × g for 15 min and the resulting supernatant containing cell wall proteins was recovered, whereas the pellet collected was comprised of cell walls. The cell wall residue was washed three times with water, twice with 70% (v/v) ethanol at 70°C, and 100% acetone. The resulting precipitate containing the cell wall was air dried for 24 h and used for cellulose, pectin and lignin determination. The supernatant was desalted and concentrated using centrifugal filter devices (10K cut-off), according to the manufacturer’s protocol (Microcon; Millipore, Billerica, MA, United States). The recovered supernatant that had not passed through the filter contained the extracted cell wall proteins. Soluble protein content in cell wall extracts was estimated using the colorimetric method of Bradford (1976).

The cell wall residue was further fractionated into the pectin-enriched fraction and a hemicellulose-enriched fraction. A washing step with 90% (v/v) dimethyl sulfoxide (DMSO) for 24 h at RT was applied to eliminate starch contamination in cell walls (Kubacka-Zęebalska and Kacperska, 1999). The resulting pellet was washed free of residual DMSO by several washings with 96% (v/v) ethanol and evaporated to dryness. The cell wall preparations were extracted with 10 mL of 0.5% (v/v) ammonium oxalate at 100°C for 1 h. The ammonium oxalate soluble fraction was dialysed for 50 h against deionised water, evaporated to dryness under reduced pressure and weighed. This dry mass represented the pectin-enriched fraction. The ammonium oxalate insoluble pellet was dissolved in 10 mL 4 M KOH supplemented with 20 mM NaBH₄ for 12 h at RT. The alkaline-soluble fraction was acidified to pH 5.5 with acetic acid, dialysed for 50 h against deionised water, evaporated to dryness under reduced pressure and weighed. This dry matter represented the hemicellulose-enriched fraction. The obtained fractions were utilized for uronic acids and neutral sugar determination and FTIR analysis.

Total cellulose content of the extracted cell walls was measured using the colorimetric Anthrone assay, according to the method of Updegraff (1969). All the polysaccharides, except crystalline cellulose, were released from the cell walls (10 mg) by incubating dry material in a 1 mL mixture of HNO₃ and CH₃COOH (1:8) for 1 h at 100°C. The homogenate was centrifuged at 12,500 × g for 10 min and the pellet was washed twice with water. The resulting pellet containing cellulose was dissolved in 1 mL 67% H₂SO₄ (v/v). The samples were transferred to 0.8 mL 0.2% cold Anthrone reagent in concentrated H₂SO₄. After boiling for 5 min, blue color development was observed. Cellulose content was measured spectrophotometrically at 620 nm against cold Anthrone reagent. The results were presented as equivalents of glucose used in the standard curve.

Pectins were estimated as uronic acid content in cell walls and in the pectin-enriched fraction, according to the colorimetric method established by Blumenkrantz and Asboe-Hansen (1973). The samples were dissolved in water and digested with 1.2 mL 0.0125 M Na₂B₄O₇ in concentrated H₂SO₄ at 100°C for 5 min. After cooling the samples, 20 μL 0.15% meta-hydroxybiphenyl in 0.5% NaOH was added for pink color development. Pectin content was estimated spectrophotometrically at 520 nm against blank samples. The results were presented as equivalents of galacturonic acid used as the standard.

Neutral sugar content was measured in the pectin-enriched fraction, according to the phenol sulfuric acid method of Dubois et al. (1956). The samples (2 mg) were diluted in water and hydrolyzed in 300 μL concentrated H₂SO₄, followed by addition of 50 μL 5% phenol. After 30 min incubation at room-temperature (RT), yellow color development was observed and absorbance was measured at 490 nm against blank sample. The neutral sugar content was presented as equivalents of glucose used as the standard.

Total lignin content in cell walls was determined by the acetyl-bromide method (Johnson et al., 1961) as described in Hatfield et al. (1999). Dry cell wall material (approximately 20 mg) was dissolved in 1 mL 25% acetyl-bromide in acetic acid for 2 h at 50°C. The samples were further dissolved in 5 mL 2 M NaOH, 0.75 M hydroxylamine-HCl, diluted with acetic acid and incubated overnight at RT. Lignin content was measured spectrophotometrically at 280 nm; the results were presented as equivalents of coniferyl alcohol.

