Structure of a new xanthylium salt derivative

TETRAHEDRON LETTERS Tetrahedmn Letters 40 (1999) 5869-5872 Pergamon Structure of a N e w Xanthyfium Salt Derivative Nour-Eddine Es-Safi "t, Christine Le Gueraev&, Benogt Lab-rbe, H$1~ne Fulerand, V&roaique Cheyuier sad Michel Moutouaet ISW-INRA lnstitut des Preduits de la Vigne, Unit~de RecherchesBiopolymheset Ar0mm, 2 Place Viala, 34060Montpelliercedex2, France. t l~coleNormaleSup~rietwe,~ de Chimie Organiqueet d'Etudes Physicochimiques, AvenueOued Akreuch, B.P. 5118 Tak_~___mjmRabat, Marec. Received 28 April 1999; accepted 15 June 1999 Almraet. A new polyphanoliccompoundwith a xanthylium skeleton has been synlhesised from (+)catechin and glyonylicacid. Its structural elucidationwas achievedby UV, MS and NMR spectrmcepies. Its formation involvedglyoxylicacid-mediated dimerisation of (+)-~_t~in I 8ivin8 thus ~mpoued 2, followed by cyclisation to a xanthene derivative 3 which by oxidation afforded the xanthylimn compound 4. The detection and structmal detmnination of the xanthene compound confirmed the postulated mechanic. © 1999 Published by Elsevier ScienceLtd. All rights reserved. Fiavanols are polyphenolic compounds found in many plants, fruits and beverages such as fruit juices, beer and wine. They have attracted much attention in relation to their physiological activities and their role has become an important issue in the reiatiorLqhip between health and hmmn diet. During conservation, phenolic compounds usually undergo progressive changes which affect sensorial properties like colour, taste end colloidal stability 1'2. Various mechanisms have been suggested to explain such transformations. Processes involving either direct condensation between fiavanols and anthocyanins giving rise to compounds with a yellow brown hue 3'4 5-11 (xanthylium salts) or reactions mediated by acetaldehyde with the formation of violet pigments have been studied in model solution systems. While the process involving acetaldehyde is well documented, little is known about the formation and the structures of xanthyllum salts. However, it is believed that xanthylium nuclei might occur in condensed tannins (flavanol polymers), and that such chromophores may contn'bute to the high absorbanee o f wine pigments in the 400-500 ran region4'12'13. In this paper, we describe the synthesis, the isolation and the chemical structure analysis of a new yellowish xanthyfium compound named NJ2 4 formed by interaction between (+)-cateehin I and glyoxylie acid (figure 1). OH HO.. t-~.. L _|/~,. * ~OH S /OH HO °'T"" y v HO" H 0 -o. , 1' 5' OH 1 I-I0" ~f~ OH OH OH HO 2 F~,ure 1 Fax. : 33 4 99612683 e-mail : msafi~msam.inra.fr 0040-4039/99/$ - see front matter © 1999 Published by Elsevier Science Ltd. All rights reserved. S0040-4039(99)01156-9 PII: OH OH 4 5870 Incubation of (+)-catechin 1 with tartaric acid in iron catalytic medium yielded colourless compounds exhibiting maxima at 280 nm and yellowish compounds with maxima in the region of 440 - 460 nm as reported 14"15 . . . 14 . earfier . As previously indicated , the obtained colourless compounds consisted of (+)-catechin units bridged by a methine carboxylic acid group (compound 2). Formation ofthe yellowish compounds was rchtcd to the disappearance of the colourless ones suggesting that the former were probably formed by evolution and further rearrangement of the latter. This prompted us to prepare them livm individually isolated colourless products. Thus, the major colourless dimer 2 was isolated by high performance liquid chromatography at the semipreparative scale and was further incubated in pH 3.5 solution, and appearmr~ of two new compounds referred as to NJ2 and NJ3, initially absent in the mixture were observed. Their retention times and UV-visl~ole spectra showing absorption maxima around 440 and 460 rim, respectively are the same as those