Seasonal variation of biosynthetic path selectivity of flavonoids, and 1 content and antioxidant activity of metabolites in Tetrastigma 2 hemsleyanum Diels et Gilg

This work investigated the seasonal variations in biosynthetic path selectivity of nine individual 32 flavonoids, and content and antioxidant activity of three major metabolites in Tetrastigma 33 hemsleyanum . The results revealed that under conditions of precipitation (2.0~6.6 mm), temperature 34 (17.5~24.1°C), humidity (67.3~80.2%) and sunshine duration (3.4~5.8 h) in April and May, the 35 total content of flavonoids reached higher levels between 282 and 394.5 μg/g; in the second half of 36 April with the highest accumulation of flavonoids, the production selectivity (PS) of isoorientin, 37 orientin, rutin, isoquercitin, kaempferol-3-0-rutinoside, astragalin, quercetin, apigenin and 38 kaempferol were 0.30, 0.06, 0.07, 0.07, 0.00, 0.04, 0.38, 0.05 and 0.03, respectively; also according 39 to the reaction network consisting of three main pathways for flavonoids syntheses in T. 40 hemsleyanum , the selectivity of reaction Path II to synthesize quercetin, isoquercitin and rutin in 41 April, May and June fluctuated between 0.52 and 0.66, much higher than that of reaction Path I or 42 Path III.


UPLC analysis of individual flavonoids 169
Dry powder of T. hemsleyanum tuber (5 g) was extracted by reflux with 80% methanol of 75 mL 170 at 85°C for 120 min. The supernatants from two repeated extractions were blended and evaporated in 171 a rotary evaporator at 65°C. Then, the as-obtained extract was dissolved in binary solvent of 172 water/methanol at 10: 1 (mL/mL) to give a 55 mL solution, which was again mixed with triple 173 volumes (165 mL) of anhydrous diethyl ether and ethyl acetate, respectively. After extraction 174 repeated twice, the supernatant of anhydrous diethyl ether extraction and that of ethyl acetate 175 extraction were blended, and then the solvents in the blended solution were removed by rotary 176 evaporation to give the product of flavonoids. Finally, the product was dissolved in 10 mL methanol 177 and then filtered through a 0.22 m nylon filter for UPLC analysis. 178 The quantitative analysis of T. hemsleyanum tuber extract was performed on an ultra high 179 performance liquid chromatography system (UPLC) (Agilent 1290, Santa Clara, CA, USA). 180 Chromatographic conditions included injection volume of 10 µL, flow rate at 1 mL/min, column 181 temperature in 32°C, detection wavelength at 280 nm and a column of 1.8m, 4.6 mm ×100 mm, i.d.  containing 2.7 mL ethanol, and then 1 mL sodium nitrite solution of 5% was added. After stood for 6 201 min, the mixture was blended with 1 mL aluminum nitrate solution of 10%. After another 6 min, the 202 blended mixture was mixed with 4% sodium hydroxide solution of 10 mL and followed with 10 mL 203 distilled water. After the last 15 min, the absorbance of final mixture at 510 nm wavelength was 204 measured by UV-1800 spectrophotometer (Shimadzu, Japan) in comparison to a standard curve 205 regression of rutin (y=0.0121x-0.0098, r²=0.99), and the content of phenolic compounds were 206 expressed as rutin equivalents with mg/g dry weight. All samples were tested in five replicates to 207 give the data with p<0.05. 208

Measurement of polysaccharides 209
Aqueous solution (50 mL) containing 2 g powder of T. hemsleyanum tuber was mixed with a 210 certain amount of glacial acetic acid to give pH=3, and then the mixture was extracted by reflux at 211 82°C for 1.5 h. Next, the extraction solution was separated by centrifugation at 3500 rpm for 15 min, 212 and the supernatant of 0.2 mL was added to a test tube to determine the content of polysaccharides by 213 UV-1800 spectrophotometer (Shimadzu, Japan) at 490 nm wavelength with using the phenol-214 sulfuric acid method (Jain et al., 2017). Finally, the determined content was compared with a 215 standard curve regression of glucose (y= 0.0629x+0.0047, r²= 0.9995). All samples were measured 216 by five repetitions to give the data with p<0.05. 217

