Tcuahcdron ~eucn. Vol. 33, No. 51, PP. 78957898.1992 Printed in Great Britain
0040-4039/92 $5.00 + 1 Porgamon Press Ltd
A Radical Approach to the Synthesis of (+)-Supinidine
Yeun-Min Tsai,* Bor-Wen Ke, Chain-Ting Yang and Chao-Hsiung Lin
Department of Chemistry, National Taiwan University, Taipei 107. Taiwan, Republic of China
Key Words: radical cyclization: a-sulfonyl radical: k)-supinidine; pyn~hzidi~. alkaloid: allylstannane
Ablroct: Intramolecular addition of a-sulfonyl radicals to triple bonds followed by addition of tin radicals to the resulting allylsulfones gave pyrrolizidine skeletons. Subsequent manipulations led to a formal synthesis of Q- supinidine.
pyrrolizidine alkaloids are an interesting class of compounds that exhibit a wide range of pharmacological activities.t-5 A veIy common structural subunit of the necine bases features an allylic alcohol moiety such as shown in the most important necine, retronecine (1). The necine bases of this type differ in the degree and stereochemistry of hydroxylation especially at C-6 and C-7 as in crotanecine (2) and supinidine (3).5 Coincide with our interest in free radical cyclization reaction6 involving a-sulfur functionalities,7*8 we felt that one might use this strategy in the construction of the allylic alcohol subunit. In this letter we wish to report the realization of this radical approach in the formal synthesis of &)-supinidine (3), the most simple necine base of its class.
1 mtronecine 2 Cr0tanecine 3 supinidine
As shown in scheme I, our original plane involved a key radical cyclization reaction to construct the pyrrolizidine skeleton followed by an allylic sulfoxide rearrangement9 to generate the desired allylic alcohol structure. Thus, alkylation of succinimide (4; scheme II) with 2-bromoethyl phenyl sulfidelo gave the imide 5
Scheme I
n = 0.1.2
7896
Scheme II
0 0
4
bC 5 X,Y = 6 X,Y = H,OEt
=o
0 7 X=H,Y=OH 8 X=Y=H e,f hori )
gcl; z:y;h ;_=;
11 X = Cl, ’ n = 1 12 X =Cl, n =2 13 X=H, n=l 14 X=H, n=2 15 16(a) NaH, DMF; BrCHzCHzSPh, room temp., 20 h (b) NaBH4, H+, EtOH (c) H-CW-SiMes (1.5 equiv), rrBuLi (1.5 equiv), -70 ’ C, 70 min (d) NaBHsCN, MeOH, pH 3, room temp., 1 h (e) NCS, CCl4, room temp., 16 h (f) MCPBA (3.5 equiv), CH,Clz, room temp., 4 h (g) PhSH, ZnCl,, Ccl,, room temp., 15 min (h) BusSnH (2.2 equiv), AIBN (0.1 equiv), PhH, 80 “C (i) BusSnH (1.5 equiv), AIBN (0.1 equiv), PhH, 80 “C
in 79% yield based on the bromide. Reduction of the imide 5 with sodium borohydride gave the lactam 6 with no problem.1
I
However, attempted amidoalkylation t2-14 of 6 with trimethylsilylacetylene orbisnimethylsilylacetylene in the presence of a Lewis acid met with failure. Fortunately, trimethylsilyacetylide addition to 515 followed by reduction of the resulting alcohol 7 with sodium cyanoborohydride produced the sulfide 8 in 59% yield with the recovery of 32% unreacted 5 (91% yield based on reacted 5). Chlorination of 8 with N-chlorosuccinimide (NCS) followed by substitution of the Cl atom in the resulting chlorosulfide 9 with a thiophenoxy group gave the dithioacetal
10
in 95% yield.7Radical cyclization carried out by the slow addition (4 h) of tributyltin hydride (0.4 M in benzene; 1.5 equiv) to a benzene solution of 10 (0.4 M) heated at 80 oC. Subsequent heating for another 18 h gave instead of the desired allylic sulfide 17, the two pynolizidines 15 (28%) and 16 (29%). In addition was isolated the uncyclized reduction product 8 (16%) together with unreacted 10 (6%). The formation of 15 and 16 indicates that the initial cyclization proceeded as expected to give 17 (Scheme III). Further reduction of the allylic sulfide
Scheme III
7897 Scheme IV b C \ e O 22
I
d SiMe, ’ 23 (a) PyHBr*Brl, CH$&, 0 ‘C, 4 h (b) NaOH, H20,80 “C, 12 h (c) PhSeBr (1.5 equiv), CH,CI,, -78 ‘C (d) PhSeBr (1.2 equiv), CH$l,, room temp., 1 h (e) PhSeBr (2.5 equiv), CH#,, -78 ‘C + room temp.moiety in 17 with tributyltin hydride gave 16. 16~7 Addition of the tributyltin radical to the vinyl silane moiety in 17 followed by p-scission of the carbon-sulfur bond gave 15. 18.19 This result indicates that the above mentioned two processes are competing with the formation of 17 from 10. In contrast, the chlorosulfoxide 11 derived from 9 gave only 15 and 13 under the same reaction conditions.8 Although the annoying over- reduction was eliminated, the extra chiral center in sulfoxide caused complication of structure identification and product isolation. We found that the reaction of chlorosulfone 12 derived from 9 (84%), with excess tributyltin hydride gave 1520 in 72% yield together with 14.8 With the lesser amount of the stannane the primary cyclization product 18 could be isolated along with 15.
