Photocyclisation of enamides. Part 24. Total synthesis of (±)-isofumigaclavine B and (±)-lysergic acid

摘要

J. CHEM. SOC. PERKIN TRANS. I 1985 941 Photocyclisation of Enamides. Part 24.' Total Synthesis of ( )-lsofumigaclavine B and (+_)-Lysergic Acid2 lchiya Ninomyia," Chiyomi Hashimoto, Toshiko Kiguchi, and Takeaki Naito Kobe Women's College of Pharmacy, Moto yamakita, Higashinada, Kobe 658,Japan Total syntheses of two ergoline-types of alkaloids, (f)-isofumigaclavine B (14) (for the first time) and methyl (4-)-lysergate (21) and methyl (+)-isolysergate (22), via a route involving reductive photocyclisation of the enamide (2) followed by glycol formation and oxidative cleavage of the dihydrofuran ring, are described. In Part 223 we have reported the establishment of a potential synthetic route to the despyrrole analogues of lysergic acid and isofumigaclavines, aiming at its application to the respective alkaloids.We now report novel total syntheses of isofumiga- clavine B and lysergic acid according to the synthetic methodology involving reductive photocyclisation of the enamide (2) followed by ring opening of the dihydrofuran ring of the photocyclised lactam (3). Previously, lysergic acid has been regarded as a central target alkaloid not only from the synthetic viewpoint but also because of its remarkable pharmacological potential, and it has been synthesized by six different method~logies,~.~ which are divided into two types depending upon the starting compounds used, i.e.route A from n R the indole derivatives (I) and the route B from the indoline derivatives(11).However, the synthesis of the alkaloid according to route A has only been successful when attempted by Oppolzer et a1.'"'' but the method has achieved some success in the synthesis of D-saturated ergo line^.^ These results reflect that the high reactivity of the indole ring and the feasibility of aromatisation of ring c might have been the major problems to overcome in the case of the synthesis starting from the indole derivatives (I).Therefore, this suggests that a synthetic approach by route B would be advantageous, in the case of the synthesis of ergolenes having a double bond in ring D, compared with route A. Therefore we chose our approach toward these ergolene alkaloids by starting from the enamide (2) which was prepared from the indoline derivative (1) and 3-furoic acid according to the methodology developed for the despyrrole analogue^.^ Acylation of the enamine prepared from 1 -benzoyl- 1,2,2a,3- tetrahydrobenzcdindol-4(5H)-one (1) and methylamine with furan-3-carbonyl chloride in the presence of triethylamine afforded the enamide (2) in 96 yield, which showed the n.m.r.peak for an olefinic proton at 6 6.45 as a broad singlet and broad and strong i.r. absorptions in the region of 1 640-1 620 cm-'.Irradiation of the enamide (2) in benzene-methanol (5: 1) in the presence of excess of sodium borohydride (10 mol equiv.) at 4-10°C with a high-pressure mercury lamp afforded the crystalline products in 81 yield, which were found to be a mixture of three compounds (3), (4), and (5) in the proportions 10:4: 1 by high-performance liquid chromatography (h.p.l.c.).t 0x0 BzNL 22 Bz N **" (3) (4) (5) OH e RN (6) R=H (8) R=H (7) R=Bz (9) R =Bz CHO Bz N t Although only two photocyclised lactams (3)and (4) were reported in the preliminary communication (ref.2), three isomers (3)-(5) have now been isolated by repeated column chromatography. These three products (3H5) were separated by repeated recrystallisation and column chromatography and were estab- lished as the expected products by reductive photocyclisation, from their mass spectra which exhibited a molecular ion peak at m/z 386, and from their i.r. absorptions at 1 660-1 620 cm-' (2NCO). Contrary to the cases of the model compounds3 (the despyrrole analogues of the alkaloids) in which the stereochemistry of the products was influenced by the solvent used, the proportions of the lactams (3)--45)were not influenced by the solvent, either by the amount of methanol or by the use of acetonitrile in place of benzene.Their stereochemistries were established from their n.m.r. spectra and by comparison of those of the model compound^.^ The DIE-C~Sfusion of three lactams (3x5) was determined from the coupling constants between the respective 3a- and 1 lc-H lo-1 1 Hz in (3)+5) and from the conversion of compound (3) into the 8,9-cis- disubstituted ergoline derivative (16). The trans relationship between 1Ib- and llc-H was deduced from their coupling constants of 10 Hz in (3)--(5) and by mechanistic considerations of enamide photocyclisation that have been well documented I as proceeding via conrotatory cyclisation of the 6 z-electron system in the enamide, thus forming a trans cyclic intermediate.The c/D-ring junctions were determined from the coupling constants between 5a- and llb- H lo Hz in (3) and (5), thus suggesting their C/D-trans fusion, while the value of 6 Hz in (4) suggested a cis fusion. Furthermore, the relative configuration between 5a-and 6a-H as being cis was also deduced from the signal pattern of 6-Ha, which appeared as a sharp quartet with J 12 Hz at 6 1.68 in (3)and at 6 1.55 in (4), thus suggesting their cis-1,3-diaxial relationship where ring c would take a half-chair conformation as shown in Figure 1. This deduction coincides well with that Figure 1.reached by Rebek et aL4"for the analogous compounds. On the other hand, the n.m.r. signals of hydrogens on rings D and E in (5) closely resemble those of (3)but the signals for 6-H, overlap at 6 ca. 2.38-2.02. This behaviour can be readily explained by assuming the conformation of ring c is a half-boat form which would have a trans relationship of two hydrogens at the 5a- and 6a-positions as shown in Figure 1. Previou~ly,4c*~~*~only those compounds having these two hydrogens at the 5a- and 6a-position in a cis-1,3-diaxial orientation had been isolated and this is the first time that a compound having a trans configuration between hydrogens at positions 5a and 6a has been prepared.Of the photocyclised lactams (3)-45), the c/D-trans-lactam (3),which was obtained predominantly, was used for the total syntheses of isofumigaclavine B and methyl lysergate. Ring opening of the dihydrofuran ring of the photocyclised lactam (3) was accomplished according to the procedure described previ~usly.