Two genetically stable pure lines of Chinese cabbage yellow cotyledon mutant 19YC-2 and normal green cotyledon near isogenic line (wild type) 19GC-2 were used and provided by Cruciferae Research Group of Northwest A&F University (Yangling, China). The seedlings were grown in Caoxinzhuang vegetable experimental farm (Yangling, China). Five individual seedlings were selected for phenotype statistics at the cotyledon stage, the two-leaf stage, the four-leaf stage, the six-leaf stage, the rosette stage, and the heading stage, respectively. The shoots were harvested to measure the fresh weight at different stages and then dried at 65°C with 12 h for measuring the dry weight. Three independent replicates were performed for each sample at every stage.

For different light intensity treatments, the germinated seeds of 19YC-2 and 19GC-2 were sowed in plant growth chamber with 16 h light (24°C, 216 μmol·m⁻²·s⁻¹), 8 h dark (16°C), and 55% relative humidity. After the cotyledon of seedlings were fully expanded, the seedlings were separately subjected to the conditions under low-intensity light (72 μmol·m⁻²·s⁻¹) and normal intensity light (216 μmol·m⁻²·s⁻¹) conditions, with the same other conditions. The leaf color observation and sampling were performed at the cotyledon stage, the first-leaf stage, the two-leaf stage, the three-leaf stage, and the four-leaf stage, respectively. The cotyledons and leaves for RNA extraction were collected, immediately frozen in liquid nitrogen and stored at −80°C.

Transmission electron microscopy (TEM) was performed to observe the chloroplast ultrastructure of the cotyledons and leaves of 19YC-2 and 19GC-2. The cotyledons were used from 19YC-2 and 19GC-2 at the cotyledon stage. The 1^st leaves of 19YC-2 and 19GC-2 were separately collected at the two-leaf stage and the four-leaf stage. The samples in size with 1 mm² were fixed in a mixture of 2% glutaraldehyde fixative in 0.2 M phosphate buffered saline (PBS). The fixed samples were then dehydrated and embedded in resin as previously decreased (Nie et al., 2023). Ultrathin sections were obtained and examined using a transmission electron microscopy (Hitachi HT7800, Japan).

The color difference parameters of cotyledons and leaves were measured by a colorimeter (Konica Minolta CR-400, Japan). A CIE Lab color system was used to examine the L*, A* and B* values. The L* value represents leaf brightness. The A* value represents the red and green, where a positive value indicates red and a negative value indicates green. The B* value represents the yellow and blue, where a positive value indicates yellow and a negative value indicates blue.

The contents of photosynthetic pigments were determined as previously reported methods (Lichtenthaler, 1987; Yang et al., 2023). The fresh leaves with 0.3 g were cut and mixed in a 50 mL tube containing 10 mL 95% ethanol in the dark for 24 h. The absorbance was measured by a M200pro microplate reader (Tecan, Switzerland) at wavelengths of 665 nm, 649 nm, and 470 nm, respectively. The contents of carotenoid, chlorophyll a, and chlorophyll b were calculated according to the reported formula by Yang et al. (2023). A Dual-PAM-100 Chlorophyll Fluorometer (Walz, Germany) was used to monitor the chlorophyll fluorescence parameters, including maximum quantum efficiency of photosystem II (Fv/Fm), effective quantum efficiency of photosystem II (Fv’/Fm’), and electron transport in photosystem II [ETR(II)] and nonphotochemical quenching (NPQ). Three individual seedlings were used in each replicate, and three independent replicates were performed for each sample.

