Kale plants with different leaf colors (purple, green, and pink) were purchased from the same online supplier (Taobao). The three color groups were confirmed to be from the same cultivar and were purchased in the same batch. The plants were approximately 4 - 6 weeks old at the time of sampling and had 3 - 4 fully expanded young leaves. Therefore, the three color groups were treated as different color phenotypes of the same cultivar under the same experimental conditions. The observed differences in leaf coloration were considered to be primarily associated with differences in anthocyanin accumulation. However, other physiological or environmental factors may also have contributed to the observed differences among the color groups.

For each color group, six individual plants were selected. Two plants of the same color were transplanted into one pot, resulting in three pots per color group. Each pot was approximately 15 cm in diameter and contained commercially available nutrient-rich soil. The pots were placed in a controlled indoor environment and maintained for three days at 24˚C under a 12-hour light/12-hour dark photoperiod. Light intensity was kept consistent during the daytime, and all plants received equal amounts of water daily to maintain adequate soil moisture. After three days, samples were collected from the third to fourth fully expanded leaves of each plant between 9:00 and 11:00 AM to reduce variation associated with diurnal changes. The distal portion of each leaf was carefully excised using sterilized scissors, immediately transferred into tubes, rapidly frozen in liquid nitrogen, and stored at −80˚C until further analysis.

Total RNA was extracted from plant samples using the Plant RNA Kit (Omega, Cat. No. R6827-02) following the manufacturer’s instructions. The procedure was as follows:

All 18 samples (6 of green, 6 of pink, 6 of purple) were removed from liquid nitrogen, and the samples were pulverized with a bead mill to ensure that all the leaf tissues were ground effectively. (Purpose: to break open cells and release RNA while maintaining low temperature to inhibit RNase activity and prevent RNA degradation.) The ground tissue was transferred into 1.5 mL RNase-free microcentrifuge tubes. 500 µL of RCL Buffer was added to each tube, and the samples were vortexed thoroughly, then kept on ice to prevent RNA degradation. Samples were incubated at 55˚C for 2 minutes, followed by centrifugation at 10,000 × g for 5 minutes. The supernatant was transferred to a gDNA Filter Column placed in the corresponding collection tube, and centrifuged at 14,000 × g for 2 minutes. The filter column was discarded, and an equal volume of RCB Buffer was added to the collection tube; the solution was mixed thoroughly by pipetting. 700 µL of the mixture was transferred to a HiBind RNA Mini Column placed in a collection tube and centrifuged at 12,000 × g for 1 minute. The flow-through was discarded, and the step was repeated until all the sample solution had passed through the column. 400 µL of RWF Wash Buffer was added to the column, followed by centrifugation at 10,000 × g for 30 seconds. 500 µL of RNA Wash Buffer was added, followed by centrifugation at 10,000 × g for 30 seconds; this step was repeated once. The flow-through was discarded, and the column was centrifuged at maximum speed for 1 minute. The column was placed into a new 1.5 mL microcentrifuge tube, and 50 µL of RNase-free water was added to the column membrane. After standing for 5 minutes, the tube was centrifuged at maximum speed for 1 minute. The column was discarded, and the RNA samples were stored at −20˚C until further use.

cDNA was synthesized using the EvOM-MLV RT Mix Kit with gDNA Clean for qPCR (AG, Cat. No. AG11728) according to the manufacturer’s instructions. To eliminate any residual genomic DNA in the RNA samples, preventing false-positive signals during qPCR, the residual genomic DNA was removed by using the reaction mixture prepared as Table 1. The reaction was incubated at 42˚C for 2 minutes. The reverse transcription reaction was performed in a reagent system as Table 2 in a PCR thermocycler under the following conditions: 37˚C for 15 minutes, followed by 85˚C for 5 seconds. After completion, cDNA was stored at −20˚C.

Table 1.Reagent system of removing residual genomic DNA.

Table 2.Reaction system of reverse transcription.

Quantitative PCR was performed using the SYBR® Green Premix Pro Taq HS qPCR Kit (AG, Cat. No. AG11701), with ACT used as the internal reference gene. A total of four pairs of primers were designed the NCBI online primer tools (Table 3).

The reaction mixture was as prepared as Table 4. After preparing the reaction mixture, the PCR plate was sealed with a single-sided adhesive film and centrifuged at 3,700 × g for 5 minutes to remove air bubbles and ensure uniform distribution (Table 5).

