Journal Pre-proof

Comparative study of vegetative and reproductive growth of different tea varietiesresponse to different fluoride concentrations stress

Peidi Yang, Zhen Liu, Yang Zhao, Yang Cheng, Juan Li, Jing Ning, Yang Yang,Jian'an Huang

PII: S0981-9428(20)30270-9

DOI: https://doi.org/10.1016/j.plaphy.2020.05.038

Reference: PLAPHY 6218

To appear in: Plant Physiology and Biochemistry

Received Date: 24 February 2020

Revised Date: 28 May 2020

Accepted Date: 28 May 2020

Please cite this article as: P. Yang, Z. Liu, Y. Zhao, Y. Cheng, J. Li, J. Ning, Y. Yang, Jian'. Huang,Comparative study of vegetative and reproductive growth of different tea varieties response to differentfluoride concentrations stress, Plant Physiology et Biochemistry (2020), doi: https://doi.org/10.1016/j.plaphy.2020.05.038.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the additionof a cover page and metadata, and formatting for readability, but it is not yet the definitive version ofrecord. This version will undergo additional copyediting, typesetting and review before it is publishedin its final form, but we are providing this version to give early visibility of the article. Please note that,during the production process, errors may be discovered which could affect the content, and all legaldisclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Masson SAS.

Author Contributions Yang yang and Huang jianan designed the project. Yang peidi analyzed the data and wrote the manuscript. Yangpeidi, Liu zhen, Zhaoyang Chengyang Lijuan performed experiments. Yang peidi contributed samples, materials, and data.Li juan helped with the data analysis and examined the results.

Title: Comparative Study of Vegetative and Reproductive Growth of Different 1 Tea Varieties Response to Different Fluoride Concentrations Stress 2 3 Peidi Yang1,2, Zhen Liu2, Yang Zhao2, Yang Cheng1,2, Juan Li1,Jing Ning 1, Yang 4 Yang2*, Jian’an Huang1* 5 6 1, National Research Center of Engineering Technology for Utilization of Botanical 7 Functional Ingredients, Collaborative Innovation Center of Utilization of Functional 8 Ingredients from Botanicals, Key Lab of Tea Science of Education Ministry, Hunan 9 Agricultural University, Changsha, 410128; 10 2, Hunan Sub Center of National Tea Improvement Center, Tea Research Institute of 11 Hunan Academy of Agricultural Sciences, Changsha, 410125 12 13 *Correspondence should be addressed to E-mail: yangyangsir@126.com and 14 jian7513@sina.com 15 16 Abstract 17

Background：：：：The amount of fluoride accumulation in tea leaves was gradually 18 increase as the matures of tea plants, and the excessive fluoride intake can threaten 19 people's health. Based on years of field investigations, a low level of fluoride variety 20 Xiangbo Lǜ (XBL) and a high level of fluoride variety Zhenong 139 (ZN139) were 21 selected. 22

Results：：：：In this study, the root, 1st and the 5th leaf of the two-year-old tea trees were 23

used for morphological, physiological and comparative transcriptomics analysis to 24 understand the different features of “XBL” and “ZN139” under fluoride stress 25 conditions. The color of the 1st and 5st leaves of XBL were yellower, the activity of 26 peroxidase, catalase and antioxidant enzyme were lower than ZN139 under the 27 high-fluoride stress. Transcriptomics analysis indicated that core genes involved in 28 photosynthesis rates regulation showed no significantly exchanged expression, the 29 co-downregulation of magnesium ions transportation, while the ROS scavenging, 30 vegetative growth and self-compatibility between the two varieties were different. 31 Crucial genes’ expression were also identified by the real-time RT-PCR. 32 Conclusion: 33 The tea tree is one of the few plants that has a high-fluoride content, but the different 34 varieties respond differently to fluoride stress. High-fluoride tea tree varieties, such as 35 ZN139, have stronger ROS scavenging abilities through the use of both their 36 non-enzymatic and enzymatic antioxidant systems which act by increasing the 37 expression levels of inositol-1-monophosphatases and peroxidases, among others. 38 ZN139 can also compensate for the decrease in photosynthetic rate that is associated 39 with the ionic imbalance caused by the reduced consumption of light energy during 40 long-periods of high fluoride stress. Reproductive development was protected in 41 ZN139 by the up-regulated expression of S-locus glycoprotein, Mildew resistance 42 locus o and Phospholipase D under fluoride stress, while the vegetative development 43 of low-fluoride varieties such as XBL was retarded. More starch and cellulose were 44

redistributed to glucose by increasing the expression levels of glycosyl transferases 45 and hydrolases to provide more energy for processes involved in the response and 46 tolerance towards fluoride stress. 47 48 Keywords: Tea, Comparative transcriptomics analysis, Fluoride, Vegetative growth, 49 oxidation-reduction reaction. 50 51 52

Introduction 53

Tea tree (Camellia sinensis (L.) O. Kuntze) is one of the plants with the highest 54 fluoride content, and can absorb and accumulate fluoride through soil, water and air 55 [1]. In the same soil environment, the fluoride content of tea plants is 10-100 times 56 higher than other plants, but tea plants can still grow normally [2, 3]. Therefore, tea 57 plant is considered as a fluoride-collecting plant. But fluoride is not one of the 58 essential elements of tea plant, and high concentration of fluoride will cause a certain 59 degree of influence on the growth of tea plant and tea quality [4]. The fluoride content 60 of different organs was different, the mature old leaves was higher than that of young 61 leaves, followed by stems and roots [5]. Roots of the tea plant will transport the 62 fluoride into the leaves and stems when the accumulation of fluoride is small in the 63 whole plants. As the fluoride accumulation increases, the fluoride in the leaves begins 64 to migrate to other organs, but it does not inhibit the absorption of fluoride by the tea 65 plant. This unique absorption mode making tea plant leaves accumulate a large 66 amount of fluoride, the tea plant can maintain normal physiological conditions 67 without being poisoned by fluoride [6, 7]. 68 Tea plants under the natural growth environment can accumulate a considerable 69 amount of fluoride without poisoning, but the application of high concentration of 70 fluoride by external sources still retard the development of tea plants through 71 effecting the normal photosynthesis and cellular respiration, vegetative growth of 72 roots, leaves and terminal buds [8–11]. Recent studies discovered that tea plants can 73 improve their tolerance to fluoride by regulating the synthesis of secondary 74 metabolites (such as the lignin, flavonoids and catechins) [4, 12], antioxidant enzyme 75 activity (such as the Superoxide dismutase (SOD) and catalase (CAT)) [9] and ion 76 transport (such as the aluminum) [10]. The study of Jiaojiao Zhu et al identified that 77 free amino acids and the total catechins decreased significantly, while caffeine was 78 notably induced by exogenous fluoride (4-day treatment with 16mg/L fluoride 79 solution)[13]. Similarly, Qing-Sheng Li et al found that the fluoride uptake by tea 80 plant is highly associated with Al, Ca and anion channels [14–16]. However, up to 81 now, few studies have been conducted to analysis the different regulatory networks 82 for the in fluoride tolerance between different tea varieties. 83 Based on the comparative data of fluoride content in leaves of spring tea for five 84 consecutive years, “Xiang Bo Lǜ” (XBL) and “Zheng Nong139” (ZN139), which 85 respectively had the lowest and highest fluoride contents in the 28 common varieties 86

of tea plants. To further discover the different regulatory networks, roots, 1st leaf and 87 the 5th leaf of the XBL and ZN139 plants treated with the different concentration of 88 fluoride (0, 10, 50mg/L) were used for morphological analysis and transcriptome 89 analysis in this study. Based on the transcriptome analysis, both of the vegetative 90 growth and reproductive growth regulate network genes showed significant different 91 expression profiles. A hypothesis on the regulation of the growth and reproduction of 92 tea plants by external application of fluoride was proposed. 93 94 Results 95 The normal growth of tea plant was affected by external application of fluoride 96

