Synthesis of some stable ozonides with anti-malarial activity

摘要

J. CHEM. SOC. PERKIN TRANS. I 1992 325 1 Synthesis of some Stable Ozonides with Anti-malarial Activity Luiz-Claudio de Almeida Barbosa,a David Cutler,b John Mann,b Geoffrey C. KirbyC and David C. WarhurstC a Departamento de Quimica, Universidade Federal de Vicosa, 36570, Vicosa -MG, Brazil Department of Chemistry, Reading University, Whiteknights, Reading RG6 2AD, UK Department of Medical Protozoology, London School of Hygiene and Tropical Medicine, Keppel Street, London WCIE 7HT, UK We describe the synthesis of variously substituted 8,9,10,11-tetraoxatricyclo5.2.1.1 2*6undecan-4-ones. Several of these stable ozonides exhibited activity (IC,'s of 10-20 pg ~rn-~)against a c h Ioroquin e -resist a nt st r a in of the ma Ia ria parasite Plasm0 dium falciparum.Several years ago we synthesised (completely by accident) the . ozonide 1by ozonolysis of 2,4-dimethyl-8-oxabicyclo3.2.lloct-6-en-3-one 3.' This ozonide was unusually stable and survived complete structure elucidation including an X-ray structure determination. The recent intense research activity centred OR upon the anti-malarial agent artemisinin 2 (quingaosu, ex. Arternisiu unnua), prompted us to carry out a simple computer modelling comparison of structures 1 and 2t (see Fig. 1). The T;-o 0 . R '* 1 R=H 0 2 0 Fig. 1 Minimised structures of the ozonide 1 and artemisinin 2. , Qo R '* 3 R=H 5 4 R = alkyl O,R 6a R=H 7 b R=CHZPh c R=Me d R=C(S)SMe root mean square difference calculations, based on all five oxygens of both structures, gave 'fits' of ca.0.5 A. This ii -Po 6Me1ii c1 Me, iii 0 0 OMe .OMe 10 ----OMe -21 11 -0Me 9 )OMe160 '*d Scheme 1 Reagents and conditions: i, 10 K,CO,-MeOH (98); ii, BrCH,CH(OMe),-benzene (64); iii, LDA/NaI/THF-HMPA (1 :1); iv, ozone-DCM at -5 "Cfor 30 min (ca. 70) encouraged us to have compound 1 evaluated for antimalarial activity which it was found to exhibit to high degree (ca. 10 pg cm-')).$ The synthesis and evaluation of other, more complex ozonides is the subject of this communication. Most of the substrates for ozonolysis were prepared from 1-substituted 2,4-dimethyl-8-oxabicyclo3.2.loct-S-en-3-ones4,available from the cycloadditions of the oxyallyl cation 54 with the requisite furans.Ozonolyses were usually carried out in dichloromethane (DCM) at -5 to 0 "C, and the ozonides either crystallised directly from the reaction mixture or were isolated, following evaporation of the DCM, and recrystallised. All the ozonides were apparently stable at room temperature, and gave clean m.p.s 100 "C.In their 'H NMR spectra they exhibited signals at 6 5.4-5.75 (2 s, l-H and 7-H), together with the other expected proton signals. Interestingly, when the ketone 3 was reduced (NaBH, in ethanol) to the axial alcohol 6a, and the benzyl and methyl ethers prepared 6b, c, these gave intractable polymer t We thank James Crabbe for the computer modelling carried out using Desktop Molecular Modeller, version 2, OUP, 1991. $ Compounds were tested in vitro against Plasmodium falciparum (Kl strain; chloroquine resistant; from Thailand) at concentrations up to 500 pg ern-,, and the results were recorded as iC,:s in pgern-,.See ref. 3 for details of the experimental protocol. 