~ Lithium aluminium hydride reduction of the c/D-trans-lactam (3)afforded the diamine (6) as a result of accompanying debenzoylation. This diamine (6) was then re- benzoylated with benzoyl chloride to give the N-benzoate (7) in 74 yield. A similar reduction of a mixture of the c/D-trans-and -cis-lactams (3) and (4) (5 :1) with lithium aluminium hydride J. CHEM. SOC. PERKIN TRANS. I 1985 gave a mixture of the corresponding N-debenzoylated amines (6) and (8)which were also rebenzoylated to yield a mixture of two benzoates (7) and (9) in 61 and 9 yield respectively; these benzoates were separated by column chromatography.Their n.m.r. spectra, which showed, for (7), 6 2.22 (td, J 10 and 2 Hz, 5a-H) and 1.43 (9, J 10 Hz, Ha,), and for (9),6 2.22 (br d, J 11 Hz, 5a-H) and 1.44 (4, J 11 Hz, 6-Ha,), established their structures. Hydroxylation of the dihydrofuran ring in the amine (7) was first carried out by using an equimolar amount of osmium tetraoxide followed by treatment with hydrogen sulphide gave the glycols (10) in 53 yield. The yield of the glycols (10) was improved to 83 by reaction with a catalytic amount of osmium tetraoxide and trimethylamine N-oxide as the oxidising agent.* The product (10) was found to be a mixture of two cis-glycols in the ratio 3 :2 which was estimated from the peak area of the two alcoholic proton signals in the n.m.r.spectrum, though it remained to be determined which isomer was predominant.* This glycol mixture was also observed in the case of despyrrole derivative^.^ The mixture of cis-glycols (lo), without separation, was then treated with sodium metaperiodate in methanol-water (1 :1) to cleave the glycol to afford the 9a- hydroxy-8P-aldehyde (1 1) as the sole unstable product upon work-up with sodium carbonate. This base treatment after the glycol cleavage brought about a complete isomerisation of the ring-opened 8-formyl group from the unstable 8a-orientation to the stable 8P-configuration. The structure of the 9a-hydroxy- 8P-aldehyde (11) was established from its spectral data v,,,.3 500 (OH) and 1 720 cm-' (CHO); 6 9.80 (br s, CHO) and 2.35 (s, NMe). On the other hand, similar glycol cleavage of a mixture of the glycols (10) but without the base treatment afforded a mixture of two epimeric aldehydes (11)and (12) with the ratio ca. 1 :1, which was estimated from the peak area of the 8-formyl protons in the n.m.r. spectrum (12); 6 9.93 (br s, CHO). The structures of these aldehydes (11)and (12)weTe later unambiguously established by their conversion into the stable synthetic intermediates (15) and (16). The aldehydes (11) and (12) bear all the functional groups required for the synthesis of ergot alkaloids having an ergoline nucleus.Total Synthesis of ( +)-Isofumigaclavine B (14).-Wolff-Kishner reduction of either the 9a-hydroxy-8P-aldehyde (11) or a mixture of the two epimeric aldehydes (11) and (12) gave the N-debenzoylated 8P-methyl-9a-01 (13) as the sole product in 40-47 yield, and which showed characteristic n.m.r. signals S 3.34 (t, J 10 Hz, 9-H) and 1.06 (d, J 7 Hz, 8-Me), thus confirming its structure as 2,3-dihydroisofumigaclavine B (ergoline numbering). Dehydrogenation of the indoline (13) with manganese dioxide gave ()-isofumigaclavine B (14) in 45 yield, which was identical with the natural alkaloid upon comparison of their i.r. and n.m.r. spectra and R, values.' Alternatively, conversion of the indoline (13) into ( +)-isofumigaclavine B (14) could be conveniently achieved by treatment with benzeneseleninic anhydride lo in 50 yield.Thus, we succeeded in the first total synthesis of the alkaloid (4 )-isofumigaclavineB. Novel Total Synthesis of Methyl (+)-Lysergute (21) and Methyl ( +)-lsolysergate (22)-The conversion of the aldehydes (11) and (12) into the desired unsaturated esters (17) and (18) was accomplished according to the procedure established previ~usly.~Oxidation of the 9a-hydroxy-8P-aldehyde (11) with chromium trioxide and sulphuric acid in methanol-acetone solution gave the 9a-hydroxy-8P-methyl * Although we reported that the glycol was described as a major product in the preliminary communication (ref.2), careful reinvestig- ation of the n.m.r. spectrum of a mixture of the glycols (10)showed the ambiguity of their assignment. J. CHEM. SOC. PERKIN TRANS. I 1985 R' Me (13) R'=Me, R2 = H (14) (16) (15) R'=C02Me, R =Bt 2CO,Me C0,Me (17) R=BZ (18) R=Bt (19) R= H (20) R= H C0,Me A ZMe (22) (23) ester (15)in 38 yield. Similar oxidation of a mixture of the two epimeric aldehydes (11) and (12) (ca. 1:1) gave the 9a-hydroxy- 8P-methyl ester (15) and 8a-methyl ester (16) in 11 and 14 yield respectively, which were separated by p.1.c. (preparative layer chromatography). The structures of these esters (15) and (16) were determined mainly from their n.m.r. spectra which suggested an equatorial orientation of the 8P-ester group in (15) S 4.04 (t, J 10 Hz, 9-H), 3.27 (dd, J 12 and 5 Hz,7-H,,), and 2.57 (t, J 12 Hz, 7-Hax), and an axial configuration of the 8a-ester group in (16)S 3.86 (br dd, J 11 and 4 Hz, 9-H), 3.48 (dd, J 12 and 3 Hz, 7-H,,), and 2.47 (dd, J 12 and 3 Hz, 7-Hax).Dehydration of the epimeric esters (15) and (16) was performed by heating with phosphoryl trichloride and 85 phosphoric acid in pyridine to afford a mixture of the two epimeric unsaturated esters (17) and (18)* in identical ratios (3:1), with the 8P-ester (17) as the major product in 44-55 yield from the respective esters (15)and (16). Recrystallisation of the above mixture afforded the homogeneous 8P-ester (17), the n.m.r. and i.r.spectra of which were completely identical with those of the N-benzoate of methyl 2,3-dihydrolysergate presented by Ramage. However, attempts to isolate the 8a-ester (18)on h.p.1.c. was unsuccessful due to ready isomerisation to the 8P-ester (17) during the course of purification. Furthermore, the facts that the homogeneous @-ester (17) yielded a 3: 1 mixture of two epimeric esters (17) and (18)upon being heated in methanol and that the same 3: 1 mixture was * Although the 8fLester (17) was described as the sole product in the preliminary communication (ref. 2), the epimeric -ester (18) was also detected by n.m.r. spectroscopy and h.p.1.c. formed during the course of purification of compound (17) supported the isomerisability of the respective esters (17) and (18)to afford the equilibrium mixture as reported by Rebek et ~l.~".?Furthermore, the n.m.r.peaks ascribable to the minor component (18)present in the 3: 1 mixture completely correlate with those of the N-benzoate of methyl 2,3-dihydroisolysergate reported by Rebek et aL4"*t Since the 8P-ester (17) had already been converted into ()-lysergic acid by Woodward et a1. and Ramage et a1., the present work formally completed the synthesis of (+)-lysergic acid. Selective removal of the N-benzoyl group from the 3: 1 mixture of the two epimeric esters (17)and (18)was achieved by alkylation with triethyloxonium tetrafluoroborate 4e followed by hydrolysis under mild acidic conditions to afford the indolines (19) and (20) in the ratio 5:2 in 55 combined yield.The n.m.r. spectra and the ratio of two isomers (19) and (20) were found to be identical with those of authentic samples reported by Ramage et a1.4' Dehydrogenation of the above mixture of indolines (19) and (20) with benzene seleninic anhydride lo furnished methyl ( f )-lysergate (21) and methyl (+)-isolysergate (22),in the ratio 3:2 in 60 yield, which were separated by h.p.1.c. The respective isomers (21) and (22) were identical with natural alkaloids 1on comparison of their n.m.r. spectra and R, values; thus we had completed a novel total synthesis of (+)-lysergic acid. Equilibrium and Stereochemical Study of the Unsaturated Esters (17)-(22).-Through our results in the synthesis of lysergic acid and its derivatives based on the synthetic route developed for their despyrrole analogue^,^ we have succeeded in the synthesis of eight compounds having analogous structures, i.e.a 9-ergolene-8-methyl ester and its despyrrole analogues, and this offers us a good opportunity of establishing their stereochemistry and of analysing the relationship between the isomerisation between respective epimers and their conform- ation, details of which have not been fully clarified even though a couple of syntheses of lysergic acid have been achieved. By taking advantage of the fact that we possessed two pairs of epimers, we investigated the isomerisation of these 9-ergolene-8- carboxylates (17) and (18), and (21) and (22), and their despyrrole analogues (23)once we had isolated the respective epimers and had measured their n.m.r.spectra at 200 MHz. As described above, the 8P-ester (17) and 8a-ester (18) underwent isomerisation in methanol to yield identical equilibrium (3:l) mixtures, with epimer (17) as the major component, which were detected by h.p.1.c. (Microporasil; 5 MeOH-CH,Cl,). This isomerisation mixture ratio coincided with that reported by Rebek et ~l.~", though Ramage el a1. stated that the isomerisation only occurred with pure compound (Is).?Attempts to isolate a pair of epimers of the N-nor derivatives were unsuccessful due to their instability, though it has previously been recognised 4e*4e that compounds (19) and (20) exist as an equilibrium mixture.Methyl ()-lysergate (21) and methyl (f)-isolysergate (22) were separated by h.p.1.c. using Microporasil in 3 MeOH- CHC1, and their structures were completely assigned by 'H n.m.r. spectroscopy for the first time as described later in this t Although the 8a-methyl ester (18) had been already synthesized both by Ramage et and Rebek ef al,,4etheir respective n.m.r. spectra and the equilibrium experiment results did not coincide. Comparison of the n.m.r. spectrum of an equilibrium mixture (3: 1) of (17) and (18), which we prepared, with those reported by Professors Ramage and Rebek clearly established that the n.m.r. spectrum of compound (18)quoted by the latter should be correct. $ The samples were gifts from Dr. P. A.Stadler, Sandoz Ltd. Table 1. 'H N.m.r. data for methyl (+)-lysergate (21) in CDC1, (200 MHz) 6 (J in Hz) 6 on irradiation (observation) NH 2-H 5-H 7-Heq 7-Hax 8-H 9-H CO,Me 7.95 (br) 6.95 (t, J 2)3.55 (dd, J 14, 6) 2.74 (ddd, J 14, 12, 2) 3.24 (m) 3.31 (br dd, J 11, 5) 2.72 0, J 11)3.77 (m) 6.64 (br s) 3.80 (s) 7.95 (lose J 2) 2.74 (lose J 2) 3.24 (lose J 6) 2.74 (lose J 14) 6.95 (lose J 2) 3.55 (lose J 14) 3.24 (lose J 12) 6.64, 3.77, 3.55, 2.74 (sharpeningof peaks) 6.64 (sharpening of peaks; J 11, 5 apparent) 3.77 (lose J 5) 2.74 (change)" 3.77 (lose J 11) 3.31 (lose J 11) 6.64, 3.31, 3.24 (sharpening of peaks) 2.74 (change)" 3.77, 3.31, 3.24 (sharpening of peaks) NMe 2.64 (s) " Irradiation at 6 2.74 (4-H,,) changed the signal patterns of 8-H and 7-He, in addition to that of 4-He, as a result of double irradiations both at 4-Ha, and 7-Ha, (6 2.72).paper. The respective epimers were each found to give an equilibrium mixture (3 :2), with epimer (21) the major component, upon being heated in methanol. In addition to the results obtained from the despyrrole analogues(23) described in a previous paper,13 the results given in this paper would indicate that, as a generalisation, isomerisation of the 8-ester group in the ergoline and benzoCflquinoline