To identify the most appropriate tea tree species for the morphological analysis and 97 transcriptome analysis, 28 common varieties of tea plants such as “Long Jing”, “Jin 98 Guanyin” and “Da Jianye”, included the “XBL” and “ZN139” et al. were selected to 99 detect the fluoride content in spring tea for five consecutive years (from the 2012 to 100 2016) (Supplemental Table S1). XBL showed the lowest standard deviation and the 101 mean content of fluoride (96.00 ±13.87 mg/kg) in the buds. The fluoride contents of 102 three varieties（Rucheng Baimao, Fuding Dabai and ZN139）were significantly higher 103 than that of XBL, and ZN139 had the highest mean content of fluoride in buds 104 (198.65±56.56 mg/kg). Further determination of the fluoride content in the different 105 seasons and tissues of XBL and ZN139 showed both of that content in summer tea 106 trees and the 5-leaf (327.33 mg/kg and 430 mg/kg) were the highest (Figure 1A), 107 while the content in roots showed no significant differences (stable at about 30mg/kg, 108 not shown in the figure 1A ). Based on these basic statistical results, the plants of 109 XBL and ZN139 were selected to cultivate under the different concentration of 110 fluoride (0, 10, 50mg/L). The development of roots under both of the normal and the 111 high concentration of fluoride showed no obvious differences compared with each 112 other, but the leaves (especially the 1st leaf) color of XBL were yellow than ZN139 113 under the 10 and 50mg/L (Figure 1B), that was just coincident with the relatively high 114 contents of chlorophyll a (Ch a) and chlorophyll b (Ch b) in the 1st leaf and 5th leaf of 115 ZN139 (Figure 1C). It is worth noting that, the amount of pigment in plants was 116 always responsible for the rate of photosynthesis, except the non-photochemical 117 quenching coefficient (qN) (Figure 1D, green), all of the maximum quantum yield of 118 photosystem II (Fv/Fm) (Figure 1D, red) and other parameters such as non-regulated 119 energy dissipation quantum yield Y(NO), the actual photochemical efficiency of PS II 120 in the light(Y(II)), photochemical quenching (qP), quantum yield Y (NPQ) et al. 121 showed no difference to each other under the stress of different fluoride content 122 (Supplemental Figure S1). Fluoride stress also decreased the content of catechins 123 (epigallocatechin (EGC), catechin (C), epicatechin (EC), epigallocatechin gallate 124 (EGCG) and epicatechin gallate (ECG) ) and improved the content of caffeine in the 125 1st and 5th leaf of XBL and ZN139 compared to the normal culture environments 126 (Supplemental Table S2). 127

Global transcriptome analysis 128 To identify the different regulate networks under the normal condition and fluoride 129 stress of different tea trees, different concentration of fluoride were used to culture the 130 plant of XBL and ZN139 for the transcriptome sequencing (Figure 2A) and DEGs 131 (log2 (relative expression level) ≥1 and≤-1, padj≤0.05) were identified. The DEGs 132 between these two different tea trees without the fluoride stress were compared firstly. 133 The root of ZN139 had the more specific DEGs than the 1st leaf and 5th leaf, and the 134 clustering analysis results of the relative expression levels of the overlapped genes of 135 these three different tissues showed the expression profile of DEGs in leaves were 136 similar to each other than the roots (Figure 2B). That conformed the tissue-specific 137 expression profiles and demonstrated the accuracy of the transcriptome analysis data. 138 Further GO enrichment analyses were performed with these specific DEGs (1037 in 139 1st leaf, 1163 in 5th leaf and 4165 in root). The up-expressed genes in the 1st and 5th 140 leaves showed the similar expression profiles and highly enriched in the cell 141 communication (GO:0007154), pollen-pistil interaction (GO:0009875), recognition of 142 pollen (GO:0048544) biological process (Figure 2C, red). The roots up-expressed 143 genes were enriched in the drug transmembrane transport biological process 144 (GO:0006855), the iron ion binding (GO:0043021) and oxidoreductase activity 145 (GO:0005506) function (Figure 2C, green), while the down-expressed genes were 146 enriched in the response to oxidative stress process (GO:0006979). At the same time, 147 further comparations were summarized in Figure 2D. Compared with the normal 148 condation, the root of XBL under the fluoride stress showed the most number of 149 DEGs compared to the normal condition, 4110 and 3788 genes were respectively 150 upregulated after the treatment of 10 mg/L and 50mg/L fluoride (2947 genes showed 151 the same trend, orange star), 4281 and 3858 genes were respectively repressed (3196 152 genes showed the same trend, orange star). And the 5th leaf of XBL after the treatment 153 of 10 and 50 ml/L fluoride showed more down-expressed genes than other tissues and 154 treatments, and had the most same expression trend genes at the same time (Figure 2D, 155 blue star). 156 157 5th leaf specific DEGs under the fluoride stress 158 Based on the calculate results of the DEGs in different tissues and treatments showed 159 in Figure 2D. XBL had the most DEGs in the 5th leaf, which showed significant lower 160 Cha content, than that of ZN139 under the fluoride stress (Figure 1C). Two venn 161 diagram showed in Figure 3A identify the specific DEGs (sDEG) between the 5th leaf 162 of XBL and ZN139 under the 10 or 50 mg/L compared with the 0 mg/L fluoride 163 respectively, and both GO enrichment analysis of these sDEG were similar to each 164 other (Figure 3A). Further Venn analysis of these sDEG in the 5th leaf of XBL 165 between the 10 or 50 mg/L compared with the 0 mg/L fluoride discovered amount of 166 same genes (6 co-up and 579 co-down expressed genes),The tentative GO analysis 167 were enriched in reproductive process and negative regulation of cellular amide 168 metabolic process and the interact graph of these biological process were linked 169