Ct value measures the number of cycles needed for the detection of gene expression, with lower Ct values representing higher expression levels.

qPCR quality control

RNA quality was assessed by measuring the A260/A280 ratio, which ranged from 1.8 to 2.1 across samples. Each biological sample was analyzed in three technical replicates. Amplification specificity was assessed using melting curve analysis, and each primer pair produced a single sharp melting peak. ACT Ct values ranged from 20.71 to 25.88 across all samples and were used as the reference for normalization. However, reference-gene stability was not formally evaluated using a dedicated stability analysis. No-template controls (NTCs) and no-reverse-transcription controls were not included in the qPCR analysis, which limited independent assessment of potential contamination and genomic DNA amplification.

Table 3.qPCR primers.

Table 4. qPCR reaction mixture preparation.

Table 5. qPCR amplification conditions.

Approximately 50 mg of plant leaf tissue was weighed and transferred into a 2 mL microcentrifuge tube. A total of 600 µL of extraction solution (methanol:HCl = 99:1, v/v) was added, and the samples were incubated overnight at 40˚C in the dark with gentle shaking. After incubation, 400 µL of deionized water and 400 µL of chloroform were added to each sample, followed by gentle inversion to mix thoroughly. The samples were then centrifuged at 12,000 rpm for 2 minutes. The supernatant was collected, and absorbance was measured at 530 nm and 657 nm using a spectrophotometer. Anthocyanin content was calculated using the following formula:

Results are expressed as cyanidin-3-glucoside equivalents (mg C3GE/g FW) based on the molar extinction coefficient (ε = 26,900 L·mol⁻¹·cm⁻¹) of cyanidin-3-glucoside. Six biological replicates were analyzed per color group.

A total of 5 mL of dd water was added into the measurement chamber. The background electrical conductivity was measured and ensured to be below 10 μS/cm before starting the experiment. The measurement chamber was inserted into the aluminum slot with slight pressure to confirm proper positioning and stable contact. A magnetic stirring bar was placed into the chamber to maintain continuous mixing during the measurement. The conductivity probe was also inserted into the solution, ensuring that it was completely immersed and did not attach to the walls of the chamber. Before adding the sample, the instrument parameters were set as: Conductivity range: 0 - 10 μS/cm, Ft parameter: 0 - 100. Leaf samples (approximately 1 cm in diameter) were cut using scissors and fixed in the frame, then inserted into the sample holder. The leaf samples were positioned to maintain a full contact with the measurement solution for effective electrolyte release. After sample insertion, the instrument parameter was adjusted to: Ft = 2000. The experiment was then started from room temperature (approximately 25˚C) up to 70˚C at a linear heating rate of 3˚C /min. During the measurement, the solution was continuously stirred at a low speed, and electrical conductivity was monitored and recorded in time. Excessive stirring was avoided to prevent fluctuations in the readings. Six biological replicates were analyzed per color group, with three leaf discs per replicate. The critical temperature (Tcrit) was determined as the inflection point of the conductivity-temperature curve, where the rate of electrolyte leakage reaches its maximum. The curve was fitted using a Logistic equation of the form Y = k/(1 + a*e^(−b*x)), where Y is the relative conductivity, x is the temperature, and k, a, and b are fitted parameters. The changes in electrical conductivity were used to evaluate membrane stability and heat tolerance of the kale plant samples.

In this study, the expression levels of the anthocyanin biosynthesis-related genes in three kale color groups (green, pink, and purple) were quantitatively analyzed using qPCR technology, with six biological replicates per color group and three technical replicates per sample. Actin (ACT) was employed as the internal reference gene, and the green sample was used as the control group to calculate relative expression levels. The expression levels of the ANS4 gene showed significant variation across the different colored kale samples, as calculated using the 2^(-ΔΔCt) method. Green Samples (Control Group): The expression of ANS4 was lowest comparing to the others in the green samples. The initial relative expression level for this group was defined as the baseline (approximately 1.0). Pink Samples: The expression of ANS4 was significantly higher in the pink samples comparing with the green samples, with average expression levels approximately 10-fold higher than those of the green samples. Purple Samples: The highest expression levels were observed in the purple samples. Based on the calculated data, the expression of ANS4 in the purple samples was markedly upregulated (P < 0.05), reaching a relative expression level more than 40-fold that of the green sample (Figure 1).