derivatives is relatively easy in methanol and leads to the respective equilibrium mixtures. Additionally, it was also found that all the 8P-esters, (17), (19), and (21), have larger R,values than their -ester counterparts, (18), (20), and (22), on h.p.l.c., thus providing a convenient method for their identification.Then we analysed the n.m.r. spectra of methyl (f)-lysergate (21) and methyl (f)-isolysergate (22). As a result of the spin- decoupling experiments on the n.m.r. spectrum of methyl (+)-lysergate (21), the following assignments of signals as summarised in Table 1 were reached. Bailey and Grey had reported14 a detailed study of the conformation of NN-dimethyl-lysergamide and NN-dimethyl- isolysergamide by n.m.r. spectroscopy. Upon taking these results into consideration, the present assignments for the esters (17)--(20) allowed us to suggest that methyl ( )-lysergate (21) has its c and D rings in the half-chair form as shown in Figure 2.(22) R'= H, R2=COZMe Figure 2. The reasons for the above conclusion are as follows; we observed an allylic coupling between 2-H and 4-H,,, thus J. CHEM. SOC. PERKIN TRANS. I 1985 Table 2. 'H N.m.r. data for methyl (+)-isolysergate (22)in CDCl, (200 MHz) 6 (J in Hz) 6 on irradiation (observation) NH 2-H 4-He, 4-Ha, 5-H 7-H, 7.96 (br) 6.92 0, J 2) 3.44 (dd, 2.75 (ddd, J 14, 5) J 14, 11, 2) 3.22 (m) 3.37 (m)b 7.96 (lose J 2) 2.75 (lose J 2) 3.22 (lose J 5) 2.75 (lose J 14)" 6.92 (lose J 2) 3.44 (lose J 14) 3.22 (lose J 11) 6.58, 3.44, 2.75 (sharpening of peaks) 6.58 (sharpening of peaks) 2.75 (change)" 7-Ha, 8-H 9-H 2.74 (m)3.32 (m) 6.58 6.58 (sharpening of peaks) 2.75 (change)" 3.22 (sharpening of peaks; J 4 apparent) C0,Me NMe (br, d, J 4)3.74 (s) 2.59 (s) Irradiation at 6 2.75 (4-Ha3 changed the signal patterns of 8-H and 7-He, in addition to that of 4-He, as a result of double irradiations both at 4-H,, and 7-H,, (6 2.74).* The signal pattern of 7-Hc, was described as a broad doublet (J 12 Hz) in our recent paper (ref. 13). suggesting a cu. 90" dihedral angle between these two hydrogens, the signal for 4-H,, appearing at much lower field than that for 4-Ha, by virtue of its being almost coplanar with ring B. The observation of an allylic coupling between 5-H and 9-H would suggest a near perpendicular orientation of 5-H with respect to the double bond at the 9,lO-position. Further, it has been shown that methine hydrogens in structurally related fragments which are gauche to the nitrogen lone pair are deshielded by about 0.6-4.8 p,p.m.relative to those orientated trans-axially which are found at 6 cu. 3.2.15 Therefore, the chemical shift of 5-H, 6 3.24, suggests its trans-axial orientation to the nitrogen lone pair. Such an arrangement can only be realised if ring D exists in that half-chair conformation in which C-5,-10, -9, and -8 are more or less in the same plane. As a result, C-7 is situated above this plane on the same side as 5-H while the nitrogen and its lone pair are directed below. The existence of a long-range W-shaped coupling between 7-He, and 9-H suggests the half-chair conformation of the ring D where two hydrogens, 7-He, and 9-H, are in the same plane. Furthermore, the pseudo-equatorial orientation of the 8-methyl ester group is also suggested from the following data; there are couplings with J 5 Hz between 7-H,, and 8-H and J 11 Hz, between 7-H,, and 8-H, a homoallylic coupling between 5-H and 8-H, and a small coupling with J cu.1 Hz between 8-H and 9-H. The above conclusion on the conformation of methyl (+)-lysergate (21) coincided with that proposed by Stoll et all6 for lysergic acid. Next, wecarried out a spin-decoupling analysis on the n.m.r. spectrum of methyl (+)-isolysergate (22). As shown in Table 2, the chemical shifts of all the protons are assigned. However, the chemical shift of 8-H appeared very close to that of 7-He, and the signal for 7-H,, overlapped with that of 4-Ha,, thus rendering the mutual coupling constants between 8-H, 7-Ha,, and 7-H,, indeterminable.In addition, the very close chemical shifts of 5-H and 8-H made the existence of a homoallylic coupling uncertain. On the other hand, the chemical shifts of 2-H, 4-H, and 5-H and their signal patterns are very similar to those observed for compound (21) and similar allylic couplings are observed between 2-H and 4-H,, and 5-H and 9-H as for (21). In particular, the almost identical chemical shift of 5-H, 6 J. CHEM. SOC. PERKIN TRANS. I 1985 Taw 3. 'HN.m.r. data (6; J in Hz) for compounds (17)-(22) in CDCl, (200 MHz) (17) (18)" (19) (20) 2-H 2-H, 4.34 (br) 3.70 4.M.10 (m) 3.67 3.82-3.60 (m) 3.67 (t, J 7) 6.95' 6.92' 3-H 3.42 (m)(t, J 10) 3.45-3.30(t, J 11) (m) 3.46-3.10 (m) 4-Hq 4-Hm 5-H 7-Hq 7-Ha, 8-H 9-H C0,Me NMe 2.57 (m) 1.40 3.04 (br d, J 12) 3.28 (br dd, J 11.5, 6) 2.69 (t, J 11.5) 3.68 (m) 6.55 (br s) 3.76 (s) 2.51 (s) (9, J 12) 2.51 (d, J 6.8)b 1.41 2.89 (br dd, J 11.7, 2) 3.45 (d, J 11.7) 2.58 (dd, J 11.7, 5) 3.18 (br t) 6.55 (d, J 3.2) 3.74 (s) 2.48 (s) (q, J 11.7) 2.62-2.43 (m) 1.43 1.42 3.05 2.89 (br d, J 12) 3.46-3.10 (m) 3.45 2.70 2.62-2.43 (m) (t, J 12)3.82-3.60 (m) 3.46-3.10 (m) (s,J 12) (4, J 12) (br d, J 12) (br d, J 12) 6.60-6.48 (m) 3.77 (s) 3.74 (s) 2.53 (s) 2.50 (s) 2.52 (d, J 5Ib 1.42 (q, J 11.5)2.90 (br dd, J 11.5, 2.5) 3.44 (4 J 12)2.57 (dd, J lZ4.7) 3.31-3.10 (m) 6.53-6.49 (m) 3.72 (s) 2.52 (s) 3.55 (dd, J 14, 6) 2.74 (ddd, J 14, 12, 2) 3.24 (m) 3.31 (br dd, J 11, 5) 2.72 (t, J 11)3.77 (m) 6.64 (brs) 3.80 (s) 2.64 (s) 3.44 (dd, J 14, 5) 2.75 (ddd, J 14, 11, 2) 3.22(m) 3.37 (m) 2.74 (m) 3.32(m) 6.58 (br d, J 4)3.74 (s) 2.59 (s) " We assigned the respective protons from the data reported by Rebek et Reported" signal patterns are different from the expected ones.Assigned from the n.m.r. spectrum of a mixture of compounds (19) and (20) in the ratio 5:2. 