Boron and Ca²⁺ concentrations in absolute dry mass samples from leaves were determined by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES). The analysis was performed in an authorized Laboratory of Quality Investigation of Horticultural Products at the Horticultural Research Institute (Skierniewice, Poland).

The amount of phenolics bound to cell walls was measured using the Folin method (Forrest and Bendall, 1969) as described earlier in Solecka et al. (1999). Phenolics were released from the cell wall preparations by alkaline hydrolysis. Extractions were carried out by vortexing samples in 0.4 mL 0.5 M NaOH and incubating for 24 h at RT. After neutralization with HCl, the homogenates were centrifuged at 13,000 × g for 20 min. The supernatants containing phenols were added to 0.2 N Folin reagents (Sigma). Thereafter, 0.6 mL 7.5% Na₂CO₃ was added and samples were incubated for 30 min at RT in darkness. The phenolic content was determined spectrophotometrically at 750 nm against a blank sample; ferulic acid was used as a standard.

Enzymatic activity of pectinases was determined in cell wall protein extracts. The activity of PME was estimated following the colorimetric method described by Solecka et al. (2008). The reaction mixture was composed of 0.2 M NaCl, 0.015% (w/v) Methyl Red as a pH indicator, and contained 0.5% citrus pectins as the substrate for PME with different grades of methylesterification (8.9, 30, and 90%, respectively). The pH of the reaction mixture was titrated to three different pH values: 5.0, 6.8, and 8.5. The reaction performed by PME was traced after addition of 50 μL of the cell wall protein extracts, and the changes in color from yellow to red were detected using a spectrophotometer at 525 nm. A calibration curve was obtained by adding 1–200 nEq H⁺ to the reaction mixture.

The activity of PG was determined using the spectrophotometric method of Gross (1982) with minor modifications. The reaction was induced by adding 0.5% polygalacturonic acid in 50 mM sodium acetate buffer (pH 5.2) to cell wall protein extracts and incubated at 37°C for 2 h. The reaction mixture was diluted with 50 mM sodium acetate buffer (pH 5.2), and 30 μl cyanoacetamide was added to label the released reducing end-groups. After inactivation of the enzymes by boiling for 10 min, the absorbance was recorded at 320 nm. The activity was calculated based on galacturonic acid released by PG.

The activity of enzymes engaged in cell wall lignification was determined in cell wall protein extracts. POX activity was monitored spectrophotometrically by using guaiacol as the substrate (Veljovic-Jovanovic et al., 2001).

The activity of phenylalanine ammonia-lyase (PAL) was measured using the spectrophotometric method by Havir and Hanson (1968), as described by Solecka and Kacperska (1995). The reaction mixture was composed of 0.016 mM L-phenylalanine, 0.15 mM borate buffer (pH 8.8), 3.6 mM NaCl, and 50 μL of extracted proteins. Production of cinnamic acid was monitored after incubation of the samples at 37°C for 3 h. The enzymatic reaction was inhibited by adding 500 μL 6 M HCl and the increase in absorbance was measured at 290 nm against a blank sample. To eliminate non-specific reactions during the incubation time, additional tests were performed using reactions without the addition of L-phenylalanine as the substrate, or with extracts deactivated by boiling.

To determine XTH/XET protein levels, 20 μg of cell wall proteins were separated in 12% sodium dodecyl sulphate (SDS)–polyacrylamide gel electrophoresis. Anti-XTH/XET (Agrisera, Vännäs, Sweden) were used as primary antibodies (diluted 1/500) and anti-mouse antibodies (Bio-Rad, Hercules, CA, United States) were used as the secondary antibodies. Immuno-blotting was performed according to the standard protocol. Bands for XTH/XET located at 33 kDa were determined on the basis of a pre-stained protein marker (Bio-Rad) as the reference. The relative protein levels were quantified by densitometry analysis using Image-Lab 5.2. software (Bio-Rad).