A total of 18 cDNA libraries were constructed using six samples from 19GC-2 and 19YC-2, including the cotyledons (CK0 and T0) at the cotyledon stage, and the 4^th leaves after 72 μmol·m⁻²·s⁻¹ (CK1 and T1) and 216 μmol·m⁻²·s⁻¹ (CK2 and T2) light treatments at the four-leaf stage, with three independent biological replicates. The library construction and sequencing for RNA-seq analysis were completed on Illumina sequencing platform by Genedenovo Biotechnology Co., Ltd (Guangzhou, China). Clean reads were mapped to the reference genome of Chinese cabbage V3.0 (http://brassicadb.cn). Pearson correlation coefficients (PCC) and principal component analysis (PCA) were used to evaluate the correlation between samples and repeatability. The abundance of gene expression was quantified as fragments per kilobase of transcript sequence per millions (FPKM) value. The differentially expressed genes (DEGs) were identified using the DEseq 2 R package with the absolute value of log2 fold change ≥ 1 and FDR < 0.05 (Love et al., 2014). TB tools software was used to draw heat maps and Venn diagram. Functional annotation of gene was performed according to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases, and enrichment analysis was performed at P-value ≤ 0.05. The data of RNA-seq were deposited in NCBI Sequence Read Archive (SRA; http://www.ncbi.nlm.nih.gov/Traces/sra/) database under the accession number: BioProject PRJNA1005457.

Total RNA extraction, cDNA synthesis, and qRT-PCR analysis were performed as previously reported methods (Nie et al., 2019). The qRT-PCR analysis was conducted on an iCycler iQ5 Real-Time PCR Detection System (Bio-Rad, USA). Three biological replicates were performed for each sample. The specific primer sequences were designed using Beacon Designer 8.0 software ( Supplementary Table S1 ). The BrEF-1α gene (Bra031602) was selected as the internal control. The relative expression level of gene was calculated using 2^−ΔΔCT method.

To construct a co-expression network and identify the modules of highly correlated DEGs based on the gene expression profiles and pigment contents, weighted gene co-expression network analysis (WGCNA) was performed using the R package with the soft threshold of 8 and the minimum module size of 50 (Langfelder and Horvath, 2008). Correlation analysis between the identified modules and the contents of photosynthetic pigments were determined by Pearson correlation coefficients (PCC).

All the data were analyzed using SPSS Statistics 24.0. The expression data were presented as the means ± SEM. Independent samples t-test was used to determine statistically significant difference.

The leaf phenotypes of Chinese cabbage yellow cotyledon mutant 19YC-2 and green cotyledon line 19GC-2 were observed at different growth stages. The cotyledons and leaves of 19GC-2 were normal green from the cotyledon stage to the heading stage. The cotyledons of 19YC-2 were obvious yellow at the cotyledon stage ( Figure 1A ) and began to turn green at the two-leaf stage ( Figure 1B ). The 1^st leaves of 19YC-2 also exhibited yellow at the two-leaf stage ( Figure 1B ), while it gradually turned green at the four-leaf stage ( Figure 1C ). Interestingly, the newly unfolded leaves of 19YC-2 were still yellow from the four-leaf stage until the rosette stage ( Figures 1C-E ). At the heading stage, all the outer leaves of 19YC-2 were close to green ( Figure 1F ). In addition, the fresh weight and dry weight of shoots in 19YC-2 were significantly lower than that in 19GC-2 at each stage ( Figures 1G, H ).

Observation of chloroplast ultrastructure by TEM showed that there were fewer chloroplasts per cell in the cotyledons of 19YC-2 (approximately 11 chloroplasts) than that in 19GC-2 (approximately 20 chloroplasts) ( Figures 2A, B ). The chloroplasts in 19GC-2 were regular and oval in shape, and the obvious grana lamellae and starch grains were observed ( Figure 2C ). However, the chloroplasts in 19YC-2 were irregular and shrunken in shape and exhibited the lack of grana lamellae and starch grains ( Figure 2D ). Moreover, at the two-leaf stage, the chloroplasts in the 1^st leaves of 19GC-2 were normal in shape and structure ( Figure 2E ), while the 1^st leaves of 19YC-2 were still yellow and exhibited abnormal chloroplasts and a lower number of grana lamellae and starch grains ( Figure 2F ). Remarkably, at four-leaf stage, the 1^st leaves of 19YC-2 had gradually turned green, and the chloroplasts exhibited the normal structures with obvious grana lamellae and increased starch grains ( Figure 2H ), which was almost similar to the normal chloroplasts in 19GC-2 ( Figure 2G ).

