The National Institute of Horticultural and Herbal Science in Korea grew the Chinese cabbage cultivar ‘Chunkwang’ (Brassica rapa subsp. pekinensis) in a greenhouse under natural light conditions. Seeds of the cultivar were purchased from SAKATA KOREA (Seoul, Korea). After sowing, the seedlings were grown in a greenhouse for 6 days. On the 6th day, they were transferred to a growth chamber with light intensity at 350 μmol m⁻² s⁻¹, temperature at 25°C/20°C (16 h Light/8 h Dark), and humidity at 75%/60% (16 h L/8 h D). The seedlings were allowed to adapt for 5 days under long-day conditions, reaching the 11-day-old seedling stage with fully expanded first and second true leaves. Subsequently, half of the plants underwent heat treatment. This treatment was conducted in a growth chamber set to 40°C/35°C (16 h L/8 h D) and 75%/60% (16 h L/8 h D) humidity.

Samples were collected at 0, 1, 2, 4, 5, 6, 7, and 9 days after heat treatment by harvesting the first and second true leaves. The samples were immediately stored at −80°C.

For short-term heat treatment experiments, the seeds were sown in plug trays filled with 110 mL of commercial potting soil (Baroker potting mix, Seoul Bio Co., Ltd., Korea) per cell. After 14 days, the seedlings were transplanted into 450 mL pots and provided with sufficient water. The plants were acclimated in the greenhouse for 2 days, and then transferred to a growth chamber (Controlled Environment Rooms, Environmental Growth Chamber, USA) set at 340 μmol m⁻² s⁻¹ light intensity, constant temperature of 21°C (day/night), and a 16/8-h light/dark photoperiod for 5 days of adaptation. For heat treatment, half of the potted plants at the seven-leaf stage (23 days old as the vegetative stage) were transferred to a growth chamber set at 40°C. Leaf samples were collected at 1, 3, 6, 9, 12, and 15 h, taking the 3rd and 4th true leaves from the bottom. The collected leaf samples were immediately flash-frozen in liquid nitrogen and stored at −80°C for RNA extraction.

Total RNA was extracted using the TRIzol Reagent method (Invitrogen, USA) according to the manufacturer’s instructions [23]. Approximately 100 mg of leaf samples were pre-processed with 1 mL of TRIzol Reagent, followed by the addition of 100 μL of chloroform to disrupt the cell membranes. The subsequent steps were carried out using the Qiagen RNeasy Mini Kit, starting from the RNeasy Mini spin column purification.

To obtain pure RNA, the extracted RNA was treated with DNase I (Qiagen) to remove any DNA contamination. The quality and purity of the RNA were verified using a NanoDrop 2000 spectrophotometer and 1% agarose gel electrophoresis.

For transcriptome sequencing, four biological replicates were processed. Library preparation was conducted using the TruSeq Stranded Total RNA Library Prep Plant Kit (Illumina), ensuring depletion of ribosomal RNA and inclusion of coding and non-coding transcripts. The prepared libraries were sequenced on the Illumina platform, generating paired-end reads with a read length of 101 bp.

The resulting raw sequence data were quality-checked using FastQC (v0.11.9) [24] and subjected to adapter trimming and low-quality read removal using Trimmomatic (v0.39) [25]. High-quality reads were aligned to the Brassica rapa reference genome (GCF_000309985.2_CAAS_Brap_v3.01) using HISAT2 (v2.1.0) [26]. The resulting alignments were processed using SAMtools [27]. Gene/transcript abundance was calculated with StringTie (v2.1.3b) [28], and normalized expression values were represented as Transcripts Per Kilobase Million (TPM).

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to identify functional categories and pathways associated with DEGs. GO enrichment analysis was conducted for the three main categories: Biological Process (BP), Molecular Function (MF), and Cellular Component (CC), using the DAVID bioinformatics resources (https://david.ncifcrf.gov/).

KEGG pathway analysis was carried out using the KEGG Mapper tools (http://www.kegg.jp/kegg/pathway.html). DEGs were mapped to KEGG pathways, and enrichment was analyzed based on pathway-specific gene mapping and pathway completeness using the species-specific KEGG database (Brassica rapa reference genome: GCF_000309985.2). Statistical significance of enriched pathways was determined using Fisher’s exact test, with results adjusted for multiple testing using Bonferroni correction and false discovery rate (FDR) methods.

