The C, N, and H contents were determined using a FlashSmart elemental analyzer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a thermal conductivity detector (TCD). The samples were burned at 950°C in an oxygen atmosphere (99.999% purity). A reference material, 2,5-bis(5-tert-butyl-2-benzo-oxazol-2-yl) (C = 72.52% ± 0.22%, N = 6.51% ± 0.09%, and H = 6.09% ± 0.08%), was used, with the ranges of the calibration curve being 0.008–0.172 mg for nitrogen (R² = 0.9997), 0.089–1.914 mg for carbon (R² = 0.9996), and 0.007–0.161 mg for hydrogen (R² = 0.9995). Calibration curves were also checked for cysteine (C = 29.98% ± 0.28%, N = 11.66% ± 0.16%, and H = 5.03% ± 0.13%).

The amount of total Si was determined by X-ray fluorescence (XRF) spectroscopy, which was performed using an Olympus VANTA C spectrometer (Waltham, MA, USA) equipped with a 40-kV X-ray tube rhodium (Rh) anode, silicon drift detector (SSD), and 3-mm beam collimator.

The morphology of intermediates and elemental composition was investigated using a scanning electron microscope (SEM) Quanta 200 FEI (Amsterdam, The Netherlands) at 30-kV electron acceleration voltage, large field detector (LFD), and ×200 magnification (Dusevich et al., 2010). The elemental composition was assessed using an Octane Super 60-mm² energy dispersive X-ray (EDX) detector (EDAX, Ametek Inc., Mahwah, NJ, USA).

The biogenic SiNPs were analyzed by transmission electron microscopy (TEM), and the elemental distribution was determined using the scanning transmission electron microscopy (STEM) mode coupled with the EDX detector X-MaxN 80T (Oxford Instruments, Abingdon, UK). TEM analysis was performed on Cu films coated with carbon film (Carbon Type-B, 200 mesh; Ted Pella, Newington, NH, USA) using a TECNAI F20 G2 TWIN Cryo-TEM transmission electron microscope (FEI, Austin, TX, USA) with an electron accelerating voltage of 200 kV. The grids were initially hydrophilized for 10 sec (model 1020; Fischione Plasma Cleaner, Export, PA, USA) in advanced vacuum and a binary gas atmosphere with O₂ (25%)/Ar (75%) (Padgett et al., 2017). This procedure hydrophilizes the support based on carbon film and promotes the uniform deposition of nanoparticles. TEM analysis was performed in bright field (BF) mode.

Dynamic light scattering (DLS) analysis was assessed using AMERIGO™—Particle Size & Zeta potential Analyzer (Cordouan Technologies, Pessac, France), as per manufacturer’s instructions, using the high concentration head, which is based on the Dual Thickness Controller (DTC) technology. The analysis of the results was performed using the AmeriQ 3.2.3.0 software (more details can be found in Data Sheet 3 in the Supplementary Material ).

To quantify the soluble Si released from SiNPs, 17.5 mL ultrapure water (SiNPs-W) or 50 mM NaCl solution prepared in ultrapure water (SiNPs-S) was poured over 18.25 mg SiNPs (the same ratio as in Section 2.4.4). The samples were analyzed both non-ultrasonicated (SiNPs-W and SiNPs-S) and ultrasonicated (SiNPs-Wu and SiNPs-Su) using a probe sonication (Sonics VCX 750, model CV334, Sonics & Materials, Newtown, CT, USA) with the following parameters: 20 kHz, 5 min, 80% amplitude, pulse 15 sec, and pause 30 sec (as in Section 2.4.2). The ultrasonication was performed at the beginning of the experiment. The released soluble silicon was assessed after ultracentrifugation of the samples at 400,000 rcf (CP100NX Ultracentrifuge, Hitachi Koki, Tokyo, Japan). The determination of soluble silicon was performed using the Supelco Silicate Test kit (1.00857.0001) (Uţoiu et al., 2018), with a concentration range of 0.5–500 mg/L Si (MQuant^®Supelco, Merck, Darmstadt, Germany), according to the protocol provided by the manufacturer. The standard curve was carried out starting from a stock solution of 1,000 mg/L SiO₂ (Supelco, Merck, Darmstadt, Germany).

A mixture design was applied with three factors, alginate, glycerol, and sorbitol, using the Design-Expert 11 software for the preparation of films (Trica et al., 2023). The films were prepared as described in Supplementary Table 1 . After drying, the films were tested by dynamic mechanical analysis (DMA) Q800 (TA Instruments, New Castle, DE, USA). Tensile test specimens (L, length; W, width; δ, thickness) were obtained and mounted in the instrument. The traction tests that followed were performed in the following conditions: rectangular geometry, isothermal for 5 min, 30°C, and ramp force 0.2 N/min from 0.001 N to 18 N (Florian et al., 2018). The traction tests yielded Young’s modulus and the ultimate tensile strength (UTS) for each film.

For the dispersion of SiNPs in alginate, we used probe sonication (Sonics VCX 750, model CV334, Sonics & Materials, Newtown, CT, USA) with the following parameters: 20 kHz, 5 min, 80% amplitude, pulse 15 sec, and pause 30 sec (Retamal Marín et al., 2017; Liu et al., 2022).

The mung bean seeds were transferred to a sterile tube. In the first step, the seeds were sterilized with 95% ethanol for 3 min and gently mixed in the tube. After the ethanol removal, 5% sodium hypochlorite solution was added and stirred in the tube for another 3 min. After removing the hypochlorite solution, the seeds were washed with sterile double-distilled water (Sauer and Burroughs, 1986).

