CsMYB1 integrates the regulation of trichome development and catechins biosynthesis in tea plant domestication.

Tea trichomes synthesize numerous specialized metabolites to protect plants from environmental stresses and contribute to tea flavours, but little is known about the regulation of trichome development. Here, we showed that CsMYB1 is involved in the regulation of trichome formation and galloylated cis-catechins biosynthesis in tea plants. The variations in CsMYB1 expression levels are closely correlated with trichome indexes and galloylated cis-catechins contents in tea plant populations. Genome resequencing showed that CsMYB1 may be selected in modern tea cultivars, since a 192-bp insertion in CsMYB1 promoter was found exclusively in modern tea cultivars but not in the glabrous wild tea Camellia taliensis. Several enhancers in the 192-bp insertion increased CsMYB1 transcription in modern tea cultivars that coincided with their higher galloylated cis-catechins contents and trichome indexes. Biochemical analyses and transgenic data showed that CsMYB1 interacted with CsGL3 and CsWD40 and formed a MYB-bHLH-WD40 (MBW) transcriptional complex to activate the trichome regulator genes CsGL2 and CsCPC, and the galloylated cis-catechins biosynthesis genes anthocyanidin reductase and serine carboxypeptidase-like 1A. CsMYB1 integratively regulated trichome formation and galloylated cis-catechins biosynthesis. Results suggest that CsMYB1, trichome and galloylated cis-catechins are coincidently selected during tea domestication by harsh environments for improved adaption and by breeders for better tea flavours.

Introduction

Tea is the second most popular nonalcoholic beverages consumed after water, due to its rich flavours and numerous health benefits (Yang & Hong, 2013). The buds and young leaves used as tea processing materials are usually covered by a high density of trichomes, also called ‘Cha Hao’ (Li et al., 2020). The presence of long and densely spaced trichomes on the apical buds and young leaves of tea plan is an important trait associated with enhanced plant resistance against many biotic and abiotic stresses, and also regarded as one of the most important visible evaluation criteria for superior tea quality. Two thousand years of Chinese tea utilization and culture tells that tea trichome critically contributes to tea flavours and health functions; indeed, many nutritional and metabolomic studies have confirmed the essential contributions of trichomes to tea qualities (Schilmiller et al., 2008; Liao et al., 2019; Li et al., 2020; Sun et al., ).

Similar to these in Arabidopsis and cotton, tea trichomes are unicellular, unbranched and nonglandular type of surface hairs (Pesch & Hulskamp, 2009). Trichomes play an array of protective roles in plants against herbivore or other predators, microbial pathogens and abiotic stresses (Balkunde et al., 2010; Bleeker et al., 2012; Gonzales‐Vigil et al., 2012; Luu et al., 2017). Tea plant trichomes are also regarded as the first physical barrier for plant adaption to adversary environments, protecting tea plants from access by herbivores and pathogens, reflecting UVB and high lights to avoid damages, reducing transpiration and preventing water loss in leaf under high temperature, or protecting leaf from freezing (Schilmiller et al., 2008; Ning et al., 2016; Li et al., 2020; Yu et al., 2021). A previous study has shown that tea trichomes also synthesized and accumulated a large array of specialized metabolites, including catechins, theanine, caffeine, flavonols, saponins and diverse terpenoid‐ and lipid‐derived volatiles, that show toxic or repellent activities to herbivores and inhibitory effects on microbial pathogens to build up chemical defences (Li et al., 2020). The metabolites identified in tea trichomes are similar to those present in leaves, but many of them are quantitatively enriched in trichomes by several folds higher than in leaves (Li et al., 2020). Despite the importance of trichomes to tea plants in both defences and tea quality, little is known about the molecular bases for trichome development and numerous metabolite synthesis inside tea trichomes, as well as their coordinative regulation (Liu et al., 2018; Li et al., 2020; Yu et al., 2021). While most modern tea cultivars of both assamica and sinensis types are covered with long and densely spaced trichomes, many wild tea relative plants in section Thea, such as Camellia taliensis, C. crassicolumna and C. tachangensis, are covered with much less trichomes or even glabrous (eFloras, 2018; http://www.efloras.org/florataxon.aspx; Liu et al., 2020). Recent chemical studies also demonstrated that galloylated cis‐catechins contents are significantly lower in these wild tea relatives in section Thea than in current tea cultivars, and most of these wild tea relatives meanwhile contained significantly higher levels of polygalloylated glucoses (e.g. 1,2,4,6‐tetra‐O‐galloyl‐β‐d‐glucose and acylated derivatives, but epigallocatechin (EGC) or other nongalloylated catechins at lower levels or galloylated trans‐catechins gallocatechin gallate (GCG)) (Yagi et al., 2009; Jin et al., 2018; Meng et al., 2019; Zhu et al., 2019; Xia et al., 2020). The genetic casuals for these interesting differences in types and contents of catechins and trichome indexes are also unclear.

The unicellular and nonglandular trichomes in Arabidopsis leaf have been well documented, in terms of their initiation and formation from epidermal cell differentiation under positive and negative regulations by a series of transcription factors (TFs) (Pesch & Hulskamp, 2009). An MYB‐basic helix‐loop‐helix (bHLH)‐WD40 protein (MBW) transcriptional complex, formed by the positive R2R3‐type MYB TFs GLABRA1 (GL1) and MYB23, the bHLH TFs GLABRA3 (GL3) and ENHANCER OF GLABRA3 (EGL3), and a WD40‐repeat protein TRANSPARENT TESTAGLABRA1 (TTG1), is positioned in the regulatory centre (Bernhardt et al., 2003; Zhao et al., 2008; Chopra et al., 2019; Zhang et al., 2019). This MBW complex activates the downstream HD‐Zip TF gene GLABRA2 (GL2) and other effectors to initiate trichome formation (Johnson et al., 2002; Zhao et al., 2008; Balkunde et al., 2010). This MBW complex also activates repressor TFs R3 MYB‐domain TF genes, such as CAPRICE (CPC) and ENHANCER OF TRYAND CPC1 (ETC1). The R3 MYBs can fine‐tune the MBW effects, by competitive binding to activator MBW complex (Zhao et al., 2008). Thus, these positive activators and negative repressors together participate in protein–protein interactions to form MBW complexes in dynamic manners to amplify and feedback fine‐tuning regulatory effects. While several selective epidermises in a field of equivalent cells are precisely positioned for either trichome or hair‐less cell fates by collaborative regulation of GL1‐MBW complexes (Pesch & Hulskamp, 2009), GL1‐ and GL1‐like TFs from different plants also regulate metabolic genes involved in the biosynthesis of different types of primary and secondary metabolites (Pesch & Hulskamp, 2009; Matías‐Hernández et al., 2017). The pleiotropic effects of the trichome initiation‐ and development‐regulating genes on metabolic genes may imply the requirement for a collaborative regulation of epidermal cell development and metabolism for biosynthesis of cell contents (Lashbrooke et al., 2015; Zhao et al., 2020). As breeding of the new tea plant varieties with long and densely spaced trichomes has been one of the goals for genetic improvement, a profound understanding of the molecular bases for tea plant trichome development and metabolite biosynthesis is highly expected.

