13C nuclear magnetic resonance spectra of some limonoids. Part I. The structure of procerin, an extractive fromCarapa procera

摘要

1974 43713C Nuclear Magnetic Resonance Spectra of Some Limonoids. Part I.The Structure of Procerin, an Extractive from CarapaproceraBy David A. H. Tay1or.t Department of Chemistry, University of York, York YO1 5DDThe structure (8) is suggested for procerin, a limonoid of the utilin group extracted from the timber of Carapaprocera (Meliaceae), on the basis of chemical and spectroscopic properties, including 13C n.m.r. spectra. Tablesare given of the 13C n.m.r. spectra of a number of other limonoids.DURING studies in Ibadan on extractives from theMeliaceae, many compounds have been obtained whichare too complex, or obtained in too small quantities, forus to have been able to determine their structures bychemical and spectroscopic means, including lH n.ni.r.One method of dealing with such problems is directcrystallographic analysis, without a heavy atom.Thedetermination of the structure of utilin (la) is an ex-ample. Unfortunately this method is not generallyapplicable; even when it is, it requires an able crystallo-graphic collaborator.It was hoped that application of 13C n.m.r. methodsmight answer at least some of these problems, and forthis purpose a collection of spectra of limonoids of knownstructure has been made. The carbon resonances inthese were assigned on the basis of the residual splittingin off-resonance decoupled spectra, data from theliterature, and the comparison of the spectra of closelyrelated compounds. The spectra of a number of deriva-tives of methyl meliacate (2) are given in Table 1 (fornomenclature see ref.2). Table 2 contains the spectraof some more complex compounds with the 1,29-cyclo-meliacate structure (3) typical of utilin. Some deriva-tives of methyl angolensate (4) are listed in Table 3, andderivatives of meliacolide (5) (for nomenclature see ref. 3)in Table 4. In some cases where resonances of thesame multiplicity have similar chemical shifts, there isno certain basis for assignment, and the Tables in thesecases give probable assignments only. Some of thesedata have been used already to reach a decision betweentwo structural proposals for fissinolide.4This paper reports the use of 13C n.m.r. in structuraldiagnosis of procerin, an extractive first obtained inIbadan from the bark of Carapa procera.5 The 13C n.m.r.spectrum showed resonances due to 41 carbon atoms.Integration of the lH n.m.r.spectrum indicated 52hydrogen atoms. The highest mass peak in the massspectrum (m/e 848-3099) corresponds to the formulaC4,H,,O1,. The lH and 13’C n.m.r. spectra showed thepresence of two methoxycarbonyl groups, the usualp-substituted furan ring, and one tertiary hydroxy-group. Quantitative hydrolysis gave 8 mol of acidper mol of procerin, and since the 13C spectrum showedonly seven carbonyl carbon atoms this suggests thepresence of a hydrolysable orthoester system, as int Present address: Department of Chemistry, University ofH. R. Hamson, 0. Hodder, C. W. L. Bevan, D. A. H.E. K. Adesogan and D. A. H. Taylor, J. Chem. SOC.