closely (Figure 3B). The next hierarchical classification of these genes primarily 170 involved in the negative regulation of translation (GO:0048544) or the recognition of 171 pollen (GO:0048544), such as the NB-ARC transcription factors and S-locus 172 glycoprotein. Expression profiles in XBL and ZN139 under the different treatment 173 were cluster showed in Figure 3C based on the RNA-seq results. Then each of the 174 three genes, which showed the of most significant down-expression in the 5th leaf of 175 XBL, were selected to perform the real-time RT-PCR testing (Figure 3D). 176 177 Root specific DEGs under the fluoride stress 178 The root of XBL showed the most DEGs compared with the leaf or the root of ZN139 179 under the fluoride stress (Figure 2D, blue stars). The preliminary GO analysis of the 180 1163 or 1085 specific up or down-expressed genes in the root of XBL with the 181 treatment of 10mg/L fluoride solution (Figure 2D) showed no significant enrichment. 182 Based on this, the Venn diagram the DEGs of the root in XBL and ZN139 under the 183 50mg/L compared with the 0 mg/L fluoride were showed in Figure 4A. Based on the 184 GO enrichment analysis, the up-expressed sDEG (Group a) of ZN139 were enriched 185 in the oxidative stress response (GO:0006979, such as the peroxidase (POD)), the 186 down-expressed sDEG (Group c) of XBL were enriched in the metabolic of glucan 187 (GO:0006073, such as the Mlo family) and the defense response (GO:0006952, such 188 as the cellulase), and the co-down expressed genes (Group b, such as the Mg2+ 189 transporters) played important roles in the cation transport (GO:0006812). Referenced 190 to the RNA-Seq results and GO enrichment assay, 3 POD (CSA020762, novel.9069, 191 CSA027568), 3 Mg2+ transporters (novel.69252, novel.29417, novel.34030), 3 Mlo 192 family genes (novel.43039, novel.47207, CSA032247) and 3 cellulose synthase 193 (novel.58702, novel.47564, novel.58609), which had significantly altered expressions 194 after the treatment of fluoride (Figure 4B), were selected to identify their expression 195 levels in the root of XBL and ZN139 after the treatment of 0, 10 or 50mg/L fluoride 196 solution by the RT-PCR (Figure 4C). 197 198 Fluoride changes the expression of self-incompatibility and glycolytic pathway 199 genes 200 Tea tree is a kind of plants with the self-incompatibility (SI). GO analysis in Figure 201 3C&4B showed DEGs were highly enriched in the reproduction process and the 202 pollen-pistil interaction. Genes played roles in the SI process, such as the S-locus 203 glycoprotein (SLG) [17] or Mildew resistance locus o (Mlo) family [18] showed 204 significant high expression in ZN139 than XBL after the treatment of fluoride in the 205 5th leaf or root respectively. Considering amounts of SI process genes were highly 206 expressed in the tender leaves, and the 1st leaf of ZN139 had more DEGs after the 207 treatment of fluoride than that of XBL (Figure 2D), genes belongs to the SI were 208 collected to compare their expression profiles in the 1st leaf of ZN139 and XBL after 209 the treatment of fluoride. The expression of abundant of SLG, Mlo and Phospholipase 210 D (PLDα-1) [19] were up-regulated in ZN139 to improve the pollen germination and 211

pollen tube elongation (Figure 5A). Another significant different change was 212 observed in the carbohydrate metabolism, especially the glycolytic pathways in the 5th 213 leaf (Figure 5B). A amount of glycosyl transferases (such as CSA009048 and 214 CAS0032715) and glycosyl hydrolases (such as CSA005984 and CSA033012) were 215 significantly up-expressed, while two hexokinases (CSA006152 and novel.72752) 216 were repressed in XBL under the treatment of 10 or 50 mg/L fluoride compared with 217 that of ZN139 (Figure 5B), which indicated XBL accumulate less starch (transfered to 218 the glucose) than ZN139 under the fluoride stress. 219 220 Discussion 221 Effects of high concentration of fluoride on the photosynthesis of tea plants 222 Tea trees was one of the woody plants with the ability of collecting fluoride from the 223 growing environment. Tea trees could grow normally on soils with high fluoride 224 content and showed high tolerance to fluoride, but the treatment with high 225 concentration of fluoride solution will affect the development of tea trees. ZN139 and 226 XBL were took as the experimental material to study the effects of different 227 concentrations of fluoride solution on the growth by hydroponic experiments [20]. 228 The 1st or 5th leaf color of XBL was yellow than that of ZN139 under both of the 10 229 and 50 mg/L treatment, and the 5th leaf color of ZN139 was relative green and grown 230 well (Figure 1B). That was just consistent with the content trend of pigments (Figure 231 1C). All these evidences showed that the high content of fluoride affected on the color 232 of leaves (especially the 1st leaf) via reducing the content of pigments. 233 But it is worth noting that qP, F0 and Fv/Fm were similar between ZN139 and XBL 234 under the different fluoride stress (Figure 1D&S1). At the same time, GO or KEGG 235 analysis of RNA-Seq DEGs between the ZN139 and XBL suffer the same fluoride 236 stress did not enriched in the photosynthesis. 237 The expression of 28 genes (including electron transporters, photosystem II oxygen 238 evolving complex members and photosystem I reaction center members ), which 239 played roles in photosynthesis, were detected in the RNA-Seq data, while only 240 CSA036179 (photosynthetic electron transporter in photosystem II) was repressed in 241 the 5th leaf of XBL (1.09-fold) after a 10-day treatment with 50mg/L fluoride 242 solutions. This contrasted with previous research that found fluoride stress can inhibit 243 physiological and biochemical processes by disrupting photosynthesis [9]. The 244 long-term treatment (10-day) and the high fluoride stress (top to 50mg/L) in this study 245 may be two of the reasons for the inconsistencies with the results of previous studies 246 (1-4 days, 16mg/L)[13, 14]. Tea trees under high stress for long durations may be able 247 to alleviate the effect of fluoride stress on their photosynthetic rate through other 248 mechanisms. The photosynthetic activity (based on qP) of these two tea trees showed 249 slight difference under the same planting environment. But the ZN139 had lower 250 non-photochemical quenching (qN), which is the parameter of light protection 251 (consumed the light energy for heat) and had relationship with the dynamics of 252 thylakoid membrane [16, 17]. And the treatment of 10 or 50mg/L fluoride did not 253

cause the change of trend, which was same with 0 mg/L. Traditional research believes 254 that the intake of fluoride will affect photosynthesis, mainly because fluoride can 255 combine with magnesium ions (Mg2+) in chloroplasts, which caused the distortion of 256 thylakoids [9, 10]. Based on the RNA-Seq and real-time RT-PCR results, the 257 expression of Mg2+ transporters in both of the XBL and ZN139 were repressed with 258 the increasing fluoride concentration, and genes such as the novel.69252 in ZN139 259 was down-regulated more significantly (Figure 4&C). AtMGT10 (AT5G22830), the 260 ortholog of novel.69252 in Arabidopsis, was essential for chloroplast development 261 and highly expressed on the member of chloroplast, and the efficiency of 262 photosynthesis was reduced in part of its mutants [23, 24]. Nagata et.al found the 263 presence of aluminum-fluoride complexes (the form of fluoride transported in the 264 woody part) in mature leaves of tea trees [25]. The content of fluoride in leaves and 265 roots of ZN139 was significantly higher than that of XBL, indicating that the content 266 of aluminum ions is higher, and decades studies showed that the Mg2+ transport was 267 inhibited by Al3+ [26]. Therefore, it is speculated that ZN139 makes up the decrease in 268 photosynthetic rate caused by the ionic imbalance via consuming less light energy for 269 heat. 270 271 Differences in antioxidant properties 272 Catechins were the main component (70%) of the tea polyphenols [27], and these 273 have been closely linked to the taste and quality of tea and have been studied 274 vigorously in regards to their antioxidant activity [28]. The statistical results of the 275 amount of each individual catechins induced by fluoride stress, shown in the 276 Supplemental Table S2, was consistent with the prior studies [13]. Certain genes 277 played important roles in the biosynthesis of catechins, for example, two flavone 278 synthases (FNS, novel.11716 and CSA001198) and one chalcone isomerase (CHI, 279 CSA024690) were highly induced by the fluoride stress in both of XBL and ZN139, 280 and the same trend in expression was also identified in the research of Jiaojiao Zhu et 281 al [13]. Disruption of active oxygen balance was a common form of plants response 282 to the abiotic stress, and the appearance of reactive oxygen species (ROS) may cause 283 irreversible damage to the development of plants [29]. Plants could eliminate 284 excessive reactive oxygen species within a certain concentration of fluoride threshold 285 by regulating its own metabolism [30, 31]. In our study, POD (such as the 286 CSA020762 and other 6 genes) were specifically up-expressed in the ZN139, 287 especially under the 50 mg/L fluoride stress (Figure 4C), that just coincident with the 288 determination of POD enzyme activity showed in the Supplemental Figure S2. And 289 the activity of superoxide dismutase (SOD) and catalase (CAT) showed the same 290 trend which demonstrated that the root of ZN139 had stronger ability in ROS 291 scavenging than XBL. But these three enzymes have different trends under different 292 fluoride stress and in different tissues (Supplemental Figure S2). The POD enzymes 293 were induced by the fluoride stress clearly, especially in the root of ZN139 suffering 294 50mg/L fluoride stress. The leaf of XBL (especially in the 1st leaf) had relative higher 295