Mid-infrared spectroscopy measurements were performed on cell wall material ground into a fine powder with an agate mortar and pestle and washed successively with 70% ethanol, 1:1 (v/v) mixture of chloroform and methanol, and acetone. Spectral data were collected using Fourier Transform Infrared (FTIR) spectrometer (Nicolet iZ 10 module, Thermo Fisher Scientific Inc., United States) equipped with a deuterated triglycine sulfate (DTGS) detector, KBr beam splitter, and a diamond single bounce Attenuated total reflectance (ATR) accessory. Two-hundred and fifty-six interferograms were recorded and added together between 4,000 and 400 cm^-1 at 2 cm^-1 resolution. The spectra were baseline corrected, normalized to a constant area between 1,760 and 880 cm^-1, and smoothed at a final resolution of 4 cm^-1 by apodization using the Blackman–Harris 3-term function and a zero-filling factor of 2. Collection, pre-processing, and analysis of spectral data were performed by means of OMNIC software (v. 9.0, Thermo Fisher Scientific Inc.). The spectra were subjected to multiple univariate and multivariate analysis. The latter was performed by unsupervised principal component analysis (PCA) (Unscrambler X, v.10.1, CAMO, Norway).

The cell walls were isolated and purified to reveal cellulose microfibrils according to the method of Davies and Harris (2003), with minor modifications. The leaf tissues were homogenized in liquid nitrogen, and aliquots of the resulting powder were homogenized in 1% SDS, three times. The resulting residue was incubated successively with 50 mM Trans-1,2-diaminocyclohexane N,N,N′,N′-tetraacetic acid (CDTA, adjusted to pH 7.0 with KOH) at 70°C for 30 min (three times) and 50 mM Na₂CO₃ overnight at 4°C to wash out pectic polysaccharides. Finally, the cell wall deposit was washed three times with warm water. After each incubation, the suspension was centrifuged (5,000 × g for 10 min) and the supernatant was discarded. The final pellet containing cell wall probes was resuspended in water and used for Atomic Force Microscopy analysis (AFM).

AFM measurements were done with an Agilent 5500 atomic force microscope (Agilent, United States) in contact mode. The 205 nm long triangular cantilever B of SNL-10 Bruker AFM probe (Bruker Probes, United States), with 2 nm tip radius was used with force const. of 0.12 N/m. Aqueous dispersions of the cell wall were layered on a freshly cleaved mica (SPI Supplies, United States) and then allowed to dry in air. Imaging was carried out in the air whilst the samples remained visibly moist. If possible images were recorded at a resolution of 256 × 256 or 512 × 512 and 1 ln/s scanning speed. Topography, phase, and amplitude images were recorded with PicoView software (Agilent, United States) and analyzed with Gwyddion software (Nečas and Klapetek, 2012). The images were flattened to remove tilt or bow in each scan lie and exported to TIFF format.

Ultrastructural imaging of mesophyll leaf cells were performed as described in Podgórska et al. (2013), and the thickness of the cell walls was measured in micrographs obtained by transmission electron microscopy (TEM). Briefly, small sections of plant leaves were fixed in 2.5% (v/v) glutaraldehyde in 50 mM sodium cacodylate buffer (pH 7.3) for 4 h at RT. The tissues were post-fixed in 2% (v/v) OsO₄ solution in 50 mM sodium cacodylate buffer (pH 7.3) overnight at 4°C. Thereafter, the samples were stained in 1% (w/v) uranyl acetate in 70% (v/v) ethanol for 1 h. This was followed by further dehydration in a graded ethanol series (70–100%) and acetone series up to 100% acetone. The plant material was progressively fixed in Epoxy embedding medium. The tissues were polymerized in 100% EPON at 60°C for 72 h. The blocks containing the embedded sample were sectioned (90 nm) on a RMC PowerTome microtome (Boeckeler Instruments, Tucson, AZ, United States) and mounted on uncoated copper grids (300 meshes). The sections were examined using a JEM 1400 transmission electron microscope (JEOL Co., Japan) and a high-resolution digital camera (CCD MORADA, SiS-Olympus, Germany) at the Laboratory of Electron Microscopy, Nencki Institute of Experimental Biology, Warsaw, Poland. The thickness of a double layer of cell walls was determined in micrographs using Image Processing and Analysis in Java software (ImageJ). The measured zone included the primary cell wall and the middle lamella of two adjacent cells.