The same samples used for RNA extraction were processed for analysis. Each sample (100 mg) was homogenized with 1 mL of PBS (potassium phosphate buffer saline, pH 7.0) and centrifuged at 13,000 rpm for 30 min at 4°C. The supernatant was collected for further analysis. Protein quantification was conducted using the Bradford method [29]. MDA and H₂O₂ contents were measured using the EZ-Lipid Peroxidation (TBARS) Assay Kit and the EZ-Hydrogen Peroxide/Peroxidase Assay Kit (DoGenBio Co., Korea), respectively, following the manufacturer’s instructions.

The activities of antioxidant enzymes, including ascorbate peroxidase (APX), catalase (CAT), peroxidase (POD), and superoxide dismutase (SOD), were analyzed using the following assay kits: Ascorbate Peroxidase Activity Assay Kit (Elabscience Biotechnology Inc., USA), EZ-Catalase Assay Kit, EZ-Hydrogen Peroxide/Peroxidase Assay Kit, and EZ-SOD Assay Kit (DoGenBio Co., Korea), respectively. The assays were performed according to the manufacturer’s protocols.

The synthesized cDNA was used for qPCR analysis using EvaGreen qPCR Master Mix (BIOFACT, Daejeon, South Korea) and a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Relative mRNA levels were determined by normalizing the PCR threshold cycle number of each target gene with that of the reference gene, Actin [30]. In the qPCR analysis, three technical repeats were measured for each biological replicate analyzed. The primers used for qPCR analyses are presented in Table S1.

In this study, we used the Chinese cabbage cultivar ‘Chunkwang’ (Brassica rapa subsp. pekinensis) to compare gene expression profiles of young seedlings under heat stress using RNA-seq analysis.

First, to investigate the effects of heat stress on the growth and physiological changes in 11-day-old Chinese cabbage seedlings, we compared the length of the first true leaf in the control group (O_group, 24°C) and the heat-treated group (H_group, 40°C) over a period of 9 days. Plants exposed to heat stress exhibited smaller leaf sizes compared to those grown under optimal conditions in the control group.

Under optimal temperature conditions (24°C), leaf growth exhibited a consistent progression, characterized by a gradual increase in the length of the first true leaf throughout the experimental period. In contrast, the group subjected to heat treatment at 40°C exhibited inhibited growth, as evidenced by a slower rate of leaf elongation compared to the control group, which resulted in relatively smaller leaf sizes from the initial stages of the treatment (Fig. 1a). Notably, at 9 days after heat treatment (DAH), the average leaf length of the heat-treated group was approximately 39% shorter than that of the control group (Fig. 1b).

Following a period of 5 days of heat treatment, the cotyledons of the treated plants displayed significant curling and initiated a yellowing process (Fig. 1a) [12]. These observations indicate stress-induced damage, and further analyses were conducted at 2 DAH (short-term heat stress) and 5 DAH (long-term heat stress) to investigate physiological changes under heat stress. These time points were selected to capture distinct stages of prolonged heat stress, with 2 DAH representing an early transitional phase beyond short-term responses (≤24 h) before visible phenotypic damage, and 5 DAH corresponding to the onset of visible stress symptoms (Fig. 1a).

ROS scavenging system and the activity of antioxidant enzymes were compared between the control and heat-treated groups at different time points.

SOD activity increased by approximately 18% in the heat-treated group compared to the control group at 2d and by 41% at 5d (Fig. 2a), while POD activity decreased by approximately 32% in the heat-treated group at 2d (Fig. 2b). In contrast, CAT activity showed no significant differences between the control and heat-treated groups at both 2d and 5d (Fig. 2c). Ascorbate peroxidase (APX) activity increased by 28% in the heat-treated group at 2d and by 73% at 5d (Fig. 2d). H₂O₂ concentration in the heat-treated group decreased by approximately 65% at 2d and by 61% at 5d, showing lower levels than in the control group (Fig. 2f). Malondialdehyde (MDA) content showed no significant differences at 2d, but increased by approximately 33% in the heat-treated group at 5d (Fig. 2e). These results indicate that 2-day and 5-day heat stress have distinct effects on plants.