'Signal for 2-H. 3.22, with that in compound (21), 6 3.24, and the existence of a long-range W-shaped coupling between 9-H and 7-H,, can only be explained by assuming a conformation for compound (22)as shown in Figure 2. Therefore, the 8-methyl ester group would take a pseudo-axial orientation as supported by the coupling constant of 4 Hz between 8-H and 9-H. Furthermore, the above conclusion on the conformation of methyl (+)-isolysergate(22) coincides with that proposed by Stoll et for isolysergic acid.Based upon the assignment of the n.m.r. spectra of compounds (21) and (22), we then compared the spectra of compound (17) and the mixture of esters (19) and (20). Though .~~Rebek et ~1 have isolated compounds (18) and (20) and described the chemical shifts and signal patterns of the peaks, they have not assigned the respective peaks. Therefore, upon comparison of the n.m.r. spectra of compounds (21)and (22),we have succeeded in the assignment of the respective peaks as summarised in Table 3. It is clear that the n.m.r. spectra, particularly the chemical shifts and the signal patterns of protons in ring D, which is the common structural moiety in the compounds compared in this study, are very similar depending on the type of compound, as shown by the 8P-esters (17),(19), and (21),and the -esters (18), (20), and (22).Thus, we conclude that all these compounds have the same conformation under the conditions of the measurements and further we assigned the orientation of the substituents as follows. The orientation of the 8-ester group would not have any effect on the chemical shift of 9-H but would alter its signal pattern. Thus the signal for 9-H, which was coupled with 8-H, appeared as a broad singlet in the spectrum of the 8P-ester, while it appeared as a broad doublet, J 34 Hz, in the spectrum of the 8a-ester. In addition the n.m.r.signals for the methyl protons in the ester moiety and the N-CH, group appear at slightly higher field in the 8a-esters than in the 8p-esters. In conclusion, the stereostructures of the epimers, including the orientation of the ester group at C-8, can now be deduced from the n.m.r. behaviour of three proton groups, i.e., of the olefinic proton 9-H, and the methyl protons of the methyl ester at C-8 and the N-CH, group. Experimental'H N.m.r. spectra were measured with JEOL PMX-60 and Varian XL-200 instruments for solutions in deuteriochloroform unless otherwise stated (tetramethylsilane as internal reference), i.r. spectra for solutions in chloroform with a Hitachi 215 spectrophotometer, and mass spectra with a Hitachi M-80 machine.H.p.1.c. was carried out on a Waters Associates instrument with a U.V. detector at 254 nm using a Microporasil column (30 cm x 7.8 mm i.d.). M.p.s were determined with a Kofler-type hot-stage apparatus. Extracts from the reaction mixtures were washed with water and dried over anhydrous sodium sulphate. The photochemical reactions were carried out by irradiation at 4-10 "C with a high-pressure mercury lamp (300 W) (Eikosha PIH-300). Ether refers to diethyl ether. N-(1 -BenzoyI- 1,2,2a,3-tetrahydrobenzcd)indol-4-yI)-N-me thy Furan- 3 -carboxam ide (2)-Anhydrous met h ylamine gas was bubbled into a boiling solution of l-benzoyl-l,2,2a,3-tetrahydrobenzcd)indol-4-(5H)-one (1)ti (2 g) in benzene (200 ml) under a nitrogen stream for 5 h; water was removed as it formed.The mixture was refluxed further to remove the excess of methylamine by the bubbling nitrogen. To the resulting ice- cooled, stirred solution were added dropwise triethylamine (1 g) and then a solution of freshly prepared furan-3-carbonyl chloride (1.13 g) in anhydrous benzene (30 ml). After being refluxed for 2 h the mixture was ice-cooled and diluted with benzene, washed, and dried. The solvent was removed under reduced pressure and the residue was recrystallised from methanokther to afford the enamide (2) (2.66 g, 96), m.p. 130-131 "C; v,,,. 1 640-1 620 cm-' (2NCO); 6 (inter alia)6.45 (1 H, br s, 5-H) and 3.27 (3 H, s, NMe) (Found C, 74.8;H, 5.15; N,7.45. C2,H,,N20, requires C, 75.0 H, 5.25; N, 7.3).Reductive Photocyclisation of the Enamide (2).-A solution of a mixture of the enamide (2) (1.5 g) and sodium borohydride (1.5 g) in benzene-methanol (5:1; 900 ml) was irradiated for 3 h. The reaction mixture was washed, dried, and evaporated to give a viscous oil which was triturated with ethanol to afford crystals (1.22 g, 81). These were found to be a mixture of three lactams (3), (4), and (5) in the proportions 10:4: 1 by h.p.1.c. 1 500 p.s.i.; 5 MeOH-MeCN (0.4 ml min-I); (3)R, 45.6 min; (4) R, 49.6 min; (5)R, 48.8 min). Repeated recrystallisation of the crystals from methylene dichlorideethyl acetate afforded pure (3af3,5ap,6ap,l lba,l lci3)-8-benzoyl-5,5a,6,6a,7,8,1lb,l lc-octah ydro- 5-methyFur03,2 -c indo lo4,3-fg quino Iin-4(3 aH)- one (3) (813 mg, 5473, m.p.204-206 "C; v,,,. 1 660-1 620 cm-' (2NCO); 6 (inter alia) 6.46 (1 H, t, J2 Hz, 2-H), 5.37 (1 H, t, J 2 Hz, 3-H), 4.88 (1 H, dd, J 11 and 10 Hz, 1 lc-H), 4.44 (1 H, br, 7-H,), 3.87 (1 H, dt, J 11 and 2 Hz, 3a-H), 3.76 (1 H, t, J 12 Hz, 7-H2, 3.63 (1 H, br dd, J 12 and 10 Hz, 5a-H), 3.40 (1 H, m, 6a-H), 3.08 (3 H, s, NMe), 3.03 (1 H, t, J 10 Hz, llb-H), 2.68 (1 H, br d, J 12 Hz, 6-He,), and 1.68 (1 H, q, J 12 Hz, 6-Ha,); m/z 386 (M+)(Found: C, 72.95; H, 5.8; N, 7.05. C,,H,,N20,~~AcOEt requires C, 73.2; H, 6.0; N, 6.75). The residue obtained from the mother liquor was chromatographed on silica gel. The first fraction eluted with chloroform gave (3aP,5aP,6aa,ll ba,l 1cP)- 8-benzoy1-5,5a,6,6a,7,8,11b,l lc-octahydro-5-rnethylfuro3,2-c-indoZo4,3-fgquinolin-4(3aH)-one(5) (81 mg, 573, m.p. 250- 252 "C (decomp.) (from methylene dichloride-thy1 acetate); v,,,.1 660-1 620 cm-' (2NCO); 6 (inter alia) 6.38 (1 H, t, J2 Hz, (1 H, br, 7-HJ, 4.17 (1 H, dt, J 10 and 2 Hz, 3a-H), 3.73 (1 H, t, J 10 Hz, 7-H,), 3.60 (1 H, m, 6a-H), 3.28 (1 H, q, J 10 Hz, 5a-H), 3.00 (3 