Since 15 can be. obtained so easily, we decided to convert 15 to Q)-supinidine (3). Treatment of 15 with pyridinium bromide perbromide (scheme IV) gave 19 in 98% yield as a mixture of steteoisomers. Note that it is the allylstannane moiety that reacts and the chemoselectivity is excellent.21 When 19 was subjected to hydrolytic condition the allylic alcohol 2117 was obtained directly in 42% yield along with 31% of 20.22 We believe that the formation of 21 involves a silicon directed SN2’ displacement23 of the allylic bromide followed by Brook rearrangement24 and desilylation of the resulting ally1 silyl ether. Alternatively, treatment of 15 with 1.5 equivalent of phenylselenium bromide at -78 Oc gave a 9/l mixture of 22 and 23 in quantitative yield. Pure 22 reacted with phenylselenium bromide (1.2 equivalent) at room temperature to afford 23 in 97% yield.25 More conveniently, one can perform this interesting transformation by direct reaction of 15 with 2.5 equivalent of phenylselenium bromide at -78 oC. and then warming up to room temperature to give 23 in quatitative yield as a 2/l mixture of diastereomers. Under the same conditions for hydrolysis of 19, the bromide 23 was converted to 21 in 59% yield. 25 Since 21 has been transformed to &)-supinidine (3).‘7 our method constitutes a formal synthesis of @-supinidine (3).
In summary, although the final synthesis deviated from our original plane, our method still provides an interesting entry to the synthesis of supinidine. Further application of this method directed to the synthesis of the more complicated necine bases is under way.
7898
Acknowledgements. Financial support by the National Science Council of the Republic of China is
gratefully acknowledged (Grant 81-0208-M-002-01).
References and Notes
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
Mattocks, A. R. in Chemistry and Toxicology of Pyrrolizidine Alkaloids; Academic Press: New York, 1986.
Robins, D. J. Nat. Prod. Rep. 1991,8,213 and references cited therein.
Ikeda, M.; Sato, T.; Ishibashi, H. Heterocycles 1988.27, 1465, and references cited therein. Robins, D. J. Chem. Sot. Rev. 1989,18, 375.
Robins, D. J. frog. Chem. Org. Nat. Prod. 1982,41, 115. Curran, D. P. SynLett 1991, 63 and references cited therein.
Tsai, Y.-M.; Chang, F.-C.; Huang, J.; Shiu, C. L. Tetrahedron Lett. 1989,30, 2121.
Tsai, Y.-M.; Ke, B.-W.; Lin, C.-H. Tetrahedron Lett. 1990.31, 6047.
Evans, D. A.; Andrews, G. C. Act. Chem. Res. 1974, 7, 147.
This bromide was prepared (83% yield) by slow addition of an ethanolic solution of sodium thiophenoxide to ethylene dibromide (1.3 equiv) in ethanol.
Hubert, J. C.; Wijnberg, J. B. P.; Speckamp, W. N. Tetrahedron Lett. 1975,31, 1437.
Zaugg, H. E. Synthesis 1984,85, 181.
Casara, P.; Metcalf, B. W. Tetrahedron Lett. 1978, 1581.
Rena&I, P.; Seebach, D. Atlgew. Chem. lnt. Ed. Engl. 1986.25, 843.
Wr6be1, I. T.; Cybulski, J.; Darbrowski, Z. Synthesis 1977, 686.
Gutierrez, C. G.; Summerhays, L. R. J. Org. Chem. 1984,49, 5206.
For the preference of an errdo-cyclic double bond in similar system, see: Burnett, D. A.; Choi, J.-K.; Hart, D. J.; Tsai, Y.-M. J. Am. Chem. Sot., 1984,106, 8201.
For radical additions to allylsulfones, see: (a) Ueno, Y.; Aoki, S.; Okawara, M. J. Am. Chem. Sot. 1979,101, 5414. (b) Smith, T. A. K.; Whitman, G. H. J. Chem. Sot. Perhitz TransJ 1989, 313,
319. (c) Giese, B.: Linker, T. Synthesis 1992, 46.
For a review about the fragmentation process, see: Curran, D. P. Syrrthesis 1988,489.
lH NMR spectrum (300 MHz, CDCl3) of 15: 6 0.02 (s, 9 H, SiCHj), 0.86 (t. J = 7 Hz, 9 H, CH3 of butyl), 1.20-1.55 (m, 19 H), 1.6-1.78 (m, 1 H, COCH$&), 2.16-2.36 (m, 2 H, COC&C&), 2.55- 2.75 (m, 1 H, COC&), 3.68 (br d, J = 15 Hz, 1 H, NC&C=), 4.35 (br d, J = 15 Hz, 2 H, NC&C!= and NCH), 5.07 (s, 1 H, =CH). All new compounds mentioned give satisfactory tH NMR, 13C NMR, IR and HRMS.
This provide a rare case for direct comparison of the reactivity of allylsilane and allylstannane. Alcohol 20 is a mixture of at least three diastereomers with a ratio of 2/3/5 determined by lH NMR integration of the signals at vinylic position.
For the silicon directing effect, please see: Fleming, I. in Comprehensive Organic Chemistry: Barton, D.; Ollis, W. D.; Jones, D. N., Eds.; Pergamon Press: New York, 1979; Vol. 3, pp 546.
Brook, A. G. Act. Chem. Res. 1974.7, 77.
Exploration of this new reaction is under way in our laboratory.