Chinese cabbage ‘ZDQ’ was selected as the test material. This hybrid was developed in 2009 by the laboratory through the crossing of the cytoplasmic male sterile line Ogu28-13-2A (female parent) with the inbred line Byd06-2-1-2 (male parent). The biofertilizer, NSB, was produced and provided by Aupro Ecological Industry Operation Co., Ltd. (Hangzhou, China).

The content analysis of elements and organic matter in NSB mother liquor includes quantification of the concentrations of total nitrogen, total phosphorus, and total potassium in the solution. This analysis is being performed by NorminKoda Biotechnology Co., Ltd. (Wuhan, China). Additionally, the fungal community structure analysis of NSB mother liquor was conducted by Biomarker Technologies (Wuhan, China).

The pH value of non-rhizosphere and rhizosphere soil samples collected from both treated and non-treated groups was measured using a Sartorius PB-21 pH meter (Sartorius, Germany). The procedure involved drying the collected soil samples, sieving them through 60 mesh, 100 mesh, and 120 mesh filters, and then precisely weighing 1.0 g of the dried soil. This soil was then mixed with 8.0 mL of distilled water at a 1:8 ratio, thoroughly swirled, and left to stand for 1 h. The pH of the resulting supernatant was subsequently determined using the pH meter. Soil DNA was extracted with a Fast DNA® spin kit for soil (MP Biomedicals, Solon, OH, USA) and stored at −20°C.

After extracting DNA from the soil samples, Biomarker Technologies Co. Ltd. (Wuhan, China) utilized the Illumina NovaSeq platform to evaluate the fungal diversity present. The ITS sequences obtained were aligned was the fungal Unite database v8.0, which served as the reference database. The analysis employed specific primers, with the forward primer sequence being 5′- CTTGGTCATTTAGAGGAAGTAA-3′ and the reverse primer 5’-GCTGCGTTCTTCATCGATGC-3′. The fungal diversity assessment was then performed on the BMK cloud platform. The resultant raw sequence data has been archived in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database, which is accessible online. For the readers’ convenience, detailed information regarding each SRA submission, including the number of reads and the Phred score for the samples, has been compiled in the Supplementary materials. These details are presented in Supplementary Table S1 for the read counts and Supplementary Table S2 for the Phred scores, respectively.

Clarity was ensured by assigning the following labels to the samples: CK for non-rhizosphere soil, SH₂O for rhizosphere soil from the purified water group, and SNSBS for rhizosphere soil from the 30× diluted NSB group. The raw data underwent quality control procedures, such as the elimination of low-quality reads and length-based filtration, to produce high-quality reads. These high-quality reads were then organized into Operational Taxonomic Units (OTUs) through the use of Vsearch v2.4.3, which were later identified as features in the subsequent analysis. Taxonomic annotation was carried out utilizing these features, and the samples were categorized and examined at various classification levels based on the outcomes of the feature analysis (Robert Edgar, 2013). A Venn diagram was then used to analyze the number of unique features (fungal types) present in each group at different OTU classification levels. Additionally, a histogram was generated at the genus level to visualize the distribution of the top 10 most abundant fungal species. Diversity indices (Shannon, Simpson, Ace, and Chao1) were calculated using QIIME2 software¹ with OTU as the classification level. Functional composition analysis was performed using FUNGuild (Nguyen et al., 2016) to create a bar chart of the top 10 functional abundances at the Guild classification level. To explore potential relationships between plant weight, NSB application, and fungal communities, Pearson correlation analysis was performed on the top 10 most abundant fungal genera. A correlation heatmap was then generated to visualize these relationships. Similarly, a network diagram was constructed to depict the correlations among the top 20 most abundant fungi.

The description and determination of agronomic traits for Chinese cabbage followed the “Specification for the Description of Germplasm Resources of Non-bearing Chinese Cabbage” released by the National Crop Germplasm Resources Platform and the National Crop Science Data Center. Six agronomic traits were investigated (Supplementary Table S3). These specific indicators included plant height, plant breadth, plant weight, blade quantity, maximum leaf length, and maximum leaf width. The parameters were assessed and documented at the time of harvest. Specifically, the effect of NSBS on the growth and nutritional quality of Chinese cabbage was evaluated when the seedlings reached 77 days old, while the effect of NSBS pretreatment on the antioxidant enzyme capacity of Chinese cabbage under high temperatures was assessed when the seedlings were 28 days old.