The coating process of the mung bean seeds was performed using a MINI-GLATT fluidized bed granulator (Glatt, Binzen, Germany) with some modifications compared to our previous work (Trica et al., 2023). A solution composed of 474.6 mg alginate, 175 mg glycerol, and 260.4 mg sorbitol in 35 mL sterilized double-distilled water was used, according to the optimum ratio between the three components found after the DMA/Design-Expert analysis. Five coating experimental variants were set up using 15 g of mung bean seeds and the abovementioned AGS solution, as follows: 1) Variant V0, AGS solution without SiNPs; 2) Variant V1, AGS solution + 1% SiNPs (w/w of total AGS organic substance); 3) Variant V2, AGS solution + 2% SiNPs (w/w); 4) Variant V3, AGS solution + 3% SiNPs (w/w); and 5) Variant V4, AGS solution + 4% SiNPs (w/w). The SiNP concentration was calculated as percent of the total weight of the organic fraction (910 mg) from the solution used for each coating. The film on each seed contained 1%, 2%, 3%, and 4% SiNPs.

As the process was carried out immediately after seed sterilization, a 2-min drying time of the seeds was set up in the granulator before coating. In addition to the five experimental coating variants, a control experiment with uncoated mung bean seeds, with water instead of the filmogenic solution spraying, was carried out. The settings of the granulator were a peristaltic pump solution flow rate of 4 mL/min, 2-bar spraying nozzle pressure, 40°C coating temperature, 45°C–50°C drying temperature, and 20–40 m³/h air volume for spray-drying. The parameters were adapted from our previously published study by modifying the debit of the pump, the nozzle pressure, and the air volume for spray-drying (Trica et al., 2023). Each coating took place for 15 min, representing the time required for all the AGS ± SiNP solution, or water in the case of control, to be sprayed inside the granulator. In the following, AGS film refers to the film without SiNPs.

Water activity for the six experimental sets was measured (Baldet et al., 2009) using LabMaster-aw neo equipment (Novasina AG, Lachen, Switzerland), as per the manufacturer’s instructions. Three measurements were determined for each experimental set.

SEM micrographs of uncoated and coated seeds were obtained using a TM4000Plus II tabletop electron microscope (Hitachi, Tokyo, Japan) at 5-kV electron acceleration voltage, backscattered-electron (BSE) detector, standard (M) vacuum mode, and ×600 magnification. The elemental composition was determined using an EDX detector AZtecLiveLite (Oxford Instruments, Abingdon, UK). The adjustments were made based on the methodology proposed by the manufacturer. The average width of the film was determined with the ImageJ software (Schindelin et al., 2012; Trica et al., 2023).

Ten uncoated (C), AGS film-coated (V0) mung bean seeds, and mung bean seeds coated with the AGS film embedding different concentrations of SiNPs (V1, V2, V3, and V4) described in Section 2.4.4 were grown in both the absence and the presence of 50 mM NaCl. The seeds were placed in sterilized Magenta™ vessel GA-7 (Sigma-Aldrich, St. Louis, MO, USA), in which autoclaved 0.4% agar ± 50 mM NaCl had been previously poured and then set in a growth chamber with controlled light and temperature (ALGAETRON AG230, Photon Systems Instruments, Drásov, Czech Republic). The conditions for germination and seedling growth were set to a 16-h light/8-h dark cycle with 26°C temperature and 1,340 μE light intensity for the light period and 22°C temperature for the dark period (Trica et al., 2023). The experiment was stopped, and the seedlings were analyzed after 4 days after being placed into the agar medium. The analysis of the seedlings was performed at the end of the experiment (approx. 3-day-old seedlings): plant measurements, alpha-amylase activity, chlorophyll and carotenoid content, hydrogen peroxide (H₂O₂) content, l-proline content, nitric oxide (NO) content, lipid peroxidation [thiobarbituric acid reactive substance (TBARS)], activity of ROS scavenging enzymes, ROS detection and semi-quantitation in mung bean leaves using 2′,7′-dichlorodihydrofluorescein diacetate staining, and H₂O₂ detection in mung bean leaves using 3,3′-diaminobenzidine (DAB) staining. For the other analysis, the specific time point is specified in the corresponding subsection. Each experimental variant was tested in triplicate.

Mung bean seedlings were photographed, and the root and stem lengths were determined using the ImageJ software (Trica et al., 2023).

The investigation of α-amylase activity was performed following Hostettler et al. (2011), with each sample accompanied by a control in which the enzyme–substrate reaction was stopped by introducing the reagent before the substrate and was subjected to the same treatment. For the preparation of plant extracts, a solid:liquid ratio of 2:1 was used (2 g of fine powder was obtained by grounding the seedlings in liquid nitrogen and mixed with 1 mL of double-distilled water). The extracts were vortexed and centrifuged at 20,000 rcf and 4°C for 10 min, and the supernatant was incubated with the substrate (soluble 0.45% w/v starch solution in a 44 mM KH₂PO₄ + 2 mM CaCl₂ solution) in a 1:1 ratio. The samples were incubated for 10 min in a water bath at 30°C, and the reaction was stopped with Bernfeld reagent (Bernfeld, 1951) containing 1% 3,5-dinitrosalicylic acid in a 1:2 extract:reagent ratio. Subsequently, the samples were incubated in a digital dry bath (BSH1004, Benchmark Scientific, Sayreville, NJ, USA) at 95°C for 5 min. After cooling, the absorbance was read at 540 nm using a microplate reader (CLARIOstar BMG Labtech, Ortenberg, Germany) against a standard maltose curve prepared from a stock solution of 1 mg/mL maltose.

Chlorophyll and carotenoid contents were determined using the equations described by Lichtenthaler and Buschmann (2001) for pure methanol solvent. Briefly, 50 mg of mung bean leaves was homogenized with 2 mL of pure methanol. The supernatant was read with a microplate reader for quantification of chlorophyll a, chlorophyll b, and carotenoids after centrifugation of the homogenate at 6,000 rcf for 15 min at 4°C.

The experiment was performed following Zandonadi et al. (2016) with some modifications regarding the agar concentration and the investigation of proton pump activity through image analysis. On the third day of seedling growth, a solution of 0.04 g/L bromocresol purple in double-distilled water was prepared and brought to pH 6.7 with 1 N NaOH solution. After the addition of agar to a final concentration of 0.75%, the solution was autoclaved and poured into sterile Petri dishes with a diameter of 90 mm. Mung bean seedlings were placed with the root partially embedded into the agar gel after the gel reached a temperature of approximately 30°C. The image acquisition was carried out after 24 h. The experimental variants were tested in triplicate. Semi-quantitative analysis of specific extracellular H⁺ level (seH⁺) was performed using the ImageJ software by measuring the area of the yellow region (adjust color threshold, make binary, and remove outliers) and the intensity of the yellow color (live histogram) and the root area and by applying the following formula:

where A_y is the area of the yellow region, I is the color intensity of the yellow region, and A_r is the root area. The total extracellular H⁺ level, i.e., without normalization to the root area, was also calculated.

