An ancient whole-genome duplication event and its contribution to flavor compounds in the tea plant (Camellia sinensis)

Evidence of an ancient WGD in the tea plant genome

We first studied the genomic syntenic relationships among the five species, i.e., tea plant, kiwifruit, grape, persimmon, and rhododendron (Supplementary Fig. S1). With a total of 1342 syntenic blocks covering 21,873 gene anchor pairs (Supplementary Fig. S1e), the syntenic blocks between tea plant and kiwifruit showed a 2 : 4 syntenic relationship. For example, CsChr04 and CsChr09 in tea plant matched AcChr04, AcChr26, AcChr27, and AcChr28 in kiwifruit (Fig. 2a). In addition, a total of 849 syntenic blocks covering 16,827 homologous gene anchor pairs were detected between tea plant and rhododendron, showing a significant 2 : 2 syntenic relationship (Supplementary Fig. S1d). For example, CsChr01, CsChr05, and CsChr07 in tea plant matched C024955.1, C024956.1, C024958.1, and C024961.1 in rhododendron (Fig. 2b). Tea plants and grapes showed 2 : 1 syntenic relationships (Fig. 2c), and the same relationship was found in the analysis of wild tea plants DASZ and grapes (Supplementary Fig. S1g, h). These results indicated that tea plants experienced WGD after the WGT.

**Fig. 2: Syntenic relationships and the Ks distribution suggested an ancient WGD event in the tea plant genome.**

The distribution of the Ks of the syntenic paralogs of tea plant, kiwifruit, grape, and rhododendron also revealed a similar result (Fig. 2d). We constructed the Ks distributions for the tea plant, rhododendron, persimmon, kiwifruit, and grape (Fig. 2d and Supplementary Fig. S4). The distribution of Ks showed that tea plant had a Ks peak at 0.425, which is on the left side of the grape peak at 1.088, suggesting that tea experienced an additional WGD after the WGT-γ. Rhododendron had a Ks peak similar to that in tea plants. Only kiwifruit showed two peaks, indicating that kiwifruit experienced additional WGDs that were not experienced by tea plant or rhododendron.

Molecular dating of this ancient WGD

We have shown that tea plants experienced only one WGD after the WGT-γ. However, was this WGD tea plant specific or shared with kiwifruit, or even more species? It cannot be clearly determined through only the Ks distribution shown in Fig. 2d. Therefore, we constructed phylogenetic trees to determine whether this WGD is tea plant specific. We finally constructed 2798 single-copy nuclear gene-based phylogenetic trees covering tea plant (order Ericales, family Theaceae), rhododendron (order Ericales, family Ericaceae), kiwifruit (order Ericales, family Actinidiaceae), and outgroups (grape or coffee) (see “Methods” and Supplementary Table S2). Among them, type I (supporting tea plant, rhododendron, and kiwifruit sharing the WGD) (Fig. 3a) had 1021 trees, accounting for 36.5% of all phylogenetic trees. Type II (supporting tea plant experiencing the WGD independently) (Fig. 3b) had 471 trees, accounting for 16.8% of all trees. The rest of the trees were types other than type I or type II. We determined the proportions of the two types (Fig. 3e). The value of type I/(type I + type II) was 68.5%, whereas type II/(type I + type II) was 31.8%; the proportion of type I trees was more than twice that of type II trees. Another question was whether the close relative persimmon (D. lotus, order Ericales, family Ebenaceae) shared this WGD or not. A total of 168 single-copy nuclear gene-based phylogenetic trees covering tea plant, persimmon, and outgroups (grape or coffee) were constructed (Supplementary Table S2). Among them, 67 (40%) phylogenetic trees (type III) (Fig. 3c) clearly supported that tea plants and persimmons shared the WGD. Only ten (6%) phylogenetic trees (type IV) (Fig. 3d) supported that the two species did not share the WGD (Fig. 3f). Therefore, this evidence showed that the tea plant shared this WGD with rhododendron (family Ericaceae), kiwifruit (family Actinidiaceae), and persimmon (family Ebenaceae). According to the phylogenetic taxonomy of the order Ericales²⁶, this WGD was shared by at least 17 families in three sections: Polemonioids, Primuloids, and core Ericales (PPC). Hence, this WGD was termed PPC-WGD in this study (Fig. 4b).

**Fig. 3: Phylogenetic trees of syntenic genes showing that the WGD was shared by tea plant, kiwifruit, rhododendron, and persimmon.**

**Fig. 4: The PPC-WGD shared by four families within the order Ericales and the evolution of tea plant chromosomes.**

In addition, we inferred chromosome evolution in tea plants. The chromosomes were plotted before and after PPC-WGD based on the syntenic gene blocks within the tea plant genome (Supplementary Fig. S8). Prior to PPC-WGD, nine ancestral chromosomes (2n = 18) were reconstructed. PPC-WGD then produced 18 ancestral chromosome intermediates (2n = 36). After three fissions and six fusions, the number of tea plant chromosomes reached the current number, 15 (2n = 30) (Fig. 4a). The evolution of tea plant chromosomes from 7 chromosomes (prior to WGT-γ) to 9 chromosomes showed that almost all chromosomes were formed by a series of fusions, interchanges, and insertions (Supplementary Fig. S6).

