The chemistry of the compositae. Part XXXI. Absolute configuration of the sesquiterpene lactones centaurepensin (chlorohyssopifolin A), acroptilin (chlorohyssopifolin C), and repin

摘要

1976 1663The Chemistry of the Compositae. Part XXX1.l Absolute Configurationof the Sesquiterpene Lactones Centaurepensin (Chlorohyssopifolin A),Acroptilin (Chlorohyssopifolin C), and RepinBy Antonio G. Gonzalez,* Jaime Bermejo, Jose Luis Breton, Guillermo M. Massanet, Beatriz Domin-guez, and Juan M. Amaro, Departamento de Quimica Orgdnica y Bioquimica de la Universidad de LaLaguna, lnstituto de Quimica de Productos Naturales del CSIC, Tenerife, SpainDirect comparison of the sesquiterpene lactones chlorohyssopifolin A (6c). from Centaurea hyssopifolia, andcentaurepensin, from Centaurea repens showed them to be identical. The former was converted, by dehalogen-ation (zinc-copper couple) and subsequent saponification, into the deacyldihydrocynaropicrin (1 Oa), a derivativeof cynaropicrin (1 l j ) , the absolute stereochemistry of which for the centres of asymmetry C-1 ( a - H ) , C-5 (a-H),C-6 (p-H), and C-7 (a-H) has been rigorously established.This stereochemistry is the same as has been found todate in all the guaianolides for which absolute configurations are known, and is the opposite of that previouslyassigned to centaurepensin. Correlation between structure (6c) and the lactones acroptilin and repin. reported tohave structures (4c and d) (C-I P-H). suggests that in these substances the fusion of rings A and B is also cisC-1 or-H and not trans.IN previous papers on Centaurea hyssopifolia Vahl,2 wereported the isolation of five sesquiterpene lactones con-taining chlorine, which we called chlorohyssopifolinsA-E. Their structures, determined from spectroscopicdata and chemical transformations, were assigned as (lband a), (2b), and (lh and i), respectively.At the same time, two lactones new to the literaturewere described. One of them, ~entaurepensin,~ isolatedfrom C.repens L., was assigned the structure andabsolute configuration (3c), on the basis of X-ray dif-fraction analysis, and the other, acroptilin, isolated fromA croptilon repens (Centaurea picris) and from C. hyrcanicaB ~ r m , ~ was assigned the structure (4c). The formerproved to be identical with our chlorohyssopifolin A onthe basis of i.r. and n.m.r. spectra, m.p., mixed m.p. andchromatographic behaviour. The physical constants ofthe latter agree with those of chlorohyssopifolin C, sug-gesting that these two compounds are also identical.7Another sesquiterpene lactone, r e ~ i n , ~ was also isolatedfrom C.hyrcanica and assigned the structure and absoluteconfiguration (4d). The trans-fusion of rings A and B(lpH, 5aH) was assigned from a detailed analysis of its1i.m.r. spectrum. Acroptilin and repin can be convertedinto the same derivative (5a). The stereochemistry atC-1, -5, -6, and -8 must therefore be the same in bothlactones, since these centres are not affected in thechemical transformation.Acroptilin :4 1n.p. 197-199", [a]= 92.3'; chlorohyssopifolinC : 2 b m.p. 197-199, [a]= 100". The lH n.m.r. spectrum of thediacetate of acroptilin is identical with that of the diacetate ofchlorohyssopifolin C.We were unable to obtain an authenticsample of acroptilin for