To explore the effects of different light intensities on leaf color changes in Chinese cabbage yellow cotyledon mutant 19YC-2, the two light intensities of 72 μmol·m⁻²·s⁻¹ and 216 μmol·m⁻²·s⁻¹ were separately used to treat the seedlings from fully expanded cotyledons ( Figure 3A ). Under 72 μmol·m⁻²·s⁻¹ light condition, both 19YC-2 and 19GC-2 displayed thinner leaves. All the leaves of 19YC-2 gradually turned green until the four-leaf stage, which had few differences with the leaf color in 19GC-2 ( Figure 3B ). Under 216 μmol·m⁻²·s⁻¹ light condition, the heart leaves of 19YC-2 remained yellow at the four-leaf stage ( Figure 3B ). Further analysis of color difference showed that the values of L *, A *, and B * in 19YC-2 cotyledons were significantly higher than that in 19GC-2 at the cotyledon stage ( Figure 3C ). After the treatment under 72 μmol·m⁻²·s⁻¹ light, there was no significant difference in color difference parameters in the 4^th leaves between 19YC-2 and 19GC-2 at the four-leaf stage ( Figure 3D ). However, under 216 μmol·m⁻²·s⁻¹ light condition, the significant differences in color difference parameters in the 4^th leaves between 19YC-2 and 19GC-2 were found at the four-leaf stage ( Figure 3E ). The results of color difference analyses were consistent with the observation of leaf color in 19YC-2 and 19GC-2 after different light intensity treatments.

Analyses of pigment contents showed that the chlorophyll content in 19YC-2 cotyledons was significantly lower than that in 19GC-2 ( Figure 4A ), and the carotenoid content in 19YC-2 cotyledons was significantly higher than that in 19GC-2 ( Figure 4B ). Furthermore, after the treatment using 72 μmol·m⁻²·s⁻¹ light, the chlorophyll and carotenoid contents in the 4^th leaves of 19YC-2 had no significant difference compared to that in 19GC-2 at the four-leaf stage ( Figures 4A, B ). Under 216 μmol·m⁻²·s⁻¹ light condition, the chlorophyll content was significantly lower, and the carotenoid content was significantly higher in the 4^th leaves of 19YC-2 than that in 19GC-2 at the four-leaf stage, which was consistent with the changes in pigment contents at the cotyledon stage ( Figures 4A, B ). Additionally, the ratio of chlorophyll/carotenoid in the 4^th leaves of 19YC-2 under 72 μmol·m⁻²·s⁻¹ light was nearly equal to that in 19GC-2, while the significant lower ratio was found in the 4^th leaves of 19YC-2 under 216 μmol·m⁻²·s⁻¹ light than that in 19GC-2 ( Figure 4C ), which may be responsible for the leaf color differences in 19YC-2 exposed to different light intensities. More importantly, the carotenoid content in the heart leaves of 19YC-2 was gradually decreased with the growth of seedlings ( Figure 4D ). The decrease trend of the level of carotenoid in 19YC-2 after 72 μmol·m⁻²·s⁻¹ light treatment was more rapid, resulting in the nearly equal level of carotenoid with the 4^th leaves of 19YC-2 to that in 19GC-2. The results indicated that low-intensity light treatment caused the rapid decrease in carotenoid content in 19YC-2 may be one of the major reasons for the leaf color changes from yellow to green.

Analyses of chlorophyll fluorescence parameters in Chinese cabbage yellow cotyledon mutant 19YC-2 showed that at the cotyledon stage, the values of Fv/Fm, Fv’/Fm’, and ETR (II) were significantly lower in 19YC-2 cotyledons than that in 19GC-2 ( Figure 4E-G ), while the NPQ had no difference between 19YC-2 and 19GC-2 ( Figure 4H ). The observation revealed that the yellow cotyledons in 19YC-2 exhibited obvious photoinhibition, which may be related to the abnormal chloroplasts and lower chlorophyll content in 19YC-2 cotyledons. After 72 μmol·m⁻²·s⁻¹ light treatment, there was no significant difference in the values of Fv/Fm, Fv’/Fm’, and ETR (II) in the 4^th leaves between 19YC-2 and 19GC-2 at the four-leaf stage ( Figures 4E-G ), while the NPQ was significantly higher in 19YC-2 ( Figure 4H ), indicating that low-intensity light activated the NPQ and induced the raised energy dissipation to defense photodamage in 19YC-2. After 216 μmol·m⁻²·s⁻¹ light treatment, the values of Fv/Fm, Fv’/Fm’, and ETR (II) were also significantly lower in the 4^th leaves of 19YC-2 at the four-leaf stage ( Figures 4E-G ), and the NPQ had no difference between 19YC-2 and 19GC-2 ( Figure 4H ), which was consistent with the results at the cotyledon stage.

