Photoinduced transformations. Part 38. Photoreactions of 17-ethoxycarbonylmethylene-etiojerva-5,13(17)- and -5,16-diene-3β,11β,20ξ-triol 3,20-diacetate 11-nitrites

摘要

J.C.S. Perkin I Photoinduced Transformations. Part 38.l Photoreactions of 17-Ethoxycarbonylmethylene-etiojerva-5,13(17)-and -5,16-diene-3@,1 I@,205-triol 3,20-Diacetate II-Nitrites By Hiroshi Suginome," Sanji Sugiura, Norihisa Yonekura, Tadashi Masamune, and Eiji Osawa, Depart-ment of Chemistry, Faculty of Science, Hokkaido University, Sapporo 060, Japan Photolysis of one of the title nitrites (5) in benzene with a Pyrex-filtered light afforded a complex mixture in- cluding as major products 1 7-ethoxycarbonylmethylene-etiojerva-5,13( 17)-diene-3P.I 1 P.20t-triol 3.20-diacetate (4). 17-ethoxycarbonylmethylene-I 1a-aza-c-homoetiojerva-5,8,1 I,12( 14),13(17)-pentaene-3~,20~-diol3.20-diacetate 11 a-oxide (6). and 17-ethoxycarbonylmethylene-11 a-aza-c-homoetiojerva-5.11 a,l3(17) -triene- 3p.11 a.20E-triol 3.20-diacetate 11 a-oxide (7) ; in contrast, photolysis of N-acetyl-(22S,25S)-5a-veratr-l3(17)-enine-3P.11 P.23P-triol 3.23-diacetate 11 -nitrite (1 ) having a 1 2a-hydrogen is known to give (22S,25S)-N-acetyl- 11a-aza-c-homo-5a-veratr-I 1 a.l3(17)-diene-3~.11 a,23p-triol 11 a-oxide (2) as the sole product.The formation of the nitrone (7) having an a-oriented hydroxy-group supports the mechanism of the rearrangement previously pro- posed. An estimation by an empirical force-field method with a model hydrocarbon indicates that the five- membered ring of the nitrite (5) is more strained than that of the nitrite (1) by ca. 6 kcal mol-'. Thus, the difference in the prodllcts obtained from the nitrites (I) and (5) may be attributable to a difference in the polar effects of the C-17-substituents which influence the degree of stabilization of the polar transition state.The result of the photolysis of 17-ethoxycarbonylmethylene-etioJerva-5,16-diene-3~,11P,ZOE-triol 3.20-di- acetate 11 -nitrite (1 0) is also described. WE have previously reported the formation of an a-hydroxy nitrone (2) via photoinduced rearrangement of a steroidal cyclopentyl nitrite (1). On the basis of several pieces of experimental evidence,26pc the re-arrangement was shown to proceed via the pathway depicted in Scheme 1. An interesting feature of this photorearrangement was its stereo- and regio-selectivities. Thus, in this photo-rearrangement, a single nitrone with an a-oriented 1l-hydroxy-group was produced exclusively and no product derived from the nitroso-aldehyde (E) formed by combination between NO and the C-17 end of the ally1 radical (B), was obtained.An alternative mode of the nitrone formation, which involved an intra-molecular migration of 12a-hydrogen of the nitroso- aldehyde (C) to the formyl carbonyl, was excluded on the basis of the result of the cross-over experiments with unlabelled and 15N-labelled nitrites2" In connection with this pathway, we were interested in a problem of how the configuration of the C(12)-H of Part 37, H. Suginome, N. Yonekura, T. Mizuguchi, and T. Masamune, Bull. Chem. SOC.Japan, 1977, 50, 3010. the nitrite (1) influences the products and their stereo- chemistry in this photoreaction. In the present study, the photolysis of the title nitrite (5), which has truns- fused hydrind-6-en-1-01 nitrite part structure (F), is described. The result is compared with that in the photolysis of the nitrite (1)with cis-fused hydrind-6-en- 1-01 nitrite unit (H) reported previously.2 The photo- lysis of the related nitrite having trans-fused hydrind- 5-en-1-01 part structure (G) is also described.RESULTS Preparations and Photolysis of the Nitrites (5) and (10) (Schemes 2 and 3).