H, s, NMe), 2.97 (1 H, t, J 10 Hz, llb-H), and 2.38-2.02 (2 H, m, 6-Hz); m/z 386 (M') (Found: C, 74.3; H, 5.75; N, 7.25. C24H22NZ03requires C, 74.6; H, 5.75; N, 7.25). The second fraction eluted with chloroform gave (3aP,5aa,6aa,ll ba,l 1cP)- 8-benzoy1-5,5a,6,6a,7,8,11b, 1 lc-octahydro-5-methyrfuro3,2-c-indolo4,3-fgquinoZin-4(3aH)-one(4) (325 mg, 22), m.p. 2 19-221 "C (from methylene dichloride-ethyl acetate); vmax.1 6-1 620 cm-' (2NCO); 6 (inter alia) 6.41 (1 H, t, J 3 Hz, 2-H),5.34(1H,t,J3H~,3-H),4.59(lH,t,JlOH~,11c-H),4.38 (1 H, br, 7-H,), 3.85 (1 H, dt, J 10 and 3 Hz, 3a-H), 3.74 (1 H, t, J 11 Hz, 7-H,), 3.66 (1 H, m, 5a-H), 3.43 (1 H, m, 6a-H), 3.30(1 H, dd, J 10 and 6 Hz, llb-H), 3.10 (3 H, s, NMe), 2.38 (1 H, m, 6-He,), and 1.55 (1 H, q, J 12 Hz, 6-Ha,); m/z 386 (M') (Found: M+, 386.1637. C2,H,,N,03 requires M, 386.1629). (3ap,5aP,6ap,ll ba,l lcP)-8-Benzoyl-3a,4,5,5a,6,6a,7,8,1lb,l lc- decah ydro- 5-methylfuro3,2-c indolo 4,3 -fg quinoline (7) .-To an ice-cooled solution of the c/D-trans-lactam (3) (970 mg) in anhydrous ether-tetrahydrofuran (THF) (1 :2; 150 ml) was added lithium aluminium hydride (970 mg) in small portions.The mixture was refluxed under a nitrogen stream for 2 h. Usual work-up afforded the N-debenzoylated amine (6) v,,,. 3 400 cm-' (NH); 6 (inter alia) 4.85 (1 H, t, J 2.5 Hz, 3-H) and 2.40 (3 H, s, NMe), which was without further purification rebenzoylated by refluxing in benzene (95 ml) with benzoyl chloride (640mg) in the presence of triethylamine ( 500 mg) to afford the N-benzoate (7) 688 mg, 74 from (3), m.p. 187-189 "C (decomp.) (from benzene-ether); v,,,. 1 640 cm-' (NCO); 6 (inter alia) 6.47 (1 H, br s, 2-H), 4.98 (1 H, br s, 3-H), 4.68 (1 H, t, J 10 Hz, llc-H), 4.34 (1 H, br, 7-H,), 3.68 (1 H, t, J 11 Hz, 7-Ha), 3.44-3.22 (2 H, m, 3a- and 6a-H), 3.14 (1 H, dd, J 13 and 7 Hz, 4-He,), 3.14 (1 H, t, J 10 Hz, llb-H), 2.60 (1 H, dd, J 13 and 4 Hz, 4-H,,), 2.46 (3 H s, NMe), 2.42 (1 H, m, 6-He,), 2.22 (1 H, td, J 10 and 2 Hz, 5a-H), and 1.43 (1 H, q, J 10 Hz, 6-H,,) (Found C, 77.5; H, 6.45; N, 7.65.C,,H,,N,O, requires C, 77.4; H, 6.5; N, 7.5). Similar reduction of a mixture of the C/D-trUnS- and cis-lactams (3)and (4) (5:1; 970 mg) with lithium aluminium hydride gave a mixture of the corresponding N-debenzoylated amines (6)and (8),which were also rebenzoylated as described above and separated by column chromatography on silica gel to afford the N-benzoate (7) (569 mg, 61) and (3aP,5aa,6aa,l lba,l lcp)-8-benzoyl- 3a,4,5,5a,6,6a,7,8,11 b, 1 lc-decahydro-5-methyIfuro3,2-c-indolo4,3-fgJquinoline (9) (83 mg, 979, m.p. 172-174 "C (decomp.) (from benzene); v,,,.1 640 cm-'(NCO);6 (inter alia) 6.43 (1 H, br s, 2-H), 4.94 (1 H, br s, 3-H), 4.46 (1 H, t, J 10 Hz, llc-H), 4.36 (1 H, br, 7-Ha), 3.70 (1 H, t, J 10 Hz, 7-HP), 3.48 (1 H, m, 3a-H), 3.42-3.22 (2 H, m, 6a- and llb-H), 2.92 (1 H, ddd, J 11, 7, and 3 Hz, 6-He,), 2.79 (2 H,d, J 8 Hz, 4-H,), 2.50 (347 mg, 53); v,,,.(Nujol) 2-H),5.33(1H,t,J2H~,3-H),5.10(1H,t,J10H~,ll~-H),4.52(inter alia) 6.49 (0.6 H, d, J 5 Hz, OH), 6.25 (0.4 H, S(CD,),SO J. CHEM. SOC. PERKIN TRANS. I 1985 11 Hz, 6-H,,) (Found: M+,372.1843. C2,H2,N202 requires M, 372.1837). Hydroxylationof the N-Benzoate (7).-(a) To a solution of the N-benzoate (7) (600 mg) and pyridine (1ml) in anhydrousTHF (15 ml) was added a solution of osmium tetraoxide (410 mg) in anhydrous THF (3 ml) at -30 "C.The reaction mixture was kept at -30 "C for 2 h, and then the solvent was evaporated off to give a solid which was washed with ether and dissolved in methylene dichloride*thanol (1:1; 20 ml). Hydrogen sulphide was bubbled vigorously into the solution at 0 "C for 1 min and the black precipitate was removed by filtration. The filtrate was evaporated and the residue was crystallised from methanol- ether to give a mixture of two cis-glycols (10) in the ratio 312 3 370 (OH) and 1 640cm-' (NCO); m, OH), 5.41 (0.6 H, d, J 5 Hz, OH), 5.16 (1 H, m, 2-H), 4.73 (0.4 H, d, J 6 Hz, OH), 2.24 (3 H, s, NMe), 1.97 (1 H, m, 5a-H), and 1.21 (1 H, br q, J 11.5 Hz, 6-Ha,); m/z 406 (M'). (b) A solution of osmium tetraoxide (1mg) in t-butyl alcohol (0.5 ml) was added to a mixture of the N-benzoate (7) (480 mg), trimethylamine N-oxide dihydrate (195 mg), pyridine (0.1 ml), water (0.8 ml), and t-butyl alcohol (10 ml).After being refluxed under a nitrogen stream for 1 h, the solution was cooled and treated with 20 aqueous sodium hydrogen sulphite (50 ml). To the concentrated reaction mixture was added water and the precipitate thus formed was collected and washed successively with water and with methylene dichloride to give a mixture of two cis-glycols (10)in the ratio 3:2 (435 mg, 83). This mixture was used in the next step without further purification. (3~)-2,3-DihydroisofumigaclauineB (13).*-A mixture of the glycols (10) (100 mg) and sodium metaperiodate (100 mg) in methanol-water (1 :1; 80 ml) was stirred at room temperature for 1 h, and then sodium carbonate (100 mg) was added to the reaction mixture which was then stirred for a further 1 h.The reaction mixture was repeatedly extracted with methylene dichloride. The combined extracts were washed, dried, and evaporated to give the 9a-hydroxy-8P-aldehyde (11) vmax. 3 500 (OH), 1 720 (CHO), and 1 640 em-' (NCO); 6 (inter alia) 9.80 (1 H, br s, CHO) and 2.35 (3 H, s, NMe) which was without further purification dissolved in ethylene glycol (5 ml). To the resulting solution were added sodium hydroxide (40 mg) and hydrazine hydrate (0.5 ml). The mixture was then refluxed under a nitrogen stream for 30 min. The reaction mixture was diluted with water and repeatedly extracted with methylene dichloride. The combined extracts were washed, dried, and evaporated to give a residue which was purified by p.1.c.on silica gel to afford the 8P-methyl-9a-01 (13) (30 mg, 4779, m.p. 182-183 "C (decomp.) (from benzene); v,,,.