comparison.3 The footnote on p. 158 of vol. 3 of the Chem. SOC. SpecialistPeriodical Reports ' Terpenes and Steroids,' which states thatcentaurepensin and chlorohyssopifolin A are not identical, wasbased on the results of a study of ours, subsequently disproved,which we forwarded to the author prematurely.Part XXX, A. G. GonzAlez, J. Bermejo, G. M. Massanet,J . M. Amaro, and B. Dominguez, in the press.A. G. GonzAlez, J. Bermejo, J. L. Breton, G. M. Massanet,and J. Triana, Phytochemistvy, 1974,13, 1193.J. Harley-Mason, A. T. Hewson, 0. Kennard, and R. C.Pettersen, J.C.S. Chem. Comna., 1972, 460.Centaurepensin is widespread in the genus Centaurea.Recently, it has been found in C.nigra,6 C. soZstitiaZis,7and C. ZinifoZia (in the last together with acroptilin 8 ) .From the fact that chlorohyssopifolin A and centaur-epensin are identical $ and that chlorohyssopifolin C canbe converted into chlorohyssopifolin A,% we deduced thatboth chlorohyssopifolins have a secondary hydroxy-group on C-3 and a p-chloro-a-hydroxyisobutyryl group,rather than an OH on C-2 and an a-chloro-p-hydroxyiso-butyryl group, as we had erroneously assumed. Conse-quently, chlorohyssopifolins A, B, D, E, and C have thestructures (6c, a, h, and i) and (4c; with 1aH) and notthose previously reported.The absolute configuration indicated for centaur-epensin is worthy of comment. Its main feature is thetx-orientation of the C(7)--C(ll) bond, which is the re-verse of that found so far in all sesquiterpene lactoneswith absolute configurations rigorously established andwhich might originate biogenetically from some of thevarious possible conformations lo of the cation (7).Thep-orientation of the 'I,ll-bond, assigned on the basis ofchemical considerations and conf ormational analysis, hasbeen confirmed in many sesquiterpene lactones by X-raydiffract ion analysis. l1In the case of the centres C-1, -5, -6, and (probably) -11(when there is an asymmetric centre), the absolute con-figuration ascribed to centaurepensin is the opposite ofthat in all the guaianolides found in the Centaurea with aR. I. Evstratova, V. I. Sheichenko, and K. S. Rybalko,Khim. prirod. Soedinenii, 1973, 9, 161.R.I. Evstratova, K. S. Rybalko, and V. I. Sheichenko, Khim.fivirod. Soedinenii, 1972, 8. 451.A. G. GonzAlez, J. Bermejo, I. Cabrera, and G. M. Massanet,Anales de QuCm., 1974, 70, 74.J . Sakakibara, personal communication.A. G. GonzAlez, J. M. Amaro, J. Bermejo, and G. M.Massanet, unpublished results.9 S. M. Kupchan and J. E. Kelsey, Tetrahedron Letters, 1967,2863, and references quoted therein.lo J. B. Hendrickson, Tetrahedron, (a) 1959, 7 , 82; (b) 1963.10,1837; (c) W. Parker and J. S. Roberts, Quart. Rev., 1967, 21, 331;11 W. KZyne and J. Buckingham, ' Atlas of Stereochemistry,Chapman and Hall, London, 1974, pp. 97-1001664 J.C.S. Perkin Idefinitely established stereochemistry.12 Consequently, widely used in halogen elimination rea~ti0ns.l~ Depend-the lactone (8) derived from centaurepensin must be the ing on the reaction time, six different compounds, separ-mirror image of the corresponding lactone obtained from able by column chromatography after interrupting thecynaropicrin ( l l j ) , the absolute configuration of which at reaction at the most convenient moment (after 24 