Silicon modifies C:N:P stoichiometry, and increases nutrient use efficiency and productivity of quinoa

Recognizably, silicon has a beneficial effect on plant growth and productivity. In this respect, it is also known that the C, N and, P stoichiometric ratios and nutrient conversion efficiency allow identifying the interactions between elements while helping to understand the role Si plays in plant growth. This study aims to investigate whether increasing Si concentrations (0, 1, 2, and 3 mmol L−1) supplied in the nutrient solution is uptaken by quinoa, modifies the C:N:P stoichiometry while increasing nutritional efficiency and crop productivity as well. Our results revealed that the Si supply by promoting a decline in the C levels, associated with greater uptake of N and P, especially decreased the C:N and C:P ratios, favoring the C metabolism efficiency, and modulated the N and P use efficiency for biomass accumulation. This improved nutritional performance and greater use efficiency of C directly favored quinoa productivity. The future perspective is to encourage new field studies with this species to adjust silicon fertilization management to different soils aiming at enhancing quinoa productivity on a sustainable basis.


Results
Si accumulation. As the supplied Si concentrations increased, Si accumulation increased linearly, based on the dry mass of different plant organs. For the 3 mmol L -1 Si concentration, Si accumulation per plant reached 175, 104, 79, and 49 mg in the leaf, grains, root, and stem, respectively. Therefore, at the highest Si concentration, the Si accumulation in the organs decreased as follows leaf > grain > root > stem (Fig. 1).

Concentrations of silicon and nutrients.
The studied Si concentrations of 0, 1, 2, and 3 mmol L −1 are referred to as Si-0, Si-1, Si-2, and Si-3, respectively. Si supplied in the nutrient solution increased significantly (p < 0.05) the Si concentration in the evaluated plant organs (Supplementary Table S1).
Also, the supplied Si changed the carbon (C), nitrogen (N), and phosphorus (P) levels distinctly in the several plant organs. As the supplied Si increased, especially Si-3, the C content decreased by 11, 4 and 5%, in the root, stem and leaf, respectively (Fig. 2a, Supplementary Table S1).
The results indicated that for all the plant organs, except for the stem, the concentrations of Si-2 and Si-3 promoted a 13 and 15% increase of N levels, respectively. The applied Si increased P content up to 19% in leaf tissue, while P availability decreased in the root and stem for Si-3, with the highest levels recorded for Si-0 (Fig. 3, Supplementary Table S1). C:N:P stoichiometric responses. Si application significantly altered all stoichiometric ratios (C:N, C:P and N:P), distinctly for each plant organ (Fig. 2b, Supplementary Table S1). Particularly for Si-2 and Si-3 concentrations, the root tissue exhibited a higher C:P (13 and 11%, respectively) and lower C:N (21 and 23%, respectively) stoichiometric ratios compared to the control (Si-0).
The C:N and N:P ratios in the stem decreased and increased by 12 and 16%, respectively. Further, in the leaves, the C:N ratio varied slightly (from 7 to 6) while the C:P ratio had a sharper decrease (from 47 to 36) as the applied Si concentrations increased. In the grains, the C:N ratio decreased by 14% between the highest Si concentration (Si-3) applied and the control (Si-0). The average values of the C:N and C:P ratios determined per plant organ decreased as follows root > stem > grains > leaf.  www.nature.com/scientificreports/ Nutrient use efficiency. Figure 4 shows that the C, N and P use efficiency increased following a quadratic fitting (Supplementary Table S2), for the supplied Si concentrations.
Biomass production. The applied Si affected positively the biomass production of the shoot (stem and leaves), following a quadratic fitting and, consequently, a linear increase in the biomass of quinoa grains (Fig. 5, Supplementary Table S2). The Si application increased the grain biomass by approximately 45%.

Discussion
The Si supplied in the nutrient solution via root increased Si accumulation in quinoa leaves (Fig. 1). Likewise, there are reports in the literature on the increasing Si levels in the leaves of crops such as wheat 5 , tomato 25 , and sugar cane 26 for Si applied via root. The quinoa plants were able to transfer close to 43% of the total Si absorbed to the leaves, while the root and stem concentration was kept at relatively low levels. This higher Si level in the leaves compared to the root indicates that quinoa should be considered a Si-accumulating plant since it was able to accumulate more Si in the leaves than in the roots.