To investigate the transcriptomic response of Chinese cabbage to heat stress, RNA-seq was performed on samples collected at 2 DAH and 5 DAH as well as from control plants grown under optimal conditions. The results showed that more than 55 million high-quality reads were generated per sample, with over 94% of the reads successfully mapped to the reference genome (Table 1).

The raw sequencing reads were processed and subjected to quality control using FastQC [24], and trimming was performed with Trimmomatic to remove low-quality bases and adapter sequences [25]. The trimmed reads were then aligned to the Brassica rapa reference genome (GCF_000309985.2_CAAS_Brap_v3.01) using HISAT2 [26]. The high mapping rates (94.74%–97.41%) and minimal levels of unmapped reads confirmed the reliability and quality of the data, ensuring suitability for downstream analyses. Transcript abundance was calculated using StringTie [28], and differential expression analysis was conducted using edgeR [31], identifying significant DEGs with |log₂ fold change| ≥ 2 and p-value < 0.05. The high-quality reads produced in this study have been deposited in the NCBI BioProject database (GenBank accession number: PRJNA1184374).

To evaluate the overall transcriptomic changes induced by heat stress, multidimensional scaling (MDS) analysis was conducted based on the gene expression profiles of all samples. The MDS plot (Fig. 3) reveals clear clustering between the heat-treated (H_group) and control (O_group) samples. Control samples (O_group) exhibited a close clustering, reflecting a consistent pattern of gene expression under normal conditions. In contrast, the heat-treated samples (H_group) were distinctly separated from the control samples along the primary component axis. The x-axis (Component 1) accounted for 71.4% of the total variance, while the y-axis (Component 2) explained 20.7% of the variance, indicating transcriptomic changes induced by heat stress. Notably, the 2-day heat-treated sample (2-H) exhibited a clear separation from the 2-day control sample (2-O) in clustering. Furthermore, the separation between the 5-day heat-treated sample (5-H) and the 5-day control sample (5-O) was even more pronounced than that between the 2-H and 2-O samples, indicating that short-term and long-term heat stress have distinct effects on gene expression.

RNA-seq analysis identified a substantial number of DEGs at both 2 and 5 days after heat treatment, consistent with prior transcriptomic studies in Brassica species under abiotic stress [32,33]. At the 2-day time point (2-H vs. 2-O), a total of 4137 genes were up-regulated, while 3928 genes were down-regulated. At the 5-day time point (5-H vs. 5-O), 5738 genes were up-regulated, and 6048 genes were down-regulated (Fig. 4a). The increased number of DEGs at day 5 indicates an intensified transcriptional response under prolonged heat stress.

Subsequently, a Venn diagram illustrated the overlap and distinct sets of DEGs across the two time points. Among these, 2507 genes were consistently up-regulated, and 2133 were consistently down-regulated across both time points (Fig. 4b), suggesting the presence of shared heat-responsive gene sets that may contribute to sustained thermotolerance in Chinese cabbage.

The functional categorization of up-regulated DEGs under heat stress exhibited distinct patterns at day 2 (2-H vs. 2-O) and day 5 (5-H vs. 5-O). Gene Ontology (GO) analysis of up-regulated genes at day 2 revealed that in the BP (Biological Process) category, RNA modification (GO:0009451) was the most enriched term, with 96 genes involved. In the CC (Cellular Component) category, mitochondrion (GO:0005739) showed the highest level of enrichment, encompassing 142 genes. In the MF (Molecular Function) category, protein binding (GO:0005515) was the most significantly enriched term, with 568 genes involved (Fig. 5a). At day 5, the enrichment pattern had shifted, with translation (GO:0006412) becoming the most enriched BP term, involving 163 genes. Mitochondrion (GO:0005739) remained the dominant CC term, increasing to 196 genes, and protein binding (GO:0005515) continued to be the most enriched MF category, now with 676 genes (Fig. 5a). These results suggest a transition from early regulatory processes to enhanced protein synthesis and energy metabolism during prolonged heat exposure.