h),Hoq;)R CH2 0'0(10)a; R = Hc; R = CO$Me(OHI-~H2Cld; R = CO-CMe.CH2.0e;R=CO.CMe:CH2b; R =CO-CClMe.CH2-OH( 8 ) 0HOCH,CH2 0HO0(11 1f ; R = CO.CMeIOH). CH,g; R=COPrlh; R = CO.CMe(OEt ).CH2.0Hi ; R = C0.CMe(OH).CH2+OHj; R =CO-C(:CHZ).CH~*OHthe above-mentioned centres has also been rigorouslyestablished. l3In order to prepare the lactone (8), chlorohyssopifolinA was dehalogenated with zinc-copper couple, a reagentla (a) A.G. GonzAlez, B. Garcia, and J. L. Bretbn, Anales deQuim., 1971, 66, 1245; (b) 2. Samek, M. Holub, K. Vokak, B.Drodz, G. Jommi, A. Gariboldi, and A. Corbella, CoZZ. Czech.Chem. Comm., 1972, 37, 2611; (c) A. G. Gonzalez, J. Bermejo,and G. M. Massanet, Anales de Quim., 1973, 69, 1333; ( d ) A. G.GonzAlez, J. Bermejo, and M. Rodriguez, ibid., 1972, 68, 333;(e) W. E. Thiessen and A. Hope, Acta Cryst., 1970, 266, 554.were obtained. In all these compounds, saturation ofthe 11,13-double bond had taken place. This is re-flected in their lH n.m.r. spectra (Table) by the absenceof the characteristic signals due to the olefinic protons ofthe a-methylene-y-lactone grouping and the appearanceof a doublet (J 7 Hz) at 6 ca.1.25 due to the methyl groupl3 A. Corbella, P. Gariboldi, J. Jommi, 2. Samek, M. Holub,1* S. M. Kupchan and M. Maruyama, J. Org. Chem., 1971, 36.D. Drodz, and E. Bolszyk, J.C.S. Chem. Comm., 1972, 386.1187, and references quoted therein1976 1665a to the lactone carbonyl group. The first compoundformed is identical with dihydrochlorohyssopifolin A(Sc), obtained by hydrogenation (NaBH,) of (6c). Thereduction (Zn-Cu) of the conjugated methylenic doublebond is, therefore, stereoselective, giving rise to anH-33.904.02q(4.5, 6.5)4.204.204.204.554.554.554.55apart from confirming the position of the secondaryhydroxy-group on C-3, suggests that the absolute con-figuration of the previously mentioned centres in chloro-hyssopifolin A is the same as in cynaropicrin and not theopposite, as it should be if the absolute configurationlH N.m.r.data (6 values; solvent CDC1,; 60 MHz; J/Hz in parentheses)H-64.804.684.904.704.654.504.204.204.184.15H-8 b5.225.165.35.205.205.205.205.205.20H-13 05.70 (3.5)6.05 (3.5)5.62 (4).6.24 (3.5)5.70 ( 3 4 ,6.07 (3.5)1.25 (7)1.20 (7)1.32 (7)1.25 (7)1.25 (7)1.28 (7)1.25 (7)H-145.03,5.215.02,5.244.99,5.155.02,5.174.90,5.124.95,5.055.10,5.155.05,5.155.00,5.155.00,5.12H-153.05d (5),3.15d (5)3.14d ( 4 4 ,3.34d (4.5)3.87d (12),4.30d (12)3.88 (12),4.23 (12)3.87 (12),4.25d (12)5.255.35,5.435.35,5.425.37,5.425.32,5.403’-H,3.76d (11.5),3.96d (11.5)2.87d (6),3.22d (6)3.78d (12),3.99d (12)3.63d (11.5),3.88d (11.5)5.65 (2),6.163.59d (12),3.86d (12)5.65 (2),6.15s1.45s1.17d (6.5)2‘-Me d1.54s1.66s1.60s1.50s1.96d (2)1.50s1.95d (2)1.45s1.17d (6.5)t Solvent (CD,),CO; 100 MHz.a q (10.5, 9).b m. e d. d Acyl group designated C(l’)-C(2’)Me-C(3’).Data from ref. 5 (100 MHz).a-oriented methyl group, as in the reduction withNaBH4.15 The n.m.r. and mass spectra of the secondlactone formed (9e) clearly show that it is an ester ofmethacrylic acid {prominent mass spectral peaks atM f - C,H60, and m/e 69 [COC(:CH,)-CH,]; n.m.r.signals for a vinylic methyl group (6 1.95, d) and a con-jugated methylene group (6 5.65 and 6.15)).In the suc-ceeding four lactones the C(4) (OH)*CH,Cl group has beenconverted into an