The metabolic activity during germination was determined according to Pouvreau et al. (2013) with some modifications in order to adjust the experiment for larger seeds. The experiment was conducted in 24-well plates. Three seeds were placed in each well, and each experimental variant was performed in triplicate. For the experimental set without salt and the salt-stress set, 500 µL of sterile double-distilled water and 50 mM NaCl solution, respectively, were pipetted into each well. The 24-well plates were placed into the growth chamber for 24 h using the same parameters as in Section 2.5.1. Next, the seedlings were washed three times with double-distilled water and incubated in the dark with 770 µL of 0.45 mg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution for 24 h at 24°C. The solubilization of the tetrazolium salt was carried out by adding 1.4 mL of buffer prepared from 10% Triton X-100 and 0.04 M HCl in isopropanol and keeping the sealed plates for 24 h in the dark at 30°C and 100 rpm. Afterward, the liquid was transferred to 96-well plates and read at 570 and 630 nm. The absorbance at 630 nm was subtracted from the absorbance at 570 nm for turbidity correction.

l-Proline was assessed after a previously described method (Igiehon and Babalola, 2021) with minor modifications related to the ratio between the seedling and the sulfosalicylic acid, as well as to the centrifugation time. Briefly, 0.35 g of mung bean seedlings previously ground in liquid nitrogen was mixed with 5 mL of 3% sulfosalicylic acid. After 20 min of centrifugation at 1,520 rcf, 2 mL of supernatant was mixed with 2 mL of acid ninhydrin reagent and incubated at 100°C in a water bath (OLS Aqua Pro Shaking Water Bath, Grant Instruments, Cambridge, UK) for 1 h. The samples were transferred on ice for a few minutes, 4 mL of toluene was added, and after vigorous stirring, the samples were left to sit for 30 min in the dark for phase separation. The absorption of the organic phase was read at 520 nm (Ocean Optics UV-VIS-NIR, Orlando, FL, USA).

The NO content was determined according to Liu et al. (2007). Briefly, the mung bean seedlings were homogenized in 40 mM HEPES buffer (pH = 7.2) with a ratio of 1:10 (w/v). After the centrifugation of the homogenates at 5,000 rcf for 10 min at 4°C, the NO content was assessed using the Griess reagent system.

Hydrogen peroxide content was determined according to Önder et al. (2020). Briefly, the seedlings frozen in liquid nitrogen were homogenized with 0.1% trichloroacetic acid (TCA). After centrifugation of the homogenate at 15,000 rcf for 15 min at 4°C, the supernatant was mixed with 10 mM K₂HPO₄ buffer, pH = 7, and 1 M KI. The reaction mix was incubated for 30 min at room temperature in the dark. Next, the samples were read at 390 nm using a microplate reader. The standard curve was subjected to the same steps starting from a stock solution of 9.8 mM H₂O₂.

The mung bean leaves were stained with DAB for the detection of H₂O₂ after a previously described method (Obrien, 2012).

Lipid peroxides were dosed using the TBARS method (Heath and Packer, 1968). After being frozen in liquid nitrogen, 0.1 g of mung bean seedling powder was mixed with 0.1% TCA at a solid/liquid ratio of 1:10. The extracts were vortexed and centrifuged at 20,000 rcf and 4°C for 10 min, and the supernatant was mixed with TBARS reagent (0.5% thiobarbituric acid in 20% trichloroacetic acid) at a 1:2 ratio. The samples were incubated at 95°C for 15 min in the BSH1004 digital dry bath. After cooling the samples on ice, the absorbance was read at 532 nm and 600 nm for turbidity correction. The standard curve was prepared from a stock solution of 200 µM MDA according to Aguilar Diaz De Leon and Borges (2020).

The activity of key enzymes in ROS metabolism was determined according to Chen and Zhang (2016), with some modifications. For enzyme extraction, 0.25 g of mung bean seedlings previously frozen in liquid nitrogen and ground into a fine mixture was mixed with 2 mL of phosphate-buffered saline (PBS), pH 7.8. The samples were vortexed for 3 min and centrifuged at 20,000 rcf and 4°C for 20 min.

Briefly, for 1) superoxide dismutase (SOD) activity, the enzyme extract was mixed in a 1:19 ratio with the reaction mix consisting of 100 mM PBS, pH 7.8, 1 mM EDTA disodium salt, 130 mM methionine, 750 µM nitroblue tetrazolium (NBT), and 20 µM riboflavin. The mixture was transferred to the ALGAETRON light-controlled chamber at 74 µE for 9 min. The absorbance was measured using the microplate reader at 560 nm. The samples were followed by two controls in which PBS was added as enzyme extract replacement. One of the controls was subjected to light at the same time as the samples, and one control was kept in the dark. One unit of SOD activity was expressed as the amount of enzyme that inhibits 50% of NBT photoreduction. The SOD activity was expressed as units (U) per gram of seedling fresh weight instead of per mg protein.

For 2) guaiacol peroxidase (G-POX) activity, the enzyme extract was mixed in a 1:19 ratio with the reaction mix prepared by mixing 56 µL of 0.2% (v/v) guaiacol and 76 µL of 30% H₂O₂ stock solution in 100 mL of 100 mM PBS, pH 7. One unit of G-POX was expressed as the production of 1 nmol of tetra guaiacol/min (extinction coefficient ε = 26.6 mM⁻¹ cm⁻¹ at 470 nm). The absorbance increase at 470 nm was monitored for 5 min in the microplate reader.