For determining cellulose localization, the upper epidermis was removed from fresh plant leaves and the tissues were cut into small sections (approximately 5 mm). The samples were stained with 0.5 mg/mL Calcofluor White (Sigma) for 5 min at RT. The leaf pieces were washed in water three times and placed on a glass slide. A NIKON A1R MP confocal laser scanning system (Zeiss LSM-510) was used to detect fluorescence with a 60× (numerical aperture 1.2) water immersion objective. Calcofluor White fluorescence was induced with the 404 nm helium-neon laser at an emission of 490 nm. In these micrographs, the cell diameters were determined, measuring the length and width of each cell. The resulting cell area was calculated assuming an ellipsoid cell shape, using the following formula: A = πab, where a is the leaf length and b the leaf width, and A is the cell area.

The localization pattern of pectins and xyloglucan in leaf tissues was studied using monoclonal antibodies, which bind to specific cell wall epitopes. Hand-cut sections of the leaf stalk and the leaf blade were fixed in 0.5% glutaraldehyde with 4% paraformaldehyde in sodium phosphate buffer (pH 7.4) for 2 h. Thereafter, the tissues were washed three times in sodium phosphate buffer for 10 min and dehydrated in a graded ethanol series of 30–100% (v/v), for 10 min in each grade. Further, the specimens were progressively fixed in Steedman’s wax dissolved in ethanol at 40°C. The plant material was polymerized in 100% Steedman’s wax at RT. The wax blocks containing the embedded samples were sectioned (20 μm) on a RMC PowerTome microtome and mounted on Poly-L-Lysine coated glass slides. After rehydratation of the samples in decreasing ethanol concentrations of 96–30% (v/v), slides were used for immunolocalization. Also fresh leaf blades were embedded in 4% agar in 0.9% NaCl and sectioned (150 μm) using a vibratome. The fixed specimens and fresh leaf sections were washed with tris-buffered saline (TTBS) buffer and incubated with a blocking buffer containing: 5% bovine serum albumin (BSA) in TTBS and 0.1% Tween20 for 60 min at RT. The sections were immunolabelled with the primary-antibodies LM19, LM20 or LM25 (1:50, Plant Probes, Inc., United Kingdom) in 0.5% BSA in TTBS overnight at 4°C with gentle shaking. After six washes with TTBS buffer, a secondary antibody (1:100, Anti-rat-Alexa IgM, Fluor 488, Molecular Probes, Leeds, United Kingdom) was applied for 1 h at RT. Finally, the sections were rinsed with TTBS buffer and mounted with Vectashield (Vector, United States) anti-fading medium. The fluorescence was observed with NIKON A1R MP confocal laser scanning system (Zeiss LSM-510) at 10× and 60× magnifications and photographed. The fluorescence was measured using the 488 nm line of an argon-krypton laser and the emission was recorded for 500–545 nm. Negative controls were performed by omitting the primary antibody step, obtaining no fluorescence signal in each control frame for all slides stained.

The mRNA level of cell wall enzymes was quantified by real-time polymerase chain reaction (qRT-PCR) analysis. Total RNA was extracted using the RNeasy Plant Mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. First strand cDNA synthesis was performed using 1 μg total RNA as a template with reverse transcriptase following the RevertAid H minus first strand cDNA synthesis kit (Fermentas, Thermo Fisher Scientific, Inc., Waltham, MA, United States), according to the manufacturer’s protocol. Thereafter RNA digestion was performed as described by Escobar et al. (2004). For all qRT-PCRs, the iTaq SYBR Green supermix was used (Bio-Rad) at 2× diluted reactions. New primer pairs were designed, and the sequence of at least one of the primer pairs covered an exon-exon border. The primer sequences were as follows: FER (At3g51550): 5′-TGCCGTCACTTCTCGTTTGC-3′ and 5′-CTTGCTCGGACATTGGGTTG-3′; THE1 (At5g54380): 5′-TCAAGAAGGCGGTAATGGACA-3′ and 5′-CAACCCCAAGCAACGAACTC-3′; EXP1 (At1g69530): 5′-GAAGAGTGCCGTGCGTGAG-3′ and 5′-TAGGTTGAAGTAAGAGTGTCCGTTT-3′; EXP17 (At4g01630): 5′-GCCTCTGGTACAATGGGTGG-3′ and 5′-TGAGCAAAGTTCGGTGGACA-3′; PG1 (At3g26610): 5′-CCAATGAAACCAACGGCACT-3′ and 5′-TGACTTTGACTCCATCCGAATC-3′; Putative-PG (At2g43890): 5′-ACAGGGTTCTGGAGTGAAGATTAGTC-3′ and 5′-TCGCTTGGCACGGATTACT-3′.