exocyclic methylene group, C(4):CH2,which is reflected in their n.m.r. spectra by the absence ofthe quartet (AB system) centred at 6 ca. 4.05 and thepresence of two new signals between 6 5.3 and 5.5.These four lactones differ from each other solely in theacyl group, since saponification of each yields the samelactone (loa). The n.m.r. and mass spectra indicate thepresence of an a-hydroxy-p-chloroisobutyryl group in(1Oc) ( 6 1.50 [s, CH3*C(OH)] and 3.59 and 3.86 (ABsystem, CH2Cl) ; M+ - C4H,C10,}, an a-hydroxyiso-butyryl group in (10f) ( 6 1.45 [6 H, s (CH,),C(OH)];M+ - C,H803 and m/e 59 (C,H,O)}, a methacryloylgroup in (loe), and an isobutyryl group in (log) ( 6 1.20[6 H, d, (CH,),CH] ; M+ - C4H,0, and m/e 71 (C,H,O)).After a sufficiently long reaction time, all the other fivelactones were converted into compound (log).The lactone (lOa), obtained by the saponification oflactones ( ~ O C , e, f, and g), turned out to be identical withthe deacyldihydrocynaropicrin prepared from deacyl-cynaropicrin (1 la) by stereoselective reduction of theconjugated methylene group. This surprising result,assigned to centaurepensin were correct.We feel, there-fore, that a reconsideration of the latter might well be in-teresting.In the light of the transformation of chlorohyssopifolinA into chlorohyssopifolin C (acroptilin) and the con-version of acroptilin and repin into the derivative (5a), itmay be assumed that the absolute configuration of thesetwo lactones at C-1 is the same as that of (10a) and, hence,that the fusion of rings A and B is cis (laH, 5ctH), and nottrans as rep~rted.~EXPERIMENTALM.p.s were determined with a Kofler hot-stage apparatus.1.r.spectra were taken for solutions in chloroform, U.V.spectra in ethanol, and 60 MHz lH n.m.r. spectra in CDC1,(Me,Si as internal reference). Optical rotations weremeasured for solutions in chloroform. Silica gel (0.05-0.2mm) was used for column chromatography.Reduction of Chlmohyssopifolin A with Zinc-CopperCoupZe.-A mixture of chlorohyssopifolin A (1.6 g), zinc-copper couple (40 g), and ethanol (250 ml) was refluxed for5 days. After 24 h,six different compounds had been formed.Most of themixture was filtered through Celite and concentrated undervacuum; the residue (600 mg) was chromatographed on acolumn of silica gel. The rest of the mixture was left toThe reaction was monitored by t.1.c.l6 S. B. Mathur, S. V. Hiremath, G. €3. ICulkarni, G. R. Kelkar,S. C. Bhattacharyya, D. Simonovic, and A. S. Rao, Tetrahedron,1972, 21, 35561666 J.C.S. Perkin Ireact for another 4 days, after which all six compounds hadbeen converted into compound (1 Og) .Elution with benzene-ethyl acetate (1 : 1) produced firsta mixture of two lactones, which could be separated bypreparative t.1.c. with benzene-ethyl acetate (8 : 2) (twoelutions). The compound of higher R p value was recrystal-lized from acetone-petroleum to give 15-chloro-3,Pdi-hydroxy-8-methacryZoyZoxygzcai-10( 14)-en-l2,8oluctone (9e) ;m.p.163-167"; [aID 32.7 (c 1.1); v,, 3 540 (OH), 1770(y-lactone), 1710 (ester), and 1635 cm-1 (C=C); mle (no&I+) 298 (M+ - 86, 5%) and 69 (100%) (Found: C, 59.7;H, 6.45; C1, 9.0. C,,H,,C10, requires C, 59.3; H, 6.5; C1,9.25%).The compound of lower RF value (an oil) (10e) could notbe crystallized; [aID 