of pLements formed in reaction between (+)-catechin and tartaric acid in iron catalysed medium15. LC/ESI-MS analysis of the major yellowish compound NJ2 both in positive and negative ion modes showed m/z values at 615 and 617, resp~ively. The fact that the molecular weight of compound NJ2 was 20 mass units lower than that of the coiourless dimer 2 ([M-H]" at m/z = 635), suggested that the yellowish compound may be formed by a dehydration followed by an oxidation process. The dehydrated product (Mr -618) was actually detected by LC/MS analysis at m/z 617, in the negative ion mode. The loss of a water molecule may be achieved either between the carboxyl group and a neighbouring hydroxyl group like those at the 7 position of the A or A' rings yielding, after oxidation, a lactonised product, or between the two 7-OH groups, giving a xanthylium salt 4 after an oxidation process. The UV-vi~'ble spectrum of NJ2 showed two maxima at 273 and 444 ran in addition to a shoulder at 308 nm. These maxima were I~hocbromically shifted to 283, 496 and 325 nm by addition of NaOH (figure 2), as reported earlier for xanthylium salts4'13'16. 2.5 i it 2 1.5 1 0.5 0 250 300 350 400 Wavelength 450 500 550 600 (am) Figure 2: UV-Visible spectra o f compounds NJ2, NJ3 and NJ2 + NaOH. The fragment ions at 571,463 and 419 obtained by mass spectroscopy, in the negative ion mode, can be atm'buted respectively to the loss of carboxyl residue (-44 mass units), the loss of a hydroxyvinylphenol group (CffhO3)obtained by retro-Diels-Alder fission (-152 mass units) and finally the ion obtained after both fragmentations (-196 mass units), in agreement with the proposed structure 4. The structure of NJ2 was further elucidated fiom IH and I~C chemical shift a s s ~ n e n t s (Table 1). Unlike that of its colourless precursor which was reported to give two catechin systems14, the H spectrum of compound N J2 presented only 8 protons, corresponding to one cat3echin spin system, thus indicating the existenceofsome symmetry in the structure. This was confirmed by 1D CNMRspcctrum, which showed only 5871 17 signals (15 corresponding to one catachin skeleton and 2 constituting the bridge moiety). Besid~l,14the disappearance of the proton signal previously attributed to that of the mathine bridge was observed , in agreement with the proposed oxidised structure. Position 2C, 2C' 3C, 3C' 4¢xC, 44zC' 415C, 4[IC' 4aA, 4aA' 5A, 5A' 6A, 6A' 7A, 7A' 8A, 8A' 8aA, 8aA' 9D 10 ~ IH (ppm); m; J (Hz) 5.16; broad d; J = 3.3 4.18; m 2.46; dd; J = 3.6, 17.35 2.66; dd; J = 3.3, 17.35 6.90; s - I'B, I'B' - 2'B, 2'B' 3'B, 3'B' 4'B, 4'B' 5'B, 5'B' 6'B, 6'B' 6.66; broad s 6.65; broad d; J = 8.0 6.54; broad d; J = 8.0 5 13C 82.7 64.4 24.5 24.5 107.9 170.5 95.6 156.8 104.8 154.9 149.4 166.1 128.9 113.5 145.2 145.3 115.7 117.0 Table 1. IH and 13C assignements of compound 4 in DMSO-de--TFA (9:1) The signals which resonate at 2.46, 2.66, 4.18 and 5.16 ppm could be readily assigned to H-4, H-3 and H-2 of the C and C' ~ while those located at 6.65 and 6.54 ppm were assigned to H-5' and H-6' after COSY exp~. Assignem~ of the two remalnln~ signals (6.66 and 6.90 ppm) was achieved using TOCSY expr~i,m~. The broad singiet located at 6.66 ppm with correlates with both H-5' and H-6' was thus atm'imted to H-2' while the singiet at 6.90 ppm which gives no scalar correlation was assigned to H-6. The ROESY spectnan also showed that protons H-2 and H-3 gave correlations with only two aromatic protons : one at 6.66 ppm which was already attrihaed to proton H-2', and the other at 6.54, which can be either H-5' or H-6'. Since the latter is in a more favourable position to give such coupling, the signal at 6.54 ppm was attn'buted to H-6", whereas the broad aromatic doublet resonating at 6.65 ppm was assigned to H-5'. After the proton resonances had been assigned, all the corresponding carbons were