Three spectral combinations, white LEDs (W, B:G:R = 34:46:20, Figures 1A-a), white plus red LEDs (WR, B:G:R = 19:25:56, Figures 1A-b), and white plus blue LEDs (WB, B:G:R = 70:21:9, Figures 1A-c) were applied and wheat seedlings were planted in a plant factory for 43 days. The spectral characteristics of different light treatments were measured with a portable spectroradiometer (Plant Lighting Analyzer, V 2.00, China), and showed in Figure 1. Light intensity was kept constant at 450 ± 20 μmol m^–2 s^–1 throughout the experiment by moving up the light sources above the plants every week, where the photosynthetic photon flux density (PPFD) was measured with a photo/radiometer (Plant Lighting Analyzer, V 2.00, China). Lighting time per day was set to 12 h (12/12 h light/dark).

Winter wheat seeds (T. aestivum L., cv. ‘JiMai22’) were sterilized by 5% hydrogen peroxide (H₂O₂) for 30 min, then washed adequately 4 times in distilled water, and soaked in distilled water for 12 h. The seeds were imbibed in the dark on moistened germination paper at 25°C until plumule exposed 1 cm, and then vernalized for 15 days at 4°C. Before planting, germinated seeds grew for 3 days under white LEDs light (spectrum same as white LEDs, Figures 1A-a) for uniform growth (PPFD = 100 μmol m^–2 s^–1). The germinated seeds were planted into sponge cubes (2.0 cm × 2.0 cm × 2.0 cm) and hydroponically grown in a plant factory. Eighty-one plants spaced 8 cm apart were planted in a hydroponic box (80 cm × 80 cm × 10 cm). Each hydroponic box was used for one light treatment, and three hydroponic boxes as biological replicates in each light environment. During the experiment, wheat was grown for 43 days and randomly rotated every 5 days. A half-strength Hoagland’s solution was used, the pH of which was kept between 5.5 and 6.5, and renewed weekly (Hoagland and Arnon, 1950). Temperature during the experiment was maintained at 22 ± 1°C, and relative humidity was controlled at 65 ± 5%.

Eight seedlings were taken from each treatment to measure plant height, leaf area and shoot dry weight at 13, 22, 31, and 40 days after planting (DAP). Plant height, length and width of all leaves of the samples were measured by a ruler, and the leaf area was calculated according to the coefficient method (Leaf area = Length × Width × 0.76) (Li et al., 2018; Guo et al., 2022b). Shoot fresh weight of the entire plant was determined by an electronic precision balance (CP224C, OHAUS, United States). After oven-dried at 105°C for 30 min, the samples were kept at 80°C for 72 h, and shoot dry weight was then measured by the electronic precision balance.

Fresh samples of 30 wheat seedlings (10 plants from the same hydroponic box as a biological repeat) were collected from each treatment at 43 DAP, and immediately frozen in liquid nitrogen, and then stored in a laboratory refrigerator at −80°C. The chemical used in the experiment were all supplied by Sigma-Aldrich (United States).

Soluble sugar content was determined as described by Hanft and Jones (1986). Fresh sample (0.1 g) was ground in liquid nitrogen. After adding 3 ml deionized water, the grounded sample was heated in a boiling water bath for 20 min, centrifuged at 4,500 g for 10 min, and the supernatant was taken. A total of 3 ml deionized water was added to the residue, extracted twice repeatedly. After combining the resulting supernatants, deionized water was added to adjust the volume to 10 ml. Extract solution (1 ml) was mixed with 4 ml anthrone solution (2 g L^–1), and then soaked in a water bath at 45°C for 12 min. After cooling in the dark, the absorbance of the mixture was detected at 625 nm by a microplate reader (Multiskan FC, Thermo Scientific™, United States), and soluble sugar content was calculated using a sucrose standard curve with concentrations ranging from 0 to 2 mg ml^–1 (R² = 0.9997).

Soluble protein content was measured by the Coomassie brilliant blue colorimetry method (Bradford, 1976). A total of 0.1 g fresh sample was homogenized in 2 ml distilled water. The mixture was centrifuged at 3,000 g for 10 min, and 1 ml supernatant was mixed with 5 ml Coomassie brilliant blue G-250 solution (0.1 g L^–1). After 2 min, the absorbance of the mixture was detected at 595 nm by the microplate reader, and soluble protein content was calculated from standard curves (R² = 0.9968).