The cotyledons of 19GC-2 (CK0) and 19YC-2 (T0) and the 4^th leaves of 19GC-2 (CK1 and CK2) and 19YC-2 (T1 and T2) after 72 and 216 μmol·m^-2·s^-1 light treatment were used for constructing six cDNA libraries and RNA-seq analysis ( Figures 5A, B ). After filtering the original data, the average clean data ranged from 38.84 to 44.02 million were obtained in six cDNA libraries, the average GC contents were ranged from 47.17% to 48.46%, and the average Q30 values were ranged from 92.85% to 93.46% ( Supplementary Table S2 ). These obtained reads were mapped to the reference genome of Chinese cabbage with the average comparison efficiency ranged from 91.75% to 92.78% ( Supplementary Table S3 ). PCA analysis validated the high reliability among three biological replications in the six sample groups ( Figure 5C ). Statistics of the number of DEGs showed that 3015 DEGs were identified in the comparison group of CK0 vs T0 (1376 upregulated and 1639 downregulated DEGs), 1084 DEGs were identified in the comparison group of CK1 vs T1 (391 upregulated and 693 downregulated DEGs), and 817 DEGs were identified in the comparison group of CK2 vs T2 (427 upregulated and 390 downregulated DEGs) ( Figure 5D ). Additionally, 169 common DEGs were obtained among three comparison groups of CK0 vs T0, CK1 vs T1, and CK2 vs T2, 345 common DEGs were obtained among the comparison groups of CK0 vs T0 and CK1 vs T1, 314 common DEGs were obtained among the comparison groups of CK0 vs T0 and CK2 vs T2, and 290 common DEGs were obtained among the comparison groups of CK1 vs T1 and CK2 vs T2 ( Figure 5E ).

Functional annotation and enrichment analysis showed that these DEGs were classified into 48 (CK0 vs T0), 45 (CK1 vs T1), and 43 (CK2 vs T2) GO subcategories ( Supplementary Figure S1 ). Moreover, 662 (CK0 vs T0), 224 (CK1 vs T1) and 171 (CK2 vs T2) DEGs were annotated on 119, 94, and 89 KEGG pathways ( Supplementary Figure S2 ), respectively. Among these pathways, the metabolic pathway (ko01100) and biosynthesis of secondary metabolites (ko01110) showed significant enrichment in all the three comparison groups. Importantly, the pathways of photosynthesis antenna proteins (ko00196) and carotenoid biosynthesis (ko00906) showed significant enrichment in the comparison groups of CK0 vs T0 and CK2 vs T2, while the two pathways were not found in the comparison group of CK1 vs T1. In addition, DEGs related to porphyrin and chlorophyll metabolism (ko00860) was enriched only in the comparison group of CK0 vs T0, while no DEGs were found in CK1 vs T1 and CK2 vs T2. Furthermore, gene set enrichment analysis (GSEA) also showed that the pathways of photosynthesis antenna proteins (ko00196) and carotenoid biosynthesis (ko00906) exhibited striking enrichment in the comparison groups of CK0 vs T0 and CK2 vs T2, while not in CK1 vs T1 ( Supplementary Figure S3 ). The results of enrichment analyses suggested that DEGs encoding photosynthesis antenna proteins and the differential expressed carotenoid biosynthesis genes may be directly related to the leaf color difference between 19YC-2 and 19GC-2 under different light intensity.