-The C-12epimer of the nitrite (1) is not easily accessible and 17-ethoxycarbonylmethylene-etiojerva-5,13( 17)-diene-3P, 1lp,20t-triol 3,20-diacetate 1l-nitrite (5), which possessed the required structural unit (F),was chosen as the model since the parent triol (3) has 2 (a) H.Suginome, N. Sato, and T. Masamune, Tetrahedron Letters, 1969, 3353; Tetrahedron, 1971,27, 4863; (b)H. Suginome,T. Mizuguchi, and T. Masamune, J.C.S. Chem. Comm., 1972, 376; H. Suginome, T. Mizuguchi, S. Honda, and T. Masamune, J.C.S. Perkin I, 1977, 927; (c) H. Suginome. T. Mizuguchi, and T. Masamune, Tetrahedron Letters, 1971, 4723; J.C.S. Perkin I, 1976, 2365. 1978 613 h hv365nm-fi -*a@ AcO Ac 0 (1 1 (A) *NO -I caged (E1 0- (2 SCHEME1 already been prepared from jervine via 10 steps.3 Partial acetylation of the trio1 (3), obtained by the reported proced~re,~with acetic anhydride and pyridine at room temperature afforded the 3,20-diacetate (4), m.p.144-146 "C. The electron-impact mass spectrum of the di- acetate (4) exhibited no molecular ion. However, two distinct fragment ions at nz/e 414 (41) and 354 (lOO~o), due to elimination of one and two acetic acid units from the molecular ions, were observed. Kitrosation of the diacetate (4) by the usual method afforded an amorphous ll-nitrite (5). The nitrite (5) in benzene was photolysecl with a Pyrex-filtered light gener- ated by a 100 W high-pressure Hg arc until photodecom- position was complete. In contrast to the 12a-nitrite the photoreaction of the 12P-nitrite (5) was found to result in a complex mixture of several products as revealed by t.1.c. Separation of the products by extensive preparative t.1.c. afforded the parent Ilp-01 (4) (12:4,), compound, (6) (6y0), and compound (7) (6) in the order of decreasing mobilities in t.1.c.together with several minor ill-defined products. In a mobile fraction, a product having various spectral properties attributable to the corresponding 1l-ketone was obtained. However, the structure was not completely confirmed because of lack of material. The electron-impact mass spectrum of the product (6), m.p. 132-135 "C, exhibited a molecular ion peak at im/e 483. The i.r. spectrum has shown a broad band due to OAc and ester group at 1 738 cm-l but no absorption due to the presence of a hydroxy-group. The U.V. spectrum of the product (6) in ethanol exhibited an intense maximum at 206 nm (E 12 700) and two maxima at 266 nm (E 6 590) and 294 nm (E 7 760) with inflection at 304 nm.These absorptions are ascribable to a substituted pyridine N-oxide chromophore. This spectral evidence together with the following 1i.m.r. spectrum supported by double-irradiation studies were in agreement with the pyridine N-oxide structure (6) : apart from the signals due to 3a-H, 3-OAc, 20-OAc, 20E-H, 18-H, 19-H, and CO,Et, the n.m.r. spectrum of the product (6) exhibited three characteristic signals. One of them appeared at T 1.53 as a sharp 1H singlet. The chemical shift corresponds to ar-H of a pyridine N-oxide nucleus. Another 1H signal at 7 4.32 is assignable to the C-6 olefinic proton. The shape of this signal differed from that for the 6-H of A5-steroids and appeared as a diffuse triplet.This suggested a change of conformation of the B-ring in going from the nitrite (1) to the product (6). The third distinct signal appeared as a broad 2H singlet at T 6.74. By double-irradiation studies it was confirmed that these signals are due to 7ar-and 7P-H. Thus, ir- radiation of the 6-H signal caused a change of the broad singlet at T 6.74 to a singlet (Wt 4.5 Hz). On the other hand, irradiation of the broad signal at T 6.74 decoupled T. Masamune, A. Murai, and