( C ) , 1968,Natal, Durban, South Africa.Taylor, and T. G. Halsall, Chem. Comm., 1970, 1388.1974.utilin. This was confirmed by the presence of a 13Cresonance at 118.6 p.p.m., characteristic of an orthoestercarbon atom (cj. utilin, 119.3; bussein, 119.1) and aresonance due to an orthoacetate methyl group at 8 1.82in the lH spectrum (cj. utilin, 1.72; bussein 1-63). ThelH spectrum also shows four acetate methyl groupsEtMe-CH. I CO, c. I :6 -“OR202C R’( 1 )1 /9 a; R =-c-c’ , R ~ = ACI Me MeMe Me ‘-A -C ; R’=-C= C /H , R2= PriCOI Me Me( 3 ) co@ 0 R,*od R2 R2( 5 1C02Me( 4 13 1 2 a; Rf = 0, $= H ,a - 0 A c , R = R i H ; a’>’b; R1= 0 ,R2= H,a-OAc ,R3= R k HC ; R’= R2= H,a-OAc,R3=OAc,RL=Hd ; R1= R2= H,a-OAc,R R4= OACe ; R1= R2= 0, R3= RL= H; lactone6 2.05, 2.15 (6H), and 2-36, three nuclear methylgroups (8 1.26, 1.16, and 0-82), and a primary methylgroup (8 1.15).Analysis of the volatile acids from thea ; R = 0, R = Hb; R1=O, $= OHC ; R1= H f3-OH, R2=Hd; R1= 0 , R2= H,3)4-seco-f ; R’= R ~ = 0, R: O A ~ , RL= HN. S. Ohochuku and D. A. H. Taylor, J. Chem. SOC. (C),D. A. H. Taylor and F. W. Wehrli, J.C.S. Perkin I , 1973,C . W. L. Bevan, J . W. Powell, and D. A. H. Taylor, J. Chem.1969, 864.1599.SOC. (C), 1963, 980J.C.S. Perkin Ihydrolysis (by lH n.m.r.6) identifies this as due to apropionic acid residue.Together, these oxygen functions, two methoxycarb-onyl groups, one furan ring, one tertiary hydroxy-group,four acetates, one propionate, and one orthoacetate,account for all the oxygen atoms in procerin.Togetherpresent has been opened as in C.O.C. (6a), a transforma-tion product of mexicanolide.7The lH n.m.r. spectrum of procerin shows four signalsdownfield of those due to the methoxycarbonyl groups,in addition to those due to the furan ring. These are allsinglets (6 5.35, 5.64, 5.80, and 6.74). Since the TABLE 1Derivatives of methyl meliacate (2) ; 13C n.m.r. spectra (solvent CDCl, ; 6c values ; Me,Si standard)Carbon atom12345678910111213141516172021 *2223 *30OCH,MeCOCCCH,-.)I ,-Im -29 X-. rl212.758.0210.949.450-532.3173.5133-840.254.318.628.837-8125.332.9169.780.6120.4142.7109.8141.536.452.121-917.817.817.46 6 5 : :213.0 212.957.1 81.7211.4 108.650.0 41.948.8 53.332.6 32.5173.5 174-433.1 63.340.0 41.552.2 62.318.6 19.626.8 32.038.2 36.9164.2 165.6113.9 118.6168.3 166.581.0 81.1119.9 119.9143.0 143.1109.8 110.0141.2 141.437.7 47.052.2 52.022-0 22-619-0 21.718.6 18.818.0 16.8y1 23,.20 q".H210.655.3209.848.945.431.9173.4124-6150.952.621.930.136.9157.6111.3165-280.411 9.8143.0109.8141-234.052.224-521-416.016.4c0 c)Y217.662.378-438.248.033.4174.1131.740.862.918.829.238.1127.733.9170.180.6120.6142.7109.8141-633-461.9213.348.978.237.446-732.3173.8125.5151.951.222.330.936.9158.1109.9166.480.7120.0143.0110.0141.230-252.0214.053.878.438.748.932.7173.7136-240.562.021.633.037.6160.6112.4164.979.7120.2143.1110.2141-4128.852.021 7.478.085.739.052.233.3174.0133.040.952.218.829.238-3125.733.5169.980.6120.6142.9109.8141.844-262-2A0c,0 3 ce218.677.386.938-747.133.1174.257.841-650-719.930.035-246-433.1170.276.7120.8143.1109.8141-635.152.2169.6 169.6 170.1 169.6 170.023.2 25.9 22.2 22-7 23-621.1 26.2 22.0 21.2 22.220.5 20.7 21-2 19.9 20.718.0 16.7 20.7 18.1 19.916.7 16.0 15.6 16.8 17.7Other (extranuclear)* These rows may be intercl-m - 3?