CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) 367 Antioxidant Quinonemethide Triterpenes from Salacia campestris by Paulo R. F. Carvalho, Dulce H. S. Silva, Vanderlan S. Bolzani, and Maysa Furlan* Instituto de Química, Universidade Estadual Paulista, CP 355, CEP14801-900, Araraquara, SP, Brazil (phone: ‡ 5516-201-6661; fax: ‡ 55-16-222-7932; e-mail: maysaf@iq.unesp.br) A new quinonemethide triterpene named as salacin, has been isolated from the root bark of Salacia campestris in addition to the known pristimerin, maytenin, 20a-hydroxymaytenin, and netzahualcoyene. Salacin was identified on the basis of NMR-spectral and mass spectrometric analysis. The free-radical scavenging activities of the quinonemethide triterpenes salacin (1), pristimerin (2), maytenin (3), 20a-hydroxymaytenin (4), and netzahualcoyene (5) towards DPPH have been evaluated and showed absorbance variation (DA) of 19, 20, 39, 28, 55, and 10%, respectively, having rutin (74% at 50 mm) and BHT (7% at 50 mm) as standard compounds. 1. Introduction. ± Quinonemethide triterpenes are natural products of occurrence restricted to Hippocrateaceae and Celastraceae families. Several representatives of this class are claimed to be medicinally useful [1] [2], as antimicrobial [3 ± 5], anticancer [4] [6] [7], antimalarial [8], and spermicidal compounds [9]. Some Salacia species, e.g., S. reticulata, S. oblonga, and S. chinesis, have long been used in India, Sri-Lanka, and China, respectively, as traditional medicines for treating rheumatism and skin diseases, and for their anti-inflammatory properties as well. Recent investigations have confirmed the hepatoprotective and antioxidant properties of S. reticulata, and disclosed the presence of catechins as the active principle [10] and aldose reductase inhibition by tingenin B, a quinonemethide from S. chinensis [11]. The antioxidant activity of organic compounds is dependent of few structural features that include, in most cases, the presence of phenolic OH-groups. Thus, flavonoids, lignans, phenylpropanoids, and other aromatic compounds have been the focus of a large number of publications dealing with antioxidant properties [12]. Antioxidant terpenes are rare, and the best known examples include rosmanol and carnosic acid derivatives, abietane diterpenes from sage (Salvia officinalis) and rosemary (Rosmarinus officinalis), in which the phenolic moiety has been produced during their biosynthesis [13] [14]. The antioxidant activities of the quinonemethide triterpenes celastrol, pristimerin, and acetylcelastrol have been previously assessed through demonstration of their lipid-peroxidation-inhibition effect in rat liver mitochondria [15]. On the other hand, no data are available on the free-radical scavenging activity of quinonemethide triterpenes. Useful information regarding the antioxidant activity of synthetic and natural products can be obtained from the measurement of their free-radical-scavenging properties, e.g., in dilution assays on the reduction of the stable free radical 1,1-diphenyl-2-picrylhydrazyl (DPPH) [16]. In the course of our studies on biosynthesis and bioactive compounds from Celastraceae and Hippocrateaceae species [12] [17], we describe herein the occurrence  2005 Verlag Helvetica Chimica Acta AG, Zürich 368 CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) of several quinonemethide