S. Numata, Tetrahedron, 1969, 25, 3145; T. Masamune and T. Orito, Bull. Chem. SOC.Japan, 1972, 45, 1888. J.C.S. Yerkin I the diffuse triplet at T 4.32 to give a sharp singlet. These double-irradiation studies proved that the signals at T 4.32 and 6.74 are coupled, the latter being assigned to 7a and 7P-H and thus supporting our assignment of the signal at T 4.32.Moreover, the irradiation at T 7.48 caused changes of the diffuse triplet at T 4.32 and the singlet at T 6.74 to a clear triplet (J = 3.6 and 3.6 Hz) and a clear doublet (J = 3.6 Hz) respectively. This indicated weak allylic coupling between 4-H and 6-H and homoallylic coupling between 4-H and 7-H. Further, it gave the coupling constants of Js,,tLand Js.,B as 3.6 Hz. The other signals in the n.m.r. spectrum of the product (6) are shown in Table 1. The mass spectral fragmentation pattern fully supported the assigned structure (6). The spectrum exhibited distinct fragment ion peaks at m/e 467 (660,/,), 407 (42y0), 352 (39), and 149 ( 109yo).The peak at m/e 467 is due to the loss of hV benzenei 0-0-(6) (7) + (Ll SCHEME2 oxygen from the molecular ion which is one of the mass spectrometric features of aromatic N-o~ides.~ The peak at m/e 407 is the one which corresponds to the loss of oxygen and acetic acid from the niolecular ion. The u.v., ix., mass, and n.m.r. spectra of the product (7) were fully in accord with an a-hydroxy-nitrone structure corresponding to the nitrone (2). Thus, in the U.V. spectrum of the product (7) in ethanol there were two maxima at 215 (E 6 600) and 294 (E 10 800). In its i.r. spectrum, the product (7) sliowetl a broad band at around 3 400 cm-' due to an associated OH ant1 a band at 1 740 H. Budzikiewicz, C.Djerassi, and 11. H. Williams, ' Mass Spectrometry of Organic Compounds,' Holden--Day, Inc., San Francisco, 1967, p. 388. 1978 615 cm-l due to OAc and C0,Et. The signal positions and the molecular ion peak but fragment ions at m/e 412 (M' -coupling constants of the n.m.r. spectrum are shown in 2MeC0,H) (15.3), 352 (52.4), and 178 (100). The Table 1. The ll-H signal appeared as a broad doublet at last two fragment ions are represented as (a) and (b) T 5.10 with J = 7.5 Hz. Irradiation at T 8.1 caused a (Scheme 4). collapse of this signal to a singlet. This defined the Nitrosation of the diacetate (9) afforded an oily nitrite TABLE 1 Chemical shifts (T) values and coupling constants (Hz) of protons in the etiojervanes and the photoproducts lla-H or 7a-and 1lP-H or 3a-I1 6-H 7P-H ll-H 16-H 18-H 19-H 20-H OAC OEt 6.46 (s) 4.72(bs) a 5.62 * a 8.03 (s) 8.70 (s) 4.95 (s) 5.75 (q),8.71 (t)J = 7.2 HZ 6.50 (bs) 4.72 (bs) a 5.70 (ni) U 8.07 (s) 8.65 (s) 4.94 (s) 5.75 (q),8.65 (t)J=7Hz ca.4.67 (bs) a ca. 5.57 t a 8.02 (s) 8.70 (s) 4.09 (s) 7.84 (s), 5.78 (q),8.74 t) 5.45 (bs) 7.95 (s) J = 7.2 H 5.41 (bs) 4.65 (bs) a 3.68 (t) a 8.27 (s) 8.88 (s) 4.12 (s) 7.85 (s), 5.78 (q),8.68 It) J = 7.5 HZ 7.97 (s) J = 7.2 HZ 5.33 (bs) 4.32 (bt) 6.74 (bs) : 1.53(s) a 7.65 (s) 8.51 (s) 3.85 (a) 7.78 (s), 5.81 (q),8.71 (t)7.90 (s) J= 7.2 HZ 5.35 (bs) 4.54(bs) a 5.10 (bd) $ a 7.68 (s) 8.82 (s) 3.99 (s) 7.84 (s), 5.76 (q),8.71 (t) 7.95 (s) J = 7.2 HZ a a a U a 7.33 (s) 8.85 (s) 3.68 (s) 7.82 (s), 5.74 (q),9.82 (t)J = 7.2 HZ 6.49 (bs) 4.72 (bs) a * 4.11 (bs) 8.82 (d) 8.71 (s) 5.35 (s) 7.95 (s) 5.76 (q),8.74 (t) J = 6.9Hz J = 7.2 HZ 6.50 (bs) 4.75 (bs) a 5.70 (m) 4.11 (bs) 8.82 (d) 8.72 (s) 5.36 (s) 5.76 (q),8.72 (t)J = 6Hz J = 7.2 HZ 5.35 (bs) 4.69 (bs) a * 4.02 (bs) 8.76 (d) : 8.67 (s) 4.45 (s) 7.81 (s), 5.78 (q),8.71 (t) $ (W, = 12.0) J = 6.9 HZ 7.93 (s) J = 7.2 HZ 5.41 (bs) 4.68(bs) a 3.83 (bt) 4.01 (bs) 8.99 (d) 8.90 (s) 4.49 (s) 7.82 (s), 5.81 (q),8.71 (t)J = 6.9 HZ 7.95 (s) J = 7.2 HZ 5.32 (bs) 4.53 (bs) a 3.95 (bd) 8.63 (d) : 8.89 (s) 4.45 (s) 7.81 (s), 5.77 (q),: 8.71 (t)$ J = 6.9 HZ 7.94 (s) J = 7.2 HZ a Unassignable.Data in literat~re.