attn'buted from the short-range HSQC ~ . The assignment of quaternary carbons was obtained from a long-range HMBC experiment. On the basis of ~ s e amlyses, attrflmtions of the various proton and carbon chemical shiRs were achieved, allowing to establish that the synthesised compound NJ2 was xanthylium salt 4. From a mechanistic point of view, the formation of compound 4 may result from ~lisation of the colourless dimer giving a xanthene structure as previonsl~y reported for 9-methyl-xalRhene17. The olRained ~ 3 is tbe~ oxidised to xanthylhan salt as sbown in figure 3. The detection ofthe ~Anthene 3 compound (Mr: 618) among the products formed from NJ2 precursor 2 by LC-MS analysis confnm~ thus the proposed mechanism. Moreover the intermediate ~ derivative was obtained by reduction of the xanthylium NJ2 and its structure elucidated by UV, MS and NMR spectroscopies. 5872 4- HO,/-~O 2 • ,1 / / " " O H 03 HO f '~'O 4 Figure 3: Mechanism ofxanthylium 4 formation from the colourless dimer 2 via the xanthene 3 derivative. Com~ar~ to the results obtained with the xanthylinm salt 4, an additional singlet was observed at 4.86 ppm in the H NMR spectrum and atm~outed to the proton H-9. This proton correlated, in HSQC e ~ t , with a carbon located at 34.45 ppm which was then attnl~uted to C-9. In the HMBC experiment spectrum, correlations with C-7, C-8 and C-10 were observed, confirming the structure of the xanthene compound and thus offering another argument to support the proposed structure 4 for the yellow compound NJ2. The formation of such yellow xanthylium compounds in wine-like model solutions suggests their poss~le contributions in colour evolution and browning observed during conservation and ageing of grape derived foods. In addition, more polymerised compounds where xanthene and xanthylium nuclei are incorporated were also detected in model solution system containing (+)-catechin and glyoxylic acid, aRer 24 hour incubation. This indicates the implication of such derivatives in the polymeric pigments respons~le for the high absorption around 450 nm formed during wine ageing. Our results also indicate that other reaction pathways conm'butdng to browning compete with polycondensation reactions and offer new information and support to the conm'bution of xanthylium salts in colour evolution and browning. They finally open perspectives for fiLrtherinvestigations of similar compounds. A number of properties such as temperature stability and copigmentation, in addition to their poss~le use as food colorants, remain of a high interest. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Somers, T. C. Phytochemistry 1971, 10, 2175-2186. Hnslam,E. Phytochemistry 1980, 19, 2577-2582. Liao, H.; Cai, Y.; Haslara, E. J. Sci. FoodAgric. 1992, 59, 299-305. Jurd, L. and Somers, T. C. Phytochemistry 1970, 9, 419-427. Es-Safi, N.; Fulcrand H.; Cheynier V.; Moutounet, M. J. Agric. Food Chem. 1999, 47, 2096-2102. Timberlake, C. F. and Bridle, P. Am. J. Enol. Vitic. 1976, 27, 97-105. Baranowski, E. S. and Nagel, C. W. J. Food Sci. 1983, 48, 419-421. Bakker, J., Picinelfi, A. and Bridle, P. Vitis 1993,32, 111-118. Escribano-Bailon, T., Dangles, O., and Brouillard, R. Phytochemistry 1996, 41, 1583-1593. Fraug3la-Arich~E. M., G u f r ~ M. T., Rivas-GonT.qlo,J. C. and Santos-Buelgeg C. J. Agric. Food Chem. 1997, 45, 2262-2266. Rivas-Gonzalo, J. C., Bravo-Haro, S. and Santos-Buelga, C. J. Agric. Food Chem. 1995, 43, 1444-1449. Jurd, L. Am. J. Enol. Vitic. 1969, 20, 191-195. Hrazdina, G. and BorzelL A. J. Phytochemistry 1971, 10, 2211-2213. Fulcrand, H., Cheynier, V., Oszmians~ J. and Moutounet, M. Phytochemistry 1997, 46, 223-227. Os27ninrmki~J., Cheynier, V. and Moutounet, M. J. Agric. Food Chem. 1996, 44, 1712-1715. Dangles, O. and Brouillard, R. NewJ. Chem. 1994, 18, 287-294. White, J. and Foo, L. Y. Tetrahedron Lett. 1990, 31, 2789-2792.