triterpenes in S. campestris (Hippocrateaceae) root barks, including the new salacin (1), which had its structure determined through MS and NMR analysis, along with the known pristimerin (2), maytenin (3), 20a-hydroxymaytenin (4), and netzahualcoyene (5). In addition, the free-radical scavenging activities of the isolated quinonemethide triterpenes towards DPPH were measured and compared to those of the pentacyclic triterpene friedelin, which co-occurs with the quinonemethides in S. campestris, and to standard antioxidants BHT and rutin. The variation in the free-radical-scavenging activities due to structural features of the isolates is also discussed. 2. Results and Discussion. ± A CH2Cl2 extract of Salacia campestris root bark was subjected to silica-gel column chromatography. The fractions obtained were further separated by silica gel (CC) to give the quinonemethide triterpenes 1 ± 5. Salacin (1) was shown to have the molecular formula C28H38O4 by analysis of its ESI mass spectrum and elemental analysis. Its IR spectrum revealed absorption bands at 3490, 1725, and 1655 cm 1 characteristic of OH, ester CˆO, and conjugated CˆO groups, respectively. The 1H- and 13C-NMR spectra (Table) of 1 showed signals assignable to quinonemethide moiety at d(H) 6.30 (d, J(7,6) ˆ 7.0, H C(7)), a doublet at d(H) 6.47 (J(1,6) ˆ 1.0, H C(1)), and a doublet of doublets at d(H) 6.96 (J(6,7) ˆ 7.0, J(6,1) ˆ 1.0, H C(6)). The assignments of quinonemethide moiety were confirmed by interpretation of a DQCOSY spectrum, which showed mutual correlation between such H-atoms. The presence of one deshielded Me singlet at d(H) 2.20, assigned to Me(23), associated to a Me doublet at d(H) 0.89 (J ˆ 7.0), confirmed a norquinonemethide skeleton. CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) 369 Table. 1H- and 13C-NMR-Spectral Data for Salacin (1). At 500 and 125 MHz, resp., in CDCl3 . d in ppm, J in Hz; multiplicities determined by DEPT-1358, and assignments based on HMQC and HMBC experiments. Position 13 1 1 2 3 4 5 6 7 8 9 10 11a 11b 12a 12b 13 14 15a 15b 16a 16b 17 18 19a 19b 20 21 22 23 25 26 27 28 29 120.5 n.o. a) 146.5 118.2 127.5 133.9 118.2 170.4 42.7 165.2 33.8 6.47 (d, J(1,6) ˆ 1.2) ± ± ± ± 6.96 (dd, J(6,1) ˆ 1.2 and J(6,7) ˆ 7.0) 6.30 (d, J(7,6) ˆ 7.0) ± ± ± 2.12 (m) 2.20 (m) 1.69 (m) 1.63 (m) ± ± 1.72 (m) 1.53 (m) 1.45 (m) 1.56 (m) 1.61 (m) 2.24 (m) 1.85 (m) 2.10 (m) 3.81 (dd, J(21,20) ˆ 3.5, J(21,22) ˆ 2.0) 3.70 (d, J(22,21) ˆ 2.0, 1 H ) 2.15 (s) 1.41 (s) 1.24 (s) 0.68 (s) 1.14 (s) 0.89 (d, J(29,20) ˆ 7.0) C 29.9 45.0 40.3 28.5 36.5 30.6 45.9 24.2 31.5 74.0 72.2 10.0 38.7 21.6 22.1 27.4 18.2 H a) Not observed. The presence of two OH groups was confirmed by signals at d(H) 3.81 (dd, J(21,20) ˆ 3.5, J(21,22) ˆ 2.0, 1 H) and 3.70 (d, J(22,21) ˆ 2.0, 1 H), attributed to H C(21) and H C(22), respectively. The assignments of these two hydroxymethine H-atoms were corroborated by interpretation of the DQCOSY spectrum. The spin system derived from Ha C(21) and Ha C(22) was readily recognized by starting with the 1 H doublet of doublets at d(H) 3.81 assigned to Ha C(21), which showed a crosspeak with the 1 H doublet at d(H) 3.70 assigned to Ha C(22). The relative configuration of Hb C(21) and Ha C(22) was deduced from J values between Hb C(21) and Hb C(20) (3.5 Hz), and Hb C(21) and Ha C(22) (2.0 Hz), which corroborated the axial ± equatorial orientation of H C(20) and H C(21), and the equatorial orientation of H C(22) (Fig. 1). 