POD activity, and activity intensity was higher when treated with 10mg/L than when 296 treated with 50mg/L fluoride solution. However, in the absence of fluoride stress, 297 ZN139 showed higher SOD enzyme activity (up to 230U/g) than XBL. The activity 298 of CAT were similar to each other in these two tree plants without the fluoride stress, 299 the root of XBL had higher CAT activity suffering the 10mg/L fluoride stress. When 300 treated with 50mg/L fluoride solution, the CAT enzyme activities in the leaves and 301 roots of ZN139 were significantly higher than that of XBL. Taken together, XBL and 302 ZN139 was different in the antioxidant properties. The activity of SOD was naturally 303 occurring without the fluoride stress in ZN139. Low concentration of fluoride 304 treatment induced the exchange of ROS scavenging ability in XBL than ZN139. 305 Nevertheless, the high activity of POD in the root and CAT, SOD in the whole plants 306 of ZN139 than XBL under high concentration of fluoride treatment is crucial to 307 reduce the concentration of ROS caused by fluoride stress. Non-enzymatic 308 antioxidant systems also provide other crucial responses for antioxidation protection. 309 Previous research in maize, rice and other tea tree species has shown that the 310 non-enzymatic antioxidant content of tea changed with ROS accumulation, for 311 example, there was a reduction in glutathione (GSH) content and an increase in 312 Ascorbic Acid (AsA) content [32, 33]. Based on the RNA-Seq data, the expression of 313 3 hydroxyacylglutathione hydrolases (CSA018082, novel.69111, CSA000172) showed 314 significantly higher expression after the treatment with the fluoride solution than in 315 the 5th leaf of XBL in the 5th leaf (4.03-fold, 1.00-fold and 15.13-fold). At the same 316 time, nine inositol-1-monophosphatase (GPPs, key enzymes for AsA biosynthesis) 317 were expressed in the leaves or roots, and 8 of these were highly expressed in the 5th 318 leaf and roots of ZN139 after treatment with the 50 mg/L fluoride solution 319 (Supplemental Table S3). All of these expression profiles coincided with a reduction 320 in GSH content and the accumulation of AsA which corroborates previous research 321 on tea trees under the fluoride stress. This conclusion also just explained the 322 phenomenon that ZN139 grows better than XBL and has higher fluoride content in 323 leaves under the high fluoride environment. 324 325 Effect of fluoride stress on self-affinity 326 Self-incompatibility may be controlled either by the sporophyte or by the 327 gametophyte [34], and tea tree is the later one. Previous researches studied well on the 328 Arabidopsis and other Brassica plants, but short on the tea tree. The incompatibility 329 for effective pollination is determined by the homology between secreted 330 glycoproteins (SLG)/membrane-bound receptor protein kinases (SRK) and S-locus 331 protein 11 (SP11)/S-locus cysteine-rich protein (SCR) [17, 34–36]. SRK could 332 interact with the M-locus protein kinase (MPLK) or regulated the expression of arm 333 repeat-containing 1 (ARC1) to affect the pollen germination and the elongation of 334 pollen tubes [37, 38]. ARC1 is the up-stream regulator of Phospholipase D (PLDα-1) 335 and glyoxalase (GLO1), and setting rate of their RNA-interfering materials decreased 336 obviously [19, 39]. There were also the mildew resistance locus o (Mlo) clade III 337

family genes (such as the AtMLO5/7/8/9/10 in Arabidopsis), which played important 338 roles in the defense response and the pollen tube reception, are highly expressed in the 339 roots, young leaves and flowers [40]. Based on the GO enrichment, RNA-seq and 340 RT-PCR analysis results, genes played roles in the SI were highly specific expressed 341 in the 1st and 5th leaf of ZN139 under the normal condition (Figure 2C), and were 342 repressed in the leaf or root of XBL under the fluoride stress, such as the SLG (Figure 343 3C&D), Mlo (Figure 4A&B) and PLDα-1 (Figure 5A) to affect the pollen-pistil 344 interaction, pollen germination and tube elongation. Based on these evidences, it can 345 be shown that, unlike the self-incompatibility of XBL which is significantly increased 346 by high fluoride pressure, ZN139 can maintain a more normal self-pollination under 347 both of the normal and high fluoride pressure environment to ensure its own breeding 348 needs. The main propagation of tea trees comprises two main methods: the vegetative 349 (cutting, grafting, and tissue culture) and the generative (seeds) methods [41]. This 350 finding reminds us that ZN139 can be seeded on the high-fluoride soils, while the 351 vegetative methods are more suitable for the non-fluorinated tea trees such as XBL in 352 this study. Further research will combine the group evolution analysis [42] with 353 resequencing data to cluster varieties with leaves contained fluoride content in the 354 same season. The selective elimination analysis (such as based on the Fst) and 355 transcriptome differential analysis will be use to further explore and identify the key 356 genes in the regulation of tea plant reproduction development. 357 358 Effects of fluoride stress on vegetative growth 359 Redistribution of carbohydrates retarded the vegetative growth was a common 360 response mode for plants to abiotic stress [43]. Based on the GO cluster showed in 361 Figure 2C, there was no DEGs enriched in the plant growth, development or the 362 glycometabolism without the fluoride stress. While this situation changed obviously 363 after the treatment of fluoride: novel.23647 and other genes with the NB-ARC domain 364 were repressed in the 5th leaf of XBL obviously (Figure 3C); novel.58702 and other 365 cellulases were repressed in the root of XBL; glycosyl transferases (such as 366 CSA009048) and glycosyl hydrolases (such as NOVEL.25420) were induced 367 expressed by the 10mg/L and 50mg/L fluoride solution treatment (Figure 5B). 368 Disease resistance protein RPS family genes were the orthologs of novel.23647 in 369 Arabidopsis with the NB-ARC domain [44], which played important roles in the 370 defense resistance and cell death, and the mutant of part of RPSs showed the dwarfing 371 plants or growth slowly [45, 46]. Cellulose synthase was crucial for the cell wall 372 formation, such as the CesA7 in Arabidopsis or rice, whose repression could retarded 373 the vegetative growth via negative effecting the cellulose biosynthesis [38, 39]. 374 UDP-glucose is the primary substrate for cellulose synthase [49], and the content of it 375 was dynamically related to the content of glucose and starch in the glycolysis pathway 376 [50] (Figure 5B). Glycosyl hydrolase is the rate-limited enzyme from cellulose to 377 D-glucose, glycosyl transferase is contribute to the utilization of starch glycogen and 378 the biosynthesis of glucuronoxylan and D-glucose [51, 52]. Based on the RNA-Seq 379