Pigment content (chlorophyll a, chlorophyll b, total chlorophyll, and carotenoids) was determined using the acetone-ethanol mixed extraction method described by Ling and Chang-bin (2018). Weigh 0.1 g of leaves lyophilized powder and put it in a glass test tube. Add 10.0 mL 1:1 (V: V) acetone-ethanol solvent; Cover the tube plug, mix it upside down 15 times until it is completely dissolved, and soak it for 24 h under dark light. After extraction, centrifuged at 4,000 rpm at 25°C for 10 min; The 2.0 mL supernatant was extracted and the acetone ethanol mixture was used as blank control. The optical density (OD) was determined by an ultraviolet rays (UV) spectrophotometer (Shimadzu UV-1800, Tokyo, Japan) at the optical wavelength of 470 nm, 663 nm, and 645 nm, respectively. Calculate the concentration of chlorophyll a, chlorophyll b, total chlorophyll, and carotenoids according to the formula:

Chlorophyll a concentration : C a = 13.95 × A 665 − 6.88 × A 649 , unit mg ⋅ L − 1 ;

Chlorophyll b concentration : C b = 24.96 × A 649 − 7.32 × A 665 , unit mg ⋅ L − 1 ;

Total chlorophyll concentration : C t = C a + C b = 6.63 × A 665 + 18.08 × A 649 , unit mg ⋅ L − 1 ;

Carotenoids concentration : C x . c = ( 1 , 000 × A 470 − 2.05 × C a − 114.8 × C b ) / 245 , unit mg ⋅ L − 1 .

Vitamin C content was measured by high-performance liquid chromatography (HPLC) on an LC-20AT system (Shimadzu, Japan) following the method of Sun et al. (2018). 0.10 g of leaves lyophilized powder was added to 8.0 mL of 1% oxalic acid solution pre-cooled at 4°C under the condition of low temperature and light protection, then the light shading vortex was 15 s until the powder was completely suspended, and the powder was cooled in ice for 15 s. After standing in ice for 2 min, centrifuged at 4°C and 7,000 rpm for 10 min. Absorbed the supernatant with a 1.0 mL sterile syringe, passed water through a vinyl chloride filter, and injected into the sample bottle. HPLC conditions: Hypersil C18 column (5 μm, 250.0 mm × 4.6 mm), column temperature at 30°C, 0.1% oxalic acid solution mobile phase, 1 mL‧min⁻¹ flow rate, the OD value was determined at 243 nm wavelength.

Free amino acid content was determined using the ninhydrin method as described by Winters et al. (2002). Weigh 0.1 g of leaves lyophilized powder into a 10.0 mL centrifuge tube and add 10.0 mL 10% acetic acid. Free amino acid extract was obtained by whirlpool and centrifugation. Reaction system: 1.0 mL extract solution, 3.0 mL ninhydrin solution, and 0.1 mL of ascorbic acid were added. The mixture was thoroughly mixed in hot water at 65°C for 3 min and then left for 15 min away from light. After absorbing 2.0 mL reaction solution, the OD value was measured by a UV spectrophotometer at 570 nm wavelength. 0.0 mg·mL⁻¹ leucine standard solution mixed with ninhydrin hydrate was used as blank zero adjustments. C_A is the concentration of free amino acids in the reaction solution (mg·mL⁻¹).

Soluble sugar content was determined using the anthrone-concentrated sulfuric acid colorimetry method with slight modifications as described by Liu et al. (2017). 0.1 g of leaves lyophilized powder and an equal amount of activated carbon (for removing pigment effects) were added to a 15.0 mL centrifuge tube with 10.0 mL of 80% ethanol; After the powder was fully swirled and completely suspended, it was placed in a water bath and bathed in water for 30 min at 85°C. After cooling to room temperature, centrifuged at 6,500 rpm at 24°C for 5 min. After the supernatant was transferred to a 50.0 mL volumetric bottle for repeated extraction and transfer 3 times, it was determined to be 50.0 mL with 80% ethanol, which was the soluble sugar extraction solution to be measured. Reaction system: 1.0 mL extraction solution was added into anthracnose-concentrated sulfuric acid solution 3.0 mL, thoroughly mixed in the water bath at 90°C for 5 min, and then left for 20 min away from light. The OD value was measured at 620 nm wavelength by a UV spectrophotometer.