The qRT-PCRs were performed using a thermo cycler (IQ5 RT-PCR detection system, Bio-Rad), using 60°C as the annealing temperature. Transcript abundance was quantified by comparing the quantification cycle (Cq) values of the target gene relative to those of the reference gene PP2A (At1g13320, Czechowski et al., 2005) according to the method described by Pfaffl (2001). The transcript levels of each gene were expressed in relation to those in control plants (set as 1).

The results were expressed as the mean values ± standard deviations from 3 to 20 measurements taken from 3 to 5 independent plant cultures. Experimental data from plants grown on NO₃^– or NH₄⁺ was statistically analyzed with Student’s t-test using Microsoft Excel (Microsoft). Results shown with an asterisk were significantly different at a p-value < 0.05.

During NH₄⁺ nutrition plants showed overall smaller leaves in their rosettes than NO₃^– treated controls (Figure 1A). The leaf area was seven-times smaller during growth on NH₄⁺ than on NO₃^– (Figure 1B). Further, the thickness of the leaf lamella was 30% less in plants grown on NH₄⁺ as the only N source (Figure 1C).

The maximum stress that could be borne by the tissues (tensile strength) increased by more than 40% in fresh leaves and 70% in rehydrated frozen-thawed leaves of plants grown on NH₄⁺ than in the controls (Figure 1D). The increase in tensile stiffness (the effective Young’s modulus) of the leaf tissue was about 60 and 90% higher in fresh and rehydrated frozen-thawed leaves, respectively, than in the control (Figure 1E). Additionally, in plants grown on NH₄⁺ a higher ratio of the stiffness to the strength of fresh leaves (9.4 vs. 8.5) and for cell-wall preparations obtained after freezing and thawing (6.07 vs. 5.45).

Changes in morphology of cell wall were observed in CLSM and TEM. In transverse sections through leaf cell walls, the middle lamella and primary cell wall were identified and the thickness of both layers of the cell wall was determined. The cell walls of NH₄⁺ grown plants were around 40% thinner than in the controls (Figure 2D). There were no other changes in leaf cell wall ultrastructure seen in response to the N treatments (Figure 2C). In addition, the cell walls were traced in vivo using fluorescent dyes. Calcofluor White is a stain that reveals all-â-glucan-containing cell walls. Therefore, it is highly effective in visualizing cell wall disassembly. Transverse sections of palisade mesophyll leaf cells revealed slightly thinner cell wall of plants grown on NH₄⁺ than in the controls (Figure 2A). Moreover, it was noticed that the cells of NH₄⁺-treated plants were smaller. The size of individual cells in the leaf blade was up to five times lower in plants growing on NH₄⁺ than in controls (Figure 2B).

The AFM technique is widely used to study cell walls revealing layered fibrous structures displaying microfibril arrangement (Kirby et al., 1996, 2006; Kirby, 2011; Posé et al., 2012). The rigidity of the cellulose network allows imaging in direct contact between AFM probe and cell wall surface even in the air (Kirby et al., 1996). The bright and dark areas in the topography image correspond to peaks and troughs of the sample surface. A fibrous structure can be seen in both variants (Figure 3). The fibers in cell walls from NH₄⁺ grown plants are more tightly arranged, they run in the same direction, whereas in the case of the NO₃^– treatment the fibers are crooked and scattered.