$-lrl0c, 2m2 16-856-577-438.148.433.0174.1138.441.649.922.430.136.845.234.4170.877.6;120.7142.9109.7141-8122-862.2169.622422.020-420.416.8ianged.L8i? ->.c, 5I-!rs, m+ c 4> x 320 m217.066-479.238.748-732.7174.060.141.449.219-229.936-946.233.9169.477.0120.7142.9109.7142456.752.1168.7169.221-820.7( ."816.722-4( 4R?8,8p.'4PIc, ce 2xc,u -r: A m214.066.878.839.248.633.4174-260.442.548-2 z 9.433-036.446.933.7172.077-0120.1143.1110.2141-063.652.3169.826-222-620.720.716.6?..I w -mX$cc,AeCt,213.267.8210.049.453.473.2174.9134.144.454.518.628.738.1126.033.1169-680.6120-4142.9109.7141.236.452.1ri-d 221.020-517.617-66Qe-7Lw I 2wTlh+- pCQ212.749.480.338.446.573.6175.734.944-951.218.626.838.2164.3113.0170.281.2119.9143-1109.8141-235.453.3169-823.823.120.718.417.1- La .-ldY 2;$'X?e ar(79.244-0219.846.350-233.8174.835.040.342.817.527.138.1164.8113.2170.181.2119.9142-9109.8141-235.761-927-725.621.717.9216.5 218.967.6 66-678.6 77-439.0 38-149.0 47-772.8 33.4175.9 173.8138.3 133.845.6 41-160.4 51.721-2 20.129-6 36.736.7 43.145.1 133.734.6 35-6168.5 172.076.7 79.0121.4 123-9143.1 143-0109.2 110.7140-6 142.1123-6 34-253-2 56-2,51-8,51.7166.9 170-423.0 23-722.8 21.121.2 20.816.5 19.414-6 16.611.6139.0(4(4127-7with the normal C,, nucleus of a limonoid, they alsoaccount for all the carbon atoms.It is significant that procerin contains no lactonesystem.Either therefore it belongs to the carbocyclic-ring-D group of limonoids, or else the lactone ring usually(I D. H. Calam and D. A. H. Taylor, J. Chem. SOC. (C), 1966,949.n.m.r. spectrum shows no double bonds (except those inthe furan ring), thebe must correspond to the carbinolprotons of four secondary ester groups, the other oxygenfunctions being tertiary.Similar carbinol singlets are characteristic of sub-7 E.K. Adesogan, C. W. L. Bevan, J. W. Powell, andD. A. H.Taylor, J. Chem. SOC. (C), 1966, 21271974 439stituents in the bicyclononane system of methyl meliacatederivatives (e.g. 2-hydroxyfissinolide and utilin l) andthey are not known to be produced by any other nucleus.We therefore assume as a working hypothesis the presenceTABLE 2spectra (solvent CDCl, ; aC values: hle,Si standard)Derivatives of methyl 1,29-cyclomeliacate (3) ; 13C n.1n.r.Carbonatom12345678910111213141616172021 *2223 *2930OCH,CO,RMeCO,MeC(0H) 0,Other (extranuclear)CCH,Utilin84.274.984.151-141.034.1174.389.687.244.164.630-138-784.637.6168.977-1120.6142.3110.1141.440.070.051.6170-6172-4168.0119.3(la)67*8(~)67*8(d)40*3(d)26.1 (t)20-318.717.717.016.616.415.113.411.3Busseinhydrolysisproductacetatemethylester (7b)83.977.782.945.642.433.6172.686-886.146.669.368.738.636.427.6169-276.7121.3143.2109.8140.639.970.161-7170-5170.6119.677.121.121.020.719.916.815-614.2Bussein84.379.683-044-943.633.3172.790.884.945.770.068.946.741-4176.2168.773.9122.0142.7109.9141-339.970.051-8170.5169.2168.4119.177.1183*3(~)36*9(d)30*3(d)26*6(t)20-920.720-720.319.718.417.816.816.914.211.6(74Procerin84.283.179.146.437.033.3172.689.088.246.922.827.844.886.761.9171.878-4122.9142.6109.7141.238.671.162.1, 61.6170.9, 169.6169-4, 169.2168.1118-6(8)27*0(t)21.421.421.021.018.416-413.411.68.6* These rows may be interchanged.of a bicyclononane system.The partial synthesis ofmexi~anolide,~ which is believed to follow the biochemicalroute, depends for the formation of the bicyclononanesystem on the presence of a lactone ring D. It thereforeseems unlikely that we have a bicyclononane containinga carbocyclic ring D, and more likely that procerin hasan opened lactone ring similar to that in C.O.C. (6a).This conclusion is supported by the lH n.m.r. spectrumof procerin.In the lH n.m.r. spectra of all