Soluble protein content was measured using the Coomassie brilliant blue method of Bradford (1976). Weigh 0.1 g of leaves lyophilized powder into a 15.0 mL centrifuge tube and add 10.0 mL distilled water. The vortex of the vortex instrument ensured that the powder was completely suspended and lasted for 30 s; After the whirlpool was finished, the supernatant was dissolved into soluble protein extract by centrifugation at 5,500 rpm at 24°C for 5 min. Reaction system: 1.0 mL extract, then 5.0 mL Coomassie bright blue solution, fully oscillating and mixing, and standing for reaction for 15 min away from light; The OD value was determined by a UV spectrophotometer at 595 nm wavelength after absorbing 2.0 mL reaction solution. 0.0 mg·mL⁻¹ bovine serum protein standard solution and Coomassie brilliant blue mixture were used as blank zero. The soluble protein content was calculated according to the standard regression curve of bovine serum protein OD value (the concentration gradients of the standard protein solution were 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, and the concentration unit was mg·mL⁻¹).

Leaf curl rate was calculated according to the formula by Chen et al. (2021): Leaf curl (%) = (maximum leaf width minus natural leaf width) / maximum leaf width × 100%.

The method for measuring membrane permeability followed the protocol described by Hong et al. (2008). After cleaning and drying the fourth functional leaf with deionized water, use a hole punch with a diameter of 4.0 mm to avoid the leaf veins and leaf edges, beat the leaves into small leaf rounds, and put 0.5 g leaves into a test tube with 5.0 mL deionized water, and vacuumed to ensure that all the leaves were submerged in the water. After standing for 15 min, the initial conductivity Ta of the extract was measured by a Model DDS-11A conductivity meter (Leici, Shanghai, China). Then put it in a boiling water bath for 15 min, took it out cooled to room temperature, and measured the final conductivity Tb of the extract. The relative conductivity of mycelia was calculated as follows: Relative conductivity % = Conductivity/Final conductivity × 100.

Five independent biological replicates were utilized to measure leaf curl rate and fresh weight above ground per plant in the study investigating the impact of NSBS pretreatment on Chinese cabbage under high temperatures. Additionally, three independent biological replicates were employed to assess other physiological traits.

Malondialdehyde (MDA) content was determined following the method proposed by Yuan et al. (2010), fresh leaf samples weighing 1.0 g were combined with 10.0 mL of a 5% trichloroacetic acid (TCA) solution and homogenized. The resulting homogenate was transferred to a 15.0 mL centrifuge tube. Centrifugation was performed at 4°C and 5,000 rpm for 10 min using a large low-temperature centrifuge to separate the supernatant extract. A volume of 2.0 mL of the supernatant was taken and mixed with 2.0 mL of a 0.6% 2-thiobarbituric acid solution. The mixture was then subjected to boiling in a water bath at 100°C for 30 min. Subsequently, centrifugation was repeated under the same conditions for 10 min, and the OD value of the resulting supernatant was measured at wavelengths of 450 nm, 532 nm, and 600 nm.

Specific extraction buffers were applied for homogenizing the considered leaf samples (0.5 g) as per the known procedure. Activities of antioxidant enzymes, including SOD, peroxidase (POD), and CAT, were measured using the methods described by Fathi et al. (2023). The absorbance of the reaction mixtures was measured at 560 nm. The amount of protein required to cause 50% inhibition in the photochemical reduction of nitroblue tetrazolium (NBT) was designated as one unit of SOD and was expressed as U·g⁻¹ FW. The mixture’s absorption of POD activity and CAT activity was measured at 470 nm and 240 nm, respectively. An absorbance variation of 0.01 units‧min⁻¹ was then applied as one unit of POD and CAT activities, which was expressed as U‧g⁻¹‧min⁻¹ FW.

SOD activity ( U .g − 1 FW ) = ( Δ A 560 × Vt ) / ( W × V s × 0.05 × t ) .

POD activity ( U .g − 1 ⋅ min − 1 FW ) = ( Δ A 470 × Vt ) / ( W × V s × 0.01 × t ) .

CAT activity ( U ⋅ − 1 ⋅ min − 1 FW ) = ( Δ A 240 × Vt ) / ( W × V s × 0.01 × t ) .

Δ A refers to the change occurring in absorbance, Vt denotes the total volume of the extracted enzyme solution, W stands for the sample’s fresh weight, t indicates the reaction time, and Vs describes the volume of the crude enzyme extract.