Measurement of steroids 218
Steroids in the dry powder of T. hemsleyanum tuber was extracted with ethanol as solvent by 219 ultrasonication and followed by centrifugation at 2500 rpm for 8 min. Then, 20 mL sample solution 220 was mixed with triple volume petroleum ether (60 mL), and the supernatant was analyzed by sulfate-221 phosphate-ferric method (SPF) (Zhou et al., 2015) with minor modification. In detail, the 222 chromogenic reagent of SPF was first prepared by dissolving 2.5 g FeCl3•6H2O into 85% phosphoric 223 acid of 100 mL, of which 4 mL solution was dissolved again in 100 mL sulfuric acid for use. 224 Subsequently, 6 mL assay mixture containing 0.2 mL solution of phytosterols in ethanol and 3 mL 225 SPF chromogenic solution was shaken to be homogeneous, and then determined by UV-1800 226 spectrophotometer (Shimadzu, Japan) at 560 nm wavelength. Finally, the content of steroids was 227 determined by comparing with a standard curve regression of Beta-sitosterol (y=0.0125x+0.0256, r 2 228 = 0.99). All samples were analyzed with five repetitions to give the data with p<0.05. August and a minimum of 4.1 or 0°C in early February, and X3 reached a maximum or minimum of 273 33.9 or 7.2°C in late July or late January respectively; X5 or X6reached a maximum of 85.2 or 68.9% 274 in late January and a minimum of 64.1 or 36.9% in early February; also X7 reached a maximum of 275 8.1 h in late July and a minimum of 0.9 h in late January. In general, the summer from June to 276 August was hot, more rainy and full of sunshine, and the winter from January to February was cold, 277 less rain and lack of sunshine, while air humidity was the highest in the second half of January and 278 the lowest in the first half of February. 279

Influence of seasonal climate on biosynthesis of flavonoids
Herein nine important individual flavonoids including isoorientin, orientin, rutin, isoquercitin, 281 kaempferol-3-0-rutinoside, astragalin, quercetin, apigenin and kaempferol were identified, and the 282 UPLC chromatogram of the standard sample was displayed in Fig and detected from T. hemsleyanum (Fig. 1), probably due to their rapid conversion to downstream 347 products or due to the limitation of sensitivity in UPLC analysis. Overall, the full biosynthetic course 348 involves three types of reactions including hydroxylation, dehydrogenation (formation of double band in 349 ring) and glycosylation. Along PathI, naringenin is converted to apigenin (Y8) to luteolin in series, further 350 into isoorientin (Y1) and orientin (Y2) in parallel or by potential cross-path to quercetin (Y7) in parallel; 351 along Path III, naringenin is converted to dihydrokaempferol to kaempferol (Y9) in series, and then to 352 astragalin (Y6) further into kaempferol-3-0-rutinoside (Y5) or by potential cross-path to quercetin (Y7); 353 also along Path II, naringenin could be converted to eriodictyol to dihydroquercelin to quercetin (Y7) in 354 turn. As the intersection point of Path I, II and III, in particular, quercetin (Y7) is the only one which could 355 be synthesized from naringenin through any of the three Paths, and after formed, it could be converted 356 into isoquercetin (Y4) and followed to rutin (Y3). Besides the main biosynthetic chains above, there are 357 two potential cross-paths, including that dihydrokaempferol on Path III could be converted into 358 dihydroquercetin on Path II, and eriodictyolon path II could be converted into luteolin on Path I. 359 There are some interests to quantitatively compare the seasonal selectivity of three main 360 biosynthetic paths I, II and III of flavonoids without considering cross-paths. To do so, the production 361 selectivity of the jth flavonoid compound (PSj) could be defined as 362 where Yj is the content of the jthflavonoid, and j=1, 2, …., 9. 364 Then the selectivity (RPS) of reaction path I, II and III could be evaluated by 365 RPS I = PS 8 + PS 1 + PS 2 366 RPS II = PS 7 + PS 4 + PS 3 367 RPS III = PS 9 + PS 6 + PS 5 368 The as-calculated data of RPSI, RPSII and RPSIII with seasonal variation were illustrated in Fig. 5. As 369 seen, the selectivity of path I (RPSI) was higher than that of path II or path III (RPSII or RPSIII) 370 during the period from 1 September to 31 December, while the selectivity of path II (RPSII) was 371 higher than that of path I or path III (RPSI or RPSIII) during the period of 1 January to 31 August 372 except for the first half of February and the second half of March. In February 1-15, RPSIII reached 373 its maximum of 0.39 slightly higher than that of RPSII (0.36), simply attributed to the highest Y6 374 higher than RPSI or RPSIII, determinately because quercetin gave the highest content (Y7) and 383 production selectivity (PS7) among the nine flavonoids during the period. Giving a consideration of the 384 potential cross-paths in the biosynthetic network of flavonoids (Fig. 4), quercetin might be also generated 385 from luteolin in the Path I or kaempferol in the Path III. In other words, as the intersection point of Path 386 I, II and III, quercetin (Y7) was the only one which could be synthesized from naringenin through any of 387 three different biosynthetic paths, and therefore became the most abundant flavonoid compound in T. 388 hemsleyanum during the period of April to June particularly. In brief, there must be the diversity of 389 reaction mechanism and the difference of enzyme catalytic activity for biosynthesis of flavonoids in 390 T. hemsleyanum, and the production selectivity of every flavonoid compound will depend on its 391 synthetic pathway, enzyme activity and environmental factors. 392