The results indicate that Si content in the quinoa leaves varied from 2.1 to 9.4 g kg −1 (Supplementary Table S1) depending on the applied Si concentration, this accumulation below 10 g kg −1 classifies quinoa as an intermediate www.nature.com/scientificreports/ Si accumulator 27 . Previous studies have shown that the Si uptake, as a monosilicic acid, by the roots of eudicots plants occurs passively, transported in the transpiratory flow, and permanently deposited as hydrated amorphous silica in the leaf tissue 3 . Thus, the supplied Si accumulated especially in the quinoa leaf tissue due to the intensity of the transpiration rate in the leaves and the balance between the Si available in the nutrient solution and the element demand by the plant 28 . The varying Si accumulation in quinoa grains may also depend on the level of Si available in the nutrient solution. Additionally, it has been suggested that the varying Si contents in grains can be genetically controlled 5,29 . We believe that this unprecedented result for quinoa should promote further research on the species and its interaction with Si. The increased Si availability in quinoa organs alters the C, N and P contents, similar behavior has been reported for other plant species in studies that proved that Si partially controls the C flow while altering the N and P levels 8,30,31 . Our results indicate that increased Si supply reduced C content in most plant organs (Fig. 3), thus indicating a negative correlation between Si and C concentrations. The low C content in the tissues may indicate that the Si incorporated in the vegetal tissues replaced part of the C in the organic compounds of the cell wall, such as lignin and cellulose 7,32 . The energy cost for incorporating Si in structural compounds is less than the inclusion of C in organic compounds, due to the intrinsic permeability of the lipid bilayers 33 , also contributing www.nature.com/scientificreports/ to better plant growth. The decreasing C levels in different plant organs have not only been confirmed under normal growing conditions but also in plants grown under abiotic stress conditions, such as salinity 12,34 and water deficit 35 , which induces greater energy expenditure by the plant and Si can mitigate this consumption, benefiting these plants.
The partial control of the C flow by Si modified the N and P concentrations, therefore, more N was available in the plant organs, except for the stem (Fig. 3). Our results demonstrate that the N content in the plant organs is affected by the supplied Si concentration, given the large difference observed between the moderate (Si-2) and high (Si-3) concentrations (Fig. 2a, Supplementary Table S1). In this sense, a greater Si availability can improve certain photosynthetic parameters and promote N uptake, as indicated in previous research 26 , contradicting other results that reported a negative correlation between N and Si in the leaf tissue 36 . Also, it is possible to suggest the possible role of Si in reducing the permeability of the plasma membrane, increasing root activity, and resulting in higher nutrient uptake 33 .
Furthermore, P concentrations increased only in the leaves with added Si (Fig. 3), suggesting a possible nutrient translocation from the root and stem to this tissue. A similar result was also reported for wheat 5 and sugar cane 35 , where adding Si in the nutrient solution increased P availability in the leaves while Si involvement in the metabolism of C and P was indicated, with subsequent changes in the nutrient stoichiometry and use efficiency. Studies in the literature suggest that Si changes the positive regulation of the expression of transporter genes for P uptake while increasing the exudation rates of organic acids that play a role in improving P uptake and availability 20 .  www.nature.com/scientificreports/ Changes in nutrient concentrations interfered with the nutrient ratios in the plant. Stoichiometric ratios of nutrients are the object of study because the nutrient concentration in plant tissues is affected by biomass production; also interfering in ecological interactions due to the important role of these nutrients in the biological and biochemical functions of plants 37 . Our results indicated that Si replaces and decreases C, increases the N availability, while the P concentrations differ for each plant organ. This result indicates that the C:N and C:P ratios changed in the different organs, while the N:P ratio was more representative of the grains (Fig. 2b). The C:N ratio in the different tissues varied because the applied Si caused the C levels to decrease followed by increasing N contents. However, the decrease of the C:N ratio in the leaves was more pronounced because the C concentration in this organ decreased more significantly, while no significant trend was observed in the grain C contents. The general decreasing pattern of the C:N ratios due to increasing Si concentrations may be related to the decreasing C levels in structural compounds, as previously indicated 7,8 .
