Biochemical Systematics and Ecology 33 (2005) 831e839 www.elsevier.com/locate/biochemsyseco Chemosystematic value of cyclopeptide alkaloids from Heisteria nitida (Olacaceae) Hesham R. El-Seedi a,b,*, Sonny Larsson a,*, Anders Backlund a a Division of Pharmacognosy, Department of Medicinal Chemistry, Biomedical Centre, Uppsala University, Box 574, SE-751 23 Uppsala, Sweden b Department of Chemistry, Faculty of Science, El-Menoufia University, Shebin El-Kom, Egypt Received 14 May 2004; accepted 24 December 2004 Keywords: Heisteria nitida; Olacaceae; Cyclopeptide alkaloids; Chemosystematics 1. Introduction Previously two cyclopeptide alkaloids have been isolated from Heisteria nitida Engl. (Olacaceae), namely integerrenine 1 and the unusual oxide anorldianine 27-N oxide 2, see Fig. 1 (El-Seedi et al., 1999). In this work we report the isolation of anorldianine 3 (Fig. 1) from this plant, and discuss some chemosystematic implications of these alkaloids. Anorldianine has previously only been reported from Canthium arnoldianum Hepper [misspelled as Canthium anorldianum throughout the reference, hence giving the alkaloid the name anorldianine], also known as Psydrax arnoldiana Bridson (Rubiaceae). This alkaloid has a unique substructure containing proline (Dongo et al., 1989). Cyclopeptide alkaloids have been found in several * Corresponding authors. Tel.: C46 18 471 44 96/97; fax: C46 18 50 91 01. E-mail addresses: hesham.el-seedi@fkog.uu.se (H.R. El-Seedi), sonny.larsson@fkog.uu.se (S. Larsson). 0305-1978/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.bse.2004.12.023 832 H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 Fig. 1. The chemical structures of the cyclopeptide alkaloids isolated from Heisteria nitida. families. No extensive investigations into their physiological role seem to have been performed, but reviews report antibacterial and antifungal activities (e.g. Gournelis et al., 1997), and vignatic acid A is lethal to the larvae of the weevil Callosobruchus chinensis (Sugawara et al., 1996). The structural type that contain 14 atoms in the macrocyclic part have been found in nine families: Olacaceae, Celastraceae, Phyllanthaceae, Pandaceae, Fabaceae, Rhamnaceae, Urticaceae, Malvaceae and Rubiaceae (El-Seedi et al., 1999; Gournelis et al., 1997; Sugawara et al., 1996). The variations in the macrocyclic entity can be grouped as shown in Fig. 2, and the distribution of these groups among the plant families is given in Table 1. 2. Material and methods 2.1. Plant material The bark of H. nitida was collected by Dr Felipe Ghia in 1992 at the Reserva Biologica, Jatun Sacha, Provincia del Napo, Ecuador. Voucher specimens are deposited in the Herbario Economica, Escuela Politecnica Nacional, EPN, Quito, Ecuador (G. F. 539), and in the herbarium of the Botany Section, Museum of Evolution at Uppsala University (UPS), Sweden. 2.2. Extraction and isolation The bark of H. nitida was dried at 40  C in a dark ventilated hood before grounding. The material (440 g) was extracted with light petroleum ether (40e60  ) three times with occasional stirring and filtered, followed by three extractions with methanol during 4 days each. The extracts were evaporated to give 3.3 g and 34 g of a gelatinous and an oily material, respectively. The methanol extract was partitioned between ethyl acetate and water. This resulted in 6.7 g of an ethyl acetate soluble fraction, and a water phase which was freeze-dried to give 27 g of a crude material