For 3) catalase (CAT) activity, the enzyme extract was mixed in a 1:19 ratio with the reaction mix prepared by adding 155 µL of 30% H₂O₂ in 100 mL of 100 mM PBS, pH 7. One unit of CAT (µmol/min) was defined as the decomposition of 1 µmol H₂O₂/min, pH 7, 25°C. The decrease of absorbance at 240 nm was monitored for 5 min in the microplate reader. The extinction coefficient 43.6 M⁻¹ cm⁻¹ of H₂O₂ at 240 nm was considered for calculations.

For 4) glutathione reductase (GR) activity, the enzyme extract was prepared by mixing 0.15 g of mung bean seedlings previously frozen in liquid nitrogen with 1.5 mL of buffer (pH = 7.2, 50 mM Tris-HCl, 1 mM EDTA, 5% glycerol, and 5 mM MgCl₂). The homogenates were centrifuged at 16,000 rcf for 20 min at 4°C. The reaction mix consisted of enzyme extract and buffer (pH = 7.2, 0.05 M Tris-HCl, 0.25 mM NADPH, 1 mM EDTA, and 1 mM oxidized glutathione) mixed in a 1:1 ratio. One unit of GR (nmol/min) was defined as the amount of GR that oxidizes 1 nmol of NADPH/min at 25°C. The absorbance was monitored for 20 min at 340 nm (Labudda et al., 2020).

2′,7′-Dichlorodihydrofluorescein diacetate (H₂DCFDA) was used in order to visualize the intracellular ROS according to Georgescu et al. (2018). Briefly, the mung bean leaves were incubated in the dark at room temperature for 30 min with 2.5 μM H₂DCFDA solution prepared in double-distilled water from a stock solution of 10 mM H₂DCFDA in dimethyl sulfoxide (DMSO). Afterward, the leaves were washed with double-distilled water. Intracellular ROS were visualized using CELENA® X High Content Imaging System (Logos Biosystems, Annandale, VA, USA). Mean fluorescence intensity was measured in the ImageJ software using the mean intensity of the green channel with respect to the RGB color model. For the detection of total ROS in the seedlings, the seedlings were subjected to the same treatment used for leaf labeling. After the washing step, the seedlings were homogenized in PBS solution (1 seedling:5 mL PBS) using a mortar and pestle. The homogenate was centrifuged for 10 min at 4°C and 19,000 rcf. The fluorescence intensity was read using the microplate reader (485-nm excitation and 530-nm emission) and normalized to fresh seedling weight.

For the determination of Si in freeze-dried mung seedlings, 0.2 g of homogeneous powder was used for inductively coupled plasma–optical emission spectroscopy (ICP-OES) analysis. The resulting ash at 650°C was subsequently mineralized with 65% HNO₃ and 30% H₂O₂ (8:2) in a Multiwave 3000 microwave digester (Anton Paar, Graz, Austria). After sample digestion, the obtained solution was transferred to a plastic volumetric flask and brought to a final volume of 25 mL with ultrapure water. The blank sample was prepared under the same conditions (Huang et al., 2004).

In the absence of saline stress, the AGS film (variant V0) significantly increases the shoot and root length, and the addition of SiNPs in the AGS film (variants V1–V4) does not induce other significant changes compared with the V0 variant ( Figure 8A ). The AGS film does not significantly affect the activities of α-amylase, the antioxidant enzymes, SOD and G-POX, but the levels of lipid peroxidation (TBARS) and chlorophyll are slightly decreased; the carotenoids, GR activity, and H₂O₂ level are slightly increased in the absence of salt. SiNPs do not induce further changes in α-amylase, TBARS, G-POX, and SOD, except for the SOD activity, which increases slightly at 3% and 4% SiNPs ( Figure 10C ). The l-Pro is slightly decreased by the AGS film, but not by SiNPs ( Figure 10A ). A lower l-Pro level than for the control is observed in the presence of salt as well.

The most significant effects of the AGS film in the absence of salt are 1) the decrease of the specific extracellular H⁺ level (seH⁺) into the medium, an effect that is alleviated by SiNPs in a dose-dependent manner, reaching control values at 3% and 4% SiNPs ( Figure 9B ), and 2) the increase of NO level. The overall metabolic activity quantified by the MTT assay also shows a decrease in the presence of AGS compared to the control in the absence of salt. There is no further significant change in the MTT assay in the presence of SiNPs. The MTT test is based on the reduction of the yellow-colored compound MTT to a dark-blue formazan (Pouvreau et al., 2013). The reduction of MTT by mitochondrial enzymes (especially succinate dehydrogenase) is an index of cell/mitochondrial integrity.

The most significant effects of SiNPs in the absence of NaCl are on the chlorophyll content, which increases, on the specific extracellular H⁺ level, in which case it reverses the effect of the AGS film at SiNP concentrations higher than 2% in the film ( Figure 9 ), and, on the GR activity and H₂O₂ level that significantly decrease in the presence of SiNPs ( Figure 10 ). There is a hormetic behavior of the carotenoids induced by SiNPs in the absence of salt. With respect to NO, the SiNPs first induce a decrease at 1% SiNPs, followed by an increase at higher SiNP concentrations.

The tested NaCl concentration of 50 mM significantly affected almost all parameters measured, acting as an abiotic stress factor: the shoot length, the root length, chlorophyll, carotenoids, and the metabolic activity decreased. l-Pro, the lipid peroxidation, and ROS level significantly increased. The α-amylase slightly decreased, but not statistically significant.

Under saline stress, most of the biochemical and physiological effects of the AGS film and SiNPs on mung seedlings are much more significant than in the absence of salt. The only parameter that is not changed by AGS or SiNPs compared with the control is the α-amylase ( Figure 8B ). The coating effects on the seedling growth are not as significant as in the absence of salt, and the growth of seedlings does not reach the unstressed control ( Figure 8A ). Starch is a mixture of amylose and amylopectin, which is found as insoluble granules in chloroplasts (temporary storage) and amyloplasts (long-term storage). During the metabolism of starch, metabolic energy is produced by the enzymatic removal of glucose from the non-reducing ends of the amylose and amylopectin cluster, with α-amylase being one of the most important enzymes responsible for starch hydrolysis (Bernfeld, 1951; Hostettler et al., 2011; Tomasik and Horton, 2012). All other parameters, photosynthetic pigments, metabolism, the l-Pro level, lipid peroxidation, the activities of the antioxidant enzymes, and ROS level are significantly influenced by both the AGS film and SiNPs. Lipid peroxidation involves the attack of free radicals against lipids, particularly polyunsaturated fatty acids (PUFAs) (Ayala et al., 2014).