~ Data obtained in pyridine. * Coincided with ethoxy protons. t Coincided with 3u-H.:Confirmed by decoupling; see text on details of decoupling studies on compds. (8) and (9). chemical shift of 9a-H of the nitrone (7) as ca. t 8.1. The (10) which was photolysed by a procedure comparable to signal due to 19-H was found to shift 0.03 p.p.m. upfield in that used for the nitrite (5). In this case, two major and pyridine relative to deuteriochloroform. This behaviour several minor products were also produced. By preparative parallels that of the nitrones obtained previously.2a.b On t.l.c., two major products, 3p,20t-dihydroxy- 17-ethoxy- this basis, an a-configuration is assigned to the 1 1-OH carbonyImethylene-etiojen-a-5,16-dien-11-one 3,20-di-group of the nitrone (7). In the mass spectrum of the acetate (11) and the parent llp-01 (9) were obtained in 13 w benzene-t, SCHEME3 nitrone (7), the molecular ion was absent but prominent and 17 yields.The former was confirmed by a direct fragment ions at m/e 485 (Mf -H20),467 (M+-2H20) comparison with a specimen obtained by Jones oxidation (16), 407 (M+-2H20 -CH,CO,H) (97), and 352 of the llp-01 (9). (1000/,) were observed. An Estimation of the Difference of Steric Energy for the Partial acetylation of 17-ethoxycarbonylmethylene-etio-Five-membered Ring of the Nitrites (1) and (5) by a Force-jerva-5,16-diene-3@,1 lp,20<-triol (8) afforded the corres-field Method.-One of the most important factors which ponding 3,20-diacetate (9) (Scheme 3). influences the ease of p-scission of the ll-oxyl radicals The Inass spectrum of the diacetate (9) has shown no would be strain in the five-membered ring.An estimation has been made of the difference between the degrees of strain for the five-membered c-rings of 17-ethoxycarbonyl-methylene-etioJerva-5,16-diene-3p,1Ip,20-triol 3,20-di-acetate 1I-nitrite (5) and (22S,25S)-N-acetylveratr-13(17)-enin-llp-yl nitrite (1) by a force-field method. In order to calculate the difference, three models having structures C02Et (a) m/e 352 (b) m/e 178 SCHEME4 (24) were used. Examination of Dreiding models corresponding to the structures (12), (13), and (14) indicated that the c,D-rings of the C/D trans-structures (12) and (13) were rather rigid while in a model corresponding to structure (14) two major conformers (14a) and (14b) with Me , Me (12) 5a-H (14 1 (13) A5 (14b) regard to the C,D rings were possible.Thus, the conformer (14a) can undergo conformational conversion into an alternative conformer (14b) by inversion at C-16. In the * The possibility that the observed llp-01 was formed via a P-scission-recombination process is ruled out since it would lead to a more stable lla-01.' (a)N. L. Allinger, M. T. Tribble, M. A. Miller, and D. H. Wertz, J. Amer. Chem. Soc., 1971, 93,1637; (h)N. L. Allinger and J. T. Sprague, ibid., 1972, 94, 5734. J.C.S. Perkin I conformer (14a) the distance between 8P-H and 166-H is much closer than that in the conformer (14b). Allinger's force field was used in conjunction with pattern-research nULE 2 Calculated energies of moclel hydrocarbons * Sum of stcric energy components of Heat of C-8, -9, -11, -12 formation Strain energy and -14 (kcal mol-') (kcal Inol-l) (kcal mol-l) (12) --45.96 23.04 18.64 (13) .