The C:N and C:P ratios represent the plant capacity for photosynthetic fixation of C through the availability of N or P 38 . Therefore, the higher N content in the leaf tissue demonstrated its role in the photosynthetic activity of fixing the C that forms part of the plant organic compounds 18 . The supplied Si decreased C while increasing P content, thus decreasing the C:P ratio in the leaves (Fig. 2b). However, the N:P ratio changed especially due to the greater availability of N in the leaf and grains, interfering with plant biomass production and composition, as previously indicated 9 . The lower C:N and C:P ratios point to the role Si plays in maintaining the balance between C and nutrients, N and P 39 . To this end, the increased Si uptake and accumulation contributes to maintain and regulate the homeostasis of stoichiometric compositions in tissues, increasing the structural compounds of C with important metabolic function by favoring the physiological and biochemical processes of the plant.
Our research indicates that a better understanding of the Si effects on quinoa growth and stoichiometric C:N:P ratios requires considering the different responses to the nutrient use efficiency (Fig. 4). The use efficiency of C increased from 0.087 ± 0.003 (Si-0) to 0.120 ± 0.003 (Si-1), 0.154 ± 0.002 (Si-2) and decreased to 0.151 ± 0.00 (Si-3). This result is attributed to higher Si incorporation by the plant tissues that acted mainly in the photosynthetic process instead of replacing the C of the organic compounds 7,35 . The supplied Si also affected N use efficiency by increasing biomass production when N concentration in the plant organs was higher as well. The N use efficiency increased from 0.97 ± 0.03 (Si-0) to 1.30 ± 0.05 (Si-1), 1.56 ± 0.05 (Si-2), and decreased to 1.43 ± 0.02 (Si-3). This increased N use efficiency probably resulted from changes in the primary metabolism, which stimulated the translocation of amino acids to the absorbing tissues 40 , supported also by the greater equilibrium in the composition of amino acids in the quinoa tissues 41 . Similar results have been previously reported for wheat plants 5 , indicating a significant increase in biomass production. The P use efficiency increased from 8.50 ± 0.24 (Si-0) to 10.80 ± 0.33 (Si-1), 14.88 ± 0.29 (Si-2), and decreased to 13.15 ± 0.09 (Si-3). This referred increase may have been due to the reduced energy cost in the substitution of Si for C in organic compounds.
Hence, the results show that Si improved the C, N, and P use efficiency since the plant produced more biomass per unit of N and P absorbed and reduced the C:N and C:P ratios, allowing us to accept our hypothesis that Si increases the use efficiency of C, N and P. Si promoted high nutritional efficiency with contributions from physiological processes because it is known that Si increases photosynthetic efficiency and water use 42 . Therefore, the greater N availability caused biomass production to increase in plant organs related to growth and storage such as leaves and grains since, in these tissues, C compounds were partially replaced by Si, in addition to P translocation and accumulation; these results are supported by previous research in other species 5,7,26 . Thus, the Si absorbed by the quinoa plants changed the C:N and C:P stoichiometric ratios, especially, by increasing the use efficiency of C, N, and P with a subsequent increase of biomass and productivity observed for the added 2.7 and 3 mmol L −1 Si concentrations, respectively.
In conclusion, these results demonstrate that quinoa Si uptake and accumulation decreases as follows leaf > grain > root > stem, whereas Si availability in the plant increases the biomass and grain productivity by changing the nutrient stoichiometric ratios, i.e., decreasing C concentrations and improving P and N use efficiency. Therefore, highlighting the biological importance of Si to quinoa plants, which are very tolerant of different stresses and present a high nutritional value, we can infer that possibly the plant defense mechanisms are strongly associated with Si, thus, indicating the need for further studies with this species on different topics related to plant nutrition.