the ring D* E. K. Adesogan and D. A. H. Taylor, J. Chem. SOC. (C),1970, 1710. ' T. D. Connollv. I. M. S. Thornton. and D. A. H. Tavlor.TABLE 3Derivatives of methyl angolensate (4) ; 13C n.m.r.spectra (solvent CDC1,; 6~ values; Me,Si standard)Carbonatom12346678910111213141616172021 *2223 *30OCH,CCH,Methyl MethylMethyl 6-hydroxy- 3 P-hydroxy-angolensate angolensate angolensate77.8 79.2 76.040-4 40.1 37.7212.7 212.2 72.549.0 49.8 50450.9 61.9 60-533-7 73-3 34.0174.3 177-1 174.6146.4 146.6 146.643.9 48.6 32.546.0 46.7 43.924.7 25.1 24.030-1 39.7 29.642-4 42-4 39.881.1 81.3 80.034.6 34-6 37.7170.4 170.6 170-480.2 80.3 80.0121.6 121.6 121.0143.3 143.3 142.7110.7 110.7 109.9141-4 141.3 140.7112.3 112.3 110.862-9 64.2 51.926-6 25.9 27.122.6 24.9 22.022.6 24.6 14.414.8 16.0 13.9* These rows may be interchanged.(44 (4b) (44TABLE 4Methylivorensate( 4 473441-5169-683.851.934.1173.3146.743.447.624.032.041.681.336-7170.079.3120.7142-6109.9140.9111-956.228.822.622.013-8Derivatives of meliacolide (5) ; I3C n.m.r.spectra(solvent CDC1,; 6~ values; Me,Si standard)Dihydro-Carbon Gedunin gedunin Khivorinatom12346678910111213141616172021*2223*MeCOCCH,167.0126-9203-844.039-614.9 t73-242-646.040.017.7 t18.3 t38.769.866.9167.478.2120.6143.0109.8141.4169.827.226.023.321-721-019.733.747.8216.746.638.915.073.742.044.037.417.4 723-8 t38.869.856-7167.678.4120-6142-9109.8141.1169-826.126.921-120-918-015.7(6472.326.6 t76.936-937-022-3 t73.842-2236.340.714.4 t26.1 t38-869.766.4167.678-4120.6143.0109.9141.1169.9169.7169-627-421.621.321.121-118.317-316.411p-Acetoxy-khivorin( W76-126-2 t76.636.037.822.9 t73.041.939.741.667.037-637.868.764.7167.178.3120.0143.3109.8141.2169-9169.6169.227-421.821.621.421.221.220.117.6( x 2)( x 2)7-OXO- 7-OXO-gedunin khivorin -(64166.0126-3203.246.247.636.7 t208.363.463.740-017.1 t32.1 t37.765.764.6166.978.0120.3143.1109.8141.027.020.920.719-717.4w72.526.8 t76.837.346.8209-462.944.440.616-9 t32-2 t37.366.463.3167-378.2120.4143.7109.8141.0169.6170.136-7 t36.921.121.120.820.816-916.8. I ' Chem: Comm., 197f,'17.* These rows may be interchanged. t Assignment doubtful440 J.C.S. Perkin Ithis structure (utilin, 6 2-77 and 2.17, J 10 Hz; bussein,6 3-15 and 2.55, J 10 Hz). The doublet of doubletswhich commonly arises from the 15-protons has a largercoupling constant, typically 19 Hz. We thereforeconsider that these further observations strongly supportthe hypothesis that procerin contains a bicyclononanesystem, and that this is present in a 1,29-cyclomeliacateskeleton (3), which has been modified by the opening ofring D to give a 8-acyloxy-methyl ester.To identifyprocerin completely, it is necessary to locate in thisnucleus the oxygen functions already mentioned, i.e.four secondary esters (including that already located atC-17), one tertiary ester, an orthoacetate, and a tertiaryh ydroxy-group.Normally, orthoacetates are resistant to alkalinehydrolysis, as in bussein. The orthoacetate in procerinis hydrolysed, as shown by the results of total acidtitration. The orthoacetate in utilin is similarly hydro-lysed; this is because one of the points of attachmentis C-14, p to the lactone carbonyl group, where it can belost by elimination to give an +-unsaturated ester anda hemi-orthoacetate, which is hydrolysed by alkali.We therefore assign one oxygen of the orthoacetate inprocerin to C-14, to