Fresh sample (0.2 g) was ground into homogenate with 5 ml trichloroacetic acid solution (TBA, 5%, w/v), and centrifuged at 15,000 g for 15 min. The supernatant was used for measurement of malondialdehyde (MDA) and GSH contents.

The method described by Hameed et al. (2021) was used to quantify MDA content, with slight modification. 1 ml supernatant was mixed with 1 ml TBA solution (0.67%, w/v), incubated in boiling water for 30 min and then quickly cooled in an ice-bath. The mixture was centrifuged and the absorbance was detected at 450, 532, and 600 nm. MDA content was determined using the following equation:

where V is the total volume of supernatant and W is the fresh weight of the sample.

Glutathione content was measured according to the method by Qiu et al. (2014) with slight modification. A total of 2 ml supernatant was added to 4 ml phosphate buffer (200 mM, pH 7.0) and 0.4 ml 5, 5-dithiobis (2-nitrobenzoic) solution (0.396 mg ml^–1, pH 7.0). After 5 min at 30°C, the absorbance was determined at 412 nm by the microplate reader and GSH content was calculated based on the GSH standard curve with concentrations ranging from 0 to 80 μg ml^–1 (R² = 0.9976).

Ascorbic acid content was determined according to the method by He et al. (2020). Fresh sample (0.1 g) was extracted with 1 ml oxalic acid EDTA solution (1%, w/v), and centrifuged at 4,000 g for 10 min. 0.2 ml supernatant was mixed with 0.8 ml oxalic acid EDTA solution (1%, w/v), 0.1 ml phosphate-acetic acid solution (3%, w/v), 0.2 ml vitriol (5%, v/v), and 0.4 ml ammonium molybdate solution (5%, w/v). After 15 min, the absorbance of mixture solution was measured at 705 nm by the microplate reader, and ASA content was calculated based on the standard curve with concentration ranging from 0 to 1 mg ml^–1 (R² = 0.9977).

Fresh sample (1.0 g) was ground in liquid nitrogen, and extracted with 25 ml methanol solution (80%, v/v) ultrasonically for 20 min, and centrifuged to obtain supernatant. The above process was repeated three times, and all supernatants were pooled and volume was adjusted to 100 ml with methanol solution (80%, v/v). The methanol extract was stored at −20°C for determination of the contents of flavonoids, triterpenes, and polyphenols, and the free radical scavenging abilities of 2,2′-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) and 2,2-diphenyl-1-picrylhydrazyl (DPPH).

The total flavonoids content was measured using the method described by Guo et al. (2013). A total of 2 ml extract, 0.2 ml NaNO₂ solution (5%, w/v), 0.2 ml Al(NO₃)₃ solution (10%, w/v), and 2 ml NaOH solution (1 M) were mixed adequately. The absorbance of the mixture was measured at 510 nm by the microplate reader after standing for 15 min. Total flavonoids content was expressed as milligram rutin equivalent gram^–1 in fresh weight (mg RE g^–1 FW) through the calibration curve of rutin with the concentration ranging from 0 to 20 μg ml^–1 (R² = 0.9974).

The total triterpenes content was measured by the method of Jiang et al. (2021). A total of 100 μl extract were mixed with 100 μl vanillin-acetic acid solution (2.5%, w/v) and 200 μl perchloric acid solution. After incubation at 60°C for 15 min, 650 μl glacial acetic acid was added and adequately mixed. The absorbance of mixture was detected at 550 nm by the microplate reader after standing for 10 min. The total triterpenes content was expressed as milligram ursolic acid equivalent gram^–1 in fresh weight (mg UAE g^–1 FW) and calculated from the standard curve (concentration ranging from 20 to 120 μg ml^–1, R² = 0.9918).

The total polyphenols content was measured using the method described by Sarker and Oba (2020). A total of 100 μl extract was mixed with the Folin–Ciocalteu reagent (2 N, 50 μl). A total of 400 μl Na₂CO₃ solution (1 M) and 1 ml deionized water were added after 5 min. The mixture was incubated at 37°C for 90 min and the absorbance was measured at 740 nm by the microplate reader. The content of the total polyphenols was expressed as gallic acid equivalent standard of fresh weight (mg GAE g^–1 FW) through the calibration curve of gallic acid with concentration ranging from 0 to 50 μg ml^–1 (R² = 0.9983).