Ammonium treatment did not affect the level of cellulose, pectins, and lignin in cell walls (Figures 4A,B,D). Further the content of homogalacturonan and neutral sugars in the pectin-enriched fraction showed no statistically relevant changes (Figure 4C), but a slightly higher percentage of neutral sugars in cell walls of NH₄⁺ grown plants could be observed. The total content of phenolics in the cell wall was more than 60% higher in leaves of NH₄⁺ treated plants than in the controls (Figure 4E). More than 70% lower content of Ca²⁺, but 45% higher content of boron was detected in tissues of NH₄⁺-treated plants than in the controls (Figures 4F,G).

Infrared spectra (1,800–800 cm^-1) obtained by FTIR analysis of cell walls isolated from leaves of plants grown on NO₃^– and NH₄⁺ nutrients revealed that the amide I + II region (1,700–1,480 cm^-1) characterized by peaks at 1,627 and 1,523 cm^-1 did not differ significantly in area as well as in the shape (Figure 5A), thereby indicating the lack of striking evidence for alteration in the structural proteins owing to the N source used. The band centered at 1,627 can be ascribed to C=O stretching in the COO^- group of non-methylesterified pectins (Jones et al., 2005). However, the presence of absorbance in the region near 1,523 cm^-1, which is generally absent from the spectra of pectins, evidences the contribution of proteins in the Amide I band. The most prominent FTIR spectral dissimilarities between the plant nutrition variants were observed in the polysaccharide region between 1,200 and 900 cm^-1 (Figure 5B). The major bands located at 1,152, 1,098, and 1,020 cm^-1, can be assigned to C-O-C asymmetric stretching, and C-C/C-O stretching vibrations, respectively. The integrated peak area was lower for the bands at 1,098 and 1,020 cm^-1 in NH₄⁺- than in NO₃^–-treated plants. Differential (NO₃^–-NH₄⁺) spectra (Figure 5C) provided additionally a positive peak at 1,044 cm^-1 and a negative peak at 1,181 cm^-1, assigned to C-O and C-C stretching in the ring of xyloglucan (Kačuráková et al., 2000) and to C-O stretching and/or guaiacyl ring breathing, respectively. Unsupervised multivariate analysis (PCA) confirmed that spectral features in the polysaccharide region distinctly separated the NH₄⁺ and NO₃^–-treated plants in the clustering pattern (Figure 5D). Plants grown on NH₄⁺ showed negative, whereas NO₃^–-treated plants positive scores for the principal component (PC) 1, explaining 89% of the total spectral variability. The PC1 loading factor plot (inset in Figure 5D) showed positive association with peak at 963 cm^-1, additionally confirming its discriminating role between the NH₄⁺ and NO₃^–-treated plants (Figure 5C). This band, however, could not be unequivocally assigned. This band was present both in the NH₄⁺ and NO₃^–-treated plants and was markedly elevated in the latter.

Because the peak at 963 cm^-1 dominated the spectral variability within the polysaccharide range and thus could mask the contribution of other compounds we performed also the PCA for the shifted range of spectra between 1500 and 1000 cm^-1. This spectral region also discriminated the nutrient treatments as revealed by the score plot in Figure 5E. The corresponding loadings (Figure 5F) showed negative correlation for the bands between 1180 and 1130 cm^-1, putatively assigned to C-O stretching of C-OH groups and to C-O-C stretching of glycosidic links and/or to ether-linked phenolic compound. Since the loading feature did not resemble that typical for polysaccharides this data may indicate increased amount of ether-linked phenolic compounds in NH₄⁺ plants compared to the NO₃^– control.