For down-regulated genes, GO analysis indicated a consistent repression of signaling-related processes. On day 2, protein phosphorylation (GO:0006468) was the most suppressed BP term (182 genes), while the extracellular region (GO:0005576) and protein kinase activity (GO:0004672) were the leading CC and MF terms, respectively (Fig. 5b). By day 5, repression intensified, with protein phosphorylation (GO:0006468) expanding to 324 genes, membrane components (GO:0016020) emerging as the most affected CC category (1224 genes), and DNA-binding transcription factor activity (GO:0003700) dominating the MF category (267 genes) (Fig. 5b).

KEGG pathway analysis further highlighted these temporal shifts. Among up-regulated genes on day 2, the top enriched pathways were “Protein Processing in the Endoplasmic Reticulum” (81 genes), “Plant Hormone Signal Transduction” (71 genes), and “Spliceosome” (64 genes). On day 5, “Ribosome” (186 genes) became the most significantly enriched pathway, alongside sustained enrichment of “Protein Processing in the Endoplasmic Reticulum” (118 genes) and “Spliceosome” (117 genes) (Fig. 5c).

In contrast, down-regulated DEGs were predominantly enriched in pathways associated with energy metabolism and hormonal signaling. On day 2, the top suppressed pathways included “Plant Hormone Signal Transduction” (92 genes), “Ribosome” (81 genes), and “Oxidative Phosphorylation” (56 genes). By day 5, “Plant Hormone Signal Transduction” remained significantly repressed (146 genes), followed by “Photosynthesis” (72 genes) and “Plant-Pathogen Interaction” (67 genes) (Fig. 5d). Collectively, these findings indicate a coordinated shift from early stress perception and regulatory adjustments to metabolic reprogramming and energy conservation strategies under prolonged heat stress.

Transcription factors were classified based on the DEG analysis to evaluate their differential expression under heat stress conditions. A total of 24 TF families were identified as being upregulated by more than twofold on both day 2 and day 5, whereas 16 families were found to be downregulated (Fig. 6).

This is broad representation of TF families highlights the extensive transcriptional reprogramming in Chinese cabbage in response to prolonged heat stress. Previous studies have demonstrated that TF families such as HSF, NAC, bZIP, HD-ZIP, MYB, and bHLH are critically involved in plant responses to various abiotic stresses, including heat stress [19,20,21,22].

Next, we selected TFs with |FC| > 10 on both day 2 and day 5 (Table 2) to identify those that are highly sensitive to heat stress. Among these, four TFs belonging to the HSF family were identified, and the TF with the highest FC value on day 5 (LOC103862791) was designated as BrHSF B-1. For the NAC family, two TFs (LOC103829311 and LOC103875105) were identified and named BrNAC13 and BrNAC59, respectively. The bZIP family included one TF (LOC103853909), which was named BrbZIP43. For the HD-ZIP family, one TF (LOC103842010) was identified and designated as BrATHB-12. The C2H2 family included one TF (LOC103859816), which was named BrAZF2. For the COL family, two TFs (LOC103874939 and LOC103868662) were identified and named BrCOL10 and BrCOL8, respectively. The bHLH family included one TF (LOC103874966), which was designated as BrbHLH78. Lastly, the MYB family included one TF (LOC103855087), which was designated as BrMYB34. The selected TFs span diverse functional families, indicating that heat stress modulates a broad array of transcriptional regulators. In total, six up-regulated TFs and four down-regulated TFs were selected for further expression pattern analysis.

To validate the RNA-seq results and investigate the expression patterns of 10 selected transcription factors with significant fold change values on day 2 (2d) and day 5 (5d) in Chinese cabbage, time-course qPCR was conducted under both prolonged heat stress and short-term conditions. The analysis was performed using RNA extracted from the leaves of 11-day-old (seedling stage) and 21-day-old (vegetative stage) plants.

During prolonged exposure, spanning 9 days, transcription factors previously identified as up-regulated in the RNA-seq analysis exhibited substantial increases in transcript levels. These up-regulated genes, including BrHSF B-1, BrNAC13, BrbZIP43, BrATHB-12, BrAZF2, and BrCOL10, showed marked induction in the heat-treated group (H_group), with pronounced increases observed at 2d and 5d following heat treatment. In the H_group, expression levels increased by 10- to 135-fold at 2d compared to the control group (O_group), and further escalated to 45- to 775-fold at 5d relative to the O_group (Fig. 7a–f). Among these, BrHSF B-1 displayed the most pronounced increase, being markedly upregulated, reaching approximately 120-fold at 2d, 1013-fold at 5d, and peaking at 1273-fold at 7d relative to 0d in the H_group. Meanwhile, the O_group exhibited no noticeable changes in BrHSF B-1 expression at either 2d or 5d compared to 0d (Fig. 7a).