The expression profile of bHLH1revealed striking differences associated with leaf coloration in kale. While the green samples exhibited baseline expression levels, the gene was significantly up-regulated in both pink and purple samples. Specifically, bHLH1 expression in pink samples was approximately 15-fold higher than that of the green control. The purple samples displayed the highest transcript abundance, with a 21-fold increase relative to the control group. The statistical significance of these elevations was confirmed by independent t-tests (P < 0.05) (Figure 1).

The relative expression levels of DFR4 also exhibited significant variation correlating with leaf pigmentation. Using the green samples as a control (baseline expression of 1), a progressive increase in transcript abundance was observed. The pink samples showed a notable 15-fold elevation compared to the control. Notably, the purple samples displayed a dramatic 45-fold increase in expression levels relative to the green samples. DFR4 expression showed an increasing pattern across the green, pink, and purple groups, with higher expression observed in the more intensely colored groups. (P < 0.05 for pink, purple and green) (Figure 1).

(a) (b) (c)

Figure 1. Relative expression levels of ANS4, bHLH1,andDFR4 in green, pink, and purple kale. Asterisks indicate significant differences compared to the green group (P < 0.05). Data are mean ± SD.

The total anthocyanin content was quantified to investigate the physiological basis of the distinct leaf colorations in the three kale variants (Figure 2). The results revealed that Anthocyanin content increased with the intensity of leaf pigmentation across the three color groups. In the green kale samples, anthocyanin levels were minimal, serving as the baseline for this study (average: 0.23 0.04 mg/g FW). In contrast, the colored variants showed a substantial increase in pigment deposition. The anthocyanin content in the pink samples was measured at an average of 3.33 0.51 mg/g FW. The most prominent accumulation was observed in the purple samples, which reached an average concentration of 18.09 mg/g FW. Statistical analysis confirmed that these differences were highly significant (P < 0.05), demonstrating that the progressive intensification of leaf pigmentation from green to pink and purple was consistent with substantial anthocyanin accumulation (Figure 2).

Figure 2. Total anthocyanin content in green, pink, and purple kale leaves. Asterisks indicate significant differences compared to the green group (P < 0.05). Data are mean ± SD.

To evaluate the heat tolerance of different kale variants, the relationship between temperature and electrolyte leakage (measured as conductivity) was analyzed to identify the critical thermal points (Figure 3). The Conductivity Curvature peak represents the critical temperature (T_crit) at which cell membrane integrity begins to fail, leading to rapid electrolyte efflux. As illustrated in Figure 3, significant variations in membrane thermostability were observed among the three samples. The green samples exhibited the highest thermal stability, with a critical point identified at 63.26 degree. The critical temperature for the pink samples was recorded at 59.58 degree, which is significantly lower than that of the green control. Similarly, the purple samples revealed a critical point of 59.29 degree, a value closely aligned with the pink samples but notably lower than the green variant.

Figure 3. Heat resistance test and cell membrane stability analysis of green, purple, and pink kale leaves.

Figure 4. Critical temperatures of green, pink, and purple kale leaves. Asterisks indicate significant differences compared to the green group (P < 0.05). Data are mean ± SD.

The thermal stability of the cell membranes across the three kale varieties was further evaluated by determining the critical temperature (T_conductivity), defined as the threshold where membrane integrity is compromised (Figure 4). Consistent with the conductivity-temperature profiles, the green kale samples demonstrated the highest thermal threshold, with a T_conductivity of approximately 63.26˚C. In contrast, both the pink and purple samples exhibited significantly lower critical temperatures, recorded at 59.58˚C and 59.29˚C, respectively. Statistical analysis revealed that the reduction in thermal tolerance for both colored variants was highly significant compared to the green control (P < 0.05) (Figure 4).

Through the data analyzed in qPCR process, there is an positive association between leaf coloration intensity and the expression of anthocyanin-related genes [14]. The darker the color of the leaves of kale, the expression of anthocyanin-related genes will increase accordingly [15]. The green samples with less color content show relatively low expression of anthocyanin-related genes, and as the color gets darker, the purple samples indicate the highest expression of anthocyanin-related genes. The results of this study suggest that the expression levels of anthocyanin biosynthesis-related genes (ANS4,bHLH1,and DFR4) increase with the intensity of leaf coloration in kale [16]. This relationship can be explained by the role of these genes in the anthocyanin biosynthesis pathway. bHLH1 serves as a transcription factor regulating the expression of anthocyanin-related genes. ANS4 and DFR4 are employed as key enzymes for anthocyanin synthesis in kale [17]. The increased expression ofbHLH1 likely played a regulatory role in anthocyanin biosynthesis by activating ANS4 and DFR4, which encode key enzymes responsible for anthocyanin synthesis. Therefore, their regulation promotes more intense darker coloration, which is extensively observed in the purple samples.

