Yanling Yinbiancha, with characteristic white edge on its leaves and a green central main vein, exhibits weak growth, robust cold tolerance, but it shows low tolerance to high temperatures and drought, which makes it susceptible to burning during summer. Through numerous rounds of asexual propagation and experimental cultivation since 2003, its performance traits have been established as stable. Rare instances of branch emergence and mutation resulted in the appearance of the green leaves, with the green leaves being larger than their albino counterparts and displaying a heightened growth potential in subsequent years. Yanling Yinbiancha, planted under natural conditions, leaves with albino edges and entirely green leaves were collected (one bud and two leaves were harvested on the morning of 24 April 2022) for further physiological, transcriptomic and metabolic characterization.

The tea cultivar Yanling Yinbiancha (C. sinensis L. O. Kuntze), exhibiting a characteristic albino leaf phenotype, was utilized in this study. Plants were grown under natural field conditions at the Tea Research Institute, Hunan Academy of Agricultural Sciences, Changsha, China. For sample collection, leaves were harvested at the one bud and two leaf stage, corresponding to the spring flush period. Specifically, the first and second fully expanded leaves along with the apical bud were collected from branches. Leaves displaying albino edges as well as entirely green leaves were sampled from separate branches of the same cultivar exhibiting these distinct phenotypes.

Sampling was conducted in the morning hours on 24 April 2022 to obtain intact RNA and metabolites. The freshly harvested leaf samples were immediately frozen in liquid nitrogen after detachment and subsequently stored at −80°C until further analysis. The sampled green and albino leaves were utilized for detailed physiological, transcriptomic and metabolomic characterization to elucidate the factors influencing the albino phenotype in this tea cultivar.

The leaf samples from two genotypes were collected and weighed. 0.2 g of leaves (first and second leaf from a single node) were cut into thin strips and centrifuged using 25 mL of 95% ethanol in a 50 mL centrifuge tube. The samples were put in a dark place at room temperature for 16 h. When all the leaves turned white, absorbance was measured at 665 nm, 649 nm, and 470 nm with a UV-2102C spectrophotometer (Unico, China). Each sample has three biological replicates, and each has three technical replicates.

The calculation formulas are as follows: C h l a   ( m g g ) = 13.95 A 665 − 6.88 A 649 × V 1000 W C h l b   ( m g g ) = 24.96 A 665 − 7.32 A 649 × V 1000 W C a r t o   ( m g g ) = 1000 A 470 − 2.05 C h l a − 114.8 C h l b / 245 × V 1000 W

Metabolic profiling was outsourced to Wuhan Metware Biotechnology Co., Ltd. (https://www.metware.cn), a well-established and renowned service provider specializing in metabolomics analysis. The meticulous execution of the metabolic profiling involved a step-by-step approach as described below:

Firstly, cryo-preserved leaf samples were carefully collected and weighed with precision, ensuring consistency in the amount of material used for analysis. These samples were specifically chosen to represent both the green and albino leaves of the tea plants, forming the basis for our comparative study. Next, metabolites were extracted from the leaf samples using 70% methanol as the solvent. Methanol extraction is a highly efficient method that allows us to retrieve a wide range of metabolites from the plant tissues. Subsequently, the methanol extracts underwent state-of-the-art liquid chromatography mass-spectrometry/M.S. analysis (LC-MS/MS) using an advanced UPLC Shim-pack UFLC SHIMADZU CBM30A system, coupled with an Applied Biosystems 6500 QTRAP mass spectrometer. LC-MS/MS is a powerful technique known for its high-throughput capabilities, sensitivity, and accuracy in detecting a diverse array of metabolites.

To identify the detected metabolites, a comprehensive approach was adopted, utilizing Metware’s extensive metabolite database in conjunction with publicly available metabolite databases. This comprehensive strategy ensured the reliable identification of metabolites based on their mass spectra and retention times. The quantification of the identified metabolites was performed based on the LC-MS/MS data. This crucial step allowed us to precisely measure the abundance of each metabolite in the samples, facilitating quantitative comparisons between the green and albino leaves. To assess the differences in metabolite accumulation patterns between the two leaf types, we employed orthogonal partial least squares discriminant analysis (OPLS-DA), a powerful multivariate statistical method. OPLS-DA aided in identifying significant changes in metabolite levels, providing valuable insights into the variations between the green and albino leaves under study.