results, both of the glycosyl hydrolases and transferases were induced specific 380 induced in the 5th leaf of XBL. And two hexokinases (such as the CSA006152 and 381 novel.72752) were repressed in XBL to reduce the consumption of D-glucose [53]. 382 These studies indicated that carbohydrates were tend to convert into the D-glucose to 383 generate more energy for the stress response (such as the active transport of ions, 384 production of ATP, NADPH and heat energy (qN)) [54], thereby retarding the 385 vegetative growth of plants via reducing starch accumulation and cellulose synthesis. 386 In addition, increasing the ATP or its analogues were proved to inhibited pollen 387 germination and pollen tube elongation [55, 56], that was just consistent with the high 388 self-incompatibility of XBL under the fluoride stress compared with ZN139. 389 Determining the starch, cellulose, and soluble sugars content in the leaves and roots of 390 tea trees after different periods after fluoride stress will be the focus of future studies, 391 to further explore the core regulatory genes in the fluoride stress response. 392 393 394 Conclusions 395 The tea tree is one of the few plants that has a high-fluoride content, but the different 396 varieties respond differently to fluoride stress. High-fluoride tea tree varieties, such as 397 ZN139, have stronger ROS scavenging abilities through the use of both their 398 non-enzymatic and enzymatic antioxidant systems which act by increasing the 399 expression levels of inositol-1-monophosphatases and peroxidases, among others. 400 ZN139 can also compensate for the decrease in photosynthetic rate that is associated 401 with the ionic imbalance caused by the reduced consumption of light energy during 402 long-periods of high fluoride stress. Reproductive development was protected in 403 ZN139 by the up-regulated expression of S-locus glycoprotein, Mildew resistance 404 locus o and Phospholipase D under fluoride stress, while the vegetative development 405 of low-fluoride varieties such as XBL was retarded. More starch and cellulose were 406 redistributed to glucose by increasing the expression levels of glycosyl transferases 407 and hydrolases to provide more energy for processes involved in the response and 408 tolerance towards fluoride stress. 409 410 Materials and Methods 411 412 Plant materials 413 Plants were grown on the 20°C under nature day condition. For RNA extraction and 414 content measurement, fresh tissues of the root, the 1st and the 5th leaf after the 415 treatment of 0mg/L, 10 mg/L and 50 mg/L Fluoride solution (NaF) were collected. 416 417 Determination of fluoride content 418 0.15g of tea powder of leaves or roots (accurate to 0.0001g), add 20ml water, boil for 419 30min then remove and cool. Add 25ml total ion strength adjustment buffer then 420 mixed with water to 50ml. Determination of fluoride content was performed by the 421

microwave digestion and fluoride ion-selective electrode method [57]. Three 422 biological replicates, 10 plants for each repeat. Significant differences between the 423 fluoride content in XBL and other varieties were evaluated using the paired Student’s 424 t test ( P <0.01 (**) and P <0.05 (*)). 425 426 Determination of CAT, SOD, POD enzyme activity 427 The activity of CAT were determined according to the method of Johansson [58]. To 428 standardize enzyme activity of SOD, two procedures based on pyrogallol autoxidation 429 called 325 nm method and 420 nm method are compared according to the method of 430 Jianrong [59]. The activity of POD were determined according to the method of 431 Vetter[60]. All bars represent means± s.d., three biological replicates (5 plants for 432 each repeat). Significant differences (Student’s t test, P <0.01 (**) and P <0.05 (*)). 433 434 435 Determination of Photosynthetic Pigment, Chlorophyll Fluorescence, Catechins 436 and caffeine 437 Pigments was extracted with 95% alcohol under dark conditions, and the contents of 438 Ch a, Ch b, and Car were quantified spectrophotometrically, according to the method 439 of Wellburn [61]. The determination of F0 and Fm, Y(II), Y(NPQ), Y(NO), NPQ, qN, 440 qP, qL and ETR was executed with a pulse modulation chlorophyll fluorescence 441 spectrometer (LI-COR Inc., Lincoln, NE, United States). Plants were dark-adapted for 442 20 min before measurement. All bars represent means± s.d., three biological 443 replicates (5 plants for each repeat). Significant differences (Student’s t test, P <0.01 444 (**) and P <0.05 (*)). The determination of catechins and caffeine in green tea using 445 HPLC were performed according to the method of Wang [13, 62]. Three biological 446 replicates, 10 plants for each repeat. Significant differences between the content of 447 catechins and caffeine after the treatment of the 0mg/L fluoride solution and 10mg/L 448 or 50mg/L fluoride solution were evaluated using the paired Student’s t test ( P <0.01 449 (**) and P <0.05 (*)). 450 451 Transcriptome analysis 452 All samples were sent to Novogene (Beijing, China) to perform extraction, 453 sequencing and bioinformatics analyses. mRNA was purified from total RNA using 454 poly-T oligo-attached magnetic beads. Fragmentation was carried out using divalent 455 cations under elevated temperature in NEBNext First Strand Synthesis Reaction 456 Buffer(5X). First strand cDNA was synthesized using random hexamer primer and 457 M-MuLV Reverse Transcriptase (RNase H-). Second strand cDNA synthesis was 458 subsequently performed using DNA Polymerase I and RNase H. Remaining 459 overhangs were converted into blunt ends via exonuclease/polymerase activities. 460 After adenylation of 3’ ends of DNA fragments, NEBNext Adaptor with hairpin loop 461 structure were ligated to prepare for hybridization. In order to select cDNA fragments 462 of preferentially 250~300 bp in length, the library fragments were purified with 463

AMPure XP system (Beckman Coulter, Beverly, USA). Then 3 μl USER Enzyme 464 (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37°C for 15 m in 465 followed by 5 min at 95 °Cbefore PCR. Then PCR was performed with Phusion High 466 -Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer. At last, 467 PCR products were purified (AMPure XP system) and library quality was assessed on 468 the Agilent Bioanalyzer 2100 system. 469 470 Differential expression analysis 471 Different concentration of fluoride (0, 10, 50 mg/L) were used to culture the plant of 472 XBL and ZN139 for 240 hours, then the roots, 1st and 5th leaf were extracted for the 473 transcriptome sequencing. Differential expression analysis of 18 groups (three 474 biological replicates per condition) was performed using the DESeq2 R package 475 (1.16.1). Varieties-specific expression genes without fluoride stress (XBL vs ZN139 476 root,1st or 5th leaf under the 0mg/L ;DEGs induced by low and high fluoride stress in 477 XBL (XBL root 50mg vs 0mg,10mg vs 0mg; XBL 1st leaf 50mg vs 0mg,10mg vs 478 0mg; XBL 5th leaf 50mg vs 0mg,10mg vs 0mg); DEGs induced by low and high 479 fluoride stress in ZN139 (ZN139 root 50mg vs 0mg,10mg vs 0mg; ZN139 1st leaf 480 50mg vs 0mg,10mg vs 0mg; ZN139 5th leaf 50mg vs 0mg,10mg vs 0mg) et al were 481 compared in this study. DESeq2 provide statistical routines for determining 482 differential expression in digital gene expression data using a model based on the 483 negative binomial distribution. The resulting P-values were adjusted using the 484 Benjamini and Hochberg’s approach for controlling the false discovery rate. Genes 485 with an adjusted P-value ≤0.05 found by DESeq2 were assigned as differentially 486 expressed. The selection of different expression genes was based on the FDR≤0.001 487 and absolute value of Log2 (Relative expression level, compared with WT) ≥1 and 488 ≤-1. 489 490 GO and KEGG enrichment analysis of differentially expressed genes 491 Case special deep analysis based on DEGs including GO and KEGG pathway 492 enrichment analysis were done by our group. Gene Ontology (GO) enrichment 493 analysis of differentially expressed genes was implemented by the cluster Profiler R 494 package, in which gene length bias was corrected. GO terms with corrected P-value 495 less than 0.05 were considered significantly enriched by differential expressed genes. 496 KEGG is a database resource for understanding high-level functions and utilities of 497 the biological system, such as the cell, the organism and the ecosystem, from 498 molecular-level information, especially large-scale molecular datasets generated by 499 genome sequencing and other high-through put experimental technologies 500 (http://www.genome.jp/kegg/). We used cluster Profiler R package to test the 501 statistical enrichment of differential expression genes in KEGG pathways. 502 503 Plant RNA extraction and real-time RT-PCR 504