consisting mainly of carbohydrates. H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 833 Fig. 2. The five substructures used as patterns for cyclopeptide alkaloids with 14 atoms in the macrocyclic part of the molecule. Displayed names are taken from groups proposed by Gournelis et al. (1997). The ethyl acetate fraction (6.5 g) was subjected to SEPARO column chromatography on 20 g of silica gel 60, using gradient elution with hexane:dichloromethane and ethyl acetate:methanol as previously described (El-Seedi et al., 2003). The eluted fractions were evaluated by TLC on silica gel using dichloromethane:methanol (98:2) as eluent, giving 16 main fractions. Fraction 7 (76 mg) showed presence of alkaloids as detected using Dragendorff’s reagent. This fraction was extracted with 0.1 M hydrochloric acid, and subsequently made alkaline with solid sodium bicarbonate before re-extracted with chloroform. Further purification over a Sephadex LH-20 834 H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 Table 1 The distribution of structural types for cyclopeptide alkaloids with 14 atoms in the macrocyclic substructure, with the number of investigated genera and species in each family Ordera Santalales Celastrales Malpighiales Fabales Rosales Malvales Gentianales a b Familya Olacaceae Celastraceae Pandaceae Phyllanthaceae Fabaceae Rhamnaceae Urticaceae Malvaceae Rubiaceae Genera:species 1:1 1:1 1:1 2:2 1:1 13:31 1:1 2:5 3:4 Structural typeb 1 2 3 4 5 C C ÿ C ÿ C C C C ÿ ÿ C ÿ C C ÿ ÿ ÿ C ÿ ÿ ÿ ÿ ÿ ÿ ÿ C ÿ ÿ ÿ ÿ ÿ C ÿ ÿ ÿ ÿ ÿ ÿ ÿ ÿ C ÿ ÿ ÿ Order and family according to APGII (APG, 2003). The signs denote presence (C) and no reported structure (ÿ), respectively. column using dichloromethane as eluent afforded anorldianine (10 mg) (Dongo et al., 1989), reported for the second time from natural origin. 2.3. Structural elucidation Our spectral data are deviating from that previously reported (Dongo et al., 1989), and accordingly extensive analysis with NMR and MS has been performed. The relative intensities of the m/z fragments from the electron impact mass spectrometry are given in Table 2. Dongo et al. (1989) assigned the structure based only on 1H, 13C and HeH COSY NMR spectral analysis, which could explain the difference in chemical shifts. The complete assignment of 1H NMR is presented together with the 13C NMR data in Table 2 The relative intensities of the m/z fragments of anorldianine electron impact mass spectroscopy experiment m/z Relative intensity 485 440 439 328 287 194 189 166 135 114 97 70 8 38 75 23 11 30 100 22 25 32 65 98 H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 835 Table 3 1 H and 13C NMR assignments of anorldianine, recorded at 400 MHz and 100.6 MHz, respectively Pos. 13 C shift dCa 1 3 4 5 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 28, 29 30 31 32 157.1 (s) 83.6 (d) 54.1 (d) 171.3 (s) 62.8 (d) 167.1 (s) e 126.7 (d) 116.5 (d) 131.6 (s) 130.4 (d) 120.6 (d) 120.9 (d) 132.1 (d) 28.9 (d) 15.6 (q) 20.4 (q) e 172.0 (s) 68.2 (d) 36.2 (t) 26.4 (d) 23.4 (q) 22.9 (q) 42.2 (q) 27.9 (t) 23.6 (t) 46.7 (t) 1 H shift dH (multi. JHeH, Hz)b e 5.01 4.75 e 4.23 e 5.00 6.64 6.40 e 6.96 7.07 7.29 7.10 1.37 1.24 0.99 8.45 e 2.50 1.49 1.65 0.94 0.92 2.12 1.46 1.91 3.86 (dd, J3,17 Z 1.5; J3,4 Z 6.9) (dd, J4,3 Z 6.9; J4,20 Z 9.8) (dd, J7,23 Z 6.9 and 9.8) (d, J9,10 Z 9.7) (dd, J10,11 Z 7.3; J10,9 Z 9.7) (d, J11,10 Z 7.3) (dd, J13,16 Z 1.5; (dd, J14,15 Z 1.5; (dd, J15,14 Z 1.5; (dd, J16,13 Z 1.5; (m) (d, J18,17 Z 6.7) (d, J19,17 Z 6.7) (d, J20,4 Z 9.8) J13,14 Z 7.3) J14,13 Z 7.3) J15,16 Z 8.8) J16,15 Z 8.8) (dd, J22,23 Z 5.8 and 8.3) (m), 1.31 (m) (m) (d, J25,24 Z 6.4) (d, J26,24 Z 6.4) (s) (m), 2.32 (m) (m), 1.69 (m) (m), 3.36 (m) HSQC-DEPT via 2,3JCeH HMBC correlations qCa CH CH CaO CH CaO NH CHa CHa qCa CHa CHa CHa CHa CH CH3 CH3 NH CaO CH CH2 CH CH3 CH3 N(CH3)2 CH2 CH2 CH2 H-3, H-14, H-15 H-17, H-18, H-19 H-3, H-17 H-3, H-4, H-7 H-30, H-31 H-7, NH-9, H-30 e NH-9, H-11 NH-9, H-10 H-10, H-11, H-13, H-16 H-11, H-14 H-13 H-14, H-16 H-11, H-15 H-3, H-18, H-19 H-3, H-19 H-3, H-18 e e H-28, H-29 H-25, H-26 H-25, H-26 H-23, H-26 H-23, H-25 e H-7, H-32 H-7, H-30 H-30, H-31 The sample was dissolved in CDCl3 with TMS as internal standard. a Assignments were based on HETCOR, HSQC-DEPT and HMBC experiments. b Assignments were based on 1He1H COSY, NOESY and TOCSY experiments. Table 3. The most important NOESY correlations of anorldianine are shown in Fig. 3 and in Fig. 4 the most important HMBC correlations are presented. 2.4. Confirmation of proline As a method of confirming the structure, the presence of proline after acid hydrolysis was investigated. Anorldianine (500 mg) was hydrolysed with 6 M hydrochloric acid (300 ml at 100  C during 21 h) in a sealed glass tube. After evaporation, esterification was performed with 300 ml of 1 M hydrochloric acid in methanol (at 100  C for 40 min). The methyl esters was N-acetylated by adding methanol:acetic anhydride (4:1) in 300 ml of water at room temperature and a reaction time of 60 min. The product was dissolved in methanol and analysed by GCeMS using an HP-5, 25 m fused silica WCOT column with a temperature 836 H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 Fig. 3. The most important NOESY correlations of anorldianine. program of 140  C for 3 min and 230  C for 6 min. The MS of the peak at retention time 8.2 min was identical to an authentic sample of N-acetylproline methyl ester. 3. Chemosystematic significance The systematic placement of Santalales has not been possible to deduce. In the past it has been associated with rosid taxa, of which many today are placed among asterids, e.g. the Apiales (Backlund and Bremer, 1997), and Icacinaceae within Aquifoliales (Kårehed, 2001). A recent phylogenetic study utilizing the sequences of the 18 S, rbcL, atpB and 26 S genes, however, show weak support for a placement of Santalales as sister group to the asterids (Soltis et al., 2003). Plotting the five structural subgroups on the proposed ordinal relationships of the core eudicots (APG, 2003), as in Fig. 5, raise interesting implications. The type 3 cyclopeptides have a seemingly restricted distribution, including only Santalales and Gentianales, and could hence be interpreted as supporting an asterid placement. This connection is strengthen if taking into account that Heisteria olivae Steyerm., have been shown to contain scopolamine (Cairo Valera et al., 1977), usually associated with the asterid family Solanaceae, and the hypothesis merits even further investigation. Fig. 4. The most important HMBC correlations of anorldianine. H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 837 Fig. 5. The ordinal classification of APG II (APG, 2003) with the presence of cyclopeptide alkaloids of substructures 1e5 given after the name. Thus far it has not been shown that Heisteria is parasitic, a feature otherwise common in Santalales. The possibility that presence of alkaloids due to such a lifestyle cannot be disregarded. Accumulation of host species secondary metabolites have been demonstrated for the mistletoes Amyema sp. [sic!] (Loranthaceae) (Boonsong and Wright, 1961), and Viscum cruciatum Sieber (Martı́n Cordero et al., 1993), as well as for Osyris alba L. (Santalaceae) (Woldemichael and Wink, 2002). 838 H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 The cyclopeptides, with a reduced macrocycle corresponding to our substructure type 2, have a restricted distribution within the rosid clade. Biosynthetic rationale implies that these reduced macrocycles can be thought of as precursors of the more abundant styryl-type macrocycles. The presence of a retained carboxylic function in the vignatic acids, placed in type 2, of Fabales (Sugawara et al., 1996), further suggest a biosynthetic series from the carboxylic acid via the saturated carbon chain to the styryl-function. 4. Conclusion The pattern of cyclopeptide alkaloids with macrocycles of 14 atoms suggests a closer relationship between Santalales and asterids than hitherto thought. The scarce data on their biosynthesis (Baig et al., 1993) may also be complemented from a selection of study organism on phylogenetic grounds. The low amounts of cyclopeptide alkaloids in plants, usually in the range of 0.0002e1% (Gournelis et al., 1997) imply that their presence in many cases may have been overlooked. The disjunct distribution could of course also be explained by differential gene expression, gene loss or simple convergence in response to similar ecological factors (e.g. discussed in Wink, 2003). A phylogenetic approach to investigate this suggests that the chemosystematic value of cyclopeptide alkaloids would increase with a greater sampling of asterid taxa. This while the biosynthetical investigations of these alkaloids may gain more from a greater sampling of rosid taxa, in particular, outside the already well studied Rhamnaceae. Acknowledgements We are very grateful to The Swedish Institute for a fellowship to H. R. E., and generous financial support from the International Foundation of Science (Grant in Aid F/3334-1). References APG [Angiosperm Phylogeny Group], 2003. Bot. J. Linn. Soc. 141, 399. Backlund, A., Bremer, B., 1997. Plant Syst. Evol. 207, 225. Baig, M.A., Banthorpe, D.V., Coleman, A.A., Tampion, M.D., Tampion, J., White, J.J., 1993. Phytochemistry 34, 171. Boonsong, C., Wright, S.E., 1961. Aust. J. Chem. 14, 449. Cairo Valera, G., de Budowski, J., Delle Monache, F., Marini-Bettòlo, G.B., 1977. Atti Accad. Naz. Lincei 62, 363. Dongo, E., Ayafor, J.F., Sondengam, B.L., Connolly, J.D., 1989. J. Nat. Prod. 52, 840. El-Seedi, H.R., Gohil, S., Perera, P., Torssell, K.B.G., Bohlin, L., 1999. Phytochemistry 52, 1739. El-Seedi, H.R., Ringbom, T., Torssell, K., Bohlin, L., 2003. Chem. Pharm. Bull. 51, 1439. H.R. El-Seedi et al. / Biochemical Systematics and Ecology 33 (2005) 831e839 839 Gournelis, D.C., Laskaris, G.G., Verpoorte, R., 1997. Nat. Prod. Rep. 14, 75. Kårehed, J., 2001. Am. J. Bot. 88, 2259. Martı́n Cordero, C., Gil Serrano, A.M., Ayuso Gonzalez, M.J., 1993. J. Chem. Ecol. 19, 2389. Soltis, D.E., Senters, A.E., Zanis, M.J., Kim, S., Thompson, J.D., Soltis, P.S., Ronse De Craene, L.P., Endress, P.K., Farris, J.S., 2003. Am. J. Bot. 90, 461. Sugawara, F., Ishimoto, M., Le-Van, N., Koshino, H., Uzawa, J., Yoshida, S., Kitamura, K., 1996. J. Agric. Food Chem. 44, 3360. Woldemichael, G.M., Wink, M., 2002. Biochem. Syst. Ecol. 30, 139. Wink, M., 2003. Phytochemistry 64, 3.