Here, we integrated metabolic and transcriptomic profiling, trichome phenotyping of tea plant varieties and various molecular and genetic approaches to identify and characterize key regulator genes of trichome formation and galloylated cis‐catechins synthesis inside the trichomes. An R2R3 MYB TF CsMYB1 was mapped and characterized as a critical trichome regulator, by forming a ternary complex with CsGL3 and CsWD40. In addition, CsMYB1 also functioned as an activator of galloylated cis‐catechins biosynthesis genes. Interestingly, both trichome and galloylated cis‐catechins are domestication traits in tea plant cultivars, but scarcely present in wild tea relatives. These domestication traits may result from harsh environment selection for improved adaption to biotic and abiotic stress and breeders’ selection for better tea flavours. A 192‐bp insertion in CsMYB1 promoter was found only in modern tea cultivars, but not in a wild tea relative, indicating that CsMYB1 is evolved as domestication gene underlying the formation of these traits.

Materials and Methods

Plant materials and growth conditions

Camellia sinensis L. cv Fudingdabai, a trichome‐enriched variety, and other tea plant varieties are grown at Guohe tea plantation (Hefei, Anhui Province, China), Dongzhi tea plant resource garden (Dongzhi, Anhui) and Jinhua National Tea Resource Center (Jinhua, Zhejing) and were used for gene cloning, phenotype survey and transcriptome and metabolomic analyses. The gl1‐2 (CS3126; Col‐0 background), ttg1‐1 (CS3099) and gl3 egl3 mutants (CS6516; Landsberg erecta) (Dai et al., 2016) were used for transformation. For anthocyanin accumulation test, 1‐wk‐old Arabidopsis seedlings were transferred to the one half‐strength MS medium supplemented with 3% sucrose for 1–3 d under normal conditions. All the Arabidopsis plants were grown at 22°C with 16 h : 8 h photoperiod at c. 120 μmol m−2 s−1.

Targeted metabolite profiling

Anthocyanins were determined using a spectrophotometer as described previously (Zhao et al., 2011). Tea catechins were profiled with high‐performance liquid chromatography (HPLC) and LC‐MS according to the method described previously (Li et al., 2020). Briefly, c. 0.05 g of tissues was powdered and then mixed with 1 ml 80% methanol extraction solution. The tissues were sonicated at room temperature for 1 h and then shaken overnight at 100 rpm. Following centrifugation at 13 680  for 10 min, supernatants were filtrated using Nylon membrane (0.45 μm). Catechins were measured by HPLC and ESI‐qTOF‐MS according to the method described previously (Li et al., 2020). Target catechins components were identified by comparing the retention time and molecular weight to authentic compounds. Standard curves were calculated using HPLC and puerarin to quantify the compounds, and a flavone C‐glycoside was used as internal standard in LC‐MS analyses.

Cloning and construction of binary vectors for plant expressing

The open reading frames (ORFs) of CsMYB1, CsWD40, CsGL3s and CsCPC were amplified from leaf cDNAs prepared from C. sinensis cv Fudingdabai using gene‐specific primers (Supporting Information Table S1), cloned into pDONR221 by BP recombinase (Invitrogen, Life Technologies) and then subcloned into pB2GW7 by LR recombinase (Invitrogen, Life Technologies, Waltham, MA, USA) for Arabidopsis transformation. The ORF of CsMYB1 was amplified and subcloned into pK7WFG2 vector fusion with green fluorescent protein (GFP) protein and driven by 35S promoter for tea plant hairy root transformation.

Transformation of Arabidopsis and hairy roots of tea plants

Agrobacterium tumefaciens GV3101 strain harbouring pB2GW7‐CsMYB1, pB2GW7‐CsWD40, pB2GW7‐CsGL3a and pB2GW7‐CsCPC were used for the genetic complementation of Arabidopsis gl1, ttg1‐1, gl3 egl3 mutants and wild‐type. Transformants were selected on half‐strength Murashige & Skoog (MS) medium containing 7 μg l−1 phosphinothricin and were further verified with genomic PCR and quantitative RT‐PCR (qRT‐PCR). The T3 progeny plants were used for the analysis.

For tea plant hairy root transformation, the pK7WFG2‐CsMYB1 construct was transformed into Agrobacterium rhizogenes strain ATCC 15834 by electroporation. The selected positive transformants were used to transform 3‐month‐old tea seedlings, which were pretreated with acetosyringone, when the hypocotyl parts close to the roots were wounded, as described by Alagarsamy et al. (2018).

RNA extension and qRT‐PCR

Total RNA isolation and qRT‐PCR were conducted as previously described (Li et al., 2016). Briefly, 1 μg of total purified RNA from each sample was used for cDNA synthesis using the SuperScript III kit (Invitrogen). qRT‐PCR was performed by an iQ5 real‐time PCR machine with gene‐specific primers (Table S1). Data were collected and analysed by the Manager software (Bio‐Rad). The relative expression of each gene was calculated after being normalized to housekeeping genes CsACTIN and CsETF.

Yeast two‐hybrid assays

The ORFs of CsWD40 and CsGL3s were amplified and subcloned in frame with the GAL4 BD into the pGBKT7; CsWD40, CsMYB1 and CsCPC were subcloned in frame with the GAL4 AD into the pGADT7 vector as previously described (Li et al., 2016). The resulting constructs were co‐transformed into the yeast strain AH109 with empty vectors as a negative reference. The protein interactions were confirmed by the growth experiment on the SD medium without Trp, Leu, His and Ade (SD/‐T‐L‐H‐A).

Bimolecular fluorescence complementation

For bimolecular fluorescence complementation (BiFC) assays, ORFs of CsMYB1, CsGL3a and CsCPC were amplified and subcloned in frame with the YFPN into the pFGC‐YN173 vector, and ORFs of CsGL3a and CsWD40 were amplified and subcloned in frame with the YFPC into the pFGC‐YC155 vector. The resulting constructs were then transformed into A. tumefaciens strain GV3101 for transient assays. About 3‐wk‐old Nicotiana benthamiana leaves were used to perform infiltration as described previously by Li et al. (2016). After 48 h of infiltration, yellow fluorescent protein (YFP) fluorescence was imaged using a Leica DMi8 M laser scanning confocal microscopy system (Leica Microsystems, Wetzlar, Germany).

Electrophoretic mobility shift assay

The ORFs of CsMYB1 were subcloned into pGEX‐4‐T, and then, the construct was translated into the Escherichia coli BL21 cells. The positive BL21 strains were induced by 0.75 mM isopropylb‐d‐thiogalactopyranoside at 16°C for 20 h, and the GST‐CsMYB1 recombinant proteins were purified by immobilized glutathione beads (Sangon Biotech Co. Ltd, Shanghai, China) as described previously (Gao et al., 2021). The Electrophoretic mobility shift assay (EMSA) was performed in 6.6% nondenatured polyacrylamide gel using the EMSA kit (Beyotime, Shanghai, China) according to the method described by Gao et al. (2021).