Photosynthetic antenna proteins had been documented to be capable of binding pigments, turning their excitation energies, and directing energy transfer to the reaction center of photosystem (Ruban, 2015). In this study, a total of 17 DEGs were identified and enriched in the pathway of photosynthesis antenna proteins ( Supplementary Table S4 ). Expression analysis showed that 15 DEGs exhibited downregulated expression in the comparison group of CK0 vs T0, and 8 DEGs exhibited downregulated expression in the comparison group of CK2 vs T2, while no DEGs were identified in the comparison group of CK1 vs T1. Remarkably, the expression levels of six transcripts, including BrCAB1 (BraA03g018080.3C), BrCAB13 (BraA10g012410.3C), BrCAB36 (BraA02g037860.3C), BrLHCB1.3 (BraA07g010770.3C, BraA07g010780.3C), and BrLHCB4.2 (BraA05g037360.3C), were significantly downregulated in both CK0 vs T0 and CK2 vs T2. To validate the differential expression of these DEGs, 5 genes were selected for qRT-PCR analysis. The results showed that all the selected genes were significantly decreased in T0 compared with CK0 and in T2 compared with CK2 ( Figure 6 ), which was consistent with the results from RNA-seq analysis.

The pathway enrichment analysis of DEGs indicated that carotenoid metabolic pathway may be associated with the leaf color differences in 19GC-2 and 19YC-2 under different light intensity treatment. As expected, a total of 25 DEGs were identified and involved in carotenoid biosynthesis ( Figure 7 ; Supplementary Table S5 ). Among these DEGs, six upregulated and 16 downregulated genes were found in the comparison group of CK0 vs T0, nine upregulated genes were found in the comparison group of CK1 vs T1, and five upregulated and three downregulated genes were found in the comparison group of CK2 vs T2 ( Supplementary Table S5 ). Moreover, two genes of BrPDS (BraA08g023090.3C) and BrLCYE (BraA01g034180.3C) were significant upregulated expression in the comparison groups of CK0 vs T0 and CK2 vs T2, and three genes of BrAAO3 (BraA04g019480.3C), BrCCD4 (BraA01g010560.3C), and BrNCED3 (BraA05g032610.3C) were significant downregulated expression in the comparison groups of CK0 vs T0 and CK2 vs T2.

Further qRT-PCR analysis showed that the expression levels of BrPDS and BrLCYE genes that were the upstream genes of carotenoid biosynthesis, were increased in the three comparison groups and showed significant increases in CK0 vs T0 and CK2 vs T2 ( Figures 8A, B ). The expression levels of BrAAO3, BrCCD4, and BrNCED3 genes that were the downstream genes of carotenoid biosynthesis, were significantly decreased in CK0 vs T0 and CK2 vs T2 ( Figures 8C-E ). The results indicated that the upregulation of upstream genes and the downregulation of downstream genes were accordant with the changes in carotenoid contents. In addition, the results of qRT-PCR analysis in gene expressions were almost identical to the results of RNA-seq analysis (R² = 0.7784) ( Figure 8F ), demonstrating the reliability of RNA-seq data.

To further investigate the relationships of gene differential expression and carotenoid content, WGCNA was performed to identify the highly correlated modules of DEGs. As a result, 19 modules named from A to S and labeled by different colors were identified ( Figure 9 ). The A (r = 0.666), J (r = 0.714), and K (r = 0.728) modules containing 97, 1371, and 1828 DEGs, respectively, were highly positively related to the changes in carotenoid content, and the D (r = 0.597) module containing 175 DEGs was highly negatively related to the changes in carotenoid content. In addition, the upregulated expressed BrPDS and BrLCYE genes were assigned to the positively related J module, which was consistent with the positive regulation of the two genes in carotenoid synthesis. The downregulated expressed BrAAO3 and BrCCD4 were assigned to the negatively related F (r=−0.307) module, and the downregulated expressed BrNCED3 was assigned to the negatively related M (r=−0.393) module. The downregulated expressed BrCAB1 was assigned to the negatively related F (r=−0.307) module, and the downregulated expressed BrCAB13, BrCAB36, BrLHCB1.3 and BrLHCB4.2 were assigned to the negatively related M (r=−0.393) module. The results suggest that these DEGs identified in these highly correlated modules may play potential roles in regulating the pathway of carotenoid biosynthesis and metabolism.