-19.50 24.32 20.83 (144 -48.72 20.28 14.97 (14b) -48.50 20.50 15.32 * Calculations were performed at the Hokksiclo Lhivcrsity Computing Centre.energy-minimization techriique to calculate tlie energies relevant to the present discussion ; these are shown in Table 2 DISCUSSION The foregoing experiments proved that the iiitrones with an a-oriented liydroxy-group at the a-position of the N-oxide group are formed in the photo-reaction of c-nor-D-homo-steroid 11p-01 nitrites with a 13(17) double-bond, irrespective of the 12-H configuration. This result is consistent with our proposed path for the nitrone formation depicted in Scheme 1.Namely, the cyclization of the nitroso-aldeliyde '(C) and/or (C') may lead to the formation of the intermediate D or D' with the a-oriented hydroxy-group since the presence of the lop-methyl may prevent the formation of a less stable intermediate with a p-oriented Iiydroxy-group.2b The present study has shown that, whereas photolysis of the llp-01 nitrite (1) resulted solely in p-scission of the oxyl radical to give the nitrone as the sole product, the 1lp-01 nitrite (5)afforded several products including, as major components, the parent 11p-ol, compound (8) with a pyridine N-oxide nucleus, and the nitrone (9).This contrast in the plioto-products is of interest since the formation of a significant amount of the llp-01 indicates that p-scission of 11p-oxyl radical generated from the nitrite (5) is more dificult than for the oxyl radical generated from the nitrite (l).* hlariy factors 879 may influence the relative ease of p-scission lo of the llp-oxyl radicals generated from tlie c/u cis-(1) and trans-nitrite (2). However, the most important ones would be the stabilities of both the ally1 radical and the product ketone generated, or of the transition state leading to them, and the degree of five-membered ring strain. IVith regard to the former 6 (a) E. 31. Engler, J. D. Ilndose, and P. v. I<. Sclileyer, J. Amev.Chem. Soc., 1973, 95, 8005; (h) J. D. Andose, and K. Illislow, ibid., 1974, 96, 2168. 7 N. Sat0 and T. Masamune, Tetvaliedvon I-ettevs, 1967, 1557; BdL. Clzenz. Soc. ,Japan, 1969, 42, 216. 8 C. Walling and A. I'adwa, J. Awier.. CAem. Soc., 1963, 85, 1593. 9 F. D. Greene, M. L. Sa.itz, 1'. D. Osterholtz, H. H. Lau, W. N. Smith, and P. A'I. Zanet, J. Oyg. Chent., 1963, 28, 55. 10 P. Gray and A. Williams, Chem. Rev., 1959, 59, 239. factor, the importance of polar effects in the transition state of p-scission has been noticed by Walling l1 and Kochi l2 and their colleagues. Of these factors, an estimation of the difference of the steric energy of the five-membered ring of etiojerva-5,13(17)-diene and 12ct-etiojerva-5,13(17)-dienewith the model systems by a force-field method indicated that the sum of steric energy components of the five-membered ring carbons of the C/D trans-isomer is some 6 kcal mol-l larger than that of the C/D cis-isomer. Thus, we should expect that provided that ring strain is the most im- portant factor for cleavage, since the nitrite (5) has a more strained five-membered ring, tf e 11,12-bond may cleave more easily than that of the nitrite (1).In fact, the reverse is found in the present exFeriments. -(B) -__t products en-1-01 nitrite system is not sufficient to cause p-scission of the oxyl radical generated. EXPERIMENTAL For instruments used and general procedures see Part 30.13 Preparation of 17-Ethoxycarbonylmethylene-etiojerva-5,13( 17)-diene-3p, 1lp,2O~-trioZ (3) and 17-Ethoxycarbonyl- methylene-etiojerva-5,16-diene-3~,11~,205-triol (8).-These compounds were prepared by a described 10-step pro- cedure from jervine.Preparation of 17-Ethoxycarbonylmethylene-etiojerva-5,13( 17)-diene-3p, lIp,205-triol 3,20-Diacetate (4) .