Promoter transactivation assays

Approximately 2.0‐kb sequences upstream the translation start codons of the CsMYB1 genes from various modern tea cultivars (GenBank accession no. MW837263) and the wild tea relative C. taliensis (GenBank accession no. MW837264), and genes CsGL2 (GenBank accession no. MW837262), CsSCPL1A (GenBank accession no. MW837263) and CsANR (GenBank accession no. MW837261), were amplified from tea genomic DNA using the special primers (Table S1) and then subcloned into p2GWL7 to generate the luciferase reporter constructs. The effecter genes CsMYB1, CsWD40, CsGL3a/CsTT8 and CsCPC were cloned into p2GW7 by LR reaction to generate the effector constructs. Arabidopsis mesophyll protoplasts were isolated and transfected according to the method described previously (Yoo et al., 2007). Luciferase activities were examined using Dual‐Luciferase Reporter Assay System (Promega) according to the method described previously (Li et al., 2016). The Renilla luciferase gene was used as an internal control to normalize the reporter gene expression values.

Subcellular localization of the TFs

The ORFs of the genes CsMYB1, CsWD40, CsGL3a and CsCPC were amplified and cloned into pK7WGF2 by LR reaction. Three‐week‐old N. benthamiana leaves were infiltrated using A. tumefaciens EHA105 harbouring the resulting constructs as described previously (Li et al., 2016). Three or four days after infiltration, the GFP fluorescence was imaged with confocal microscopy.

Treatment on tea plants

For hormone treatment experiments, the detached branches with tender tea shoots were sprayed with either 3 mM abscisic acid (ABA) or 100 μM methyl jasmonate (MeJA) solution, while the controls were treated with distilled water (Shi et al., 2015). The bud and the first leaf were harvested at 0, 12 and 48 h after treatment for metabolism and transcriptome analysis. For insect biting experiments, tea geometrids (Ectropis oblique) at the third larval stage were starved for 6–8 h and placed onto young tea plant leaves. Tea geometrids were removed when one‐third of the leaves were consumed. The leaves were collected at 0 and 72 h after treatment for further analysis.

Antisense oligonucleotides assay

Gene suppression of CsMYB1 in tea plants using antisense oligonucleotides (asOND) was performed as described previously (Xie et al., 2014) with slight modifications. Briefly, asOND was designed and selected using the Soligo software (https://sfold.wadsworth.org/cgi‐bin/soligo.pl) (Table S1), and tender tea shoots with one leaf were cut off and partially inserted into Eppendorf Tubes containing 2 ml of 20 μM asOND solution and the solution with sense oligonucleotides (sense OND) of the gene as the control. After 72 h of treatment, the tissues were harvested for metabolism and transcriptome analysis.

Bioinformatic analysis

The GenBank accession numbers for the reported genes are CsMYB1 (MN604029), CsCPC (MN604030), CsWD40 (MN604035), CsGL3a (MN604034), CsGL3b (MN604033) and CsGL3c (MN604032). The first leaves from different tea varieties (n = 60) were isolated, and then, the abaxial surface of the leaves was scanned by a microscope. The trichome phenotype (1‐mm2 area) was manually investigated (≥ 30 leaves of each variety) using the ImageJ software. In the tea genome, a total of 1610 TFs were identified (Wei et al., 2018). Association mapping between the trichome phenotypes and the expression patterns of all TF genes and resulting heatmap was performed by the R package (Koutecky, 2015). The ‘corrplot’ in the R package was used for correlation analysis and drawing diagrams. Briefly, different tissues and developing leaves of tea plants were used for catechins content and transcriptome analysis, respectively. Then, the correlation analysis was performed using the expression levels of MYB genes with the contents of EGCG and ECG or the expression levels of CsANR and CsSCPL1A in different tissues or in developing leaves. Phylogenetic trees were constructed with the Mega7.0 software using the neighbour‐joining method with 1000 bootstrap replications, and multiple alignments were constructed using the ClustalW2 program.

Genetic variation analysis of the CsMYB1 in Camellia species

SNPs and indels of CsMYB1 in CDS region, as well as the up and down 4 kb regions, were extracted from the data generated from 81 cultivated and wild tea plants (Xia et al., 2020). The genetic divergence and nucleotide diversity of CsMYB1 were calculated using Vcftools (http://vcftools.sourceforge.net/) based on all the SNPs/indels identified.

GUS activity assays

The promoters of CsMYB1 amplified from cultivars (SCZ and ZH) and wild tea relatives were subcloned into pBI121 vector to generate GUS reporter constructs. The resulting constructs were transformed into A. tumefaciens strain GV3101 for transient expression assay in the 3‐wk‐old N. benthamiana leaves. About 3 d after infiltration, the GUS activity was assayed using the leaf staining method as described previously (Wang et al., 2021).

Statistical analysis

All data were obtained from at least three independent experiments. The confidence limit 95% or 99% was defined as the significant difference between two‐tailed data in Student's t‐test. For the yeast two‐hybrid (Y2H) assay, BIFC and plant phenotypic display, only representative pictures were shown.

Results

Association of trichome phenotype with a critical MYB regulator CsMYB1

To understand how tea plant trichomes are formed and vary morphologically in tea plant varieties and the underlying genetic factor regulating trichome phenotype, we surveyed the trichome phenotypes on more than 100 tea plant varieties, including these well‐known core germplasms with much longer and higher‐density trichomes, as well as these varieties with shorter and low‐density trichomes.

Both trichome length and density are quantitative traits, with continuous distributions in examined tea plant varieties (Fig. S1). Considering the varying length and density of trichomes on the young leaves of different tea plant varieties, a trichome index (Density × Length) was introduced to more accurately evaluate and quantify the trichome phenotypes in tea plants. We measured the large variations in the length and density of trichomes on the apical bud, the first leaf and the second leaf of these tea plants and found some tea plants with long trichomes but in low density, while others with short trichomes but at high density on leaf (Fig. S2; Table S2).

RNA‐seq transcriptome profiling on apical buds and the first leaves of 60 tea plant varieties were used to conduct the association mapping between trichome phenotypes and expression levels of TF genes. Association analyses of expression patterns of all TF genes (1609) in the tea plant genome with trichome phenotypes revealed a candidate MYB TF, TEA029017.1 (CsMYB1), strongly related to the long and high‐density trichome phenotype (r = 0.81, P = 5.11E−15) (Fig. 1a; Table S3). Meanwhile, we analysed all MYB family genes highly expressed in trichomes (Table S4) and CsMYB1 was one of the 10 MYB genes with the highest expression levels in trichomes (Fig. 1b). Heatmap data showed that the CsMYB1 was the gene with the highest expression level in young leaves where trichomes are mostly present (Figs 1c, S3a). Phylogenetic analyses showed that CsMYB1 shares high sequence similarity with other MYBs known to regulate trichome initiation, such as GaMYB2 in cotton and GL1 in Arabidopsis, and some others are predicted to regulate anthocyanin and proanthocyanin biosynthesis (Figs 1d, S4a–c). The distribution of trichomes on different tissue surfaces showed that the trichomes exhibited a tissue‐specific distribution. The trichome index was higher on the young tissues, such as the apical bud and the first leaf, but lower on the old leaf (Fig. 1e). qRT‐PCR data confirmed that CsMYB1 is highly expressed in apical buds and young leaves and stems with high trichome index (Fig. 1e). These data strongly indicated that CsMYB1 could be associated with trichome formation in tea plants.