Finally, the identification of Differentially Accumulated Metabolites (DAMs) was based on defined criteria, with metabolites displaying |Log₂ Foldchange| ≥ 1 and VIP ≥1 considered as key indicators of differences in abundance between the green and albino leaves. These DAMs hold great significance in unraveling the metabolic changes associated with albinism in tea plants, shedding light on potential mechanisms underlying this intriguing phenotype.

Total RNA was extracted from the samples with the RNA Extraction kit (TIANGEN, Beijing, China). The RNA quality and concentration were assessed with agarose gel electrophoresis and NanoDrop2000 spectrophotometer. Quality testing, library construction, and sequencing for each sample were done at Metware Biotechnology Co., Ltd. (https://www.metware.cn), following the company’s standard procedures. Low-quality data were removed for downstream analysis, and high-quality clean reads were used for transcriptome quantification. The clean reads were localized to the reference genome (Xia et al., 2020) using Hisat2 (Sirén et al., 2014). Reads per kilobase mapping (FPKM) for all genes were used to determine gene expression values, and further screening for differentially expressed genes (DEGs) was performed. The identified DEGs were further enriched by KEGG analysis. Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were applied using Tbtools software (Chen et al., 2020). Heat maps were generated using the OmicStudio tools at https://www.omicstudio.cn/tool.

Yanling Yinbiancha, a cultivar of C. sinensis (L.) O. Ktze, was used in this study. The Yanling Yinbiancha plant exhibits a characteristic white edge on its leaves, with a green central main vein. To further characterize this phenomenon, we estimated physiological parameters, including chlorophyll a, chlorophyll b, and carotenoids from leaf samples in Yanling Yinbiancha with albino and green leaves. The mean performance of each trait was estimated, and parameters depicted significant differences between albino and green genotypes (Figures 1A, B).

In comparing green leaves to albino leaves, there was a significant increase observed in the levels of total chlorophyll, chlorophyll a, chlorophyll b, and carotenoids in the green leaves. Specifically, the green leaves exhibited a 55% increase in total chlorophyll, a 57% increase in chlorophyll a, a 50% increase in chlorophyll b, and a 64% increase in carotenoid content compared to the albino leaves. Interestingly, the ratio of chlorophyll a to chlorophyll b (Chla/Chlb) was found to be higher in albino leaves when compared to green leaves. This suggests that albino leaves have a relatively higher proportion of chlorophyll a in relation to chlorophyll b compared to green leaves. These findings indicate that the presence of green pigmentation in leaves is associated with higher levels of total chlorophyll, chlorophyll a, chlorophyll b, and carotenoids, highlighting their role in photosynthetic processes.

For metabolome profiling, leaf samples from Yanling Yinbiancha and its mutant branch with green leaves were collected and subjected to subsequent analysis. Metabolome profiling resulted in the identification of 1,046 metabolites belonging to subclasses alkaloids (61), amino acids and derivatives (83), flavonoids (307), lignans and coumarins (38), lipids (89), nucleotides and derivatives (63), organic acids (67), phenolic acids (209), tannins and terpenoids (30), and others (99) (Figure 2B; Supplementary Table S1). The accumulation pattern of metabolites showed significant differences in the metabolic profile of albino and green leaves. We further identified differentially accumulated metabolites Albino vs. Green (A vs. G) comparison (127 DAMs) (Figure 2B; Supplementary Table S1). Among 127 DAMs, 101 up-accumulated and 26 down-accumulated metabolites were identified (Figure 2D). Amino acids, alkaloids, flavonoids, and phenolics were among the metabolites accumulated differentially (Figure 2E).