Contents of three major metabolites 394
The contents of phenolic compounds (YTPH), sterols (YTS) and polysaccharids (YTP) in 24 395 samples of T. hemsleyanum tubers have been measured to give 72 data in total. As previously 396 presented in Table 1, their contents in any sample followed the order of YTPH> YTP> YTS, and varied 397 in the range of 29.1±1.2~77.7±3.2 mg/g, 9.9±0.9~21.7±0.9 mg/g and 0.73±0.09~1.7±0.1 mg/g, 398 respectively. By mass, polysaccharids were the most abundant and sterols were the rarest in T. 399 hemsleyanum. Also it should be noted that the content YTPH was contributed by various phenolic 400 compounds including flavonoids and phenolic acids, and thus its seasonal variation was different 401 from that in the total content of nine individual flavonoids (Fig. 2). 402

Antioxidant activities of three major metabolites 403
Chemistry and biochemistry of antioxidants have significant interests in academic research and 404 practical application, because oxidative stress plays a pivotal role in pathogenesis of cardiovascular 405 diseases, neural disorders, diabetes, cancer and aging (Ksouri et al., 2012). Plants or parts are 406 commonly perceived as excellent and safe sources of antioxidant compounds, and their antioxidant 407 activity usually evaluated by common methods such as DPPH (2, 2-diphenyl-1-picrylhydrazyl)-RSA 408 assay, Table 2 presented 72 data of the DPPH radical scavenging for three major metabolites in 24 409 samples of T. hemsleyanum tubers. As seen, the DPPH radical scavenging rate (%) of phenolic 410 compounds (ZTPH), sterols (ZTS) and polysaccharides (ZTP) in any sample followed the order of ZTPH> 411 ZTP> ZTS, and they varied in the range of 39.9-95.1%, 35.4-43.5% and 5.5-9.8%, respectively. 412 Clearly, sterols with very low content displayed much lower antioxidant activity than phenolic 413 compounds or polysaccharides. 414

Inter-correlation between metabolite contents and environmental factors 416
To disclose inter-correlation between 3 response contents of major metabolites (YTPH, YTP, YTS) 417 and 7 environmental factors (Xi), Redundancy and Person analysis have been carried out. By 418 directionally ranking the statistical significances of various variables and responses at the same time, 419 redundancy analysis could preliminarily evaluate the correlation or influence of various 420 environmental factors to or on metabolites. As showed by the RDA plot in Fig.6A, most of the 421 climatic factors showed a negative correlation to the metabolites, and the coefficients of all 422 correlations were less than 0.8. The data in Table 3      Quercetin formed from pathII and also potentially from path I or III, and successively converted to isoquercetin to rutin; (C) The flavonoids formed in path Ⅲ, as illustrated in Fig. 4.  Table 1; YTPH, YTS and YTP indicated in Table 1; ZTPH, ZTS and ZTP indicated in Table 2 Fig.1