There are several correlations between parameters in both the absence and especially the presence of salt stress, as determined by Pearson’s correlation analysis ( Supplementary Tables 8, 9 ). In the absence of salt, the growth parameters correlate negatively with l-Pro, TBARS, leaf ROS, and total ROS; the pigments correlate positively with eH⁺ and negatively with GR activity and leaf ROS, the last two being positively correlated to each other; MTT correlates positively (+) with seH⁺ and negatively (−) with G-POX; l-Pro, TBARS, and ROS correlate positively; H₂O₂ correlates positively with GR activity. These correlations are in general straightforward: ROS induce lipid peroxidation (TBARS), production of l-Pro, and activation of GR, which induces the production of the antioxidant GSH. Stress and consequently ROS affect growth and metabolism, which are positively correlated with each other. It is worth mentioning that chlorophyll b correlates (+) additionally with seH⁺ (marginally significant), which indicates a higher degree of correlation of chlorophyll b with the metabolism of the seedling than for the other pigments. In the presence of salt, the number of correlations increases. The growth parameters correlate additionally (+) with the photosynthetic pigments, NO, SOD, and G-POX and (−) with H₂O₂ and GR activity; α-amylase correlates (−) with H₂O₂; the pigments correlate additionally (−) with l-Pro and TBARS; SOD, G-POX, and NO are (+) correlated with each other and correlate (−) with ROS and GR. MTT retains only the (+) correlation with seH+, both of them being non-significantly correlated with the other parameters. CAT activity does not correlate significantly with any of the other parameters. The correlations confirm the expected outcome, i.e., the salt stress induces ROS that affect the seedling growth and metabolism, activating the defense mechanism represented by l-Pro, GR activity, and CAT. The film coating efficiently reduces ROS by activating the antioxidant enzymes SOD and G-POX and significantly induces NO production while reducing the l-Pro content and GR activity, whereas the CAT activity remains high when the film with SiNPs is applied. In this context, in addition to the ROS level, increased GR and l-Pro can be seen as markers for saline stress, whereas increased SOD, G-POX, and pigments, accompanied by reduced ROS, indicate biostimulant activity. The CAT activity has a rather heterogeneous behavior.

Most of the effects observed are in accordance with recently reported works involving Si and alginate (Al Murad and Muneer, 2022; Mir et al., 2022; Mukarram et al., 2022) and/or as expected, although there is no similar formulation available until now, to the best of our knowledge. In the case of Si, there is an increased interest in its application under different formulations to mitigate the abiotic stress on plants, including salt stress. Recent studies investigated the effects of either radicular or foliar application on different plant species and cultivars at different stages of development, with limited information on the effects induced from the first stages of development, at least in the case of beans (Farhangi-Abriz and Torabian, 2018; Ahmad et al., 2019; Naguib and Abdalla, 2019; Rady et al., 2019; El Moukhtari et al., 2022; Shen et al., 2022; Ellouzi et al., 2023; Peña Calzada et al., 2023). Dhiman et al. and El Moukhtari et al. recently reviewed the multifaceted and interesting effects of Si on plants and the missing links for fully understanding the mechanism of action (Dhiman et al., 2021; El Moukhtari et al., 2022). Si applied as sodium silicate or commercial chemical SiNPs increased the salt stress tolerance of Lathyrus odoratus by seed priming, inducing lower levels of proline, lipid peroxidation, ROS, and Na⁺ uptake in plants and increasing the antioxidant enzymatic activities and K⁺ and K⁺/Na⁺ ratio (El-Serafy et al., 2021). Similar results were obtained in the case of Hordeum vulgare L (Ellouzi et al., 2023). Most studies, including on mung beans, obtained similar results, except when it comes to the level of proline, which either decreases or increases in the presence of Si. Some other differences compared with other studies, such as the effect on the antioxidant enzymes, were found in the case of cucumber, where the enzymatic activities were decreased by Si compared to control (Gou et al., 2020). There is no systematic experimental comparison between plant species, timing, seedling age, and Si forms of application, to the best of our knowledge, and many studies do not make a comparison between Si effects in the absence and presence of salt stress in order to distinguish between biostimulant and other types of effects. In our case, AGS in particular and also SiNPs decrease the lipid peroxidation and the l-proline content, with saturation at 3% SiNPs. AGS and SiNPs partially compensate the decrease in the metabolic activity and seH⁺, but SiNPs present a maximum at 1%–2% after which the dose becomes too high or other mechanisms compensate and the effect reverses. There is an increase of SOD and G-POX activities induced by AGS and further by SiNPs with saturation after 2% SiNPs. The increase of the antioxidant enzymatic activities is negatively correlated with the ROS level, except for the GR activity, which is positively correlated. Silicon was shown to be involved in the activation of enzymes that lead to ROS scavenging and consequent decrease of lipid peroxides content, improvement of membrane integrity, and prevention of electrolyte leakage (Liang et al., 2003, 2006) and also in increasing the activity of enzymes involved in respiration processes for a higher ATP content to be hydrolyzed by specific enzymes for nutrient uptake (Zhang et al., 2011).

The data indicate that the AGS film and SiNPs act as plant biostimulants when applied as a homogeneous coating film on mung seeds using a fluidized bed granulator [bottom-spray (Wurster) process]. Most effects of AGS and Si are more significant and with a higher degree of correlation under salt stress than under normal conditions in our study.

The decrease of seH⁺ and overall metabolic activity induced by AGS alone in the absence of salt and the decrease of seH⁺ by NaCl was a somewhat surprising result. Some of these data cannot be compared with other studies due to the lack of information in this respect. Our hypothesis is that we could have in fact several types of actions induced by AGS that overlap, increasing the complexity of the system: 1) a possible initial effect of (osmo)priming induced by the film before germination, probably correlated with the biostimulant effect observed; 2) a weak stressor effect during germination induced by the film; and 3) a possible plant growth promotion effect of the film.