Flavonoid content was determined using the aluminum chloride colorimetric method of Koca and Karaman (2015). Fresh leaf samples were placed in a 15.0 mL centrifuge tube, followed by 10.0 mL 60% ethanol, and centrifuged at 24°C and 6,500 rpm for 10 min to obtain flavonoid extract. Flavonoid content determination reaction system: 2.0 mL extract, 0.3 mL 5% NaNO₂, 0.3 mL 10% AlCl₃, 4.0 mL 4% NaOH solution, standing for 30 min away from light; The OD value of 0.0 mL rutin standard liquid and reaction system mixture was determined by a UV spectrophotometer at 510 nm wavelength.

Three independent biological replicates were utilized for each experiment to evaluate biochemical characteristics.

Each experiment was confirmed with three biological and technical replicates. Photographs were captured using a Canon EOS 600D camera. Graphs were plotted using GraphPad Prism 8, and the resulting figures were assembled using Microsoft PowerPoint 2019. Data were represented as means ± SD, calculated by SPSS27.0 analysis software (IBM, Chicago, IL, USA). Further analysis of variance was performed to determine statistical significance. A p < 0.05 was statistically significant.

NSB application significantly promoted Chinese cabbage growth (Figure 1). The plant’s height, width, and weight were significantly increased in the WSF and NSBS groups compared to the H₂O group at 77 days old. Notably, plant weight in the NSBS group showed a remarkable 2.66-fold increase (Figures 1B,C). However, plant height and breadth did not differ significantly between the WSF and NSBS groups (Figure 1B). Similar trends were observed for blade quantity, maximum leaf length, and maximum leaf width (Figures 1D,E). All three measurements increased significantly in both the WSF and NSBS groups compared to the H₂O group. The NSBS group displayed a significant 1.21-fold increase in blade quantity. Additionally, the maximum leaf length and width in the NSBS group (18.60 cm and 15.70 cm, respectively) were significantly higher than those in the H₂O group (15.11 cm and 10.42 cm). Interestingly, the maximum leaf length did not differ significantly between the WSF and NSBS groups (Figure 1E). Chlorophyll content (a, b, and total) and carotenoids all increased significantly in the WSF and NSBS groups compared to the H₂O group (Figure 1F). Notably, chlorophyll a, b, and total chlorophyll content did not differ significantly between the WSF and NSBS groups. However, the NSBS group displayed a slightly lower carotenoid content (1.42 mg∙g⁻¹) compared to the WSF group (1.63 mg∙g⁻¹).

NSB application significantly enhanced the quality of Chinese cabbage (Figure 2). Vitamin C content increased remarkably in the NSBS group, showing a 1.36-fold and 1.06-fold increase compared to the H₂O and WSF groups, respectively (Figure 2A). Free amino acid content also increased significantly in both the WSFWSF and NSBS groups compared to the H₂O group, with the NSBS group showing the largest increase (2.16-fold) relative to the H₂O control (Figure 2B). Soluble sugar content, however, did not differ significantly among the groups (Figure 2C). Finally, the NSBS group displayed a significant increase (1.33-fold) in soluble protein content compared to the H₂O group (Figure 2D).

The findings of the study demonstrated that NSBS significantly improved the growth and nutritional value of Chinese cabbage, even more effectively than WSF alone (Figures 1, 2). To understand these effects, the ingredients of NSB mother liquid were investigated. It was found that with a pH of 7, NSB primarily consists of 0.322 g·L⁻¹ total nitrogen, 19.794 mg·L⁻¹ total phosphorus, 8.000 mg·L⁻¹ total potassium, and 48.000 mg·L⁻¹ organic matter content (Supplementary Table S4). The dominant fungal groups observed in the NSB were Ascomycota, accounting for 70.23% of the population, and Mortierellomycota, representing 16.35% of the total (Supplementary Table S5). To further explore the effect of soil by irrigating NSB, soil properties, and fungal communities in the rhizosphere compared to non-rhizosphere soil. The non-rhizosphere soil was named the control group (CK), the rhizosphere soil from the water treatment group was named SH₂O, and the rhizosphere soil from the NSBS treatment group was named SNSBS. Interestingly, there were no significant differences in pH between the control, SH₂O, and SNSBS groups. Two methods, Ace and Chao1 indices, were utilized to estimate the diversity of fungal species in the soil. Both methods showed the highest diversity in the control group and the lowest diversity in the SNSBS group (Table 1). In simpler terms, the control group had the most fungal species, followed by the SH₂O group and then the SNSBS group. Similarly, two other indices (Shannon and Simpson) indicated that the control group harbored the most diverse fungal community, followed by the SH₂O group and then the SNSBS group.