(Nujol) 3 330 cm-' (NH); 6 7.41 (1 H,d, J~Hz, 13-H),6.56(1 H,d, J8 12-H),7.01(1 H, t, J~Hz, Hz, 14-H),3.66(1 H,m,2-H,), 3.34(1 H, t, JlOHz,9-H),3.22- 3.10 (2 H, m, 2-Ha and 3-H), 2.99 (1 H, dd, J 12 and 4.5 Hz, 7- Heq), 2.82 (1 H, t, J 10 Hz, 10-H), 2.46-2.38 (1 H, m, 4-He,), 2.40 (3 H, s,NMe), 2.28 (1 H, br t, J 11 Hz, 5-H),2.18 (1 H, t, J12 Hz, 7-H,,), 2.00 (1 H, m, 8-H), 1.44 (1 H, q, J 11 Hz, 4-Ha,), and 1.06 (3 H, d, J 7 Hz, CMe) (Found: M+, 258.1756.Cl6Hz2N,O requires M, 258.173 1). Oxidation of the glycols (10) (100 mg) with sodium meta- periodate (53 mg) as described above but without treatment with sodium carbonate gave a mixture of two epimeric aldehydes (11) and (12) in the ratio ca. 1: 1 v,,,. 3 500 (OH), 1 720 (CHO), and 1 640 cm-' (NCO); 6 (inter alia) 9.80 (0.5 H, br s, P-CHO), 9.67 (0.5 H, br s, a-CHO), and 2.35 (3 H, s, NMe). Wolff-Kishner reduction of this mixture gave the 8P-methyl-9a- (3H,s,NMe),2.22(1H,brd,JllHz,5a-H),and1.44(1H,q,J* Ergoline-type numbering is used for compounds (llH20). J. CHEM. SOC. PERKIN TRANS. I 1985 01 (13)(25 mg, 40)as the sole product which was identical with that obtained above. (+)-Isofwnigaclauine B (14).-(a) To a solution of the Sp-methyl-9a-01 (13) (18 mg) in chloroform (10 ml) was added manganese dioxide (40 mg).The mixture was stirred at room temperature for 18 h and was then filtered and the residual solid was washed thoroughly with chloroform. The combined chloroform layers were evaporated to give a residue which was purified by p.1.c. on silica gel to afford (+)-isofumigaclavine B (14) (8 mg, 45), m.p. 271-273 "C (decomp.) (from methanol) lit.,g" 222-224 "C (sublimed.); lit.,9b 278-282 "C (decomp.); vmax.(NuJol)3 495 cm-' (NH); 6 (CDCl,-CD,OD) 7.65 (1 H, d, J8 Hz, 12-H), 7.20(1 H,d, J8 Hz, 14-H), 7.11 (1 H, t, J8 Hz, 13- (1 H, m, 4-H,,), 2.35 (3 H, s, NMe), 2.12 (1 H, br t, J 10 Hz, 5-H), and 1.41 (1 H, q, J 12 Hz, 4-Ha,) (Found: M', 406.1911. C,,H,,N,O, requires M, 406.1891).Dehydration of the 9a-Hydroxy-8P-ester (15).-To a solution of the 9a-hydroxy-8P-ester (15)(42 mg) in anhydrous pyridine (1 ml) was successively added 85 phosphoric acid (0.01 ml) and phosphoryl trichloride (0.2 ml). l1 The solution was warmed at 60 "C under a nitrogen stream for 45 min. The reaction mixture was poured into ice-water, made alkaline by the addition of aqueous sodium carbonate solution, and extracted with methyl- ene dichloride. The organic layer was washed, dried, and evaporated. Purification of the residue by p.1.c. on silica gel gave a solid (22 mg, 55) which was found to be a mixture of two epimers (17) and (18)in the ratio 3: 1 by h.p.1.c. l 500 p.s.i.; 5H),6.92(1H,d,J1.5H~,2-H),3.54(1H,t,JlOH~,9-H),3.41(1 H, dd, J 14 and 4 Hz, 4-H,,), 3.03 (1 H, t, J 10 Hz, 10-H), 2.96 (1 MeOH-CH,Cl, (0.5ml min-'): (17) R, 50min; (18)R, 42.8 min.H, dd, J 12 and 4 Hz, 7-H,,), 2.79 (1 H, ddd, J 14,11, and 1.5 Hz, Recrystallisation of the solid from ethyl acetate afforded homo- 4-Ha,), 2.53 (3 H, s, NMe), 2.44 (1 H, ddd, J 11, 10, and 4 Hz, 5-geneous methyl (3p)- 1-benzoyl- 1,2-dihydrolysergate (17), m.p. H),2.18(1H,t,J12Hz,7-Ha,),2.00(1H,m,8-H),and1.13(3H,165-168 "C (lit.? 165-168 "C); v,,,. 1 725 (C0,Me) and d, J 7 Hz, CMe). The i.r. and n.m.r. spectra and R, values of (+)-1 640cm-' (NCO). The i.r. and n.m.r. spectra are superimposable (14) were found to be identical with those of natural isofumigaclavine B (Found: M', 256.1571. C16H2,N20 requires M,256.1574).(b) A solution of the 8P-methyl-9a-01 (13) (10 mg) and benzeneseleninic anhydride (7.2 mg) in THF (1 ml) was stirred at 40 "C under a nitrogen stream for 2 h. After being diluted with methylene dichloride, the solution was washed successively with aqueous sodium carbonate and water, dried, and evaporated to give a residue which was purified by p.1.c. on silica gel to afford (f)-isofumigaclavineB (14) (5mg, 50). Methyl (3P,Sp,9a)-1-Benzoyl-2,3-dihydro-9-hydroxy-6-methylergoline-8-carboxylate (15).-To a stirred, ice-cooled solution of the homogeneous 9a-hydroxy-8p-aldehyde (1l), which was prepared from the glycols (10) (200 mg), in methanol- acetone (1 :4; 50 ml) was added dropwise a solution (0.5 ml) of 4.0n4-chromiurn trioxide-sulphuric acid in water.The mixture was stirred at 0°C for 1 h. The precipitate thus formed was filtered and washed with methylene dichloride. The filtrate was washed successively with aqueous sodium carbonate and water. Upon removal of the solvent, the solid obtained was purified by column chromatography on silica gel to give the 9a-hydroxy-SB- methyl ester (15) (75 mg, 38), m.p. 22G228 "C (from methanol); v,,,. 1 720 (C0,Me) and 1 630 cm-' (NCO); 6 7.83 (1 H, brd, J8 Hz, 12-H), 7.6-6.90(7 H, m,ArH),4.36(1 H, m, 2-H,), 4.04 (1 H, t, J 10 Hz, 9-H), 3.78 (3 H, s, OMe), 3.65 (1 H, t, J 10.5 Hz, 2-Ha), 3.40 (1 H, m, 3-H), 3.27 (1 H, dd, J 12 and 5 Hz, 7-H,,), 2.98 (1 H, ddd, J 12, 10, and 5 Hz, 8-H), 2.80 (1 H, t, J 10 Hz, 10-H), 2.57 (1 H, t, J 12 Hz, 7-Ha,), 2.42 (3 H, s, NMe), 2.40 (2 H, m, 5-H and 4-H,,), and 1.47 (1 H, q, J 12 Hz, 4-Ha,) (Found: C, 71.05; H, 6.45; N, 6.95.C,,H,,N,O, requires C, 70.9; H, 6.45; N, 6.9). Oxidation of an Epimeric Mixture of the Aldehydes (11)and (12).-A mixture of the epimeric aldehydes (11)and (12) (ca. 1:l), which was prepared by metaperiodate oxidation of the glycols (10) (290 mg) according to the procedure given above but without sodium carbonate treatment, was oxidised with chromium trioxide as described above. Separation of the crude product by p.1.c. on silica gel afforded the 9a-hydroxy-8P-ester (15)(33 mg, 11) and methyl (3P,Sa,9a)-l-benzoy1-2,3-dihydro-9-hydroxy-6-methylergoline-8-carboxylate(16)(40 mg, 14), m.p.195-197 "C (from ethyl acetate); v,,,. 