Due to their capacity to reduce the water potential, sorbitol and glycerol are two of the organic compounds used for seed osmopriming, by which seeds are treated with these osmolytes and then washed, and the seeds are dried before germination. In this way, the pre-germination metabolism is activated by small doses of available water, and the germination, plant growth, and tolerance to stress are improved (Melekote Nagabhushan et al., 2022). As the equilibrium between the positive and negative activities of these polyols depends on their concentration, usually, concentrations up to a maximum of 100 mM, depending on the context, are used for obtaining osmoprotectant effects. Higher concentrations can exert toxic effects, and the osmolytes can be used to simulate drought stress, but the effects are not as pronounced as those of PEG, and sorbitol was reported to upregulate the antioxidant enzymes in rice leaves (Hsu and Kao, 2003). In some cases, negative and positive effects were reported in the absence and presence of other stress forms, respectively (Wang et al., 2022).

The positive outcome observed in our study on the plant development indicates an (osmo)priming effect of the AGS film. After germination, the AGS film is dissolved in the medium, and the organic components probably diffuse slowly and gradually through the agar medium away from the seed. Therefore, we believe that the direct effect should be less pronounced than the (osmo)priming effect. All three components of AGS could contribute to this priming effect, the organic compounds from the film by inducing a low external water potential (osmopriming) and alginate additionally by the presence of small quantities of Na⁺, acting as a weak stressor. Seed priming by osmotic adjustment (osmopriming) is already a commercially available method and was recently reviewed (Melekote Nagabhushan et al., 2022).

The inhibition of the metabolic activity could be the result of a certain degree of stress level of the AGS film (maybe a combination of osmo-stress and Na⁺ stress from alginate), but the stress does not surpass the beneficial effects. In the presence of salt, the AGS film has a different behavior, slightly alleviating the metabolic inhibition induced by salt. One explanation could be that the film, especially alginate, retards the Na⁺ diffusion to the plant cells. More optimized compositions in the future, depending on the contribution of each component, could have even better effects.

Proton pumping seems to be involved in osmoregulation, and exogenous osmolytes have been shown to affect membrane permeability, activity of proton pump enzymes, and proton efflux (Reinhold et al., 1984; Lew and Spanswick, 1985; Van Bel and Patrick, 1985; Shanko et al., 2003; Chiu et al., 2006), but the subject has not been sufficiently investigated. The effects of osmolytes on proton pump enzymes investigated in vitro seem to be rather an indirect process resulting from changes in membrane structure (Chiu et al., 2006). Proton pump activity can be considered a marker of membrane integrity and plant metabolism status. In order to maintain the transport of different mineral nutrients in symport with H⁺ ions, proton pumps use the ATP energy (Zhang et al., 2011; Cosse and Seidel, 2021).

The effects of salt on the activities of proton pumps (PPs) either from the plasma membrane (PM) or from tonoplast (T) are in some cases contradictory. Most of the studies available until now reported stimulation of PM and T proton pumps (Kłobus and Janicka-Russak, 2004; López-Pérez et al., 2009; Janicka-Russak et al., 2010, 2013), but some studies reported no effect (in the case of salt-resistant plant species) or inhibition of proton pumps, which was time-dependent in some cases (Shan'ko and Babakov, 2002; Zörb et al., 2005; Kabała and Kłobus, 2008; Pitann et al., 2009; Wakeel et al., 2010). Tonoplast communicates with the plasma membrane through the cytosol, and the processes in one can affect the processes in the other (Horaruang et al., 2020). We have to emphasize here that most quantitative PP activities have been investigated in vitro, either in extracted vacuoles or in plant tissue sections. There are no quantitative assays available for in vivo estimation of PP activities, but only qualitative (the acidification of agar medium) on which our semi-quantitative assay is based as well. This assay, based on medium acidification, reflects the interplay between the activities of several ion transporters (proton pumps and Na⁺/H⁺ exchangers), membrane passive diffusion, and other in vivo processes, which could affect the H⁺ equilibria. The assay has not been tested before for investigating the effects of salt stress or osmolytes, to the best of our knowledge. Under high levels of exogenous Na⁺ ions, both the proton pumps and Na⁺/H⁺ antiporters are upregulated (Miranda et al., 2017). The PM proton pumps expel H⁺ from the cells and maintain the PM potential gradient necessary for the Na⁺/H⁺ antiporter (SOS1) to efficiently exchange extracellular H⁺ with intracellular Na⁺.

Previous studies suggest that an excess of salt inhibits the proton pump due to an increased level of ROS and lipid peroxides that have a negative effect on membrane integrity (Liang et al., 2006; Cuin et al., 2011). Moreover, continuous active extrusion of Na⁺ would increase the gradient, leading to a more and more energetically expensive process cycle that in the long run will amplify salt toxicity and irreversibly affect the plant. Halophytes might be able to better adapt and tolerate salt stress by being more capable of reducing the influx and/or sequestering Na⁺ in vacuoles rather than expelling Na⁺ through the plasma membrane (Cuin et al., 2011).

In our study, we established a semi-quantitative in vivo-induced seH⁺ independent of root area, using medium acidification and ImageJ analysis. Both the AGS film and NaCl induced inhibition on the specific activity when applied individually, but the total acidification of the medium was inhibited only by NaCl due to the reduced root area. The AGS has a positive effect on root length, which compensates for the reduced seH⁺, and as a consequence, the overall medium acidification is similar to the control. In the case of the AGS film, the small quantities of Na⁺ ions present in alginate could be partially responsible for the reduced seH⁺ observed. We propose that the extracellular H⁺ level is the result of the fine regulation between proton pumps and Na⁺/H⁺ antiporter, i.e., the level by which each of them is stimulated and the new equilibria that are established in time. In our case, the intracellular uptake of H⁺ through the antiporter seems to overpass the extracellular efflux by proton pumps in the presence of the AGS film or NaCl. Our hypothesis is that during the initial phases of salt stress, the plant is able to keep the pH gradient necessary for Na⁺ efflux from the cells, but as the stress is prolonged or becomes more severe, the pH gradient is reduced probably because of changes in transporter activities. There is a need for more in-depth studies and more precise in vivo assays in order to accurately establish the overall effects. Mathematical modeling and system biology could also significantly contribute in this respect (Horaruang et al., 2020).