allow the same mechanism to operatein this case.We consider the other points of attachmentto be C-8 and C-9, as in utilin, since this is known to givea stable, naturally occurring group.We assign two of the secondary esters to C-3 and C-30,in accord with the usual substitution pattern. For theseto give rise to IH n.m.r. carbinol singlets, it is necessaryfor tertiary oxygen atoms to be located at C-2 and C-9,and a tertiary oxygen atom is also expected a t C-1,resulting from the ring A bridging. From comparisonwith bussein, the singlets a t 6 5.35 and 5-64 may beascribed to H-3 and H-30, respectively.This leaves one secondary ester group to be located.Three lines of evidence suggest that this is at C-14.(a)This is the only location left in the nucleus where an esterwould give rise to a carbinol singlet, except C-29 whichis believed to be unsubstituted. (lH and 13C n.m.r.;see above). (b) The chemical shift of the remainingsinglet (6 6-74) is unusually large, suggesting locationnext to a carbonyl group. (c) The usual C-15 protonn.m.r. signals (utilin, 6 3.35 and 2.53, J 19 Hz) are absentin procerin, although this might also be due to theopening of the lactone ring.An alternative possibility is that the ester is at C-6,as in many known limonoids. This substitution norm-ally gives rise to a broad H-6 singlet, owing to weakcoupling with H-5.The singlet at 6 6-74 is much sharperthan this, showing no sign of weak coupling. Further,there is a two-proton doublet at 6 2-26 ( J 7 Hz) and whatappears to be a one-proton triplet at 6 2-88 ( J 7 Hz).Although these have not been shown to be coupled, andone band of the supposed triplet is obscured by the ABlactones of the limonoid group, and of C.O.C. (6a), theresonance due to H-17 is characteristically broadened loin comparison with other singlets in the same spectrum.The band at 6 5.80 in the spectrum of procerin is broad-ened in this way, and could correspond to H-17, the otherR Of R’( 6 ) ( 7 ) .a;R = 0 a;R’=Pr’ a n d Pr‘CH2(mixture)R2= Ac, R3= Pr =1 2 3 b;R =Me, R = H , R = H 2‘7 b;R = H,p-OAcOH‘ I(8)singlets being much sharper.There is no signal detect-able in the 6 3 4 region which can be assigned to H-17in a carbocyclic ring D. The chemical shift of 6 5-80assigned to H-17 is similar to that in the lactones, andrequires an acyloxy-substituent at C-17, not an ether asin C.O.C. (6a)Derivatives of the methyl meliacate group containingthe bicyclononane system normally have a l-oxo-group.The 13C n.m.r. spectrum of procerin shows that no ketoneis present. The only known natural bicyclononaneswithout a l-ketone are members of the group of 1,29-cyclo-compounds, which so far contains utilin (la):entandrophragmin (1 b) ,l candollein (Ic) ,11J2 bussein(7a),l3 and the closely related phragmalin.14 The pre-sence of this bridge in procerin is suggested by the factthat only three nuclear methyl groups are shown by the1H n.m.r.spectrum, instead of the usual four in methylmeliacate. It is supported by the 13C n.m.r. spectrum ofprocerin, which is similar in detail to those of utilin andof bussein hydrolysis product acetate methyl ester (7b),and in particular shows a triplet at 38.6 p.p.m. whichmay be ascribed to C-29 in this bridged ring structure(cj. utilin, 40.0; bussein, 39.9 p.p.m.). A doublet ofdoublets (6 3.04 and 2-74, J 13 Hz) in the lH