-The trio1 (280mg) in pyridine (5ml) in the presence of acetic an- hydride (0.75ml) was stirred for 13 h at room temperature. The reaction mixture was worked up as usual to afford the In view of the complexity of the molecule3, with which we are dealing. the observed difference of the reactivities U' of two nitrites (1) and (5) cannot be attributed to any single factor. However, having excluded the factor of five-membered ring strain, the observed preference for opening of the five-membered ring of the nitrite (1) over that of the nitrite (5) might be chiefly ascribable to a difference in the degree of the stabilization of the ally1 radical intermediate (B) or of the transition state (J) leading to them.Thus, the difference in the products between the nitrites (1) and (5) may, perhaps, be rationalized in terms of the difference of the polar effectsl1?l2 of the C-17 substituents which influence the degree of stabilization of the polar transition state (Jb). The path which leads to formation of the product (6) is not easily explained, there being, at present, no evidence on which to base a mechanism. The product (6) is, however, most probably formed zkz dehydration and oxidation of the nitrone (7) during the photo- reaction and not simply during the work-up stage since the latter was stable at room temperature.The form- ation of a heteroaromatic AT-oxide analogous to the product (6) was not observed in the photoreaction of the nitrite (1). Although it is not clear why the pyridine N-oxide is formed, expecially in the photolysis of the nitrite (5),it is possible that the nitrone ring in the nitrite(s) is destabilized by the 17-substituent. The results of the photolysis of the nitrite (10) indicate that the strain in the five-membered ring of hydrind-5- l1 C. Walling and P. J. Wagner, J. Anaer. Chein. Snc., 1964, 86, 3368. l2 J. D. TSacha and J. K. Kochi, J. Org. Chem., 1965, 30, 3272. amorphous diacetate.This was recrystallized from ether- n-hexane to afford crystals of the diacetate (295mg), map. 144-146 "C (Found: C, 68.15; H, 8.1. C2,H3,0, requires C, 68.35; H, 8.05), vmx. 3 507 (OH) and 1721 and 1 744 cm-l (OAc, C0,Et); oiID2O -1141.7" c 0.83,(MeOH); n.m.r. data are given in Table 1. 17-Ethoxycarbonylmethylene-etiojerva-5,13( 17) -diene- 3p,1lp,20~-triol 3,20-Diacetate Nitrite (5).-The diacetate (196 mg) was dissolved in pyridine (3 ml) and nitrosyl chloride in pyridine was added dropwise at ca. -20 "C. The progress of the reaction was monitored by t.1.c. The solution was poured into water and the nitrite was extracted with methylenedichloride. The methylene dichloride solu- tion was washed with water three times and dried (Na,SO,).The solvent was removed by rotary evaporator at room temperature to yield an amorphous nitrite (226mg) which was immediately subjected to photolysis, vmaX.(CHCl,) 1 735br (OAc, C02Et), 1644,and 824 cm-1 (ONO); n.m.r. data are given in Table 1. PhotoZysis of the Nitrite (5).-The nitrite (352 mg) dis-solved in dry benzene (15.5ml) was subjected to photolysis with 100 W high-pressure Hg arc lamp, through Pyrex, under nitrogen. The progress of the reaction was monitored by t.1.c. (Spots due to the nitrite were detected by diphenylamine-sulphuric acid.) All the nitrite decomposed after 9.5 h and the solvent was removed by a rotary evapor- ator at room temperature to afford a complex mixture of brown oily products (372 mg).This was subjected to preparative t.1.c. (solvent, benzene-ether 5 : 1). Seven fractions A-G were obtained in the order of decreasing mobilities on t.1.c. The most mobile fractions A (18mg) and B (14mg) were not identified. The fraction C (43mg) l3 H. Suginome, T. Tsuneno, N. Sato, N. Maeda, T. Masamune, H. Shimanouchi, Y. Tsuchida, and Y. Sasada, J.C.S. Perkin I, 1976, 1297. was purified again by t.1.c. to afford two fractions (C, and C,). The more mobile fraction