Alginate is usually applied to plants as alginate oligosaccharides and more often by foliar spraying than by root treatment. Alginate oligosaccharides can be absorbed in plants and act as internal modulators of plant metabolism. These applications usually lead to increased metabolic activity and increased l-Pro levels. Only a few reports on alginate used as seed coating exist until now, and not all of them refer to metabolic activity. Therefore, a direct comparison with other studies is difficult to make.

In our case, the short time frame of the experiment (4 days) probably resulted in only small quantities of alginate oligosaccharides, considering also the lack of microorganisms in the agar medium and previous reports that claim it takes 120 days for 20% of alginate-based hydrogels to be decomposed in soil (Song et al., 2020). Therefore, if alginate contributes to the coating effects observed, the mechanism could be different than the mechanisms reported for alginate oligosaccharide-based bioproducts, especially since the latter have been mainly applied foliarly. As alginate resides close to the seed during germination and growth due to its application as a coating film and swells in the presence of water, it could retain water close to the seed, at least during the first stages of germination. How much water was available for the seed was difficult to estimate. The observation that the AGS film induced an increase in the root length without an increase in the α-amylase activity, in the absence of salt, can be explained and supports the idea that the alginate polymer acted mainly by enhancing the water-holding capacity of the seeds rather than inducing α-amylase activity. The seed germination and radicle growth depend on water availability (Bewley, 1997). Water-absorbent polymers regulate water availability to seeds and water uptake (Pedrini et al., 2017) and improve seedling/radicle growth (Pathak and Ambrose, 2020; Amirkhani et al., 2023). Alginate-derived oligomers enhance germination by amylase induction (Hu et al., 2004). The effect on the root length under salt stress was only marginally significant and more correlated with the α-amylase activity than in the absence of salt, for both the AGS film and SiNPs. In the case of alginate, the specific contribution each of the of priming/biostimulant, weak stressor, and/or plant growth promoting activity to the overall alginate effect is less evident. Nevertheless, it would imply that alginate could have positive effects on plants even in the absence of generation of significant amounts of oligosaccharides. The few previous studies that applied intact alginate for a short period of time support this hypothesis, with alginate having, for example, positive effects on sweet corn (Z. mays var. saccharata) germination in the absence or presence of osmotic stress (Behboud et al., 2021).

A recent study used alginate films reticulated with Ca²⁺ ions to coat bean seeds (Phaseolus vulgaris L.) (De Castro et al., 2020). In comparison to our film coating, the crosslinked film delayed seed germination, which indicates that reticulation has important disadvantages. They speculated that the delay observed is related to the film forming a barrier that decreases the absorption of H₂O and the passage of O₂. STEM EDX analysis revealed some contents of Ca²⁺ in the alginate used for SiNP dispersion and for coating ( Figure 4F ), but the amounts are not sufficient to significantly reticulate the film. Although our film is probably able to quickly break down at germination because it is not significantly reticulated, we believe that it interacts with and slows down the transport of exogenous Na⁺ during germination and seedling growth process, therefore slightly alleviating the drastic decrease in metabolic activity induced by salt. The mechanism could involve changes in the surrounding environment of the roots and would be different than that of alginate oligosaccharides that are able to be transported into plants, influencing the physiology of the plants.

Our results support the hypothesis that in the absence of stress, alginate acts as a proton pump inhibitor. One explanation could be the formation of a polymeric barrier consisting of chains of β-d-mannuronic (M) and α-l-guluronic (G) acids linked by glycosidic bonds 1→4, monomers that are sequentially distributed in each repetitive subunit (Loureiro Dos Santos, 2017). In the presence of Ca²⁺, alginate forms a hydrogel with an open network, in which divalent cations interact cooperatively and ionically with the G monomers (Bierhalz et al., 2014). Therefore, this polymeric barrier could inhibit the proton pump conformational changes and the homeostasis of the whole system. Additionally, a hypothesis for counteracting the effect produced by alginate, by including SiNPs in the seed coating, could involve the intercalation polymerization of silicic acid and the formation of large chains that lose the ionic bonds established between Na⁺ ions and alginate, and dislocate Na⁺ ions, acting as a crosslinker that creates gaps in the polymeric network, allowing the pump to resume its activity.

With respect to l-Pro, the induced decrease by AGS could have two explanations: 1) osmoprotection by the sorbitol and maybe glycerol uptake from the medium into the seedlings, so the seedlings do not need a high level of l-Pro, and 2) the form of alginate application, e.g., as intact alginate and application at seed. As mentioned above, alginate could act directly on the availability and transport of sodium ions and therefore could reduce the saline stress, therefore reducing the l-Pro needed to protect the plant. The application of alginate oligosaccharides usually leads to an increase in l-Pro levels. It is also possible that alginate alone could increase l-Pro and that this effect is overcome by the opposite effect of sorbitol and glycerol.