n.m.r.spectrum may be ascribed to the two C-29 protons inN. S. Ohochuku and J. W. Powell, Chem. Comm., 1966, 422.11 G. A. Adesida and D. A. H. Taylor, Phytochemistry, 1967, 8,1429.l2 K. Wragg, unpublished results.l3 R.Hanni and Ch. Tamm, J.C.S. Chem. Comm., 1972, 1253.l4 R. R. Arndt and W. H. Barschers, Tetrahedron, 1972, 28,23331974system ascribed to 29-H,, these correspond to the posi-tions of the 6-H,, 5-H system in the spectra of manymethyl meliacate derivatives.We consider that the tertiary acetate is more likelyto be at C-2, leaving the tertiary hydroxy-group at C-1,on grounds of steric hindrance, and because 2-acetylderivatives are known, e.g. 2-acetoxyfissinolide.8There is no evidence to show where the propionateester system is situated. In view of the number ofnatural products with more complex ester groups at C-3and acetates at other positions (e.g. utilin, bussein, andmethyl 30-acetoxy-3P-isobut yryloxy- l-oxomeliacate 15)it is perhaps more likely that the propionate is at C-3.We therefore consider that procerin is represented by thestructure (8), in which the acyl groups are most probablydistributed as shown.Assignments of the 13C n.m.r.spectrum of procerin, based on this structure, are in-cluded in Table 2.EXPERIMENTALIsoY'ation of Procerin.-The milled bark of Carapa proceya(10 kg) was extracted with refluxing light petroleum (b.p.60-80°). The solid which precipitated from the cold,concentrated extract was recrystallised from methanol togive a mixture (130 mg), m.p. 260-270", not separated onfurther crystallisation. This (70 mg) was kindly separatedby t.1.c. by Dr. J. D. Corinolly, to give procerin (43 mg) anda second compound (24 mg). Previous extractions ofCnrupa bark had given only procerin.KCrystallised from methanol, procerin had m.p.296-301",8~ (CDC1,; 60 and 100 MHz) 7-64 and 7-39 (in, 2 x a-furan),6.74(s), 6-46 (m, p-furan), 5-80(s), 5-64(s), 5.35(s), 3.9 (s, OH),3-72(s) and 3.66(s) (2 x OMe), 3.04 and 2-74 (dd, J 13 Hz),2.86 (t, J 8 Hz), 2.26 (2H, d, J 8 Hz), 2.36, 2.15 (6H), 2.05,1-82, 1.26, and 0.82 (all s, 8 x CMe tert.), and 1.15 (t, J 7.5Hz, CMe primary) (Found: lMf, 848.3099. Calc. forC4,HK2Ol,: M , 848-3103). The second compound, m.p.285-288", had M+ 424 (probably C,,H,,O,).Quantitative Hydrolysis of Procerin.-Procerin (13.7 mg)was dissolved in methanol (10 ml) and aqueous 0 . 1 ~ -sodium hydroxide (10 ml) and refluxed for 0.5 h. Titrationagainst aqueous 0.1N-hydrochloric acid (phenolphthalein;8.7 ml required) revealed the production of 8.06 mol of acidper mol of procerin. The solution was treated with more0-1N-hydrochloric acid (total 10 ml) and distilled to dryness.The residue was treated with water (10 ml) and distilled todryness again. Titration of the combined distillates with0.0 1N-sodium hydroxide (phenolphthalein) required 8- 1 ml,equivalent to the production of 5-01 mol of volatile acid forevery mol of procerin. The neutralised solution from thetitration was evaporated to dryness, and the residue wasdissolved in D,O. lH N.m.r. spectroscopic examination ofthe solution showed the presence of acetic and propionicacids in the ratio 10 : 1.I thank the S.R.C. for a grant for 13C n.m.r. studies, andDr. F. W. Wehrli for advice and assistance in assigning thespectra.3/1586 Received, 27th J d y , 19731l5 D. A. H. Taylor, J . Chem. SOC. (C), 1969, 2439