The total RNA was extracted by Trizol reagent (TAKARA, Japan). An amount of 500 505 ng total RNA was used to inverse transcription with Transcriptor First Strand cDNA 506 Synthesis Kit (TAKARA, Japan). Diluted the cDNA 10 times and then as the 507 template for real-time RT-PCR. RT-PCR reaction system: SYBR® Premix Ex 508 TaqTM (2×) 5 μl, PCR Forward Primer (10 μM) 0.2 μl, PCR Reverse Primer (10 μM) 509 0.2 μl, cDNA template 1μl, with distilled deionized water up to 10 μl. The PCR 510 process as following conditions: 95°C for 15s, 56°C for 30s, and 72°C for 30 s, 40 511 cycles. All bars represent means± s.d., three biological replicates (5 plants for each 512 repeat). Significant differences (Student’s t test, P <0.01 (**) and P <0.05 (*)). 513 Primers used for the real-time RT-PCR were listed in the Supplemental Table S4. 514 515 Figure legends 516 Figure 1. Phenotypes and the determination of photosynthesis index of XBL and 517 ZN139 under the fluoride stress. 518 A, The fluoride content of the terminal bud and 1ST leaf, 4th and 5th leaf of XBL and 519 ZN139 in spring, summer and autumn. B, Phenotypes of the 1ST, 5th leaf and root of 520 XBL and ZN139 under the different concentration of fluoride solution (0, 10 and 50 521 mg/L). C, Contents of chlorophyll a and chlorophyll b and carotenoids in the 1st, 3rd 522 and 5th leaf of XBL and ZN139 under the different concentration of fluoride solution 523 (0, 10 and 50 mg/L). D, F0, Fv/Fm and qN in the tea trees used in D. Fv, maximal 524 fluorescence; Fm, variable fluorescence; qN, non-photochemical quenching 525 coefficient. All bars represent means± s.d., three biological replicates, 5 plants for 526 each repeat for C&D, 10 plants for each repeat for A. Significant differences 527 (Student’s t test, P <0.01 (**) and P <0.05 (*)). 528 529 Figure 2 Overall differentially expressed genes (DEGs) between the XBL and 530 ZN139 under the normal condition 531 A, Schematic diagram for the material selection of detail transcriptome sequencing. B, 532 Venn diagram of the overlap between the DEGs in root, 1st and 5th leaf, and the 533 clustering analysis results of the relative expression levels of the different groups. C, 534 GO enrichment analysis results of the tissue-specific expression gene. Red for the 535 biological process and green for the molecular function. D, The histogram of the 536 statistical results for the DEGs in root, 1st and 5th leaf between the XBL and ZN139 537 without the fluoride stress (Blue: specific DEGs after the treatment of 10mg/L 538 fluoride; Gray: specific DEGs after the treatment of 50mg/L fluoride; Orange: DEGs 539 had the same expression trends after the treatment of 10 and 50mg/L compared to the 540 normal condition). 541 542 Figure 3 XBL 5th leaf specific DEGs analysis 543 A, Venn diagram of the DEGs of the 5th leaf of XBL and ZN139 after the treatment of 544 10 mg/L or 50 mg/L fluoride solution compared with the 0 mg/L, with the GO 545 enrichment analysis clustering results of the specific DEGs of XBL (up-expressed 546

(read) and down-expressed (green)). B, Venn diagram of the specific DEGs of the 5th 547 leaf of XBL after the treatment of 10mg/L or 50 mg/L fluoride solution compared to 548 the 0 mg/L, with GO enrichment and interactive graph analysis results of the 6 549 co-down and 579 co-up expressed genes. C, The clustering analysis results of genes, 550 which belonged to the translation and recognition of pollen biological process, 551 relative expression levels of the different groups based on the results of RNA-seq 552 analysis. D, Expression profiles of genes (Real-time RT-PCR results) selected in C 553 (green points). All bars represent means± s.d., three biological replicates. Significant 554 differences (Student’s t test, P <0.01 (**) and P <0.05 (*)). 555 556 Figure 4 XBL root specific DEGs analysis 557 A, Venn diagram of the overlap and the GO enrichment analysis results of the DEGs 558 between the root of XBL and ZN139 after the treatment of 50mg/L compared with the 559 0mg/L fluoride. B, DEGs involved in the oxidative stress (Group a), cation transport 560 (Group b), the metabolic of glucan and the defense response (Group C). The absolute 561 values (RNA-seq) of log2 (50mg/L/0mg/L) ≥1 and FDR < 0.001 were used as the 562 criteria for DEGs. The color of the box represents up (red) and down 563 (green)-regulated genes. C, Expression profiles of 15 genes (Real-time RT-PCR 564 results) in the root of XBL or ZN139 after the treatment of 50 and 0 ml/L fluoride. All 565 of the expressions were normalized by the expression level of Actin in XBL without 566 the treatment of Fluoride. All bars represent means± s.d., three biological replicates. 567 Significant differences (Student’s t test, P <0.01 (**) and P <0.05 (*)). 568 569 Figure 5 Differential expression of representative genes involved in 570 self-incompatibility and carbohydrate metabolism process. 571 A, DEGs in the self-incompatibility process. B, DEGs in carbohydrate metabolism. 572 The absolute values of log2 ≥1 and FDR < 0.001 were used as the criteria for DEGs. 573 The color of the box represents up (red) and down (green)-regulated genes and the 574 value in the box is the log2 of the genes in the 1st and 5th leaf of ZN139 and XBL 575 under the treatment of 10 or 50 mg/l compared with the 0mg/L fluoride. 576 Figure 6 The proposed model for the fluoride stress on the development of tree 577 plants. 578 579 580 Supplemental data 581 Supplemental Figure S1 Y(II), Y(NPQ), Y(NO), NPQ, qN, qP, qL and ETR in the 582 1st, 3rd and 5th leaf of XBL and ZN139 under the different concentration of fluoride 583 solution (0, 10 and 50 mg/L) . All bars represent means± s.d., three biological 584 replicates (10 plants for each replicate). Significant differences (Student’s t test, P 585 <0.01 (**) and P <0.05 (*)). 586

Supplemental Figure S2 The SOD, CAT and POD enzyme activity determination 587 results. All bars represent means± s.d., three biological replicates (5 plants for each 588 replicate). Significant differences (Student’s t test, P <0.01 (**) and P <0.05 (*)). 589 Supplemental Table S1 Fluoride content of 28 tea tree varieties in spring for 5 years. 590 The 1st and the 5th leaves were selected for the determination of fluoride content, each 591 assay was designed with three biological replicates, 10 plants for each biological 592 replicate. 593 Supplemental Table S2 The amount of individual catechins and caffeine induced by 594 fluoride stress. Each assay was designed with three biological replicates, 10 plants for 595 each biological replicate, labeled with a letter are significantly different at P < 0.05 by 596 Duncan’s test. 597 Supplemental Table S3 The expression profiles of GPP. 598 Supplemental Table S4 Primers used for the real-time RT-PCR assay. 599 600 Acknowledgements 601