An alternative or simultaneous mechanism in the case of l-Pro could be related to the time frame of the treatments, the development and metabolic stage of the seedlings, and the type of the seeds. This hypothesis comes from previous and present observations related to the effect of Si on l-Pro levels in the presence of saline stress, which could also apply to alginate. The influence of Si on l-Pro is somewhat controversial. In the literature, there are reports of both increased and decreased accumulation of l-Pro following treatment with Si under salt or drought stress (Zhu et al., 2020; Basilio-Apolinar et al., 2021). Research on okra (Abelmoschus esculentus), Glycyrrhiza uralensis, P. vulgaris, and Capsicum annuum found that Si protected the plants against saline stress by inducing higher content of free l-Pro (Soares Pereira et al., 2013; Abbas et al., 2015; Rady et al., 2019; Cui et al., 2021), but the decreased level of l-Pro induced by Si was reported for wheat and soybean (Lee et al., 2010; Bybordi, 2015; Singh et al., 2022). In a recent study, the modulation of l-Pro by Si in cucumber seedlings was found to be time-dependent, i.e., L-Pro increased compared to control during the first days after NaCl + Si treatment, and started to decrease 6 days after the treatment (Zhu et al., 2020). The authors speculated that Si first acted by one mechanism, boosting the l-Pro content to alleviate the salt stress, after which, probably because of the alleviation of stress injury, a new metabolic equilibrium might be established to keep l-Pro at adequate levels in plant tissues. The decrease in l-Pro could be also an outcome of the inhibition of Na⁺ and Cl⁻ uptake by Si, an inhibition reported in previous studies (Shi et al., 2013; Zhu and Gong, 2014; Al Murad and Muneer, 2022). l-Pro, an essential amino acid, is involved in various signaling pathways under stress conditions (Liang et al., 2013), in addition to its osmoprotection function. In our case, alginate and SiNPs induced a decrease of l-Pro in mung beans 4 days after germination, when the treatments were applied before germination. Therefore, we propose that there is more than one mechanism of action for alginate and Si and that multiple parameters are involved and have to be taken into consideration: plant type, mode of biostimulant application, plant development stage, time frame of biostimulant action, etc.

Complex behavior can be found also in the antioxidant enzymes. SOD, G-POX, and CAT are some of the main enzymes involved in ROS scavenging. SOD leads to the production of hydrogen peroxide through the dismutation of the superoxide anion. Hydrogen peroxide is subsequently converted by G-POX and CAT into water and oxygen (Liang et al., 2003). GR, however, contributes to maintaining ROS homeostasis by keeping the balance between the oxidized (GSSG) and reduced (GSH) forms of glutathione, thereby using glutathione for parallel scavenging (Couto et al., 2016). As several enzymes are implicated in keeping ROS levels under control, their gene expression is tightly regulated, and each enzyme is more or less active depending on the context. In the literature, there is no homogeneous trend in the behavior of the enzymes, and it is difficult to model all situations in a general systemic model due to the multitude of different conditions. The antioxidant system is in continuous dynamics, and processes such as germination and seedling establishment impose significant changes (Chen and Arora, 2011). (Osmo)priming and biostimulants have significant effects on the antioxidant enzymes. Why certain enzymes are sometimes activated while others with similar functions are inhibited is not yet very clear at the mechanistic level.

Most of our data are as expected; almost all activities are kept at the base level of the control irrespective of coating in the absence of salt. The reduction of H₂O₂ in the absence of salt did not need the activity of catalase; this task is accomplished by GR, which maintains a high GSH/GSSG ratio, G-POX, and probably other enzymes involved in the glutathione–ascorbate cycle such as ascorbate peroxidase. Moreover, from Pearson’s correlation, the H₂O₂ level was significantly correlated with the GR activity. Interestingly, salt stress activated catalase and slightly inhibited G-POX. This is probably because catalase is much more efficient than G-POX, reducing two molecules of H₂O₂ in one reaction. Catalase was reported to be the most responsive enzyme at pretreatment of maize with H₂O₂ (Gondim et al., 2012), indicating that high levels of H₂O₂ need high expression of catalase. Increased CAT activity was obtained previously for six genotypes of mung bean treated with 100 and 150 mM NaCl, and the CAT activity decreased at 200 and 250 mM NaCl (Alharby et al., 2019). In another study, 150 mM NaCl induced inhibition of all enzymes in mung seedlings when analyzed after 18 days of seedling growth (Dong et al., 2022), which could suggest that the plants are severely affected at high and/or long salt stress.

The reduction of CAT and increase of G-POX activity in AGS-coated seedlings compared to control, under salt stress, could be related to the proposed reduced availability of Na⁺ ions to plants, (osmo)priming effects, and/or biostimulant effects of AGS. Practically, the seedling condition seems to almost return to the control in the absence of salt. SOD seems to be more sensitive, as this enzyme is more responsive to the positive effect of AGS than the other two enzymes. This is probably because superoxide is more harmful than H₂O₂, and it is critical to be kept under control. SOD is responsive to the highest doses of SiNPs tested (3%–4%) in the absence of salt ( Figure 10D ). An intriguing effect is the reduction of the GR activity by SiNPs. Previous studies reported a decrease, an increase, or no change of GR by Si in plants at different development stages and subjected to abiotic stress, such as salt, draught, and/or heavy metals (Siddiqui et al., 2014; Zhang et al., 2017; Das et al., 2018). In fact, most antioxidant enzymes have a heterogeneous behavior previously reported by various studies, and the exact causes and mechanisms behind these observations are not fully elucidated. In our case, future planned, more in-depth investigations hopefully will give more answers. As mentioned before, SiNPs were very efficient in stimulating all the other enzyme activities and reducing the ROS levels.

Considering the sizes of the SiNPs obtained in this study, the effects observed were probably induced mostly by the released Si(OH)₄, as observed in Figure 5 . As mentioned in the Introduction, the classical belief that only the soluble forms of Si can be translocated to plants has been challenged by new studies that propose translocation of small NPs as well (Kurczyńska et al., 2021). The average sizes of our SiNPs would make most of them difficult to translocate to the seedlings. Nevertheless, we proved that the seedlings do get enriched in Si upon coating the seeds with SiNPs compared with the controls without SiNPs ( Table 2 ); therefore, the translocation definitely exists under one form and/or another.

There is a need for more in-depth studies in order to fully understand the mechanism of the different components of our tested formulation. Transcriptomics and proteomics are promising tools that can provide information on the interactions and effects at the molecular level (Dhiman et al., 2021), which we plan to address in the following studies. It is also important to continue the investigation in real-life conditions such as the coated seeds planted on a growth substrate, grown in greenhouses or on fields, in order to assess the full effects of bean production yield and quality under the possibility of having multiple abiotic and biotic stress forms. As we have recently shown, we are able to control the thickness of the film (Trica et al., 2023), and optimizing the thickness, but also the SiNP dose, might give an even better output than the one presented in the current study.