Furthermore, a total of 56 DEGs belonging to transcription factors (TFs) were identified in these highly correlated modules by WGCNA ( Figure 10 ). Among these differentially expressed TFs, most of these TFs were upregulated expression in the comparison group of CK0 vs T0, and seven genes were downregulated expression. Correlation analysis showed that the differential expressions of 17 TFs were strongly correlated with the changes in carotenoid contents ( Figure 10B ; Supplementary Table S6 ). For instance, the gene expression of BraA08g033820.3C (BrRF2b), BraA08g001340.3C (BrRAP2-12), BraA04g017150.3C (BrERF008), and BraA06g042530.3C (BrR2R3-MYB) had highly positive correlation (r > 0.91) to the changes in carotenoid contents, and the gene expression of BraA08g003010.3C (BrSCL3) and BraA07g009390.3C (BrbHLH77) had highly negative correlation (r < −0.9) to the changes in carotenoid contents. Moreover, many genes were identified as the transcription factor families of BrERF/AP2 (BraA08g001340.3C, BraA04g017150.3C, BraA10g029700.3C, BraA09g009130.3C, and BraA09g040790.3C), BrR2R3-MYB (BraA06g042530.3C), BrbHLH (BraA10g001670.3C, BraA09g011350.3C, and BraA07g009390.3C), and BrWRKY (BraA04g018060.3C and BraA02g028540.3C) ( Supplementary Table S6 ). In general, these identified DEGs and TFs that highly related to carotenoid biosynthesis and metabolism may play important roles in regulating leaf color changes in Chinese cabbage yellow cotyledon mutant 19YC-2.

Leaf color formation is closely related to the contents of photosynthetic pigments (Terry et al., 2001). Leaf color mutants are easy to recognize in leaf phenotypes and often used as valuable materials for investigating pigment biosynthesis and enriching the germplasms in genetic breeding (Deng et al., 2012). Various leaf color mutants usually impact the accumulation and metabolism of photosynthetic pigments and have been characterized in many horticultural crops (Wang et al., 2004; Zhao et al., 2020). In this study, the Chinese cabbage yellow cotyledon mutant 19YC-2 is a distinctive mutant that displays obvious yellow phenotype in the cotyledons. Like the other reported yellow leaf mutants, the Chinese cabbage yellow cotyledon mutant has special excellent characters and can be used as a selective biomarker for hybrid breeding. More importantly, yellow cotyledon phenotype can be directly selected at the earlier cotyledon stage for hybrid purity identification, which shortens the breeding time and improves the breeding efficiency.

Most yellow leaf phenotype of mutant is mainly determined by the changed contents in chlorophyll and carotenoid and the disrupted chloroplast development (Sakamoto, 2003; Cheng et al., 2022). Previous study reported the reduced chlorophyll content in the yellow leaf mutant of ginkgo (Ginkgo biloba) (Sun et al., 2022). In the green leaf mutant of Populus deltoides Marsh, the leaf color was associated with the lower levels of chlorophyll and carotenoid (Zhang et al., 2019). The yellow-green leaf color in wheat (Triticum aestivum L.) was consequence of the decreased chlorophyll content and abnormal chloroplast development (Wu et al., 2018). In addition, chlorophyll deficient was also showed to cause yellow leaf phenotypes in watermelon (Xu et al., 2023), and the chlorophyll deficient mutant in Pak-choi exhibited the obvious yellow leaves and the reduced total chlorophyll content (Zhang et al., 2017). As expected, our studies also showed the changes in chlorophyll and carotenoid contents in Chinese cabbage yellow cotyledon mutant 19YC-2. The significantly higher carotenoid content and lower chlorophyll content were found in the cotyledons of 19YC-2, which may be responsible for the yellow cotyledon phenotypes and was similar to the previous reports (Zhang et al., 2015; Sun et al., 2022; Xu et al., 2023). It is known that chloroplast development also affects the accumulation of chlorophyll and carotenoid and leaf color changes (Pogson and Albrecht, 2011; Zhao et al., 2020). The thylakoid is an important part of chloroplast structures and essential for chloroplast function (Allen and Forsberg, 2001; Pogson and Albrecht, 2011). Many studies in leaf color mutants demonstrated that the variations in thylakoid structure caused the decreases in grana numbers and thylakoid lamella, resulting in the damaged chloroplast ultrastructure (Nie et al., 2021; Wang et al., 2022; Zhang et al., 2022b). In the current study, observation of leaf color found that the 1^st leaves of 19YC-2 gradually turned green with the growth of seedlings. Observation of chloroplast ultrastructure showed that the cotyledons and the yellow 1^st leaves of 19YC-2 exhibited the abnormal chloroplast and thylakoid structures with the absence of grana lamellae and starch grains, while the normal chloroplast in shape and structure were observed in the turn-green 1^st leaves of 19YC-2 at the four-leaf stage. The abnormal chloroplast ultrastructure may contribute to the yellow phenotypes in cotyledons and leaves of 19YC-2 mutant, which was consistent with many other reports (Nie et al., 2021; Wang et al., 2022; Zhang et al., 2022b). The results suggest that the changes in the contents of photosynthetic pigments in yellow cotyledon mutant 19YC-2 may be attributed to the abnormal chloroplast development, which can impair photosynthetic efficiency (Pogson and Albrecht, 2011).