This study was supported by the Hunan Natural Science Foundation(2018JJ3266)，602

National Natural Science Foundation, joint fund for regional innovation and 603 development (U19A2030), Key Research and Development Projects of Hunan 604 Province (2018NK2031), National Modern Agricultural Technology System 605 Construction Project (CARS-19). 606 607 Author Contributions 608 Yang yang and Huang jianan designed the project. Yang peidi analyzed the data and 609 wrote the manuscript. Yangpeidi, Liu zhen, Zhaoyang, Chengyang ,Ning Jing and 610 Lijuan performed experiments. Yang peidi contributed samples, materials, and data.Li 611 juan helped with the data analysis and examined the results. 612 613 References 614 1. Shu WS, Zhang ZQ, Lan CY, Wong MH. Fluoride and aluminium concentrations 615 of tea plants and tea products from Sichuan Province, PR China. Chemosphere. 2003. 616 2. Xie ZM, Ye ZH, Wong MH. Distribution characteristics of fluoride and aluminum 617 in soil profiles of an abandoned tea plantation and their uptake by six woody species. 618 Environ Int. 2001. 619 3. Fornasiero RB. Phytotoxic effects of fluorides. Plant Sci. 2001. 620 4. Li C, Ni D. Effect of fluoride on chemical constituents of tea leaves. Fluoride. 621 2009. 622 5. Ruan J, Wong MH. Accumulation of fluoride and aluminium related to different 623 varieties of tea plant. Environ Geochem Health. 2001. 624 6. Arnesen AKM. Availability of fluoride to plants grown in contaminated soils. Plant 625 Soil. 1997. 626

7. Zhang L, Li Q, Ma L, Ruan J. Characterization of fluoride uptake by roots of tea 627 plants (Camellia sinensis (L.) O. Kuntze). Plant Soil. 2013. 628 8. Baroni Fornasiero R. Fluorides effects on Hypericum perforatum plants: First field 629 observations. Plant Sci. 2003. 630 9. Li C, Zheng Y, Zhou J, Xu J, Ni D. Changes of leaf antioxidant system, 631 photosynthesis and ultrastructure in tea plant under the stress of fluorine. Biol Plant. 632 2011. 633 10. Fluorides in the environment: effects on plants and animals. 2004. 634 11. Wang LX, H. TJ, B. X, J. YY, J. L. Variation of Photosynthesis, Fatty Acid 635 Composition, ATPase and Acid Phosphatase Activities, and Anatomical Structure of 636 Two Tea (Camellia sinensis (L.) O. Kuntze) Cultivars in Response to Fluoride. 637 ScientificWorldJournal. 2013:1–9. 638 12. Wang YS, Gao LP, Shan Y, Liu YJ, Tian YW, Xia T. Influence of shade on 639 flavonoid biosynthesis in tea (Camellia sinensis (L.) O. Kuntze). Sci Hortic 640 (Amsterdam). 2012. 641 13. Zhu J, Pan J, Nong S, Ma Y, Xing A, Zhu X, et al. Transcriptome analysis reveals 642 the mechanism of fluoride treatment affecting biochemical components in Camellia 643 sinensis. Int J Mol Sci. 2019. 644 14. Li QS, Li XM, Qiao RY, Shen EH, Lin XM, Lu JL, et al. Data descriptor: De 645 novo transcriptome assembly of fluorine accumulator tea plant camellia sinensis with 646 fluoride treatments. Sci Data. 2018. 647 15. Ruan J, Ma L, Shi Y, Han W. Uptake of fluoride by tea plant (Camellia sinensis L) 648 and the impact of aluminium. J Sci Food Agric. 2003. 649 16. Zhang XC, Gao HJ, Wu HH, Yang TY, Zhang ZZ, Mao JD, et al. Ca2+ and CaM 650 are involved in Al3+ pretreatment-promoted fluoride accumulation in tea plants 651 (Camellia sinesis L.). Plant Physiol Biochem. 2015. 652 17. Iwano M, Takayama S. Self/non-self discrimination in angiosperm 653 self-incompatibility. 15:78–83. 654 18. Jones DS, Yuan J, Smith BE, Willoughby AC, Kumimoto EL, Kessler SA. 655 MILDEW RESISTANCE LOCUS O Function in Pollen Tube Reception Is Linked to 656 Its Oligomerization and Subcellular Distribution. 2017;:pp.00523.2017. 657 19. Scandola S, Samuel MA. A Flower-Specific Phospholipase D1 is a Stigmatic 658 Compatibility Factor Targeted by the Self-Incompatibility Response in Brassica 659 Napus. Soc Sci Electron Publ. 2018. 660 20. Lin ZH, Chen LS, Chen RB, Zhang FZ, Jiang HX, Tang N, et al. Root release and 661 metabolism of organic acids in tea plants in response to phosphorus supply. J Plant 662 Physiol. 2011. 663 21. Bilger W, Björkman O. Role of the xanthophyll cycle in photoprotection 664 elucidated by measurements of light-induced absorbance changes, fluorescence and 665 photosynthesis in leaves of Hedera canariensis. Photosynth Res. 1990. 666

22. Schreiber U, Bilger W, Neubauer C. Chlorophyll Fluorescence as a Nonintrusive 667 Indicator for Rapid Assessment of In Vivo Photosynthesis. In: Ecophysiology of 668 Photosynthesis. 1995. 669 23. Li L, Tutone AF, Drummond RSM, Gardner RC, Luan S. A novel family of 670 magnesium transport genes in Arabidopsis. Plant Cell. 2001. 671 24. Sun Y, Yang R, Li L, Huang J. The Magnesium Transporter MGT10 Is Essential 672 for Chloroplast Development and Photosynthesis in Arabidopsis thaliana. Molecular 673 Plant. 2017. 674 25. Nagata T, Hayatsu M. Aluminium kinetics in the tea plant using (27)Al and (19)F 675 NMR. 1993;32:771–5. 676 26. Ishijima S, Uda M, Hirata T, Shibata M, Kitagawa N, Sagami I. Magnesium 677 uptake of Arabidopsis transporters, AtMRS2-10 and AtMRS2-11, expressed in 678 Escherichia coli mutants: Complementation and growth inhibition by aluminum. 679 Biochim Biophys Acta - Biomembr. 2015. 680 27. Liu M, Tian HL, Wu JH, Cang RR, Wang RX, Qi XH, et al. Relationship between 681 gene expression and the accumulation of catechin during spring and autumn in tea 682 plants (Camellia sinensis L.). Hortic Res. 2015. 683 28. Dobashi Y, Hirano T, Hirano M, Ohkatsu Y. Antioxidant and photo-antioxidant 684 abilities of catechins. J Photochem Photobiol A Chem. 197:141–8. 685 29. Chen C, Twito S, Miller G. New cross talk between ROS, ABA and auxin 686 controlling seed maturation and germination unraveled in APX6 deficient Arabidopsis 687 seeds. Plant Signal Behav. 2014. 688 30. Noctor G, Reichheld JP, Foyer CH. ROS-related redox regulation and signaling in 689 plants. Seminars in Cell and Developmental Biology. 2018. 690 31. Choudhury FK, Rivero RM, Blumwald E, Mittler R. Reactive oxygen species, 691 abiotic stress and stress combination. Plant J. 2017. 692 32. Ibrahim HM. Selenium Pretreatment Regulates the Antioxidant Defense System 693 and Reduces Oxidative Stress on Drought-Stressed Wheat (Triticum aestivum L.) 694 Plants. Asian J Plant Sci. 2014;13:120–8. 695 33. Luo YL, Song SQ, Lan QY. Possible Involvement of Enzymatic and 696 Non-enzymatic Antioxidant System in Acquisition of Desiccation Tolerance of Maize 697 Embryos. Acta Bot Yunnanica. 2009;31:253–9. 698 34. Self-Incompatibility. In: Encyclopedia of Genetics, Genomics, Proteomics and 699 Informatics. Dordrecht: Springer Netherlands; 2008. p. 1781–2. 700 doi:10.1007/978-1-4020-6754-9_15325. 701 35. Wang CL, Zhang ZP, Oikawa E, Kitashiba H, Nishio T. SCR-22 of 702 pollen-dominant S haplotype class is recessive to SCR-44 of pollen-recessive S 703 haplotype class in Brassica rapa. Hortic Res. 2019. 704 36. Yasuda S, Wada Y, Kakizaki T, Tarutani Y, Miura-Uno E, Murase K, et al. A 705 complex dominance hierarchy is controlled by polymorphism of small RNAs and 706 their targets. Nat Plants. 2016. 707