C, (8 mg) was not identified. The i.r. and n.ni.r. spectra suggested that the amorphous compound C, (6) (11 mg) was probably the corresponding ll-ketone but this was not confirmed. Fraction D (67 mg) was a mixture of several minor compounds and none of the compounds was identified.Fraction E (157 mg) was a mixture of the N-oxide (6), the 11/3-01 (4), and the nitrone (7). The mixture was subjected to preparative t.1.c. twice, and three compounds, (6) (21 mg), (4) (41 mg), and (6) (20 mg) were obtained in order of decreasing mobilities. The N-oxide (6) was recrystallized from ether to yield an analy- tical specimen, m.p. 132-135 "C (Found: C, 66.6; H, 7.0; N, 2.8. C27H36NO7 requires C, 66.8; H, 7.3; N, 2.9). The 11/3-01 was recrystallized from ether. The nitrone (7) was recrystallized from ether to yield an analytical specimen, m.p. 128-132 "C (Found: C, 63.8; H, 7.6; N, 2.6. C,,H,,NO, requires C, 64.1; H, 7.8; N, 2.8). Fraction F (29 mg) and fraction G (75 mg) were complex mixtures and were not identified.Preparation of 3/3,20-Dihydroxy- 17-ethoxycarbonyl- methyZene-etiojerva-5,16-dien-1 1-one, 3,20-Diacetate ( 1 1).-Jones reagent was added dropwise to the 3,20-diacetate (9) (51 mg) in acetone (3 ml). Excess of chromic acid was decomposed with 5 sodium hydrogensulphite and the solution was extracted with ether. The ethereal solution was twice washed with water and dried (Na2S0,). After removal of the solvent, the residue (49 mg) was purified by preparative t.1.c. and recrystallized from n-hexane-ether to afford the ketone (11) (11 mg), m.p. 167-169 "C (Found: C, 68.45; H, 7.75. C27H3@7 requires C, 68.6; H, 7.7y0), vmaX. 1740br cm-l (CO,Et, five-membered ring ketone and OAc); n.m.r. data are given in Table 1. Preparation of 17-Ethoxycarbonylmethylene-etiojerva-5,16-diene-3/3, 11(3,20-lrioZ 3,20-Diacetate (9) .-The trio1 (349 mg) in pyridine (8 ml) in the presence of acetic anhydride J.C.S.Perkin I (1.0 ml) was stirred for 12 h at room temperature. The reaction mixture was worked up to afford a crude diacetate. This was recrystallized from ethyl acetate-n-hexane to afford the diacetate (117 mg) as the first crop, n1.p. 141- 143 "C (Found: C, 68.35; H, 8.1. C2,H3,0, requires C, 68.35; H, 8.05), v,,,. 1 735 (C0,Et and OAc) and 3 527 cm-' (OH); aID2O -140.2" (c, 0.90, methanol); n.m.r. data are given in Table 1. 17-Ethoxycarbonylrnethylene-etiojerva-5,16-diene-3/3, 1 1/3,20~-trioZ 3,20-Diacetate Nitrite (10).-The diacetate (265 mg) in pyridine (5 ml) was nitrosated as for the iso- meric 3,20-diacetate.An oily nitrite (283 mg) obtained was immediately subjected to photolysis; n.m.r. data are given in Table 1. PhotoZysis of the Nitrite (lo).-The nitrite (283 mg) in dry benzene (13.0 ml) was subjected to photolysis under the same conditions as those for the isomeric nitrite (5). All the nitrite decomposed after 7.5 h. Removal of the solvent left a residue. T.1.c. of the product indicated that it was a mixture of several products including two major products. This was subjected to preparative t.1.c. (solvent, benzene-ether 3 : 1). Six fractions A-F were obtained in order of decreasing mobilities. The most mobile fraction A (34 mg), once purified by preparative t.l.c., was recrystallized from ether-n-hexane to yield the ll-ketone (6 mg), m.p. 160-163 "C. Fraction B (44 mg) was again purified by preparative t.1.c. and recrystallized from n-hexane-ether to yield the Ilp-01 (2 mg). Fractions C (24 mg), D (32 mg), E (18 mg), and F (65 mg) were a complex mixture of unidentified compounds. We thank Mrs. Tomoko Okayama for the measurements of 100 MHz n.m.r. spectra and the spin decoupling studies. 7/988 Received, 10th June, 19771