370 CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) Fig. 1. Selected HMBC correlations and 1H-NMR coupling constants for salacin (1) The 13C-NMR data were in agreement with the quinonemethide skeleton due to the signals related to conjugated CˆC bonds in rings A and B at d(C) 170.4, 146.5, 133.9, 127.5, and 120.5, assigned to C(8), C(3), C(6), C(5), and C(1), respectively. The signals at d(C) 74.0 and 72.2 were attributed to two hydroxymethine C-atoms. The 13C-NMR signal at d(C) 18.2 attributed to Me(29) group corroborated a 30-nor-skeleton once quinonemethides with C(29)-nor-skeleton (with b-orientation of the Me group at C(20)) shows the shielded signal for C(30) around d(C) 10.6 ± 11.5 [1b]. The positions of these groups were unambiguously assigned by interpretation of the gHMBC spectra (Fig. 1). The signal at d(C) 74.0 assigned to C(21) showed correlation with the signal at d(H) 0.89 (H C(29)); and the signal at d(C) 72.2, assigned to C(22), showed correlation with the signal d(H) 1.14 (H C(28)). The analysis of gHMQC data was used to assign the signals for all H-bearing C-atoms and allowed the structural determination of the new quinonemethide triterpene salacin (1). Pristimerin (2) has been previously isolated from Salacia beddomei [18] and from S. kraussii [19]. Maytenin (3) has been described from S. crassifolia [4] and S. macrosperma [20]. 20a-Hydroxymaytenin (4) has been reported from S. macrosperma [20] and netzahualcoyene (5) from S. reticulata [21]. Analysis of NMR data of compounds 2 ± 5 and comparison with those reported in the literature allowed the identification of these quinonemethide triterpenes. The radical-scavenging activity of compounds 1 ± 6 has been evaluated towards the stable free radical DPPH, which exhibits an absorption maximum at 517 nm. The freeradical-scavenging activity of the quinonemethide triterpenes is due to the dienonephenol moiety, which displays great stability to the quinonemethide radical as soon as it has been formed, after hydrogen radical donation to DPPH. Netzahualcoyene (5, DA ˆ 55  3% at 50 mm) was more effective than quinonemethides 1 ± 4 (DA ˆ 19  2, 20  1, 39  2, and 28  3%, resp., at 50 mm) (Fig. 2). Such behavior seems to be associated with the more extended conjugation, which, in the case of compound 5, is spread through rings A ± D and should, therefore, stabilize the netzahualcoyene radical more effectively than the radicals of compounds 1 ± 4, in which the conjugated system is restricted to rings A and B. Conversely, friedelin (6, DA ˆ 10  2% at 50 mm), which lacks the dienone-phenol system, was not effective in promoting DPPH reduction and, CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) 371 therefore, appears as a poor radical scavenger. Structural variations at ring E do not apparently interfere with the free-radical-scavenging activity, although the presence of free hydroxyls and/or carboxyls on ring E might be important in biological systems due to interaction of these groups with cell membranes, affecting lipid peroxidation processes [15]. The comparison of quinonemethide triterpenes and standard compound rutin (DA ˆ 74  4% at 50 mm) evidences the reported major roles of some structural features of flavonoids as the catechol moiety present in ring B of rutin and the C(2)ˆC(3) bond for the enhancement of the free-radical-scavenging activity [22]. On the other hand, the synthetic antioxidant BHT (DA ˆ 7  1% at 50 mm) exhibited the weakest antiradicalar activity as expected due to its slow reaction rate with DPPH [23]. The use of Salacia and other Hippocrateaceae or Celastraceae species as folk medicines for purposes that include the treatment of gastric ulcer might be related to their content in antiradicalar compounds as flavonoids, which have been described elsewhere (e.g., from Maytenus aquifolium [12] and Salacia reticulata [10]) and the quinonemethide triterpenes described in this work. Fig. 2. Free-radical-scavenging activity of quinonemethide triterpenes 1 ± 5 and friedelin (6) isolated from S. campestris, and reference compounds BHT and rutin Experimental Part General. Silica gel (Merck; 230 ± 400 Mesh) was used for column chromatography (CC) unless otherwise stated, and solvents were redistilled prior to use. 1H- and 13C-NMR spectra were recorded on a Varian Unity-500 spectrometer at 500 and 125 MHz, resp., with CDCl3 as solvent and TMS as reference. IR Spectra were obtained on a Nicolet spectrometer. ESI-MS were recorded on a VG Platform-II spectrometer. Elemental analysis was performed on a Perkin-Elmer 2400-CHN apparatus. Plant Material. Salacia campestris root barks were collected in Fazenda Canchin at the Universidade Federal de Sao Carlos, SaÄo Carlos, SP and identified by Prof. Dr. Maria Helena de O. Antunes. The voucher specimens (No. 2845) are deposited at Herbarium of Departamento de BotaÃnica da Universidade Federal de SaÄo Carlos - UFSCar, SaÄo Carlos, SP, Brazil. Extraction and Isolation. Root barks of S. campestris were dried at 408 for 24 h in an oven with forced air circulation. The dried and powered root bark of S. campestris (485 g) was extracted with CH2Cl2 . The resulting CH2Cl2 extract was filtered and concentrated in vacuum to afford a red gum (100.9 g), which was submitted to CC on silica gel (300 g) and eluted with hexane/AcOEt gradient followed by AcOEt/MeOH gradient. The hexane/AcOEt 1 : 1 fraction (11.6 g) was subjected to silica-gel CC with hexane/AcOEt gradient to give 45 372 CHEMISTRY & BIODIVERSITY ± Vol. 2 (2005) fractions. Fraction 6 (0.296 g) was applied onto prep. TLC plates eluted with hexane/AcOEt/AcOH 97 : 2 : 1 to yield Fr. 6-3 (0.093 g), which was further purified by prep. TLC (hexane/AcOEt/AcOH 90 : 9 : 1) to give maytenin (3; 0.030 g) and netzahualcoyene (5; 0.046 g). Fr. 31 was subjected to prep. TLC eluted with CHCl3/ MeOH/AcOH 99 : 1 : 1 to give pristimerin (2; 0.012 g) and 20a-hydroxymaytenin (4; 0.043 g). Fr. 27 was submitted to two stepwise prep. TLC (hexane/AcOEt/AcOH 60 : 39 : 1 and then 70 : 29 : 1) to yield Fr. 27-3-2 (0.010 g), which was further purified by prep. TLC (hexane/AcOEt/AcOH 70 : 29 : 1) to yield the new quinonemethide triterpene salacin (1; 0.0055 g). Salacin (1). Amorphous solid. UV (MeOH): 420. IR (KBr): 3490, 2930, 1725, 1655, 1288, 1080. 1H- and 13 C-NMR: see the Table. ESI-MS: 439 (100, [M ‡ 1]‡ ). Anal. calc. for C28H38O4 : C 76.71, H 8.68, O 14.61; found: C 76.71, H 8.68, O 14.61. Reduction of 1,1-Diphenyl-2-picrylhydrazyl Radical (DPPH). The free-radical-scavenging capacities of quinonemethide triterpenes, friedelin, and standard compounds rutin and BHT were evaluated with respect to their ability to reduce the stable free radical DPPH. Stock solns. of each compound consisted of 1 mg solubilized in MeOH (10 ml). Several dilutions of each compound from 5  10 6 mol l 1 to 100  10 6 mol l 1 were then prepared in MeOH, and to each sample a soln. of DPPH (0.004%) was added. Tests were performed in triplicate, absorbance at 517 nm was determined after 30 min on a Milton Roy 20D spectrophotometer, and the percentage of activity was calculated [16]. This work was supported by grants provided by FAPESP (proc. 97/10184-7). P. R. F. C. thanks FAPESP for providing a scholarship; M. F. and V. S. B. are grateful to CNPq for research fellowships. The authors are grateful to Prof. Dr. Paulo Cezar Vieira for collecting the plant material. REFERENCES [1] a) A. A. L Gunatilaka, Triterpenoid Quinonemethides and Related Compounds (Celastroloids), Prog. Chem. Org. Comp. 1996, 67, 1; b) H. C. Fernando, A. A. L. Gunatilaka, Y. Tezuka, T. Kikuchi, Tetrahedron 1989, 45, 5867. [2] F. Calzada, R. Mata, R. Lopez, E. Linares, R. Bye, V. M. Barreto, F. Del Rio, Planta Med. 1991, 57, 194. [3] S. S. Bhatnagar, P. V. Divekar, J. Sci. Ind. Res. 1951, 10 B, 56. [4] O. Gonçalves de Lima, J. S. de B. Coelho, E. Weight, I. L. DAlbuquerque, D. A. Lima, M. A. de M. Souza, Rev. Inst. AntibioÂt. 1971, 11, 35. [5] L. Moujir, A. M. Gutierrez-Navarro, A. G. Gonzalez, A. G. Ravelo, J. G. Luis, Biochem. Syst. Ecol. 1990, 18, 25. [6] O. Ngassapa, D. D. Soejarto, J. M. Pezzuto, N. R. Farnsworth, J. Nat. Prod. 1994, 57, 1. [7] A. T. Sneden, J. Nat. Prod. 1981, 44, 503. [8] K. Pavanand, H. K. Webster, K. Yongvanitchit, A. Kun-Anake, T. De-Chatiwongse, W. Nutakul, J. Bansiddhi, Phytother. Res. 1989, 3, 136. [9] G. A. S Premakumara, W. D. Ratnasooriya, Contraception 1992, 45, 239. [10] M. Yoshikawa, K. Ninomyia, H. Shimoda, N. Nishida, H. Matsuda, Biol. Pharm. Bull. 2002, 25, 72. [11] T. Morikawa, A. Kishi, Y. Pongpiriyadacha, H. Matsuda, M. Yoshikawa, J. Nat. Prod. 2003, 66, 1191. [12] J. Corsino, D. H. S. Silva, M. V. B. Zanoni, V. da S. Bolzani, S. C. França, A. M. S. Pereira, M. Furlan, Phytother. Res. 2003, 17, 913. [13] M.-N. Maillard, P. Giampaoli, M.-E. Cuvelier, Talanta 1996, 43, 339. [14] D. A Markovic, Z. Djamarti, R. M. Jankov, D. M. Markovic, L. M. Ignjatovic, Mikrochim. Acta 1996, 124, 219. [15] H. Sassa, K. Kogure, Y. Takaishi, H. Terada, Free Rad. Biol. Med. 1994, 17, 201. [16] M. S. Blois, Nature 1958, 181, 1199. [17] a) J. Corsino, A. C. Alecio, M. L. Ribeiro, M. Furlan, A. M. S. Pereira, I. B Duarte, S. C. França, Phytochemistry 1998, 9, 245. b) J. Corsino, P. R. F. Carvalho, M. J. Kato, L. R. Latorre, O. M. M. F. Oliveira, A. R. Araujo, V. da S. Bolzani, A. M. S. Pereira, S. C. França, M. Furlan, Phytochemistry 2000, 55, 741. [18] A. Hisham, G. J. Kumar, Y. Fujimoto, N. Hara, Phytochemistry 1995, 40, 1227. [19] J. N. Figueiredo, B. Raz, U. Sequin, J. Nat. Prod. 1998, 61, 718. [20] G. C. S. Reddy, K. N. N. Ayengar, S. Rangaswami, Indian J. Chem., Sect. B 1981, 20, 197. [21] B. Dhanabalasingham, V. Karunaratne, Y. Tezuka, T. Kikuchi, Phytochemistry 1996, 42, 1377. [22] S. A. B. E. Van Acker, D.-J. Vanden Berg, M. N. J. L. Tromp, D. H Griffioen, W. P Van Bennekom, W. J. F. Van der Vijgh, A. Bast, Free Rad. Biol. Med. 1996, 20, 331. [23] V. Bondet, W. Brand-Williams, C. Berset, Lebensm.-Wiss. Technol. 1997, 30, 609. Received March 12, 2004