1710 (C0,Me) and 1 630cm-' (NCO); 6 7.90( 1 H, br d, J8 Hz, 12-H), 7.64-6.90(7 H, m,ArH),4.24(1 H,m,2-H,), 3.86(1 H, brdd, Jl1 and4Hz,9-H), 3.80(3 H, s, OMe), 3.64 (1 H, t, J 12 Hz, 2-Ha), 3.48 (1 H, dd, J 12 and3Hz,7-HC,),3.35(1H,m,3-H),3.17(1H,t,JlOHz, 10-H), 3.01 (1 H, m, 8-H), 2.47 (1 H, dd, J 12 and 3 Hz, 7-Ha,), 2.40 with those of the authentic sample reported by Ramage er a1. and Rebek et d4' From the n.m.r. spectrum of the equilibrium mixture (17) and (18)(3:l),the following peaks were assignable for methyl (3P)-l- benzoyl-2,3-dihydroisolysergate(18):6 3.74 (s, OMe), 2.89 (br d, J 12 Hz, 5-H), 2.48 (s,NMe), and 1.41 (4, J 12 Hz, 4-Ha,). These peaks were found to coincide with those which were present in the n.m.r.spectrum of compound (18)reported by Rebek et aL4' Each isomer gave the same equilibrium mixture in methanol (17):(18) 3: 1). Dehydration of the 9a-Hydroxy-8a-ester (16).-A mixture of the 9a-hydroxy-8a-ester (16)(40 mg), anhydrous pyridine (1 ml), 85 phosphoric acid (0.01 ml), and phosphoryl trichloride (0.2 ml) was worked up in the same manner as given for the dehydration of compound (15) to give a mixture of compounds (17) and (18)(17 mg, 44) in the same ratio (3: 1) as described for the dehydration of the 9a-hydroxy-Sp-ester (15). Methyl (3P)-2,3-Dihydrolysergate(19) and Methyl (38)-2,3- Dihydroisolysergate (20).-To a solution of an epimeric mixture of the esters (17) and (18)(3:l; 20 mg) in methylene dichloride (4 ml) were added successively sodium carbonate (10 mg) and excess of triethyloxonium tetrafluoroborate at room temperature under a nitrogen stream.After being stirred for 18 h, the reaction mixture was treated with 5 hydrochloric acid (20 ml) for 15 min, neutralised with saturated aqueous sodium hydrogen carbonate, and extracted with methylene dichloride. The extract was dried and evaporated to give a residue which was purified by p.1.c. on silica gel to afford a solid mixture of debenzoylated esters (19)and (20) (8 mg, 55) in the ratio 5:2, v,,,. 3 400 (NH) and 1 730 cm-l (C0,Me). The i.r. and n.m.r. spectra and the ratio of the two isomers (19) and (20) are identical with those of the authentic sample reported by Ramage et a1. This mixture was used in the next step without further purification.Methyl (+)-Lysergate (21) and Methyl (f)-Isolysergate (22).-A solution of an epimeric mixture of the esters (19) and (20)(5:2; 5 mg) and benzeneseleninic anhydride (3 mg) in THF (1 ml) was stirred at 40 "Cunder a nitrogen stream for 2 h. Usual work-up as in the manner given for compound (14) afforded a mixture of methyl (f)-lysergate (21) and methyl ()-isolysergate (22) (3 mg, 60) in the ratio 3:2 (Found M+, 282.1389. C, ,H *N202requires M, 282.1368). This mixture was separated by h.p.1.c. SO0 p.s.i.; 3 MeOH-CHCl, (0.8 ml min-')I: methyl (+)-lysergate: Rt 35.2 min, methyl ()-isolysergate: R,32.8 min. The n.m.r. spectra of (21) and (22) are superimpsable with those of natural methyl lysergate and methyl isolysergate respectively.Each isomer gave the same equilibrium mixture in methanol (21) :(22) 3 :21. Equilibrium Experiments.-Two solutions, one of methyl ( )-lysergate (21) (1 mg) and the other of methyl (+)-isolysergate (22) (1 mg) in methanol (1 ml each), were refluxed under a nitrogen stream. After 3 h, the reaction mixtures were found to isomerise to an equilibrium mixture of these two esters with identical ratio (3: 2) with methyl ()-lysergate (21) the major product. The ratio was not changed even after prolonged refluxing. The isolated 8P-ester (17) was also isomerised in a similar manner as above to give an equilibrium mixture of the 8P-ester (17) and 8a-ester (18) in the ratio 3: 1.Acknowledgements We thank the Ministry of Education, Science, and Culture (Japan) for a research grant, Professor R. Ramage, UMIST, for gifts of spectra of the authentic samples, and Professor M. Ohmomo, The University of Tsukuba, and Dr. P. A. Stadler, Sandoz Ltd., for their gifts of natural alkaloids. We are also grateful to Sir Derek H. R. Barton, I.C.S.N., for his advice and encouragement. References 1 Part 23, T. Naito, Y.Tada, Y.Nishiguchi, and I. Ninomiya, J. Chem. SOC.,Perkin Trans. 1, 1985, in the press. 2 Preliminary communication, T. Kiguchi, C. Hashimoto, T. Naito, and I. Ninomiya, Heterocycles, 1982, 19, 2279. J. CHEM. soc. PERKIN TRANS. I 1985 3 I. Ninomiya, C. Hashimoto, T. Kiguchi, and T. Naito, J. Chem. SOC., Perkin Trans. I, 1984, 291 1. 4 (a) E. C. Kornfeld, E.J. Kornefeld, G. B. Kline, M. J. Mann, D. E. Morrison, R. G. Jones, and R. B. Woodward, J. Am. Chem. Soc., 1956,78,3087; (b)M.Julia, F. Le Goffic, J. Igolen, and M.Baillarge, Tetrahedron Lett., 1969, 1569; (c)R. Ramage, V. W.Armstrong, and S. Coulton, Tetrahedron, 1981, 37, supplement 1, 157; (d) W. Oppolzer, E. Francotte, and K. Battig, Helu. Chim. Acta, 1981, 64, 478; (e)J. Rebek, Jr., D. F. Tai, and Y.K. Shue, J. Am. Chem. Soc., 1984, 106, 1813. 5 M. Natsume and H. Muratake, Heterocycles, 1981, 16, 1481; W. Oppolzer, J. I. Grayson, H. Wegmann, and M. Urrea, Tetrahedron, 1983,39, 3695. 6 D. E. Nichols, J. M.Robinson, G. S. Li, J. M.Cassady, and H. G. Floss, Org. Prep. Proced. Int., 1977, 9, 277. 7 P. A. Stadler, A. J. Frey, F. Troxler, and A. Hofmann, Helu. Chim. Acta, 1964, 47, 756. 8 R. Ray and D. S. Matteson, Tetrahedron Lett., 1980, 21, 449. 9 (a) S. Ohmomo, Bokin Bobai, 1982, 10, 253; (b) B. Arnoux, M.A. Merrien, C. Pascard, J. Polonsky, and P. M. Scott, J. Chem. Res. (S), 1978, 210. 10D. H. R. Barton, X.Lusinchi, and P. Millet, Tetrahedron Lett., 1982, 23, 4949. 11 Z. Horii, T. Watanabe, T. Kurihara, and Y.Tamura, Chem. Pharm. Bull., 1965, 13, 420. 12 S. Smith and G. M. Timmis, J. Chem. SOC.,1936, 1440. 13 T. Kiguchi, C. Hashimoto, T. Naito, and I. Ninomiya, Heterocycles, 1984,22, 1719. 14 K. Bailey and A. A. Grey, Can. J. Chem., 1972,50, 3876. 15 T. A. Crabb, R. F. Newton, and D. Jackson, Chem. Rev., 1971,71,109. 16 A. Stoll, T. Petrzilka, J. Rutschmann, A. Hofmann, and H. H. Giinthard, Helv. Chim. Acta, 1954, 37, 2039. Received 6th August 1984; Paper 411392