Numerous environmental factors, such as light, have been described to affect the development and biogenesis of chloroplast via photomorphogenic pathways (Xu et al., 2021; Chen et al., 2023b). Long-term deficient light can lead to degradation of chlorophyll and leaf yellowing (Lee et al., 2014). Different light intensities can affect the chloroplast formation and the synthesis and degradation of photosynthetic pigments (Kong et al., 2016). The mutants of lacking photosynthetic electron transport in Arabidopsis exhibited a variegated phenotype under high light intensity and a green phenotype under low light intensity (Rosso et al., 2009). The yellow leaf mutant in pepper showed the increase in carotenoid content under high light intensity (Liu et al., 2023). Our previous study also found that low-intensity light treatment caused the turn-green leaves in Chinese cabbage yellow cotyledon mutant 19YC-2. Therefore, to investigate the underlying physiological mechanism of leaf color change in 19YC-2 under low-intensity light treatment is significant in this study. We found that the leaves of 19YC-2 showed higher carotenoid and lower chlorophyll contents under normal light condition than that under low-intensity light condition. The observation was coincident with that the leaves of 19YC-2 exhibited yellow exposed to normal light but turned green under low-intensity light. These results indicated that low-intensity light treatment primarily changed the carotenoid and chlorophyll contents and then induced the transition from yellow to green in the new leaves of 19YC-2. Meanwhile, RNA-seq analysis and KEGG pathway enrichment revealed that many DEGs were identified in the significantly enriched pathway of carotenoid biosynthesis in CK0 vs T0 (the cotyledon stage) and CK2 vs T2 (normal light treatment), but not in CK1 vs T1 (low-intensity light treatment), which indicated that the carotenoid biosynthesis and metabolism may be closely related to leaf color changes in 19YC-2. However, the pathway related to porphyrin and chlorophyll metabolism was enriched only in CK0 vs T0, but not in CK1 vs T1 and CK2 vs T2, implying that different light intensity treatments may have no significant effects on the DEGs related to chlorophyll biosynthesis in yellow cotyledon mutant 19YC-2. In addition, the existence of photosynthetic antenna proteins can increase the rate of energy delivery and photosynthetic efficiency (Pogson and Albrecht, 2011; Ruban, 2015). In the current study, the expression levels of DEGs encoding photosynthesis antenna proteins, such as BrCAB1, BrCAB13, BrCAB36, BrLHCB1.3, and BrLHCB4.2, were significantly decreased in CK0 vs T0 and CK2 vs T2, suggesting the suppressed photosynthesis in the yellow cotyledons and yellow leaves in 19YC-2.