As shown in Figure 2, the highest content of anthocyanin occurs in purple samples, medium in pink samples, and lowest in the green samples. This reveals a positive association between leaf coloration intensity and anthocyanin content in the leaves [15]. This discovery can be explained by anthocyanin being a plant pigment that controls the coloration of the leaves, with most anthocyanin concentrated in the leaves, consistent with previous studies [18]. Darker colored leaves are associated with higher levels of anthocyanin, regulated by the upregulation of genes involved in the biosynthesis pathway.

The heat resistance measurements showed that the green group had the highest Tcrit (63.26˚C), whereas the pink and purple groups had lower Tcrit values of 59.58˚C and 59.29˚C, respectively. These results indicate greater membrane thermostability in the green group under the tested conditions [19][20]. The lower Tcrit values observed in the pink and purple groups coincided with their higher anthocyanin contents. Anthocyanins are known to be sensitive to temperature and may undergo degradation under high-temperature conditions [21][22]. Therefore, the observed association between higher anthocyanin content and lower membrane thermostability may be related to differences in pigment stability. However, because the specific cultivars and genetic backgrounds of the plants were not available or controlled, the lower Tcrit values cannot be attributed solely to anthocyanin accumulation. Other physiological differences among the kale groups may also have contributed to the observed variation in heat tolerance [23][24].

This study investigated the relationship between anthocyanin-related gene expression, pigment accumulation, and heat tolerance in kale and found significant differences among green, pink, and purple samples. The qPCR results showed that the upregulation of the anthocyanin biosynthesis pathway was associated with more intense leaf coloration, which helps explain the higher anthocyanin content observed in purple samples [25]. In addition, anthocyanin content was closely related to leaf coloration, indicating that darker leaves generally accumulate higher levels of anthocyanin [26]. However, in the present study, higher anthocyanin-related gene expression and anthocyanin accumulation coincided with lower Tcrit values, indicating lower membrane thermostability under the tested conditions [27]. Green leaves showed relatively higher heat tolerance, while pink and purple leaves, which contained higher anthocyanin levels, exhibited lower tolerance to elevated temperatures [28]. This pattern may be related to the thermal instability of anthocyanins, although other physiological differences among the kale groups may also have contributed to the observed variation [29].

The initial experimental design aimed to investigate the relationship between anthocyanin-related gene expression and antioxidant capacity in kale, with the goal of exploring the nutritional value of different kale varieties [30]. However, due to time constraints and the unavailability of the antioxidant capacity assay kit, the research focus was shifted from anthocyanin gene expression and antioxidant capacity to anthocyanin gene expression and heat tolerance, as the latter could be more readily measured in the laboratory. As a result, the relationship between anthocyanin gene expression and antioxidant capacity in kale remains to be further elucidated. The study also analyzed only a limited number of anthocyanin biosynthesis-related genes, which may not fully represent the complexity of the anthocyanin biosynthetic pathway.

Future studies could further investigate the relationship between anthocyanin-related gene expression and antioxidant capacity to provide a more comprehensive understanding of the nutritional characteristics of kale. RNA sequencing could also be used to examine global gene expression changes under heat stress and identify pathways associated with differences in heat responses among kale groups [26]. Protein-level analyses, such as Western blotting, could be conducted to determine whether differences in transcript abundance are reflected at the protein level for key enzymes involved in anthocyanin biosynthesis, such as DFR4 and ANS4[31]. The findings of this study may provide a preliminary reference for investigating kale responses to high-temperature environments such as those found in Guangdong [32]. However, the observed differences in Tcrit were obtained under the specific experimental conditions used in this study and may provide only limited implications for kale cultivation under the high-temperature conditions of Guangdong [27]. Further studies using larger sample sizes, independent experimental units, longer-term heat treatments, and field or greenhouse experiments are needed to evaluate the cultivation performance of different kale color phenotypes under high-temperature conditions.