To gain insights into the molecular mechanisms underlying the observed differences in pigmentation between green and albino leaves, transcriptome profiling was conducted on the collected leaf samples. This technique allowed for the characterization of the expression profiles of genes that showed differential expression between the two leaf types. We constructed six libraries and FPKM-based expression profiles of 51,202 genes were quantified. A total of 42.01 Gb of data was generated. A total of 308,901,608 raw reads were obtained, and after filtering, 286,071,524 clean reads were kept (Supplementary Table S2). The quality check was performed to confer the reliability and reproducibility of the data. Q20 and Q30 were estimated to be over 96% and 91%, respectively (Supplementary Table S2). Moreover, GC contents ranged from 44.03% to 44.54%. All the transcriptome data were subjected to principal component analysis to identify the corresponding variation. PC1 and PC2 covered 35.39% and 19.49% of the variation, and replicates from each group were clustered, depicting the significance of the transcriptome data sets for downstream analysis (Supplementary Figure S1).

The expression profile of genes in both albino and green variants were compared, and 1,412 differentially expressed genes (DEGs) were identified (Figure 2A). Among them, 814 genes were upregulated in the green variant compared to the albino, and 598 were downregulated (Figure 2C). Go enrichment analysis of DEGs suggested significant enrichment of GO terms associated with flavonoid biosynthesis, photosynthesis, light harvesting in photosystem 1, and response to oxygen levels (Figure 3).

To comprehend the molecular basis of albinism in Yanling Yinbiancha, we compared metabolomic and transcriptomic profiles of both albino and green variants. As it is evident from previous studies that albino leaves lack chlorophyll contents considerably. Physiological data also suggested the same where we quantified total chlorophyll in both variants, and albino leaves were identified with less chlorophyll contents. Moreover, a decrease in chlorophyll contents can simultaneously result in reduced carotenoids in leaves. Therefore, we screened the DEGs and identified the DEGs associated with chlorophyll biosynthesis, carotenoid biosynthesis, and stay green trait based on the annotation information (Supplementary Table S3; Figure 4A). The genes related to chlorophyll biosynthesis, including CLH2, CB4A, CB24, SGR, PSBC, ARR1, AB3C, HEM11, SUOX (TEA000669), CRD1, and CIP7, were identified with upregulated expression pattern in albino leaves compared to green leaves. While LHCA3, PORA, CB23, CB11, CHLH, and SUOX (TEA000535) were identified with downregulated expression patterns in albino leaves compared to green leaves. It is worth noting that SGR (protein STAY-GREEN) was significantly suppressed in green leaves, indicating that chlorophyll biosynthesis was positively regulated in green leaves. Moreover, two genes, NHL6 and AOX4 (associated with carotenoid biosynthesis showed upregulated expression patterns in albino leaves, and one gene (CCD1) was identified with a downregulated expression pattern in albino leaves. The annotation information and expression profile of these suggested candidacy for further functional verification of these genes and their role in conferring albinism in C. sinensis.

Moreover, we identified 105 transcription factors as DEGs. The expression profile of TFs (81) suggested a significant increase in expression pattern in albino leaves compared to green leaves, while 23 were downregulated in albino leaves (Figure 4B). The identified transcription factors belong to AP2, AUX, B3, bHLH, bZIP, C2C2, GARP, CSD, GRAS, GRF, MYB, NAC, and WRKY families.

Previously published reports suggested an indirect relationship between albinism and flavonoid accumulation patterns. Lack of chlorophyll in albino leaves makes them vulnerable to light stress resulting in increased flavonoid accumulation in leaves. Therefore, we further explored the transcriptome and metabolome datasets to identify differentially accumulated flavonoids and DEGs associated with flavonoid biosynthesis. The most prominent DEGs associated with flavonoids biosynthesis included ANR, ARR1, DMR6, VINSY, and VSR1, showing upregulated expression patterns in albino leaves, while JOX1, PKSB, MY123, MYB1, GSTFC, LAC15_1, UFOG3, SAT, CYQ32_1, HST and SAT_1 were upregulated in green leaves (Figure 5A). The differential expression profile of these genes suggested regulation of flavonoids in albino and green leaves, which was further confirmed by flavonoid accumulation patterns in both types (Figure 4B). Flavonoids including, Epicatechin-epiafzelechin, Pinocembrin-7-O-(2″-O-arabinosyl)glucoside, Chrysoeriol-6-C-glucoside-5-O-glucoside, Myricetin-3,7,3′-trimethyl ether, Epitheaflavic acid-3-O-Gallate, Apigenin-7-O-(2″-Sinapoyl) glucuronide, O-MethylNaringenin-8-C-arabinoside, Isorhamnetin-3-O-arabinoside, Phloretin-2′-O-(6″-O-xylosyl) glucoside, and 8,8′-Methylenebiscatechin were up-accumulated in albino leaves while Phloretin, 3,4,2′,4′,6′-Pentahydroxychalcone-4′-O-glucoside, Phloretin-2′-O-glucoside (Phlorizin), Dihydrocharcone-4′-O-glucoside, Myricetin-3-O-(6″-malony)glucoside, Sexangularetin, Apigenin-7-O-glucuronide, and Quercetin-3′,4′-dimethyl ether were up-accumulated in green leaves.