Species distribution and diversity indices (Figure 3) confirmed that the SNSBS group had the lowest fungal abundance and diversity compared to the control and SH₂O groups. This suggests that NSBS application suppressed the diversity and richness of fungal communities in the rhizosphere soil. The number of fungal types (OTUs) present in the rhizosphere soil was analyzed (Figure 4A). The control group had the most diverse fungal community, with 1,470 OTUs detected. Among these, 795 were unique to the control group, while 586 and 417 OTUs were shared with the SH₂O and SNSBS groups, respectively. The SH₂O group exhibited a total of 1,553 OTUs, with 842 unique OTUs and 453 shared with the SNSBS group. Finally, the SNSBS group had the least diverse community, with 894 OTUs, including 352 unique OTUs and 328 shared with both the control and SH₂O groups. Upon examination of the ten most abundant fungal genera across the groups, excluding unclassified fungi, Olpidium (Olpidiomycota), Thermogenes, Scutellinia, Fusarium, Kernia, Enterococcus, Sortariomycetes, Lasiobolidium (Ascomycota), and Basidiomycota were identified (Figure 4B). Interestingly, the fungal communities in the SH₂O and SNSBS groups were more similar in terms of species abundance (Figure 4C). However, the relative abundance of specific fungi varied. For example, Olpidium was most abundant in the SNSBS group and least abundant in the control group. Conversely, the control group had the highest relative abundance of Basidiomycota and Fusarium, while the SNSBS group had the lowest.

Fungal communities in the rhizosphere soil were dominated by Ascomycota across all groups (Table 2). In the control group, Ascomycota was the sole dominant fungal phylum. In the SH₂O and SNSBS groups, Ascomycota and Olpidiomycota together comprised over 80% of the identified fungal groups. Notably, the relative abundance of Ascomycota was significantly lower in the SNSBS group compared to both the control rhizosphere soil and non-rhizosphere soil (p < 0.05). Interestingly, within these dominant phyla, Olpidiomycota reached the highest relative abundance (33.4%) in the SNSBS group, while Basidiomycota showed the lowest abundance (1.5%) across all groups. Several genera exhibited distinct abundance patterns across the groups (Table 2). Notably, Poaceascoma, Cladosporium, Exserohilum, Microascus, Lasiobolidium, Gibellulopsis, Myrmetridium, and Lomentospora displayed no significant differences between the CK and SH₂O groups, while their relative abundance significantly decreased in the SNSBS group. Ascobolus and Thermogenes had the highest relative abundance in the SH₂O group, followed by the CK group, and the lowest in the SNSBS group. Similarly, Scutellinia showed the highest abundance in the SH₂O group, with no significant difference observed between the CK and SNSBS groups. Microascus displayed the opposite trend, with the highest relative abundance in the CK group, followed by the SH₂O group, and the lowest in the SNSBS group. Enterococcus showed the highest relative abundance in the SNSBS group, followed by the CK group, and the lowest in the SH₂O group. Finally, Olpidium followed a similar pattern to Enterococcus, reaching the highest relative abundance (33.39%) in the SNSBS group, followed by the SH₂O group, and the lowest abundance (3.62%) in the CK group.

Functional prediction identified the top five abundant fungal groups: undefined saprotrophs (decomposers of dead organic matter), fungal parasites, animal parasites, dung saprotrophs, and plant pathogens (Figure 5).

The control group harbored the highest relative abundance of saprotrophic fungi, followed by the SH₂O group and then the SNSBS group. Conversely, the relative abundance of parasitic functional groups showed the opposite trend, with the SNSBS group having the highest abundance (Figure 5A). Interestingly, correlation analysis revealed a significant positive correlation between individual plant weight and the abundance of oil chytrids (p < 0.001) in the SNSBS group (Figure 5B). In contrast, plant weight showed a significant negative correlation with Basidiomycota fungi, and the application of NSBS also negatively correlated with Basidiomycota abundance. Further analysis using a correlation network of the top 20 most abundant genera identified several interesting relationships (Figure 5C). The relative abundance of Chytrid fungi displayed a significant negative correlation with Basidiomycota and a significant positive correlation with Byssochlamys abundance. Additionally, a positive correlation existed between Thermophilic fungi and Fusarium, as well as between Enterococcus and Obturator.