Carotenoid biosynthesis is documented to be influenced by the changes in light and is stalled in darkness (Yuan et al., 2015; Xu et al., 2021). Carotenoids are synthesized in plastids, which is regulated by the complex and highly conserved biosynthetic pathway that has been well established and involves the transcription of many biosynthetic genes (Nisar et al., 2015; Yuan et al., 2015). Among these key genes that encode carotenogenic enzymes, PDS and ζ-carotene desaturase (ZDS) genes are responsible for the conversion of uncolored phytoene to red-colored lycopene and have been identified and studied in many colored fruits and crops (Fanciullino et al., 2007; Yuan et al., 2015). The high expression of PDS gene had been found to be correlated with the accumulation of carotenoid during tomato fruit ripening and citrus fruit development (Pecker et al., 1992; Kato et al., 2004; Rodrigo et al., 2004). The subsequent two biochemical pathways, lycopene converted to lutein and zeaxanthin, were mediated by LCYE and LCYB genes (Nisar et al., 2015). These carotenoid pigments are abundant not only in the yellow fruits and flowers, but also in leafy vegetables exposed to light (Yuan et al., 2015). Studies showed that LCYE mutation caused lutein deficiency and higher β-carotene accumulation (Pogson et al., 1996), and the high expression of LCYE gene promoted lutein accumulation in tomato and strawberry (Zhu et al., 2015; Yuan et al., 2022). Exploring the expression changes of these carotenoid biosynthesis related genes and their potential function can contribute to further understand the mechanism of leaf color changes in Chinese cabbage yellow cotyledon mutant 19YC-2. In this study, RNA-seq analysis showed that BrPDS (BraA08g023090.3C) and BrLCYE (BraA01g034180.3C) genes were upregulated expression in the yellow leaves of yellow cotyledon mutant 19YC-2, indicating the putative regulatory roles of BrPDS and BrLCYE in leaf color changes in 19YC-2.

Furthermore, the downstream genes in the pathway of carotenoid metabolism, such as BrAAO3 (BraA04g019480.3C), BrCCD4 (BraA01g010560.3C), and BrNCED3 (BraA05g032610.3C), were found to be downregulated expression in the cotyledons and leaves of 19YC-2 when displayed yellow. In the downstream process of carotenoid metabolism, a list of enzymes can cleave carotenoid to produce phytohormone abscisic acid (ABA) and strigolactones, and the downregulated expression of these downstream genes had been extensively demonstrated to be involved in the catabolism of carotenoid (Nisar et al., 2015; Yuan et al., 2015). The expression of NCED gene in ginkgo was increased in yellow leaves in autumn, while decreased in green leaves in summer (Sun et al., 2022). Previous studies in zucchini reported that NCED and CCD genes were downregulated in yellow and orange peels, which was agreement with the accumulation of yellow lutein (Obrero et al., 2013; Niu et al., 2023). These evidences suggest that the differential expression of carotenoid-related genes combining the alteration of carotenoid content favor the leaf color changes in yellow cotyledon mutant 19YC-2.

Many putative regulators have been identified to function in transcriptional regulation of carotenoid biosynthetic genes in photosynthetic tissues, which is highly influenced by various environmental factors (Stanley and Yuan, 2019). In the current study, many differentially expressed TFs were identified to be highly correlated with the changes in carotenoid contents by WGCNA, and belonged to the families of BrERF/AP2, BrR2R3-MYB, BrbHLH, and BrWRKY. The BrERF/AP2 family contains many TFs that had been characterized to participate in regulating the process of carotenoid metabolism (Yuan et al., 2015; Stanley and Yuan, 2019). In Arabidopsis leaves, RAP2.2 (a member of ERF/AP2 family) was reported to interact with SINAT2 and positively regulate PSY and PDS during carotenoid biosynthesis (Welsch et al., 2007). In tomato, an AP2a gene was found to be a major regulator of the upstream genes of carotenoid biosynthesis, mediating fruit ripening (Chung et al., 2010; Karlova et al., 2011). We also found that two DEGs, BraA08g001340.3C and BraA04g017150.3C, were identified to BrERF/AP2 family and highly positively related to the changes in carotenoid contents, which provide important references for exploring the functions of the two TFs. Meanwhile, the TFs of MYB family had been demonstrated to involve in the transcriptional regulation of leaf-color mutation in ginkgo (Wu et al., 2019). In kiwifruit, MYB7 gene was found to activate the promoter of LCYB gene that modulates carotenoid accumulation (Ampomah-Dwamena et al., 2019). In the current study, a BrMYB gene (BraA06g042530.3C) was upregulated expressed in the cotyledons and leaves of yellow cotyledon mutant 19YC-2 and highly positively related to carotenoid accumulation. Overall, these differentially expressed TF genes may play potential important roles in regulating carotenoid biosynthesis mediating leaf color change in yellow cotyledon mutant 19YC-2.