To estimate possible differences in the degree of esterification of pectins, FTIR spectra were performed on the pectin-enriched fraction (Figure 5G). The carbonyl C=O stretching vibration of the ester group at about 1735 cm^-1 (McCann et al., 2001) was characterized by similar absorbance (normalized to the area between 1800 and 900 cm^-1) for both nutrient variants. The NO₃^– grown plants showed a slightly more visible shoulder about 1600 cm^-1 and higher absorbance at 1418 cm^-1, corresponding to vibrations of COO-group of polygalacturonic acids (Séné et al., 1994; Jones et al., 2005). However, the ratio of the absorbance at 1600 to 1730 cm^-1, that is indicative of the degree of pectin esterification (Filippov and Kohn, 1975), did not differ significantly. The analysis of the hemicellulose-enriched fraction (Figure 5H) showed bands at 1150, 1080 and 1035 cm^-1, the latter two are tentatively due to complex C-OH, C-O-C and C-C stretching vibrations, which may be attributed to xyloglucans (Kačuráková et al., 2000). The absorbance area of the band centered at 1035 cm^-1, normalized to the height of the maximum peak for the whole spectrum at 3310 cm^-1 (Figure 5H, inset) was higher for the NH₄⁺ treated plants compared to NO₃^– grown controls (0.0301 vs. 0.0285, respectively), the difference was statistically not significant though. Some interfering impact due to presence of arabinogalactans to the absorbances cannot be excluded.

The major cell wall hemicelluloses are xyloglucan. Immunofluorescence studies displayed the xyloglucan-specific motifs in cell walls recognized by the monoclonal antibody LM25. The level of xyloglucan, indicated by green fluorescence of the LM25-treated sections, was higher in mesophyll leaf cells from NH₄⁺-treated plants than in the controls (Figures 6B,C). In addition, a higher fluorescence signal was observed preferentially in the epidermis of the petiole in NH₄⁺ grown plants (Figure 6A). Regions of highly methylesterified pectins in cell walls were labeled by the LM20 antibody-dependent fluorescence. The LM20 dependent labeling was relatively high in mesophyll leaf tissues during NH₄⁺ nutrition than in the control (Figures 6B,C). Fluorescence showed a similar pattern in the petiole between different N treatments (Figure 6A). Also epitopes of low-methylesterified pectins in cell walls were labeled by the LM19 antibody. The resulting fluorescence in leaf cells showed a similar intensity between different nitrogen growth regimes (Figures 6B,C). Negative controls (fixed micro-sliced sections of Arabidopsis leaves treated with secondary antibodies) showed only lignin autofluorescence.

Activities of cell wall enzymes engaged in pectin hydrolysis were determined. PME is responsible for demethylesterification of pectins, which can be utilized by PG. Because the performance of PME is dependent on apoplastic pH, the activity of enzymes was measured in different pH regimes. The response of PME activity to NH₄⁺ treatment showed a similar trend when using as the substrate pectins with a low (8.9%), medium (30%), or high (90%) degree of methylesterification (Supplementary Figure S1). Therefore, only results for the degree of 30% methylesterification were selected for presentation. The activity of PME in the cell walls of NH₄⁺-treated plants was decreased by more than 60% at pH 5.0 than in the control, but no differences in enzyme activity were observed at pH 6.8 (Figure 7A). At pH 8.5, enzyme activity was below the detection level for all utilized pectins (results not presented). The activity of PG was 3 times lower in NH₄⁺-treated plants than in the controls (Figure 7B).

Moreover, enzymes that are able to cross-link wall polymers were analyzed. The protein level of XTH/XET was unchanged during growth on different N sources (Figure 7E). The activity of POX in cell walls was 30% increased in plants treated with NH₄⁺ than in the controls (Figure 7C). PAL is the main enzyme in the pathway for the synthesis of phenolic compounds. The activity of PAL was more than twofold higher in tissues of NH₄⁺-treated plants than in the controls (Figure 7D).

The expression of some proteins that are engaged in cell wall loosening was determined. The expression of EXP1 and EXP17 was lower in NH₄⁺-treated plants (Figure 8A). However, only the lowered expression of EXP17 was significant as compared to that of the controls. Further, the expression of both PG was decreased in NH₄⁺-treated plants (Figure 8B), but only the fivefold decrease in the expression of the putative PG protein was significantly altered as compared to that of the control.

The expression of two of the cell wall sensor receptor-like kinases was determined. FER is thought to be associated with regulation of signaling during cell elongation. Approximately 40% lower expression of FER was determined in plants growing on NH₄⁺ as the source of N (Figure 9). THE1 is called a cell wall integrity sensor kinase. The expression of THE1 was unchanged between different N-source treatments (Figure 9).