BrNAC13 expression in the O_group remained largely unchanged at 2d and increased moderately by 5-fold at 5d. In contrast, expression in the H_group was markedly upregulated, reaching approximately 45-fold at 2d and 346-fold at 5d, before peaking at 887-fold at 9d (Fig. 7b). BrbZIP43 followed a similar trend, showing little change in the O_group at 2d but increasing by 5-fold at 5d. In the H_group, expression was markedly upregulated, recording an increase of about 124-fold at 2d and 243-fold at 5d, and reaching a peak of 283-fold at 7d (Fig. 7c). BrATHB-12 expression in the O_group showed minimal variation at 2d and a modest 5-fold increase at 5d. Conversely, expression in the H_group was markedly upregulated, reaching approximately 32-fold at 2d and 607-fold at 5d (Fig. 7d). BrAZF2 in the O_group exhibited a slight increase of 1.3-fold at 2d and 3.7-fold at 5d, whereas in the H_group, expression was markedly upregulated, with levels surging by 14-fold at 2d and dramatically climbing to 743-fold at 5d (Fig. 7e). BrCOL10 expression in the O_group remained unchanged, whereas in the H_group, expression was markedly upregulated, increasing by about 72-fold at 2d and peaking at 236-fold at 5d (Fig. 7f).

In contrast, the down-regulated genes identified by RNA-seq, including BrNAC59, BrCOL8, BrbHLH78, and BrMYB34, generally exhibited lower expression levels in the H_group compared to the O_group. BrNAC59 expression increased modestly in both groups during the early stages, by approximately 1.7-fold at 2d and 7.2-fold at 5d in the H_group relative to 0d, with no clear difference between the two groups. However, divergence became apparent at 9d, when expression increased 73-fold in the O_group but only 43-fold in the H_group, indicating a 1.7-fold lower level in heat-treated plants (Fig. 7g). BrCOL8 expression declined by approximately 14 percent in the O_group and 55 percent in the H_group at 2d relative to 0d, resulting in a 1.9-fold greater reduction in the H_group. At 5d, expression in the O_group increased by about 2.3-fold, whereas in the H_group it decreased by 13-fold, leading to a 31-fold difference between the groups (Fig. 7h). BrbHLH78 expression was reduced by 31 percent in the O_group and by 78 percent in the H_group at 2d, corresponding to a 3.1-fold greater decrease in the H_group. Although the difference between the groups narrowed to approximately 1.17-fold at 5d, a substantial gap reappeared at 6d, with expression in the H_group approximately fourfold lower than in the O_group (Fig. 7i). BrMYB34 exhibited the most pronounced suppression, with expression in the H_group decreasing 37-fold relative to the O_group as early as 1d. Transcript levels remained low until 6d, after which a gradual increase was observed at 7d and 9d (Fig. 7j).

Additionally, to examine the expression profiles of transcription factors during the vegetative stage under short-term heat stress, qPCR analysis was performed using leaves from 21-day-old plants collected at 0, 1, 3, 6, and 12 h after heat treatment. The analysis revealed that all six transcription factors previously identified as upregulated under prolonged stress reached their peak expression levels as early as 1 h following heat exposure (Fig. S1a–f). In contrast, the four transcription factors classified as downregulated during prolonged heat stress exhibited variable expression patterns under short-term conditions. BrNAC59 and BrbHLH78 consistently showed lower expression in heat-treated plants compared to controls throughout the treatment period, with BrNAC59 declining by approximately 70% and BrbHLH78 by 93% at 12 h (Fig. S1g,i). Meanwhile, BrCOL8 and BrMYB34, which were suppressed under prolonged stress, exhibited transient upregulation at early time points following heat exposure (Fig. S1h,j). However, as the prolonged and short-term heat treatments were conducted at different developmental stages, direct comparisons of expression patterns between these conditions should be made with caution.