ABTS radical scavenging ability was measured using the method by Jiang et al. (2021). Equal volume of 7.4 mM ABTS solution and 2.6 mM K₂S₂O₈ solution were mixed at room temperature in the dark for 20 h to generate ABTS+ stock solution. A total of 5 mM PBS solution (pH 7.4) was used to dilute ABTS+ stock solution to absorbance of 0.75 at 734 nm to prepare working solution. A total of 200 μl of working solution was mixed with 40 μl extract. After incubation in the dark for 6 min, the absorbance was read at 734 nm by the microplate reader. The ABTS radical scavenging ability was calculated from the Trolox calibration curve (concentration ranging from 0 to 0.05 mg ml^–1, R² = 0.999) and expressed as Trolox equivalent standard of fresh weight (mg TE g^–1 FW).

2,2-Diphenyl-1-picrylhydrazyl radical scavenging ability was also measured using the method by Jiang et al. (2021) with small adjustment. A total of 200 μl DPPH solution (0.15 mM) was blended with 100 μl extract in a 96-well-plate. After incubation at 37°C for 30 min, the absorbance was read at 517 nm by the microplate reader. The result was expressed as ASA equivalent per gram of fresh weight (mg AE g^–1 FW) based on the standard curve (concentration ranging from 0.005 to 0.4 mM, R² = 0.9976).

Fresh sample (0.5 g) was ground in liquid nitrogen, and extracted into a homogenate with 3 ml of ice-cold extraction buffer, which contained phosphate buffer (50 mM, pH 7.8) and EDTA (1 mM). The homogenate was centrifuged at 15,000 g for 15 min at 4°C, and the supernatant was used to measure the activities of SOD, POD, CAT, and APX.

Superoxide dismutase activity was measured based on the method proposed by Giannopolitis and Ries (1977), with slight modifications. A total of 3 ml reaction mixture was composed of 0.1 ml supernatant, phosphate buffer (50 mM, pH 7.8), EDTA (0.1 mM), methionine (130 mM), nitroblue tetrazolium (NBT, 0.4 mM), and riboflavin (0.02 mM). The reaction was started by placing the reaction tube under two 40 W fluorescent lamps and stopped after 20 min by removing the reaction tube from the light source. Light reactions with no light and no supernatant were used as calibration standards. The absorbance of the reaction mixture was read at 560 nm. One unit of SOD activity (U) was defined as the amount of enzyme required to produce 50% inhibition of NBT reduction.

Peroxidase activity was measured based on the method of Qiu et al. (2011) with slight changes. Added 0.1 ml supernatant to a mixture consisting of phosphate buffer (50 mM, pH 7.0), guaiacol (0.05%, w/v), EDTA (0.1 mM), and H₂O₂ (2%, v/v). By tracking the increase in absorbance at 470 nm over 5 min and quantified by the amount of tetraguaiacol formed using its molar extinction coefficient (26.6 mM^–1 cm^–1). One unit of POD activity (U) was defined as a decrease in absorbance value of 0.01 per minute.

Catalase activity was measured with the method of Cakmak and Marschner (1992). A total of 0.1 ml supernatant was added to a mixture of phosphate buffer (100 mM, pH 7.0), EDTA (0.1 μM), and H₂O₂ (0.1%, v/v). The reaction was measured by tracking the decrease in absorbance at 240 nm for 3 min and quantified by its molar extinction coefficient (39.4 mM^–1 cm^–1). One unit of CAT activity (U) was defined as a decrease in absorbance value of 0.01 per minute.

Ascorbate peroxidase activity was measured with the method of Nakano and Asada (1981). A total of 3 ml reaction solution consisted of 0.1 ml supernatant, phosphate buffer (50 mM, pH 7.8), EDTA (0.1 mM), ascorbate (0.5 mM), and H₂O₂ (0.1 mM). The resulting mixture was quantified by tracking the increase in absorbance at 290 nm for 1 min and by its molar extinction coefficient (2.8 mM^–1 cm^–1). One unit of APX activity (U) was defined as an increase in absorbance value of 0.01 per minute.

The data were subjected to one-way ANOVA using SPSS statistical software (PASW statistical version 26.0, SPSS Inc., Chicago, IL, United States). Duncan’s new multiple range test was applied to identify differences between means at the ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001 level. Correlations between growth, quality characteristics, antioxidant properties of wheat seedlings and light environments were uncovered through Pearson correlation analysis based on the percentage composition of light quality in the treatments (Figure 1B). All figures were created using OriginPro 2022 (OriginLab Corp., United States) or Adobe Illustrator CC 242 2019 (Adobe Inc., United States).