A conjoint analysis of the transcriptome and metabolome was performed to investigate the relationship between genes and metabolites. KEGG pathway enrichment analysis was conducted to identify co-enriched pathways of differentially expressed genes (DEGs) and differentially accumulated metabolites (DAMs). The analysis revealed that DEGs and DAMs were enriched in 53 KEGG pathways, including biosynthesis of secondary metabolites, flavonoid biosynthesis, and biosynthesis of amino acids (Supplementary Table S5; Figure 6C). The Pearson correlation coefficient (PCC) was calculated to evaluate the correlation between DEGs and DAMs, and a nine-quadrant diagram was generated to show the fold changes of DEGs and DAMs with a PCC >0.8. The I and III quadrants showed a positive correlation between DEGs and DAMs, while the VII and IX quadrants showed a negative correlation (Figure 6A). A clustered heatmap was constructed to visualize the DAMs related to DEGs, which were classified into several categories, with alkaloids, amino acids, and flavonoids being the largest categories (Figure 6B).

We further explored the correlation and developed a network of all the co-regulated DEGs and DAMs (Figure 6D). Furthermore, each KEGG pathway and its associated DEGs and DAMs were explored to understand the regulatory relationship of DEGs and DAMs. We focused on flavonoid biosynthesis and photosynthesis pathways (Figures 6E, F). In the flavonoid biosynthesis pathway, our study identified three coregulated metabolites, namely, pme1201 (Phloretin), Lmsn003297 (3,4,2′,4′,6′-Pentahydroxychalcone-4′-O-glucoside), and mws2118 (Phloretin-2′-O-glucoside (Phlorizin)), which were influenced by 12 differentially expressed genes (DEGs) including FL3H, PKSB, FAP3, VSR1, JOX1, ANRCH, SAT, BEATH, VINSY, SAT, and ANR with one unknown gene (Supplementary Table S5). Among these DEGs, FAP3, ANR, FL3H, ANRCH and VSR1 displayed a significant positive correlation with the three metabolites, while PKSB, SAT showed a significant negative correlation with their accumulation pattern (Figure 6E). In contrast, we only identified one metabolite, MWSmce668 (ATP; Adenosine 5′-Triphosphate), positively associated with TEA023668 (PSBY; Photosystem II core complex proteins) in the photosynthesis pathway (Figure 6F). Nonetheless, we also identified YCX91, ATPB, FER1, PSAEB, PSAK, PSAL, PSAN, PSBC, PSBW, and PSB28 with one unknown gen as DEGs in the photosynthesis pathways (Supplementary Table S5).

To understand the regulation patterns of candidate genes in two contrasting genotypes (G and A), we conducted qRT-PCR analysis for all the candidate genes, using the primer sequences provided in Supplementary Table S6. The expression profile of the genotype was considerably different among two genotypes (Figure 6A). Among 18 candidate genes, 11 depicted positive expression patterns in albino leaves (A), while seven showed negative expression patterns (Figure 7A). Furthermore, the qRT-PCR results also support the observed variations in the whole transcriptomic datasets. Notably, a significant positive correlation (r² = 0.87) was observed between RNA-seq and qRT-PCR data (Figure 7B). This finding provides further support for the reliability and accuracy of the qRT-PCR method for assessing gene expression levels in different genotypes.