Following the discovery that applying NSB to soil significantly increased osmoregulatory substances in Chinese cabbage, this study investigated its potential to enhance the crop’s thermotolerance. The performance of four groups were compared: H₂O-OT (pretreated with purified water at optimal temperature), H₂O-HT (pretreated with purified water at short-term high temperature), NSBS-OT (pretreated with NSBS solution at optimal temperature), and NSBS-HT (pretreated with NSBS solution at short-term high temperature) (Figure 6).

Pretreatment with NSBS solution effectively mitigated the negative effects of high temperatures on Chinese cabbage seedlings (Figure 6). Compared to the control group pretreated with water at optimal temperature (H₂O-OT), the H₂O-HT group (pretreated with water at high temperature) exhibited a significantly higher leaf curl rate (Figure 6A). While the NSBS-HT group (pretreated with NSBS at high temperature) also showed increased leaf curling under high temperatures compared to the NSBS-OT group (pretreated with NSBS at optimal temperature), it was still considerably lower than the H₂O-HT group. High temperatures also significantly reduced fresh weight in all groups (Figure 6B). The H₂O-HT group displayed a 25% decrease in fresh weight compared to the H₂O-OT group, whereas the NSBS-HT group only experienced a 9% decrease. Due to high temperatures causing leaf curling and weight reduction, the extent of leaf damage was examined (Figures 6C–E). Pretreatment with NSBS helped to mitigate leaf damage in Chinese cabbage under high temperatures (Figures 6C–E). There was no significant difference in total chlorophyll content between the H₂O-OT and NSBS-OT groups. However, compared to the H₂O-OT group, the H₂O-HT group showed a significant decrease (87%) in total chlorophyll content, while the NSBS-HT group exhibited a 91% decrease. Notably, NSBS pretreatment helped to alleviate the decrease in total chlorophyll content after 24 h of heat exposure. Interestingly, NSBS pretreatment did not significantly affect the relative electrical conductivity of leaves. Following 24 h of heat exposure, the relative conductivity of the NSBS-HT group was significantly lower than that of the H₂O-HT group (Figure 6D). This pattern mirrored the response of MDA content to high temperatures, with the NSBS-HT group exhibiting a notable decrease in MDA accumulation in the leaves compared to the H₂O-HT group (Figure 6E).

To further investigate the impact of NSBS on Chinese cabbage under high temperatures, the activity of antioxidant enzymes and their corresponding antioxidants in the leaves of Chinese cabbage seedlings was determined. As shown in Figure 7, the NSBS-OT group had significantly higher SOD and POD activities compared to the H₂O-OT group. After 24 h of heat exposure, SOD activity in the NSBS-HT group showed no significant difference compared to the H₂O-HT group (Figure 7A). However, POD activity in the NSBS-HT group was significantly higher than that in the H₂O-HT group (Figure 7B). CAT activity in the NSBS-OT group also did not show a significant difference compared to the H₂O-HT group. Regarding antioxidant compounds, high temperatures caused a notable increase in flavonoid accumulation, with the NSBS-HT group exhibiting a higher total flavonoid content compared to the H₂O-HT group (Figure 7D). In contrast, vitamin C content significantly decreased after 24-h heat exposure, although the NSBS-HT group still showed a higher concentration compared to the H₂O-HT group (Figure 7E). Following the identification of these differences in antioxidant enzyme activity and antioxidant content, osmoregulatory substance content under the influence of NSBS pretreatment was examined (Figures 7F–H). The NSBS-OT group exhibited a notable 1.23-fold increase in free amino acid accumulation compared to the H₂O-OT group. Furthermore, even after 24 h of heat exposure, the NSBS-HT group maintained a significantly higher free amino acid content compared to the H₂O-HT group (Figure 7F). Interestingly, NSBS pretreatment did not significantly alter the concentration of soluble protein, as evidenced by the lack of notable differences between the H₂O-OT and NSBS-OT groups (Figure 7G). However, NSBS did mitigate the decline in soluble sugar content under high temperatures. Specifically, compared to their respective control groups (H₂O-OT and NSBS-OT), soluble sugar content decreased by 0.73-fold in the H₂O-HT group and 0.85-fold in the NSBS-HT group (Figure 7H). Therefore, NSBS plays an important role in maintaining the leaf quality of Chinese cabbage under high temperatures.