As showed in Figure 2, growth of wheat seedlings was significantly influenced by light qualities. The highest plant height of wheat seedlings was observed under WR and lowest was observed under WB at 13, 22, and 31 DAP (Figures 2A–C). Leaf area, shoot fresh and dry weight per wheat seedling was highest under WR at 13 and 22 DAP (Figures 2E,F,I,J,M,N). Shoot fresh weight was higher under WR than that under W at 31 DAP (Figure 2K). Difference in plant height and shoot fresh weight among the three light treatments disappeared at 40 DAP (Figures 2D,L), and difference in leaf area and shoot dry weight disappeared at 31 and 40 DAP (Figures 2G,H,O,P).

The contents of soluble sugar and soluble protein in wheat seedling leaves and stems were shown in Figure 3. Soluble sugar contents under WB increased 22.3% in leaves and 65.0% in stems compared with those under W (Figures 3A,B). Soluble protein content under WB decreased 33.5% in leaves than that under W (Figure 3C). Soluble sugar contents under WR increased 54.1% in stems (Figure 3B), and decreased 31.8% in leaves compared than those under W (Figure 3A). Soluble protein contents were 36.8% lower in leaves and 15.2% lower in stems than those under W (Figures 3C,D).

The contents of total flavonoids, triterpenes and polyphenols in leaves and stems of wheat seedlings grown under different LEDs treatments were provided in Figure 4. Compared with W, WB enhanced total flavonoids content in leaves by 19.3% (Figure 4A), and total triterpenes content in stems by 29.4% (Figure 4D). Compared with W, WR, and WB resulted in a significant decrease in total triterpenes and polyphenols content in leaves (Figures 4C,E). There was no significant difference in the contents of total flavonoids and polyphenols in stems among all treatments (Figures 4B,F).

Contents of MDA, GSH, and ASA in leaves were higher than those in stems (Figure 5). MDA contents of leaves and stems under WB and WR were significantly lower than those under W (Figures 5A,B). Contents of GSH and ASA in leaves and stems were the highest under WB (Figures 5C–F). GSH content in stems increased 98.6% under WR than that under W (Figure 5D). ASA content in stems decreased 11.0% under WR than that under W (Figure 5F). Contents of GSH and ASA in leaves under WR and W were not significantly different (Figures 5C,E).

The activities of antioxidant enzymes in leaves and stems were affected by light qualities (Figure 6). Compared with W, WB increased SOD activity of leaves by 72.1% (Figure 6A), and decreased SOD activity of stems by 19.8% (Figure 6B). SOD activity under WR decreased 40.9% in leaves and 18.3% in stems compared with those under W (Figures 6A,B). POD and CAT activities in leaves and stems under WR and WB were significantly higher than those under W (Figures 6C–F). APX activities in leaves under WR and WB were higher than that under W (Figure 6G). No significant difference in stem APX activity was observed among the three treatments (Figure 6H).

The ABTS and DPPH radical scavenging capacities in leaves were higher than those in stems (Figure 7). Compared with W, WR reduced ABTS radical scavenging capacity in leaves by 26.7% (Figure 7A). Compared with W, WB, and WR increased ABTS radical scavenging capacity in stems by 32.1 and 37.9%, respectively (Figure 7B), and DPPH radical scavenging ability in leaves by 130.6 and 218.5%, respectively (Figure 7C). There was no significant difference in DPPH radical scavenging ability in stems among all the three treatments (Figure 7D).

To explore the effects of light spectral qualities (blue, green, and red) of light treatments, correlation between light spectral qualities and growth characteristics (Figure 8), nutritional and antioxidant characteristics (Figure 9) of wheat seedlings was analyzed. Among the measured parameters of growth characteristics, blue light was significantly negatively correlated with most parameters, and red light was significantly positively correlated with most parameters (Figure 8). The situation was different when it came to parameters in terms of nutritional characteristics and antioxidant properties (Figure 9). In leaves, blue light was significantly positively correlated with SS (soluble sugar), TF (total flavonoids), GSH, ASA, SOD, POD, APX, and ABTS, and significantly negatively correlated with MDA. In stems, blue light was significantly positively correlated with TT (total triterpenes), GSH, ASA, POD, and CAT. Conversely, red light was significantly negatively correlated with SS, TF, TT, TP (total polyphenols), GSH, SOD, and ABTS in leaves, and was significantly negatively correlated with SP (soluble protein) and ASA in stems. In addition, there was a significant positive correlation among TF, GSH, ASA, SOD, POD, and APX in leaves. MDA was significantly negatively correlated with TF, GSH, ASA, POD, APX, and DPPH.