Methods
Experimental conditions. The experiment was conducted in a hydroponic growing system in a greenhouse at the São Paulo State University, in Jaboticabal, Brazil. Seeds of quinoa cv. BRS Piabiru 43 , were obtained from the Brazilian Agricultural Research Corporation of the Ministry of Agriculture, Livestock and Food Supply, Brazil. This research was not conducted with endangered species and was conducted in accordance with the is in accordance with the Declaration of IUCN Policy on Research Involving Endangered Species. During the experiment, the maximum and minimum temperatures and relative humidity were measured daily using a thermohygrometer (Supplementary Figure S1). The experiment was carried out as randomized blocks, with four treatments and five repetitions. The Si concentrations of 0 (Si-0), 1 (Si-1), 2 (Si-2), and 3 (Si-3) mmol L -1 were supplied in the nutritive solution via root, using potassium silicate and sodium stabilized with sorbitol as Si sources [Si = 107.9 g L −1 ; K 2 O = 16.44 g L −1 ; Na 2 O = 60.7 g L −1 ]. The K concentration in the nutrient solution was adjusted with potassium chloride in all treatments.
Ten quinoa seeds were sown in plastic pots containing 6 dm 3 of medium-grained sand, which was previously washed with running water, and then with deionized and distilled water 44 . As plants developed, only one plant per pot was maintained. After emergence, Si was added to the modified nutrient solution 45 , Fe source changed from Fe-EDTA to Fe-EDDHMA. The nutrient solution pH was adjusted to 5.5 ± 0.2, using a 1.0 mol L −1 solution of either HCl or NaOH. Sorbitol plays an indispensable role as Si stabilizer in aqueous solutions by decreasing www.nature.com/scientificreports/ the polymerization process at pH below 7. The ionic concentration of the nutrient solution was increased during the culture cycle, starting with 10% during the first 10 days after emergence, followed by 25% in leaf development (up to six true leaves), 50% since the emergence of inflorescence, 80% in flowering and 100% in fruit development until the end of the experiment 46 .
Plant analysis. When the plants reached physiological maturity, the biomass was separated into root, stem, leaves, and grains. All plant tissues were washed in deionized water with a detergent solution (0.1%), HCl solution (0.3%), and deionized water. The plant material was dried in a forced-air circulation oven at 65 ± 5 °C to constant weight, and dry weight was determined. The C and N contents in the dry mass were determined by the dry combustion method (1000 °C) in an elemental analyzer (LECO Truspec CHNS) calibrated with the LECO 502-278 wheat standard (C = 45.00% and N = 2.68%). Total P concentrations were measured using the molybdenum antimony colorimetric method 47 . The Si content was determined by wet digestion, hydrogen peroxide (H 2 O 2 ) and sodium hydroxide (NaOH) were added, and the reaction was induced in an autoclave at 123 °C. The silicon was determined by the colorimetry method 48 with hydrochloric acid, oxalic acid and ammonium molybdate in a spectrophotometer at 410 nm.
From the nutrient concentrations in the plant tissue (g kg −1 ) and dry matter (g per plant), the nutrient accumulation was calculated and expressed as mg per plant. The C:N:P stoichiometric ratios were determined in different plant organs and expressed in g kg -1 . Nutrient use efficiency was calculated as the total plant dry weight (TDW) divided by the C, N and P contents (g TDW mg −1 of each element) 49 : Statistical analysis. After testing normality (Kolmogorov-Smirnov test) and homogeneity of variance (Shapiro-Wilk test), all data were submitted to analysis of variance and, when the F test was significant (p < 0.05), the data were fitted to the linear or quadratic polynomial regression model opting for the model with the highest coefficient of determination (R 2 ). Pearson's correlation test was performed for the sample size n = 20. The statistical analyses were performed using SAS Version 9.1 (SAS Institute, Cary, NC, USA). The log response values for the numbers were calculated by ln of the mean values (Si treatment/control) in the following equation 50 : ln [ X (Si-treatment) /X (control) ) ± (s 2 (Si-treatment) /n (Si-treatment) X 2 (Si-treatment) + (s 2 (control) /n (control) X 2 (control) ]. If the ln response rate is 0, the Si treatments and control do not differ. If it is less than 0, the Si treatments are lower than the control and if the ln response rate is greater than zero, the Si treatments are higher than the control.