38.4; vmX. 3 490, 1 760, 1 720, and 1 620cm-1; m/e 332 (M+, 5%), 246 (Mf - 86, 27), and 69 (100).Acetylation with acetic anhydride-pyridine produced aliquid monoacetate.Two more compounds eluted from the column wereseparated by fractional crystallization from acetone-petroleum. One of them was identical with 1 l,l&dihydro-chlorohyssopifolin A (9c) (m.p., mixed m.p.and i.r.spectra); m.p. 167-170"; [a]= 9.8 (c 1.0 in MeOH); v-3 530, 1 770, and 1 730 cm-l; wz/e (no M+) 298 (M+ - 138,41%) and 93 (100) (Found: C, 52.45; H, 6.1; C1, 16.35.Calc. for CI,H2,C1207: C , 52.3; H, 5.95; C1, 16.05%).The other was the final product, 3-~~ydroxy-&isobutyryZ-oxyguuiu-4( 15), 10( 14)-dien-12,6-oZuctone (log) ; m.p. 203-208"; [aID 61.8 (G 3.56); vmX. (KBr) 3 240, 1 780, 1 715, and1 630 cm-1; m]e 334 (M+, 22%), 246 (M+ - 88, 75), and 71(100) (Found: C, 67.9; H, 7.85. ClgH2605 requires C, 68.3;H, 7.85%).After this, a mixture of (9c) and the p-chloro-a-hydroxy-isobutyryl com9ound (1Oc) was eluted and separated byfractional crystallization from ethyl acetate-petroleum;(1Oc) is a liquid, [a], 90" (c 1.3) (Found: C, 59.5; H, 6.65;C1, 9.1.C,,H2,C106 requires C, 59.3; H, 6.5; Cl, 9.25%);vn- 3 520, 1765, 1730, and 1630 cm-l; mle 384 (M',21%), 246 (M+ - 138, 66), and 93 (100).Finally, an oily compound, the a-hydroxyisobulyrjd de-rivative (lor), was eluted. T.1.c. showed i t to be homo-geneous; [a],, 45.5 (c 1.56); vmx. 3 580, 1 765, 1 720, and1 635 cm-1; m/e 350 (M+, 7%), 246 ( M f - 104, IS), and 59(100) (Found: C, 64.9; H, 7.55. ClgH2,06 requires: C,65.15; H, 7.5%).Dihyd~ockZorohysso~~~oZin A (9c) .-Sodium borohydride(300 mg) was added to compound (6c) (50 mg) dissolved inmethanol (40 ml) and the solution was stirred for 10 min at0 "C. The solution was evaporated and the residue acidifiedwith aqueous 5 yo hydrochloric acid and continuouslyextracted with ethyl acetate.The resulting solid waschromatographed on silica gel, yielding the product (Sc),m.p. 168-170".Dihydrodeacylcynaropicyin from Compounds (log) and(1 la) .-A solution of compound (log) (100 mg) in methanol(6 ml) was treated overnight with aqueous 5% potassiumcarbonate (10 ml) at room temperature. The methanol wasevaporated off under vacuum and the residue acidified with5% sulphuric acid (10 ml) and extracted with ethyl acetate.The crude product was eluted from a column of silica gelwith benzene-ethyl acetate (3 : 2). Crystallization fromethyl acetate-petroleum gave a solid, m.p. 135-137"; [a],72.8 (c 1.34 in MeOH); v,,, (KBr), 3 330, 1 730, and 1 630cm-1; mle 264 (M+, 13%), 246 (9), 218 (9), and 200 (11)(Found: C, 68.4; H, 7.7. C,,H200, requires C, 68.15; H,7.65%), identical with that obtained by saponifying a mix-ture of (lOc, e, f, and g) under the same conditions.A solution of compound (1 la) (80 mg) in methanol (50 ml)was treated with sodium borohydride (480 mg) . Under thesame conditions as above, a solid (60 mg) [m.p. 135-137';[a]= 74.8 (c 1.02 in MeOH)], was obtained, identical with thatproduced from (log) (mixed m.p., i.r. and n.m.r. spectra, andchromatographic behaviour) .We thank Professor Gariboldi for a sample of deacyl-We also thank Professor Pascual for help and cynaropicrin.advice in preparing this manuscript in its English form.[S/lSS Received, 26th January, 1976