Fig. 1

CsMYB1 is a candidate gene for regulating trichome formation in tea plants. (a) Pearson correlation analysis (r) between global transcription factors (TFs) (n = 1609) expression and leaf trichome density in cultivated tea accessions (n = 60). TEA029017.1 (CsMYB1) with highest correlation is highlighted by a red dot. (b) Heat map analysis of top 10 MYB TFs with the highest expression levels in trichomes. The expression level (FPKM) of each gene is shown in the heatmap boxes. The corresponding homologous genes of those TFs are shown. TEA029017.1 and its homologous GaMYB2 are indicated with blue colour. (c) The expression pattern of top 10 MYB TFs in developmental leaves and stems. The expression level (log10(FPKM)) of each gene is shown in the heatmap boxes. (d) Phylogenetic tree of those MYB TFs with others related to flavonoid biosynthesis and trichome formation. The numbers at the nodes indicate the bootstrap value with 1000 replicates. The bold TEA IDs highlighted their corresponding homolog genes in tea plants. (e) Trichome distribution and the CsMYB1 expression patterns in tea tissues. Trichome index was calculated by trichome length × trichome density. More than 30 samples of each tissue were observed and documented. Bars, 500 μm. Expression data are from at least three biological replicates and are expressed as means ± SD. CsACTIN was used as an internal control to normalize the expression data.

Trichome index variation as a result of tea plant domestication

As the universal presence but with the large variations in trichomes of natural tea plant populations including landraces and elite cultivars, the relationship between trichome index and CsMYB1 expression variations was examined between tea cultivars. First, six tea plant varieties with obviously extreme trichome phenotypes, either longest or shortest among their averages, were chosen (Fig. 2a). The tea cultivars with significantly longer and higher‐density trichomes, such as FDDB, DJBH and TYDY, than the cultivars with short and low trichome density, such as HJG, LJ and HK, were on abaxial sides of leaves (Fig. S3b). The trichome index of FDDB, DJBH and TYDY was significantly higher than that of HJG, LJ and HK (Fig. 2b), which is consistent with the high expression levels of CsMYB1 in FDDB, DJBH and TYDY (Fig. 2c). Furthermore, trichome indexes in 10 assamica‐type tea plants were also generally larger than those in 10 sinensis‐type plants on average (Fig. 2d); correspondingly, the CsMYB1 expression levels in the young leaves of these assamica‐type tea plants were also higher than those in sinensis‐type tea plants (Fig. 2e).

Fig. 2

Trichome morphology and density variation in tea plant varieties. (a) Morphological phenotype observation of tea germplasm resources with more trichomes (FDDB, DJBH and TYDY) and fewer trichomes (HJG, LJ and HK). Bars, 250 μm. (b, c) Quantification of the trichome indexes (b) and expression level (FPKM) of CsMYB1 (c) in those tea cultivars. All data are from at least three biological replicates and are expressed as mean ± SD (*, P < 0.05; **, P < 0.01; Student’s t‐test). (d, e) Comparison between the trichome indexes (d) and expression levels of CsMYB1 (e) of Assamica and Sinensis‐type plants. Ten Assamica and Sinensis cultivars, respectively, were used for investigating the trichome indexes and gene expression levels. All data are shown with red dots. The horizontal lines mean the median value (black bold lines) or the variation range (black thin lines). The minimum and maximum values are highlighted using vertical lines. (f) Variation of trichome index and catechins contents among representative tea plants of Sect. Thea (L.) Dyer. The catechins include two major galloylated cis‐catechins, Epigallocatechin‐3‐gallate and Epicatechin gallate. The data are expressed as means ± SD. Differences between tea plants were analysed with two‐factor ANOVA (lowercase letters indicate significance).

We next examined a broader scope of Camellia plants for more evidence of the involvement of CsMYB1 in the regulation of trichome formation, by investigating the trichome phenotypes in wild tea relative plants beyond modern tea cultivars. In contrast to modern tea cultivars, many wild tea relatives in Section Thea are covered with fewer trichomes (eFloras, 2018; http://www.efloras.org/florataxon.aspx; Liu et al., 2020). Recent chemical studies further demonstrated significant differences in the types and contents of catechins between wild tea relative and modern tea cultivars, such as galloylated cis‐catechins and polygalloylated glucoses. Among the varieties of Ser. Gymnogynae, Ser. Pentastyla, Ser. Quinquelocularis and Ser. Sinenses Chang, only the two varieties in Ser. Sinenses Chang, C. sinensis var. assamica (assamica type) and C. sinensis var. sinensis (sinensis type), were widely cultivated for making teas (Wei et al., 2018). The trichome indexes and galloylated cis‐catechins contents of non‐Ser. Sinenses Chang were lower than those in Ser. Sinenses Chang (Fig. 2f). The trichome indexes and galloylated cis‐catechins contents of assamica types are higher than those of sinensis types, Camellia  pubescens and Camellia angustifolia (Fig. 2f).

Natural variation in CsMYB1 expression is associated with trichome and galloylated cis‐catechins phenotypes in Camellia plants

To further understand the genetic causes of the significant variations in trichome indexes of these Camellia plants, we searched the genetic variations in c. 4 kb of promoter regions upstream from CsMYB1 gene in 81 diverse cultivated and wild tea plant accessions, whose genomes were re‐sequenced recently (Xia et al., 2020). Since wild ancestors of cultivated tea plants have not been identified yet, we therefore used C. taliensis, which is a close wild relative of C. sinensis located at the basal lineage of the section Thea, as the ‘wild’ tea plant to investigate their genetic diversity (P. Wang et al., 2020; X. Wang et al., 2020; Xia et al., 2020; Zhang et al., 2020, 2021). More than 1340 SNPs and Indels in CsMYB1 gene region were identified, which were further used to estimate the diversity in different tea plant populations. We also found that wild tea relatives possessed the highest divergence, with landraces having less, while the cultivated group had the lowest divergence (Fig. 3a). This result indicates that CsMYB1 may be involved in the selection during domestication, which was well supported by the level of nucleotide diversity in each tea plant group (Fig. 3b). Interestingly, significant variations were detected in the promoter regions of CsMYB1 gene from the cultivated and wild tea plants (Fig. 3b,c).

Fig. 3

Genetic variations of the CsMYB1 in Camellia plants. (a) Nucleotide diversity (π) and genetic divergence (F ST) between the wild, ancient and cultivated tea plants. The circle size indicates π, and the numbers marked between each subpopulation indicate the Weir and Cockerham weighted F ST values. (b) Nucleotide diversity of the wild (grey), ancient (blue) and cultivated (orange) tea plant populations. Black boxes mean the exons of CsMYB1. (c) The variation of the CsMYB1 promoter sequence in cultivars and wild. Twelve cultivars (Asama and Chinses), and a wild tea relative (Camellia taliensis) were used for the CsMYB1 promoter sequence isolation. The promoter sequences of CsMYB1 were sequenced by Sanger Sequencing. The promoter sites are shown according to the reference sequence (SCZ). The 192‐bp insertion found in the promoter of CsMYB1 in modern tea cultivars is shown as a red dash, and the cis‐elements in the insertion region of CsMYB1 promoter are indicated by lines. (d) CsMYB1 expression level (FPKM) in wild and cultivar (Cul) tea plant accessions. (e) Catechins contents in wild and cultivar (Cul) tea plant accessions. (f) Trichome index in wild and cultivar (Cul) tea plant accessions. The horizontal lines in boxes mean the median value (black bold lines) or the variation range (black thin lines). The minimum and maximum values are highlighted using dotted lines. (g) Constructs for transient expression of CsMYB1 promoters. (h, i) The activity test of CsMYB1 promoters from wild and cultivar (SCZ and ZH) in tobacco leaves. All data are from at least three biological replicates and are expressed as means ± SD (**, P < 0.01, Student’s t‐test).

Furthermore, a 192‐bp insertion was found exclusively in the CsMYB1 promoter regions of cultivated tea plants, but not in the wild plant accession C. taliensis (Figs 3c, S5). This 192‐bp insertion contains many cis‐elements that could enhance CsMYB1 gene expression (Fig. 3c; Table S5). Further analyses showed that this insertion is correlated with CsMYB1 expression level, leaf trichome index and galloylated cis‐catechins contents (Fig. 3d–f). We further tested the promoter activity of CsMYB1 in the GUS transient expression assay system in tobacco leaves. The 1.7 kb promoter sequences of CsMYB1, CsMYB1 and CsMYB1 were used to drive the GUS reporter gene (Fig. 3g). The promoter of CsMYB1 without the 192‐bp insertion exhibited significantly lower GUS activation compared with the promoter of CsMYB1 and CsMYB1 with the insertion (Fig. 3h,i). Thus, the 192‐bp insertion in the promoter region resulted in the elevation of CsMYB1 expression level and lead to the higher trichome distribution in the cultivated tea plants.

CsMYB1 regulated trichome formation in tea plants

To verify the functions of CsMYB1 in trichome formation, genetic complementation of Arabidopsis gl1 mutant was performed. While the gl1 mutant leaves and stems display defective in trichome development compared with the wild‐type, the overexpression of CsMYB1 in the gl1 mutant restored the trichome phenotypes to wild‐type (Figs 4a, S4d). Gene expression analysis also showed that CsMYB1 could activate some downstream TF genes, such as AtGL2 (Fig. 4b), AtGL3 and AtCPC (Fig. 4c), which play important roles in trichome formation in Arabidopsis (Johnson et al., 2002; Ishida et al., 2007). These data suggested that CsMYB1 is an orthologue of AtGL1 and can form a MBW complex to activate downstream genes.

Fig. 4

CsMYB1 positively regulated tea leaf trichome development. (a) Genetic complementation of the Arabidopsis gl1 mutant by CsMYB1. (b, c) CsMYB1 significantly activated downstream transcription factor (TF) AtGL2 (b) and AtGL3 and AtCPC (c) expression in gl1 mutant. (d) CsMYB1 expression was significantly reduced in the tea leaves treated with antisense oligonucleotides (asOND)‐CsMYB1 (CsMYB1‐KD) as compared with sense OND (sOND). (e, f) The effect of CsMYB1 downregulation on the expression the downstream target TFs CsTTG2 and CsGL2 (e), and CsGL3s and CsCPC (f). (g) Representative picture of transgenic tea hairy roots expressing CsMYB1 and GFP control. The hairy roots are indicated by the white arrows. (h) qRT‐PCR confirming the overexpression of CsMYB1 in transgenic tea hairy roots. (i) The expression of CsGL3a, CsCPC, CsTTG2 and CsGL2 in the CsMYB1 overexpression hairy roots. All data are from at least three biological replicates and are expressed as means ± SD (*, P < 0.05; **, P < 0.01; Student’s t‐test). CsACTIN and AtACTIN were introduced to normalize the expression data.

In tea plants, an antisense oligodeoxynucleotide (asOND)‐interfering experiment was performed to create CsMYB1‐knockdown (CsMYB1‐KD) in tea leaves to verify the function in trichome formation. CsMYB1 transcript levels were significantly reduced following 72 h of asOND treatment compared with sense OND treatment, as verified with qRT‐PCR (Fig. 4d). Meanwhile, the expression of putative targets of CsMYB1, including the HD‐Zip TF gene CsGL2, the WRKY TF gene CsTTG2, the bHLH gene CsGL3 and the R3 MYB gene CsCPC, were significantly repressed in CsMYB1‐KD leaves (Fig. 4e,f). To further characterize CsMYB1, we created Agrobacterium‐mediated tea plant transgenic hairy roots with ectopic overexpression of CsMYB1 (CsMYB1OE) (Fig. 4g). Even though the overall transformation rate was low (c. 10%), four independent hairy root lines were obtained and verified with qRT‐PCR (Fig. 4h). Results showed that in comparison with GFP roots, the expression of CsGL2, CsTTG2, CsGL3 and CsCPC in CsMYB1OE were all up‐regulated at different levels (Fig. 4i).

Regulation of tea plant trichome formation by MBW complexes

In Arabidopsis, AtGL1 regulated trichome formation by forming MBW complex with AtGL3/EGL3 and AtTTG1 to activate AtGL2 expression. In both Arabidopsis and tea plants, we found that CsMYB1 could activate the expression of GL3, CPC and GL2 genes (Fig. 4). Thus, we further explored the CsMYB1‐directed MBW complex and the functions of their partners in trichome formation in tea plants. First, we identified the functions of tea GL3, WD40 and CPC TFs in trichome initiation (Figs S6, S7). Then, subcellular localization assay confirmed this, except that GFP‐CsWD40 was observed in both nucleus and cytoplasm, whereas GFP‐CsMYB1, GFP‐CsGL3a and GFP‐CsCPC were all observed only in the nucleus of tobacco epidermal cells (Fig. 5a). Both yeast Y2H and BiFC assays showed the protein interactions among MYB (CsMYB1 and CsCPC) and WD‐40 (CsWD40), with bHLHs (CsGL3a, CsGL3b, or CsGL3c) (Figs 5b,c, S8). Therefore, CsMYB1 or CsCPC, CsGL3 and CsWD40 indeed formed a ternary transcriptional regulatory complex. As GL2 is a critical regulator for plant trichome formation, the EMSA was performed to examine the binding of CsMYB1 to the promoter of CsGL2 gene. The electrophoretic mobility shift assay results showed that CsMYB1 could directly bind to the promoter of CsGL2 (Figs 5d, S9a–e).

Fig. 5

CsMYB1‐CsGL3‐CsWD40 complexes activated GL2 promoter. (a) Subcellular localization of GFP‐CsMYB1, GFP‐CsGL3a and GFP‐CsWD40 in tobacco epidermal cells. Bars, 100 μm. (b) Yeast two‐hybrid (Y2H) assay to check the MBW complex formation. CsMYB1, CsGL3s and CsWD40 in Matchmaker Gold Y2H vectors pGBKT7 and pGADT7 with reading frame in fusion with the GAL4 BD and AD, respectively, were tested in different combinations for their interaction. The interaction was verified with growth in SC‐Trp‐Leu‐His. Representative photographs from at least three experiments. (c) Bimolecular fluorescence complementation verification of the formation of MBW complexes in tobacco epidermal cells. Bars, 100 μm. (d) The binding of CsMYB1 to CsGL2 promoter detected by the electrophoretic mobility shift assay. The competitor represents the putative motif without the biotin label. The concentration of competitors with different rations with a biotin‐labelled motif is 5×, 10× or 50×. The biotin signal was indicated by a black arrow. (e) The activation or repression of the CsGL2 promoter in transient expression assay using Arabidopsis protoplasts with Renilla luciferase activity as reference. The data are from at least three biological replicates and are expressed as means ± SD.

To further test the regulatory functions of tea plant MBW complexes, we conducted a transactivation assay with both CsGL2 promoters (Fig. S9f). Co‐expression of CsMYB1, CsWD40 and CsGL3a activated the CsGL2 promoter‐driven luciferase reporter gene by > 5‐fold compared with no‐effecter assay; meanwhile, the MBW activation of CsGL2 promoter could be completely repressed by CsCPC (Fig. 5e). These results indicated that CsMYB1/CsCPC‐CsGL3‐CsWD40 formed a ternary regulatory complex to regulate trichome formation via directly regulating downstream effectors in tea plants.

CsMYB1 regulates galloylated cis‐catechins biosynthesis in tea leaves

Tea plants contain multiple types of catechins, such as galloylated catechins (epigallocatechin‐3‐gallate (EGCG), epicatechin gallate (ECG) and gallocatechin gallate (GCG)) and nongalloylated catechins (catechin (C) and epicatechin (EC)) (Fig. 6a) (Lashbrooke et al., 2015; Zhao et al., 2020). However, the regulatory mechanism of galloylated catechins biosynthesis in plants is unknown (Jiang et al., 2018; P. Wang et al., 2020; X. Wang et al., 2020).

Fig. 6

CsMYB1 was involved in galloylated cis‐catechins (EGCG and ECG) biosynthesis processes in tea leaves. (a) Biosynthesis pathway for galloylated cis‐catechins in tea plants. Phenylpropanoid metabolic fluxes mainly flow into cis‐catechins (EGCG and ECG) biosynthesis, and fewer flow into trans‐catechins (C and GC) biosynthesis. Some key biosynthetic enzymes are highlighted with the blue colour and black solid arrows. Other related enzymes are represented with a dashed arrow. ANR, anthocyanidin reductase; ANS, anthocyanidin synthase; C, catechin; DFR, dihydroflavonol 4‐reductase; ECG, epicatechin gallate; EGCG, epigallocatechin‐3‐gallate; GC, gallocatechin; LAR, leucoanthocyanidin reductase; SCPL1A, type 1A serine carboxypeptidase‐like acyltransferases; β‐G, β‐glucogallin. (b) Pearson correlation analysis of the expression of candidate MYB genes and the contents of galloylated cis‐catechins (EGCG, ECG), as well as the expression (FPKM) of galloylated cis‐catechins biosynthetic genes, CsANR and CsSCPL1A, from eight different tissues and developing leaves, including the apical bud, the young leaf, the mature leaf, the old leaf, flower, fruit, root and stem; developing leaves included the apical bud, the first leaf, the second leaf, the third leaf, the fourth leaf and the old leaf. The sizes of circles indicate the degree of correlation. Green indicates positive, and red, negative correlation. (c) Enhanced CsMYB1 transcription and catechins contents in the leaves of tea plant by abscisic acid (ABA) and methyl jasmonate (MeJA) treatment or insect attack. Data were from three biological replicates and are expressed as means ± SD. The difference between treatments and control is significant when *, P < 0.05; **, P < 0.01 in Student’s t‐test. CsACTIN was used to normalize the expression data.

In total, 10 R2R3‐type MYB TFs were predicated in the tea genome as the homologs of Arabidopsis TT2 and MYB5 regulating catechins biosynthesis (Fig. 6b). However, correlation analysis showed that only CsMYB1 and TEA031375 (CsMYB5b) had the highest correlation with the galloylated cis‐catechins contents and the expression of corresponding biosynthetic genes anthocyanidin reductase (CsANR) and serine carboxypeptidase‐like 1A (CsSCPL1A) (Figs 6b, S10; Table S6). As CsMYB5b has been reported to regulate catechins biosynthesis (P. Wang et al., 2020; X. Wang et al., 2020), and CsMYB1 had a much higher expression level than CsMYB5b did in tea leaves, CsMYB1 could be another key regulator of catechins biosynthesis (Fig. S11a). These data in Figs 2(e) and 3(e) also indicate the positive correlation between the trichome distribution and galloylated cis‐catechins (mainly EGCG and ECG) contents. Moreover, when the apical buds and the young leaves of tea plants were sprayed with ABA, MeJA or attacked by insects, CsMYB1 transcripts increased drastically. Meanwhile, the levels of galloylated cis‐catechins (EGCG and ECG) also increased (Fig. 6c), implying that CsMYB1 was also regulated by ABA, MeJA or insect attack and mediated these stress‐induced biosynthesis of galloylated cis‐catechins. These observations were consistent with previous reports (Tables S7, S8) (Shi et al., 2015; Xia et al., 2020).

We also found that the contents of galloylated cis‐catechins (EGCG, ECG) and the expression of corresponding biosynthetic genes, such as ANR, anthocyanidin synthase (ANS) and SCPL1A, were declined in CsMYB1‐KD leaves compared with sense OND treatments (P < 0.01) (Figs 4d, 7a,b). On the contrary, the overexpression of CsMYB1 in tea seedling hairy roots significantly increased the contents of galloylated cis‐catechins, as verified with MS/MS (Figs 4g, 7c,d, S10). The CsANR and CsSCPL1A play crucial roles in galloylated cis‐catechins syntheses, and their coding genes' promoters also contained many MYB binding sites, such as MYB1AT and MYBST1 (Fig. 7e). To uncover how CsMYB1 regulate galloylated cis‐catechins syntheses, the EMSA and promoter activation assays were performed. The EMSA results clearly showed that the CsMYB1 could directly bind to the promoters of CsANR and CsSCPL1A (Figs 7e,f, S9e). Promoter activation assays also showed that the co‐expression of CsMYB1, CsWD40 and CsGL3a/CsTT8 could activate the promoters of galloylated cis‐catechins biosynthetic genes, CsANR or CsSCPL1A, by more than 10‐fold compared with no‐effecter assays (Fig. 7g–i). These results indicated that CsMYB1 interacted with CsWD40 and CsGL3a/CsTT8 and formed ternary regulatory complexes to regulate galloylated cis‐catechins biosynthesis via directly regulating biosynthetic genes in tea plants.

Fig. 7

CsMYB1 promoted catechins biosynthesis in tea leaves. (a) The repressed biosynthesis of catechins (EGCG, ECG) in CsMYB1‐KD leaves compared with control. (b) The effect of CsMYB1 downregulation on the expression of key genes involved in catechins biosynthesis. (c) HPLC chromatographs of major catechins in tea hairy roots expressing CsMYB1 and GFP control. (d) The effects of ectopic expression of CsMYB1 on catechins in tea transgenic hairy roots as compared with GFP control. The roots of similar statuses were chosen for comparison. (e) The selected sites from ANR and SCPL1A promoters for the electrophoretic mobility shift assay (EMSA) are indicated through red colour. (f) The binding of CsMYB1 to the putative motifs in the ANR and SCPL1A promoters was detected by the EMSA. The competitor represents the putative motif without the biotin label. The concentration of competitor with different rations with a biotin‐labelled motif is 5×, 10× or 50×. The biotin signal is indicated by a black arrow. (g–i) The activation of the ANR (h) and SCPL1A (i) promoters in transient expression assay using Arabidopsis protoplasts with Renilla luciferase activity as reference. All data are from at least three biological replicates and are expressed as means ± SD (*, P < 0.05; **, P < 0.01, Student’s t‐test). CsACTIN was used to normalize the expression data.

Discussion

In addition to discouraging insect predators, reflecting UV‐ and high light, insulating the plant body from high temperature and protecting the leaves from microbial attack, trichomes on tea plant buds and young leaves also contribute to tea flavours and health functions because of their rich metabolites. However, the mechanism underlying the trichome formation and biosynthesis of tea characteristic secondary metabolites is not fully understood (Zhao et al., 2020). By correlation analyses of transcriptomes and trichome phenotype, we identified and defined CsMYB1 as a critical regulator for trichome formation and the biosynthesis of galloylated cis‐catechins.

CsMYB1‐centred MBW complex plays a role in tea trichome formation

MBW regulatory complexes are responsible for co‐ordinarily regulating the spatial positioning of trichomes and their density on leaves (Pesch & Hulskamp, 2009). The surface hair (trichome or root hair) patterning is regulated by both positive and negative regulators in complex ways. AtGL1 acts as a key factor in the protein–protein dynamic interaction network with other regulators moved between trichome formation and hair‐less cells where they determinate the fates of the cells. The initiation and development of unicellular trichomes in Arabidopsis are regulated by AtGL1 with a WD40 TF TTG1 and bHLH TFs GL3/EGL3 (Serna & Martin, 2006; Pesch & Hulskamp, 2009).

From an association study on tea plant trichome phenotypes and TF gene expression, we found several TFs were highly associated with trichome phenotypes. Among them, an AtGL1‐like MYB TF, CsMYB1 caught our attention with the highest correlation efficiency. We showed that the unicellular unbranched trichomes on tea plant leaves are also regulated by an MBW ternary complex composed of CsMYB1, CsGL3 and CsWD40. These MBW complexes can activate downstream TF genes such as CsGL2 and CsCPC, whose homologs in Arabidopsis function to feedback inhibition, activation or diversification of MBW regulatory functions (Marks et al., 2008; Zhao et al., 2008; Won et al., 2009; Bruex et al., 2012; Pesch et al., 2015; Chopra et al., 2019). Similarly, we observed inhibitory effects of CsCPC on CsMYB1‐regulatory complex’s activation of GL2 gene, in both transgenic Arabidopsis and tea plants. Moreover, both asOND knockdown of CsMYB1 on shoot tips and overexpression of CsMYB1 in transgenic hairy roots showed CsMYB1‐dependent regulation of MBW downstream effector genes CsGL2, CsTTG2, CsGL3s or CsCPC, which are important TFs regulating trichome or root hair formation. Thus, CsMYB1 was defined as a critical trichome formation regulator in tea plants.

CsMYB1 owns the six solvent‐exposed amino acid residues ([D/E]Lx2[R/K]x3Lx6Lx3R) that specify the interaction with bHLH proteins and are conserved in the R2R3 MYBs regulating trichome formation through an MBW regulatory complex mechanism (Serna & Martin, 2006) (Fig. S4a,b). The regulators in the same subclade with CsMYB1, including cotton GhMYB2 and GhMYB109, Arabidopsis AtGL1, and their functional equivalents AtMYB23 and AtMYB66 (WER), are all known trichome‐ or root hair‐formation regulators (Serna & Martin, 2006). CsMYB1 is far different from R2R3 MYBs in another clade, including tomato trichome regulators MIXTA and MIXTA‐like R2R3 MYBs, Arabidopsis AtMYB106, cotton GhMYB25 that have mutated amino acids in the bHLH‐binding domain (Lashbrooke et al., 2015) (Fig. S4c). MIXTA and MIXTA‐like R2R3 MYBs may specifically regulate cuticle formation in reproductive organs and trichomes (Oshima et al., 2013; Lashbrooke et al., 2015). However, CsMYB1 is much closer to these R2R3 MYBs that regulate flavonoid biosynthesis, through the formation of MBW complexes, such as MtPAR, AtPAP1 and MtLAP1 (Li et al., 2016), indicating an evolutionary trend of CsMYB1 towards combining regulatory functions in both trichome formation and secondary metabolite synthesis.

CsMYB1‐CsGL3‐CsWD40 complex regulates biosynthesis of galloylated cis‐catechins

The GL3/EGL3, TTG1 and their orthologs have been well described in regulating multiple secondary metabolite biosynthesis in many plants. However, not much is reported about the function of GL1 and its orthologs on plant secondary metabolism. A previous study provided the reliable evidence that GL1 could directly bind to promoter regions of biosynthetic genes involved in many metabolism processes, such as lipid metabolism (Morohashi & Grotewold, 2009). Here, our study showed that the tea trichome regulator CsMYB1 directly activated key genes involved in galloylated cis‐catechins biosynthesis.

Even though galloylated cis‐catechins biosynthesis is not fully understood, ANR and SCPL1A acyltransferases were shown to play crucial roles in the last step of galloylated cis‐catechins biosynthesis (Wei et al., 2018; Ahmad et al., 2020). However, little is known about the regulatory mechanisms of a galloylated cis‐catechins biosynthesis. Because of the lack of a galloylated catechins biosynthesis pathway in model plants, for example, Arabidopsis and tobacco, tea plant TFs were shown to function as the nongalloylated catechins regulators when heterologously expressed. For example, the CsMYB5a, CsMYB5b and CsMYB5e could significantly promote C, EC and polymeric PA biosynthesis in tobacco flowers (Jiang et al., 2018; P. Wang et al., 2020; X. Wang et al., 2020). In this study, ectopic expression of CsMYB1 could promote anthocyanin biosynthesis and related gene expression level in Arabidopsis (Fig. S12a,b), and in tea plants, CsMYB1 expression pattern also was correlated with multiple metabolic processes (P < 0.01), such as catechins biosynthesis (Figs S11b, S13). In the tea genome, at least 10 MYB TFs were predicted to regulate catechins biosynthesis. However, only CsMYB5b (P. Wang et al., 2020; X. Wang et al., 2020) and CsMYB1 (this study) were firmly predicted as regulators for the biosynthesis of the galloylated cis‐catechins based on the correlation analysis integrating metabolism profiling and gene expression pattern. Given the higher expression of CsMYB1 than the CsMYB5b in young leaves, CsMYB1 might be a major regulator of galloylated cis‐catechins biosynthesis.

Indeed, both the EMSA and promoter transactivation assay supported that the CsMYB1‐mediated MBW complexes could directly activate galloylated cis‐catechins synthesis genes, including CsANR and CsSCPL1A. Knockdown or overexpression of CsMYB1 in tea plants could significantly down‐ or upregulate the galloylated cis‐catechins, respectively. Moreover, the expression patterns of CsMYB1 in tea plant shoot tips were responsive to insect attacks and stress hormones and were also tightly correlated to the levels of galloylated cis‐catechins. The CsMYB1‐CsGL3/CsTT8‐CsWD40 complexes play critical roles in regulating the production of the specialized metabolites in tea plants in response to environmental or developmental cues (Lepiniec et al., 2006; Li et al., 2016). As an essential target of MBW complexes, GL2 not only critically regulates trichome formation but also diversify the MBW functions to other metabolic pathways (Ohashi et al., 2003; Tominaga‐Wada et al., 2009; Khosla et al., 2014). As a repressor, GL2 negatively regulates anthocyanin biosynthesis by directly repressing the MBW activator complex, such as the production of anthocyanin pigment 1 (PAP1), PAP2, MYB113, MYB114 and TT8 (Wang et al., 2015). Likewise, CsMYB1 can also modify tea plant characteristic metabolites through CsGL2. For the tea trichomes that especially contribute to tea plant defence capacity and nutritional quality (Schilmiller et al., 2008; Li et al., 2020), CsMYB1 may be used for a molecular marker for breeding new tea plant varieties with better quality traits.

Variation in CsMYB1 expression attributable to the different trichome index and galloylated catechins contents in Camellia varieties

Tropical tea plants are subject to attacks by a wide range of microbial pathogens and herbivores, which caused more than 380 diseases and production losses (Lehmann‐Danzinger, 2000). While most other Thea section plants, such as C. tachangensis, C. taliensis and C. angustifolia, have glabrous leaves or have short and rare trichomes, almost all current tea plant cultivars, both assamica and sinensis types, have long and densely spaced trichomes (eFloras, 2018; http://www.efloras.org/florataxon.aspx; Li et al., 2020; Liu et al., 2020). These significant differences were precisely quantified and correlated with a trichome index in this study (Fig. 2). Furthermore, several studies have revealed the lower galloylated cis‐catechins contents and higher hydrolysable tannins (polygalloylated glucose and derivatives) contents in these wild tea relatives as compared to those in most modern tea cultivars (Meng et al., 2019; Liu et al., 2020). Not only do these trichomes confer insect and pathogen resistance and help tea plants survive against both abiotic and biotic stresses, but also their enriched special chemicals also contribute to tea flavours (Li et al., 2020). Thus, the thousands of years of tea plant cultivation in adversary environments and farmer selection history for better tea flavour have likely driven tea plant domestication (Zhang et al., 2021). During the plant domestication, traits such as higher levels of galloylated cis‐catechins (EGCG and ECG) and lower plant height may dominate in most contemporary tea plant cultivars or elites (Xia et al., 2020; Yu et al., 2020; Zhang et al., 2021). Our data here show that the presence of long and densely spaced trichomes on the young leaves of tea plant and high contents of galloylated cis‐catechins are the two correlated domestication traits in tea plants; they might be the targets of tea plant breeding for enhanced resistance against herbivores and other stresses, as well as for better tea flavours (Fig. 2; Yagi et al., 2009; Jin et al., 2018; Liao et al., 2019; Zhu et al., 2019; Liu et al., 2020; Zhao et al., 2020). Our study showed that, as an indispensable regulator gene for tea plant trichome formation and galloylated cis‐catechins accumulation, CsMYB1 is likely selected from the wild tea relative to modern tea plant cultivars, although other mechanisms may also play roles in the domestication of these significant traits in tea plants (Fig. 3).

In summary, we characterized CsMYB1 as a trichome regulator in tea plants and further showed that CsMYB1 regulated both trichome formation and biosynthesis of galloylated cis‐catechins (Fig. 8). Although only some wild tea relatives and more than 20 assamica‐ and sinensis‐type tea cultivars were screened with PCR for the variation in CsMYB1 promoters in this study, and there were some other SNP/Indels in CsMYB1 promoters from different varieties, the drastic differences in trichome index and galloylated cis‐catechins contents in wild tea plants and modern tea cultivars are closely related to CsMYB1 expression levels. Thus, these two correlated domestication traits in Camellia plants are largely attributable to CsMYB1, although other epistatic factors may be also involved. CsMYB1 transcription responded to multiple biotic and abiotic stresses. As a domestication gene, CsMYB1 may be selected by both natural adverse biotic and abiotic stresses and farmers for better tea flavour contributed by higher levels of galloylated catechins in tea leaves: the CsMYB1–MBW complex regulated both trichome development regulator genes CsGL2 and CsCPC, as well as the key galloylated cis‐catechins biosynthetic genes CsANR and CsSCPL1A. This study lays a foundation for the upcoming studies that explore in detail the molecular mechanisms of trichome formation and genetic improvement of major secondary metabolites in tea plants, for both plant resistance and tea qualities of both sensory perception and health promotion.

Fig. 8

Working model for the regulatory functions of CsMYB1 in tea plants. CsMYB1 interacts with CsGL3 and CsWD40 to form an MBW complex, which directly activates CsGL2 and then trichome development in tea plants. CsCPC activated by the MBW could feedback inhibit trichome formation by competitively forming CsCPC‐CsGL3‐CsWD40 complex with CsMYB1 to repress CsGL2 expression. The CsMYB1‐directed MBW complex also upregulates catechins biosynthesis genes, such as ANS, ANR and SCPL1A, and mostly promotes the production of galloylated cis‐catechins (such as EGCG and ECG), but less significantly affect trans‐catechins biosynthesis (such as C and GC). The oval shapes in different colours are marked with various transcription factors or metabolic enzymes. The black arrows show transcriptional downstream targets or effects, and the red blunted arrow denotes inhibitory regulation, and the orange arrows show biosynthesis pathways, in which the solid arrows indicate one‐step reactions, while the dashed arrow represents a multiple‐step reaction.

Author contributions

JZ conceived and designed the experiments; Penghui Li, JF, YX, YS, YZ, ZY, WT, XZ, JY, DT, Ping Li and HZ performed experiments; Penghui Li, YX, WT, QW and EX analysed the data; JZ, Penghui Li and SW wrote the manuscript.

Fig. S1 Normal distribution test of the tea trichome density and length phenotype in a natural population.

Fig. S2 Trichome appearances on leaves of different tea plant varieties.

Fig. S3 Expression analysis of ten MYB transcription factors in different tissues of tea plants.

Fig. S4 Functional analysis of CsMYB1 protein in trichome formation.

Fig. S5 Sequences of CsMYB1 promoter in the wild and cultivar tea plants.

Fig. S6 Identification of CsWD40, CsGL3s and CsCPC in tea plants.

Fig. S7 Functional identification of CsWD40, CsGL3s and CsCPC in trichome formation.

Fig. S8 Negative controls of bimolecular fluorescence complementation assay in Fig. 5(c).

Fig. S9 Identification of CsGL2 from tea plants for trichome formation.

Fig. S10 Verification of catechins compounds by the HPLC and LC‐MS.

Fig. S11 Expression and correlation analysis of CsMYB1 in tea plants.

Fig. S12 Regulation of flavonoid biosynthesis by CsMYB1 overexpression in Arabidopsis.

Fig. S13 MYB‐binding sites in the promoters of the predicted target genes.

Click here for additional data file.

Table S1 List of primers used in this study.

Table S2 The trichome indexes and catechins contents of tea cultivars used for correlation analysis.

Table S3 Association analysis of the expression of all the TF genes in the tea genome with trichome phenotypes.

Table S4 Expression levels of MYB family genes in trichomes and leaves.

Table S5 Analysis of cis‐acting regulatory DNA elements in the 192‐bp sequence of the CsMYB1 promoter in modern tea cultivars.

Table S6 Gene expression and catechins contents in tea plants.

Table S7 The expression patterns of genes related to flavonoid biosynthesis in tea plants under MeJA treatment.

Table S8 The expression patterns of genes related to flavonoid biosynthesis in tea plants under tea geometrid biting treatment.

Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Click here for additional data file.