To infer the occurrence time of the PPC-WGD, the Ks = t/2r method, which has been widely applied to calculate the occurrence time of WGDs in many articles^25,28,32, was used. The results showed that the time of the PPC-WGD was ~63 MYA (Fig. 4b), which is very close to the mass extinction at ~66 MYA at the Cretaceous-Paleogene (K-Pg) boundary³³.

Contributions of PPC-WGD to characteristic secondary metabolites in tea

Caffeine, theanine, and catechins are the three most characteristic secondary metabolites in tea, playing important roles in creating tea flavor. At present, a series of studies have revealed that tandem duplication is the main reason for the accumulation of these special secondary metabolites, such as caffeine¹⁰ and catechins⁴, in tea plants. However, whether the PPC-WGD contributed to the amplification of genes related to these special secondary metabolites in tea plants is unclear. The biosynthesis of catechins involves regulation by many key enzymes, including phenylalanine ammonia lyase (PAL), leucoanthocyanidin reductase (LAR), anthocyanidin reductase (ANR), and many other key enzymes³⁴ (Fig. 5a). Our analyses showed that a pair of LAR genes (Fig. 5b), a pair of CHALCONE SYNTHASE genes (Fig. 5c), two pairs of PAL genes (Fig. 5d, e), a pair of FLAVONOL SYNTHASE (FLS) genes (Fig. 5f), a pair of type 1A SERINE CARBOXYPEPIDA-LIKELTRANSFERASE (SCPL 1A) genes (Fig. 5g), a pair of ANR genes (Fig. 5h), and a pair of ANTHOCYANIDIN SYNTHASE (ANS) genes (Fig. 5i) were gene pairs with strong syntenic relationships produced by the PPC-WGD. The expression profile of the duplicated genes showed that most of these genes had at least one copy and sometimes both copies had high expression in the apical bud and leaf organ. In addition, only one copy of LAR (TEA026458.1), PAL (TEA003137.1), and ANS (TEA015769.1) had low expression in plant organs; under the different temperature treatments, both copies of most genes showed high expression in plant organs or under the temperature treatments. For example, two copies of ANR were highly expressed in apical buds and young leaves (Fig. 5h). The FLS copy TEA016601.1 had high expression in flowers, whereas TEA0010328.1 had high expression in the third mature leaf at severe low temperature and moderate low temperature (Fig. 5f). Although the two copies of most PPC-WGD gene pairs did not have high expression in the same organs or the same temperature treatments, the two copies were expressed at higher levels in different organs or under different treatments, showing that the two PPC-WGD copies contribute to the biosynthesis of catechins in different organs and under different temperatures in tea plants.

**Fig. 5: PPC-WGD contributed to the expansion of genes involved in the biosynthesis of catechins in tea plants.**

Theanine, a nonprotein amino acid found in Camellia plants that accounts for ~70% of the total free amino acids in the new shoots of tea plants, is closely correlated with tea quality³⁵. The main route for theanine biosynthesis progresses from glutamine to theanine³⁶, including catalytic enzymes such as glutamate synthase (GOGAT), glutamine synthetase (GS), arginine decarboxylase (ADC), glutamate dehydrogenase (GDH), and theanine synthase (Fig. 6a). Our analysis showed that a pair of GOGAT genes, a pair of GS genes, a pair of ADC genes, and a pair of GDH genes in tea plant are anchor pairs duplicated by the PPC-WGD (Fig. 6). Expressional analyses showed that the two copies of GOGAT had high expression, both in different organs and under different temperature treatments. Other duplicates had high expression in tea plant organs or under different temperature treatments (Fig. 6). Together, these results suggested that PPC-WGD probably contributed greatly to the development of theanine biosynthesis.

**Fig. 6: PPC-WGD contributed to the expansion of genes involved in the biosynthesis of theanine in tea plants.**

Caffeine (1,3,7-trimethylxanthine), a common ingredient found in tea, coffee, and cocoa, is an important flavor substance in tea that has many benefits for human health³⁷. The main steps of caffeine biosynthesis involve three methylation steps from xanthosine to caffeine^38,39. The tea plant pathway from xanthosine nucleosides to caffeine mainly depends on a continuum of three N-methyltransferases (NMTs), including xanthosine methyltransferase, 7-methylxanthine methyltransferase, and 3,7-dimethylxanthine methyltransferase (Fig. 7a). Our analyses showed that a pair of NMT genes in the tea plant was duplicated through the PPC-WGD (Fig. 7b). The expression profiles of the two duplicates showed relatively low expression in all tea plant organs, but this could be due to the specific spatial and temporal expression patterns or the induced expression of NMT genes under specific circumstances (Fig. 7b).

**Fig. 7: PPC-WGD contributed to the expansion of genes involved in the biosynthesis of caffeine in tea plants.**

We then compared the gene numbers from those pathways, i.e., the catechins, theanine, and caffeine biosynthesis pathways, Supplementary Table S1) in persimmon and rhododendron. Only a few or even no homologous genes were found in rhododendron and persimmon, Supplementary Table S3), showing that although rhododendron and persimmon shared this PPC-WGD with tea plant, the tea plant was better able to the genes that participate in these pathways.

— to www.nature.com