3252 J. CHEM. SOC. PERKIN TRANS. I 1992 materials but no stable ozonides. Similarly, when the carbonyl group was removed completely, via the xanthate 6d and reaction with tributyltin hydride, this too gave only polymeric material upon ozonolysis. We suggest that an interaction between an oxygen lone pair and the carbonyl x electron system helps to reduce the nucleophilicity of the bridgehead oxygen, thus stabilising the ozonides against destruction by this nucleophilic centre.In addition, we have prepared a number of ozonides that more closely resemble the lactol ethers 7 derived from artemisinin, which have been shown to have greater anti- malarial activity.' Synthesis of the bicyclic alkenes 8 and 9 and their respective ozonides 10 and 11 (Scheme 1) are repre-sentative. These compounds had the expected structural and spectroscopic properties,* and their minimised structures more closely resembled 3 than did compound 1. All the ozonides were evaluated in vitro for anti-malarial activity using a multi- resistant strain of Plasmodium falciparum from Thailand.3 The results are shown in Table 1. Clearly the simpler compounds possess the best activity, and there is a suggestion that increasing the lipophilicity of the compounds could be bene- ficial.Further compounds are being prepared for evaluation. Typical Experimental for 0zonization.-Ozone was passed through a solution of the alkene (2 mmol) in light petroleum (35 cm3) held at 0 "C.After ca. 10 min the reaction was judged * Compound 8 had the following 'H NMR characteristics (CDCl,, 400 MHz): 6 0.94 (d, 3 H, J 7 J values in Hz, 2-Me), 1.24 (s, 3 H, l-Me), 1.53 (d, 1 H, J 13.7, Hi), 2.29 (dd, 1 H, J 13.7 and 9.6, Hh), 2.84-2.95 (m, 1 H, Hg), 3.35 (s, 3 H, OMe), 3.57 (d, 1 H, J 12.2, Hf), 4.12 (d, 1 H, J 12.2, He),4.78(d,1 H,J4.1,Hd),4.86(dd, 1 H, J5.2and 1.8, Hc),6.11 (d, 1 H, J6.1, Hb) and 6.36 (dd, 1 H, J6.1 and 1.9, Ha). Compound 9 had the following 'H NMR characteristics (CDCl,, 400 MHz): 0.96 (d, 3 H, J 7,2-Me), 1.16 (s, 3 H, 1-Me), 1.54 (dd, 1 H, J 12.8 and 2.2, Hi), 2.142.19 (m, 1 H, Hh), 2.942.97 (m, 1 H, Hg), 3.48 (2, 3 H, OMe),3.91and4.00(2d,2H,J12.7,HeandHf),4.59(dd,1H,J9.6and 2.3, Hd), 4.87 (dd, 1 H, J5.1 and 1.8, H"), 6.07 (d, 1 H, 56.1, Hb) and 6.38 (dd, 1 H, J 6.1 and 1.8, Ha).The ozonides 10 and 11 possessed similar NMR spectra to the parent alkenes except for the disappearance of the alkene signals, and appearance of the expected singlets in the region 6 5.4-5.7. Mass spectral data (chemical ionisation) for the ozonides showed M+ +NH, at 304.1396 for 10 and 304.1400 for 11 (Cl3Hl,O7 +NH, requires 304.1390). Table 1 Anti-Malarial activities of ozonides I (R = H) 18.2 1 (R = Et) 11.9 (R = C7H15) 26.2 I (R = CH,OAc) Inactive (500) 1 R = CH(OMe)Ph 10 Inactive (500) 250 11 Inactive (500) "For comparison: artemisinin 2 IC,, pg cm-, and quinine IC,, 0.18 pg ~m-~.Each result is the mean of two determinations; errors were lo). to be complete (by TLC analysis), and the solvent was removed under reduced pressure to leave the ozonide as a white crystalline solid, which could (usually) be recrystallised from dichloromethane-ether(1 :1). Acknowledgements We thank the British Technology Group and the Wellcome Trust for generous financial support and Dr. Oliver Howarth for numerous NMR studies. David Warhurst is supported by the PHLS. References 1 W. J. Cummins, M. G. B. Drew, J. Mann and E. B. Walsh, J. Chem. SOC., Perkin Trans. I, 1983, 167. 2 D. L. Klayman, Science, 1985,228,1049;W.4. Zhou and X.-X. Xu, in Studies in Natural Products Chemistry, ed. Atta-ur-Rahman, 1989, vol. 3, Elsevier, Amsterdam; A. R. Butler and Y.-L. Wu, Chem. SOC. Rev., 1992, 21, 85. 3 R. M. Ekong, G. C. Kirby, G. Patel, J. D. Phillipson and D. C. Warhurst, Biochem. Pharmacol., 1990,40,297. 4 J. Mann, Tetrahedron, 1986,42,4611. 5 A. J. Lin, L-q. Li, S. L. Anderson and D. L. Klayman, J. Med. Chem., 1992,35, 1639 and refs therein. Paper 2/05889H Received 4th November 1992 Accepted 4th November 1992