37. Kakita M, Murase K, Iwano M, Matsumoto T, Watanabe M, Shiba H, et al. Two 708 distinct forms of M-locus protein kinase localize to the plasma membrane and interact 709 directly with S-locus receptor kinase to transduce self-incompatibility signaling in 710 Brassica rapa. Plant Cell. 2007. 711 38. Indriolo E, Tharmapalan P, Wright SI, Goring DR. The ARC1 E3 ligase gene is 712 frequently deleted in self-compatible brassicaceae species and has a conserved role in 713 arabidopsis lyrata self-pollen rejectionw. Plant Cell. 2012. 714 39. Sankaranarayanan S, Jamshed M, Samuel MA. Degradation of glyoxalase i in 715 Brassica napus stigma leads to self-incompatibility response. Nat Plants. 2015. 716 40. Kessler SA, Shimosato-Asano H, Keinath NF, Wuest SE, Ingram G, Panstruga R, 717 et al. Conserved molecular components for pollen tube reception and fungal invasion. 718 Science (80- ). 2010. 719 41. Diby L, Kahia J, Kouamé C, Aynekulu E. Tea, Coffee, and Cocoa. In: 720 Encyclopedia of Applied Plant Sciences. 2016. 721 42. Yu Y, Fu J, Xu Y, Zhang J, Ren F, Zhao H, et al. Genome re-sequencing reveals 722 the evolutionary history of peach fruit edibility. Nat Commun. 2018. 723 43. Liu C, Sun Q, Zhao L, Li Z, Peng Z, Zhang J. Heterologous expression of the 724 transcription factor EsNAC1 in arabidopsis enhances abiotic stress resistance and 725 retards growth by regulating the expression of different target genes. Front Plant Sci. 726 2018. 727 44. Swiderski MR, Birker D, Jones JDG. The TIR domain of TIR-NB-LRR resistance 728 proteins is a signaling domain involved in cell death induction. Mol Plant-Microbe 729 Interact. 2009. 730 45. Zhang XC, Gassmann W. Alternative splicing and mRNA levels of the disease 731 resistance gene RPS4 are induced during defense responses. Plant Physiol. 2007. 732 46. Qi D, de Young BJ, Innes RW. Structure-function analysis of the coiled-coil and 733 leucine-rich repeat domains of the RPS5 disease resistance protein. Plant Physiol. 734 2012. 735 47. Zhong R, Morrison WH, Freshour GD, Hahn MG, Ye ZH. Expression of a mutant 736 form of cellulose synthase AtCesA7 causes dominant negative effect on cellulose 737 biosynthesis. Plant Physiol. 2003. 738 48. Tanaka K, Murata K, Yamazaki M, Onosato K, Miyao A, Hirochika H. Three 739 distinct rice cellulose synthase catalytic subunit genes required for cellulose synthesis 740 in the secondary wall. Plant Physiol. 2003. 741 49. Krystynowicz A, Koziołkiewicz M, Wiktorowska-Jezierska A, Bielecki S, 742 Płucienniczak A. Molecular basis of cellulose biosynthesis disappearance in 743 submerged culture of Acetobacter xylinum. Acta Biochim Pol. 2005;52:691–8. 744 50. Lao NT, Long D, Kiang S, Coupland G, Shoue DA, Carpita NC, et al. Mutation of 745 a family 8 glycosyltransferase gene alters cell wall carbohydrate composition and 746 causes a humidity-sensitive semi-sterile dwarf phenotype in Arabidopsis. Plant Mol 747 Biol. 2003;53:647–61. 748

51. Lee C, Zhong R, Richardson EA, Himmelsbach DS, McPhail BT, Ye ZH. The 749 PARVUS gene is expressed in cells undergoing secondary wall thickening and is 750 essential for glucuronoxylan biosynthesis. Plant Cell Physiol. 2007;48:1659–72. 751 52. Shao M, Zheng H, Hu Y, Liu D, Jang JC, Ma H, et al. The gaolaozhuangren1 752 gene encodes a putative glycosyltransferase that is critical for normal development 753 and carbohydrate metabolism. Plant Cell Physiol. 2004;45:1453–60. 754 53. Jang JC, León P, Zhou L, Sheen J. Hexokinase as a sugar sensor in higher plants. 755 Plant Cell. 1997. 756 54. Liu C, Wang B, Li Z, Peng Z, Zhang J. TsNAC1 is a key transcription factor in 757 abiotic stress resistance and growth. Plant Physiol. 2018. 758 55. Reichler SA, Torres J, Rivera AL, Cintolesi VA, Clark G, Roux SJ. Intersection of 759 two signalling pathways: extracellular nucleotides regulate pollen germination and 760 pollen tube growth via nitric oxide. J Exp Bot. 60:2129–38. 761 56. Zhou J, Fan C, Liu K, Jing Y. Extracellular ATP is involved in the initiation of 762 pollen germination and tube growth in Picea meyeri. Trees - Struct Funct. 2015. 763 57. Lin J, Yang J. Determination of Fluoride Content in Plant Leaves by Microwave 764 Digestion and Fluoride Ion-Selective Electrode Method. Guangdong Trace Elem Sci. 765 2008. 766 58. Johansson LH, Håkan Borg LA. A spectrophotometric method for determination 767 of catalase activity in small tissue samples. Anal Biochem. 1988. 768 59. Liu J, Jing T. A comparison between two procedures based on pyrogallol 769 autoxidation for determination of SOD activity. 1996;17. 770 60. Vetter JL, Steinberg MP, Nelson AI. Enzyme Assay, Quantitative Determination 771 of Peroxidase in Sweet Corn. J Agric Food Chem. 6:39–41. 772 61. Wellburn AR, Lichtenthaler H. Formulae and Program to Determine Total 773 Carotenoids and Chlorophylls A and B of Leaf Extracts in Different Solvents. In: 774 Advances in Photosynthesis Research. 1984. 775 62. Wang H, Helliwell K, You X. Isocratic elution system for the determination of 776 catechins, caffeine and gallic acid in green tea using HPLC. Food Chem. 2000. 777 778

• A low-level and a high-level fluoride varieties were selected for

the morphological and transcriptome analysis to explore their

different regulate network for the fluoride stress.

• High-fluoride tea tree varieties had stronger ROS-scavenging

ability than low-fluoride varieties....

• High-fluoride tea tree varieties could maintain the normal growth

and reproduction under high-fluoride stress

• Redistribution of carbohydrates retards the vegetative growth of

low-fluoride tea tree varieties under high-fluoride stress.

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships

that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered

as potential competing interests: