sign in sign up
<<
Home , Acetate, Anhydrous, Convergent synthesis, Deoxygenation, Ferrier rearrangement, Glycol cleavage

1262 COMMUNICATIONSTO THE EDITOR Tol. i9

methylation,s was converted into the trans product followed by crystallization from - IV, m.p. 202-204", C, 78.9; H, 8.33, in 69%.yield. containing an equivalent of pyridine, led to 75q) Alkaline oxidation4 transformed IV into V of D-a-phenoxymethylpenicilloicacid hydrate (IV), (R = H) which was converted with diazomethane CI~HZONZO~S.H~O;m.p. 129" dec. [Found: C, into the V (R = CHa), and cyclized with 49.61; H, 5.77; N, 6.94; aZ5D + 94" (c, 1 in potassium t-butoxide in . The resulting )]. Identity with a sample prepared by keto ester was decarbomethoxylated with hydro- saponification of natural penicillin V5 was estab- chloric and acetic to give the dl- VI, lished by comparison of m.p., infrared spectra m.p. 155.5-161.5". The infrared spectrum of this (KBr), optical rotation and mixed m.p. material was indistinguishable from that of Treatment with N,N'-dicyclohexylcarbodiimide authentic 3,B-hydroxy-9,1l-dehydroandrostane-17- in dioxane-water (20 min. at 25') cyclized Was the one.g monopotassium in l(rl27, yield. By partition (S) At this stage the 3-hydroxyl group was protected as the tetra. between methyl isobutyl ketone and pH 5.5 phos- hydropyranyl (cf. ref. 3). phate buffer (two funnels) the totally synthetic (9) C. W. Shoppee, J. Ckem. SOC.,1134 (1946). crystalline potassium salt of penicillin V was iso- DEPARTMENTOF lated. The natural and synthetic potassium salts UNIVERSITYOF WISCOXSIN WILLIAMS. JOHNSON were shown to be identical by microbiological as- MADISON,WISCONSIN ALLEN,JR. DUFFS. say,6 optical rotation [synthetic, d5D -/- 223" RECEIVEDFEBRUARY 1, 1957 (c, 0.2 in water); natural, a25~+ 223" (c, 0.2 in water); reported,' Q?OD + 223" (c, 1 in water)], in- THE OF PENICILLIN V frared spectra (KBr), m.p. 263" dec. (reported,' Sir: 256-260" uncorr.), undepressed upon admixture. The ability of aliphatic to form s amide bonds in aqueous solution directly from the /\ (1) C6HjOCH2COC1 amine and carboxyl components under very mild HCl'HzNCH-CH C(CH3)z (2) HC1 //I > conditions' suggested the use of these reagents for CO XH-CHCOZH (3) CsHsN the cyclization of a pencilloic acid to a penicillin. I We have Drepared by total synthesis in good over- OC(CHa)3 all yield the penicilloic acid corresponding to peni- 11, D-CY cillin V (phenoxymethylpenicillin). By use of N ,N'- S /\ dicyclohexylcarbodiimide cyclization was effected C6HsOCHzCOXHCH-CH C( CH3)z rapidly at room temperature, thereby completing //I the first rational synthesis of a natural penicillin.2 CO NII-CHCOzH Condensation of D-penicillamine with t-butyl I (1) KOH (one equiv.) OH f phthalimidomalonaldehydate afforded the t-butyl (2) CJ~lli\i'. C.=NCJTI, ~-a-4-carboxy-5,5-dimethyl-a-phthalimido-2- thi- I17 azolidineacetate (I), C2d&4N2O6S, m.p. 161" dec. S [Found: C, 57.45; H, 6.06; N, 6.83; aZ5~ /\ + 54" (c,l in )] as described for the cor- CeHjOCHCONHCH-$If C(CII3)2 responding DL-a The a, or natural, CO-k--&HCO&I configuration of the more soluble (-water) penicillin V potassium I was established chemically by relationship to (potassium phenoxymethylpenicillinate) natural dimethyl D-cr-benzylpenicilloate. The less soluble D-y-iSOmer may be isomerized in high yield The same results were obtained using IV derived to the D-aform as in the DL-ester ~eries,~thus pro- from natural penicillin V. The entire series also has been carried through starting with DL-penicil- viding a stereochemically efficient synthesis. Hy- drazinolysis of I, followed by acidification with lamine. The crystalline DL-penicillin V potassium hydrochloric acid, produced t-butyl ~-a-4-car- salt showed 51.4% (514u/mg.) of the bioactivity hoxy-5,5-dimethyl- a-amino -2-thiazolidineacetate of natural penicillin V, indicating that L-penicillin hvdrochloride (11), Cl2H&zOnSC1, in 85% yield; V has little, if any, antibiotic activity. Cycliza- m.p. 172" dec. [Found: C, 43.83; H, 7.18; C1, tion of the penicilloate also was effected, but in lower yield, by ethoxyacetylene and a ketenimine 10.87; CX~~D+ 111" (c, 1 in methanol)]. Phenoxyacetyl chloride and triethylamine con- (pentamethyleneketene cyclohexylimines). It is verted I1 to a-t-butyl D-a-phenoxymethylpenicil- interesting to note that the entire reaction se- hate (III), Cd&sN&S, in 75% yield; m.p. quence starting with penicillamine was conducted 120-122' dec. [Found: C, 56.85; H, 6.86; N, at or below room temperature. Ci.Ti9; aZ5D+ 67" (c, 1 in methanol)]. Cleavage We are indebted to Bristol Laboratories of of the t-butyl ester with dry chloride, Syracuse, NX., for financial support, to Merck and Co., Inc., of Rahway, N. J., for the preparation (1) J. C. Sheehan and G. P. Hess, THISJOURNAL, 77, 1067 (1965). (2) Penicillamine and 2-benzyl-4-methoxymethylene-5-(4)-oxazolone (6) Kindly furnished by Eli Lilly & Company, Indianapolis, Ind. condense to form trace amounts (0.03 to 0 08% by bioassay, 0.008% (6) Synthetic potassium penicillin V had a potency of 1078 dmg. i isolated) of penicillin G (benzylpenicillin). For a recent review of this 1Oyo (107.870 i lOYc) compared to standard natural penicillin V in a reaction see Karl Folkers in "Perspectives in Organic Chemistry," plate diffusion assay carried out under the supervision of Dr. J. Lein, Sir Alexander Todd, Editor, Interscience Publishers, Inc., New York, Bristol Laboratories, Syracuse, N. Y. N. Y.,1956, p. 409. (7) E. Brand1 and H. hiargreiter, Ostew. Chenz. 2.. 65, 11 (1954). (3) J. C. Sheehan and D. A. Johnson, THISJOURNAL, 76, 168 (8) Directions for the preparation of this ketenimine were fur- (1954). nished by Dr. C. L. Stevens, TVayne University. private communica- (4) J. C. Sheehan and P. A. Cruickshank, ibid., 78, 3677 (1956). tion. March 5, 1957 COMMUNICATIONSTO THE EDITOR 1263

of substantial quantities of certain key interme- 935. diates and to Mr. Sergey V. Chodsky for technical c, assistance. $34- 3.3- DEPARTMENTOF CHEMISTRY JOHN C. SHEEHAN < a MASSACHUSETTSINSTITUTE OF TECHNOLOGY CAMBRIDGE39, MASS. KENNETHR. HENERY-LOGAN 63.2- RECEIVEDFEBRUARY 11, 1957

is- THE SALT EFFECT IN THE AROMATIC NUCLEOPHILIC SUBSTITUTION REACTION' & Sir : The effect of added neutral salts upon the veloc- ity of the second order of the -dipole aromatic nucleophilic substitution reactions of lithium, SO- dium and potassium with 2,4-dinitro- has been investigated at 25". The rates were studied in absolute methanol as a function of reactant (LiOCHs, NaOCHa, and KOCH3) in the presence of added cations (Lif, Na+, and K+) and added anions (C2H-02-, I-, Br-, C104-, C1-, and NO9-). The reaction of NaOCH3 in the presence of added LiC104.3HzO also was studied in a 50 volume % methanol- benzene solvent. I .9; For reactions without added salts, the rate con- 1.81 stants (1 mole-' sec.-l) were: LiOCHI, 0.0242; 0.0 0.025 0.05 0.10 0.15 0.20 NaOCHS, 0.0262; KOCH3, 0.0278. A consistent MOLARITY OF ADDED' SALT. pattern of salt effects is typified by the data for the LiOCH3 reaction shown in Fig. 1. At low concen- Fig. 1.-Lithium and 2,4-dinitrochlorobenzene. trations of added salt, each cation exhibits a.n in- sociation occurs for LiOCH, than for KOCH3 or dividual effect, added to that of the cation intro- NaOCH3 in methanol. Potassium salts are strong duced along with the reactant methoxide. The electrolytes in methanol with dissociation con- anions cause an additional secondary effect. The stants of about 0.1 to 0.0Z5 It is known that reaction rate increases for > C1-, Br- > potassium salts are stronger electrolytes than are I-, NOS- > Clod-. Salt effects are more pro- lithium salts in acetone.6 If a similar order of nounced in of lower dielectric constant. electrolyte strength holds for methanol solutions, The observed effects cannot be correlated with then the effect of added potassium salts on the changes in ionic strength of the reaction medium LiOCH3 Li+ + -OCH3 equilibrium would be to as found by Bolto and Miller.* supply anions which would tend to associate more A qualitative explanation of the effect of lithium readily with Li+ so that the equilibrium would be salts assumes the equilibrium shifted to provide a greater concentration of OCH3-. LiOCH3 If Li + + -0CH3 This accounts for the increase in rate of the reac- The addition of a salt providing Li+ as a common tion. Sodium salts are not as effective as potas- ion should shift this equilibrium to decrease the sium salts, and the anion effects are consistent with concentration of the reactant, OCH3-. Since the those observed in the presence of Li+ alone. effective concentration of added Li+ will depend (5) E. C. Evers and 8. 0. Knox, THIS JOURNAL, 73, 1739 (1951). on the degree to which it remains associated with (6) J. F. Dippy, H. 0. Jenkins and J. E. Page, J. Ckem. Soc., 1368 the added anion, the rate will differ with different (1939). added salts. This assumes that the ion pair reacts JOHN D. REINHEIMER WILLIAMF. KIEFFER at a negligible rate compared to that for the ion. THE COLLEGEOF WOOSTER STANLEYW. FREY A similar interpretation has been used to account ~T'OOSTER, OHIO JOHN C. COCHRAN for the variation in rate of decarboxylation of tri- EDWARDW. BARR chloroacetic acid.3 The observed effect of anions on RECEIVEDNOVEMBER 16, 1956 reaction rate thus can be interpreted to suggest that the order of attraction for lithium in THE EFFECT OF NITRATE ION ON THE YIELD OF methanol is Ac- > C1-, Br- > NO3-, I- > Clod-. HYDROGEN FROM WATER RADIOLYSIS The fact that NaOCH3 and KOCH3 react faster sir : suggests that the corresponding equilibria involving Solutions of calcium nitrate have been irradiated these methoxides is shifted more to the right, pro- in the mixed fast neutron-y-flux of the Harwell ex- viding a greater effective concentration of OCH3-. perimental reactor BEPO at a temperature of about Conductivity data4 suggest that more ion as- 80'. Nitrate concentration was varied from 15.9 (1) This research supported by the Petroleum Research Fund of to 0.037 M. The thermal neutron dose was moni- the American Chemical Society. tored using cobalt wire of high purity.l Energy (2) B. Bolto and J. Miller, Ausfrolian J. Chcm., 9, 74 (1956). deposition figures were derived using the data of (3) G. A. Hall and F. H. Verhoek, THIS JOUXNAL, 69, 613 (1947). (4) G. E. M. Jones and 0. L. Hughes, J. Chcm. Soc., 1197 (1934). (1) J. Wright, to be published. June 20, 1959 TOTALSYNTHESIS OF PENICILLINV 3089

[CONTRIBUTION FROM THE DEPARTMENTOF CHEMISTRY, MASSACHUSETTSINSTITUTE OF TECHNOLOGY] The Total Synthesis of Penicillin V

BY JOHN C. SHEEHANAND KENNETHR. HENERY-LOGAN RECEIVEDDECEMBER 4, 1958

Potassium phenoxymethylpenicillinate (VIII), synthesized totally in a series of reactions from D-penicillamine (0-11) and t-butyl phthalimidomalonaldehydate (I), has been shown to be identical to natural penicillin V (potassum salt) in physi- cal and biological properties. In the key step, the monopotassium salt of the penicilloic acid (D-CY-VII)wdS cyclized by means of N,N’-dicyclohexylcarbodiimide. A similar series of reactions starting with DL-penicillamine (DL-11)produced crystalline DL-penicillin V, which was resolved to give the hitherto unknown L-penicillin V having less than 1% of the bio- activity of the natural enantiomer. The natural and synthetic series were related stereochemically at three points, viz., IV, VI1 and VIII. A promising intermediate for the synthesis of penicillins, penicillin analogs and penicillanic , vzz., t-butyl D-a-~-carboxy-5,5-dimethyl-~-amino-2-thiazolidineacetatehydrochloride (V), was prepared. Acylation of D- and DL-WVwith phenoxyacetyl chloride, followed by cleavage of the t-butyl ester, afforded the penultimate penicilloic acids D- AND DL-a-VII.

Penicillin was discovered by Fleming in 1929.l ) have failed,“ which is not sur- The remarkable in vivo activity of the antibiotic prising in view of the known instability of the de- against a variety of pathogenic organisms was first sired product (penicillin) in the presence of acidic demonstrated by Chain, Florey and co-workers. reagents and byproducts of the reaction. The dis- Extensive degradative and physical studies during covery that aliphatic carbodiimides are capable of the wartime cooperative program between Ameri- forming amide bonds in aqueous solution directly can and British scientists culminated in the pro- from the amine and carboxyl components under posal of the fused @-lactamthiazolidine structure very mild conditiond2 suggested the use of these for the penicillins (VIII).3 This same intensive reagents for the cyclization of a penicilloic acid (VII) collaborative effort directed toward the synthesis to a penicillin (VIII). By the use of N,N’-dicyclo- of penicillin resulted, however, in the formation of hexylcarbodiimide cyclization was effected readily penicillin only in minute yield.4 at room temperature to give totally synthetic peni- In 1948, this laboratory embarked on a substan- cillin V in both the natural and racemic series. tial program, the objective of which was to devise This communication also describes the prepara- synthetic methods powerful and selective enough to tion of important intermediates for a penicillin overcome the “diabolic concatenations of reac- synthesis carried through without a blocking group tive groupings“6 and thereby to make possible a on the @-carboxylof the penicilloic acid VII. This total synthesis of the penicillins and simpler struc- feature obviates the necessity of removing a pro- tural analogs. Within a few years three new p- tective group (e.g., a benzyl ester by catalytic hy- lactam synthesese+?were developed including one drogenolysis) in a last step.g Thus the key inter- which led to the formation of a 5-phenylpeni~illin~mediate V presents attractive possibilities for syn- having many of the chemical and physical proper- thesis of a variety of natural and unnatural peni- ties of the natural penicillins. Recently, the synthe- cillins differing in the acyl on the side- sis of a biologically active “sulfonyl analog” of ben- chain amino group. zylpenicillin (penicillin G) has been reported. We Stereoisomerism of Penicilloic Acids.-In the now wish to record the first rational synthesis of a condensation of I with D-I1 two new asymmetric natural peni~illin.~~ centers are formed and it is necessary to determine Many attempts directed toward the cyclization which, if either, of the two thiazolidines formed (of of penicilloates of type VI1 with acid halide- and the four theoretically possible) corresponds in acid anhydride-forming reagents (e.g., thionyl chlo- configuration to the natural D-a-penicilloates, ride, phosphorus trichloride, and Comparisons were made at three points in the syn- thetic sequence, viz., at D-CY-IV,-VI1 and -VIII. (1) A. Fleming, Brig. J. Ex#. Patkol., 10, 226 (1929). (2) E. Chain, H. W. Florey, A. D. Gardner, N. G. Heatley, M. A. The DL-WIVisomer had been shown previously Jennings, J. Orr-Ewing and A. G. Sanders, Lancef, 239, 226 (1940). by Sheehan and Cruickshank13 to correspond in (3) H. T. Clarke, J. R. Johnson and R. Robinson, editors, “The configuration to the natural D-a-penicilloates; this Chemistry of Penicillin,” Princeton University Press, Princeton, N. J., assignment was confirmed in the present work by 1949, p. 454. (4) Penicillamine and 2-benzyl-4-methoxymethylene-5(4)-oxazo- the conversion of DL-cY-IIIinto DL-CU-VIII,the natu- lone condense to form trace amounts (0.03-0.08% by bioassay, ral diastereomer. 0.008% isolated) of penicillin G (benzylpenicillin). For a recent Assignments of configuration to D-wIV and D- review of this reaction see Karl Folkers in “Perspectives in Organic Chemistry,” Sir Alexander Todd, Editor, Interscience Publishers, y-IV were made on the basis of detailed comparison Inc., New York, N. Y, 1956, p. 409. of infrared spectra with those of DL-CY-IVand DL- (5) R. B. Woodward in “Perspectives in Organic Chemistry,” Sir yIV. The crystalline a-isomers also melted 50” Alexander Todd, Editor, Interscience Publishers, Inc., New York, N. higher than the corresponding y-isomers. The di- Y., 1956, p. 160. acid hydrate D-CX-VIIwas shown to be identical with (6) J. C. Sheehan and P. T. Izzo, THISJOURNAL, 70, 1985 (1948); 71, 4059 (1949). the penicilloate obtained from the alkaline hydroly- (7) J. C. Sheehan and A. K. Bose, ibid., 72, 5158 (1950). sis of penicillin V in physical properties, including (8) J. C. Sheehan, E. L. Buhle, E. J. Corey. G. D. Laubach and J. J. optical rotation. Ryan, ibid., 73. 3828 (1950); J. C. Sheehan and G. D. Laubach, ibid., 73, 4376 (1951). (11) Reference 3, p. 861. (9) J. C. Sheehan and D. R. HOB, ibid., 79, 237 (1957). (12) J. C. Sheehan and G. P. Hess, THIS JOURNAL, 77, 1067 (10) A preliminary communication has been published, J. C. Shee- (1955). han and K. R. Henery-Logan, ibid., 79, 1262 (1957). (13) J. C Sheehan and P. A. Cruickshank, ibid., 78, 3677 (1956). 11, D- and Dr-isomers OC(CH3);:

111, D-a-,~-y-, ILCY md DL-7-isomers ' 1, X2H4 42, aq. HCl S /\ CbHaOCH~COSHCH-CH C( CHj)i CsH 50 CHLCOC1 HC1 iiLNCII--CH C(CH3)- I 1 P I I I CO KH-CHCOJI (C2HshN CO SH-CHCOZH I I OC(CH3h OC( CHZi? VI, D- and DL-a-isomers T', D- and DL-a-isomers 1, HC1 .1 2, CBHSN S /\ C6HdOCH~COSHCH---CH C( CH3)Z I I I CO KH-CHCOsH I OH 1-11, D- and Dr,-e-twiiiers 1.111, penicillin V potassium (potassium D- and DL-phenoxy- methylpenicillinate) The third comparison was made by showing the ilarly, the DL-isomer of V gave the racemic thiazol- identity of totally synthetic D-VI11 with natural idine DL-a-VIin 79yo yield. penicillin V potassium salt (vide ixfva). Inciden- The marked lability of t-butyl toward an- tally] the conversion of natural D-a-VI1 to VI11 hydrous acids14 permits the facile cleavage of the established for the first time that the reverse re- carbo-t-butoxy group in compounds VI to give the action, the alkaline of VIII, occurred P-amino acid hydrochlorides. Treatment of meth- without epimerization. Consequently the many ylene chloride solutions of D- and DL-0-VIwith an- wartime attempts at penicilloic acid cyclization did hydrous at 0" liberated the a-car- not fail because of operating in the wrong stereo- boxyl function in almost quantitative yield. Re- chemical series. crystallization of these hydrochlorides from acetone- Synthesis of D- and DL-a-Penicilloic Acids.-The water containing an equivalent of pyridine afforded interaction of D-penicillamine hydrochloride (D- the penicilloic acids D- and DL-a-VII. Identity of 11) and t-butyl phthalimidomalonaldehydate (I) the D-a-phenoxymethylpenicilloic acid hydrate (D- in buffered aqueous ethanol af- a-VII) with a sample prepared by saponification forded directly the crystalline thiazolidine D-7-111 of natural penicillin V was established by conipari- (30%). The a-isomer, which separated only on son of m.p., mixed m.p., infrared spectra (potas- addition of water, appeared to be uncontaminated sium bromide) and optical rotation. with the y-isomer and was isolated in slightly larger Cyclization of Penicilloic Acids to Penicillin V and amount (23y0)than in the DL-series.13 DL-Penicillin V.-Cyclization of the penicilloates Additional quantities of the desired D-a-isomer of 171 as the monopotassium (or monosodium) salts I11 were prepared by heating a pyridine solution of was found to take place readily at room tempera- the y-isomer. This procedure established an equi- ture in dioxane-water solution. The p-lactam ring librium consisting of about 25y0 of D-a-111, which closure could be effected rapidly (20 min.) with one crystallized directly on cooling the solution. Addi- equivalent of N,N'-dicyclohexylcarbodiimide. tional quantities of the a-isomer could be obtained Higher yields were achieved with longer reaction by recycling the filtrate. The DL-y-isomer was also times in very dilute solutions (0.37G)1using four isomerized in pyridine to DL-CY-111. equivalents of . In an experiment Reinoval of the phthaloyl group from D-o(-I11was run for 33 hours using natural phenoxymethylpeni- accomplished by the action of hydrazine below room cilloic acid hydrate (D-~-VII)'~the yield of penicil- temperature to yield the phthalhydrazide complex (14) J. C. Sheehan and G. I). Laubach, THISJOURNAL, 73, 4752 isolated by lyophilization. A suspension of the il95l). complex in acetic acid was treated with hydro- (1:) All samples uf naiurd jihcnoxymethylj~euicilloicacid hydrate made by the alkaline hydrolysis (PI< 11.5) of penicillin V by the pro- chloric acid to afford D-CY-Viii 52% yiFld. Similar cedure described in the Experimental section contained no penicillin V treatment of the DL-isomer with hydrazine gave DL- as determined by bioassay18 and by chemical assay." a-V in 8570 yield. (16) The synthetic samples mere compared to standard natural The D-iSOIller of I-, upon treatment with one penicillin V in a plate diffusion assay carried out under the supervision uf Dr. J. Lein, Bristol Lahoratories, Syracuse, N Y. Bioassays of crude equivalent of phenoxyacetyl chloride and two samples of synthetic penicillin V lend to decrease on storage, even equivalents of triethylamine at O", afforded in 70% ;rt so, which may account for the discrepancy between chemical and yield the phenoxymethylpenicilloate D-wT'I. Sitn- micr,,hiological assay. June %U, 1959 TOTALSYNTHESIS OF PENICILLINV 3091 lin V in the partially purified product was 11% by tion in 0.19% yield. N,N'-Carbonyldiimidazole, chemical assay17 and 9% by bioassay.l6 In a a new peptide-formingreagent,zzgavea product with larger scale cyclization (4.6 g.) carried out for 22 no biological activity. l6 For cyclizations with re- hours the yield by chemical assay was 9% and by agents other than carbodiimides, however, no at- bioassay 6% and, after purification by partition tempt was made to develop optimum conditions between methyl isobutyl ketone and two phosphate for the reactions. The monosodium salt of the buffers, pure crystalline potassium phenoxymethyl- penicilloic acid corresponding to penicillin G was penicillinate was isolated in 5.4y0 yield. The natu- also cyclized with N ,N'-dicyclohexylcarbodiimide ral and synthetic potassium salts of penicillin V to penicillin G but in much lower yield. 23 were shown to be identical by microbiological assay Resolution of DL-Penicillin V.-The racemic (99.7y0 of the activity of natural penicillin V) and amine erythro-1,2-diphenyl-2-methylaminoethanol by physical properties. These yields are consid- is readily resolved since the levo-isomer forms a ered representative of the developed cyclization and sparingly soluble salt with penicillin Gz4; hence it isolation procedure, since they have been obtained follows that the optically active amine would re- consistently on a macro scale. A somewhat less solve penicillin G. Natural penicillin V formed a efficient process (first chronologically) was em- salt with the levo-aminez5which immediately crys- ployed in the totally synthetic and DL-series. tallized from water or n-. Reaction In a similar manner, cyclization of D-CY-VII,ob- of DL-penicillin V with the levo-amine in water solu- tained from D-penicillamine, proceeded rapidly (25 tion gave a salt of the D-acid but only after standing min.) at room temperature with one equivalent of many days at 5" (even with initial seeding with N,N '-dicyclohexylcarbodiimide to give a 416- traces of the natural salt). This salt was, however, 5yOl7yield of D-VIII. This sample of totally readily crystallizable from n-butyl acetate. The synthetic potassium salt of penicillin V was shown filtrate, containing the L-penicillin, was treated to be identical with the natural potassium salt by with the dextro-amineZ5in n-butyl acetate solution microbiological assay16 (108 f. 10% of the bioac- to give rapid crystallization of the L-penicillin salt. tivity of penicillin V potassium), optical rotation The D-penicillin levo-amine and L-penicillin dextro- [synthetic CY~~D+ 223" (c 0.2 in water); natural, amine salts had opposite rotations but were iden- CY~OD + 223" (C 1 in water)], infrared spectra (40 tical in other physical properties. peaks and shoulders in potassium bromide), m.p. The salt of D-penicillin from the resolution gave 263 dec. (reportedls 258-260' uncor.), undepressed the same bioassay value as the salt of natural peni- upon admixture. cillin V. The salt of L-penicillin, however, showed Formation of the P-lactam ring was also carried only 0.7% of the bioactivity expected for the salt out in the racemic series. Thus DL-CPVII cyclized of natural penicillin. It is entirely possible that a to form an optically inactive penicillin V potassium trace of D-isomer, present as a contaminant, is suf- salt, m.p. 244' dec. This salt exhibited 51% of ficient to account for the very low bioactivity ob- the bioactivity16 of natural penicillin V, strongly served, and that pure L-penicillin V might show no suggesting that the unnatural L-penicillin is devoid biological activity. of antibiotic activity. The infrared spectrum We are indebted to Bristol Laboratories of Syra- (potassium bromide) had a strong band at 5.64 p, cuse, N. Y., for financial support, to Merck and characteristic of the fused p-lactam-thiazolidine Co., Inc., of Rahway, N. J., for the preparation of carbonyl function, and was identical with the spec- substantial quantities of certain key intermediates trum of the natural potassium salt of penicillin V. and to Mr. Sergey V. Chodsky for technical assist- In parallel experiments N,N'-diisopropylcarbo- ance. diimide promoted cyclization of the D-a-penicilloate Experimentalz6 in essentially the same yield as with N,N'-dicyclo- &Butyl D- and DL-4-Carboxy-5,5-dimethyl-a-phthalimido- hexylcarbodiimide. It was also found possible to 2-Thiazolidineacetate (111).-To an ethanol solution (300 effect p-lactam closure with other amide-bond ml.) of 42 g. (0.146 mole) of t-butyl a-phthalimidomalon- forming reagents, which, like carbodiimides, have aldehydateZ7(I) was added a solution of 27.2 g. (0.146 mole) the property of being neutral themselves and giving of D-penicillamine hydrochloride2*(D-11) and sodium acetate rise to neutral by-products. Ethoxyacetylene has trihydrate (29.9 g., 0.22 mole) in 300 ml. of water. After been used by hrenslg for the formation of peptides nished by Dr. C. L. Stevens, Wayne University, private communica- in anhydrous solvents. Sheehan and Hlavka, zo tion. (22) G. W. Anderson and R. Paul, THISJOURNAL, 80, 4423 (1958). attempting to form a peptide in aqueous solution at (23) Preliminary experiments in this Laboratory by Dr. P. A. room temperature, found that phthaloylglycine and Cruickshank indicated that sodium D-a-benzylpenicilloate was cyclized ethoxyacetylene formed a reactive isolable peptide to sodium benzylpenicillinate (penicillin G) in small amounts (0.3% by intermediate, which could interact with ethyl gly- bioassay). (24) V. V. Young, J. Am. Phmm. Asroc., Sci. Ed., 40, 261 (1951); cinate to form phthaloylglycylglycine ethyl ester. W. B. Wheatley, W. E. Fitsgibbon and L. C. Cheney, J. Org. Chem.,18, Ethoxyacetylene cyclized the monosodium salt of 1564 (1953). the D-a-penicilloate in 0.29y0 yield. Pentamethyl- (25) Samples of levo- and dexfro-eryrhro-1,2-diphenyl-2-methyl- eneketene cyclohexylaminez1 also effected cycliza- aminoethanol hydrochlorides were kindly supplied by Dr. L. C. Cheney, Bristol Laboratories, Syracuse, N. Y. (17) J. H. Ford, Anal. Chem., 19, 1004 (1947). The procedure was (2fi) All melting points are corrected. We are indebted to Dr. S. hl. modified only by replacing the one volume of phosphate buffer and one Nagy and his associates for the microanalyses, and to Dr. N. A. Nelson volume of acetate buffer with two volumes of phosphate buffer; this and his associates for the infrared and ultraviolet spectra. change did not affectthe color yield from the sodium salt of penicillin V. (27) J. C. Sheehan end D. A. Johnson, Trrrs JOURNAL,76, 168 (18) E. Brandl and H. Margreiter, osferv. Chem. Zfg.,68, 11 (1954). (1954). (19) J. F. Arens, Rec. Irav. chim., 74, 769 (1955). (28) J. C. Sheehan, R. Mozingo, K. Folkers and M. Tishler, U. S. (20) J. C. Sheehan and J. J. Hlavka, J. Org. Chem., 23, 635 (1958). Patents 2,496,418 and 2,496,417; B. E. Leach and J. H. Hunter, (21) Directions for the preparation of this ketenimine were fur- Biochervi. Prepnuofions. 3, 111 (1953). 3092 JOHN C. SHEEHANAND KENNETHR. HENERY-LOGAN Vol. 81 storage for 10 hours, 18.2 g. (30%) of crystals was collected m.p. 176-177.5'. Recrystallization gave an analytical by filtration, m.p. 145' (dec., in bath at 135"). This crop sample, m.p. 177.5-178.5", aZ4D -1" (c 2 in dioxane). was essentially pure 7-isomer. Two recrystallizations from The infrared spectrum (1 ,I ,2,2-tetrachloroethane) was iden- methanol-water raised the m.p. to 145-146" dec., #D 4-22' tical with that of the DL-or-isomer, which had been shown pre- (c 1 in acetic acid). viouslyl* to correspond to configuration to the natural di- And. Calcd. for C20H24N206S:C, 57.13; H, 5.75; N, methyl D-a-benzylpenicilloate; therefore this isomer has 6.66. Found: C, 57.25; H, 5.79; N, 6.62. been designated as the D-dsomer. Addition of 75 ml. of water to the aforementioned filtrate Anal. -Calcd. for CI~H~~N~OBS:C, 58.05; H, 6.03; N, caused the slow crystallization of 9.03 g. of a-isomer as 6.45. Found: C, 57.96; H, 6.09; N, 6.28. colorless needles, m.p. 152-153" (dec., in bath 135"). Ad- &Butyl D- and ~~-a4-Carboxy-5,5-dimethyl-a-amino-2- dition of a further 60 ml. of water gave 5.63 g., m.p. 148- thiazolidineacetate Hydrochloride (V).-A solution of 14.22 151' dec. The total yield of a-isomer was 14.66 g. (24%). g. (0.0338 mole) of ~~-111in 425 ml. of purified dioxane and Three recrystallizations from methanol-water afforded a 4 ml. of water wascooled to 13" and 3.80 ml. (3.84 g., 0.0768 product with a constant melting point, 159-160" dec., CY*~D mole) of hydrazine hydrate added over a 1-minute period $62' (c 1 in acetic acid). with st.irring. The solid which precipitated was redissolved Anal. Calcd. for C2oHe~N2OeS:C, 57.13; H, 5.75; iY, by warming to 18", and the solution was maintained at 13- 6.66. Found: C, 57.45; H, 6.06; N, 6.83. 15' for 3 hours, then at room temperature for 21 hours, The condensation under similar conditions of DL-peni- after which solvent and excess hydrazine were removed by cillamine hydrochloridez8and I and the isolation of the less- lyophilization. The phthalhydrazide complex was decom- soluble 7-isomer has been described previo~sly.~~From the posed by treatment of a suspension in 310 ml. of acetic acid last crop the more soluble (in acetone-water) a-isomer has at 13" with 8.15 ml. of concentrated hydrochloric acid, also now been obtained, m.p. 184-185' dec., mixed m.p. followed by agitation at room temperature for 30 minutes. with the ?-isomer was 180-181' dec. (both in bath at 160" The lyophilized suspension was digested with 175 ml. of with pure ?-isomer, which melted at 183-184" dec.). The cold methanol; 5.2 g. (95%) of phthalhydrazide was re- two diastereoisomers had infrared spectra (potassium bro- moved by filtration. Concentration of the filtrate to 80 ml. mide) which were distinctly different in the region between yielded 1.63 g. of hydrazine dihydrochloride, m.p. 199' dec. 10 and 13 Addition of 275 ml. of ether yielded 0.93 g. of impure solid N. of m.p. 90-125" dec. The further gradual addition of 1650 Anal. Calcd. for C~UHZ&O&:C, 57.13; E, 5.75; X, ml. of ether gave 9.07 g. (82O1,) of analytically pure D-cx-V, 6.66. Found: C, 57.21; H, 6.03; N,6.55. m.p. 172" dec., CY~~D+I11' (c 1 in methanol). Isomerization of the DL-?-Isomer to the ~~-a-Isomer.29- Anal. Calcd. for CI~H23N204SC1:C, 44.10; H, 7.09; A solution of 200 g. of DL-7-111in 350 ml. of reagent-grade C1, 10.85. Found: C, 43.83; H, 7.13; C1, 10.87. pyridine was heated on a steam-bath under an atmosphere of prepurified for 22 hours. After cooling in a Similar treatment of 28.44 g. of DL-01-111in a solution of refrigerator overnight, the crystalline nL-a-isomer (80 g.) 1360 ml. of dioxane and 10 ml. of water with 7.60 ml. of was collected and washed with two 25-mI. portions of cold hydrazine hydrate yielded 18.8 g. (85%) of analytically pure pyridine and two 75-1111. portions of cold ethyl ether. DL-CPV, m.p. 170' dec. Additional DL-?-isomer (80 g.) was added to the pyridine Anal. Calcd. for C12H2&0&C1: C, 44.10; H, 7.09; filtrate which was heated on a steam-bath for 24 hours and E,8.57. Found: C,43.80; H, 7.33; N, 8.72. cooled as before to give an additional 86 g. Another 80 g. &-Butyl D- and DL-a-Phenoxymethylpenicilloate (VI).- of DL-?-isomer was added to the filtrate to give after iso- To a solution of 20 g. (0.0612 mole) of DL-~-Vand 8.6 ml. merization a further 81 g. The combined product (247 g.) (6.19 g., 0.0612 mole) of triethylamine in 800 ml. of methyl- was recrystallized from acetone-water to yield 186 g. (52%), ene chloride at Oo, there was added, simultaneously, a solu- m.p. 186-187' dec.; comparison of infrared spectra (po- tion of 8.6 ml. (6.19 g., 0.0612 mole) of triethylamine in 600 tas4um bromide) indicated that this sample was substan- ml. of chloride and a solution of 8.65 ml. (10.47 tially pure DL-a-isomer. g., 0.0612 mole) of phenoxyacetyl chloride in 600 ml. of Anal. Calcd. for C20H24N20&: C, 57.13; H, 5.75; N, methylene chloride over a period of 1.5 hours in a system 6.66. Found: C, 57.21; H, 6.03; N, 6.55. protected from moisture. After 20 hours at room tempera- ture, a small amount of solid was removed by filtration, and Isomerization of the D-?-Isomer to the D-a-Isomer .-In the filtrate washed with two 1600-ml. portions of a solution a similar manner, heating a solution of 10 g. of DL-T-III in 18 ml. of reagent pyridine for 21 hours on a steam-bath containing equal volumes of 0.1 N hydrochloric acid and yielded 2.76 g. of colorless needles. Recrystallization from saturated , then with 800 ml. of saturated sodium chloride, dried over and concen- 40 ml. of methanol-water (1:l) afforded 1.69 g. of pure trated under reduced pressure to a foam. From an ethereal D-CY-111,m.p. 161' (dec., in bath 135'), mixed m.p. with solution there was obtained 20.54 g. (797,) of crystalline authentic ~e-111was 160.5' dec., aZ4~+62' (c 1 in acetic product, m.p. 142" dec. Recrystallization from ether- acid). petroleum ether gave an analytical sample, m.p. 143' dec. Anal. Calcd. for C20H24?;206S: C, 57.13; H, 5.75; hT, 6.66. Found: C, 56.87; H, 5.87; N, 6.70. Anal. Calcd. for C~OH~SN~O~S:C, 56.59; H, 6.65; N, 6.60. Found: C, 56.48; H, 6.86; 6.53. Concentration of the pyridine mother liquors and recy- IT, cling three times gave, after recrystallization, 2.10 g., m.p. In a similar manner, phenoxyacetyl chloride and triethyl- 159-160" dec. The total yield of D-cY-IIIwas 3.79 g. (38%). amine converted D-WVinto D-cY-VIin 7oyOyield, m.p. 120- 122" dec., aZ5D$67" (c 1 in methanol). &Butyl D- and DL-4-Carbomethoxy-5,S-dimethyl-a-phthal- imido-2-thiazolidineacetate (IV).-The preparations of Anal. Calcd. for C~UHZ~~ZOBS:C, 56.59; H, 6.65; N, DL-a-and DL-Y-IVhave been described previously by Shee- 6.60. Found: C, 56.88; H, 6.86; N, 6.59. han and Cr~ickshank.~~ D- and DL-a-Phenoxymethylpenidloic Acid Hydrochloride The first crop of D-I11 (3.48 g., 8.26 mmoles) was dis- (VII.HCl).-A solution of 19.8 g. (0.0466 mole) of DL-WVI solved in 40 ml. of dioxane (heating required) and the solu- in 800 ml. of methylene chloride, cooled to Oo, was saturated tion treated with excess diazomethane. The ester was cry;- with hydrogen chloride by passing the anhydrous gas tallized from ethanol-water yielding 1.54 g., m.p. 124-125 . through the solution for 20 minutes. Storage at 5' for 29 Two recrystallizations gave an analytical sample, m.p. 126- hours gave 18.1 g. (96%) of colorless crystals, m.p. 204- 127', aZ4D -19' (c 2 in dioxane). The infrared spectrum 205' dec. (1,1,2,2-tetrachloroethane) was identical with that of the ~~-?-isomer'3;therefore this ester has been designated as Anal. Calcd. for CI&lN?OBSC1: C, 47.46; H, 5.23; the D-7-isomer. N, 6.92. Found: C, 47.37; H, 5.21; N, 7.05. Anal. Calcd. for CzlH26N206S: C, 58.05; H,6.03; N, .halogous treatment of 3.21 g. (7.56 mmoles) of D-CY-VI 6.45. Found: C, 57.79; H,6.02; N, 6.66. in 52 ml. of methylene chloride with hydrogen chloride Similar treatment of the second crop of D-I11(2.72 g., 6.46 afforded 2.87 g. (947,) of D-(Y-VII.HC~,m.p. 113-116' dec. mmoles) from the condensation afforded, after crystalliza- D- and DL-a-Phenoxymethylpenicilloic Acid (VII).-To a tion from 35 ml. of 9570 ethanol, 2.22 g. (79%) of an ester,' solution of 147 mg. of D-cY-VII.HC~(0.36 mmole) in 0.6 ml. of acetone-water"(l:2) was added 35 mg. (0.44'mmole) of (29) This method of isomerization was developed in this Laboratory pyridine. The crystalline product was recrystallized from by V. J. Grenda. acetone-water to yield 113 mg. @I%), m.p. 125' dec. An June 20, 1959 TOTALSYNTHESIS OF PENICILLINV 3093 additional recrystallization from the same solvent combina- solve in either layer. This buffer "A" was prepared from tion gave analytically pure D-a-phenoxymethylpenicilloic 100 ml. of water, 100 ml. of 1.5 Mphosphate buffer (pH 6.0) acid hydrate, m.p. 129.5"dec., CPD+94" (c 1 in methanol). and 800 ml. of saturated sulfate. Funnels 1 and This penicilloic acid hydrate was shown to be identical with 2 were extracted successively with seven 1000-ml. portions the compound (vide infra) obtained from the alkaline hy- of buffer "B," which was composed of 567 ml. of water, drolysis of natural penicillin V by comparison of m.p., 100 ml. of 1.5 M phosphate buffer (pH 6.0) and 333 ml. of mixed m.p. (129' dec.), optical rotation and infrared spec- saturated ammonium sulfate. The 7 1. of buffer "B," trum (potassium bromide). containing the penicillin V, was cooled for 1 hour in an ice- Anal. Calcd. for CI&~ONZ~~S.HZO:c, 49.74; H, 5.74; bath, covered with 300 ml. of methyl isobutyl ketone, and N, 7.25. Found: C, 49.61; H, 5.77; N,6.94. 240 ml. of 20% phosphoric acid was added gradually with swirling. The aqueous layer was extracted further with In a similar manner 73 mg. of DL-~u-VII-HCIgave, after two 80-ml. portions of methyl isobutyl ketone. The com- one recrystallization from acetone-water, 23 mg. of DL-W bined orga$c layer was immediately shaken with 400 ml. of phenoxymethylpenicilloic acid, m.p. 133' dec. Recrystalli- buffer zation gave an analytical sample, m.p. 139" dec. "A. A second countercurrent distribution was carried out in AnaE. Calcd. for CI~H~ON~O&:c, 52.17; H, 5.47; N, the same manner with two funnels (400 ml. in each phase). 7.61. Found: C, 51.62; H, 5.28; N, 7.50. The three 400-ml. portions of buffer "A" were successively Alkaline Hydrolysis of Natural Penicillin V.-To a stirred equilibrated with funnels 1 and 2 and then were discarded. suspension of 75 g. (0.214 mole) of natural penicillin V30 in The penicillin V was then removed from the organic layer 1000 ml. of water, under an atmosphere of prepurified nitro- by successive extractions with seven 400-ml. portions of gen, was added 0.5 N dropwise until the buffer "B." The 2800-ml. aqueous layer was cooled in an pH was 11-11.5, the solution was maintained at this pH by ice-bath, was covered with 400 ml. of cold ether, and 100 the gradual addition of sodium hydroxide over a period of mi. of cold 20% phosphoric acid was added in portions with 2.5 hours (the total amount of sodium hydroxide required swirling. The aqueous layer (pH 2.5) was extracted with was 900 ml., 0.45 mole); the pH then remained constant an additional 200-ml. portion of ether. The ether layer at 11.5 for another hour without the further addition of was extracted with a 50-ml. portion of cold water, which sodium hydroxide. Hydrochloric acid (450 ml. of N, 0.45 was discarded. Cold water (150 ml.) was added, and the mole) was added in portions with stirring over 15 minutes, acids were titrated with 0.05 N potassium hydroxide (17.8 and the precipitated product was collected on a filter. The ml.) to PH 7.0 (pH meter). The aqueou? phase was lyo- wet solid was taken up as quickly as possible in 850 ml. of philized to yield 327 mg. of a white solid. Most of this boiling acetone, concentrated at room temperature to a solid (323 mg., which contained 242 mg. of potassium peni- smaller volume and water added. The resulting solid prod- cillin V by chemical assay) was taken up in a solution of 1 uct was recrystallized by solution in 2800 ml. of boiling ml. of water and 0.7 ml. of acetone. Dilution to 10 ml. wit! acetone, concentration at room temperature to 300 ml. acetone afforded 79 mg. of crystalline product, m.p. 261 followed by the addition of 300 ml. of water to yield 14.9 dec. Dilution to 40 ml. with acetone gave a second crop g., m.p. 127" dec., @D $93" (c 2 in methanol). Storage of 153 mg., m.p. 257" dec. Work-up of the mother liquors of the original aqueous filtrate overnight at 5" yielded a sec- afforded a further 13 mg. The infrared spectra (potassium ond crop which, after recrystallization, amounted to 13.6 g., bromide) of these fractions were identical with the spectrum m.p. 125', CU~~D+94' (c 2 in methanol). The total yield of natural potassium phenoxymethylpenicillinate. The was 28.5 g. (35%) of material suitablel6 for cyclization to total isolated yield of crystalline potassium D-phenoxy- penicillin V. Recrystallization gave an analytical sample methylpenicillinate was 245 mg. (5.4%). Recrystalliza- of D-a-phenoxymethylpenicilloic acid hydrate, m.p. 128" tion from water-acetone gave an analytical sample, m.D. dec., (~26~+93' (c 2 in methanol). 261' dec. (reported1* 256-260' uncor.), mixed m.p. with AnaE. Calcd. for CI~H~ONZO&H~~:c, 49.74; H, 5.74; natural VI11 was 261" dec., aZ7~$224" (c 0.4 in water); N, 7.25. Found: C,49.81; H, 5.98; N, 7.16. reported1* @D $223' (c 1 in water); ultraviolet spectrum Potassium n-Phenoxymethylpenicillinate (VIII). Relay A;:? 268 mb (e 1280) and 274.5 mp (e 1060) [natural peni- Syntheses.-To a suspension of 4.64 g. (0.012 mole) ?f cillin V potassium had 268 mp (E 1270) and 274.5 mp natural D-a-phenoxymethylpenicilloic acid hydrate's in (c 1040)]. 960 ml. of purified dioxane and 600 ml. of water waq added Anal. Calcd. for CL~HI~N~O~SK:c, 49.46; H, 4.41; with stirring 24 ml. of 0.5 N (0.012 mole) sodium hydroxide; N, 7.21; K, 10.06. Found: C, 49.69; H, 4.63; N, 7.52; after 30 minutes all the solid was in solution. To this K, 10.22. stirred solution was added in one portion a solution of 9.31 A small scale cyclization (5% of above quantities, ;.e., g. (0.048 mole) of N,N'-dicyclohexylcarbodiimide in 600 232 mg., 0.6 mole, of D-CY-VII),carried out exactly as above ml. of purified dioxane. After 22 hours at room temperature except for a longer period of time (33 hours), afforded 215 the pale yellow solution was lyophilized. A suspension of mg. of crude water-soluble penicillin V, which was 11% the residue in a mixture of 350 ml. of ether and 350 ml. of pure (11% yield) by chemical assay and 9% pure (9% water was stirred vigorously for 15 minutes. The layers yield) by bioassay. were separated and a trace of insoluble solid removed by filtration. The aqueous layer was extracted with 100 ml. To a suspension of 221 mg. (0.57 mmole) of D-CU-VI1in of ether and then lyophilized to yield 3.955 g. of crude water- 8 ml. of dioxane was added 5 ml. of 0.12 N sodium hydrox- soluble penicillin V which was 6.8% pure (6.0y0 yield) by ide. To this solution was added a solution of 0.5 ml. (7 bioassayle and 10.5% pure (9.3% yield) by chemical assay.'? mmoles) of ethoxyacetylene in 5 ml. of dioxane. .4fter The crude penicillin V was purified by a two-funnel dis- storage at room temperature for 19 hours, the solution was tribution between methyl isobutyl ketone and two succes- lyophilized to yield 259 mg. of a solid containing 0.22% sive phosphate buffers containing different concentrations 0.29% yield) of penicillin V by bioassay. of ammonium sulfate. A solution of the crude penicillin To a solution of 212 mg. (0.6 mmole) of the monolithium V in 100 ml. of water and 100 ml. of 1.5 M phosphate buffer salt of DW-VII in 5 ml. of water and 8 ml. of dioxane was (pH 6 .O, dibasic sodium phosphate-monobasic potassium added a solution of 150 mg. (0.78 mmole) of pentamethylene- phosphate = 1:2) was covered with 1000 ml. of distilled cyclohexylamine21 in 5 ml. of dioxane. After 18 methyl isobutyl ketone and 800 ml. of saturated ammonium hours the solution was lyophilized to give 393 mg. of a solid sulfate (70 g./100 ml. water) was added gradually with which contained 0.099% (0.19% yield) of penicillin V by swirling.s1 The aqueous phase was then transferred to bioassay. funnel 2, equilibrated with methyl isobutyl ketone and then To a suspension of 116 mg. (0.3 mmole) of D-wVII in 4 discarded. The methyl isobutyl ketone layers (each 1000 ml. of dioxane and 1.9 ml. of water was added 0.6 ml. of ml.) in funnels 1 and 2 were then sucffssively extracted with 0.5 N sodium hydroxide. To this solution was added a two 1000-ml. portions of buffer "A, which were then dis- suspension of 49 mg. (0.3 mmole) of N,N'-carbonyldiimid- carded as was the small amount of oil which failed to dis- azolez2in 2.5 ml. of dioxane. After storage for 1 hour at room temperature, the solution was lyophilized to yield 175 (30) Kindly furnished by Eli Lilly and Company, Indianapolis, mg. of an oily solid having no biological activity.16 A Indiana. second reaction was carried out as above, except that a 195- (31) A very similar solvent system was used for the separation of mg. (1.2 mmoles) portion of N,N'-carbonyldiimidazole sus- penicillin G from its penicilloic acid by S C. Pan, Anal. Chem., 26, 1438 pended in 10 ml. of dioxane was used, and gave a product (1954). which also had no biological activity. 3094 JOHN C. SHEEHANAND KENNETHK. HENERY-LOGAN Vol. 81

Total Synthesis (First Procedure).-To a solution of 2.54 from water-acetone afforded an analytical sample of DL- g. (6.3 mmoles) of D-~-VII.HCIin 86 ml. of dioxane and 28 VIII, m.p. 244' (dec., in bath 225'). The crystalline DL- nil. of water was added with stirring 25.1 ml. of 0.5 N (12.6 penicillinVpotassium saltshowed 51.4% of the bioactivity'ouf inmoles) sodium hydroxide. To this solution of the mono- natural penicillin V, indicating that L-penicillin V has little, iodium salt of D-WVIIwas added in one portion with stirring if any, antibiotic activity. The infrared spectrum (potas- solution of 1.21 g. (6.3 mmoles) of N,N'-dicyclohexylcarbo- sium bromide) had a strong band at 5.64 p, characteristic tliimide in 53 ml. of dioxane. After 25 minutes at room of the fused p-lactam-thiazolidine carbonyl function and temperature 0.21 g. (16%) of N,N'-dicylohexylurea was was in fact identical (in the position and intensity of 40 peaks removed by filtration and the filtrate was lyophilized. The and shoulders) with the infrared spectrum of the natural residue was taken up in 100 ml. cold methanol; 100 ml. of potassium salt of penicillin T'. water was added followed by 300 ml. of 1.5 phosphate Anal. Calcd. for CI6HliN~O5SK: C, 49.16; H, 4.41; buffer (pH 6.4); the water-insoluble material was removed N, 7.21. Found: C, 49.70; H, 4.61; N, 7.14. by filtration and discarded. The aqueous layer was covered with 300 ml. of ether and was cooled in an ice-bath; 142 nil. Zeoo-erythro-l,2-Diphenyl-Z-methylaminoethanolSalt of of 20% phosphoric acid was added in portions. After ex- Penicillin V (X).-To a stirred solution of 204 mg. (1 traction of the aqueous layer (pH 2.5) with an additional mmole) of leco-erythro-l,2-diphenyl-2-methylaminoetl1- 150-ml. portion of ether, the combined ethereal layers were ar~ol.HCl*~*~~(leoa-erythro-IX.HC1) in 20 ml. of water was washed with a 125-1111. portion of cold water, which was dis- added a solution of 350 mg. (0.9 mmole) of the natural carded. Cold water (100 ml.) was added, and the acids potassium salt of penicillin V in 2 ml. of water. The product were titrated with 9.9 ml. of 0.5 N sodium hydroxide to PH crystallized spontaneously and was collected after 20 minutes 6.8. The aqueous phase was lyophilized to yield 1.38 g. of to afford 464 mg. (8970) of X, m.p. 174-176" dec. Two re- crude water-soluble penicillin V which was 6.3% pure crystallizations from methanol-ether gave an analytical (3.8% yield) by bioassay'6 and 9.0% pure (5.3% yield) by sample, m.p. 181-183" dec., wZ5~+go" (c 0.4 in N-niethyl- chemical assay." 2-pyrrolidone). This compound had 95 =t 10% (573 p/ The crude penicillin V (1.34 9.) was purified by two coun- mg.) of the theoretical bioactivity (606 p/mg.) expected for tercurrent distributions between methyl isobutyl ketone and the salt.16 phosphate buffers (as described above for the relay synthe- 4nal. Calcd. for CBIH~~X~O~S:C, 64.46; H, 6.11; N, sis) to yield 98 mg. of a colorless solid which contained 7.28. Found: C,64.17; H,6.25; N, 7.25. 6570 of potassium phenoxymethylpenicillinate (2.8% yield) The Resolution of Potassium DL-Phenoxymethylpeni- by chemical assay. Crystallization from 98% acetone gave cillinate (VIII).-To a solution of 78 mg. (0.2 mmole) of 30 mg. (1.4% yield), m.p. 263' dec. (unchanged on recrys- DL-VI11in 2.5 ml. of water was added a solution of 43 mg. tallization), of the totally synthetic crystalline potassium (0.16 mmole) of leoo-e~ythro-IX~HC1z4~z~in 2 ml. of water. salt of penicillin V. The natural and synthetic potassium Seeding the solution with traces of X failed to induce crys- salts were shown to be identical by microbiological assay, tallization, but storage at 5" for 9 days afforded 46 mg. optical rotation [synthetic, aZ5D +223' (c 0.2 in water)! (70%) of crystalline product of m.p. 168" dec. Part of natural, @D +223" (c 0.2 in water); reported'8 ~ODf223 this solid (16 mg., 0.02 mmole) was purified further by shak- (c 1 in water)], infrared spectrum (identical in 40 peaks and ing with 1.5 ml. of n-butyl acetate and 1.5 ml. of 0.7% shoulders in potassium bromide), ultraviolet spectrum [syn- phosphoric acid until two clear layers resulted. The aque- thetic, A::? 268 mp (E 1250) and 274.5 mp (e 1030); natural: ous layer was discarded, and the organic layer was washed A""t"mRI 268 mp (E 1270) and 274.5 mp (e 1040)], m.p. 263 with two I-ml. portions of water. To this solution was dec. (reported's 256-260' uncor.) and undepressed mixed added a solution of 4.6 mg. (0.02 mmole) of levo-eryfhro-IX m.p. in 0.4 ml. of n-butyl acetate. Storage at 5" overnight Potassium DL-Phenoxymethylpenicillhate (VIII) .-To a afforded 10 mg. (44%) of a crystalline salt, m.p. solution of 18.1 g. 10.045 mole) of DL-~-VII.HCIin 1800 ml. 175' dec., aZ5D +92" (c 0.2 in N-methyl-2-pyrrolidone). of dioxane and 1120 ml. of water was added with stirring Comparison of infrared spectra (potassium bromide) 179 ml. of 0.5 N (0.09 mole) sodium hydroxide. To this showed this compound to be identical with X. This com- solution was added with stirring a solution of 34.7 g. (0.18 pound showed 95 f lOy0 (576 p/mg.) of the theoretical mole of N,N'-dicyclohexylcarbodiimide in 1120 ml. of di- bioactivity expected for the salt.I6 oxane in one portion, -4fter 26 hours at room temperature The aqueous filtrate containing L-penicillin V was cov- the yellow solution was lyophilized. The residue was made ered with 6 ml. of n-butyl acetate, and 3.5 ml. of lyOphos- completely water-soluble by the procedure described for the phoric acid was added; shaking caused the solid which pre- total synthesis of D-VIII. The crude water-soluble so- cipitated to redissolve. The organic layer was washed dium salt of DL-penicillin V (5.38 9.) was 9.1% pure (bio- with two 5-ml. portions of water. To the organic layer assay16 showed the sample contained 4.2870 natural peni- was added a solution of 46 mg. (0.2 mmole) of dextro- cillin V acid, and assuming the L-form has no biological erq'thro-IX24v25in 3 ml. of n-butyl acetate. Scratching activity, the sample contains 9.1% DL-penicillin V as the initiated crystallization and storage for 3 days at 5' af- sodium salt) (3.0Y0 yield) and 9.3% pure (3.0% yield) by forded 35 mg. of crystals, m.p. 177" dec. A second crop chemical assay.'? was obtained from the filtrate, m.p. 175' dec. The total The crude DL-penicillin V (5.35 8.) was purified by two yield was 41 mg. (66%). Recrystallization from dioxane- countercurrent distributions (as described above for the ether gave an analytical sample of the dextuo-erythro-IX salt relay synthesis) to yield 595 'mg. of a white solid which of L-VIII, m.p. 175' dec., a25~-86" (c 0.4 in N-methyl-2- contained 53% (1.970 yield) of DL-potassium phenoxy- pyrrolidone). The infrared spectrum (potassium bromide) methylpenicillinate by chemical assay. The solid (585 mg.1 was identical to that of X. This salt of L-penicillin l' was taken up in a minimum volume of water and was diluted showed, 0.7% (4p/mg.) of the theoretical bioactivity (606 p/ with acetone to afford 128 mg. of crystalline product, m.p. mg.) expected for the salt of natural penicillin V.'6 239' (dec., in bath 225'); further dilution with acetone Anal. Calcd. for CalH85N30eS:C, 64.46; I-I, 6.11. gave a second crop of 165 mg., m.p. 235" dec. The total Found: C, 64.09; H, 6.41. yield of crystalline synthetic potassium DL-phenoxymethyl- penicillinate was 293 mg. (1.8%). Two recrystallizations CAMBRIDGE39, MASS. J. Am. Chem. SOC.1987, 109, 2205-2208 2205

Scheme 111. Summary of the Reductive Activation-Alkylation Reactions of N-Methylmitomycin A 10, 11

0'

+ le 1. DNA DNA - apomilosene 1- - crosslink milomydn 2. -18

OW) 5 (milomydn semiquinone) 6 (milosene semiquinone)

7 (arindtnomtlosene) complicated by uncertainties in the sequence of the various acy- 7).3bThe rough vinylogy between the two processes is indicated lations. (cf. arrows). Our data do not preclude significant alkylation Hornemann and Kohn had previously examined the reaction properties for compound 3a. They also do not define the precise of MMC with potassium ethyl xanthate (9) under reductive species involved in the remarkable transformation of 3a -+ 4.16 (Na2S204) ~0nditions.l~ Apomitosenes, arising from the They do, however, provide a basis for proposing a very concise nonstereospecific ring opening of the aziridine by nucleophile at sequence for bioactivation of mitomycins, as shown in Scheme C1,were encountered. It was implicitly presumed that reaction 111. A natural consequence of these findings is that new de- had occurred via the MMC derived version of 3a. In the N- partures in mitomycin drug development might well center on MeMMA series, we could evaluate the reactivity of the aziri- substitutions which will favor species generically related to 6. This dine-containing compounds 7 and 3a, generated as discrete entities proposition will now be pursued. by our method~logy.~~.~Reactions were conducted in aqueous pyridine. From this series of reactions, three products could be Acknowledgment. This work was supported by PHS Grant CA obtained and identified. These were the C, and Cloxanthates 28824. NMR spectra were obtained through the auspices of the 10 and 12, as well as the dixanthate 11. When reaction was Northeast Regional NSF/NMR Facility at Yale University, conducted with 3a (10 min, 0 "C) followed by subsequent air which was supported by NSF Chemistry Division Grant CHE oxidation, a 20% yield of the three products in the indicated ratio 7916210. We also thank Dr. Terry Doyle of the Bristol Phar- was obtained. When reaction was conducted on 7, a trace of 10 maceutical Co. for supplying us with mitomycin C. (ca. 5%) could be detected and ca. 95% of 7 was recovered. Supplementary Material Available: Experimental data for Maximum yield was realized from the reaction of 7 with compounds 2a,b, 3a,b, 4, 7, 8a,c and 10-12 (3 pages). Ordering Na2S204in the presence of 9. Oxidation (air) after the 10-min information is given on any current masthead page. incubation period afforded a 35% yield of 10 and a 25% yield of 11. Thus the process of reductive priming of 7 with sodium (16) The similarities of "electron flow" inherent in the formation of dithionite gave a substantially higher yield than was realized from xanthate adducts 10, 11, 12, and of ene pyrrole 4 make tempting the possibility the two-electron reduction product (3a) itself. Furthermore, that semiquinone 6 is intervening in the formation of 4. attempted reduction of 7 with dithionite (aqueous pyridine) in the absence of nucleophile 9 led to very slow reaction and the product was not 3a, but rather the ene pyrrole 4 (NMR analysis). The rate of formation of 3a is too slow for it to be the primary alkylating agent. Hence the formulation whereby the two electron reduction product, 3a, acts as the active alkylating agent, producing Stereocontrolled Construction of Key Building Blocks 10, 11, and 12, is untenable. The sequence embodied in Scheme 111, wherein mitosene semiquinone 6 alkylates nucleophile 9, for the Total Synthesis of Amphoteronolide B and accounts very well for the observed result. Further support for Amphotericin B the proposal comes from the reaction of aziridinomitosene (7) with a catalytic amount (0.3 equiv) of Na2S204in the presence of K. C. Nicolaou,* R. A. Daines, J. Uenishi, W. S. Li, nucleophile 9. Workup after 35 min yielded an 80% combined D. P. Papahatjis, and T. K. Chakraborty yield of xanthate alkylated products.15 Thus the extent ofal- kylation substantially exceeds the availability of reducing agent. Department of Chemistry, University of Pennsylvania These data in the aggregate point toward the intervention of a Philadelphia, Pennsylvania 19104 steady-state reactive intermediate (cf. semiquinone 6) as the active Received October 6. 1986 electrophile. A parallelism is noted between the intervention of semiquinone Amphotericin B1 (I, Scheme I, with @-linked mycosamine at equivalent 6 in the xanthate alkylation reactions and the in- the C- 19 hydroxyl), a clinically used antifungal agent isolated from volvement of species 5 in the C9, methoxy-ejection event (1 - Streptomyces nodosus, and its aglycon, amphoteronolide B2 (I, Scheme I), are important synthetic targets of considerable current (13) (a) Tomasz, M.; Lipman, R. Biochemistry 1986, 25, 4337. (b) intere~t.~In this paper we describe stereocontrolled constructions Compounds 8a and 8b are briefly alluded to (stereochemistry not defined) by the following patent: Matsui, M.; Yamada, Y.;Uzu, K.; Hirata, T.; Wakaki, S. US.Patent 3429894, Feb. 25, 1969. (14) Hornemann, U.; Iguchi, K.; Keller, P. J.; Huynh, M. V.; Kozlowski, (1) Isolation: Vandeputte, J.; Watchtel, J. L.; Stiller, E. T. Antibior. Annu. J. F.; Kohn, H. J. Org. Chem. 1983, 48, 5026. 1956, 587. X-ray structural determination: Mechinski, W.; Shaffner, C. P.; (15) Compounds 10,11, and 12 were obtained in yields of 8%, 23%, and Ganis, P.; Avitabile, G. Tetrahedron Leu. 1970, 3873. Ganis, P.; Avitabile, 33% respectively. Twelve percent of the starting material (7) was recovered. G.; Mechinski, W.; Shaffner, C. P. J. Am. Chem. Sor. 1971, 93, 4560. Reaction of 7 and 9 for a similar time period resulted in a 15% yield of 10 (2) Preparation from amphotericin B: Nicolaou, K. C.; Chakraborty, T. and an 85% recovery of starting material. K.; Daines, R. A.; Ogawa, Y. J. Chem. SOC.,Chem. Commun., in press. 0002-7863/87/1509-2205$01.50/0 0 1987 American Chemical Society 2206 J. Am. Chem. SOC.,Vol. 109, No. 7, 1987 Communications to the Editor

Scheme 1" Scheme 11"

HO-OWPh IX

mJ ~ \= -0-Ph Ph-0-

OH OH OH HO HO HO 1 2 Vlll Ila Ilb 4 or or

H0i."rOH I

OH HO HouonIlla: enantiomers1 Illb: k-)-xyiosel HO OH

D 001

VI1 6 6 'Reagents and conditions. (a) See: ref 6, ca. 35% overall. (b) (i) 1.0 equiv of LiBH4, THF, 0 OC, 0.5 h, then 1.1 equiv of r-BuCOCI, pyr, 3 h, 90% overall, (ii) 1.2 equiv of Me,-t-BuSiOTf, 2.0 equiv of 2,6-lutidine, CH,CI,, 0 OC, 1 h, 97%, then AcOH-THF-H20 (3:1:1), 50 OC, 2 h, 72%, (iii) 1.2 equiv of PhSSPh, 1.2 equiv of n-Bu3P, THF, 0-25 OC, 3 h then Raney Ni, EtOH, 12 h, 90% overall. (c) (i) 1.2 equiv of dihydropyran, CSA catalyst, CH2C12,0-25 OC, 3 h, 96%, (ii) 2.5 equiv of DIBAL, CH2C12, -78 "C, 0.5 h, then 1.5 equiv of CrO,. HChpyr, 1.5 equiv of NaOAc, CH2C12, 25 OC, 4 h, 75% overall. (d) (i) 2.5 equiv of EtzA1C~CCH20Si-t-BuPh2,hexane-toluene (1 :I), -78 - 0 OC, 0.5 h, 85%, (ii) 1.2 equiv of NaH, 1.2 equiv of PhCH,Br, THF, 0-25 OC, 14 h, 91%, (iii) 1.5 equiv of n-Bu4NF, THF, 0-25 OC, 95%. (e) 3.5 equiv of REDAL, Et20,0-25 OC, 3 h, 97%. (0 1.5 equiv of (-)-DET, 2.2 equiv of t-BuOOH, 1.2 equiv of Ti(iPrO),, CH2CI2, -20 "C, 16 h, 82%. (g) (i) 4.0 equiv of REDAL, THF, 0 OC, 4 h, 97%, (ii) 2.6 equiv of t-BuMe2SiC1, 3.0 equiv of imidazole, DMF, 0-25 Vlll IV VI1 OC, 12 h, 89% (iii) H,, 10% Pd(OH),-C, EtOH, 25 OC, 1 h, 95%. (h) (i) 1.2 equiv of PhCH(OMe)2, CSA catalyst, benzene, 25 OC, I h, ci 80%, (ii) 6.0 equiv of S03.pyr, 10.0 equiv of Et3N, 20 equiv of Me2S0, 1 (+)-DET I CHZCI,, 25 OC, 2 h, 95%. and key building blocks of amphoteronolide chiral auxiliaries to secure optically active materials. Thus, B (1). following the indicated disconnections (dotted lines) in Scheme of all requisite building blocks (V-VIII, Scheme I) for these I the initially generated key intermediates V-VI11 were further polyene macrolide targets. The reported strategies were based traced back to IV (VI1 and VIII), (+)-xylose (IIIa) and on the recognition of important subtle symmetry elements in these (-)-xylose (IIIb) (V and VI, respectively), or compounds IIa and complex that simplified considerably the overall syn- IIb (V and VI, respectively). Enantiomerically pure epoxide IV thetic plan and made it possible to adopt the same or enantiom- is readily available from (+)-DET (diethyl tartrate), whereas (-)- erically related materials as starting points for all four syntheses. and (+)-DET can be used as chiral inducers to build the requisite Scheme I presents the retrosynthetic analysis and strategic bond absolute stereochemistry in intermediates IIa and IIb, respectively, disconnections defining the key compounds V-VI11 as the four from the prochiral starting material IX via a Sharpless asymmetric requisite building blocks for an eventual total synthesis of am- ep~xidation.~The numbering in Scheme I traces the origin of photeronolide B and amphotericin B. This analysis uncovered selected carbon centers of amphoteronolide B, clarifying the choice certain stereochemical and symmetry elements allowing the design of starting materials. of a synthetic strategy that utilizes the readily available enan- Scheme I1 outlines the construction of building blocks VI11 and tiomers of xylose and tartaric acid as starting materials and/or VI1 from epoxide IV.s Thus, IV was expeditiously converted to intermediate 1 by the Evans aldol methodology6 (35% overall (3) Synthetic studies: (a) McGarvey, G. J.; Williams, J. M.; Hiner, R. N.; Matsubara, Y.; Taeboem, 0. J. Am. Chem. SOC.1986, 108, 4943. (b) ~~ Boscheli, D.; Takemasa, T.; Nishitani, Y.;Masamune, S. Tetrahedron Lett. (4) For excellent recent reviews of this reaction see: Rossiter, B. E. In 1985, 5239. (c) Blanchette, M. A,; Choy, W.; Davis, J. T.; Essenfield, A. P.; Asymmetric Synthesis; Morrison, J. D., Ed.; Academic: New York, 1985; Masamune, S.; Roush, W. R.; Sakai, T. Terrahedron Lert. 1984, 2183. (d) Vol. 5, p 193. Finn, M. G.; Sharpless, K. B. Ibid.; p 247. Masamune, S.; Ma, P.; Okumoto, H.; Ellingboe, J. W. J. Org. Chem. 1984, (5) Nicolaou, K. C.; Papahatjis, D. P.; Claremon, D. A,; Magolda, R. L.; 49, 2843. (e) Ma, P.; Martin, V. S.;Masamune, S.; Sharpless, K. B.; Viti, Dolle, R. E. J. Org. Chem. 1985, 50, 1440. S. M. J. Org. Chem. 1982,47, 1378. (0 Masamune, S.; Kaiho, T.; Garvey, (6) This conversion, which features a stereoselective Evans aldol conden- D. S. J. Am. Chem. SOC.1982, 104. 5521. (g) Liang, D.; Pauls, H. W.; sation (for an excellent review of this reaction, see: Evans, D. A,; Nelson, J. Fraser-Reid, B. J. Chem. SOC.,Chem. Cornmun. 1984, 1123. (h) Lipshutz, V.; Taber, T. R. In Topics in Stereochemistry; Allinger, N. L, Eliel, E. L., B. H.; Koslowski, J. A. J. Org. Chem. 1984, 49, 1147. (i) Hirama, M.; Vie, Wilen, S. H., Eds.; 1982; Vol. 13, p 1) was first reported at the 1st Cyprus M. Tetrahedron Lett. 1982, 5307. 6) Brooks, D. W.; Kellogg, R. P. Tetra- Conference on New Methods in Drug Research, Cyprus, 1983; see: Nicolaou, hedron Left. 1982, 4991. (k) Floyd, D. M.; Fritz, A. W. Tetrahedron Lett. K. C.; Petasis, N. A,; Dolle, R. E.; Li, W. S.; Papahatjis, D. P.; Uenishi, J.; 1981, 2847. Hanessian, S.; Sahoo, S. P.; Murray, P. J. Tetrahedron Lett. Zipkin, R. E. In New Methods in Drug Research; Makriyannis, A,, Ed.; Prous 1985, 26, 5631. Science: Barcelona, Spain, 1985; p 179. Communications to the Editor J. Am. Chem. SOC.,Vol. 109, No. 7, 1987 2207

Scheme 111" Scheme IV"

mO-Ph Ph-0- OH OH OH HO HO HO ("7Ilb IIa 'KY 0, Ph i 9 V HO-?Y

b] (60%)

ph2'-~o-.qp~L(35%) (35%) o~m~E"phz 0 11 'Reagents and conditions. (a) See: ref 13. (b) (i) 1.1 equiv of 0-T- OH : HO 7-0 t-BuCOCI, pyr, 0 OC, 6 h, 88%, (ii) Me2C(OMe),, CSA catalyst, 25 7a 111a: (+)-xylose ~ Illb (-)-xylose 7b OC, 1 h, 93%, (iii) H,, 10% Pd-C, EtOH, 12 h, 94%, (iv) 1.1 equiv of e (58%) t-BuPh, SiC1, 4.0 equiv of imidazole, DMF, 0-25 "C, 2 h, 88%, (v) 2.5 equiv of DIBAL, CHZClz,-78 OC, 0.5 h, 9156, (vi) 1.1 equiv of NaH, 00 ph Fll 1 1.1 equiv of PhCH2Br, THF, 0-25 OC, 3 h, 85%. (c) As in Scheme ~Meol'p~mrEuM-~o~o~~E"M*~~0-Y-Y-HO OH 111. (d) (i) 1.1 equiv of t-BuMe,SiCI, 4.0 equiv of imidazole, DMF, OK0 K 0-25 "C, 2 h, 90%, (ii) as in b part ii, above, 95%. (e) As in Scheme VI 11 10 111. Reagents and conditions. (a) (i) Me,CO, concentrated H,S04 catalyst, 25 OC, 6 h, then dilute H$04 catalyst MeOH, 25 OC, 2 h, to afford derivative 7a (35% overall yield). Removal of the 50% overall, (ii) 1.1 equiv of t-BuPhzSiC1, 4.0 equiv of imidazole, acetonide group from 7a required the use of BCI3l1 and led, after DMF, 0-25 OC, 2 h, 96%, (iii) 1.5 equiv of PhOC(S)Cl, 2.5 equiv of Wittig methylenation, to 8 (60% overall). Acetonide for- pyr, DMAP catalyst, CH,CI,, 0-25 OC, 15 h, 94% then 1.1 equiv of mation followed by hydroboration and benzylation of the resulting n-Bu,SnH, AIBN catalyst, toluene, 80 OC, 1 h, 77%. (b) (i) 1.0 equiv primary gave compound 9 (63.5% overall), which was then of BCI,, CH,Cl,-hexane (1:2), -78 OC, 10 min, 9070, (ii) 1.0 equiv of desilylated and oxidized to afford the targeted V (72% NaH then 3.2 equiv of Ph3P=CH2 (from CH,PPh,+Br- and n-BuLi in overall yield). The route to building block VI started from THF), THF, -30 - 25 OC, 4 h, 67%. (c) (i) Me,C(OMe),, CSA (-)-xylose (IIIb) and proceeded through derivative 7b, obtained catalyst, 25 "C, 1 h, 90%, (ii) 2.1 equiv of Sia,BH, THF, 0 OC, 1.5 h as described above for its enantiomer (7a). The silyl protecting then NaOH-H202 workup, 88%. (iii) 1.2 equiv of KH, 1.2 equiv of group in was then exchanged with a benzyl group, the acetonide PhCH,Br, THF, 0-25 'C, 14 h, 85%. (d) (i) 1.2 equiv of n-Bu,NF, 7b THF, 0-25 OC, 4 h, 96%, (ii) 6.0 equiv of S03.pyr, 10.0 equiv of Et3N, was removed (aqueous HCl), and the resulting lactol was converted 20.0 equiv of MezSO, CH,CI,, 25 OC, 3 h, 75%. (e) (i) same as (d) (i) to olefin 10 by a Wittig reaction (58% overall). Engagement of above, (ii) 1.1 equiv of NaH, 1.1 equiv of PhCHzBr, THF, 0-25 OG, 2 the 1,3-diol system of 10 as an acetonide, followed by hydro- h, 95%, (iii) dil HC1, DME-HzO (2:1), reflux, 1 h, 95%, then same as boration and silylation, led to compound 11 in 73% overall yield. b part ii, above. (f) same as c parts i and ii, above, then 1.1 equiv of Finally, debenzylation, oxidation of the liberated hydroxyl group t-BuMe,SiCI, 4.0 equiv of imidazole, DMF, 0-25 OC, I h, 92%. (8) (i) to the (Ru04 catalyst-NaIO,),'* methyl ester H,, 10% Pd-C, EtOH, 25 OC, 9856, (ii) 5.0 equiv of NaIO,, Ru0, formation, and nucleophilic attack by LiCH,P(0)(OMe)2 fur- catalyst, CH,CN-CC14-Hz0 (2:2:3), 25 OC, 6 h, then CHzN2,76% nished the desired keto phosphonate VI in 71.5% overall yield. overall, (iii) 2.2 equiv of (MeO), P(O)CH,Li, THF, -78 to 0 "C, 1 h, 9670. Alternative syntheses of fragments V and VI starting with the prochiral allylic alcohol IX are summarized in Scheme IV. As yield), which was reduced (LiBH,) to the corresponding primary mentioned above, (-)- and (+)-DET were utilized in conjunction alcohol, protected as a pivalate ester, silylated at the free secondary with the Sharpless asymmetric epoxidation reaction4 to induce hydroxyl, deacetonated, and deoxygenated at the liberated primary the desired asymmetry. Thus, according to our previously reported position to afford compound 2 (56.5% overall yield). Protection general method for building 1, 3, 5, ..., (2n + 1) polyols,13 IX was of the hydroxyl group in 2 as a THP ether followed by DIBAL converted to the enantiomeric triols IIa and IIb. Protecting group cleavage of the pivalate ester and PCC oxidation led to the desired manipulation of IIa as detailed in Scheme IV then led to inter- key intermediate VI11 (72% overall) in optically active form. The mediate 9 (52% overall yield), which was converted to V as already synthesis of VI1 began with attack on epoxide IV by [[(tert-bu- described above. In parallel, IIb was transformed to the protected tyldiphenylsilyl)oxy]propargyl]diethylalane (Et2A1C= derivative 11 by selective silylation and acetonide formation (85.5% CCH20Si-t-BuPh2),' affording, after benzylation and desilylation, overall yield) and thence to VI as described above. hydroxy 3 in 77.5% overall yield. REDAL reduction In conclusion, focusing on subtle and repeated structural units, of 3 gave allylic alcohol 4 (97%), which underwent smooth the described retrosynthetic analysis allows the utilization of readily Sharpless asymmetric epoxidation [(-)-DET] furnishing hydroxy available enantiomeric structures as starting points for an eventual epoxide 5 (82%). Regioselective epoxide opening of 5 (REDAL)* total synthesis of both amphoteronolide B and amphotericin B. followed by silylation and debenzylation then gave 6 (86.5% Thus, four major building blocks (V-VIII) have been synthesized overall), which was selectively protected as a six-membered in optically active forms and by highly efficient and concise se- benzylidene9 and oxidized with S03.pyr complex to afford frag- quences using (+)- and (-)-xylose and (+)- and (-)-DET as ment VI1 (76% overall yield), optically pure and suitably protected sources of chirality. The stage is now set for a highly convergent for further elaboration. total synthesis of both amphoteronolide B and amphotericin B. Scheme I11 presents the construction of the remaining building The following paper describes the accomplishment of the former blocks V and VI in enantiomerically pure form starting with the goa1.14~5 two enantiomers of xylose. Thus, (+)-xylose (IIIa) was converted Acknowledgment. We express our many thanks to Drs. George to its 1,2-monoacetonide, selectively silylated at the primary Furst and John Dykins in this department for their superb NMR position, and deoxygenated by the Robins and Wilson methodlo and mass spectroscopic assistance. This work was financially supported by the National Institutes of Health, Merck Sharp & (7) This reagent was prepared from the corresponding acetylene (hexane, -78 "C) by sequential addition of 1.0 equiv of n-BuLi (1.55 M in hexane) and (11) Tewson, T. J.; Welch, M. J. J. Org. Chem. 1978, 43, 1090. 1.0 equiv of Et,AICI (1.8 M in toluene), see: Suzuki, T.; Saimoto, H. I.; (12) Carlsen, H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org. Tomioka, H.; Oshima, K.; Nozaki, H. Tetrahedron Lett. 1982, 23, 3597. Chem. 1981, 46, 3938. (8) Minami, N.; KO, S. S.; Kishi, Y.J. Am. Chem. SOC.1982, 104, 1169. (13) Nicolaou, K. C.; Uenishi, J. J. Chem. Soc., Chem. Commun. 1982, Kishi, Y.; Finan, J. M. Tetrahedron Lett. 1982, 23, 2719. 1292. (9) For previous selective six-membered ring benzylidene formations, see: (14) Nicolaou, K. C.; Daines, R. A,; Chakraborty, T. K., following paper (a) Sinclair, H. B. Carbohydr. Res. 1969, 12, 150. (b) Ziegler, F. E.; Gilligan, in this issue. P. J. Tetrahedron Lett. 1979, 3371. (c) Schubert, T.; Welzel, P. Angew. (15) All new compounds exhibited satisfactory spectral and analytical/ Chem., In?. Ed. Engl. 1982, 21, 137. exact mass spectral data. Yields refer to spectroscopically and chromato- (10) Robins, M. J.; Wilson, J. S. J. Am. Chem. SOC.1981, 103, 932. graphically homogenous materials. 2208 J. Am. Chem. SOC.1987, 109, 2208-2210

Dohme, and Hoffmann-La Roche. Scheme I' Registry No. 1, 106799-08-0; 2, 101417-56-5; 3, 106799-09-1; 4, OH 106820-43-3; 5, 106799-10-4; 6, 106799-11-5; 7a, 106799-13-7; 7b, 106799-16-0; 8, 106862-35-5; 9, 106799-14-8; 10, 106799-17-1; 11, 106799-18-2; IIa, 81 120-67-4; IIb, 106862-36-6; IIIa, 58-86-6; IIIb, 609-06-3; IV, 81 177-24-4; V, 106799-15-9; VI, 106799-19-3; VII, 106799-12-6; VIII, 105172-28-9; IX, 69152-88-1; Et2AIC=CCH20Si- t-BuPh2, 106799-20-6; CH3PPh3+Br-, 1779-49-3; (Me0)2P(0)CH2Li, 34939-91-8; amphoteronolide B, 106799-07-9; amphotericin B, 1397- 89-3. OR Supplementary Material Available: List of Rf, [aID,IR, and 'H NMR data for compounds V-VI11 (2 pages). Ordering in- formation is given on any current masthead page.

Total Synthesis of Amphoteronolide B K. C. Nicolaou,* R. A. Daines, and T. K. Chakraborty Department of Chemistry, University of Pennsylvania Philadelphia, Pennsylvania 19104 Received October 6, 1986

Amphoteronolide B (I, Scheme I), the aglycon of amphotericin B, has recently been obtained from naturally derived amphotericin B and fully characterized by spectroscopic means.' We now report the first total synthesis of this important and long sought target in its optically active form from readily available starting materials and in a highly stereocontrolled manner.2 Ph Scheme I outlines a retrasynthetic analysis of the titled . OR Ox0 Ox0 Thus, it was envisioned that amphoteronolide B (I) could be I derived from the protected heptaenone I1 by stereoselective Me,,,, OTHP and deprotection. This maneuver then allowed disconnection of this precursor at the lactone and unsaturated sites as indicated in structure 11. The chosen strategic bond discon- nections leading to advanced intermediates I11 and IV pointed to a highly convergent synthesis and also to two powerful coupling Vlll reactions, an esterification and a keto phosphonate-aldehyde 0 VI1 condensation, in the synthetic plan to construct 11. Finally, (EtO),PW COOEt subtargets keto phosphonate carboxylic acid I11 and hydroxy IX aldehyde IV were retrosynthetically disassembled as indicated in R = Si'BuMep Scheme I, revealing building blocks V-IX as potential starting Retrosynthetic analysis of amphoteronolide B (I). points for the total synthesis. The construction of building blocks V-VI11 is reported in the preceeding paper.j Their coupling and elaboration to ampho- confirmed by X-ray crystallographic analysis (see the ORTEP teronolide B is detailed in Scheme 11. Thus, coupling of aldehyde drawing in Scheme 11) of the crystalline p-chlorobenzenesulfonate V and keto phosphonate VI under basic conditions led to the 3 prepared from 2 as outlined in Scheme 11. Compound 2 was expected conjugated enone in 94% yield, which was cleanly hy- then functionalized appropriately so as to allow its coupling to drogenated to the saturated ketone 1 (98%). Molecular models the third building block VI1 as follows. Protection of the secondary of this ketone suggested that reduction should occur from the hydroxyl of 2 with the more stable t-BuPh2Si group5 (91%) opposite side of the adjacent acetonide, particularly by a sterically followed by selective removal of the t-BuMe2Si group (84%) from demanding reagent attacking a frozen conformation of 1. Indeed, the primary hydroxyl led to compound 5 via 4. Intermediate 5 L-selectride at -120 OC produced the single diastereoisomer 2 in was then sequentially converted to iodide 6 (97%) via its mesylate 98% yield.4 The stereochemical outcome of this reduction was and then to dimethyl phosphonate 7 by displacement with sodium dimethylphosphite.6 Sulfenation of the anion of 7 then led to a diastereomeric mixture of the a-methylthio phosphonate 8 (73%; (1) Nicolaou, K. C.; Chakraborty, T. K.; Daines, R. A.; Ogawa, Y.J. ca. 1:l by 'H NMR). Condensation of the anion of 8 with Chem. SOC.,Chem. Commun., in press. aldehyde VI1 proceeded smoothly, leading to coupling product (2) For synthetic studies in this area by other groups, see: (a) McGarvey, G. J.; Williams, J. M.; Hiner, R. N.; Matsubara, Y.;Taeboem, 0. J. Am. 9 (84%; mixture of geometrical isomers, ca. 1:l by 'H NMR). Chem. Sot. 1986, 108, 4943. (b) Boscheli, D.; Takemasa, T.; Nishitani, Y.; Desilylation of 9 to the triol 10 (96%) followed by an acid-induced Masamune, S. Tetrahedron Lett. 1985, 5239. (c) Blanchette, M. A,; Choy, cyclization led to mixed cyclic ketal 11 (64%; ca. 1:l mixture of W.; Davis, J. T.; Essenfield, A. P.; Masamune, S.; Roush, W. R.; Sakai, T. anomers by 'H NMR), which was converted to the methoxy Tetrahedron Lett. 1984, 2183. (d) Masamune, S.; Ma, P.; Okumoto, H.; Ellingtoe, J. W. J. Org. Chem. 1984, 49, 2843. (e) Ma, P.; Martin, V. S.; compound 12 by exposure to NBS-MeOH (95%; ca. 1:l mixture Masamune, S.;Sharpless, K. B.; Viti, S. M. J. Org. Chem. 1982, 47, 1378. (f) Masamune, S.; Kaiho, T.; Garvey, D. S. J. Am. Chem. SOC.1982, 104, 5521. (g) Liang, D.; Pads, H. W.; Fraser-Reid, B. J. Chem. Sot., Chem. (4) At least 98% pure, as checked by 'H NMR (250 MHz). Commun. 1984, 1123. (h) Lipshutz, B. H.; Koslowski, J. A. J. Org. Chem. A variety of other reduction conditions gave mixtures of 2 and its epimer (e&, 1984, 49, 1147. (i) Hirama, M.; Vie, M. Tetrahedron Lett. 1982, 5307. u) L-selectride, THF, -78 OC, ca. 5:l ratio; L-selectride, Et,O, -78 "C, ca. 1.3:1 Brooks, D. W.; Kellogg, R. P. Tetrahedron Lett. 1982, 4991. (k) Floyd, D. ratio; Zn(BH,),, Et20,0 or -78 OC, ca. 1:1 ratio; DIBAL, CH2CIz,-78 "C, M.; Fritz, A. W. Tetrahedron Lett. 1981, 2847. ca. 2.7:l ratio; t-BuNHyBH,, THF, -40 "C,ca. 1.2:l ratio). (3) Nicolaou, K. C.; Daines, R. A.; Uenishi, J.; Li, W.-S.;Papahatjis, D. (5) Hanessian, S.; Lavallee, P. Can. J. Chem. 1975, 53, 2975. P.; Chakraborty, T. K., preceeding paper in this issue. (6) Sturtz, G. Bull. Chim. Fr. 1964, 2340.

0002-7863/87/ 1509-2208$01.50/0 0 1987 American Chemical Society Communications to the Editor J. Am. Chem. SOC.,Vol. 109, No. 7, 1987 2209

Scheme 11" 0 C w VI Ph 00 oxo Rz 0 x 0 OSi'BuMe, X 1 2, R, = H, & -OSi'BUMe2 eC 3, R, = H. R, = oso,cbH,~Jd 4, R, = SiBUPh,, R2 = OSPBuMe, ' 5,R, = SiBuPh,, =OH 6, R, -Si'&IPh,, &-I 7, R, = Si'BuPh,. R, = P(O)(OMe),

ORTEP Drawing 01 Compound 3

0 , R, = Si'BuPb. R, = Si'BuMe, kcio.R, - R,= H

P (O - RO Ph

11 R, =R,-H. X-SMe 12, R, = F$ - H, X- OMe 15, R = CH,Ph nL 13.R = H. R, =W&I.X = OMe oc14.R:=Si I BuMe,, R, = H, X OMe lr OS?BuMe, 1

A&&R* 0 G,,, OS?BuMe,

x x x O%, OH 0 X 18,R-H Me0 lS,R-SfBuMe, 17 p 0

OSI'BuMe, RTOSi'BuMe, x OSi'BuMe,

X X O''+ COOMe x X COOMe MeO (MeO),; 0 00 IV 20, R -CH,Ok I1 - 111 WC21.R-COOH Y "Reagents and conditions. (a) VI, 1.1 equiv of NaH, DME, 0 OC, then 1.0 equiv of aldehyde V, -60 - -20 OC, 4 h, 94%. (b) 5% Pd-C catalyst, H2, EtOAc, 25 OC, 3 h, 98%. (c) 5.0 equiv of L-Selectride, THF, -120 OC, 2 h, 98%. (d) 1.5 equiv of n-Bu4NF, THF, 1 h, 25 OC, then 1.1 equiv of p-C1-C6H4SO2C1, 1.5 equiv of Et,N, DMAP catalyst, CH2C12,0 OC, 4 h, 90%. (e) 1.2 equiv of r-BuPh2SiC1, 1.5 equiv of imidazole, DMF, 0-25 OC, 4 h, 91%. (f) 1.1 equiv of n-Bu,NF, THF, 0 "C, 6 h, 84%. (9) 1.1 equiv of MsCI, 1.3 equiv of Et,N, CH2C12,-15 OC, 15 min then excess NaI, acetone, 25 OC, 8 h, 97% overall. (h) 1.2 equiv of (MeO),P(O)H, 1.2 equiv of NaH, DME:DMF (3:2), 45 "C, 4 h, 97%. (i) 1.1 equiv of LDA, THF, -78 OC then 1.1 equiv of MeSSMe, -78 OC, 5 min, 73%. 6) 1.3 equiv of LDA, THF, -78 OC then 1.1 equiv of aldehyde VII, -78 - 25 "C, 2 h, 84%. (k) 3.0 equiv of n-Bu,NF, THF, 8 h, 25 OC, 96%. (I) 1.1 equiv of CSA, CH2C12(0.05 M), 0-25 OC, 20 min, 64%. (m) 1.1 equiv of NBS, 3 A molecular sieves, CH2CI2-MeOH (lO:l), 0 OC, 10 min, 95%. (n) 1.1 equiv of t-BuCOCI, pyr, 0-25 "C, 8 h, 86%. (0) 1.1 equiv of t- BuMe2SiOTf, 1.3 equiv of 2,6-lutidine, CH2C12,0 OC, 10 min, 89% then 2.5 equiv of DIBAL, CH2C12,-78 "C, 15 rnin, 98%. (p) 5.0 equiv of PDC, DMF, 25 OC, 12 h, then CH2N2, Et20, 0 OC, 82% overall. (4) 10% Pd-C catalyst, H2,absolute EtOH, 25 "C, 3 days, then 1.2 equiv of Ac20, 3.0 equiv of DMAP, CH2C12,0 OC, 15 min, 67% overall. (r) 10% Pd-C catalyst, H,, absolute MeOH, 25 OC, 2 days, 76%. (s) 5.0 equiv of imidazole, CHJN, 25 OC, 10 h, 76%. (t) 1.1 equiv of t-BuMe,SiOTf, 1.3 equiv of 2,6-lutidine, CH2CI2,10 min, 0 OC, 80%. (u) 1.1 equiv of aqueous LiOH (1.0 M), THF, 0-25 OC, 20 min, then CH2N2,Et,O, 0 "C, 98% overall. (v) 5.0 equiv of PDC, DMF, 25 "C, 12 h, then CH2N2,Et20, 0 OC, 76%. (w) 5.0 equiv of K2C03,MeOH, 0 OC, 30 min, 95% then 5.0 equiv of PDC, DMF, 12 h, 79%. (x) 3.0 equiv of CH,P(0)(OCH3)2,3.0 equiv of n-BuLi, THF, -78 "C, then add 21, -78 OC, 15 min, 62%. (y) see ref 8. of anomers by 'H NMR). Differentiation between the primary with an acetate group, leading to 16 (67% overall yield) so as to and secondary hydroxyls of 12 was achieved via monopivalate ester allow for subsequent differentiations. Removal of the benzylidene 13 (86%), which was silylated (89%) and deprotected to afford group from 16 furnished diol 17 (76%), which underwent smooth 14 (98%). PDC oxidation' of 14 followed by lactonization to 18 by treatment with imidazole (76%), thus diazomethane treatment led to methyl ester 15 (82%); the benzyl temporarily engaging the primary hydroxyl group. Subsequent ether protection of 15 was then selectively removed and replaced silylation of the remaining free hydroxyl of 18 led to the disilyl ether 19 (80%). The highly sensitive lactone functionality of 19 (7) Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 399. was then dismantled (without acetate removal) by aqueous 2210 J. Am. Chem. SOC.1987, 109, 2210-2211

Scheme 111' methyl ester 25. Thus, the total synthesis of amphoteronolide B OR (1) was accomplished.I2 Acknowledgment. We express our many thanks to Dr. C. Cimarusti, The Squibb Institute for Medical Research, for gen- erous samples of amphotericin B, and to Drs. George Furst, Patrick Carroll, and John Dykins for their superb NMR, X-ray crys- tallographic, and mass spectroscopic assistance. This work was II , R c Si'BuMe, 0 financially supported by the National Institutes of Health, Merck 22,R.H La Sharp and Dohme, and Hoffmann-La Roche. Ib Supplementary Material Available: Listing of Rj CY]^, IR, UV, OH and 'H NMR data for compounds 4, 8, 9, 14, 17, 19, 21, 11, 111, IV,and 25 and a I3C NMR spectrum of 25 (7 pages). Ordering information is given on any current masthead page.

(1 1) After chromatographic purification (silica, 25-75% MeOH in CH2C12) and spectroscopic characterization, the aglycon I was methylated 23 0 (CH2N,, Et20-Me2S0, 25 "C) back to amphoteronolide B methyl ester, identical with an authentic sample, thus further confirming its structure. IC (12) All new compounds exhibited satisfactory spectral and analytical/ OH exact mass spectral data. Yields refer to spectroscopically and chromato- graphically homogeneous materials.

Preparation and Structure of ( NEt4)2[V4S2(SCH2CH2S)6]and Its Structural and 24, R, P R,. Me xdOH 25, R, =H .R,=Me Electronic Relationship to the Li,VS2 Phases I,R,=R,=HLe Joanna K. Money, John C. Huffman, and George Christou* Reagents and conditions. (a) excess HF-pyr, MeOH, 45 OC, 48 h, 55%. (b) 0.1 equiv of CSA, MeOH, 0-25 OC, 1 h, 50% based on re- Department of Chemistry and the covered starting material (10%) and a monoacetonide (25%, as yet Molecular Structure Center, Indiana University unidentified isomeric structure). (c) 10 equiv of NaBH,, MeOH, 0 Bloomington, Indiana 47405 OC, 98%. (d) 0.1 equiv of CSA, MeOH-H,O (9:1), 0-25 OC, 97%. (e) 10.0 equiv of 1 N LiOH, HzO, 0-25 'C, 1 h, 80% (75% conver- Received October 15, 1986 sion). The polymeric sulfides of the early transition metals often and the resulting hydroxy acid was converted to the dimethyl ester display interesting magnetic and electrical properties'i2 and have 20 by sequential methylation (CH2N2),PDC oxidation, and a proven to be of considerable importance to many areas, not least second methylation (CH2N2)(76% overall yield). The acetate of which are heterogeneous catalysis3 and employment as battery was then removed from 20 (95%) and the carboxylic acid 21 was electrode^.^ A current and important challenge to the synthetic obtained by PDC oxidation of the resulting primary alcohol (79%). inorganic chemist is the preparation of soluble, discrete structural Finally, differentiation among the three carboxyl groups in 21 counterparts of the polymeric metalsulfide phases to allow parallel (anion formation at C-1, steric congestion at C-16) was observed characterization of both the reactivity characteristics in homo- in the one-step, chemoselective conversion of this intermediate geneous solution and the intrinsic properties of the basic building to the requisite keto phosphonate acid I11 by attack of dimethyl block of the extended lattice. Such efforts have resulted in (1ithiomethyl)phosphonate at C-19 (62%). The second requisite considerable progress, particularly in the chemistry of soluble key intermediate, hydroxy aldehyde IV, was constructed from molybdenum sulfide^.^ We herein report the preparation and synthetic VII13 and two units of phosphonate IX (Scheme I) as properties of the first tetranuclear vanadium-sulfur-thiolate recently described.* Finally, coupling of I11 and IV (esterification, species. We believe this complex presages a rich new area of high 70%) followed by macrocyclization (intramolecular keto phos- nuclearity V/S/SR chemistry. In addition, we describe its phonatealdehyde condensation, 70%) according to the procedures structural and electronic correspondence to the Li,VS2 polymeric recently reported from these laboratories* led to heptaenone IL9 phases (0 5 x 5 1).6 Heptaenone I1 was then converted to I as outlined in Scheme 111. Reaction of VC13, Liz& Na2edt (edt is -1,2-dithiolate), Thus, desilylation of I1 (HFspyr-MeOH) afforded triol 22 (55%), and NEt4Br in a 3:4:3:6 ratio in MeCN yields an intensely brown which was then subjected to deacetonization (CSA-MeOH) solution that, after filtration and addition of diethyl ether, deposits leading to heptahydroxy heptaenone 23 (50% yield based on ca. large black prismatic crystals of (NEt4)2[V4S2(edt)6].2MeCNin 50% conversion). Sodium borohydride reduction of 23 led, analytical purity.' A single-crystal structure determination' shows stereospecifically,' to the amphoteronolide derivative 24 (98%). the anion (Figure 1) to possess a V4S2 central core with two The 19R configuration of the reduction product was confirmed by CD studies' and by comparisons of materials derived from 24 (1) Hullinger, F. Struct. Bonding (Berlin) 1968, 4, 83. and amphotericin B.Io Finally, sequential demethylation of 24 (2) Whittingham, M. S. J. Electrochem. SOC.1976, 123, 315. (CSA-MeOH, 97% followed by LiOH hydrolysis, 80% yield based (3) Weisser, 0.;Landa, S. Sulfde Catalysts: Their Properties and Ap- plications; Pergamon: New York, 1973. on ca. 75% conversion) led to amphoteronolide B (1)" via its (4) Rouxel, J.; Brec. R. Annu. Reo. Mater. Sci. 1986, 16, 137-162. (5) Muller, A. Polyhedron 1986, 5, 323. (6) Murphy, D. W.; Cros, C.; DiSalvo, F. J.; Waszczak, J. V. Inorg. Chem. (8) Nicolaou, K. C.; Chakraborty, T. K.; Daines, R. A.; Simpkins, N. S. 1977, 16, 3027. J. Chem. SOC.,Chem. Commun. 1986, 413. (7) Crystallographic data at -155 OC: triclinic; space group Pi; a = 11.130 (9) Both synthetic I1 and degradatively derived* I1 (two methoxy anomers, (3), b = 11.424 (3), c = 10.748 (3) A; a = 112.04 (l), @ = 94.82 (1). y = chromatographically separated) were found to be spectroscopically and 93.44 (1)'; 2 = 1; R = 0.0459, R, = 0.0467, using 3290 unique intensities chromatographically identical. Compounds 11-21 and 111 were carried with F > 30(F). All non-hydrogen atoms were refined anisotropically; hy- through the sequence as mixtures of methoxy anomers. drogen atoms were located in a difference Fourier and refined isotropically. (10) Nicolaou, K. C.; Daines, R. A,; Chakraborty, T. K.; Ogawa, Y.J. Anal. Calcd for C32H70N4S,4V4:C, 33.03; H, 6.06; N, 4.82. Found: C, Am. Chem. Soc., in press. 32.77; H, 6.01; N, 4.60.

0002-7863 /87/ 1509-2210$01.50/0 0 1987 American Chemical Society Communications to the Editor 803 1

=~.~Hz,~H),O.~O(~,J=~.~HZ,~H),1.17(~,3H),3.60 thetic compound, a single crystal X-ray study was carried out (d, J = 7 Hz, 1 H)) in 52% yield from 5b. using a Syntex P21 four-circle diffractometer. A complete data Treatment of the hydroxy ketone 2b with triphenylmeth- set was collected; the published coordinates' were refined using ylenephosphorane in dimethyl (1 2 h, 55 "C)Io gave Sheldrick's SHELX-76 least-squares program. The refinement the exo-methylene compound 2d6 (mp 68.0-69.0 "C (from converged, with a residual of 0.13, using isotropic thermal hexane): NMR 8veasi(CCI4) 0.74 (d, J = 7 Hz, 3 H), 0.87 (d, parameters and without the hydrogen atoms being included. J = 7 Hz, 3 H), 0.98 (s, 3 H), 3.60 (d, J = 7.5 Hz, 1 H), 4.63 A difference Fourier synthesis showed no peaks of electron (br s, 1 H), 4.75 (br s, 1 H)) in 92% yield. Reductive cleavage density greater than 0.5e/A3. This conclusively demonstrated of 2d with lithium in ethylamine (I min, 16 "C) gave a single that the synthetic material was in fact (-)-axisonitrile-3 alcohol, lb, in 90% crude yield.' I There was no evidence that (la).17 any of the 100 epimer of lb (retention of configuration in the ring opening) was obtained in this reaction. Without purifi- References and Notes cation lb was converted into the tosylate 1c6 (mp 81-82 "C (1) B. DiBlasio, E. Fattorusso, S. Magno, L. Mayol, C. Pedone, C. Santacroce, and D. Sica, Tetrahedron, 32, 473 (1976). dMqSi (from hexane); NMR (CCI4) 0.75 (d, J = 6 Hz, 3 H), (2) (a) W. G. Dauben and E. J. Deviny. J. Org. Chem., 31, 3794 (1966);(b) E. 0.78 (d, J = 6 Hz, 3 H), 0.86 (d, J = 6 Hz, 3 H), 1.63 (br s, 3 Piers and P. M.Worster, J, Am. Chem. Soc., 94, 2895 (1972);(c) D. Caine. H),2.45(~,3H),4.52(d,J=SHz,IH),5.13(brs,lH),7.23 W. R. Pennington, and T. L. Smith. Jr., Tetrahedron Lett., 2663 (1978). (3) J. A. Marshall, W. I. Fanta, and H. Roebke, J. Org. Chem., 31, 1016 (1966); (d, J = 8 Hz, 2 H), 7.72 (d, J = 8 Hz, 2 H)) using tosyl chlo- J. A. Marshall, G. L. Bundy, and W. I. Fanta, ibid., 33, 3913 (1968). ride in pyridine (96 h, 25 "C) in 60% overall yield from 2d. (4) SMC Corp., Glidden Organics, Jacksonville, Fla. (5) D. N. Kirk and J. M. Wiles, Chem. Commun., 518 (1970). We also carried out the conversion of the model tricyclo- (6) A correct combustion analysis has been obtained for this compound. decanone 2a into (+)-spiroaxene (Id)' in a similar manner. (7) A. F. Kluge, K. G. Untch, and J. H. Fried, J. Am. Chem. SOC., 94, 7827 (1972). We suggest the abbreviation MIP (i,e., rnethoxyisopropylidine) for Methylenation of 2a as above gave 2c (NMR d~~~~i(CCI4) this derivative. 0.82 (d, J = 6 Hz, 6 H), 0.97 (s, 3 H), 4.64 (br s, 1 H), 4.47 (br (8) (a) H. J. Reich, J. M. Renga, and I. L. Reich, J. Am. Chem. Soc., 97, 5434 s, 1 H)) which upon reaction with lithium in ethylamine gave (1975).(b) K. B. Sharpless, R. F. Lauer, and A. Y. Teranishi, ibid., 95, 6137 (1973). (c) D. L. J. Clive, J. Chem. Soc., Chem. Commun., 695 (1973).(d) Id6 (NMR bh.leasi(CC14) 0.75 (d, J = 6 Hz, 3 H), 0.83 (d, J For the specific application of this sequence to conversion of an octalone = 6 Hz, 6 H), 1.72 (br s, 3 H), 5.28 (br s, 1 H); [c?]*jD +I 1.6" to a cross-conjugated dienone, see D. Caine, A. A. Boucugnani, and W. R. Pennington. J. Org. Chem., 41, 3632 (1976). (c 2.0, ether)." (9) (a) For reviews covering the photochemical conversion of bicyclic cross- Bose, Kistner, and FarberI* have reported the conversion conjugated dienones into tricyclodecenones, see P. J. Kropp, Org. Pho- tochem.. 1, 1 (1967);K. Schaffner, Adv. Photochem., 4, 81 (1966). (b) A of menthyl tosylate into neomenthylamine via SN~reaction Hanau NK-20 low-pressure mercury lamp was used as the light source. with sodium azide in aqueous dimethylformamide followed (10) R. Greenwald, M. Chaykovsky, and E. J. Corey, J. Org. Chem., 28, 1128 (19631. by lithium aluminum hydride reduction of the azide. However, \ ~--, In general, the cyclopropane ring of simple vinylcyclopropanes is not attempted conversion of ICto the azide le using their proce- cleaved by metals in liquid (see S. W. Staley, Sei. Org. Transform., dure led primarily to the formation of elimination products as 2, 97 (1972)).Compound 2c was found to be stable to excess lithium in liquid ammonia at -33 'C for 2 h. However, 2c and 2d were rapidly cleaved did the use of potassium azide in containing 18- by lithium in ethylamine and extreme care had to be exercised to prevent crown-6. It was clear that it would be necessary to carry out overreduction. To our knowledge the conversions of 2c and 2d into ldand the tosylate displacement under conditions which would be lb, respectively, represent the first examples of reductive cleavages of vinylcyclopropanes with lithium in ethylamine. However, cleavages of more favorable to an Sy2 reaction. This was accomplished by strained divinylcyclopropanes such as octamethylsemibullavene with treating IC with 3 equiv of potassium azide in benzene con- lithium in liquid ammonia have been reported (W. T. Borden, A. Gold, and S. D. Young, J. Org. Chem., 43, 486 (1978)). taining 5 equiv of 18-crown-6 (48 h, 80 "C)." The azide le A. J. Bose, J. F. Kishner, and L. Farber, J, Org. Chem., 27, 2925 (1962). showed NMR dve4si(CC13) 0.71 (d, J = 6 Hz, 3 H), 0.93 (d, We are indebted to Professor Charles L. Liotta for discussions on this point J = 6 Hz, 3 H), 0.97 (d, J = 6 Hz, 3 H), 1.77 (br s, 3 H), 3.47 and for a gift of 18-crown-6. The yield of the amine If obtained in the reduction of the azide le with (s, 1 H), 5.20 (br s, 1 H); mass spectrum (70 eV), no M+, m/e LiAIH4 in ether was surprisingly low. Other methods for accomplishing this 219 (M - N2, weak), 204 (M - "3, strong). Reduction of conversion are being explored. W. R. Hertler and E. J. Corey, J. Org. Chem., 23, 1221 (1958). le with LiAIH4 in ether at reflux gave the amine If (NMR Spectral data for (-)-axamide-3 were not reported in ref 1, 6~~~~j(CC14) 0.71 (d, J = 6 Hz, 3 H), 0.87 (d, J = 5 Hz, 6 H), We are very grateful to Dr. Donald G. VanDerveer for his assistance in 1.73 (br s, 3 H), 2.63 (br s, 1 H), 5.27 (br s, 1 H)) in 30% carrying out the X-ray work. overall yieldI4 from IC. Drur? Caine,* Howard Deutsch Using a procedure analogous to that of Hertler and Corey," School of Cheniisfrj, Georgia Im?i?u~rof Technologj If was converted into (-)-axisonitrile-3 (la) in 85% overall Atlanta, Georgia 30332 yield by treatment with a 2: I mixture of in acetic anhydride at reflux for 2 h followed by reaction with tosyl Received June 2, I978 chloride in pyridine at 25 "C for 1 h. (The deriva- tive, presumably (+)-axamide-3,l,lh was isolated as an inter- mediate in this sequence.) The synthetic material6 showed mp Stereospecific Total Synthesis of Gibberellic Acid. 97-99 "C (from hexane); NMR 8ve4si(CC14) 0.75 (d, J = 6.5 A Key Tricyclic Intermediate Hz, 3 H), 0.93 (br d, J = 6.5 Hz, 6 H), 1.75 (br s, 3 H), 3.52 (br s, 1 H), 5.14 (br s, 1 H): Id (CC14) 2120 cm-l; [c?I2j~ Sir: -71" (C 0.35, CHC13). Since the recognition of the central biological role of gib- These physical properties generally agreed with those re- berellic acid (gibberellin A3, GA3) (1) in the kingdom,' ported for (+)-a~isonitrile-3,~the enantiomer of la, except for the clarification of its chemical structure,2 and commercial the sign of the optical rotation.' However, there were small production on a large scale from the Gibberellafuji- discrepancies between the observed NMR chemical shifts, kuroi, this substance has occupied a major position in the field particularly for the methyl groups of the isopropyl group, and of natural product^.^ The of gibberellic acid from those reported by Sica and co-workers.' Therefore, verification prenyl units, though long and involved, is known in consider- of the structure of the synthetic material was desirable. Un- able detai1.'$3,4Despite extensive efforts (some 150 published fortunately, a direct comparison of the synthetic material with papers from about 25 different laboratories), the total chemical the natural product could not be made since neither a pure synthesis of gibberellic acid has not previously been a~hieved,~ authentic sample nor copies of the original spectral data were largely because the combination of overall molecular com- available to us. In order to confirm the structure of the syn- plexity, centers of high sensitivity toward many reagents, and

0002-7863/78/1500-8031$01.00/0 0 1978 American Chemical Society 8032 Journal of the American Chemical Society / 100.25 / December 6, 1978

Scliemc I

a singularly diabolical placement and density of functionality reduction of the noraldehyde so obtained with sodium boro- serves to thwart all but the most sobhisticated of approaches.6 hydride in absolute ethanol at 0 OCfor 20 min, (3) benzylation We report in this and the following publication the first total of the resulting primary alcohol via the sodium salt (excess synthesis of gibberellic acid by a route which is structurally NaH in THF) in THF with excess benzyl bromide at reflux unambiguous and stereospecific, and which employs a number for 15 h, and (4) selective cleavage of the MEM group using of crucial new synthetic method^.^ 1.5 equiv of trifluoroacetic acid in methylene chloride at 23 OC The plan of synthesis, which was derived by extensive an- for 18 h. After column chromatography pure phenol 3 was tithetic analysis, is outlined briefly in Scheme I.8 Key features obtained in 74% yield overall from 2 as a colorless oil. Oxida- of the approach as revealed in this summary include (men- tion of the phenol 3 to the yellow quinone 4 (Scheme I), mp tioned in the order of synthetic execution) (1) the stereospecific 7 1-72 OC, was effected by stirring in dimethylformamide so- generation of the cis-fused B/C ring unit by Diels-Alder ad- lution at 23 OC for 4 days with molecular in the pres- dition, (2) formation of the D ring by internal pinacol cycli- ence of 0.08 equiv of bis(salicylidene)ethylenediiminoco- zation, (3) position-specific ring contraction of ring B from six balt( 11) (salcomine)' 3,14 (>75% yield). to five members, (4) formation of ring A by internal Diels- Reaction of the quinone 4 with trans-2,4-pentadien-l-ol Alder reaction and stereospecific methylation to form a pen- (5)15 occurred upon heating in benzene solution at reflux for tacyclic lactone with all carbons in place, (5) oxidation and 30 h to afford a single crystalline adduct (6)in 91% yield.I6 isomerization at C(6) and C(7) of the pentacyclic lactone, and The next phase of the synthesis, transformation of the adduct (6) stereospecific elaboration of the complete A/B ring unit 6 to the keto aldehyde 7, was originally attempted using the of gibberellic acid.9 well-known Woodward pr0~edure.l~~However, because this The phenolic ether 2 was prepared in two steps and -75% direct method failed completely, alternative routes were ex- overall yield from 2-allyloxyanisole by (1) Claisen rear- amined.17b The most satisfactory from the standpoint of re- rangement at 230 OC for 2 h (89% of distilled product, bp producibility and ease of scale-up consisted of the following 87-97 OC at 0.6 mm, containing 93% ortho-allyllic and 7% sequence: (1) reaction of 6 with 1.1 equiv of dihydropyran and para-allylic phenol),I0 followed by (2) etherification of the 0.12 mol % p-toluenesulfonic acid in methylene chloride ( 10 resulting phenol using sequentially 1.3 equiv of sodium hydride mL/g of 6) at 0 "C for 18 h to form quantitatively the te- in tetrahydrofuran (THF) at 0 OC and 1.6 equiv of chloro- trahydropyranyl (THP) ether; (2)reduction of the THP ether methyl 2-methoxyethyl ether (MEM chloride"), first at 0 OC with 1 mol equiv of sodium borohydride in absolute ethanol at and then at 25 OC for 2 h.I2 Conversion of 2 to the phenol 3 was 0 OC for 100 min to produce the hydroxy enone 8 (100%);(3) accomplished by the sequence (1) oxidation of 2 with 3 equiv conversion of 8 to the a-methoxymethylenoxy ketone 9 by of sodium and 0.2 mole % of osmium tetroxide in 3: 1 reaction with 8 equiv of N,N-diisopropylethylamine and 4 THF-H20 at 0 "C for 0.5 h and 23 "C for 1.5 h, (2) immediate equiv of chloromethyl methyl ether in methylene chloride at

ROCH, OH OCH, / OMEM PhCH,O OCH,Ph OCH,Ph 2 -3 1 R=H 3 R:CH,0CH3 uF1

QH THPO' o, CH, QOcHs THPO'QOH H THPO~ -0 OCO,CH, OH :q OCH,Ph Communications to the Editor 8033 reflux for 9 h (100%); (4) reduction of the keto group in 9 with justment of size and pattern of functionality of the B ring of lithium aluminum hydride in ether at -10 OC for 1 h, isolation, the tricyclic intermediate 16, which was achieved as follows. and immediate mesylation of the resulting alcohol at -58 OC Treatment of a solution of 16 in acetone-water (2.5:l)with in THF with 2 equiv each of methanesulfonyl chloride and 1.3 equiv of N-methylmorpholine N-oxide and 0.05 equiv of triethylamine, followed by slow addition of saturated aqueous osmium tetroxide at 23 "C for 80 h furnished a single cis diol potassium bicarbonate and gradual warming to 0 OC over 45 (17) (89% yield after chr~matography)~~which was cleaved min to effect solvolysis, and finally chromatography of the by reaction with 1.03 equiv of lead tetraacetate in benzene at product on silica gel to give 10 in 77% overall yield from 6; (5) 5 OCfor 0.5 h. The sensitive dialdehyde 18, isolated from the selective hydrogenation18 of the enone double bond of 10 using reaction mixture simply by addition of ether, filtration through 1 .O equiv of hydrogen and 5% rhodium-on-carbon catalyst at Celite-anhydrous sodium sulfate, and concentration in vacuo 23 OC in THF; and (6) addition of a solution of the hydroge- was used directly in the next step without purification (and nation product and tert-butyl alcohol (10 equiv) in THF to 12 with minimal delay26). Treatment of 18 with 0.2equiv of di- equiv of lithium in liquid ammonia at -78 "C over 7 min, benzylammonium triflu~roacetate~'in benzene at 50 OCfor stirring at reflux for 6 h, quenching with ammonium chloride, 1 h gave after chromatography on silica gel the desired a$- and isolation by chromatography on silica gel to yield the THP unsaturated aldehyde 19 in 64% overall yield from the diol 17. ether diol 11 in 63% overall yield from 10. Oxidation of 11 with The yield of 17 from the cyclization is actually higher since 19 dipyridinechromium(V1) oxide (Collins reagent) (excess) in undergoes partial decomposition on silica gel; in practice, methylene chloride at -45 "C for 2 hand -25 OC for 1 h with therefore, 19 was used for the next step without purification. stirring in the presence of dry, acid-washed Celite gave after Reaction of 19 with 5 equiv of methylenetriphenylphosphorane treatment with powdered sodium bisulfate monohydrate in THF-hexamethylphosphoramide at reflux for 3.5 h fur- (stirring at -20 OCfor 30 min), dilution with dry ether, fil- nished the Wittig product 20 in good yield (44% overall from tration, and concentration the sensitive keto aldehyde 7 in 84% diol 17; 80% from purified keto aldehyde 19). Exposure of 20 yield as a pale yellow oil which was used as such in the next step to acetic acid-THF-water (3:l:l)at 35 OC for 40 h produced (intermediate storage at -78 OCfor a minimal period). the diene alcohol 21 cleanly without detectible cleavage of the The persistence of a cis-ring fusion in the various interme- MEM protecting group (32% overall for 5 steps from 16).The diates derived from the Diels-Alder adduct 6 could be dem- successful production of the critical tricyclic intermediate 21 onstrated chemically. Thus, reaction of 8 with N-bromosuc- in quantity set the stage for the completion of the synthesis of cinimide in THF at 23 "C afforded quantitatively the bridged gibberellic acid as described in the following communica- bromo ether 12. Further, reaction of the diol 11 with 1 equiv ti0n.2~J9 of methyl chloroformate-pyridine gave selectively the car- References and Notes bonate of the primary alcohol which was oxidized (Collins reagent) to the corresponding cyclohexanone and treated with (1) See, H. N. Krishnamurthy, Ed., "Gibberellins and Plant Growth", Wiley, New York, 1975. p-toluenesulfonic acid-methanol to produce in high overall (2) (a) P. W. Brian, J. F. Grove, and J. MacMillan, Fortschr. Chem. Org. Naturst., yield the bridged ketal 13. 18, 350 (1960); (b) J. F. Grove, 0.Rev. (London). 15, 56 (1961); (c) F. M. The pinacol cyclization of the keto aldehyde 7 to the tricyclic McCapra, A. I. Scott, G. A. Sim, and D. W. Young, Proc. Chem. SOC., London, 185 (1962); (d) J. A. Hartsuck and W. N. Lipscomb, J. Am. Chem. intermediate 14 proved to be a more difficult proposition than SOC.,85, 3414 (1963). expected from earlier studies of closely related models.9a The (3) See (a) J. R. Hanson, "The Tetracyclic Diterpenes", Pergamon Press, Oxford, 1968; (b) K. Nakanishi, T. Goto, S. Ito, S. Natori, and S. Nozoe, most satisfactory and convenient procedure involved prepa- "Natural Products Chemistry", Vol. 1, Academic Press, New York, 1974, ration of finely powdered metallic titanium under by p 265. addition of small pieces of potassium to a mixture of titanium (4) See also (a) B. E. Cross in "Progress in Phytochemistry", Vol. 1. Inter- science, New York, 1968; (b) B. Dockerill, R. Evans, and J. R. Hanson, J, trichloride in THF (8.5 equiv of Tic13 and 24 equiv of K)9a319 Chem. SOC.,Chem. Commun., 919 (1977), and papers therein cited. and then heating cautiously at reflux for 3 h, cooling to 23 OC, (5) A number of simpler gibberellins and degradation products of gibberellic and gradual addition of the keto aldehyde 7 in THF. The re- acid have been synthesized. See (a) K. Mori, M. Shiozaki. N. Itaya. T. Ogawa, M. Matsui, and Y. Sumiki, Tetrahedron Lett., 2183 (1968),and K. action mixture was stirred at 23 OC for 2.5 h, cooled to 0 OC, Mori, M. Shiozaki, N. Itaya, M. Matsui, and Y. Sumiki, Tetrahedron, 25, 1293 and very cautiously treated dropwise with anhydrous methanol, (1969), for a 52-stage synthesis of gibberellin Aq; (b) W. Nagata, T. Wak- abayashi, Y. Nayase. M. Narisada, and S. Kamata. J. Am. Chem. SOC.,93, diluted with aqueous K2C03, filtered through Celite, and ex- 5740 (1971), and 92, 3202 (1970), for a 49-step route to gibberellin AIS; tracted with 4:1 ether-methylene chloride. After isolation the (c) H. 0. House and D. G. Melillo, J. Org. Chem., 38, 1398 (1973), and K. Mori, Tetrahedron, 27, 4907 (1971), for syntheses (>30 steps) of epial- product (-10-g batch size) was chromatographed on a iogibberic acid. Associates Model 500 preparative machine which separated (6) For a discussion see R. L. Danheiser, Ph.D. Dissertation, Harvard University, the three major components easily and afforded 40% cis- 14, 1978. (7) The problem of the synthesis of gibberellic acid has provided the impetus 15% trans-14, and -10% diol 11 (which was recycled).20 for the development of many new synthetic methods,6 for example from Oxidation of either cis- or trans- 14 (or a mixture of the two) the laboratories of Stork (see G. Stork, D. F. Taber, and M. Marx, Tetra- to form the ketol 15 without appreciable glycol cleavage was hedron Left., 2445 (19781, and references therein cited), Loewenthai (see H. J. E. Loewenthal and S. Schatzmiller, J. Chem. Soc., Perkin Trans. 7, readily accomplished via an oxysulfonium intermediate, in 2149 (1975), and references cited), S. Masamune (see S. Masamune, J. accordance with previously published results.2' The most Am. Chem. SOC.,83, 1009 (1961); 86, 288 (1964)), Dolby (see L. J. Dolby and R. H. iwamoto, J. Org. Chem., 30, 2420 (1965)), Mander (see D. J. satisfactory conditions for the oxidation involved addition of Beames, J. A. Halleday, andL. N. Mander, Aust. J. Chem., 25, 137 (1972)). pinacol(s) 14 to a suspension of the complex formed from 7 House (see H. 0. House, D. G. Meiillo, and F. J. Sauter. J. Org. Chem., 38, equiv of and equiv of trichloroacetic 741 (1973)),and Ziegler (see F. E. Ziegler and J. A. Kloek, Tetrahedron 3.5 Lett..~. 315 (19741) .~ -\ -. ,I anhydride in methylene chloride at -60 "C for 1 h, stirring the Much of the work described herein has been disclosed in various lectures, mixture of reactants at -50 "C for 45 min, treatment with 3.5 e.g., at the Nichols Award Symposium, Tarrytown, N.Y., March 1977, and at the International Symposium on , Oxford, England, equiv of triethylamine at -50 to +23 OCover 2.5 h, and finally 1977. extractive isolation (ether-methylene chloride).22The ketol Certain key stages of this synthetic approach had already been tested and 15 (yield 75-80% if purified chromatographically) was con- described in previous publications from this laboratory. These include (a) internal inacol cyclization of a keto aldehyde (E. J. Corey and R. L. Carney, verted directly without p~rification~~to the MEM ether de- J. Am. &em. Soc., 93, 7318(1971), and E. J. Corey, R. L. Danheiser, and rivative 16' by treatment with 3 equiv of MEM chloride S. Chandrasekaran, J. Org. Chem., 41, 260 (1976)): (b) internal Dieis-Aider approach to the construction of the AI6 ring unit (E. J. Corey and R. L. and IO equiv of diisopropylethylamine in methylene chloride Danheiser, Tetrahedron Left., 4477 (1973)); (c) stereospecific placement at reflux for -16 h (65% yield overall from 14 after chroma- of all A ring ring (E. J. Corey. T. M. Brennan, and R. L. Carney, tographic purification on silica gel). J. Am. Chem. SOC..93. 7316 11971\1 The mixture was used as suchwith purification conveniently effected at The next stage of the synthetic approach required the ad- a subsequent stage 8034 Journal of the American Chemical Society 1 100:25 1 December 6, 1978

(11) E. J. Corey, J.-L. Gras, and P. Ulrich. Tetrahedron Lett., 809 (1976). the approach detailed in the preceding publication.' In addition (12) Structural assignments for all stable synthetic intermediates are based upon proton magnetic (lH NMR), infrared, and mass spectra deter- we disclose a new facet of the chemistry of gibberellic acid mined using purified, chromatographically homogeneous samples. In ad- which allows access to derivatives in which the C(7) substituent dition ultraviolet spectra were determined where appropriate and were also consistent with the formulations shown herein. All reactions involving air- on ring B is in the unnatural (and generally less stable2) cy or- or moisture-sensitive components were carried out in an atmosphere of ientation and which also provided useful direct correlation of drv-., nronn GA3 with a number of advanced synthetic intermediates. (13) (a) H. M. Van Dort and H. J. Geursen, Recl. Trav. Chim. Pays-Bas, 86, 520 (1967); (b) L. H. Vogt, Jr., J. G. Wirth, and H. L. Finkbeiner, J. Org. Chem., Deprotonation of the hydroxy diene 2 with 1 .O equiv of n- 34. 273 11969). butyllithium in tetrahydrofuran (THF) at -40 "C followed (14) The quinone 4could also be obtained in high yield from the phenol 3 by oxidation with 2 equiv of Fremy's salt in aqueous methanol; however, this by acylation with 1.55 equiv of trans-2-chloroacryly1 chloride3 procedure was less convenient largely because of the labor involved in at -40 "C for 0.5 h afforded the ester 3 in -80% yield (-62% preparing the reagent. Using the reactions outlined above the quinone 4 overall from the THP ether of 2).4 When 3 was heated in can be prepared reproducibly in 100-9 lots. (15) (a) R. G. Glushov and 0. Y. Magidson, Med. Prom. SSSR, 16, 27 (1962); benzene solution containing -100 equiv of propylene oxide (as (Chem. Abstr., 58, 4420 (1963)); (b) S. Oida and E. Ohki, Chem. Pharm. a hydrogen chloride scavenger) in a sealed tube at 160 "C for Bull., 17, 1990 (1969). 45 h under argon the pure crystalline lactone 4, mp 149-1 50 (16) The structure and stereochemistry of the adduct was anticipated to be that expressed by 6 on the basis of much precedent in the literature, and this "C, could be obtained in 55% yield after recrystalli~ation.~The expectation is fully confirmed by the data which follow. stereochemistry of 4 is assigned from the supposition of con- (17) (a) R. B. Woodward, F. Sondheimer. D. Taub, K. Heusler, and W. M. McLamore. J. Am. Chem. Soc., 74, 4223 (1952). (b) For a subsequent certed, a-face, "endo" internal Diels-Alder addition (there application of one of the sequences developed in the course of our in- was no evidence for the formation of an appreciable amount vestigations, see J.-L. Gras, Tetrahedron Lett., 41 17 (1977). of any stereoisomer of 4); it is supported by NMR data and (18) See S. K. Roy and D. M. S. Wheeler, J. Chem. SOC.,2155 (1963). 'H (19) See J. E. McMurry and M. P. Fleming, J. Org. Chem., 41, 896 (1976). For also by subsequent transformation to GA3. Treatment of the a detailed account of the very extensive studies carried out on the cycli- adduct 4 with 2.2 equiv of lithium isopropylcyclohexylamide zation of 7 and related substances with a wide range of reagents, see also ref 6. and 5 equiv of hexamethylphosphoramide in THF at -78 "C (20) The less polar of the isomeric pinacois, mp 97-99 OC, is probably cis-I4 for 50 min followed by reaction with 5 equiv of methyl iodide and the more polar isomer, mp 87.5-89 OC, trans-14, based upon previous at -78 to 0 "C over 12 h afforded cleanly the methylated experience with tricyclic analogues of known configuration (see, for ex- ample, ref 9a) and also on chemical data. lactone 5 (-75% yield). At this stage the MEM6 protecting (21) See, E. J. Corey and C. U. Kim, J. Org. Chem., 38, 1233 (1973). The various group was removed from 5 by stirring in dry chloroform- known chromium(Vi) reagents and many other standard oxidizing agents for afford mainly glycol fission products with substrates such as ether-nitromethane (1 55:1 by volume) with 25 equiv of finely 14. powdered anhydrous zinc bromide at 23 OC for 3 h to yield (22) This is a useful modification of the procedure of Swern; see K. Omura, A. hydroxy lactone 6 (-70% after chromatography). The IR, IH K. Sharma, and D. Swern, J. Org. Chem., 41, 957 (1976), and S. L. Huang, K. Omura, and D. Swern, ibid., 41, 3329 (1976). NMR, UV, and mass spectra and the TLC mobility of this (23) The keto1 15 is prone to 1,2 rearrangement of methylene at the bridgehead material were identical with those of a sample of optically to form the stereoisomeric a-ketoi upon exposure to base or acid or pro- active 6 obtained from natural gibberellic acid as described longed chromatography.This aiiogibberlc - gibberic type rearrangement is driven by the relief of strain in going from cis-fused BIC rings (skew-boat below. C ring) to trans-fused BiC rings (chair C ring). The occurrence of the same The synthetic (&)-hydroxy lactone 6 was resolved using a rearrangement with 17-nor-17-oxoailogibberic acid represented an in- consistency in the originally assigned stereochemistry of gibberellic acid novel procedure designed to take advantage of the lone (ter- which led us to propose the X-ray crystallographic study (see ref 2d) that tiary bridgehead) hydroxyl in 6. Exposure of 6 to a large excess eventually produced the correction of the earlier configurational assign- ment at C(9). of phosgene and 3 equiv of 4-dimethylaminopyridine in dry (24) The fl-methoxyethoxymethyl (MEM) protecting group'l was originally de- methylene chloride at 23 "C for 36 h gave, after rapid filtration veloped for this specific application. through dry Celite and concentration in vacuo, crude chloro- (25) For method see V. Van Rheenen, R. C. Kelly, and D. A. Cha, Tetrahedron Lett., 1973 (1976). formate 77 which was directly treated with (-)-a-phenyleth- (26) As might be expected the diaidehyde 18 is quite unstable (e.g., to water ylamine ( [cY]*~D-41.7" in benzene) to provide after isolation or silica gel). (27) The use of this outstanding selective reagent (a crystalline solid) for this a mixture of two diastereomeric urethanes (8) (95% total yield) very demanding step (see ref 6) was arrived at by systematic experimental which could be separated cleanly by chromatography on silica variation of secondary amine and acid components based on the idea of gel using 1 : 1 -hexane for elution (TLC R, values activating the methylene group a to the less hindered formyl group as an enamine by means of a not-too-basic, sterically discriminating secondary in this solvent system, 0.24 and 0.20). The less polar diaste- amine under almost neutral aprotic conditions. Other studies with this re- reomer, [RI2'D +59" (c 0.44, CHC13), was identical spectro- agent will be published separately. Since these investigations one of the scopically (IR. 'H NMR, mass spectrum) and chromato- undersigned has successfully applied a similar reagent to the direct ru methylenation of ; see J.-L. Gras. Tetrahedron Lett., 21 11 graphically with urethane prepared from hydroxy lactone 6 ( 19 78). from natural GAj and (-)-a-phenylethylamine which showed (28) This investigation was supported financially by the US. National Science Foundation to whom we are deeply grateful. [a]15~+61" (c 0.42, CHCI,). Reaction of this less polar (29) We are pleased to acknowledge helpful information and experimental data synthetic urethane 8 with 5 equiv of triethylamine and 3 equiv from the following colleagues: Drs. Sandor Barcza, Thomas M. Brennan, Robert L. Carney, Tetsuo Hiraoka, Masayuki Narisada. George Strunz, and of trichlorosilane in dry benzene at 25 "C for 60 h8 afforded Gerald L. Thompson. in 95% yield resolved hydroxy lactone 6, mp 21 1-212 "C, [.]*OD +162" (c 0.58, CHC13), identical in all respects (IR, E. J. Corey,* Rick L. Danheiser 'H NMR, mass spectrum, TLC high pressure liquid chro- Srinivasan Chandrasekaran, Patrice Siret matography) with the hydroxy lactone 6 derived from natural Gary E. Keck, Jean-Louis Gras GA3 which showed [.I2'D +161" (c 0.49, CHC13). Department of Chemistry, Harvard Unicersity The optically active lactone 6 was hydrolyzed to the corre- Cambridge, Massachusetts 02138 sponding hydroxy acid salt by heating at reflux with excess 1 .O Receiced September 5, I978 N aqueous potassium hydroxide for 45 min (argon atmo- sphere) and the resulting solution was treated at 23 "C with 2.07 equiv of 0.013 M sodium ruthenate9 in 1 N aqueous SO- dium hydroxide for 2.5 h. Filtration through Celite. acidifi- Stereospecific Total Synthesis of Gibberellic Acid cation to pH 3 at 0 "C, and extraction afforded upon isolation the diacid 9, spectroscopically and chromatographically Sir: identical with the diacid obtained from GA, (see below); the This communication describes the completion of the ste- corresponding dimethyl esters (from excess CHIN>in ether) reospecific total synthesis of gibberellic acid (GA3) (1) from were also identical. The formation of diacid 9 clearly proceeds a key tricyclic intermediate (2) which is readily accessible by by way of the intermediate acid aldehyde which undergoes

0002-7863/78/ 1500-8034$01 .OO/O 0 1978 American Chemical Society Communications to the Editor 8035

1 2 R:H I 3 R=C! H - c:c; " co

%OH n HO'-OH RORo~o,,

CO,R co2cn1 CO,CH, L!

-OH -OH

QAOAO '0% rs base-catalyzed epimerization to the more stable 6P-formyl function under equilibrating conditions (triethylamine catal- derivative and then further oxidation to the observed product.1° ysis) and subsequent capture of the C(6) a-oriented carbonyl Selective monoesterification of the diacid was accomplished by the 4a-carboxylic group. Reduction of 15 with -0.6 mol in THF by treatment with triethylamine (1.5 equiv) and p- equiv of lithium borohydride in dimethoxyethane at -25 "C toluenesulfonyl chloride (1 equiv) at -78 OC for 0.5 h and -50 for 1.5 h yielded, upon acidification with acetic acid, workup, OC for 2 h (to form the mixed sulfonic anhydride), subsequent and chromatography on silica gel, the lactone 6 (50%) together quenching with excess methanol at -50 OC initially, and then, with a structurally isomeric lactone (16, 16%); R, values for after warming to 23 OC, stirring for a further 2 h. Chroma- 6 and 16 were 0.74 and 0.85, respectively (silica gel plates, tography afforded the monoester 10 as a solid foam, [cY]*OD ethyl acetate-acetic acid, 95:5). The isomeric lactone 16, mp -21 O (c 4.9, THF), identical in all respects with the compound 171 OC, [(Y]20D +133O (c 8.7, CHC13), was synthesized un- obtained from GA3 as described earlier.Il From this optically ambiguously from the ester acid 10 by the following sequence: active intermediate (10)I2the synthesis of GA3 is completed (1) reaction of the tetra-n-butylammonium salt of 10 in dry by the previously describedi1 route which includes (1) hy- THF with 1 equiv of mesitylenesulfonyl chloride at -78 - 23 droxylactonization of 10 with m-chloroperbenzoic acid to form OC over 2 h to form the mixed sulfonic anhydride; (2) reduction 11;13 (2) lactone saponification and iodolactonization of 11 to of the activated 4-carboxylic group to a 4-hydroxymethyl group give the iodolactone 12; (3) in one flask, trifluoroacetylation (without isolation) at 0 OC by addition of excess sodium of 12 to 13, reduction with zinc to eliminate the 1-iodo and borohydride and reaction at 0 OC for 1 h: and (3) epimerization 2-trifluoroacetoxy substituents and bicarbonate treatment to at C(6) and concomitant lactonization of the resulting dihy- saponify the 3-trifluoroacetate forming GA3 methyl ester; and droxy ester (14) by heating at reflux with sodium methoxide finally (4) conversion of GA3 methyl ester to the free acid 1 in absolute methanol for 48 h. using sodium n-propyl mercaptide in hexamethylphosphora- Lithium borohydride reduction of either lactone 6 or 16 mideI4 at 0 OC. afforded the triol 17 (colorless, foam), [(ulZ0~-17.0' (c 3.5, The transformation of gibberellic acid 1 to the key inter- THF), oxidation of which with chromic acid (two phase, mediate 6 and several other gibberellins having the C(7) sub- ether-water) led exclusively to lactone 16 (no detectible stituent a-oriented at C(6) was achieved by the use of a novel 6)." strategy for effecting the 60 - 6a epimerization which is The research results described in this and the foregoing normally contrathermodynamic in this series. paper mark the achievement of one of the more intriguing and Saponification of the acid ester loi5by heating at reflux salient objectives in the area of organic synthesis. They also with excess 1 N aqueous potassium hydroxide for 40 min af- provide a basis for further synthetic and transformational in- forded after acidification and isolation the diacid 9 (95% yield), vestigations relating to gibberellic acid, and we hope to report [CY]"D -25.5' (c 2.4, THF), which could be reesterified with on the ongoing work in this area in due course.lx.Iy excess diazomethane to the same dimethyl ester obtained by methylation of 10 (indicating that no epimerization at C(6) References and Notes occurs in the saponification). Reaction of the diacid 9 with IO (1) E. J. Corey, R. L. Danheiser, S. Chandrasekaran. P. Siret, G. E. Keck. and equiv of triethylamine and 1 equiv of N,N'-dicyclohexyl car- J.-L. Gras, J. Am. Chem. Soc.. preceding paper in this issue. (2) (a) J. F. Grove, 0. Rev. (London), 15, 56 (1961); (b) E. J. Corey and R. L. bodiimide in THF at reflux for 7 h furnished, upon workup and Danheiser, Tetrahedron Lett., 4477 (1973). chromatography on silica gel, the anhydride 15, mp 167-1 68 (3) (a) A. N. Kurtz. W E. Billups. R. B. Greenlee. H. F. Hamil. and W. T. Pace, OC, [F]*~D+268O (c 9.3, CHCI3) (73% yield). The stereo- J. Org. Chem., 30, 3141 (1965); (b) P. K. Freeman, B. K. Stevenson, D. M. Balls, and D. H. Jones, ibid., 39, 546 (1974). chemistry of the anhydride, anticipated to be as shown in 15 (4) Satisfactory infrared (IR), proton magnetic resonance ('H NMR), and mass on geometrical grounds, was shown by methanolysis spectral data were obtained on purified, chromatographically homogeneous (CH~OH-C~HSN)and methylation of the resulting acid-ester samples of the synthetic intermediates described herein. All reactions involving air- or moisture-sensitive components were carried out in an with diazomethane to produce a dimethyl ester stereoisomeric atmosphere of dry argon. with that obtained by methylation of 9 or 10.l6The success of (5) In earlier model studies of this internal Diels-Alder step,lb satisfactory cyclization was obtained both with trans-2-chloroacrylate and propiolate the 6/3 - 6a epimerization involved in the formation of the esters. In contrast to the earlier studies the conversion of 3 to 4 failed anhydride 15 depends on activation of the C(6) carboxylic acid completely in the absence of propylene oxide. The cyclization of the 8036 Journal of the American Chemical Society / 100:25 / December 6, 1978

propiolate ester of 2 was not studied since we were unable to find conditions for its formation in high yield (from propiolic anhydride or other carboxyl- activated derivatives of propiolic acid), again in contradistinctionto earlier model studies or model experiments with other primary or secondary al- cohols. (6) E. J. Corey, J.-L. Gras, and P. Ulrich, Tetrahedron Lett., 809 (1976). (7) The chloroformate 7 reacted rapidly with water to re-form with various 6, Me alcohols to form the corresponding mixed carbonates as well as with pri- mary amines to form the expected urethanes. (8) Method of W. H. Pirkle and J. R. Hauske, J. Org. Chem., 42, 2781 eCN (1977). (9) D. G. Lee, D. T. Hall, and J. H. Cleland, Can. J. Chem., 50, 3741 (1972). As the orange solution containing the reagent was added to the reaction mixture a black precipitate developed. (10) Analogous 01 - p epimerization of 6 n-formyl derivatives in a number of related structures has been observed in these laboratories (see also ref 2b); it occurs readily and completely even under mild conditions (triethyl- amine, Oediger's base, or chromatography on silica gel). Me (1 1) E. J. Corey, T. M. Brennan, and R. L. Carney, J. Am. Chem. SOC.,93,7316 &* (1971). For another less direct synthesis of 10 from 6, see R. L. Danheiser, AX Ph.D, Dissertation, Harvard Universtiy, 1978. (12) The optically active ester acid 10 may also be obtained by resolution of 4: X=O;Y=CN racemic 10 using (-)+ 1'-napthy1)ethylamine with ethyl acetate-ether for crystallization. Y = CN (13) Hydroxy lactone 11 could also be obtained from diacid 9 by hydroxylac- tonization (1.05 equiv of in water-ethyl acetate at pH 9) Y = C02H followed by esterification with diazomethane. (14) P. A. Bartlett and W. S. Johnson, Tetrahedron Lett., 4459 (1970). (15) Derived from naturally occurring GAS. (16) The dimethyl ester correspondingto diacid 9 had R, 0.27. whereas the 6a epimer had R, 0.25 (silica gel plates with 1:l ethyl acetate-hexane); the epimeric dimethyl esters were very easily separable by high pressure liquid chromatography (l-min difference in retention time with heptane-ether- (lOO:lO:l)). (17) The naturally derived triol 17 was spectroscopically and chromatographi- cally identical with the triol obtained by lithium borohydride reduction of synthetic (f)-6. (18) We are indebted to the following colleagues for experimental help at various times: Drs. Yoshihiro Hayakawa, Thomas M.Brennan, and Robert L. Carney. Our sincere thanks are extended to Imperial Chemical industries Ltd., Merck and Co., Abbott Laboratories,and Chas. Pfizer and Co. for their generosity in donating samples of gibberellic acid. (19) This work was supported financially by the National Science Founda- tion. _-13 --14 E. J. Corey,* Rick L. Danheiser Srinivasan Chandrasekaran, Gary E. Keck, B. Gopalan stereomers. Cyanodione 4 is converted via cyano diketal 5 Samuel D. Larsen, Patrice Siret, Jean-Louis Gras (HOCH~CHZOH,p-TsOH, C6H6, reflux; 99%) to diketal acid Department of Chemistry, Harcard Unicersity 6 (KOH, H20, CzHsOH, reflux, 16 h; 90%). Treatment of Cambridge, Massachusetts 021 38 acid 6 with ethyl chloroformate in the presence of triethyl- Receiced September 5, 1978 amine, followed by 3-benzyloxypropylamine (THF, - IO "C; 88%),i3 affords amide 7 which is reduced to secondary amine 8 (LiAIH4, THF, reflux, 16 h; 99%). Treatment of amino diketal8 with HCI in methanol results A Highly Efficient Total Synthesis in slow intramolecular Mannich cyclization (3.2 M HCI, re- of (f)-Lycopodine flux, 14 days), affording a single tricyclic amino ketone (10) in 65% yield. Although compound 8, like compounds 4-7, is Sir: an equimolar mixture of C2 epimers, none of the 12-epi dia- Lycopodine (l), the archetypal Lycopodium alkaloid,' has stereomer (lycopodine numbering) has been found in the re- been known since 1881,2 although its full structure was not action product. This kinetic stereoselectivity was anti~ipated'~ established until 1 960.3 Intensive synthetic work during the and is also observed in cyclization of the analogous N-ben- 1 960s4 resulted in two total syntheses of the alkaloid which zylamine 9, which affords tricyclic amino ketone It, uncon- were communicated in 196gS5An earlier approach resulted in taminated by its diastereomer, under similar (but less strin- the synthesis of the unnatural diastereomer 12-epilycopodine gent) conditions (2.2 equiv of HCI, CH3OH, reflux, 48 h; (2).6A recent communication reports a synthesis of racemic 66%). anhydrolycod~line.~Since natural anhydrolycodoline is hy- Catalytic debenzylation of 10 (H20, CzHsOH, HCI, Hz, drogenated to 2 and 1 in a ratio of 6.5:1,8 this work constitutes Pd; 96%) affords crystalline alcohol 12 (mp 86-87 "C), which a further formal synthesis of lycopodine. We wish to com- undergoes Oppenauer oxidation (benzophenone, t-CdHgOK, municate a highly efficient stereospecific total synthesis of C6H6, reflux, 30 min)I5 with subsequent intramolecular al- lycopodine which is promising for application to the synthesis dolization and dehydration to afford racemic dehydrolyco- of some of the many other members of this important class of podineL6(13, mp 104-105 "C; A,,, 24.5 nm (t 5000)) in 72% alkaloids.' yield. Catalytic of 13 (H2, Pt, C~HSOH)affords Cyanoenone 39 is converted into cyanodione 4 by stereuse- racemic lycopodine (1, mp 130-131 "C (lit.5a mp 130-131 lective trans additionlo of lithium dimethallylcopper (ether, "C)) in 87% yield. The synthetic material produced in this -78 "C; 64%),'l followed by ozonolysis (03, CH30H, -78 manner is identical with a sample of natural lycopodine by "C; 87%), or by conjugate addition of the cuprate derived from infrared and 180-MHz 'H NMR spectroscopy. the lithiated N,N-dimethylhydrazone of acetone, followed by The efficiency of the current synthesis is demonstrated by aqueous hydrolysis ((1) THF, -78 "C, 4 h; (2) Cu2C12, THF, the high overall yield (17.7% from enone 3, 1 1 .l% from dihy- HlO, pH 7, 25 "C, 16 h; 60%).12 Both procedures afford cy- dro~rcinol'~)and by the fact that no other lycopodine diaste- anodione 4 as a separable mixture of C2 epimers, in an ap- reomer may be detected in the final product, even though proximate equimolar ratio. However, we have been unable to isomer separations are not carried out at any point during the detect, at this stage or any subsequent stage, C3-Cj cis dia- synthesis. In one continuous run, we have prepared 1.2 g of 0002-78631781l S00-8036$01.OO/O 0 1978 American Chemical Society TETRAHEDRON LETTERS Pergamon Tetrahedron Letters 43 (2002) 545–548

The synthesis of a key intermediate en route to gelsemine: a program based on intramolecular displacement of the carbonoxygen bond of a strategic oxetane

Fay W. Ng,a Hong Lin,b Qiang Tanb and Samuel J. Danishefskya,b,*

aDepartment of Chemistry, Havemeyer Hall, Columbia University, New York, NY 10027, USA bLaboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue, New York, NY 10021, USA Received 5 November 2001

Abstract—The synthesis of key intermediate 30 en route to gelsemine has been accomplished from known aldehyde 10 via oxetane 19 featuring stereospecific Claisen rearrangement and Lewis acid-catalyzed oxetane ring opening. © 2002 Published by Elsevier Science Ltd.

The appearance of the alkaloid gelsemine (1) (isolated hydroxylation with allylic transposition (en route from from Gelseminium semperverans) in the chemical litera- 6“7) could be achieved. ture goes back to 1870.1 Its structure was arrived at in 1959 through spectroscopic as well as degradative argu- The oxetane moiety in 6 would be fashioned from a ments advanced by Conroy,2a independent of a crystal- C5C16 olefinic linkage by an overall addition of a lographic determination conducted concurrently by ‘’ residue in the proper regiochemical and Lovell and colleagues.2b The keen interest which stereochemical sense at the stage of 4. The logic used gelsemine (1) has attracted from the point of view of for adding this formaldehyde element to the C5C16 total synthesis seems incongruous with the rather olefin in 5 was destined to be the key element of the sketchy and anecdotal suggestions of its potential use- program (vide infra). It was further proposed that the fulness.3 Clearly, it is the novel architecture of gelsem- nucleophilic arm of the projected oxetane displacement ine, which has provoked many interesting strategies reaction (see 6“7) would have been derived from Cur- regarding possible routes for its assembly.4 tius degradation of a suitable two carbon carboxylic acid, mounted at C20 (see structure 6). Any proposal to reach gelsemine (1) must take note of the spiroanilide arising from the quaternary center at We conjectured about the possibility of concurrent C7 (see Scheme 1). Further disconnection of the C7“ presentation of the C20 and the acetic acid N2 and O4“C3 bonds leads back to structure type 7 residues in 6 via some form of a [3,3] rearrangement. (see 7“8“1). Progression from 7“8 requires suprafa- The face selectivity issues in such a transformation cial chirality transfer of a carboxyl equivalent with would be a question for exploration. Assuming this allylic transposition from C14“C7. The underside (a matter could be resolved favorably, the prospect of face) of 7 is quite hindered and prospects for introduc- reaching the allylic alcohol moiety of 5 from a C20 tion of a hydroxymethyl group at C16 by late stage ketone virtually presented itself (see 4“5). joining of a C16C17 bond were not inviting. Our approach to solving this problem called for an Finally, in the retrosynthetic sense, it was hypothesized intramolecular displacement of a properly configured that 4 could have been derived from a divinyl cyclo- oxetane (see sequences 6“7). In this way the required propane“cycloheptadiene rearrangement (3“4). Com- C17 hydroxymethyl group is released as the pyrrolidine pound type 3 might be reached by chain extension (cf. ring is established. It was assumed that overall b o-nitrobenzylidenation) of the aldehyde linkage of sub- strate type 2. Depending on the precise nature of the structure, 2 could be a known compound (vide infra). * Corresponding author. In this and the accompanying paper, the synthesis of

0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd. PII: S0040-4039(01)02212-2 546 F. W. Ng et al. / Tetrahedron Letters 43 (2002) 545–548

Scheme 1. Synthetic plan.

(±)-gelsemine (1) is described. While the proposal pre- hyde 10.6 o-Nitrobenzylidenation7 of this aldehyde, sented above was realized in broad terms, reduction to using phosphonate 11, led to 13 presumably via divinyl- practice involved exposure to many interesting issues in cyclopropane 12. It was envisioned that ketone 15 organic synthesis. would be an attractive type of intermediate to construct the critical oxetane moiety (cf. 5). The synthesis commenced with the epoxidation of 7-t- butoxynorbornadiene (Scheme 2).5 Alumina promoted In principle, 15 could be readily obtained from alcohol rearrangement of epoxide 9 afforded the known alde- 14, which might be reached by hydroboration–oxida-

Scheme 2. Synthesis of the oxetane ring. Reagents and conditions: (a) 11, NaOMe, DMF, 0°C, 74%; (b) BH2Cl·DMS, Et2O, 0°C; 8 NaOH/H2O2, 77%, +7% regioisomer; (c) (COCl)2, DMSO, Et3N, CH2Cl2, 98.7%; (d) LiHMDS, TESCl, Et3N, THF, −78 to 0°C;

Eschenmoser’s salt, CH2Cl2, 91%; (e) MeI, CH2Cl2/Et2O; Al2O3,CH2Cl2, 95%; (f) NaBH4, CeCl3·7H2O, MeOH, 99%; (g) 9-BBN dimer, THF; NaOH/H2O2, 88%; (h) MsCl, Et3N, CH2Cl2, −78°C; NaHMDS, THF, −78°C, 91%. DMS=dimethyl sulfide;  HMDS=hexamethyldisilazane; TESCl=chlorotriethylsilane; Eschenmoser’s salt=(CH3)2N CH2I; 9-BBN=9-borabicyclo[3.3.1]- nonane. F. W. Ng et al. / Tetrahedron Letters 43 (2002) 545–548 547 tion of 13. Such a hydroboration reaction raised signifi- was converted by reduction to its allylic alcohol coun- cant questions of chemoselectivity among the two dou- terpart (23 and 24, respectively).15 These isomers were ble bonds, regioselectivity at the C5C16 olefin and face individually treated with triethylorthoacetate as selectivity. The key issue was that of regiopreference, shown.16 Remarkably, each allylic alcohol gave rise to a even assuming, as we did, that reaction would be single and identical g,d-unsaturated ester 26 (pre- directed to the more strained and more exposed sumably via 25) with the b-vinyl and a-carboxymethyl cyclopentene (C5C16) linkage. Our findings and con- functions at C20 in the required sense.17 This stereo- jectures on this kind of hydroboration as well as related chemical convergence might arise from the tendency of reactions, in a model closely related to 13, have been the enolate like component of the Claisen rearrange- discussed elsewhere.8 In the event, treatment of 13 with ment step to glide over the five-membered ring fused to “ BH2Cl·DMS followed by oxidative workup, as shown, oxetane (see 25 26). Additional cases must be evalu- afforded a 11:1 ratio of alcohol 14 with the newly ated to distinguish between possible steric or electronic introduced alcohol at C5, relative to its isomer where factors in directing the face of the migration step. the alcohol is at C16. Oxidation9 of 14 afforded ketone 15. Regardless of the reasons for this convergence, it pro- vided smooth access to a key intermediate, 26. Alkaline The campaign to install the oxetane commenced with hydrolysis of the ethyl ester function served to release dimethylaminomethylation of the silyl ether the free acid 27.18 Subjection of the latter to Curtius derived from 15.10 Following quaternization of the degradation, as practiced by Shiori, afforded urethane nitrogen, and base induced elimination, the a- 28.19 As anticipated, the hitherto robust oxetane link- methyleneketone 16 was in hand. At this stage we could age, which had survived in the sequence that started take advantage of i-face addition to both sp 2 centers (C5 with 19, was opened by the urethane nitrogen under 20 and C16). Hydride delivery at C5, in the context of a Lewis acid activation (BF3 etherate) and compound Luche reaction,11 afforded 17. Hydroboration of 17 29 was in hand. also occurred from the b-face generating diol 18.12 From this diol intermediate, the a-face oxetane (19)was In summary, we have shown the viability of a synthetic fashioned in a straightforward way as shown. strategy organized around the central idea of using an oxetane linkage to store molecular functionality in a With the critical oxetane in hand, we entered the next compact setting. In this case, the logic was used to phase of the projected plan hoping to reach a func- deliver a highly hindered hydroxymethyl function. Key tional version of allylic alcohol 5. We were anticipating selectivity issues with potentially broader ramifications a [3,3]-type rearrangement en route to structure type 6. in synthesis were resolved favorably in the regioselective Following the cleavage of t-butyl ether 19,13 the result- hydroboration of 13 and in the stereoconvergent rear- ing alcohol function in 20 was oxidized to afford ketone rangements of 23 and 24“26. The progression of 29 to 21 (Scheme 3). Emmons-type condensation14 was suc- gelsemine, requiring responses to some difficult and cessful in terms of overall yield, but led to a 3:2 mixture unanticipated challenges, is described in an accompany- of b,b-disubstituted steroisomers 22. Each compound ing paper.

Scheme 3. Construction of quaternary C7 and the pyrollidine ring. Reagents and conditions: (a) TFA/CH2Cl2,0°C, 81%; (b)

(COCl)2, DMSO, Et3N, CH2Cl2, −78°C, 81%; (c) triethylphosphonoacetate, NaH, THF, 0°C, 3:2, 92%; (d) DIBAL, CH2Cl2,

−78°C, 88%; (e) cat. propionic acid, H3CC(OEt)3, toluene, reflux, 64%; (f) NaOH/THF/EtOH, 86%; (g) diphenylphosphoryl azide,

Et3N, benzene, 25°C, reflux; MeOH, reflux; 89%; (h) BF3·Et2O, CH2Cl2, −78 to 12°C, 64%; (i) PivCl, Et3N, DMAP, CH2Cl2, 0–25°C, 92%. DIBAL=diisobutylaluminum hydride; PivCl=2,2,2-trimethylacetyl chloride; DMAP=N,N-dimethylaminopy- ridine. 548 F. W. Ng et al. / Tetrahedron Letters 43 (2002) 545–548

Acknowledgements 7. Le Corre, M.; Hercouet, A.; Le Stanc, Y.; Le Baron, H. Tetrahedron 1985, 41, 5313–5320. 8. Ng, F. W.; Chiu, P.; Danishefsky, S. J. Tetrahedron Lett. This work was supported by grants from the National 1998, 767–770. Institute of Health (grant HL25848). H.L. would like to 9. Mancuso, A. J.; Brownfain, D. S.; Swern, D. J. Org. thank the Texaco Foundation for a postdoctoral fel- Chem. 1979, 44, 4148–4150. lowship. We thank Dr. George Sukenick and Ms. Sylvi 10. (a) Kleinman, E. F. In Comprehensive Organic Synthesis; Rusli of the MSKCC NMR Core Facility for NMR Trost, B. M.; Fleming, I., Eds.; Press: New York, USA, and MS spectral analyses (NIH Grant CA08748). 1991; Vol. 2, p. 899; (b) Danishefsky, S.; Kitahara, T.; Schuda, P. F.; Etheredge, S. J. J. Am. Chem. Soc. 1976, 98, 3028–3030. 11. Luche, J.-L.; Gemal, A. L. J. Chem. Soc., Chem. Comm. References 1978, 976–977. 12. It was unlikely for the hydroboration to occur on the 1. Saxton, J. E. In The Alkaloids; Manske, R. H. F., Ed.; endo-face because the free hydroxyl group reacted with Academic Press: New York, 1965; Vol. 8, pp. 93–117. the boron reagent to give the boronate ester, which is a 2. (a) Conroy, H.; Chakrabarti, J. K. Tetrahedron Lett. poor chelating partner, see: Hoveyda, A. H.; Evans, D. 1959,6–13; (b) Lovell, F. M.; Pepinsky, R.; Wilson, A. J. A.; Fu, G. C. Chem. Rev. 1993, 93, 1307–1370. C. Tetrahedron Lett. 1959,1–5. 13. Beyerman, H. C.; Heiszwolf, G. L. J. Chem. Soc. 1963, 3. Liu, Z.-J.; Lu, R.-R. In The Alkaloids; Manske, R. H. F., 755–756. Ed.; Academic Press: New York, 1988; Vol. 33, pp. 14. Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 83–140. 863–927. 4. (a) Newcombe, N. J.; Ya, F.; Vijn, R. J.; Hiemstra, H.; 15. Winterfeldt, E. Synthesis 1975, 617–630. Speckamp, W. N. J. Chem. Soc., Chem. Commun. 1994, 16. Johnson, W. S.; Werthemann, L.; Bartlett, W. R.; 767–768; (b) Sheikh, Z.; Steel, R. W.; Tasker, A. S.; Brosksom, T. J.; Li, T.-T.; Faulkner, D. J.; Perterson, M. Johnson, A. P. J. Chem. Soc., Chem. Commun. 1994, R. J. Am. Chem. Soc. 1970, 92, 741–743. For an overview 763–764; (c) Dutton, J. K.; Steel, R. W.; Tasker, A. S.; of Claisen reaarangement see: (a) Ziegler, F. E. Chem. Popsavin, V.; Johnson, A. P. J. Chem. Soc., Chem. Rev. 1988, 88, 1423–1452; (b) Ziegler, F. E. Acc. Chem. Commun. 1994, 765–766; (d) Atarashi, S.; Choi, J.-K.; Res. 1977, 10, 227–232. Ha, D.-C.; Hart, D. J.; Kuzmich, D.; Lee, C.-S.; Ramesh, 17. After this discovery, the 3:2 mixture of allylic alcohols 23 S.; Wu, S. C. J. Am. Chem. Soc. 1997, 119, 6226–6241; (e) and 24 were subjected to the above variant Claisen Kuzmich, D.; Wu, S. C.; Ha, D.-C.; Lee, C.-S.; Ramesh, rearrangement conditions without further separation of S.; Atarashi, S.; Choi, J.-K.; Hart, D. J. J. Am. Chem. the two isomers. Soc. 1994, 116, 6943–6944; (f) Fukuyama, T.; Liu, G. J. 18. Honda, M.; Hirata, K.; Sueoka, H.; Katsuki, T.; Am. Chem. Soc. 1996, 118, 7426–7427; (g) Madin, A.; Yamaguchi, M. Tetrahedron Lett. 1981, 22, 2679–2682. O’Donnell, C. J.; Oh, T.; Old, D. W.; Overman, L. E.; 19. (a) Shiori, T.; Ninomiya, K.; Yamada, S. J. Am. Chem. Sharp, M. J. Angew. Chem., Int. Ed. Engl. 1999, 38, Soc. 1972, 94, 6203–6205; (b) Ninomiya, K.; Shiori, T.; 2934–2936. Yamada, S. Tetrahedron 1974, 30, 2151–2152. 5. 7-t-Butoxynorbornadiene is commercially available. It 20. Boron trifluoride was the superior Lewis acid for the can be prepared by the method of Story: Story, P. R. J. activation of oxetane derivatives with intramolecular Org. Chem. 1961, 26, 287–290. etheral oxygen participation. See: Itoh, A.; Hirose, Y.; 6. (a) Klumpp, G. W.; Barnick, J. W. F. K.; Veefkind, A. Kashiwagi, H.; Masaki, Y. Heterocycles 1994, 38, 2165– H.; Bickelhaupt, F. Recl. Trav. Chim. Pays-Bas 1969, 88, 2169. Boron trifluoride catalyzed the nucleophilic ring 766–778; (b) Cupas, C.; Watts, W. E.; Schleyer, P.; von, opening of oxetane derivatives, see: Xianming, H.; Kel- R. Tetrahedron Lett. 1964, 5, 2503–2507. logg, R. M. Tetrahedron: Asymmetry 1995, 6, 1399–1408. TETRAHEDRON LETTERS Pergamon Tetrahedron Letters 43 (2002) 549–551

The synthesis of (±)-gelsemine

Hong Lin,a Fay W. Ngb and Samuel J. Danishefskya,b,*

aLaboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue, New York, NY 10021, USA bDepartment of Chemistry, Havemeyer Hall, Columbia University, New York, NY 10027, USA Received 5 November 2001

Abstract—The synthesis of (±)-gelsemine has been completed from tetracyclic intermediate 2 via a stereospecific [3,3]-rearrange- ment followed by a one carbon excision to convert a d-lactam (13)toag-lactam (19). © 2002 Published by Elsevier Science Ltd.

In the preceding paper1 we reported on a synthesis of tion of the nitro group afforded the amine, which was compound 2 in furtherence of a proposed total synthe- protected with CbzCl to provide 8. Deprotection of the sis of gelsemine (1).2 Before describing how gelsemine acetate afforded 9. was reached via 2, we must place our study in context. It was possible in several settings to accomplish overall Major efforts were directed to the goal of using the C14 isomerization of the non conjugated D3(14) double bond b-OH group to introduce a b-one carbon fragment at to the conjugated D3(7) series, in which the aromatic C7 by suprafacial allylic transposition (cf. 10“11, amino function is also presented (see Scheme 1, struc- Scheme 3). Such attempts were uniformly unsuccessful. ture type 3). With this capability, a number of possibil- A particularly disappointing case is seen in the high ities for introduction of a one carbon residue between C7 and the anilino nitrogen group were surveyed (cf. 3“4, with or without concurrent participation from the primary hydroxy function). These attempts were uni- formly unsuccessful. Instead, those reactions that could be achieved, were initiated by attack of a one carbon moiety at C3 (in a Markovnikov sense) rather at C7 as would be required for spirocyclization to a 5-membered anilide.3

A key element in the overall isomerization (ie. conjuga- tion) of the double bond to the D3(7) series was allylic bromination of a D3(14) double bond isomer to intro- duce a b-bromine at C14 with reappearance of the double bond at the D3(7) position.4 This reaction was now conducted on pivaloate 2, thereby affording 5 (Scheme 2).5 Debromination of 5 (tri-n-butyltin hydride) in the presence of O2, followed by the reduc- tion of the resulting hydroperoxide with sodium boro- hydride afforded 6 with high stereoselectivity.6 Alternatively, acetolysis of 5 was accomplished, with in acetic acid, to provide 7.7 The overall stereochemical retention result in this reaction attests to the highly hindered nature of the a-face of the ‘ball- like’ surface of 5, and species derived therefrom. Reduc-

* Corresponding author. Scheme 1.

0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd. PII: S0040-4039(01)02213-4 550 H. Lin et al. / Tetrahedron Letters 43 (2002) 549–551

w Scheme 2. Reagents and conditions: (a) NBS, AIBN, h , CCl4/CH2Cl2,reflux, 60% based on recovered starting material; (b) w AIBN, Bu3SnH, dry air, h , toluene, 60°C; NaBH4,0°C; 55% based on recovered starting material; (c) AgOAc, HOAc, 52%; (d) zinc dust, THF/HOAc; (e) CbzCl, NaHCO3 (aq.), CH2Cl2, 94% for two steps; (f) K2CO3, MeOH, 90%. NBS=N-bromosuccin- imide, AIBN=2,2%-azobisisobutyronitrile, CbzCl=benzyl chloroformate.

Scheme 3. Reagents and conditions: (a) CH3C(OMe)2NMe2, m-xylene, silica gel purification, 30–40%; (b) NaOMe, MeOH, 74%. yielding transformation of 10“128 by apparent [1,2] which the extremely hindered free hydoxymethyl group transposition,9 in the context of a projected Still–Wittig on a-face (initially derived by intramolecular oxetane rearrangement.10 Apparently, formidable steric forces opening)1 was now poised to close the tetrahydropyran are arrayed against carboncarbon bond formation ring. even on the b-face of C7, and even by intramolecular means. Fortunately, it was found that the Eschenmoser Oxymercruation of 20 with Hg(OTf)2 N,N-dimethyl- 13,14 amide acetal version of the Claisen rearrangement took aniline complex in CH3NO2 afforded the desired place in the desired [3,3] sense.11 Subjection of 9 to the mercuric cyclization product which, following reductive conditions shown, led to 13 and, following deprotection demercuration (using Fukuyama’s protocol),15 fur- of the pivaloate group, 14.12 nished hydropyran 21 in 60% yield. Hydrolysis of the Cbz protected oxindole in 21,16 with 10% of NaOH in We now faced the challenging prospect of shrinking the THF afforded a 90% yield of 22. Finally, the methyl 6-membered lactam to a 5-membered spiroanilide for carbamate of 22 was reduced to an N- 17 purposes of reaching gelsemine (1). The sequence to with LiAlH4, thereby affording 1 whose spectroscopic accomplish this goal began with reduction of the imide- and chromatographic properties matched those of natu- like functionality of 13 to afford aminal 15 (Scheme 4). rally derived gelsemine. Dehydration of this aminal, as shown, furnished enam- ide 16 (50% yield over two steps). Dihydroxylation of In summary, we had set out to explore some novel 16, across the more electron-rich enamide double bond, synthetic constructions using the synthesis of gelsemine provided a trihydroxy intermediate, which was sub- (1) as an orienting, clearly defined, goal. Many interest- jected to oxidative cleavage, as shown. This degrada- ing issues of selectivity, both at regiochemical and tion provided a 45% yield of 17, containing an stereochemical levels, were resolved in favorable all-important b-face aldehyde at C7. Protection of the ways.1,18 Our findings as to reaction specificities in of 17 led to silyl ether 18. Methanolysis subtle cases merit continuing study. While confident of this compound served to accomplish N-deformyla- application of these findings to new cases would require tion and, concurrently, ring closure to a cyclic hemi- broadening of our database as to scope and limitations, aminal. The latter, following oxidation, gave rise to the finding reported here as part of the synthesis, oxindole 19. Desilylation of 19, as shown, led to 20 in already invite potentially important interpretations. H. Lin et al. / Tetrahedron Letters 43 (2002) 549–551 551

Scheme 4. Reagents and conditions: (a) DIBAL, CH2Cl2, −78°C; (b) TsOH·H2O, CH2Cl2,reflux, 50% for two steps; (c) OsO4, THF, −25°C; Na2SO3 (aq.); (d) NaIO4, THF/H2O, 45%; (e) TESOTf, Et3N, CH2Cl2,0°C, 80%; (f) NaOMe, MeOH; (g) TPAP, , NMO, CH2Cl2,4AMS, 50% for two steps; (h) TBAF/HOAc 1:1, THF, 80%; (i) Hg(OTf)2·C6H5NMe2,CH3NO2; NaBH4, 10%

NaOH, Et3BnNCl, CH2Cl2, 60%; (j) 10% NaOH, THF, 90%; (k) LiAlH4, THF, 0–25°C, 81%. DIBAL=diisobutylaluminum hydride, TESOTf=triethylsilyltrifluoromethanesulfonate, TPAP=tetrapropylammonium perruthenate, NMO=N-methylmorpho- line N-oxide. Unfortunately, the focused gelsemine target goal 7. Cf: Winstein, S.; Buckles, R. E. J. Am. Chem. Soc. 1942, became quite complicated in that its solution required 64, 2787–2790. excision of a one carbon unit from a six-membered 8. Ng, F. W. Ph.D. Thesis, Columbia University, 1997, 54. lactam to a five-membered spiroanilide (see 13“19). In The stereochemistry at C14 of the rearrangement product the end, this ring contraction was accomplished. A full (12) was not determined. account of these experiments and other interesting 9. Cf: Still, C. W.; Mitra, A. J. Am. Chem. Soc. 1978, 100, excursions directed to gelsemine (1) is planned. 1927. 10. For an overview of [2,3]-Wittig rearrangements see: (a) Mikami, K.; Nakai, T. Org. React. 1994, 46, 105; (b) Nakai, Acknowledgements T.; Mikami, K. Chem. Rev. 1986, 86, 885–902. 11. Felix, D.; Gschwend-Steen, K.; Wick, A. E.; Eschenmoser, This work was supported by grants from the National A. Helv. Chim. Acta 1969, 52, 1030–1042. Institute of Health (grant HL25848). H.L. would like to 12. Compound 14 was prepared to clarify the stereo outcome thank the Texaco Foundation for a postdoctoral fel- of the Claisen rearrangement as shown in Scheme 3. The lowship. We thank Dr. George Sukenick and Ms. Sylvi most conclusive signal was the NOE between one of the Rusli of the MSKCC NMR Core Facility for NMR a-protons in the lactam and H19 (the methine proton of and MS spectral analyses (NIH Grant CA08748). the terminal ). 13. Nishizawa, M.; Takenaka, H.; Hayashi, Y. J. Org. Chem. 1986, 51, 806–813. References 14. Newcombe, N. J.; Ya, F.; Vijn, R. J.; Hiemstra, H.; Speckamp, W. N. J. Chem. Soc., Chem. Commun. 1994, 1. Ng, F. W.; Lin, H.; Tan, Q.; Danishefsky, S. J. Tetrahedron 767–768. Lett. 2002, 43, 545–548. 15. Fukuyama, T.; Liu, G. J. Am. Chem. Soc. 1996, 118, 2. Saxton, J. E. In The Alkaloids; Manske, R. H. F., Ed.; 7426–7427. Academic Press: New York, 1965; Vol. 8, pp. 93–117. 16. Reductive demercuration was capricious when conducted 3. For specific examples of this effect see: Ng, F.W. Ph.D. Thesis, Columbia Unversity, 1997, 70, and 74. on small scale. It is currently under optimization. Hydroly- 4. For free benzylic bromination of substituted nitro- sis of 21 was performed on the material that was degraded toluene derivatives with NBS, see: Mataka, S.; Kurisu, M.; from commercially available gelsemine (1) in three steps: Takahashi, K.; Tashiro, M. Chem. Lett. 1984, 1969–1972. demethylation of 1 with PhOCOCl and Hu¨nig’s base 5. Following precedents, methylene chloride was used to provided the phenyl carbamate, which was converted to minimize double bromination, see: Offermann, W.; Vogtle, methyl carbamate (which is identical to 22) with NaOMe F. Angew. Chem., Int. Ed. Engl. 1980, 19, 464–465. in MeOH at reflux. Protection of the free oxindole with 6. Nakamura, E.; Inubushi, T.; Aoki, S.; Machii, D. J. Am. CbzCl, Et3N and DMAP in CH2Cl2 provided 21 whose Chem. Soc. 1991, 113, 8980–8982. For additional examples spectroscopic and chromatographic properties were identi- of free radical initiated oxygenation, see: Mayer, S.; Prandi, cal to those of its synthetic counterpart. J. Tetrahedron Lett. 1996, 37, 3117–3120; Moutel, S.; 17. Confalone, P. N.; Huie, E. M. J. Org. Chem. 1985, 52,79–83. Prandi, J. Tetrahedron Lett. 1994, 35, 8163–8166. The yield 18. Ng, F. W.; Chiu, P.; Danishefsky, S. J. Tetrahedron Lett. was based on the recovered starting material. 1998, 39, 767–770. Published on Web 12/11/2002

Total Synthesis of Gambierol

Isao Kadota,† Hiroyoshi Takamura,‡ Kumi Sato,‡ Akio Ohno,‡ Kumiko Matsuda,‡ and Yoshinori Yamamoto*,‡

Research Center for Sustainable Materials Engineering, Institute of Multidisciplinary Research for AdVanced Materials, and Department of Chemistry, Graduate School of Science, Tohoku UniVersity, Sendai 980-8578, Japan

Received September 27, 2002 ; E-mail: [email protected]

In recent years, there has been an explosion of interest in biologically active natural products of marine origin.1 Because of their structural novelty and toxicity, polycyclic are particularly attractive targets for synthetic chemists.2 Gambierol (1), a potent neurotoxin isolated from the cultured cells of Gambierdiscus toxicus, has 8 ether rings and 18 stereogenic centers.3 The compound shows toxicity against mice (LD50 50 µg/kg), and the symptoms resemble those caused by ciguatoxins, inferring the possibility that it is also implicated in ciguatera poisoning.1 The unique structural features have attracted the attention of synthetic chemists, and a number of strategies have been investigated.4,5 In this paper, we describe a new approach to the total synthesis of gambierol (1). Figure 1 illustrates our retrosynthetic analysis of 1. Recently, we developed a convergent method for the synthesis of polycyclic Figure 1. Retrosynthetic analysis of gambierol (1). ether frameworks via the intramolecular allylation of R-acetoxy ethers and subsequent ring-closing metathesis.4i,l On the basis of this methodology, the octacyclic ether framework of 1 was as a single stereoisomer. TBS protection and subsequent selective retrosynthetically broken down into the ABC ring segment 3 and deprotection of the primary silyloxy group afforded 15 in 82% the FGH fragment 4 via the diene 2. overall yield. The initial task of the total synthesis was the construction of the We have already succeeded in the stereoselective synthesis of octacyclic framework. The carboxylic acid 3 and the alcohol 46 the triene moiety via the Uenishi hydrogenolysis of dibromoalkene 4c were connected by Yamaguchi conditions7 to give the ester 5 in followed by a modified Stille coupling in a model study. However, 94% yield (Scheme 1). A series of reactions including desilylation the coupling reaction of Z-bromoalkenes was very slow. This with TBAF, acid-catalyzed acetal formation with 6, and acetal problem prompted us to develop an efficient method for the cleavage with TMSI/HMDS furnished the allylic stannane 7 in 77% construction of the triene side chain. After several unfruitful - overall yield.8 The ester 7 was then subjected to the modified attempts, we found that the reduction of diiodoalkenes with a Zn V V V Rychnovsky . Thus, partial reduction of 7 with DIBALH, Cu couple ga e reacti e Z-iodoalkenes in a highly stereoselecti e 15 followed by treatment of the resulting aluminum hemiacetal with manner. Thus, PCC oxidation of 15 followed by treatment of the 16 9 resulting aldehyde with CI4 and PPh3 gave the diiodoalkene 16 (CH2ClCO)2O/pyridine/DMAP, gave the R-chloroacetoxy ether 8 as a 3:2 inseparable mixture of diastereoisomers in 88% yield. in 92% overall yield. Stereoselective hydrogenolysis of 16 was carried out by using a Zn-Cu couple and AcOH in THF-MeOH Treatment of 8 with BF3‚OEt2 gave a 2:1 mixture of the desired product 2 and its epimer 9 in 87% yield.10,11 The diene 2 obtained to give the Z-iodoalkene 17 as a single stereoisomer, which was was subjected to ring-closing metathesis using the second generation then treated with DIBALH, furnishing 18 in quantitative yield. Grubbs catalyst 1012 to give the octacyclic ether 11 in 88% yield. Deprotection of the bis-silyl ether 18 with SiF4 proceeded smoothly 17 We then focused on the modification of the H ring moiety. to afford the corresponding triol. Finally, the iodoalkene obtained Hydrolysis of the benzylidene acetal of 11 led to the corresponding was subjected to the modified Stille coupling with 19 to give 18 diol. Selective protection of the primary alcohol followed by TPAP gambierol (1) in 72% yield. The synthetic gambierol exhibited oxidation of the secondary alcohol gave the ketone 12 in 77% physical and spectroscopic data identical to those reported previ- 3,5 overall yield. Hydrogenation of 12 followed by debenzylation gave ously. the saturated diol. The primary and secondary alcohols were In conclusion, the convergent total synthesis of gambierol has R protected by Pv and TIPS groups, respectively, to afford 13 in 80% been achieved using the intramolecular allylation of -chloroacetoxy 5 ether and subsequent ring-closing metathesis. The longest linear overall yield. Treatment of 13 with LiHMDS/TMSCl/Et3N gave the corresponding enol silyl ether, which was subjected to de- sequence leading to 1 was 66 steps with 1.2% overall yield, and 13 the total number of steps was 102, while the synthesis by Sasaki hydrosilylation with Pd(OAc)2 to afford the enone 14 in 92% overall yield. Stereoselective introduction of the methyl group was and Tachibana gave an overall yield of 0.57% by 71 steps, with a 5 carried out using MeMgI in toluene14 to give the tertiary alcohol total of 107 steps. The present synthesis demonstrated that this methodology is efficient and practical for constructing polycyclic † Institute of Multidisciplinary Research for Advanced Materials. ether frameworks. Application of the present strategy to the ‡ Graduate School of Science. synthesis of other marine natural products is in progress.

46 9 J. AM. CHEM. SOC. 2003, 125,46-47 10.1021/ja028726d CCC: $25.00 © 2003 American Chemical Society COMMUNICATIONS

Scheme 1 a

a (a) 2,4,6-Trichlorobenzoyl chloride, Et3N, THF, 40 °C, then 4, DMAP, toluene, 40 °C, 94%; (b) TBAF, THF, room temperature, 99%; (c) 6, CSA, CH2Cl2, room temperature; (d) HMDS, TMSI, CH2Cl2,0°C, 78% (two steps); (e) DIBALH, CH2Cl2, -78 °C, then (CH2ClCO)2O, DMAP, pyridine, -78 °C to room temperature, 88%; (f) BF3‚OEt2, MS4A, CH3CN, -40 to 0 °C, 87% (2:9 ) 2:1); (g) 10,CH2Cl2, room temperature, 88%; (h) CSA, CH2Cl2- MeOH, 30 °C; (i) TBSCl, imidazole, CH2Cl2,0°C, 80% (two steps); (j) TPAP, NMO, MS4A, CH2Cl2, room temperature, 96%; (k) H2,Pd-C, EtOAc, room temperature; (l) H2, Pd(OH)2-C, EtOAc, room temperature; (m) PvCl, DMAP, CH2Cl2, room temperature; (n) TIPSOTf, 2,6-lutidine, DMF, 65 °C, 80% (four steps); (o) LiHMDS, TMSCl, Et3N, THF, -78 °C; (p) Pd(OAc)2,CH3CN, 92% (two steps); (q) MeMgI, toluene, -78 °C; (r) TBSOTf, 2,6-lutidine, CH2Cl2, room temperature; (s) CSA, CH2Cl2-MeOH, 0 °C, 82% (three steps); (t) PCC, MS4A, CH2Cl2, room temperature; (u) CI4, PPh3,CH2Cl2,0°C, 92% (two steps); (v) Zn-Cu, AcOH, THF-MeOH, 0 °C; (w) DIBALH, CH2Cl2, -78 °C, 100% (two steps); (x) SiF4,CH2Cl2-CH3CN, 0 °C; (y) 19, Pd2dba3‚CHCl3, P(furyl)3, CuI, DMSO, 40 °C, 72% (two steps).

Acknowledgment. We thank Professor T. Yasumoto (Tohoku (5) During our study, very recently the Sasaki and Tachibana group reported 1 the first total synthesis of 1. (a) Fuwa, H.; Sasaki, M.; Satake, M.; University) for providing the H NMR spectrum of 1, and Professor Tachibana, K. Org. Lett. 2002, 4, 2981-2984. (b) Fuwa, H.; Kainuma, M. Hirama (Tohoku University) for supporting the measurement N.; Tachibana, K.; Sasaki, M. J. Am. Chem. Soc., in press. of MALDI TOF mass spectra. This work was financially supported (6) For the preparation of 3 and 4, see Supporting Information. (7) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull. Chem. by the Grant-in-Aid for Scientific Research from the Ministry of Soc. Jpn. 1979, 52, 1989-1993. Education, Culture, Sports, Science, and Technology, Japan. (8) Kadota, I.; Sakaihara, T.; Yamamoto, Y. Tetrahedron Lett. 1996, 37, 3195-3198. Supporting Information Available: Schemes for the preparation (9) Acetic anhydride was used in the original procedure, see: (a) Dahanukar, V. H.; Rychnovsky, S. D. J. Org. Chem. 1996, 61, 8317-8320. (b) of compounds 3 and 4. Experimental procedures and characterization Rychnovsky, S. D.; Hu, Y.; Ellsworth, B. Tetrahedron Lett. 1998, 39, data for all new compounds reported in Scheme 1. Copies of 1H NMR 7271-7274. (c) Kopecky, D. J.; Rychnovsky, S. D. J. Org. Chem. 2000, 65, 191-198. (d) Rychnovsky, S. D.; Thomas, C. R. Org. Lett. 2000, 2, spectra for selected compounds (PDF). This material is available free 1217-1219. (e) Jaber, J. J.; Mitsui, K.; Rychnovsky, S. D. J. Org. Chem. of charge via the Internet at http://pubs.acs.org. 2001, 66, 4679-4686. (10) The reaction of the corresponding R-acetoxy ether gave the undesired isomer 9, predominantly. Perhaps, the poor leaving ability of , References in comparison with chloroacetoxy group, would force the reaction course to proceed through the SN2 pathway. The D ring of the desired isomer 2, (1) For recent reviews, see: (a) Shimizu, Y. Chem. ReV. 1993, 93, 1685- in which all of the substituents take an equatorial position, is more 1698. (b) Yasumoto, T.; Murata, M. Chem. ReV. 1993, 93, 1897-1909. thermally stable than that of 9. The details of the mechanism are under (2) For recent reviews, see: (a) Alvarez, E.; Candenas, M.-L.; Pe´rez, R.; investigation. We appreciate a referee who pointed out the above problems, Ravelo, J. L.; Martı´n,J.D.Chem. ReV. 1995, 95, 1953-1980. (b) and we also thank Professor T. Oishi (Osaka University) and Professor Nicolaou, K. C. Angew. Chem., Int. Ed. Engl. 1996, 35, 589-607. (c) M. Inoue (Tohoku University) for suggesting the use of the chloroacetyl Mori, Y. Chem.-Eur. J. 1997, 3, 849-852. Also see: (d) Hoberg, J. O. group. Tetrahedron 1998, 54, 12630-12670. (11) For the determination of the stereochemistries of 2 and 9, see Supporting Information. (3) (a) Satake, M.; Murata, M.; Yasumoto, T. J. Am. Chem. Soc. 1993, 115, - 361-362. (b) Morohashi, A.; Satake, M.; Yasumoto, T. Tetrahedron Lett. (12) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1, 953 1998, 39,97-100. 956. (13) Ito, Y.; Hirao, T.; Saegusa, T. J. Org. Chem. 1978, 43, 1011-1013. (4) (a) Kadota, I.; Park, C.-H.; Ohtaka, M.; Oguro, N.; Yamamoto, Y. - Tetrahedron Lett. 1998, 39, 6365-6368. (b) Kadota, I.; Kadowaki, C.; (14) Feng, F.; Murai, A. Chem. Lett. 1992, 1587 1590. Yoshida, N.; Yamamoto, Y. Tetrahedron Lett. 1998, 39, 6369-6372. (c) (15) The palladium-catalyzed hydrogenolysis of diiodoalkenes with Bu3SnH was not selective, see: Uenishi, J.; Kawahama, R.; Yonemitsu, O.; Tsuji, Kadota, I.; Ohno, A.; Matsukawa, Y.; Yamamoto, Y. Tetrahedron Lett. - 1998, 39, 6373-6376. (d) Fuwa, H.; Sasaki, M.; Tachibana, K. Tetra- J. J. Org. Chem. 1998, 63, 8965 8975. hedron Lett. 2000, 41, 8371-8375. (e) Sakamoto, Y.; Matsuo, G.; (16) Gavina, F.; Luis, S. V.; Ferrer, P.; Costero, A. M.; Marco, J. A. J. Chem. Matukura, H.; Nakata, T. Org. Lett. 2001, 3, 2749-2752. (f) Cox, J. M.; Res. 1986, 330-331. Rainier, J. D. Org. Lett. 2001, 3, 2919-2922. (g) Fuwa, H.; Sasaki, M.; (17) Corey, E. J.; Yi, K. Y. Tetrahedron Lett. 1992, 33, 2289-2290. The other Tachibana, K. Tetrahedron 2001, 57, 3019-3033. (h) Fuwa, H.; Sasaki, attempts for desilylating the protected gambierol having the triene side M.; Tachibana, K. Org. Lett. 2001, 3, 3549-3552. (i) Kadota, I.; Ohno, chain, for example, the use of TBAF, TAS-F, and SiF4, resulted in failure A.; Matsuda, K.; Yamamoto, Y. J. Am. Chem. Soc. 2001, 123, 6702- as reported by Sasaki, see ref 5. 6703. (j) Kadota, I.; Kadowaki, C.; Park, C.-H.; Takamura, H.; Sato, K.; (18) Fuji Silysia silica gel was used for the purification of 1. Although the Chan, P. W. H.; Thorand, S.; Yamamoto, Y. Tetrahedron 2002, 58, 1799- detail is not clear, the use of Kanto Chemical silica gel (neutral) caused 1816. (k) Kadota, I.; Takamura, H.; Sato, K.; Yamamoto, Y. J. Org. Chem. partial isomerization of 1. See Supporting Information. 2002, 67, 3494-3498. (l) Kadota, I.; Ohno, A.; Matsuda, K.; Yamamoto, Y. J. Am. Chem. Soc. 2002, 124, 3562-3566. JA028726D

J. AM. CHEM. SOC. 9 VOL. 125, NO. 1, 2003 47 J. Am. Chem. Soc. 1988, 110, 649-651 649 to 3.10 (3) A. The Ag-C bond lengths for silver(1) alkene and parameters, semirigid body TLS refinement details, bond distances arene complexes'* range from 2.31 to 2.84 A. The following and angles, and least-squares planes analysis for the crystal bonding scheme is proposed for the eclipsed silver bis(TBC) unit. structure of 2[Ag(TBC)2(OTF)].2TBCC6H14and a general la- Four of the alkynes behave as two electron donors, while two beling diagram for all six TBC molecules (25 pages); table of alkynes are noninteracting giving an 18-electron count. The observed and calculated structure factors (32 pages). Ordering asymmetric nature of the metal-teligand interaction in the eclipsed information is given on any current masthead page. conformtion is reminiscent of the bonding in ruthenium and rhodium bis-arene comple~es.'~ In contrast, the staggered sandwich has a much smaller range of Ag-C distances for all Total Synthesis of (f)-Ginkgolide B silver-acetylene interactions leading to the conclusion that the six alkynes are donating 1.333 electrons each or that this is an E. J. Corey,* Myung-chol Kang, Manoj C. Desai, electrostatic interaction. The hardsoft acid-base theorym provides Arun K. Ghosh, and Ioannis N. Houpis an explanation for the preference of the soft silver cation for the Department of Chemistry, Harvard University soft alkyne carbons instead of the hard triflate oxygen. Cambridge, Massachusetts 021 38 The silver sandwiches form infinite chains parallel to the 2' axis Received October 5, 1987 stacked metal over metal alternating eclipsed and staggered sandwiches. The cocrystallization of staggered and eclipsed Extracts of the ginkgo tree, Ginkgo biloba, now widely rec- conformations of sandwich complexes is unusual but not un- ommended in Asian and European medicine (annual sales ca. precedented. The cocrystallization of eclipsed and staggered $500 000 000 per annum), have been found to antagonize platelet conformations of bis( 1,3,5,7-tetramethylcyclooctatetraene)uranium activating factor (PAF),' a very fundamental mammalian regu- has been reported.21 The shortest unique silversilver vectors in lator.2 Ginkgo extracts increasingly find therapeutic use in the the unit cell are 7.102 (3) and 7.115 (3) A and are approximately treatment of cerebrovascular and peripheral circulatory problems parallel to the b-axis. The five unique interplanar spacings for of the elderly and asthma. The most active anti-PAF agent in the least-squares plane defined b each TBC bound to silver range the ginkgo extract is the hexacyclic C20trilactone ginkgolide B between 3.45 (1) and 3.58 (1) (Tables XVI-XXI in the Sup- (1)3(ICso 10-7-10-s M in various tests),' which appears to an- plementary Material). All dihedralK angles between coordinated tagonize all known PAF-induced membrane events. The first total TBC planes are less than 5'. The chains of [Ag(TBC),]+ are synthesis of ginkgolide B (racemic form) is described herein. A surrounded by a tube of free TBC molecules whose planes are recent paper from these laboratories4 has reported the total nearly perpendicular to the planes of the sandwiches and to each synthesis of the related CISginkgolide, (f)-bil~balide,~by a totally other (Figure 2). The triflate anions do not directly interact with different approach. the silver ions. The closest silver-(triflate atom) distance is greater Reaction of 1-morpholinocyclopentene with dimethoxyacet- than 6.8 A. The triflate anions show considerable thermal motion aldehyde in benzene at 23 OCfor 18 h, stirring of the resulting and have been refined as semirigid bodies with TLS motion.22 The solution with 6 N hydrochloric acid at 0 "C for 30 min, extractive hexane solvent molecule is disordered and modeled as two mol- isolation and distillation (145-146 OC at 15 Torr) provided enone ecules of six carbons each at half occupancy. 2 in 70% Enone 2 was converted into the enol silyl ether Studies of the reaction chemistry of 1 including the intercon- 3 (93% yield) by reaction in ether with the cuprate reagent t- version of the staggered and eclipsed conformations and the Bu2CuCNLi2(1.5 equiv relative to 2; prepared from reaction of synthesis of other transition-metal sandwich complexes of TBC cuprous cyanide and tert-butyllithium in a 1:2 ratio at -78 OC are currently under investigation in our laboratory. for 50 min and then at -45 OC for 30 min) at -78 OC for 10 min and then at -45 "C for 30 min, followed by silylation of the Acknowledgment. J.D.F. thanks the BF Goodrich Company resulting enolate with 5 equiv each of trimethylchlorosilane and for financial support through the BF Goodrich Fellowship at triethylamine (-45 "C for 45 min, then -10 "C for 5 min) and CWRU. A.D. thanks the Algerian Government for support. extractive isolation. Addition of 3 in methylene chloride to a Acknowledgment is made to the donors of The Petroleum Re- solution of 1,3,5-trioxane (1.2 mol equiv) and titanium tetra- search Fund administered by the American Chemical Society for chloride (3.6 equiv) in methylene chloride at -78 "C (over 20 min), support of this work. further reaction (-78 OCfor 2 h and -45 "C for 1 h), and finally treatment with one-half volume of methanol (0 "C initially then Supplementary Material Available: Tables of crystal data, data 23 "C for 12 h) produced stereosele~tively~cyclopentanone 4, mp collection reduction, and refinement details, positional and thermal 25-27 OC, as a 2:l mixture of two C(11) anomeric methyl acetals (ginkgolide numbering) in 65% yield. Deprotonation of 4 with 1.25 equiv of lithium diisopropylamide (LDA) in dimethoxyethane (18)(a) Kang, H. C.; Hanson, A. W.; Eaton, B.; Boekelheide, V. J. Am. (-78 OC for 1 h, 0 OC for 20 min) and subsequent reaction with Chem. SOC.1985,107, 1979-1985. (b) Mak, T. C. W.; Ho, W. C.; Huang, N. Z. J. Orgammet. Chem. 1983, 251, 413-421. (c) Albinati, A.; Meille, S. N-phenyltriflimideIo (0 OCfor 1.5 h, 23 OC for 1 h) afforded after V.; Carturan, G. J. Organomet. Chem. 1979,182,269-274.(d) Lewandos, G. S.; Gregston, D. K.; Nelson, F. R. J. Organomet. Chem. 1976, 118, (1) (a) Braquet, P. Drugs ofthe Future 1987, 12,643-699.(b) Braquet, 363-374. (e) Ermer, 0.;Eser, H.; Dunitz, J. D. Helu. Chim. Acta 1971,54, P.; Godfroid, J. J. Trends Pharmacol. Sci. 1986, 7, 397-403. (c) Barnes, P. 2469-2475. (f) Rodesiler, P. F.; Amma, E. L. Inorg. Chem. 1972, 11, J.; Chung, K. F. Ibid. 1987,8,285-286. (d) Max, B. Ibid. 1987,8,29+292. 388-395. (9) Griffith, E. A. H.; Amma, E. L. J. Am. Chem. SOC.1971, 93, (2) Hanahan, D. J. Ann. Reu. Eiochem. 1986, 55, 483-509. 3167-3172. (h) Taylor, I. F.; Hall, E. A,; Amma, E. L. J. Am. Chem. SOC. (3) (a) Sakabe, N.; Takada, S.; Okabe., K. J. Chem. SOC.Chem. Commun. 1969, 91, 5745-5749. (i) Jackson, R. B.; Streib, W. E. J. Am. Chem. SOC. 1967, 259-261. (b) Okabe, K.; Yamada, K.; Yamamura, S.; Takada, S. J. 1967,89,2539-2543. (j)McKechnie, J. S.; Newton, M. G.; Paul, I. C. J. Chem. SOC.C 1967,2201-2206. (c) Nakanishi, K. Pure Appl. Chem. 1967, Am. Chem. SOC.1967,89,4819-4825. (k) Baenziger, N. C.; Haight, H. L.; 14, 89-113. Alexander, R.; Doyle, J. R. Inorg. Chem. 1966, 5, 1399-1401. (I) Nyburg, (4)Corey, E. J.; Su, W.-g. J. Am. Chem. SOC.1987, 109, 7534. S. C.; Hilton, J. Acta. Crystallogr. 1959, 12, 116-121. (m) Mathews, F. S.; (5) Nakanishi, K.; Habaguchi, K.; Nakadaira, Y.; Woods, M. C.; Maru- Lipscomb, W. N. J. Phys. Chem. 1959, 63, 845-850. ' yama, M.; Major, R. T.; Alauddin, M.; Patel, A. R.; Weinges, K.; Btihr, W. (19) (a) Huttner, G.; Lange, S. Acta Crystallogr., Sect. E: Strucf. J. Am. Chem. Soc. 1971, 93, 3544-3546,and references contained therein. Crystallogr. Cryst. Chem. 1972, B28, 2049. (b) Huttner, G.; Lange, D.; (6) See: Barco, A.; Benetti, S. Synthesis 1981, 199-200. Fischer, E. 0. Angew. Chem., Int. Ed. Engl. 1971, IO, 556. (c) Churchill, (7) All reaction products were liquids unlas otherwise noted. Satisfactory M. R.; Mason, R. Proc. R. SOC.London, A 1966, 292, 61. high resolution mass spectra were obtained for all stable products. (20) (a) Pearson, R. G. J. Am. Chem. SOC.1963, 85, 3533-3539. (b) (8).Dimethoxyacetaldehyde was prepared in 72% yield by ozonolysis of Fleming, I. Frontier Orbitals and Organic Chemical Reactions; Wiley-In- dimethyl acetal (CH2C12at -78 "C). See: Bestmann, H.; Ermann, terscience: New York, 1976; Chapter 3. P. Chem. Eer. 1983, 116, 3264-3266. (21) Hodgson, K. 0.;Raymond, K. N. Inorg. Chem. 1973, 12, 458. (9)The trans arrangement of the vicinal B-tert-butyl and a-oxymethyl (22) (a) Strouse, C. E. Acta. Crystallogr., Sect. A,: Cryst. Phys., Dflr., substituents in 4 was demonstrated by 'H NOE experiments involving irra- Theor. Gen. Crystallogr. 1970,26,604408. (b) Schomaker, V.;Trueblood, diation of the tert-butyl group of 4 and the isomer of 4 with the cis arrangment K. N. Acta Crystallogr., Sect. B Struct. Crystallogr. Cryst. Chem. 1968, of these substituents (available to us by an alternative synthetic route but not 24, 63-76. produced under the conditions indicated for 4).

0002-7863/88/1510-0649$01.50/00 1988 American Chemical Society 650 J. Am. Chem. SOC.,Vol. 110, No. 2, 1988 Communications to the Editor

sponding acid chloride (5 equiv of oxalyl chloride in benzene at 23 "C for 2 h) and addition of the acid chloride in toluene solution (0.2 M) over 2 h to a stirred solution of tri-n-butylamine (10 equiv) in toluene at reflux, followed by further reaction at reflux for 1 'tBu "tBu h, furnished stereospecifically (7 1-8710 yield) the tetracyclic ketone 2 3 4 9, mp 59 OC. Structure 9, which results from internal ketene- olefin cy~loaddition'~and elimination of the anomeric (under tri-n-butylammonium chloride catalysis), follows from spectroscopic data and the transformation 11 - 12 described byoMe0 be10w.l~ Addition of 9 in acetone (at -30 "C) to a stirred solution of triphenylmethyl hydroperoxide in 8: 1 acetone1 N aqueous sodium 'ell hydroxide at -30 "C over 10 min and a further reaction time of 5 6 2 h at -30 "C produced a single Baeyer-Villiger product 10, mp 163 OC, in 86% yield.]* Lactone 10 was transformed into 4- hydroxylactone 11 (ginkgolide numbering as in 1) by a two-step sequence: (1) deprotonation (1.5 equiv of sodium bis(tri- methylsily1)amide in THF at -50 OC for 20 min) followed by reaction of the resulting anion with 2 equiv of (E)-2-(phenyl- sulfonyl)-3-phenyloxaziridine16(at -50 "C for 5 min and then at -50 OC to 0 OC over 10 min) to afford the corresponding why- * tBu droxylactone (73% after SGC) and (2) exposure to a 1% solution 7 of camphorsulfonic acid (CSA) in methanol at 23 OC for 48 h to give 11, mp 155 "C (75%)." Reaction of 11 with lead tetra- acetate (4.5 equiv) and (3 equiv) in pyridine-1,2-di- chlorcethane at 5 "C under sunlamp irradiation for 10 min resulted in complete conversion to a single product, determined by 500- MHz 'H NMR analysis to be the cyclic ether 12,1s rather than the hoped for product of functionalization at C( 12). Although this result was not useful as a synthetic step, it did provide con- formation of the stereochemistry of intermediates 9, 10, and 11. 10 The required oxygen bridge between C(4) and C(12) was es- tablished by an alternative route starting from 10. Reaction of 10 with 1.2 equiv of propane-1,3-dithiol and excess titanium tetrachloride in methylene chloride at 0 "C for 10 min and then at 23 OC for 40 min produced the thioacetal-primary alcohol 13, mp 230 "C (98%), which was transformed into the aldehyde 14, mp 165-166 OC (75% yield), by treatment with pyridinium di- chromate (PDC, 1 mol equiv), powdered 3-A molecular sieves and acetic acid in methylene chloride at 0 OC for 1 h. The aldehyde 14 was converted into the bis-acetal 15 (80% overall yield 11 12 as a 2: 1 mixture of C( 12) anomers) (major anomer from SGC, mp 107 "C) by the following process: (1) oxidative dithiane isolation and silica gel chromatography (SGC) an 80% yield of cleavage by reaction of 14 with 0.5 mol equiv of in enol triflate 5. A solution of 5 and Pd(PPh3), (5.7 mol%) in 1:l methanol-methylene chloride containing ca. 1% water at -30 benzene was stirred at 16 OC for 15 min and then treated suc- OC initially then at 0 OC for 20 min and 23 OC for 40 min and cessively with a benzene solution of acetylenic OB0 ester 6" (1 (2) stirring of the resulting product with methanolic CSA at 23 equiv), n-propylamine (2.3 equiv), and 0.5 equiv of cuprous iodide, "C. The C(4)-C( 12) oxygen bridge was generated by the fol- all at 16 "C to give after 4 h at 16 OC, extractive isolation and lowing sequence: (1) deprotonation of bis-acetal 15 with use of SGC 76-84% of the coupling product (2:l mixture of anomers), 7 1.9 equiv of lithium diethylamide initially at -25 OC and then at mp 44-47 OC.I2 The triple bond of was reduced by reaction 7 0 for 15 min and subsequent oxygenation with (E)-2-(phe- with 1.5 equiv of dicyclohexylborane in tetrahydrofuran (THF) "C (0 OC for 2 h, 23 OC for 0.5 h), followed by protonolysis (acetic acid 23 OC for 16 h), and decomposition of residual boranes (H,Oz, (14) See: (a) Corey, E. J.; Desai, M. C.; Engler, T. A. J. Am. Chem. SOC. 1985,107,4339-4340. (b) Corey, E. J.; Desai, M. C. Tetrahedron Lett. 1985, 23 "C, pH 10). The resulting solution was acidified to pH 3 with 26, 3535-3538. The stereospecificity of the internal cycloaddition to form 9 1 N hydrochloric acid, brought to pH 11 (vigorous stirring, 4 h) was predicted from mechanistic con~iderationsl~~and the lesser degree of steric and reacidified to pH 3 to cleave the OB0 ester unit,13 and the screening for the pathway leading to 9. (1 5) The structure of 10 was confirmed by the conversion of 11 - 12. Use (Z)-olefinic acid 8 was isolated by extraction and removal of of tert-butyl hydroperoxide as oxidant or higher reaction temperatures led to solvent (70% yield, colorless oil). Conversion of 8 to the corre- the appearance of the position isomeric lactone as a byproduct. (16) (a) Davis, F. A.; Stringer, 0. D. J. Org. Chem. 1982,47, 1774-1775. (b) Davis, F. A.; Vishwakarma, L. C.; Billmers, J. M.; Finn, J. J. Org. Chem. (10) McMurray, J. E.; Scott, W. J. Tetrahedron Lett. 1983, 24, 979-983. 1984, 49, 3241-3243. (1 1) The preparation of 6 was carried out as follows. 4-Pentynoic acid was (17) Methyl acetal lactone 11 was obtained as a single kinetically con- converted to the acid chloride (oxalyl chloride, benzene, 23 "C) which was trolled stereoisomer (from 500-MHz 'H NMR analysis). The orientation of treated with 3-methyl-3-hydroxymethyloxetaneto form the corresponding methoxy at C( 11) follows from the strong steric shielding by tert-butyl at the aster (85% overall) and in turn rearranged with boron trifluoride in methylene opposite face of the tetrahydrofuran subunit. The orientation of the hydroxyl chloride at -20 "C to form 6 (92%). See: Corey, E. J.; Raju, N. Tetrahedron group at C(4) follows from the strong preference for formation of a cis Lett. 1983, 24, 5571-5574. 4-Pentynoic acid was obtained in 25% overall yield 5,S-fusion in the hydroxylation reaction and is further confirmed by the by the following sequence: (1) alkylation of diethyl sodiomalonate in ethanol conversion to 12. with propargyl chloride; (2) saponification with potassium hydroxide in (18) The 'H NMR spectrum of 12 (with spin decoupling) provides unam- aqueous ethanol at 23 "C; and (3) thermal decarboxylation of propargyl biguous support for this structure and the following key assignments: H12a, malonic acid at 150-170 OC over 45 min. d, 4.33 6, = 9.6 HZ; H12b, d, 3.45 6, J12a,i2b = 9.6 HZ; H7*, dd, 2.40 (12) (a) Sonogashira, K.; Tohda, Y.;Hagihara, N. Tetrahedron Lett. 1975, 6, J7a,7b = 15.1 Hz, J~,J~= 7.7 Hz; H7b d, 1.98 6, JJ~,J~= 15.1 Hz; f-Bu, S, 4467-4471. (b) Ratevelomana, V.; Linstrumelle, G. Synth. Commun. 1981, 1.07 6. In all intermediates in the synthesis which have H attached to C(8) 11, 917-923. a coupling J7b,8of 5-6 Hz is observed; the doublet for HSbthen shows there (13) Corey, E. J.; De, B. J. Am. Chem. SOC.1984, 106, 2735-2736. is no H attached to C(8). J. Am. Chem. SOC.1988, 110, 651-652 65 1 elimination of methanol from C(10)-C(11) by heating 18 under argon with 5 equiv of pyridinium tosylate and 2.5 equiv of dry pyridine in chlorobenzene at 135 "C for 16 h (80% yield)21*22and (2) enone epoxidation with triphenylmethyl hydroperoxide (5 equiv) and benzyltrimethylammonium isopropoxide (0.5 equiv) in THF at -10 "C for 3 h to give after reduction of excess hy- droperoxide by trimethyl phosphite (10 equiv) and SGC 72% of 19. Reaction of 19 with 7 equiv of the lithium enolate of tert-butyl 0 0 propionate (from LDA) in 4: 1 THF-hexamethylphosphorictri- 13 XrCHeOH 15 X=H amide, at -78 "C to -30 OC for 2 h and then at -30 "C for 10 14 X=CHO 16 X=OH h, furnished the desired aldol adduct 20 in 60% yield after SGC.23 Exposure of 20 to 4 equiv of CSA in methylene chloride (23 OC OM0 for 15 h) afforded bis-lactone 21 (82%) which was converted to the tert-butyldimethylsilyl (TBMS) ether 22 upon treatment with 2.5 equiv of TBMS triflate and 5 equiv of 2,6-lutidine in aceto- nitrile at 23 "C for l h (89%). Hydroxylation of 22 using osmium tetroxide in pyridine followed by oxidation of the resulting product with excess iodine in methanol in the presence of CaC03 (23 "C for 12 h) produced trilactone 23 (ca. 40% from 22)2' along with a small amount of the C(10) epimer. Desilylation of 23 (5 equiv 17 X=H2 18 of BF,-Et,O in methylene chloride at 23 "C for 14 h) gave 89% 18 x=o yield of (f)-ginkgolide B (l),identical with an authentic sample by 500-MHz 'H NMR, FT-IR, SG-TLC analysis in several solvent systems, and mass spectral c~mparison.~~ -tary Material Availak Spectral data for compounds 1-23 (6 pages). Ordering information is given on any current masthead page.

(21) The yields given for this and remaining steps in the synthesis are 20 probably not optimum since these reactions have been conducted only a few 21 RIH times without systematic attempts at further improvement. 0 22 R=TBMS (22) This product and also 21 and 1 are solids which decompose before melting; 22 and 23 are colorless oils. (23) A small amount of C(3) epimer (ca. 7%) was also obtained; the C(3) epimer became the major product from reaction in THF alone at -78 "C. (24) We are grateful to Dr. P. Braquet, Institute Henri Beaufour, Paris and Dr. K. Yamada of Kagoya University for samples of ginkgolide B, to Dr. Ashvinikumar V. Gavai and Dr. Yi Bin Xiang for valuable experimental assistance, and to Francis J. Hannon for determination of mass spectra. This research was assisted financially by grants from the National Institutes of Health and the National Science Foundation. 23 R=TBMS 1 RsH Asymmetric Synthesis on Carbohydrate Templates: nylsulfonyl)-3-phenylo~aziridine~~(2 equiv, 0 "C for 30 min) to Stereoselective Ugi Synthesis of a-Amino Acid form the a-hydroxylacetone 16, mp 115-1 16 "C, and (2) reaction with a solution of CSA in methylene chloride (20 mg/100 mL) Derivatives at 23 OC for 24 h to afford 17, mp 151-153 OC (75%).*O The Horst Kunz* and Waldemar Pfrengle introduction of an oxo function at C(3) was accomplished in 50% overall yield by the following transformations: (1) allylic bro- Institut fur Organische Chemie mination with 1.3 equiv of N-bromosuccinimide in carbon tet- Universitat Mainz, 0-6500Mainz rachloride (0.02 M) under external tungsten lamp irradiation at Federal Republic of Germany 10 OC for 2-3 h (monitored by SG TLC) to give a mixture of Received September 3, 1987 60% of the C(3) brominated product (Br3), 30% of the C(l) brominated product (Brl), and 10% of the 3,3-dibrominated Experiences from the of glycopeptides' re- product (Br393);(2) reaction of the mixture with 10 M vealed that carbohydrates exhibit considerable complexing abilities in acetonitrile at 23 OC for 15 min which generates a mixture of toward cations. This stimulated the concept to utilize this com- the enone 18, mp 267-268 OC (from Br3v3), the 1-nitrate ester plexation together with the high chirality of the carbohydrates (from Br3), and the 3-nitrate ester (from Br'), easily separated for the stereochemical control of reactions.' The present paper by SGC; (3) conversion of the 3-nitrate ester to 18 by nitrate reports a stereoselective formation of a-amino acid derivatives cleavage with zinc-acetic acid followed by oxidation of the re- by using the Ugi four-component c~ndensation.~ sulting C(3) alcohol with PDC in methylene chloride at 23 OC In contrast to the recently developed stereoselection methods for 5 h; (4) conversion of the 1-nitrate ester to 18 by exposure for synthesis of a-amino acids4 based on the electrophilic ami- to 20 equiv of 1,8-diazabicyclo[5.4.O]undec-7-eneand 20 equiv of water in 1O:l benzene-methanol at 23 "C for 30 h followed (I) Review: Kunz, H. Angew. Chem., Int. Ed. Engl. 1987, 26, 294. by oxidation of the resulting C(3) alcohol (from overall SN2' (2) Kunz, H.; Muller, B.; Schanzenbach, D. Angew. Chem., In?. Ed. Engl. 1987, 26, 267. Apart from chiral pool syntheses and some chiral hydride reaction) by PDC. reagents, carbohydrates have been used only in isolated cases in asymmetric The final y-lactone ring was affixed starting with epoxy ketone synthesis: BrandBnge, S.; Josephson, S.; Morch, L.; ValEn, S. Acta Chem. 19 which was obtained from 18 in the following two steps: (1) Scand., Ser. B 1981, 35, 1296. Heathcock, C. H.; White, C. T.; Morrison, J. J.; VanDerveer, D. J. Org. Chem. 1981, 46, 1296. Hoppe, I.; Schollkopf, U.; Tolle, R. Synthesis 1983, 789. Vasella, A,; Huber, R.; Knierzinger, A,; (19) The observed course of the functionalization to give 12 suggests that Obrecht, J.-P. Helv. Chim. Acta 1985, 68, 1730. Felber, H.; Kresze, G.; steric repulsion between the tert-butyl substituent at C(8) and the Prewo, R.; Vasella, A. Helv. Chim. Acta 1986, 69, 1137. Gupta, R. C.; at C(10) and C(11) causes puckering of the C(5)-C(9) ring so as to bring Slawin, A. M. Z.; Stocdley, R. J.; Williams, D. J. J. Chem. Sot., Chem. H-C(8) into proximity to O-C(4). Commun. 1986,668 and 11 16. (20) The configuration at C(11) of 17 follows clearly from chemical and (3) Ugi, I.; Offermann, K.; Herlinger, H.; Marquarding, D. Justus Liebigs NOE studies not reported herein. Ann. Chem. 1967, 709, 1 and references therein.

0002-7863/88/1510-0651$01.50/0 0 1988 American Chemical Society J. Am. Chem. SOC.1981,103, 219-222 219 The yield of higher hydrocarbons is very low. This reflects the not a prerequisite for CH4 formation.8 (Several experiments in fact that in TPDE the primary reaction products are swept from flowing Hz at loadings between 0.0036 and 0.76% Mo show that the reactor with very short contact times ( N 1.4 s).13 Although the yield of Cz/Mo is independent of loading, again suggesting the C balance (based on the collection of the above gases) is usually that multinuclear sites are not involved.) Hydrogenation is not quite good (>90%), in a few cases it is not, most notably for preceded by the of CO since the yield of COz Rh6(C0)16.15 This could be due to the formation of heavier is extremely low. Further, in most cases dissociation of CO products (including oxygenates) which are not detected.I6 followed by hydrogenation of Cads is unlikely, since the CH4 is However, due to the low pressure and short contact times, a more often formed in a temperature region in which CO is evolved if likely explanation may be the formation of some type of Cads the TPDE is carried out in flowing He. Hence, it is most likely (adsorbed carbon) which is unreactive toward H2.l7 in these systems that the direct hydrogenation of coordinated CO A notable feature of Table I is the large amounts of CH4which to CH4 is occurring at discrete subcarbonyl sites. are formed, in six cases the conversion of CO to CH4 exceeding In addition to the substantial yields of CHI during TPDE in 50%. The yield of CH4 does not correlate with the activity of the flowing Hz, it is seen (Table I) that the reaction often occurs at respective metals for catalytic methanationk8 For example, Ir, temperatures well below that required for catalytic methanation the least active of the group 8B metals for methanation, yields (200-350 OC).18 An interesting corollary to this result is that a far more CHI than Ni, one of the most active metals. Examination good catalyst for methanation should result if a metal can react of Figures 1-4 shows that the thermal stability of a supported with CO to re-form a carbonyl (or subcarbonyl) complex, since complex is the prime factor in determining the quantity of CH4 a catalytic cycle is now formed. Two of the more active catalysts, formed. Thus Ni(CO),, which decomposes at a very low tem- Ni and Fe, are in fact the metals which most readily form car- perature (Figure l), has lost most of its CO before the temperature bonyls from CO, whereas many of the catalytically less active is high enough to give a reasonable rate of methanation (Figure metals (Ir, Mn, and Cr) do not undergo this reacti~n.'~*~~ 2).19 In contrast, Ir4(CO)12does not lose its CO until 125 OC However, the most active metal for methanation, Ru, only very (Figure 3).2k TPDE in flowing H2 (Figure 4) then yields a curve slowly forms a carbonyl by exposure to CO under severe condi- for CH4 evolution which is remarkably similar to that for CO tions, although it was suggested that this may be due to adsorption evolution in He (Figure 3). Similar correlations are found for of RU(CO)~which inhibits further reaction.27 Hydrogenation of the other catalysts. carbonyl-like intermediates has been considered as a mechanism The two complexes giving the most CH4 (per complex) are for methanation and Fischer-Tropsch synthesis,z8 but currently Ir4(CO)12and OS~(CO)~~.These are the same two cluster com- favored is the dissociation of CO followed by hydrogenation of plexes which were reported to be active for homogeneous catalytic CadseZ9Thus, although the facile hydrogenation of coordinated methanatiomz2 However, in those experiments the total yield CO is now demonstrated, it is nuclear if this process is important of CH4 was only about 4 CH4/complex, whereas several times during catalytic methanation. Catalytic methanation over sup- this amount is now seen to be formed in a purely stoichiometric ported carbonyl complexes is currently being studied. reaction.z3 Hence, it is possible that the claimed catalytic reaction was in fact stoichiometric (or a heterogeneous reacti~n~~~~~)and Acknowledgment. Support of this research by the Department the ability of some cluster complexes (but not mononuclear of Energy is gratefully acknowledged. complexes) to homogeneously yield CH4 may simply reflect their enhanced thermal stability. (27) Wender, I.; Pino, P. "Organic Synthesis Via Metal Carbonyls"; In- TPDE in flowing Hz increases the CH4 yield by 25-fold over terscience: New York, 1968; Vol. 1, Chapter 1. TPDE in flowing He.8 In both He8 and Hz (Table I) it is found (28) Anderson, R. B. Catalysis 1956, 4, 29. that the quantity of CHI formed is essentially independent of the (29) Biloen, P.; Helle, J. N.; Sachtler, W. M. H. J. Caral. 1979, 58, 95. nuclearity of a complex. This is contrary to some claims that multinuclear sites are necessary to effect the reduction of COZ6 A Strategy for the Total Synthesis of Jatrophone: but consistent with more recent work suggesting that mononuclear Synthesis of Normethyljatrophone complexes can be active for methanation.8~~~Recent data for the TPDE of Mo(C0)6/Al2O3 in flowing He and Hz indicate that Amos B. Smith, III,*t Michael A. Guaciaro, sintering of mononuclear precursors to polynuclear sites is probably Steven R. Schow, Peter M. Wovkulich, Bruce H. Toder, and Tse Wai Hall (1 3) Hydrocarbon synthesis has been reported for several carbonyl cluster The Department of Chemistry complexes supported on wet A1203and heated in sealed ampules at approx- The Laboratory for Research on the Structure of Matter imately 300 OC for - l@ s.I4 (H2 is generated by the water gas shift reaction.) and The Monell Chemical Senses Center (14) Smith, A. K.; Theolier, A,; Basset, J. M.; Ugo, R.; Commereuc, D.; Chauvin, Y. J. Am. Chem. SOC.1978, 100, 2590. The University of Pennsylvania (15) Rh,(CO),, also gave a particularly poor C balance for reaction in Philadelphia, Pennsylvania I91 04 sealed ampule^.'^ Received October I, 1980 (16) Ichikawa, M. Bull. Chem. SOC.Jpn. 1978,51, 2273. (17) TPDE in flowing H, of Mn2(CO),o yielded some CH4 at temperatures Jatrophone (l), an architecturally interesting macrocyclic higher than for CO evolution in a sweep gas of He, suggesting some hydro- diterpene first isolated in 1970 by the late Professor Kupchan from genation of Cads. However, due to the generally good C balances and the fact that TPDE in Hz did not significantly improve the C balances compared extracts of Jatropha gossypiifolia L. (Euphorbiacae),' merits to TPDE in He, it appears that usually only small quantities of Cads can be consideration as a synthetic target in that it displays significant formed, and it must be fairly unreactive toward H,. inhibitory activity against a variety of cell lines, including sarcoma (18) Vannice, M. A. J. Carol. 1975, 37, 449. 180, Lewis lung carcinoma, P-388 lymphocytic leukemia, and (19) These results are in a reement with less detailed data previously published for Ni(CO)4/A1203!o Walker 256 intramolecular carcinsarcoma.2 Indeed, extracts of (20) Bjorkland, R. B.; Burwell, R. L., Jr. J. Colloid Interface Sci. 1979, this plant had long been employed in the treatment of cancerous 70, 383. growths.2 (21) For this experiment the Ir4(C0)12was dispersed at 125 OC during which time 0.2 CO/complex was evolved. The structure of jatrophone was based on both chemical and (22) Thomas, M. G.; Beier, B. F.; Muetterties, E. L. J. Am. Chem. Soc. X-ray crystallographic studies.' Central to the derived structure 1976, 98, 1296. (23) The homogeneous reactions were run at 140 OC in toluene solution Camille and Henry Dreyfus Teacher-Scholar, 1978-1983; recipient of and gave a turnover frequency of 1 X lo-' s-'. At 140 "C the extrapolated a National Institutes of Health (National Cancer Institute) Career Devel- turnover frequencies for the stoichiometric of Ir4(C0)12and opment Award, 1980-1985. OS~(CO),~are also -1 x 104 s-I. (1) S. M. Kupchan, C. W. Sigel, M. J. Matz, J. A. Saenz Renauld, R. C. (24) Bradley, J. S. J. Am. Chem. SOC.1980, 101, 7419. Haltiwanger, and R. F. Bryan, J. Am. Chem. Soc., 92,4476 (1970); S. M. (25) Doyle, M. J.; Kouwenhoven, A. P.; Cornelis, A.; Oort, B. V. J. Or- Kupchan, C. W. Sigel, M. J. Matz, C. J. Gilmore, and R. F. Bryan, ibid., 98, ganomet. Chem. 1979, 174, C55. 2295 (1976). (26) Muetterties, E. L. Science 1977, 196, 839. (2) J. L. Hartwell, Lloydia, 32, 153 (1969). 0002-7863/81/1503-219$01 .OO/O 0 1980 American Chemical Society 220 J. Am. Chem. SOC.,Vol. 103, No. 1, 1981 Communications to the Editor Central to the above scenario was the availability of an efficient, and hopefully general, method for the construction of the 3- (2H)-furanone ring system; unfortunately no such approach was available at the outset of our work. In fact, the chemistry of this increasingly important heterocycle has been little e~plored.~ Convinced that the most plausible route to 3(2H)-furanone de- rivatives would involve the acid-catalyzed cyclizationdehydration of an appropriately substituted a-hydroxy 1,3-diketone,we directed Jatrophone, R = CH, (1) our attention to model studies of the aldol-oxidation strategy Normethyljatrophone, R = H (2) illustrated in eq 1. We note in advance that this approach proved quite advantageous (vide infra).5

(1) 1 eq LDA (2)RCHO - (2) H,Ot OSiMe, OSiMe,

(1 1 With a viable route to the central structural element, the 3- (2H)-furanone ring, secure, we initiated work on the normethyl system. Our strategy here begins with the elaboration of hydroxy ketone 5b and aldehyde 6. To this end, the readily available (3) (4) a-(hydroxymethy1)cyclopentenone 76 was converted to 5b in 42% overall yield (five steps) via the dithiane adduct. Specifically, the 0 tert-butyldimethylsilyl ether of 7 was treated with lithioethyl- dithiane followed by hydrolysis (CH31/CaC03/CH3CN/H20). Subsequent removal of the tert-butyldimethylsilyl group (aqueous AcOH/THF) afforded a keto diol which was silylated to give 5b.' Aldehyde 6 in turn was prepared from 3,3-dimethyl-4-pentynoic acid (8)*in 55% yield via a straightforward sequence. The dianion derived from 8 (2.4 equiv of LDA/4.6 equiv of HMPA, -78 "C) T(,hi + was treated with excess propanal; protection of the resultant I hydroxy acid as the bis(tert-butyldimethylsilyl)derivative followed (5) ( 6) by LiAlH4 reduction in ether and Collins oxidation9 afforded the (a): R=CH, desired aldehyde (6).' (b): R= H kCOOH are the spiro-3(2H)-furanone, the macrocyclic ring, and the I cross-conjugated trienone functionalities. oo Recently, we have made considerable progress toward the Ill development of a viable synthetic strategy for the elaboration of HO J jatrophone. We now wish to announce completion of the synthesis of normethyljatrophone (2), the alicyclic backbone of this novel ( 7) (8) diterpene. Our strategy for both jatrophone and the normethyl With ample quantities of both 5b and 6 in hand, we executed system calls for the initial convergent elaboration of a spiro- the aforementioned 3(2H)-furanone synthetic protocol. To our furanone carbocyclic array (4). With all necessary carbons in place, closure of the macrocyclic ring (e.g., aldol condensation) delight, condensation of 5b with 6, followed by oxidation with followed by conversion of the C(8,9) acetylene to a trans olefin would then complete the synthesis. The strategic selection of an (4) For a recent detailed study on the synthesis and reactions of simple 3(2H)-furanones see A. B. Smith, 111, P. A. Levenberg, P. J. Jerris, R. M. acetylene functionality at C(8,9) was seen as the cornerstone of Scarborough, Jr., and P. M. Wovkulich, J. Am. Chem. SOC.,in press. this approach in that it was anticipated that the linear nature of (5) We have, for example, prepared in good to excellent yield a wide this functionality would provide the configurational control re- variety of 5-alkyl-5-aryl-and 5-alkenyl-2,2-dimethyl-3(2H)-furanonesvia this quired at C(5,6). That is, examination of molecular models of approach; see ief 4. (6) S. J. Branca and A. B. Smith. 111. J. Am. Chem. Soc.. 100. 7767 3 reveals that the Z configuration at C(5,6) should be considerably (1978); A. B. Smith, 111, M. A. Guaciaro, and P. M. Wovkulich, Tetrahedron more stable than the corresponding E. Alternatively, models of Lefr., 4661 (1978). jatrophone (1) suggested that both configurational isomers at (7) (a) The structure assigned to each new compound was in accord with C(5,6) are feasible molecular systems, the natural 2 isomer being its infrared and 220-, 250-, or 360-MHz NMR spectra. Analytical samples of all new compounds, obtained by recrystallization or chromatography (LC only slightly less strained. For final conversion of the acetylene or TLC), gave satisfactory C and H combustion analysis within 0.4% and/or linkage to the required trans olefin, we anticipated exploiting appropriate parent ion identification by high-resolution mass spectrometry. chromous sulfate, recently shown in our laboratory to effect cleanly (b) All yields recorded here are based upon isolated material which was >97% the stereospecific reduction of a-oxoacetylenes to the corresponding pure. The 250-MHz 'H NMR spectral data (CDCI,) for representative intermediates are recorded below. 13 NMR 6 1.33(s, 3 H), 1.45(s, 3 H), trans enone derivative^.^ 1.73(s, 3 H), 1.80(d, J = 2 Hz, 1 H), 2.00-2.70(m, 5 H), 2.90(d, J = 14 Hz, 1 H), 5.37(m, 1 H), 6.00(m, 1 H). 14: 6 1.16(s, 3 H), 1.24(s, 3 H), 1.70(s, (3) Although chromous sulfate reduction of acetylenic alcohols to the 3 H), 1.88(d, J = 2 Hz, 3 H), 2.02-2.70(m, 4 H), 2.32(d, J = 14 Hz, 1 H), corresponding trans allylic alcohols has been known for some time [see J. 2.64(d, J = 14 Hz, 1 H), 5.62(d, J = 14 Hz, 1 H), 5.86(d, J = 14 Hz, 1 H), Inanaga, T. Katsuki, S.Takimoto, S. Ouchida, K. Inoue, A. Nakano, M. Aiga, 5.94(m, 1 H), 6.08(m, 1 H). 2: 6 1.26(s, 3 H) 1.38(s, 3 H), 1.76(s, 3 H), N. Okukado, and M. Yamaguchi, Chem. Lett., 1021 (1979); C. E. Castro and 1.90(d, J = 2 Hz, 3 H), 5.80(m, 1 H), 5.94(m, 1 H),6.00(d, J = 16 Hz, 1 R. D. Stephens, J. Am. Chem. SOC.,86, 4358 (1964)], reduction of a-ace- H), 6.66(d, J = 16 Hz, 1 H). 1: 1.08(d, J = 7 Hz, 3 H), 1.24(s, 3 H), 1.36(s, tylenic ketones is novel. In that regard we have demonstrated that i undergoes 3 H), 1.74(s, 3 H), 1.88(br s, 3 H), 1.88(dd, J = 7 and 12 Hz, 1 H), 2.16(dd, stereospecific reduction to ii in 53% yield. J = 7 and 12 Hz, 1 H), 2.40(d, J = 14 Hz, 1 H), 2.88(d, J = 14 Hz, 1 H), 2.98(m, 1 H), 5.78(m, 2 H), 5.98(d, J = 15 Hz, 1 H), 6.44(d, J = 15 Hz, 1 H). (8) 0. K. Behrens, J. Come, D. E. Huff, R. G. Jones, Q.F. Soper, and C. W. Whitehead, J. Biol. Chem., 175, 771 (1948). (9) J. C. Collins, W. W. Hess, and F. J. Frank, Tetrahedron Left., 3363 lil /ii/ (1968). Communications to the Editor J. Am. Chem. Sot., Vol. 103, No. 1, 1981 221 Collins reagent (12 equiv), acid-catalyzed deprotection, and cy- clization-lehydration (10% HCl/THF, 5 h, room temperature), afforded spirofuranone 9a as a mixture of diastereomers. The

(111

/+O to complete normethyljatrophone, the first major plateau in our ii jatrophone synthetic program. Toward this end, both the major (11) and minor isomers were found to eliminate cleanly (9a): R = R'= 0 glycol upon exposure to toluenesulfonic acid in benzene, affording (9b): R = R'= OCH,CH,O the same crystalline solid (mp 185-186 "C) in 79% yield. The overall yield of 9a based on 5a after purification (medium-pressure 'H NMR spectrum (250 MHz) of the elimination product re- LC, hexane/ethyl acetate, 2:l) was 68%. Fortunately for our vealed a substantial downfield shift for the C(5) vinyl proton. This purposes (vide infra) the allylic silyl ether underwent oxidation result, clearly inconsistent with the chemical shift expected on to the corresponding aldehyde during the Collins procedure? Final the basis of analogy to jatrophone (6 5.80 vs. 7.20 in CDC13), conversion to keto aldehyde 4b: a crystalline solid (mp 91.5-92.5 suggested that we had in fact generated the less stable E isomer "C), was accomplished again through the agency of Collins (l2).' X-ray crystallographic analysis, demonstrated this to be reagent. With this efficient convergent approach to keto aldehyde 4b available, the stage was set to explore the key aldol cyclization (Le., 4b - 3b). Unfortunately, all attempts to effect a direct cyclization to the desired eleven-membered ring system employing a wide variety of different acidic and basic reagents met with complete failure. Success was finally achieved via a somewhat less direct approach. Consider for the moment the events required for the successful cyclization of 4b. First, chemospecific generation of the enol or enolate of the C(7) ketone and not the aldehyde (121 is required.'O Second, the enolate (enol) must undergo irreversible addition to the aldehyde. Ideal here would be the availability of the case.I2 Significantly for our purposes, furanone (12) was found an internal protecting or trapping agent. The latter stringent to undergo slow isomerization to the more stable Z isomer (13)' requirement appeared to be ideally fulfilled by the Mukaiyama' upon prolonged exposure to the above elimination conditions (2 TiC14-promoted condensation of acetals with enol silyl ethers. To weeks, 88%). this end, ketal 9b prepared from 9a (loo%, HOCH2CH20H, We had now arrived at the final step necessary for completion TsOH, C6H6,-H20, 24 h) was oxidized (85%, Collins reagent)9 of our synthetic goal. The planned one-step conversion of the to keto acetal 10, which in turn was converted to the corresponding acetylene functionality to the desired trans olefin with chromous enol silyl ether (LDA/THP, -73 OC, 2.2 equiv of Me3SiC1). ion was, however, thwarted, presumably due to transannular re- Without isolation, the long sought after cyclization was effected action~.'~Fortunately, acetylene (13) could be selectively cis with TiC14 (2.0 equiv, CHZC12, -78 OC, 5 min). Two diastereo- hydrogenated (PdS04/pyridine)14 to ci~/Z-14,'*'~which in turn was promptly isomerized (KI AcOH, room temperature, 30 min)I6 meric products (47%, 2:l) resulted, the major (11) proving to be to normethyljatrophone (2) (71% yield from 12), the latter ob- crystalline (mp 156-157 "C). To demonstrate beyond doubt that J/ the macrocyclic ring was intact, we completed a single-crystal tained as a beautiful crystalline solid (mp 135-136 "C). That X-ray analysis; the result of that study is illustrated below.12 normethyljatrophone was in hand derived from careful comparison Elimination of the elements of and conversion of the acetylene to a trans olefin were now all that were required (13) That trans annular reaction plays a significant role in the reaction of chromous sulfate with 13 derives from the observed cyclization of model system i to ii. The latter was contaminated by a small amount of iii. Ad- (10) To control generation of the specific ketone enolate, we initially se- ditional examples of this transformation will be reported in due course. lected the "Reformatsky-like" aldol condensation conditions recently intro- duced by Yamamoto; see K. Maruoka, S. Hashimoto, Y. Kitagawa, H. Yamamoto, and H. Nozaki, J. Am. Chem. Soc., 99,7705 (1977); also see J. Tsuji and T. Mandai, Tetrahedron Lett., 1817 (1978). To this end bromo aldehyde i was prepared from 9b and subjected to the Yamamoto conditions.

(il IiiI IiiiI Et2A1CI/Zn HO (14) L. Fieser and M. Fieser, "Reagents for Organic Synthesis", Wiley, 123%) New York, 1967, p 566. (1 5) Interestingly, hydrogenation (PdSO,/pyridine) of 12 afforded fura- none 15 (mp 212-213 "C) which proved to be stable to cis-trans isomerization. ti/ e(lil The structure of 15 was assigned by X-ray crystallographic analysis.I2

While successful, the above strategy does not eliminate the possibility of reversal of the aldol reaction; the latter, of course, would contribute to the modest yield of observed cyclization (Le., 23%). Indeed, isolation of keto aldehyde 4, as a major side product, is suggestive that a reverse aldol process may be occurring under the Yamamoto conditions. Furthermore, we found aldol ii to be remarkably unstable, reverting to 4b with ease! (1 1) T. Mukaiyama and M. Yayashi, Chem. Lett., 15 (1974); K. Banno and T. Mukaiyama, Chem. Lett., 741 (1975); T. Mukaiyama and A. Ishida, Chem. Lett., 312 (1975). (12) Unpublished results of B. H. Toder, a graduate student, in our lab- (16) We thank Dr. William Schreiber, International Flavors and Fra- oratory; a detailed account will be published in due course. grances, for bringing this procedure to our attention. 222 J. Am. Chem. SOC.1981, 103, 222-223

of the 250-MHz ‘H NMR spectral data of the normethyl system synthesis, Horii and co-workers succeeded in preparing tricyclic with that of authentic jatrophone’,’’ as well as by completion of amino ketone Z4 but were unable to add the final ring.5 Because a single-crystal X-ray analysis; that result is illustrated below.12

&*o H

2 1: X=OH

0%3: X=H - of the close structural resemblance between lycodoline and ly- copodine (3), we have examined the use of intermediates employed in our lycopodine synthesis6 for construction of alkaloid 1. However, the presence of the sensitive tertiary alcohol function in lycodoline precludes the use of the acidic conditions required to effect the key Mannich cyclization in the synthesis of 3.6 In this communication, we report an interesting solution to this problem, which has culminated in the first total synthesis of (f)-lycodoline. Normethyljatrophonc Amino diketal5, available in three steps (58% overall yield) from cyano enone 4,6 is treated briefly with 10% aqueous HCl, In summation, the total synthesis of normethyljatrophone has and the resulting solution is made basic with NaOH. The unstable been achieved in 15 steps and in 5.6% overall yield from cyclo- pentenone (7). Four X-ray crystallographic analyses were com- n pleted during this venture, thereby confirming the structure of 11, 12, 15 and that of normethyljatrophone (2). Studies to improve the overall sequence, as well as to effect the total synthesis of jatrophone paralleling the above strategy, will be reported in due course. Note Added in Proof. Since acceptance of the manuscript, we have successfully completed the first stereocontrolled total synthesis of both (+)-jatrophone (1)and that of its epimer (+)-epijatrophone, exploiting the synthetic strategy outlined

above; 5a and its epimer served, respectively, as starting materials. 7 8 A complete account of this effort will be forthcoming in the near future. octahydroquinoline (6) is extracted with ethyl acetate, and the resulting solution is treated with oxygen gas and then hydrogen Acknowledgment. It is a pleasure to acknowledge the support and Pd/C to obtain a mixture of alcohol 7 (mp 164-65 OC, 43%) of this investigation by the National Institutes of Health (National and hemiketal 8 (oil, 4%). This interesting autoxidation finds Cancer Institute) through Grant No. CA-22807. In addition, we precedent in the work of Cohen and Witkop on the parent octa- thank Mr. S. T. Bella of the Rockefeller University for the mi- hydroquinoline.’ In the present case, it is noteworthy that the croanalyses and the Middle Atlantic Regional NMR Facility diastereomer having the angular oxygen and the neighboring (NIH No. RR542) at the University of Pennsylvania where the acetonyl group trans predominates by a factor of 10:l. We 220- and 360-MHz spectra were recorded. postulate that this stereoselectivity arises from simple steric hindrance of approach of an oxygen molecule to the intermediate (17) We thank Dr. Matthew Suffness of the National Cancer Institutes free radical. and Dr. Jeffrey Cordell, University of Illinois, Chicago Circle-Medical The third ring is smoothly formed by heating a dilute solution Center, for providing us with a generous sample of jatrophone. of compound 7 (0.075 M) in a 5:l mixture of toluene and 3- bromopropanol at reflux for 24 h. Neutralization of the resulting hydrobromide salt (which crystallizes from the hot solution) provides amino ketone 2 (mp 165-166 “C)in 85% yield. Many other attempts to accomplish this cyclization were unsuccessful. Since the product is a hydrobromide salt, a full equivalent of HBr Total Synthesis of (f)-Lycodoline is required. However, it appears to be crucial to the success of the reaction that the acid be added exceedingly slowly. Thus, if Clayton H. Heathcock* and Edward F. Kleinman the hydrobromide salt of 7 is heated for 24 h in toluene, no cyclization occurs. 3-Bromopropanol functions as a source of Department of Chemistry, University of California HBr by slowly polymerizing under the reaction conditions. It is Berkeley, California 94720 interesting to note that 3-bromopropanol is superior to 2- Received September 5, 1980 bromoethanol for this purpose. Lycodoline (1, “alkaloid L.8”) is the second most widely oc- (3) (a) W.A. Ayer and G. G. Iverbach, Tetrahedron Lett., No. 3, 87 curring of the lycopodium alkaloids.’ It was first isolated in 1943 (1961); (b) W.A. Ayer and G. G. Iverbach, Can. J. Chem., 42,2514 (1964). by Manske and Marion from L. annotinum Linn2 and its structure (4) Z.4. Horii, S.-W. Kim, T. Imanishi, and T. Momose, Chem. Pharm. was established in 1961 by Ayer and Iverbach.’ In an attempted Bull., 18, 2235 (1970). (5) S.-W. Kim, Y. Bando, and 2.4.Horii, Tetrahedron Lett., 2293 (1978). (6) C. H. Heathcock, E. Kleinman, and E. S. Binkley, J. Am. Chem. Soc., (1) D. B. MacLean, Alkaloids, 14, 347 (1973). lod, 8036 (1978). (2) R. H. Manske and L. Marion, Can. J. Res., Sect. B, 21, 92 (1943). (7) L. A. Cohen and B. Witkop, J. Am. Chem. SOC.,77, 6595 (1955). 0002-7863/81/1503-222$01.00/0 0 1980 American Chemical Society 3442 J. Org. Chem. 1982,47, 3442-3447

6 0.9 (6 H, t, CHzCH3),1.2 (19 H, br s, RzCHz,R3CH), 3.7 (3 H, the following: bp (Kugelrohr) 187-193 "C (0.15 torr); nuD 1.4558"; m, RzCHOH, RCHzOH); IR 3400 cm-'. IR 1735 cm-'; 'H NMR 6 0.9 (9 H, t, CH,CH3), 1.3 (27 H, br s, Anal. Calcd for C14HMOz: C, 72.98; H, 13.12; mol wt, 230. RzCHz,R3CH), 2.2 (2 H, t, CH,C=O), 3.8 (2 H, S, COCHzBr), Found: C, 72.81; H, 12.92; mol wt, 230 (MS). 4.1 (2 H, d, RCHzOC=O), 5.0 (1 H, m, RzCHOC=O). Saponi- A solution of 500 mg of the mixture of 3a and 3b, 2.5 mL of fication2' gave heptanoic acid and 2-pentyl-1,3-nonanediol (7); acetic anhydride, and 0.75 mL of pyridine was kept for 24 h at bromoacetic acid was not recovered in the workup. room temperature under anhydrous conditions. Workup gave Anal. Calcd for Cz3H43Br04:mol wt, 462. Found: mol wt, 512 mg of the monoacetyl derivative: bp (Kugelrohr) 162-168 462 (MS). "C (0.2 torr); nz4D 1.4430" [lit.21 for 2-pentyl-3-acetoxynonyl 2-Pentyl-3-acetoxynonyl heptanoate (5a) and its isomer (5b), heptanoate, bp 160-164 "C (1.0 torr); n'8'5D 1.448401. a mixture, had the following: bp (Kugelrohr) 162-166 "C (0.1 torr); Anal. Calcd for CBHUO4: C, 71.83; H, 11.52. Found C, 72.05; nz4D1.4427" [lit.z1bp 160-164 OC (0.1 torr); n18,5D1.4484"]; IR H, 11.68. 1740 cm-'; 'H NMR 6 0.9 (9 H, t, CHzCH3),1.3 (27 H, br s, R&Hz, Oxidation of 3a and 3b. A mixture of 3a and 3b (390 mg) RSCH), 2.0 (3 H, 9, OCOCH3), 2.2 (2 H, t, CHzC=O), 4.0 (2 H, was dissolved in acetone (250 d)and Jones reagentm was added m, RCHzOC=O), 4.9 (1 H, m, RzCHOC=O). GC of the sample dropwise, with stirring, until a yellow color persisted. The at 195 "C showed both possible isomers in a ratio of 6:l; the 'H chromium salta were filtered, most of the acetone was removed NMR spectrum indicated the principal isomer to be 5a. at reduced pressure, water was added, and the resulting solution Anal. Calcd for C23H404: C, 71.83; H, 11.52; mol wt, 384. was extracted with ether. Drying over sodium sulfate and removal Found: C, 71.79; H, 11.53; mol wt, 384 (MS). of solvent gave 368 mg of material. Thin-layer chromatographic Saponificationz1gave heptanoic acid and 2-pentyl-1,3-nonane- analysis indicated the presence of two compounds, which were diol (7); acetic acid was not recovered in the workup. separated by chromatography on silica gel into 73 mg (18%) of 2-Pentyl-3-[ (3-hydroxynonanoy1)oxy Inonyl heptanoate a carboxylic acid and 285 mg (76%) of neutral 2-pentyl-3-oxononyl (6a), presumably mixed with its positional isomer 6b, had the heptanoate (8): bp (Kugelrohr) 138-142 "C (0.1 torr); nuD 1.4440"; following: bp (Kugelrohr) 208-214 "C (0.006 torr); nz4D1.4532"; 'H NMR 6 0.9 (9 H, t, CHzCH3), 1.3 (24 H, m, CH,), 2.1-2.8 (5 IR 3450,1735 cm-'; 'H NMR 6 0.9 (12 H, t, CHzCH3),1.3 (37 H, H, m, CHzC=O, CHC=O), 4.1 (2 H, d, CHzOCO);IR 1735,1720 br s, RzCHz, R3CH), 2.3-2.4 (2 H, m, CHzC=O), 3.9 (1 H, m, cm-' (ester and ketone C=O). R&HOH), 4.0 (2 H, d, RCH,OC=O), 5.0 (1H, m, RzCHOC=O). Anal. Calcd for C21&&: C, 74.07; H, 11.84; mol wt, 340. Anal. Calcd for C30H5805: C, 72.24; H, 11.72; mol wt, 498. Found: C, 73.76; H, 11.96; mol wt, 340 (MS). Found: C, 71.72; H, 11.55; mol wt, 498 (MS). The acid (9a) was treated with diazomethane and characterized Saponificationz1 gave 2-pentyl-1,3-nonanediol (7) and two as ita methyl ester, methyl 2-pentyl-3-(heptanoyloxy)nonanoate carboxylic acids, which were separated by chromatography into (9b): bp (Kugelrohr) 143-147 OC (0.2 torr); nuD 1.4410'; 'H NMR heptanoic acid and 3-hydroxynonanoic acid, mp 58-59 "C (lit.z6 6 0.9 (9 H, t, CHzCH3),5.0 (1 H, m, RCHOCO); IR 1735 cm-' (ester 57-59 "C). C=O). Anal. Calcd for CZzH4204: C, 71.31; H, 11.42; mol wt, 370. Acknowledgment. This work was supported in part Found: C, 71.18; H, 11.42; mol wt, 370 (MS). by a grant from the National Institute of Allergy and In- 2-Pentyl-3-(bromoacetoxy)nonylheptanoate (4a) and its fectious Diseases (AI 04769). isomer (4b) could not be obtained analytically pure. GC of the sample at 245 "C showed both isomers, in a ratio of 3:1, principally Registry No. 1, 26257-80-7;2,3021-89-4; 2 DNP, 10385-38-3;3a, 4a, as established by the 'H NMR spectrum. The sample had 49562-88-1; 3b, 55109-59-6; 4a, 82352-12-3; 4b, 82352-13-4; 5a, 82352-14-5; 5b, 82374-04-7; 6a, 82352-15-6; 6b, 82352-16-7; 7, 55109-63-2;8, 82352-17-8;9a, 82352-18-9; 9b, 82352-19-0;heptanal, (28) Bowers, A,; Halsall, T. G.;Jones, E. R. H.;Lemin, A. J. J. Chem. 111-71-7; ethyl bromoacetate, 105-36-2; heptanoic acid, 111-14-8; SOC.1953, 2548-2560. 3-hydroxynonanoicacid, 40165-87-5.

Jatrophone Analogues: Synthesis of cis- and trans -Normethyljatropholactones

Amos B. Smith, III,*l and Michael S. Malamas Department of Chemistry, The Laboratory for Research on the Structure of Matter and The Monell Chemical Senses Center, The University of Pennsylvania, Philadelphia, Pennsylvania 19104 Received February 10, 1982

This, a full account, discloses an efficient, convergent synthesis of two novel analogues of the macrocyclic antitumor diterpene jatrophone (1). We term these analogues cis- and trans-normethyljatropholactone(2 and 3, respectively). Our approach in each case begins with the bis(trimethylsily1oxy) ketone 7 and the requisite acetylenic or trans ester-aldehyde, 8 or 12a. Application of our previously developed 3(2H)-furanone synthetic protocol consisting of aldol condensation of the lithium enolate derived from 7 with the respective estepaldehydes 8 or 12a, followed by oxidation (Collins reagent) and acid-catalyzed cyclization-dehydration, affords spirofuranone 6c and 14c, respectively, in 52% and 45% overall yields. Sodium borohydride reduction, ester hydrolysis, and closure of the macrolide by employing the conditions of Mukaiyama (i.e., 1-methyl-2-chloropyridiniumiodide- /Et3N/CH3CN) in the case of spirofuranone 14a leads directly to trans-normethyljatropholactone(3), while completion of cis-normethyljatropholactone (2) requires first semihydrogenation; the latter was accomplished by employing PdSOl in pyridine as the catalyst. The overall yields of 2 and 3, based on 7, were 23% and 21 %, respectively.

In connection with a synthetic program which recently of (f)-jatrophone its epimer (1b),2and (*)-nor- culminated in the sucessful stereocontrolled total synthesis methyljatrophone (l~),~we have prepared two architec- 0022-3263/82/ 1947-3442$01.25/0 0 1982 American Chemical Society cis- and trans-Normethyljatropholactones J. Org. Chem., Vol. 47, No. 18, 1982 3443

/ R’OOC 1 2 ’0 5 6 (a1 Jatrophone, R = wMe cis-Normethyljatropholactone (a) R = CH,OH; R’= H (b) Epijatrophone, R = p-Me (b) R = CH,OH; R’= Et (cl Normethyljatrophone, R = H (c) R = CHO; R’ = Et

3 4 / “xormethyljatropholactone (a) R =n-Pr Me,SiO EtOOC (b) R = protein 7 8 turally related analogues which we term cis- and trans- normethyljatropholactone (2 and 3, respectively). We This scenario calls initially for construction of the silyl record here a full account of the synthesis of these ana- protected keto alcohol 7 and acetylenic aldehyde 8, the logues. Our interest in such systems stemmed from the former prepared previously from a-(hydroxymethy1)- Kupchan observation4 that jatrophone, when treated in cyclopentenone (9).2bv6 The requisite aldehyde (8) was neutral or buffered solution with a variety of thiols, in- cluding cysteine hydrochloride and the free thiols of pro- teins such as bovine serum alubumin and DNA-dependent ’r.. RNA polymerase, led to the formation of unstable adducts possessing the general structure 4. This reactivity was II I HO/ I suggested by Kupchan to be responsible for the pro- R nounced antileukemic activity (P-388 lymphocytic leuke- mia) and cytotoxicity (KB cell cultures) displayed by ja- 9 10 trophone.28 It was anticipated that the availability of such (a) R = H; R‘= COOH synthetic analogues, in conjunction with jatrophone, epi- (b) R = H; R’=CH,OH jatrophone, and normethyljatrophone, would allow further (c) R = COOH; R’ = CH2OH exploration of their relative antitumor properties as well (d) R = COOEt; R‘= CH,OH as definition of the site(s) of reactivity with model biologic nucleophiles. prepared from readily available acetylenic acid loa.’ Results and Discussion Reduction with LiA1H4 to the corresponding alcohol lob, followed by generation of the dianion with 2.2 equiv of From the retrosynthetic perspective, acetylenic macro- n-BuLi in THF at -40 OC for 2 h and then addition of solid lide 5 appeared to be an ideal advanced intermediate from which both 2 and 3 could in turn be elaborated via the stereocontrolled reduction of the acetylenic linkage. (6)Initially, 2-(hydroxymethyl)-2-cyclopentenone(9) was prepared via Central to this strategy was the prospect of exploiting our the keto vinyl anion equivalent methodology developed in our and Swenton’s laboratory. See: Branca, S. J.; Smith, A. B., I11 J. Am. Chem. recently developed 3(W-furanone synthetic protocol5for SOC.1978,100, 7767. Smith, A. B.,111; Guaciaro, M. A.; Wovkulich, P. construction of spirofuranone 6a; subsequent macrocyclic M. Tetrahedron Lett. 1978,4661. Also see: Manning, M. J.; Raynolds, lactonization would then afford 5. P. W.; Swenton, J. S. J. Am. Chem. SOC.1976,98,5008. Raynolds, P.W.; Manning, M. J.; Swenton, J. S. J. Org. Chem. 1980,45,4467.An altemate more economical preparation of (9)is outlined below: unpublished results (1)Camille and Henry Dreyfus Teacher Scholar, 1978-1983; National of Mr. M. Malamas of our laboratory. Institutes of Health (National Career Institute) Career Development Awardee. 0 0 11 IlI LDA/THF.-7S°C II SePh (a) Structure: Kupchan, S. M.; Sigel, C. W.; Matz, M. J.; Saenz (2) PhSeBr O,ICH,CI,, -780 c Renauld, R. C.; Haltiwanger, R. C.; Bryan, R. F. J. Am. Chem. SOC.1970, I21 - 92,4476. Kupchan, S. M.; Sigel, C. W.; Matz, M. J.; Gilmore, C. J.; Bryan, 131 LDA/THF, -780 C *OH R. F. Zbid. 1976,98,2295 (b) Synthesis: Smith, A. B., 111; Schow, S. R.; 141 HCHO Guaciaro, M. A.; Wovkulich, P. M.; Toder, B. H.; Malamas, M.; Hall, T. W., unpublished results. 0 (3)Smith, A. B., III; Guaciaro, M. A.; Schow, S. R.; Wovkulich, P. M.; Toder, B. H.; Hall, T. W. J. Am. Chem. SOC.1981,103, 219. (4)Kupchan, M. S.; Giacobbe, T. J.; Krull, I. S. Tetrahedron Lett. 1970, 2859. Lillehaug, J. R.: Kleuue. K.: Sigel. C. W.:.- Kuwhan, S. M. 9 Biochem. Biophys. A& 1973,3if 92. ’ - ’ (5)Smith, A. B., III; Levenberg, P. A.; Jerris, P. J., Scarborough, R. (7) Behrens, 0. K., Corse, J.; Huff, D. E.; Jones, R. G., Soper, Q. F.; M., Jr.; Wovkulich, P. M. J. Am. Chem. SOC.1981, 103, 1501. Whitehead, C. W. J. Bid. Chem. 1948, 175, 771. 3444 J. Org. Chem., Vol. 47, No. 18, 1982 Smith and Malamas

Table I. High-Field (250 MHz) 'H NMR Data for cis- and trans-Normethyljatropholactone(2 and 3) and (&E)- and (E,Z)-Normethyljatrophone(IC and 11, Respectively)' chemical shift

Hb 2 3 IC 11 1-Ha 2.37 (ddd, 13.4, 8.2, 6.0) 2.29 (ddd, 14.1, 7.1, 6.7) 2.15 (dd, 13.1, 8.1) 2.35 (m) 1-Hh 2.13 (ddd. 13.4. 7.4. 3.0) 2.04 (ddd, 14.1, 7.1. 3.3) 1.85 (dd. 13.1. 6.4) 2.10 (m) 2-HI 2.62 (m) 2.58 (m) 2.4 (m) 2.40 (m) 2-HL 2.62 (mj 2.58 (m) 2.2 (m) 2.40 (m) 3-H 6.14 (br s) 6.55 (br s) 5.8 (m) 6.08 (m) 5-HaC 4.98 (dq, 13.9, 1.1) 4.72 (dq, 13.8, 1.3) 5.67. (m) 6.94 (m) 5-HbC 4.65 (dq, 13.9, 2.3) 4.27 (d, 13.8) 7 -H 5.72 (d, 13.6) 5.91 (d, 16.1) 5.86 (d, 16.4) 5.62 (d, 14.0) 8-H 5.84 (d, 13.6) 6.53 (d, 16.1) 6.35 (d, 16.4) 5.68 (d, 14.0) 1%Ha 2.25 (d, 13.6) 2.31 (d, 12.9) 2.29 (d, 14.9) 2.32 (d, 13.2) 1O-Hb 2.78 (d, 13.6) 2.78 (d, 12.9) 2.77 (d, 14.9) 2.64 (d, 13.2) 15-CH,d 1.42 (s) 1.34 (s) 1.24 (s) 1.16 (s) 16-CH,d 1.30 (s) 1.21 (s) 1.13 (s) 1.24 (s) 17-CH3 1.64 (s) 1.75 (s) 1.62 (s) 1.70 (s) Chemical shifts are iven in parts per million downfield from (CH,),Si (a), with multiplicities and coupling constants (in hertz) in parentheses. % Numbering for ICand 11 has been altered so that corresponding carbons in 2, 3, IC, and 11 have the same number. Assignments for 5-H, could not be made with confidence in regard to stereochemistry. Assignments for 15-CH3 and 16-CH, may be reversed in 2, 3, and 11.

Table 11. High-Field (62.9 MHz) NMR Data for cis- methanol-water (2:l) gave acid 6a in 77.4% yield for the and trans-Normethyljatrolactone(2 and 3) and two steps. Lactonization employing Mukaiyama's proce- Normethyljatrophone (1~)~ dure'O (i.e., 1-methyl-2-chloropyridiniumiodide/Et3N/ chemical shift CH3CN, high dilution) then afforded the pivotal inter- mediate, macrolide 5, in 58% yield. carbon bfd 2 3 IC At this point all that remained to complete the synthesis 34.2 35.4 33.5 of cis- and trans-normethyljatropholactonewas the ste- 30.6 30.6 30.4 122.2 117.9 123.1 reoselective reduction of the C(7,8) acetylenic linkage. As 135.6 136.4 137.8 anticipated, semihydrogenation employing PdS04 in 61.1 62.5 140.6 pyridine" afforded cis-normethyljatropholactone (2) as a 165.6 168.8 201.1 crystalline solid, mp 101-103 OC (ether-hexane). That in 135.8 144.8 128.2 fact 2 was in hand derives from careful comparison of the 146.6 152.4 158.7 high-field 'H and I3C NMR spectra of 2 with those of la-c, 38.8 39.6 36.3 40.4 42.8 40.9 as well as with that of 11, the latter prepared in connection 184.3 184.1 182.8 with our normethyljatrophone synthesis (see Tables I and 112.1 112.4 112.0 II).~ 204.8 205.1 203.7 98.5 98.7 98.8 34.5 29.0 30.0 26.4 24.0 26.5 6.24 6.55 5.66 Chemical shifts are reported in parts per million down- field from (CH,),Si (6 ). Numbering for IChas been altered so that corresponding carbons in 2, 3, and IC have the same number. Assignments for carbons 15 and 16 ' 11 in 2 and 3 may be reversed. Multiplicity is given in 12 paren theses. (a) R = CHO; R'= Et COP,afforded after an acidic workup acid 1Oc in 80% yield. (b) R = CH,OH; R'= H This acid was then esterified with EtOH and oxidized with (c) R = CH,OH; R'= Et pyridinium chlorochromateato afford the desired aldehyde 8 in 94% yield. Turning next to the conversion of macrolide 5 to With ample quantities of both 7 and 8 in hand, we ex- trans-normethyljatropholactone (3), we initially antici- ecuted the 3(2H)-furanonesynthetic protocol. To this end, pated exploiting the well-known propensity of chromous the lithium enolate of 7, generated via addition of 1.3 equiv sulfate to effect the stereoselective trans reduction of a- of LDA in THF at -78 "C, was condensed with aldehyde acetylenic acids and esters.12 This one-step plan, however, 8 in THF. Without purification the derived aldol was was thwarted presumably due to transannular reaction^.^ subjected first to Collins oxidation (ca. 13 eq~iv)~and then Undaunted by lack of initial success, we considered next to acid-catalyzed cyclization-dehydration to afford spiro- isomerization of cis-normethyljatropholactone (2) to the furanone 6c in 51% overall yield. desired trans isomer. Such a strategy, of course, was not With an efficient, convergent approach to 6c available, the stage was next set for macrocyclic ring closure. Re- duction of 6c with 1.1 equiv of NaBH, in MeOH at -23 (10)Mukaiyama, T.;Usui, M.; Saigo, K. Chem. Lett. 1976, 49. (11) Cram, D. J.; Allinger, N. L. J. Am. Chem. SOC.1956, 78, 2518. "C for 10 min followed by hydrolysis with KzC03 in Also see private communication to L. Fieser: Fieser, L.; Fieser, M. 'Reagents for Organic Synthesis"; Wiley: New York, 1967; p 566. (12) Castro, C. E.;Stephens, R. D. J. Am. Chem. SOC.1964,86,4358. (8) Corey, E. J.; Suggs, J. W. Tetrahedron Lett. 1975, 2647. Inanaga, J.; Kataski, T.; Takino, S.; Ouchida, S.; Inoui, K.; Nakano, A.; (9)Collins, J. E.; Hess, W. W.; Frank, F. 3. Tetrahedron Lett. 1968, Okudako, N.; Yamaguchi, M. Chem. Lett. 1979,102. For preparation of 3363. this reagent see: Castro, C. E. J. Am. Chem. SOC.1961, 83, 3262. cis- and trans-Normethyljatropholactones J. Org. Chem., Vol. 47, No. 18, 1982 3445 novel; it was this approach that we had exploited to great presented in due course. advantage in our jatrophone synthesi~.~~~~However, to our dismay all attempts to effect the required isomerization Experimental Section by employing numerous acids, light (with and without Iz), Materials and Methods. Melting points were taken on a and/or transition-metal catalysts under a wide variety of Thomas-Hoover capillary melting apparatus and are corrected. Boiling points are uncorrected. All solvents were distilled prior time, temperature, and solvent regimes proved fruitless. to use. Solutions were dried over MgSO,. Unless specified Our only alternative at this point, aside from abandoning otherwise, IR and 'H NMR spectra were obtained for CC14 so- the goal of tram-normethyljatropholacbne, was to prepare lutions, the former on a Perkin-Elmer Model 337 spectropho- trans aldehyde ester 12a and then to subject this inter- tometer and the latter on either a Varian A-60A (60 MHz), a mediate to the 3(2H)-furanone synthetic protocol. Toward Varian T60A (60 MHz), or a Bruker WP 250 (250 MHz) spec- this end, reduction of the previously prepared hydroxy acid trometer. 13C NMR spectra were obtained in CDCl, on either 1Oc with aqueous chromous sulfate for 24 h at room tem- a JEOL PS-100 (25 MHz) or a Bruker WP-250 FT (62.9 MHz) perature proceeded as anticipated to yield 12b in 78% as spectrometer; signal multiplicities were determined via off- res- the sole product. Characteristic of the desired trans ole- onance decoupling. The internal standard for 'H and 13C NMR finic geometry, the 250-MHz NMR spectrum displayed experiments was Me4Si. For those compounds containing the TBDMS protecting group, an external reference of either Me& two doublets at 6 5.95 and 6.50 having a coupling constant or CHC13 (6 7.24) for CDC13 solutions was used. High-resolution of 16.2 Hz. Fischer e~terificationl~then gave alcohol 12c mass spectra were obtained at the University of Pennsylvania in 92% yield, contaminated with 8% of tetrahydrofuran Mass Spectrometry Service Center on a Hitachi Perkin-Elmer 13, the latter arising via a facile 5-exo-trigonal cyclization1* RMH-2 high-resolution double-focusing electron-impact spec- of ester 12c.15 Subsequent Collins oxidation of 12c then trometer or a VG micromass 70/70H high-resolution double-fo- led to aldehyde 12a. cusing electron impact-chemical ionization spectrometer, the latter using isobutane as the reagent gas and each interfaced with a Kratos DS-504 data system. Preparative thin-layer chroma- tography (TLC) was performed on 500- or 1000-fimprecoated silica gel plates with fluorescent indicator supplied by Analtech, Inc. Visualization was accomplished with UV light. Flash column chromatography and medium-pressure liquid chromatography were performed with silica gel 60 (0.04-0.063 mm) supplied by E. M. Merck. R'OOC Preparation of 3,3-Dimethyl-4-pentyn-l-ol(lob). To a 13 14 mixture of 100 mL ether and 3.056 g (80.52 mmol) of lithium aluminum hydride under a nitrogen atmosphere at 0 "C was added (a) R = CH,OH; R'= H 5.06 g (40.16 mmol) of 3,3-dimethyl-4-pentynoicacid in 30 mL (b) R = CH,OH; R'= Et of ether over a 30-min period. The mixture was stirred for 2 h (c) R = CHO; R'= Et after the addition. Solid sodium sulfate decahydrate was added to quench the reaction, and the resultant solid was filtered and Executioq of the 3(2H)-furanone synthetic protocol [(a) washed thoroughly with ether. Concnetrated in vacuo afforded aldol condensation, (b) Collins oxidation, (c) cyclization- 4.28 g of a colorless oil, which was distilled (Kugelrohr, -3 mm, 49 "C) to give 4.21 g (93.5%) of a colorless oil: IR (CHC13)3635, dehydration] employed above led to spirofuranone 14c as 3430 (br), 2980,2880 (s), 2120 (w), 1260 (br), 1065,1095 (w), 630 in 45% overall yield. Reduction of the aldehyde func- (w), cm-'; NMR (250 MHz, CDC13) 6 1.26 (s, 6 H), 1.72 (t, J = tionality sodium borohydride in methanol (14c with - 7 Hz, 2 H), 2.16 (6, 1 H), 2.30 (br s, 1 H), 3.85 (t,J = 7 Hz, 2 H); 14b) followed by saponification and macrolactonization of mass spectrum, m/e 112.0823 (M'; calcd for C7Hlz0112.0888). the resultant hydroxy acid (14a) again a lb Mukaiyama Preparation of Ethyl 4,4-Dimethyl-6-hydroxy-2-hexynoate afforded tram-jatropholactone (3) as a crystalline solid [mp (loa). To a mixture of 250 mL of THF and 12.47 g of acetylenic 168-170 "C (ether-hexane)], the overall yield from spiro- alcohol 10b (111.3 mmol) under a nitrogen atmosphere at -42 "C furanone 14c being 42%. Indicative of the trans geometry (acetonitrile/COz) was added dropwise 131 mL (256.15 mmol) the olefinic resonances display as two doublets (6 5.91 and of n-BuLi. The mixture was stirred 2 h at -42 "C after completion 6.53) with the characteristic trans coupling of 16.1 Hz. of the n-butyllithium addition. Solid was then added to the solution, and the thick solution was stirred for 0.5 In summation, an economic convergent synthesis of two h followed by acidification with 6 N hydrochloric acid. The analogues of jatrophone has been achieved in 20% and aqueous layer was extracted four times with ether and the solution 19% yields (2 and 3, respectively) based on 7. Biological dried. Concentration in vacuo gave 15.7 g of a colorless oil that screening data as well as NMR studies to correlate the crystallized upon standing; the latter was employed in the next solution conformation vis-&vis the bioactivity will be step without purification. A mixture of 100 mL of absolute ethanol, 6.8 g (43.58 mmol) of hydroxy acid lOc, and 1 mL of concentrated was (13) Fischer, E. J. Physiol. Chem. 1901, 33, 151. stirred at room temperature for 64 h. Dilution in water and (14) Baldwin, j. E. J. Chem. SOC.,Chem. Commun. 1976,734. Bald- extraction with ether afforded 7.95 g of colorless oil. Purification win, J. E.; Cutting, J.; Dupont, W.; Kruse, L.; Silberman, L.; Thomas, R. C. Ibid. 1976,736. Baldwin, J. E. Zbid. 1976,738. Baldwin, J. E.; Thomas, via flash column chromatography (5:l hexane/ethyl acetate) R. C.; Kruse, K. I.; Siverman, L. J. Org. Chem. 1977, 42, 3846. afforded 7.86 g (98%) of colorless oil: IR (CHCl,) 3630-3450 (br), (15) Interestingly,the propensity with which the analogous hydroxy 2980, 2930 (s), 2210 (s), 1690 (s), 1260 (br), 905 (w) cm-'; NMR enone i displayed for this 5-exc-trigonal cyclization (i - ii) precluded its (250 MHz, CDCl3) S 1.30, (9, 6 H), 1.30 (t,J = 7.0 Hz, 3 H), 1.78 utilization as an intermediate in our jatrophone approach unpublished (t, = 7.25 Hz, 2 H), 2.79 (br s, 1 H), 3.83 (t,J = 7.25 Hz, 2 H), results of Dr. S. R. Schow of this laboratory. J 4.23 (q, J = 7.0 Hz, 2 H);mass spectrum, m/e 184.1100 (M+;calcd for C10H1603 184.1100). Anal. Calcd for C10H1603: C, 65.20; H, 8.70. Found: C, 65.04; H, 8.79. Preparation of Ethyl 4,4-Dimethyl-6-oxo-%-hexynoate(8). To a mixture of 160 mL dry methylene chloride and 26.35 g (122.28 mmol) pyridinium chlorochromate under a nitrogen atmosphere at room temperature was added 9.02 g (49.02 mmol) of ester 10d in 50 mL of methylene chloride. The mixture was stirred for 3.5 i ii h at room temperature, diluted in 800 mL ether, and filtered 3446 J. Org. Chem., Vol. 47, No. 18, 1982 Smith and Malamas through Florisil(lO0-200 mesh). Removal of the solvent in vacuo of the solvent in vacuo gave 657 mg of a viscous yellow oil which and distillation (Kugelrohr, -1 mm, 110-120 "C) gave 8.47 g was purified via medium-pressure liquid chromatography [hex- (93.9%) of a colorless oil: IR (CHC13, 2980, 2930 (s), 2210 (m), ane/ethyl acetate (2:1)] to afford 409.1 mg (50.8% from 7) of 6c 1700 (s), 1260 (br), 1040 (br), 850 (w) cm-l; NMR (250 MHz, as a pale yellow oil: IR (CC14)2995, 2930 (s), 2210 (s), 1700 (s), CDC13) 6 1.31 (t, J, = 7.25 Hz, 3 H), 1.40 (9, 6 H), 2.54 (d, J = 1630 (s), 1250 (br), 1035 (m), 920 (s) cm-'; NMR (250 MHz, CDC13) 2.0 Hz, 2 H), 4.24 (4, J = 7.25 Hz, 2 H), 9.86 (t,J = 2.0 Hz, 1 H); 6 1.30 (t,J = 7.3 Hz, 3 H), 1.38 (s, 3 H), 1.39 (s, 3 H), 1.79 (s, 3 mass spectrum, m/e 182.0715 (M+;calcd for C10H1403 182.0709). H), 2.27 (m, 1 H), 2.44 (m, 1 H), 2.70 (s, 2 H), 2.80 (m, 2 H), 4.20 Preparation of trans -Ethyl 4,4-Dimethyl-6-hydroxy-2- (q, J = 7.0 Hz, 2 H), 7.33 (m, 1 H), 9.64 (s, 1 H); mass spectrum; hexenoate (12c). To a mixture of 46 mL of degassed water m/e 330.1403 (M'; calcd for C19H2205330.1467). (nitrogen purge for 2 h) and 4.2 g (26.92 mmol) of hydroxy acid Preparation of Spirofuranone 6b. To a mixture of 4 mL 1Oc was added under a nitrogen atmosphere at room temperature of methanol and 244 mg (0.74 mmol) of spirofuranone 6c under 229 mL of 0.4 M CrS04 over a 30-min period. The reaction a nitrogen atmosphere at -23 "C (CC14/C02)was added 27.97 mg solution first became green and then 10 min later turned dark. (0.74 mmol) of sodium borohydride. The mixture was stirred for The mixture was stirred for 12 h at room temperature. The dark 15 min at -23 "C before the pH was adjusted to 7 with dilute reaction solution was then basified with aqueous KOH and aqueous HC1. The mixture was extracted with ether, and the vacuum filtered from chromic hydroxide. The filtrate was next organic extract was dried and concentrated in vacuo to afford 219.6 acidified (concentrated H2SO4) to pH 1 and extracted with ethyl mg of a viscous colorless oil which was purified via medium- acetate. The organic extracts were dried over sodium sulfate, pressure liquid chromatography [hexane/ethyl acetate (2:1)] to filtered, and conentrated in vacuo to afford 3.31 g (78.6%) of 12b afford 203 mg (83.2%) of 6b as a colorless oil: IR 3500,3400 (br), as a colorless oil which was used without purification. 2980, 2930 (s), 2220 (m), 1700 (s), 1630 (s), 1260, 1100 (br), 1030 A mixture of 100 mL of absolute ethanol, 7.4 g (46.83 mmol) (m), 750 (br) cm-'; NMR (250 MHz, CDC13) 6 1.29 (t,J = 7.0 Hz, of 12b, and 0.5 mL of concentrated sulfuric acid was stirred at 3 H), 1.40 (9, 3 H), 1.43 (s, 3 H), 1.74 (s, 3 H), 2.21 (m, 1 H), 2.40 room temperature for 50 h. Dilution in water and extraction with (m, 1 H), 2.57 (m, 2 H), 2.71 (s, 2 H), 4.03 (m, 2 H), 4.20 (4, J ether afforded 8.53 g (98%) of a colorless oil. Purification via = 7.0 Hz, 2 H), 6.21 (t, J = 1.8 Hz, 1 H); mass spectrum, m/e flash column chromatography (101 hexane/ethyl acetate) afforded 332.1620 (M+; calcd for C19H24O5 332.1624). 7.89 g (90.6%) of 12c as a colorless oil: IR (CC14)3630,3400 (br), Preparation of Spirofuranone 6a. To a solution of 5 mL 2980,2930 (s), 1700,1650 (s), 1300,1200 (br), 1035 (br), 995 (w), of MeOH, 2 mL of H20and 170 mg (0.512 "01) of spirofuranone 870 (w) cm-'; NMR (250 MHz, CDClJ 6 1.1 (s, 6 H), 1.30 (t, J 6b at room temperature was added a concentrated aqueous K2C03 = 7.0 Hz, 3 H), 1.71 (t, J = 7.75 Hz, 2 H), 2.30 (9, 1 H), 3.62 (t, solution, adjusting the pH of the reaction mixture to 10.5-11.0. J = 7.75 Hz, 2 H) 4.21 (9, J = 7.0 Hz, 2 H), 5.76 (d, J = 16.2 Hz, The mixture was then stirred 15 h at room temperature under 1 H) 6.95 (d, J = 16.2 Hz, 1 H); mass spectrum, m/e 186.1311 nitrogen and diluted with 5 mL of H20,and the pH was adjusted (M+; calcd for C10H18O3 186.1290). to 2. The mixture was extracted extensively with ether, and the Anal. Calcd for Cl0Hl8O3:C, 64.51; H, 9.67. Found C, 64.42; organic fraction was dried and concentrated in vacuo to afford H, 9.69. 144.7 mg (93.0%) of acid 6a: IR 3450,3300,2500 (br),2980,2930 Preparation of Ethyl trans-4,4-Dimethyl-6-0~0-2-hex- (s), 2220 (m), 1700, 1630 (s), 1450 (s), 1200 (s),700 (br) cm-'; NMR enoate (12a). To a mixture of 120 mL of methylene chloride and (250 MHz, CDC13) 6 1.37 (9, 3 H), 1.41 (s, 3 H), 1.70 (s, 3 H), 2.05 19.03 g (88.28 mmol) of pyridinium chlorochromate under a ni- (m, 1 H), 2.40 (m, 2 H), 2.61 (m, 2 H), 2.81 (d, J = 13 Hz, 1 H), trogen atmosphere at room temperature was added 6.6 g (34.94 4.05 (m, 2 H), 5.75 (br s, 1 H), 6.20 (t, J = 1.8 Hz, 1 H), COOH mmol) of ester 12c in 40 mL of methylene chloride. The mixture not observed. was stirred for 3.5 h at room temperature, diluted in 700 mL of Preparation of the Macrolide 5. To a solution consisting ether, and filtered through Florisil(100-200 mesh). Removal of of 767 mg (3.0 mmol) of 1-methyl-2-chloropyridiniumiodide in solvent in vacuo and distillation (Kugelrohr, -1 mm, 115-130 73 mL of dry acetonitrile held at reflux was continuously and OC) gave 6.07 g (93%) of 12a as a colorless oil: IR (CC14)2980, uniformly added a solution of 230 mg (0.75 mmol) of 6a and 0.84 2930 (s), 1700 (s), 1250 (br), 1030 (br) 850 (w) cm-', NMR (250 mL (6 mmol) of triethylamine in 65 mL of dry acetonitrile over MHz, CDCl3) 6 1.25 (9, 6 H), 1.32, (t, J = 7.0 Hz, 3 H), 2.48 (d, a period of 9 h. After one additional hour at reflux, evaporation J = 2.1 Hz, 2 H), 4.24 (4,J = 7.0 Hz, 2 H), 5.81 (d, J = 16.2 Hz, of the solvent under reduced pressure afforded a residue which 1 H), 6.90 (d, J = 16.2 Hz, 1 H), 9.75 (t,J = 2.1 Hz, 1 H); mass was separated via silica gel column chromatography to afford 125.6 spectrum, m/e 184.1093 (M+ calcd for C10H16O3 184.1100). mg (57.8%) of 5 as a viscous colorless oil that crystallized upon Preparation of Spirofuranone (6c). To a solution of 0.62 standing: mp 125-127 "C; IR (CHCl,) 2980, 2930 (s), 220 (m), mL (4.38 mmol) of diisopropylamine in 3 mL of THF containing 1700, 1630 (s), 1500,1400 (m), 1180,1200 (br) cm-'; NMR (250 a few crystals of 2,2'-dipyridylamineas an indicator at 0 "C under MHz, CDC1,) 6 1.45 (s, 3 H), 1.49 (s, 3 H) 1.78 (s, 3 H), 2.20 (m, a nitrogen atmosphere was added 2.40 mL (3.41 mmol, 1.42 M 1 H), 2.41 (m, 1 H), 2.60 (d, J = 13.2 Hz, 1 H), 2.65 (m, 2 H) 2.81 solution) of n-butyllithium. The resultant deep red solution was (d, J = 13.2 Hz, 1 H) 4.64 (s, 2 H), 6.47 (t, J = 1.9 Hz, 1 H); mass stirred 30 min at 0 "C and then cooled to -78 "C. After 5 min spectrum, m/e 286.1201 (M+;calcd for C17H1804 286.1205). bis(sily1oxy) ketone 7 (765 mg, 2.44 mmol) in THF (3 mL) was Preparation of cis-Normethyljatropholactone (2). A added over a period of 5 min. The red mixture was stirred 2 h, suspension of 10 mg of 5% palladium on barium sulfate in 2 mL and then 896 mg (4.92 mmol) of aldehyde 8 in mL of THF was of pyridine was stirred under hydrogen at atmospheric pressure added over a 50-9 period. The color of the reaction mixture for 30 min, whereupon 153 mg of macrolide 5 in 200 mL of changed to orange-brown; stirring was continued for 2 min, and pyridine was added and the stirring under hydrogen continued then saturated ammonium chloride (2 mL) was added. The for 15 min. The reaction mixture was then filtered through Celite mixture was diluted with 3 mL of water and extracted thoroughly with a CH2C12wash. Removal of the solvent in vacuo yielded 2 with ether. The combined extracts were washed with water and as a viscous yellow oil that crystallized upon standing. Recrys- brine. Removal of solvent in vacuo gave 1.34 g of yellow oil which tallization from hexane/ethyl ether (1O:l) gave 144.7 mg (94.6%) was employed in the next step without purification. of a white crystalline solid: mp 101-103 OC; IR (CHC13) 2980, To a solution of 7.49 mL (92.61 mmol) of pyridine in 100 mL 2930 (s), 1700,1620 (s), 1220 (br), 1090 (m), 715 (br) cm-'; NMR of methylene chloride was added 4.41 g (44.1 mmol) of chromium (250 MHz, CDC13) 6 1.30 (9, 3 H), 1.42 (s, 3 H), 1.64 (s, 3 H), 2.20 trioxide (CrOJ. After the mixture was stirred 20 min at room (m, 1 H), 2.28 (d, J = 13.6 Hz, 1 H), 2.35 (m, 1 H), 2.60 (m, 2 H), temperature under a nitrogen atmosphere, the above hydroxyl 2.75 (d, J = 13.6 H3, 1 H), 4.71 (m, 1 H), 5.05 (m, 1 H), 5.72 (d, ketone (1.34 g) in 5 mL of CH2C12was added and the stirring J = 13.3 Hz, 1 H), 5.85 (d, J = 13.3 Hz, 1 H), 6.14 (t,J = 1.5 Hz, continued for 3 h. The organic layer was decanted from the black 1 H); mass spectrum, m/e 288.1351 (M+; calcd for C17H2004 residue and the latter washed with ether. Conventional workup 288.1361). and removal of solvent in vacuo gave 1.09 g of a light brown oil, Anal. Calcd for C17H&: C, 70.80; H, 6.95. Found C, 70.60; which was taken up in 50 mL THF with 25 mL of 10% HCl and H, 7.03. stirred at room temperature under a nitrogen atmosphere for 48 Preparation of Spirofuranone 14c. To a solution of 0.45 h. The mixture was then saturated with solid NaCl and diluted mL (3.27 "01) of diisopropylamine in 2.5 mL of THF containing with ether. The organic layer was washed and dried. Removal a few crystals of 2,2'-dipyridylamine as indicator at 0 "C under J. Org. Chem. 1982,47, 3447-3450 3447 a nitrogen atmosphere was added 1.88 mL (2.54 mmol, 1.35 M 5.75 (d, J = 16.3 Hz, 1 H), 6.20 (br s, 1 H), 7.03 (d, J = 16.3 Hz, solution) of n-butyllithium. The resultant deep red solution was 1 H); mass spectrum, m/e 334.1746 (M'; calcd for CI9Hz6O5 stirred 30 min at 0 "C and cooled to -78 "C. After 5 min bis- 334.1781). (silyloxy) ketone 7 (507.2 mg, 1.62 mmol) in 3 mL of THF was Preparation of Spirofuranone 14a. To a solution of 8 mL added over a period of 5 min. The red mixture was stirred 2 h, of MeOH, 2 mL of HzO,and 250 mg (0.748 "01) of spirofuranone and then 675.5 mg (3.67 mmol) of aldehyde 12a in 1 mL of THF 14b at room temperature was added a concentrated aqueous was added over a 50-s period. The color of the reaction mixture K2C03 solution, adjusting the pH of the reaction mixture to changed to orange-brown; stirring was continued for 2 min, and 10.5-11.0. The mixture was stirred 12 h at room temperature then saturated ammonium chloride (2 mL) was added. The under nitrogen and then diluted with 5 mL of HzO, and the pH mixture was diluted with 3 mL of water and extracted thoroughly was adjusted to 7. The mixture was extracted extensively with with ether. The combined organic extracts were washed with ether, and the organic fraction was dried and concentrated in water and dried. Removal of the solvent in vacuo gave 767 mg vacuo to afford 231.3 mg (91.5%) of 14a: IR (CHC1,) 3450-2500 of yellow oil which was employed in the next step without pu- (br) 2980-2930 (s) 1700 (s), 1450 (br), 1080 (m), 850 (m)cm-'; rification. NMR (250 MHz, CDClJ 6 1.15 (9, 3 H), 1.20 (9, 3 H) 1.65 (5, 3 To a solution of 5.6 mL (68.6 mmol) of pyridine in 65 mL of H), 2.1 (m, 1 H), 2.35 (m, 2 H), 2.55 (m, 2 H), 2.70 (m, 1 H), 3.59 methylene chloride was added 3.25 g (32.5 mmol) of chromium (br s, 1 H), 3.95 (m, 2 H), 5.18 (br I, 1 H), 5.75 (d, J = 16.2 Hz, trioxide (&Os). After the mixture was stirred 20 min at room 1 H), 6.10 (e, 1 H), 6.50 (d, J = 16.2 Hz, 1 H). temperature under a nitrogen atmosphere, the above hydroxy Preparation of trans-Normethyljatropholactone(3). To ketone (767 mg) in 5 mL of CHzClz was then added and the a solution consisting of 654 mg (2.55 mmol) of 1-methyl-2- stirring continued for 3 h. The organic layer was decanted from chloropyridinium iodide in 65 mL of acetonitrile held at reflux the dark residue, and the latter was washed with ether. A con- was continuously and uniformly added a solution of 195 mg (0.64 ventional workup and removal of solvent in vacuo gave 585 mg mmol) of spirofuranone 14a and 0.72 mL (5.12 mmol) of tri- of a light brown oil, which was taken up in 40 mL of THF with ethylamine in 55 mL of dry acetonitrile over a period of 9 h. After 20 mL of 5% aqueous HCl and stirred at room temperature under one additional hour at reflux evaporation of the solvent under a nitrogen atmosphere for 5 days. The mixture was then saturated reduced pressure followed via silica gel column chromatography with solid NaCl and diluted with ether. The organic layer was afforded 138 mg of a viscous colorless oil that crystallized upon washed and dried, and the solvent was removed in vacuo to give standing. Recrystallization [hexane/ethyl ether (lOl)]gave 104.0 355 mg of a viscous yellow oil, which was purified via medium- mg (56.8%) of 3 as a white crystalline solid: mp 168-170 "C;IR pressure liquid chromatography [hexane/ethyl acetate (2:1)] to 2980,2930 (s), 1700,1630 (s), 1460 (m), 1210 (m), 720 (br) cm-l; afford 241.3 mg (45% from 7) of 14c as a pale yellow oil: IR NMR (250 MHz, CDCLJ 6 1.21 (5, 3 H), 1.34 (5, 3 H), 1.75 (5, (CHC13) 2980, 2930 (s), 1700, 1630 (s), 1200, 1300 (br), 1040 (s), 3 H), 1.15 (m, 1 H) 2.28 (m, 1 H), 2.33 (d, J = 13 Hz, 1 H), 2.60 920, 840 (m) cm-l; NMR (250 MHz, CDC13) 6 1.12 (s, 6 H), 1.20 (m, 2 H), 2.80 (d, J = 13 Hz, 1 H), 4.31 (d, J = 14 Hz, 1 H), 4.75 (t, J = 7.25 Hz, 3 H), 1.66 (s, 3 H), 2.10 (m, 1 H), 2.35 (m, 1 H), (m, 1 H), 5.95 (d, J = 16.1 Hz, 1 H), 6.50 (d, J = 16.1 Hz, 1 H) 2.51 (d, J = 10.25 Hz, 2 H), 2.81 (m, 2 H), 4.20 (4, J = 7.25 Hz, 6.51 (t,J = 1.5 Hz, 1H); mass spectrum, m/e 288.1359 (M'; calcd 2 H), 5.75 (d, J = 16.2 Hz, 1 H), 6.94 (d, J = 16.2 Hz, 1 H), 7.20 for C17H20O4 288.1362). (t, J = 1.5 Hz, 1 H), 9.61 (s, 1 H); mass spectrum, m/e 332.1615 (M+; calcd for ClgH2405 332.1693). Acknowledgment. It is a pleasure to acknowledge the Preparation of Spirofuranone 14b. To a mixture of 4-mL support of this investigation by the National Institutes of of methanol and 290 mg (0.873 mmol) of spirofuranone 14c under Health (National Cancer Institute) through Grant Ca- a nitrogen atmosphere at -23 "C (CC14/C02)was added 33.78 mg 22807. In addition, we thank Mr. S. T. Bella of the (0.893 mmol) of sodium borohydride. The mixture was stirred Rockefeller University for the microanalyses and Drs. G. for 15 min at 23 "C before the pH was adjusted to 7 with dilute Furst and T. Terwilliger of the University of Pennsylvania aqueous HCl. The mixture was extracted extensively with ether. Spectroscopic Service Centers for aid in recording and The organic material was then dried and concentrated in vacuo interpertation of the high-field NMR and mass spectra, to afford 275.3 mg of a viscous colorless oil which was purified via medium-pressureliquid chromatography [hexane/ethyl acetate respectively. (21)] to give 231.3 mg (79.5%) of 14b as a colorless oil: IR (CHC13) Registry No. 2, 82351-40-4; 3, 82398-40-1; 5, 82351-41-5; 6a, 3450,3400 (br), 2980,2930 (s), 1700,1625 (s), 1480 (m), 1200 (br), 82351-42-6; 6b, 82351-43-7;6~, 82351-44-8; 7, 76445-18-6; 8, 82351- 1040 (s) cm-'; NMR (250 MHz, CDClJ 6 1.15 (s, 6 H), 1.22 (t, 45-9; loa, 67099-40-5; lob, 67099-41-6; lOc, 82351-46-0; 10d, 82351- J = 7.8 Hz, 3 HI, 1.72 (s, 3 H), 2.10 (m, 1 H), 2.35 (m, 1 H), 2.60 47-1;12a, 82351-48-2;12b, 82351-49-3; 1212, 82351-50-6; 14a, 82351- (br m, 4 H), 3.96 (br d, J = 2 Hz, 2 H), 4.22 (9, J = 7.8 Hz, 2 H), 51-7; 14b, 82351-52-8; 14~,82351-53-9.

Cobalt-Mediated [2 + 2 + 21 Cycloadditions En Route to Natural Products: A Novel Total Synthesis of Steroids via the One-Step Construction of the B,C,D Framework from an A-Ring Precursor

Ethan D. Sternberg and K. Peter C. Vollhardt* Department of Chemistry, University of California, and the Materials and Molecular Research Division, Lawrence Berkeley Laboratory, Berkeley, California 94720 Received February 9, 1982

The first application of the cobalt-mediated intramolecular cyclization of a&-diynenes to annulated cy- clohexadienes in natural product synthesis is described by demonstrating its feasibility in a versatile and efficient steroid synthesis, including a new total synthesis of the Torgov intermediate, 3-methoxyestra-l,3,5(10),8,14- pentaen-17-one,via a new steroid, 3-methoxyestra-1,3,5(10),8(14),9-pentaen-17-oneethylene ketal. Several model reactions en route to B-homo-7-oxa steroids allow the delineation of some stereochemical details of the tran- sition-metal-catalyzed [2 + 2 + 21 cycloaddition reaction.

We have recently developed methodology based on co- substrates which yields annulated and complexed five-l balt-mediated [2 + 2 + 21 cycloadditions of unsaturated and six-membered2 rings. We believe that this strategy 0022-3263/82/1947-3447$01,25/00 1982 American Chemical Society Published on Web 12/11/2003

A Formal Synthesis of (-)-Mycalamide A Barry M. Trost,* Hanbiao Yang, and Gary D. Probst Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080 Received September 29, 2003; E-mail: [email protected]

The mycalamides,1 onnamides,2 and theopederins3 are biologi- Scheme 1. Retrosynthetic Analysis cally and synthetically interesting natural products isolated from marine sponges. Many of these compounds possess potent antiviral and antitumor properties due to their ability to arrest protein synthesis.4 The inhibition of protein synthesis is accomplished by binding to the 80S ribosome and preventing the translocation of the nascent peptide from the A site to the P site. One of the most remarkable aspects of mycalamides A and B is their ability to change the morphology of ras-transformed rat NRK-cells back to normal cells by selectively inhibiting the biosynthesis of p21, a G protein.5 These natural products have attracted a great deal of attention from the synthetic cmmunity. Total synthesis of mycal- amide A,6a,7,8 mycalamide B,6a,9 onnamide A,6b and theopederin D9 have been recorded. In addition, studies toward their total synthesis have been reported by other groups.10 Since these molecules are - a so complex, it is not surprising that the previous total syntheses Scheme 2. Synthesis of ( )-7-Benzoylpederic Acid typically require around 30 linear steps to complete. Through the use of synthetic methodologies developed in these laboratories, we sought to streamline the synthesis. We focused our efforts on an asymmetric synthesis of (-)-7- benzoylpederic acid 2 and the azide 3 since Nakata et al.7b reported their conversion to mycalamide A in three steps (Scheme 1). By envisioning alkene 4 as a flexible building block to several members of these families including mycalamide A, a novel route for its construction was developed that involves two Pd(0)-catalyzed O-π- allyl cyclizations and a novel strategy to create 1,3-dioxan-4-ones. All of the stereochemistry on the trioxadecalin ring derives from either R-orS-pantolactone. Our strategy to synthesize acid 2 is a - ° based on the recently developed Ru-catalyzed alkene-alkyne Conditions: (a) TMS acetylene, n-BuLi, 78 C, 20 min; then Me3Al, -78 °C, 30 min, then -45 °C, 30 min; then 10, -78 °C, 15 min; then BF 11 3 coupling reaction. The stereochemistry of the pederic acid derives etherate, -78 °C, 1 h, quant. (b) 8, CpRu(CH3CN)3PF6 (10 mol %), acetone, from that of the initial chiral trans-2-butene epoxide, which is rt, 63%. (c) TBDMSOTi, Et3N, CH2Cl2, quant. (d) OsO4, NMO, acetone/ ° commercially available. While our synthesis targeted the enanti- H2O, 2 C, 96% (1/1 dr). (e) DMAP (5 mol %) Et3N, PhCOCl, CH2Cl2, - - ° ° omer, the natural series is equally accessible simply by using the 78 to 40 C, 70% (96% after f). (f) NaOCH3, MeOH, 0 C to room temperature. (g) Dess-Martin Periodianene, NaHCO ,CHCl ; then CSA mirror-image starting materials. 3 2 2 (10 mol %), MeOH, 81%, (h) silica gel. (i) CH3OTf, LHMDS, DME, -70 The synthesis of 2 commenced with epoxide 10 (Scheme 2). °C, 87%. (i) NIS, CH3CN, 0 °C, 91%. (k) (Ph3P)4Pd, Bu3SnH, 96%. (l) While the opening of 10 with the Yamaguchi protocol (lithium n-PrSLi, HMPA, rt, 94%. 12 trimethylsilylacetylide and BF3‚Et2O) gave capricious results, it tion of the benzoyl group followed by removal of the vinyl TMS 1 13 cleanly reacted with the corresponding alanate complex to afford afforded 17. The spectroscopic data ( HNMR, CNMR, [R]D)of the alkyne 9.13 A regioselective Ru-catalyzed coupling reaction 17 match those of Nakata.14a A dealkylative saponification of 17 between alkene 8 and alkyne 9 rapidly gave 7 with the carbon with n-PrSLi15 completed the synthesis of the left-hand side 2, skeleton present in 2. After protection with TBS-OTf, the less- common to the mycalamide, onnamide, and theopederin families. hindered olefin was chemoselectively dihydroxylated to furnish the The synthesis of the right-hand side azide 3 started from diol 11 as a 1:1 diastereomeric mixture. Monobenzoylation followed commercially available (R)-pantolactone, which was methylated 16 by oxidative cyclization gave the desired pyran 15 in 18% yield with Ag2O and excess CH3I without racemization (Scheme 3). together with a mixture of 14 and 15 (63%) in 1:1 ratio after After reduction with DIBAL-H, the resulting lactol was subjected chromatography on silica gel. Interestingly, diastereomer 14 was to mediated allylation reaction with 2-(chloromethyl) 17 in dynamic equilibrium with the desired isomer 15 on silica gel. in aqueous saturated NH4Cl to produce 6 with a 5:1 diastereomeric Subjecting the 1:1 mixture of 14 and 15 to silical gel column ratio favoring 6. The stereochemistry was confirmed by X-ray chromatography for three cycles provided 15 in 53% yield together analysis. Surprisingly, Pd(0)-catalyzed cyclization exclusively with a 2:1 ratio of 14 and 15 (34%). The C(7) stereochemistry, produced the eight-membered ring product 25 (79% yield) wherein which has been difficult to control in many previous syntheses,14 the primary alcohol served as the nucleophile (Scheme 4). In is under substrate control. Thus, the stereochemical outcome in the contrast, the chemoselectivity was completely switched to form the dihydroxylation step is inconsequential. Methylation without migra- tetrahydrofuran 18 in 99% yield with the addition of Et3B. Moffat-

48 9 J. AM. CHEM. SOC. 2004, 126,48-49 10.1021/ja038787r CCC: $27.50 © 2004 American Chemical Society COMMUNICATIONS

Scheme 3. Synthesis of Azide 3a to control the challenging C(7) stereocenter. The right-hand side 3 was synthesized from (R)-pantolactone. The novel features include constructing the trioxadecalin core with two Pd(0)-mediated O-π- allyl cyclizations. The first one is chemoselective, while the second one is highly diastereoselective. Furthermore, a new strategy to construct 1,3-dioxan-4-ones involving 4-methylene tetrahydro- furans28 has been developed. Three additional steps would be required to complete a total synthesis of mycalamide A. Acknowledgment. We thank Professor Y. Kishi (Harvard University) for sending us part of Dr. Hong C. Y.’s thesis, Professor T. Nakata (RIKEN) for providing us the 1HNMR of 17, 18, and 3, NIH General Medical Science (GM 13598) and NSF for their generous financial support. Mass spectra were provided by the Mass Spectrometry Regional Center of the UCSF Supported by the NIH a Conditions: (a) Ag2O, MeI, CH3CN, 58 °C, 86%, 98% ee, (b) (i) Division of Research Resources. - ° DIBAL-H, CH2Cl2, 78 C; then 2-(chloromethyl)allyl acetate, In powder, Supporting Information Available: Experimental details and sat. aq NH Cl, 62% (5/1 dr). (c) PdCl (dppf), BEt ,EtN, THF, reflux, 4 2 3 3 spectroscopic data (PDF). This material is available free of charge via 99%. (d) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °D, 90% (96% BRSM) (e) vinyl magnesium bromide, MgBr2 diethyl ether complex, CH2Cl2, -78 °C the Internet at http://pubs.acs.org. to rt, 96%. (f) n-BuLi, (Boc)2O, THF, 87% (95% BRSM). (g) (DHQD)2- PHAL, K3Fe(CN)3,K2CO3, MeSO2NH2, t-BuOH/H2O; then NalO4, THF/ References H2O, 91%. (h) m-CPBA, 30% Li2CO3,CH2Cl2, rt, 98%. (i) TBDMSOTf, (1) (a) Perry, N. B.; Blunt, J. W.; Munro, M. H. G. J. Am. Chem. Soc. 1988, TEA, CH2Cl2, -78 °C, 30 min; then DMDO, acetone, CH2Cl2, molecular - ° ° 110, 4850. (b) Perry, N. B.; Blunt, J. W.; Munro, M. H. G.; Thompson, sieves, 5 C, 68%. (j) TBAT, , THF, 50 C, 83%. (k) Tf2O, A. M. J. Org. Chem. 1990, 55, 223. (c) Simpson, J. S.; Garson, M. J.; pyridine, 0 °C; then NaNO2, DMF, rt, 75%. (I) Pd2(dba)3CHCl3, dppf, DCE, Blunt, J. W.; Munro, M. H. G.; Hooper, J. N. A. J. Nat. Prod. 2000, 63, 70 °C, 58%. (m) DIBAL-H, -78 °C; then pyridine, DMAP, Ac2O, -78 704. (d) West, L. M.; Northcote, P. T.; Hood, K. A.; Miller, J. H.; Page, °C to rt, 100% (1.6/1 dr). (n) 9-BBN, Wilkinson’s catalyst; then PCC, DCM, M. J. J. Nat. Prod. 2000, 63, 707. (2) (a) Shinichi, S.; Ichiba, T.; Kohmoto, S.; Saucy, G. J. Am. Chem. Soc. 45 °C. (o) Ph3PdCH2, toluene, -40 to -20 °C, 47% over two steps. (p) R 1988, 110, 4851. (b) Matsunaga, S.; Fusetani, N.; Nakao, Y. Tetrahedron (DHQD)2PYR, OsO4,K2CO3,K3Fe(CN)6, t-BuOH/H2O, for -AcO 74%, 1992, 48, 8369. (c) Kobayashi, J.; Itagaki, F.; Shigemori, H. J. Nat. Prod. 4.3/1 dr; for â-AcO quant., 9/1 dr. (q) Triphosgen, pyridine, DCM, -78 1993, 56, 976. (d) Vuong, D.; Capon, R. J.; Lacey, E.; Gill, J. H.; Heiland, °C, R-AcO 73%, â-AcO 84%. (r) TMSOTf, TMSN3,CH3CN, 0 °C, 68% K.; Friedel, T. J. Nat. Prod. 2001, 64, 640. (1.6/1 dr). (3) (a) Fusetani, N.; Sugawara, T.; Matsunaga, S. J. Org. Chem. 1992, 57, 3828. (b) Tsukamoto, S.; Matsunaga, S.; Fusetani, N.; Toh-e, A. Scheme 4. Divergence on Pd(0)-Catalyzed Cyclization of 6 Tetrahedron 1999, 55, 13697. (4) Burres, N. S.; Clement, J. J. Cancer Res. 1989, 49, 2935 and references therein. (5) Ogawara, H.; Higashi, K.; Uchino, K.; Perry, N. B. Chem. Pharm. Bull. 1991, 39, 2251. (6) (a) Hong, C. Y.; Kishi, Y. J. Org. Chem. 1990, 55, 4242. (b) Hong, C. Y.; Kishi, Y. J. Am. Chem. Soc. 1991, 113, 9694. Swern oxidation18 of 18 followed by vinylmagnesium bromide (7) (a) Nakata, T.; Matsukura, H.; Jian, D.; Nagashima, H. Tetrahedron Lett. 1994, 35, 8229. (b) Nakata, T.; Fukui, H.; Nakagawa, T.; Matsukura, H. 19 addition in the presence of MgBr2‚Et2O gave the allyl alcohol 19 Heterocycles 1996, 42, 869. as a single diastereomer. Carbonate formation, selective cleavage (8) Roush, W. R.; Pfeifer, L. A. Org. Lett. 2000, 2, 859. (9) Kocienski, P. J.; Narquizian, R.; Raubo, P.; Smith, C.; Farrugia, L. J.; of the exocyclic double bond, and a regioselective Baeyer-Villiger Muir, K.; Boyle, F. T. J. Chem. Soc., Perkin Trans. 1 2000, 2357 and oxidation furnished the lactone 20. The installation of the C(11) references therein. (10) For recent examples, see: (a) Breitfelder, S.; Svchlapbach, A.; Hoffmann, hydroxyl group was quite challenging. This was accomplished via R. W. Synthesis 1998, 468 and references therein. (b) Toyata, M.; Hirano, a Rubottom oxidation20 with anhydrous DMDO21 to give 21. H.; Ihara, M. Org. Lett. 2000, 2, 2031. (c) Rech, J. C.; Floreacig, P. E. Org. Lett. 2003, 5, 1495. Addition of 4 Å molecular sieves was crucial to improve both the (11) (a) Trost, B. M.; Indolese, A. J. Am. Chem. Soc. 1993, 115, 4361. (b) yield and scalability of this reaction due to the competing Trost, B. M.; Machachek, M.; Schnaderbeck, M. Org. Lett. 2000, 2, 1761. protonation/desilylation process. After removal of the TBDMS (12) Yamaguchbvi, M.; Hirao, I. Tetrahedron Lett. 1983, 24, 391. (13) Skrydstrup, T.; Be´ne´chie M.; Khuong-Huu, F. Tetrahedron Lett. 1990, group with TBAT,22 the hydroxyl stereochemistry was inverted by 31, 7145. activation as a triflate followed by treatment with NaNO in DMF.23 (14) For discussions on controlling C(7) stereochemistry, see (a) Trotter, N.; 2 Takahashi, S.; Nakata, T. Org. Lett. 1999, 1, 957. (b) Roush, W. R.; This is an efficient method to invert the stereochemistry of Pfeifer, L. A. Org. Lett. 2000, 2, 859. secondary alcohols. Obviously, many of the methods in the literature (15) Barlett, P. A.; Johnson, W. S. Tetrahedron Lett. 1970, 11, 4459. (16) Mendik, M. T.; Cottard, M.; Rein, T.; Helquist, P. Tetrahedron Lett. 1997, employing carboxylic acids as nucleophiles would introduce the 36, 6375. dilemma of having to chemoselectively differentiate three different (17) Loh, T.-P.; Gao, G.-Q.; Pei, J. Tetrahedron Lett. 1998, 39, 1457. (18) Swern, D.; Mancuso, A. J.; Huang, S.-L. J. Org. Chem. 1978, 43, 2480. ester groups. The second O-π-allyl cyclization with Pd2(dba)3 and (19) Keck, G. E.; Andrus, M. B.; Romer, D. R. J. Org. Chem. 1991, 56, 417. dppf as ligand proceeded in a highly diastereoselective fashion to (20) Rubottom, G. M.; Marrero, R. Synth. Commun. 1981, 11, 505. furnish lactone 4, which was reductively acylated24 to give (21) Murray, R. W.; Jeyaraman, R. J. J. Org. Chem. 1985, 50, 2847. (22) Pilcher, A. S.; Ammon, H. L.; DeShong, P. J. Am. Chem. Soc. 1995, quantitative yield of 22 as a separable mixture of R/â (1.6/ 117, 5166. 1) diastereomers. Hydroboration,25 one-pot PCC oxidation,26 fol- (23) (a) Albert, R.; Dax, K.; Link, R. W.; Stutz, A. E. Carbohydr. Res. 1983, 118, C5. (b) Raduchel, B. Synthesis 1980, 292. lowed by a Wittig olefination yielded the alkene 23. After (24) Kopecky, D.; Rychnovsky, S. D. J. Org. Chem. 2000, 65, 191. dihydroxylation,27 carbonate formation, and azide formation, 3 was (25) Evans, D. A.; Fu, G. C.; Hoveyda, A. H. J. Am. Chem. Soc. 1992, 114, R 6671. obtained as a 1.6/1 mixture of /â C(10) diastereomers. The spectra (26) Brown, H. C.; Kulkarni, S. U.; Rao, C. G. Synthesis 1980, 151. (1HNMR, IR) match those reported by Kishi6a and Nakata.7 (27) Crispino, G. A.; Jeong, K.-S.; Kolb, H. C.; Wang, Z.-M.; Xu, D.; Sharpless, - K. B. J. Org. Chem. 1993, 58, 3785. In conclusion, an efficient formal synthesis of ( )-mycalamide (28) (a) Trost, B. M.; Bonk, P. J. J. Am. Chem. Soc. 1985, 107, 8277. (b) A was achieved. The left-hand side 2 was synthesized from (2S,3S)- Trost, B. M.; King, S. A.; Schmidt, T. J. Am. Chem. Soc. 1989, 111, 2,3-epoxybutane. The key features include a highly regioselective 5902. Ru-catalyzed alkene-alkyne coupling reaction and a novel method JA038787R

J. AM. CHEM. SOC. 9 VOL. 126, NO. 1, 2004 49 4242 J. Org. Chem. 1990,55, 4242-4245 after treatment with potassium tert-butoxide, tetra- hydrobenzofuran 31 in good yield. We sought to further demonstrate the utility of this approach as an entry into 17: RzCH, 15: R=CH, 18. R=Ph the vast number of 3-methyl furan~terpenoids'~by em- 16 R=Ph ploying DBP in the total synthesis of (R)-menthofuran (33).19 The menthofuran precursor 32 was prepared in

CHZ& CW~CW~)~* S02Ph

0 30

R;CH, or Ph I produced dihydrobenzofuranone 23 (R = H) in 85% yield. Note that in this case, the substitution pattern about the furan ring is different from that encountered with the acyclic dicarbonyl compounds. More than likely the initial step involves 0-alkylation to give 21 as a transient spec- ies.'* Further reaction of this material with base results in cyclization to 22, which then undergoes a subsequent aromatization. A similar series of reactions was used to prepare furan 24. HKSo2Ph

33 31 (R=H) 32 (R=CH3) 21 R=H or CHI the same fashion as 31, using the commercially available (R)-3-methylcyclohexanone.Furan 32 was then treated with sodium amalgam to give (@menthofuran in 85% overall yield. In conclusion, the DBP approach is a general method for the synthesis of C-2 and C-3 substituted furans. In addition to its ease of removal, the pendant sulfone at C-4 offers a convenient and versatile site for further elaboration 22 23 (R=H) (via alkylationmor Julia coupling2'). This strategy toward 24 (R=CH3) furans clearly could be applied to more complex targets. When 2-acetylcyclohexanone was used, employing so- We are currently investigating the scope and limitations dium methoxide as the base, ring opening of the inter- of this protocol. mediate adduct 28 to give 30 takes precedence over dea- Acknowledgment. We gratefully acknowledge support cetylation, presumably as a consequence of the stability of this work by the National Institutes of Health (CA- of the anion formed. This circumstance can be avoided, 26750). Use of the high-field NMR spectrometer used in however, by the use of a formyl group in place of the acetyl these studies was made possible through equipment grants group to activate the cyclic ketone. Thus, DBP reacted from the NIH and NSF. with the sodium salt of 2-formylcyclohexanone(26) to give, Supplementary Material Available: Experimental proce-

~~ dures and spectroscopic data for new compounds (5 pages). (18) 1,3-Cyclohexanediones preferentially alkylate on the 0 atom of Ordering information is given on any current masthead page. the diketone enolate, see: Stetter, H. In Newer Methods of Preparative Organic Chemistry; Foerst, W.,Ed.; Academic Press: New York, 1964, Vol 11. Taylor, E. C.; Hanks, G.H.; McKillip, A. J. Am. Chem. SOC.1968, (19) Tori, K.; Ueyama, M.; Horibe, I.; Tamura, Y.; Takeda, K. Tet- 90, 2421. An alternate, but less likely, mechanism for the formation of rahedron Lett. 1975, 4583. 23 (or 24) could involve SN2 displacement of the allylic bromide by the (20) Magnus, P. D. Tetrahedron 1977,33, 2019. enolate carbon followed by cyclization and aromatization. (21) Julia, M.; Stacino, J. Tetrahedron 1986, 42, 2469.

Total Synthesis of Mycalamides A and B Chang Yong Hong and Yoshito Kishi* Department of Chemistry, Harvard University, 12 Oxford Street, Cambridge, Massachusetts 02138 Received April 26, 1990 Summary: A total synthesis of mycalamides A (1) and B turally, they are strikingly similar to pederin (3), the (2) was accomplished in an enantiomerically pure form, vesicatory principle of staphylinid Paeder~s.~ establishing unambiguously their absolute configuration. (1) (a) Perry, N. B.; Blunt, J. W.; Munro, M. H. G.;Pannell, L. K. J. Mycalamides A (1) and B (2) and onnamide A have Am. Chem. SOC.1988,110,4850. (b) Perry, N. B.; Blunt, J. W.; Munro, recently been isolated from marine sponges.'s2 Struc- M. H. G.; Thompson, A. M. J. Org. Chem. 1990,55, 223.

0022-3263/90/19~5-4242$02.50/0 0 1990 American Chemical Society Communications J. Org. Chem., Vol. 55, No. 14, 1990 4243 Each of these natural products exhibits remarkable bio- (OH), on C/EtOAc/room temperature then H2/Pt02/ logical activity: and we are intrigued by the acylaminal AcOH/room temperature; (5) (t-B~)(c~H,)~Sicl/ group, such as the one found at the C-10 position5 of 1-3. imidazole/CH2Clz/room temperature; (6) C6H,CH2Br/ This may play an important role in the NaH/THF/room temperature; (7) n-Bu,NF/THF/room origin of the biological activity. For example, it may be temperature] allowed the introduction of the geminal activated by eliminative cleavage of the C-10 carbon-ox- dimethyl groups at the C-14 position in 62% overall yield, ygen bond, and if so, the additional ring system found in i.e. 4 - 5. We planned to incorporate the C-17,18 glycol mycalamides and onnamide could provide an interesting via the terminal olefin 6. As direct methods did not yield means to address the question of the process of activation fruitful results,125 (aD+63.0", CHC1,) was converted into and subsequent reaction in terms of stereoelectronic ef- 6 (aD +77.9", CHC1,) in five steps [(l)Swern oxidation; fects. In this paper, we report a total synthesis of myca- (2) Horner-Emmons olefination; (3) DIBAL/CH2C12/-78 lamides A and B as the first step toward this goal. "C; (4) H2/Rh on AlZO3/EtOAc/room temperature; (5) O-O,NC6H4SeCN/P(n-BU)3/c6H6/rOOmtemperature, HO 7 followed by MCPBA treatment',] in 79% overall yield. Application of the recently developed asymmetric osmy- lation [0sO4/N,N'-bis(2,4,6-trimethylbenzy1)-(S,S)-1,2- diphenyl-1,2-diaminoethane/CH,Cl,/-90"C]14 gave a 6:l mixture of the two possible glycols, which were then transformed to the corresponding carbonates and sepa- rated on silica gel to yield the desired carbonate 7 (75% overall yield; aD +66.7", CHCl,) along with the corre- sponding undesired carbonate (13% yield). The undesired 1 mycalamide A (R=H) carbonate was recycled in 65% overall yield via the olefin 2 mycalamide (R=Me) 6.15 The C-17 stereochemistry in 7 was assigned on the 6 basis of three pieces of evidence. First, the asymmetric osmylation of 6 in the presence of the antipode of the chiral diamine yielded an inverted ratio (ca. 1:ll) of the two possible glycols. Second, Sharpless asymmetric ep- oxidation [diethyl ~-tartrate/Ti(i-PrO)~/t-BuOOH/ CHzC1,]16of the allylic alcohol prepared from 5,17 followed by DIBAL reduction,18 yielded a ca. 1:l mixture of the expected 1,2- and 1,3-glycols. The 1,2-glycol thus obtained was found to be identical with the major glycol which resulted in the asymmetric osmylation of 6 in the presence of the (S,S)-diamine. Third, the asymmetric osmylation 3 : pederin using Oppolzer's chiral a~xiliary'~yielded a >301 mixture We chose methyl a-D-glucopyranoside6 as the starting of the two possible glycols, the major product of which was material for the synthesis of right half of mycalamides correlated with 7.20 because of their obvious structural similarity. With small In order to construct the B ring, we needed to introduce modifications of the functional group transformations known in the carbohydrate literature, this substance was (11) For a large scale preparation, we used (1) CHC13/aqueous converted into the alcohol 4 ((YD+33.3", CHC13)7in over NaOH/C6H,CH2(Me),N+Cl- and (2) Li/liquid NH,/-78 OC in place of the step 3 for safety reasons. 65% overall yield.8 A seven-step sequence of routine (12) These included a vinylcuprate reaction on the tosylate and iodide synthetic reactions [ (1) Swern o~idation;~(2) Wittig re- prepared from 5. action; (3) CH2N2/Pd(OAc)2/EtzO/0oC,loJ1 (4) H2/Pd- (13) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976,41, 1485. (14) Corey, E. J.; Jardine, P. D.; Virgil, S.; Yuen, P.-W.; Connell, R. D. J. Am. Chem. SOC.1989,111,9243. For recent papers on this subject, see: (a) Hirama, M.; Oishi, T. J. Org. Chem. 1989,54, 5834. Hirama, M.; (2) Sakemi, S.; Ichiba, T.; Kohmoto, S.; Saucy, G.; Higa, T. J. Am. Oishi, T.; Ito, S. J. Chem. Soe., Chem. Commun. 1989,665. (b) Tomioka, Chem. SOC.1988,110,4851. K.; Nakajima, M.; Iitaka, Y.; Koga, K. Tetrahedron Lett. 1988,29, 573. (3) (a) Cardani, C.; Ghiringhelli, D.; Mondelli, R.; Quilico, A. Tetra- (c) Wai, J. S. M.; Marko, I.; Svendsen, J. S.; Finn, M. G.; Jacobsen, E. hedron Lett. 1968,2537. (b) Matsumoto, T.; Yanagiya, M.; Maeno, s.; N.; Sharpless, K. B. J. Am. Chem. SOC.1989,111, 1123. Svendsen, J. S.; Yasuda, S. Tetrahedron Lett. 1968,6297. (c) Furusaki, A.; Watanabe, Marko I.; Jacobsen, E. N.; Pulla Rao, Ch.; Bott, S.; Sharpless, K. B. J. T.; Matsumoto, T.; Yanagiya, M. Tetrahedron Lett. 1968, 6301. Org. Chem. 1989,54, 2264. (d) Annunziata, R.; Cinquini, M.; Cozzi, F.; (4) Mycalamides and onnamide are known to exhibit promising an- Raimondi, L.; Stefenelli, S. TetrahedronLett. 1987,28,3139. (e) Tokles, tiviral and antitumor activity: Burres, N. S.; Clement, J. J. Cancer Res. M.; Snyder, J. K. Tetrahedron Lett. 1986, 27, 3951. (f) Yamada, T.; 1989, 49, 2935. See also the references cited in ref 1 and 2. Pedrin is Narasaka, K. Chem. Lett. 1986, 131. known to be a powerful insect toxin and a potent inhibitor of protein (15) This recycling was carried out in four steps, (1) 0.5 N NaOH/ biosynthesis: see the references cited in ref 3. THF/room temperature; (2) HC(OMe)3/PPTS/room temperature; (3) isobutyric anhydride/l25 'C; (4) asymmetric osmylation. (5) The numbering of compounds used in this paper corresponds to (16) Kabuki, T.; Sharpless, K. B. J. Am. Chem. SOC.1980,102, 5974. that of mycalamides: see structures 1 and 2. (17) This allylic alcohol was prepared from 5 in three steps [(l) Swem (6) Purchased from Pfanstiehl Laboratories, Inc. oxidation; (2) Horner-Emmons olefination; (3) DIBAL reduction]. (7) Satisfactory spectroscopic data ('H and I3C NMR, MS, HR MS, (18) Finan, J. M.; Kishi, Y. Tetrahedron Lett. 1982, 23, 2719. IR, [a]~)were obtained for all the new compounds reported. (19) (a) Vandewalle, M.; Van der Eycken, J.; Oppolzer, W.; Vullioud, (8) This transformation required a six-step sequence of reactions, (1) C. Tetrahedron 1986, 42, 4035. (b) Oppolzer, W.; Barras, J.-P. Helo. 4-MeOC8H4CH(OMe)2/p-TsOH;(2) (n-B~)~snO, followed by p-TsCl/ Chim. Acta 1987, 70, 1666. NEb treatment: cf. Munavu, M.; Szmant, H. H. Org. Chem. 1976, (20) The preparation of the substrate for this asymmeric osmylation R. J. involved (1) ozonization of 6; (2) Horner-Emmons olefination; (3) basic 41, 1832; (3) MeI/Ag,O; (4) Na(Hg); (5) C6H6CH2Br/NaH; (6) hydrolysis; (4) acid chloride formation; and (5) coupling with Oppolzer's NaBH3CN/TFA cf. Johansson, R.; Samuelsson, B. J. Chem. SOC., chiral sulfonamide. The major osmylation product with the Oppolzer's Perkin Trans. I 1984, 2371. chiral auxiliary, derived from (1s)-camphor-10-sulfonyl chloride, was (9) (a) Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651. (b) Man- subjected to a five-step sequence of reactions, (1) (MeO)2C(Me)2/cam- cuso, A. J.; Huang, s.-L.; Swern, D. J. Org. Chem. 1978, 43, 2480. phorsulfonic acid; (2) LAH; (3) Swern oxidation; (4) MCPBA; (5) LAH, (10) Suda, M. Synthesis 1981, 714. to yield the glycol corresponding to 7. 4244 J. Org. Chem., Vol. 55, No. 14, 1990 Communications CHC13) in 60% overall yield. After debenzylation [H2/ ,OMPM Pd(OH)2on C/EtOAc/room temperature], 8 was treated with paraformaldehyde in the presence of hydrochloric acid at 0 "C to yield the hemiacetal 9 as a C-10 diastereomeric mixture in 86% overall yield.22 A standard procedure of 6en activation (MsC1/Et3N/DMAP/CH2Cl2/-60 "C) and displacement (n-Bu4N+N3-/CH2C12/-78"C - room tem- 4 5 6 perature) was applied to convert 9 into the azide a suitable precursor of the right half of mycalamide A, in 72% overall yield as a chromatographically inseparable 2:l C-10 diastereomeric mixture. A straightforward functional group manipulation allowed for the transformation of 10 into the azide acetate 12," a suitable precursor of the right half of mycalamide B, in five steps [(l)NaOH/aqueous Me0 p-dioxanelroom temperature; (2) 4-MeOC6H4- OBn (C6H5),CC1/Hunig base/CH2C12/room temperature; (3) 7 Me1/ NaH / DMF/ 75 'C; (4) p-TsOH/ MeOH / room tem- 8 9 X=OH perature; (5) Ac20/DMAP/CH2C12/roomtemperature] in 10 X=N, 76% overall yield. The azide acetate 12 was also a 21C-10 11 X=NH2 diastereomeric mixture, favoring the natural configuration. These diastereomers were separated on silica gel to give a-12 and 6-12, and both proved to be configurationally stable under a variety of conditions. Hydrogenation (H2/Pd on C/EtOAc/room temperature) of a-12 or 6-12 yielded the expected amine a-13 or 4-13. However, neither a-13 nor 6-13 was found to be configu- rationally stable under acidic (camphorsulfonic acid), neutral, or basic (NH3or Et3N) conditions. The ratio of 12 a-12 X=N,, Y=H 14.R =H,R =Me a- and P-13 was approximately 2:l under neutral or basic P-12 X=H, Y=N, IS R'=DMPM, R~=H conditions while approximately 1:4, disfavoring the natural a-13 X=NH2, Y=H a-configuration, under acidic conditions. Through the lH P-13 X=H, Y=NH 2 NMR studies, the same trend was found for the amines 11 derived from the azides 10. These experiments dem- onstrated that the C-10 stereochemistry of mycalamides R'O"RzoL should be addressed at the step of the amide-bond for- mation or thereafter. The left half of mycalamides, 15, was prepared in two steps [(l)3,4-(MeO)2C6H,CH2C1/NaH/DMF/room tem- perature; (2) n-PrSLi/HMPA/room temperature] from 14, one of the intermediates used in Nakatas' total synthesis CH2 of ~ederin,~~in 63% overall yield. Activation of the car- 2 boxylic acid 15 with p-toluenesulfonyl chloride (DMAP/ a- 16 R'=R =carbonate a-17 R'=Me,R 2 =Ac CH,Cl,/room temperature)26 and subsequent coupling

(22) When warmed up to room temperature, this reaction yielded the R'O"R*ok corresponding chloride, which was directly used for the next azide dis- placement reaction. The overall yield of this sequence was comparable with the one reported in the text but its reproducibility was not as high as the other. (23) HR MS (FAB,NaI): calcd for C14HaNs07 (M+ + H) 344.1456, found 344.1458. [a]~:+5.8O (c 0.97, CHC13). 'H NMR (major isomer, CDC13): 6 0.95 (3 H, s, C-14 CH,), 1.21 (3 H, s, C-14 CH3), 1.70 (1 H, m, C-16 H), 2.78 (1 H, ddd, J= 4.5, 12.6, 14.6 Hz, C-16 H), 3.01 (1 H, d, J = 3.2 Hz, C-13 H), 3.41 (3 H, 8, OCHI), 3.66 (1 H,dd, J = 3.4, 12.6 Hz, 8-16 R'=R2=carbonate C-15 H), 3.72 (1 H, dd, J= 2.0,3.2 Hz, C-12 H), 3.79 (1 H, t, J= 2.0 Hz, C-11 H), 4.21 (1 H, dd, J = 7.5, 8.5 Hz, C-18 H), 4.61 (1 H, dd, J 7.8, 8-17 R'=Me, R2=Ac 8.5 Hz, C-18-H). 4.69 (1 H. d. J = 2.0 Hz. C-10-H). 4.78 (1 H. d. J 6.6 Hz, OCH20),4.86 (1 H, m, d-17 H),5.20 (1 H, d, 2 = 6.6 Hz, OCH20). the axially disposed aldehyde group or its equivalent at IR (CCl,): 2120, 1823 cm-'. the C-11 position. This was accomplished by treatment (24) HR MS (FAB, NaI): calcd for Cl6HnN9O7Na (M+ + Na) of 7 with propargyltrimethylsilane in the presence of 396.1745, found 396.1739. 'H NMR (CDC13): 6 0.92 (3 H, s, C-14 CH& 1.08 (3 H, 8, C-14 CHI), 1.62 (1 H, ddd, J = 3.0, 9.4, 14.7 Hz, (2-16 H),

TMSOTf in acetonitrile,2*followed by ozonization and ,, ~ ~~ ,~~~ ~~,~ ~ ~ ~~~ 2.09 (3 H. s. OCOCH3. 2.26 (1 H. m.,~ C-16 ~~~ HL 3.01 (1 H. d. J = 3.6 Hz. acetalization, to furnish the dimethyl acetal 8 (aD+63.9', C-13 H),3.40 (3 H, s,CCH,), 3.41'(3 H, s, OCHd, 3.57 (I'H, m, C-17 H); 3.70 (1 H, m, C-15 H), 3.71 (1 H, m, (2-12 H), 3.87 (1 H, m, C-11 H), 4.14 (1 H, dd, J = 4.9, 12.1 Hz, C-18 H), 4.36 (1 H, dd, J 2.6, 12.1 Hz, C-18 (21) For the stereochemical outcome of this type of C-glycosidation, H), 4.63 (1 H, d, J = 2.0 Hz, C-10 H), 4.77 (1 H, d, J = 6.5 Hz, OCH*O), see: Lewis, M. D.; Cha, J. K.;Kishi, Y.J. Am. Chem. SOC.1982,104,4976. 5.21 (1 H, d, J = 6.5 Hz, OCH20). IR (film): 2119, 1739 cm-'. Propargyltrimethylsilanewas first used by Dr. Wu in these laboratories (25) Nakata, T.; Nagao, S.; Mori, N.; Oishi, T. Tetrahedron Lett. 1985, in connection with work on the conformational analysis of C-glycosides. 26, 6461 and 6465. Communications J. Org. Chem., Vol. 55, No. 14, 1990 4245 with the amines, prepared by hydrogenation of 10 or 12, natural and unnatural diastereomer^.^^^^^ Mycalamides yielded a mixture of a-16 (38% yield)27and 0-16 (40% A (1)33and B (2)34were obtained from a-16 and a-17 in yieldP or a-17 (59% yield)29and 0-17 (26% yield),30re- two steps [(l)t-BuOK/THF/room temperature for a-16 spectively. The diastereomers were readily separable by and LiOH/MeOH/room temperature for a-17; (2) chromatography to yield stereochemically pure products. DDQ/H20-CH2C12/roomtemperature] in 60% and 69% Both 0-16 and 8-17were found to isomerize to the corre- overall yields, respectively. On comparison of ‘H and I3C sponding natural diastereomers upon treatment with base NMR, IR, MS, [aID,and TLC data, the synthetic materials (t-BuOKITHFlreflux). It is interesting to note that 8-16 were found to be identical with an authentic sample of isomerized almost exclusively to the natural diastereome9 mycalamides A and B.35 while p-17 epimerized to reach a ca. 1:l mixture of the The reported synthesis has a good flexibility for the preparation of various analogues of mycalamides. In this (26) The following reagents were tested for the activation of carboxylic regard, it is also interesting to note that some of the in- acid and the order of efficiencv was found as p-toluenesulfonyl chloride termediates should be useful for the construction of the - 2,4,6-trichlorobenzeneaulfonylchloride > ethyl chloroformat;! > phenyl right half of onnamide A, Lastly, it is worthwhile to point chloroformate > sulfonyl chloride > 2,4,6-triisopropylbenzenesulfonyl out that this synthesis has unambiguously established the chloride > 2,4,6-trimethylbenzenesulfonylchloride > methanesulfonyl chloride. absolute stereochemistry of mycalamides A and B, which (27) HR MS (FAB, NaI): calcd for CNHlsNOl3Na (M’ + Na) had tentatively been assigned on the basis of their struc- 702.3099, found 702.3116. [“ID: +72.8’ (c 0.72, CHzClz). ‘H NMR tural similarity to pederin. (benzene-&): 6 0.55 (3 H. s. C-14 CH.). 0.69 13 H. s. C-14 CHd. 1.04 (3 H, d, J = 66 Hz, C-2 CHJ, 1.07 (3 H”, d, J = 6.9 Hz, C-3 CH;), 1.14 (1 H, dd, J 10.3, 13.6 Hz, C-16 H), 1.61 (1 H, ddd, J = 3.2, 10.3, 13.6 Hz, Acknowledgment. We thank Dr. Hisashi Kawasaki for C-16H),1.99(1H,dq,J=2.4,6.9H~,C-3H),2.74(1H,d,J=14.3HZ,some experiments on asymmetric osmylation. Financial C-5 H), 2.78 (1 H, d, J = 14.3 Hz, C-5 H), 2.87 (1 H, d, J = 10.4 Hz, C-13 support from the National Institutes of Health (CA-22215) H), 2.99 (1 H, br d, J = 10.3 Hz, C-15 H), 3.16 (3 H, s, C-6 OCH3), 3.24 is gratefully acknowledged. (3 H. 8. C-13 OCH,). 3.41 (1 H. dd. J = 7.1.9.8 Hz. C-11 H). 3.50 (3 H. Si Ar’OCH,), 3.52 (i’H,s, ArOCii3),’3.56 (1 H, m, C-18 H), 3.88 (1 H, dq; J = 2,4,6.6 Hz, C-2 H), 4.14 (1 H, dd, J 7.1, 10.4 Hz, (2-12 H), 4.15 (1 Supplementary Material Available: Experimental details H, s, H7),4.44 (1 H, m, C-18 HI, 4.45 (1 H, m, C-17 HI, 4.51 (1 H, d, J for the coupling reactions to form 16 and 17 and spectroscopic = 6.9 Hz. OCHVO). 4.56 (1 H. d, J = 6.9 Hz, OCH,O), 4.62 (1 H, d, J = data for the key intermediates and mycalamides A and B (13 11.3 Hz, CHzArj, 4.74 (1 H, d, J = 11.6 Hz, CH,Ar);4.80 (1 H, s, =CH2), pages). Ordering information is given on any current masthead 4.82 (1 H. 8. =CH*). 5.65 (1 H. t. J = 9.8 Hz. C-10 H). 6.70-6.97 (3 H. m, ArH), 7.11 (1 I-f, d, J = 9.8 Hz, NH). page. (28) HR MS (FAB, NaI): calcd for CNH19N013Na (M’ + Na) 702.3099, found 702.3065. [a]D: +7.6’ (c 1.01, CHzC12). ‘H NMR (benzene-d6): 6 0.64 (3 H, s, C-14 CH,), 1.02 (1 H, m, C-16 H), 1.07 (3 (32) On potassium tert-butoxide treatment, the carbonate group in 16 H. d. J 6.5 Hz. C-2 CHq). 1.18 (3 H. d. J 7.0 Hz. C-3 CHd, 1.20 (3 or the acetate group in 17 was hydrolyzed. It was observed that these --,H: s.’C-14, - CHn).’2.05(1 H:’do. J = 2.6. 7.0 Hz. C-3 H). 2.32 ti H. ddd. substances decomposed slowly under the basic conditions, so epimeriza- J = 7.1, 12.2, i4.1 Hz, C-16 Ffj, 2.62 (1 H, d, J = 2.2 Hz, C-13 H), 2.84 tion was stopped at approximately 60% completion for preparative (1 H, m, C-5 H), 2.86 (3 H, s, C-6 OCH3), 2.88 (1 H, d, J = 14.4 Hz, C-5 purposes to yield the epimerized natural diastereomer in 42% yield (67% H), 3.10 (1 H, t, J = 8.4 Hz, (2-18 H), 3.12 (1 H, m, C-11 H), 3.15 (1 H, mrrerted vielrll,__-_,. dd, J = 2.5, 12.2 Hz, C-15 H), 3.19 (1 H, m, C-12 H), 3.28 (3 H, s, ‘2-13 (33) HR MS (FAB, NaI): calcd for CZ4HI1NOoNa (M’ + Na) OCH,). 3.42 (3 H. s. ArOCHd. 3.53 (3 H. s. ArOCHd. 3.55 (1 H. t. J = 526.2626, found 526.2664 [“In: +82.4O (c 0.37, CHC1,). \H NMR (CDCld: 7.9 Hi,’C-l8 HI, 3:97 (1 H, do= 2.7,6.5Hz, C-2 HC4.11 (l’H, quintet, 6 0.88 (3 H, s, C-14 CH,); 039 (3 H, s, C-14 CH3),-1.01 (3 H, d, J = f0 J = 7.0 Hz,C-17 H), 4.40 (1 H, d, J 6.6 Hz, OCHZO), 4.57 (1 H, 9, C-7 Hz. C-3 CHA 1.21 (3 H. d. J = 6.4 Hz. C-2 CH,). 1.56 (2 H. m. C-16 H). H), 4.76 (1 H, d, J = 10.7 Hz, CHzAr), 4.81 (1 H, s, =Chz), 4.83 (1 H, s, 2.26 (1 H, iq,J = 2.4, 7.0 Hz,C-3 H), 2.38 (2 H, m, C-5 HI, 3.31 (3 H, =CH,), 4.85 (1 H, d, J = 6.6 Hz, OCHZO), 4.98 (1 H, d, J = 10.7 Hz, S, C-6 OCHS), 3.39 (1 H, dd, J 6.1,11.2 Hz, C-18 H), 3.47 (1 H, d, J CH2Ar),5.41 (1 H, dd, J = 1.8, 9.6 Hz, ‘2-10 H), 6.62-7.15 (3 H, m, ArH), 10.2 Hz, C-13 H), 3.57 (3 H, S, ‘2-13 OCHz), 3.58 (1 H, dd, J = 3.3, 11.2 8.04 I1 H. d. J = 9.6 Hz. NH). Hz. C-18 H), 3.65 (1 H, dd. J = 4.5, 7.5 Hz, C-15 H), 3.75 (1 H, m, C-17 (29) HR ’MS (FAB, ‘NaI); calcd for CSBHSNOl3Na (M’ + Na) H),3.87(1H,dd,J=6.7,9.6H~,C-llH),4.00(1H,dq,J=2.7,6.4H~, 732.3568, found 732.3570. [.ID: +38.2’ (c 0.24, CCl,). ‘H NMR (benz- C-2 H). 4.24 (1H. dd. J = 6.7. 10.2 Hz. C-12 H). 4.31 (1 H. 8. C-7 H). 4.75 ene-d6): 6 0.80 (3 H, s, C-14 CH3),0.87 (3 H, s, C-14 CH,), 0.92 (3 H, m, (1 H, si =CHp), 4.86 (1 H, s, =CHz), 4.89 (1 H, d, J = 7.0 Hz, OCHZO), C-16 H), 1.00 (3 H, d, J = 6.6 Hz, C-2 CH3), 1.08 (3 H, d, J 7.0 Hz, C-3 5.15 (1 H, d, J = 7.0 Hz, OCHZO), 5.88 (1 H, t, J = 9.7 Hz, C-10 H), 7.49 CH3), 1.65 (1 H, m, C-16 H), 1.81 (3 H, s, OCOCH3), 1.98 (1 H, dq, J = (1 H, d, J = 9.7 Hz, NH). 13C NMR (CDCl3): 6 12.02 (C-3 CH3), 13.51 2.5, 7.0 Hz, C-3 H), 2.75 (1 H, d, J = 14.2 Hz, C-5 H), 2.80 (1 H, d, J = (C-14 CH,), 17.89 ((2-2 CHq), 23.11 ((2-14 CHd, 31.99 (C-16), 33.69 (C-5), 14.2 Hz, C-5 H), 3.02 (1 H, d, J = 10.4 Hz, (2-13 H), 3.11 (3 H, s, OCH,), 41.33 (C-g), 41.62 (C-14), 48.92 (C-6 OCH,); 61.81 (C-13 OCH,), 66.48 3.28 (3 H,s, OCH3), 3.29 (3 H,s, OCH3),3.39 (3 H,s,0CH3),3.40 (1 H, (C-18). 69.82 (C-2). 71.25 (C-11). 71.57 (C-17). 72.82 IC-7). 73.75 (C-10). m, H17),3.54 (3 H, s, OCH,), 3.63 (1 H, dd, J = 7.0,g.g Hz, C-11 H), 3.89 +4.38’(C:12), ‘78.97 (c-isj,79.11 (c-I~),86.82 (ock,oi, 99.~2‘(c-sj; (1 H, dq, J = 2.7, 6.6 Hz, C-2 H), 4.21 (1 H, S, C-7 H), 4.25 (1 H, dd, J 110.61 (=CH2), 145.56 (C-4), 171.79 (C-8). IR (film): 3100-3600,2970, 6.9, 10.4 Hz, C-12 H), 4.47 (1 H, dd, J= 5.3, 12.2 Hz, C-18 H), 4.59 (1 2929,1684,1522,1470,1382,1267,1194,1173,1093,1075,1034,938,879, H. dd. J = 2.3. 12.2 Hz. ‘2-18 H). 4.60 (1 H. d. J = 7.0 Hz, OCH,O). 4.62 670 cm“. (I’H, d, J = 10.9 Hz, CHzAr),4.65 (1 H, d,’J’= 7.0 Hz, OCHzOj, 4.’76 (1 (34) HR MS (FAB NaI): calcd for C&,NO $Ja (M+ + Na) 540.2782, H, s, =CHz),4.80 (1 H, s, =CH2),4.91 (1H, d, J = 10.9 Hz, CHzAr), 5.95 found 540.2769. [(Y]D: +41.2’ (c 0.19, CHCl,). NMR (CDClJ: 6 0.88 (1 H, t, J = 9.9 Hz, C-10 H), 6.60-7.08 (3 H, m, ArH), 7.28 (1 H, d, J = (3 H. 9. C-14 CHo). 1.00 (3 H. 9. C-14 CH,). 1.04 (3 H. d. J 7.0 Hz. C-3 9.9 Hz, NH). (30) HR MS (FAB, NaI): calcd for CSH,NOl3Na (M’ + Na) 732.3568, found 732.3599. [“ID: +19.8O (c 0.35, CClJ. ‘H NMR (benz- ene&): 6 0.83 (3 H, s, C-14 CHJ, 1.16 (3 H, d, J = 6.6 Hz, C-2 CHJ, 1.19 (3 H, d, J = 7.2 Hz, C-3 CH,), 1.29 (3 H, s, C-14 CH3), 1.68 (1 H, m, C-16 H), 1.72 (3 H, S, OCOCH3), 2.04 (1 H, dq, J = 2.6, 7.1 Hz, C-3 H), 2.41 3.68 (1 H, m, (2-18 H), 3.80 (1 H, dd,J= 6.7, 9.6 Hz, C-11 H), 3.89 (1 k, (1 H, ddd, J = 3.9, 12.8, 15.5 Hz, ‘2-16 H), 2.72 (1 H, d, J = 2.1 Hz, C-13 S, C-7 OH), 4.05 (1 H, dq, J 2.8,6.6 Hz, C-2 H), 4.23 (1 H, dd, J = 6.7,

10.4-. . ~~~Hz. C-12 H). 4.31 (1 H. S. C-7 H). 4.74 (1 H. 8. =CHo). 4.84 (1 H. d. H), 2.85 (1 H, d, J = 15.3 Hz, C-5 H), 2.88 (1 H, d, J = 15.3 Hz, C-5 H), , - -~ ~~ ~ ~~ ~~ --, -. - 2.92 (3 H, s, OCH,), 3.18 (3 H, s, OCH3), 3.26 (3 H, s, OCH,), 3.29 (1 H, J = 7.0 H~,OCH,O), 4.84 (1 H, i,’=cH2jl51i2’(1 H, X J-= 7.0 Hz; m, C-11 H), 3.42 (3 H, s, OCH3), 3.48 (1H, m, C-12 H), 3.50 (1H, m, C-17 OCHZO), 5.80 (1H, t, J = 9.7 Hz, C-10 H), 7.52 (1 H, d, J 9.9 Hz, NH). H), 3.51 (3 H, S, OCHB), 3.70 (1 H, dd,J= 2.4, 11.6 Hz, C-15 H), 3.98 (1 ‘3c NMR (CDCl,): 6 12.17 (C-3 CH3), 13.70 (C-14 CHJ, 17.96 (C-2 CH3), H, dq, J = 2.6, 6,6 Hz,‘2-2 H), 4.21 (1 H, dd, J = 5.9, 11.8 Hz, C-18 H), 23.20 (C-14 CH,), 29.71 (C-l6), 33.68 (C-5), 41.27 (C-3), 41.45 (C-14),48.61 4.35 (1H, dd, J 3.4,ll.a Hz, C-18 H), 4.39 (1 H, d, J 6.7 Hz, OCHzO), ((2-6 OCH,), 56.65 (C-17 OCH,), 61.80 (C-13 OCH3), 63.58 (C-18), 69.67 4.41 (1 H, s, C-7 H), 4.68 (1 H, d, J = 11.0 Hz, CHzAr), 4.77 (1 H, s, (C-2), 70.70 (C-ll), 71.67 (C-7), 73.97 (C-lo), 74.47 (C-12), 75.57 (C-15), =CHz), 4.81 (1 H, 9, =CHp), 4.87 (1 H, d, J = 6.7 Hz, OCHzO), 5.01 (1 78.75 (C-17), 79.31 (C-13). 86.51 (OCHsO), 99.94 (C-6), 111.14 (=CHd, H, d, J = 11.0 Hz, CHzAr), 5.52 (1 H, dd, J = 1.9, 9,8 Hz, C-10 H), 145.00 (C-4), 171.75 (C-8). IR (film): 3856,2926,2854,1687,1514,1465, 6.60-7.13 (3 H, m, ArH), 8.09 (1 H, d, J = 9.8 Hz, NH). 1382, 1237, 1102, 1074, 1036 cm-’. (31) This was confirmed by starting from both the unnatural and (35) We are indebted to Professors Perry and Munro for a sample of natural C-10 diastereomers. mycalamides A and B. 4834 J. Am. Chem. Soc. 2001, 123, 4834-4836

Total Synthesis of (+)-Phorboxazole A Scheme 1

Amos B. Smith, III,* Patrick R. Verhoest, Kevin P. Minbiole, and Michael Schelhaas

Department of Chemistry, Monell Chemical Senses Center and Laboratory for Research on the Structure of Matter UniVersity of PennsylVania, Philadelphia, PennsylVania 19104 ReceiVed February 26, 2001 In 1995 Searle and Molinski reported the isolation of phor- boxazoles A (1)andB(2), isomeric oxazole-containing mac- rolides, from the marine sponge Phorbas sp. endemic to the western coast of Australia (Scheme 1).1 The relative and absolute stereochemistries of the phorboxazoles were secured via a combination of NMR analysis, degradation studies, and synthetic correlation.2 When tested against the NCI panel of 60 human tumor cell lines, the phorboxazoles displayed virtually unsurpassed -9 cytotoxicity, exhibiting a mean GI50 of 1.58 × 10 M. Although the exact mechanism of action remains unknown, studies dem- onstrate that phorboxazole A (1) arrests the cell cycle at the S phase and does not affect tubulin. Given the potent cytotoxicity and the possibility of a new mechanism of action, the phorbox- azoles were selected by the NCI for in vivo trials.2a The combination of the outstanding antimitotic activity, architectural complexity, and extreme scarcity has led to wide interest in the synthetic community.3 The first total synthesis of phorboxazole A was reported by Forsyth and co-workers in 1998;4 shortly thereafter Evans and Fitch reported the completion of 4 into aldehyde 56 and salt 6, the syntheses of which were 5 phorboxazole B. In 1997 we embarked on the synthesis of these described previously.3n,o Continuing with this analysis, discon- challenging marine natural products; subsequently we disclosed nection of subtarget 3 at C(32-33) and C(40-41) revealed vinyl - assembly of two subtargets exploiting a modified Petasis Ferrier stannane 7, vinyl iodide 8, and the bifunctional oxazole 9. union-rearrangement tactic for the stereocontrolled construction Construction of the C(40-41) linkage would entail a Stille 3n,o of the two cis-fused tetrahydropyrans. In this communication, coupling, while oxazole 9, possessing the pseudobenzylic bromide - we describe the synthesis of the C(3 28) vinyl stannane, the and the triflate moieties, was envisaged as a novel bidirectional - C(33 46) lactone, their union via a bifunctional oxazole linchpin, linchpin to unite the side chain with the macrocycle. Importantly, and completion of the phorboxazole A synthetic venture. the coupling strategy possessed considerable flexibility from the From the retrosynthetic perspective, disconnections of phor- tactical perspective (vide infra). - - boxazole A (1) at the C(1) macrolactone, the C(2 3) and C(28 Assembly of the side chain of phorboxazole began with known 29) linkages led to side chain subtarget 3 and macrolide precursor Brown allylation7 adduct (+)-10 (Scheme 2).8,9 Methylation of - 4 (Scheme 1). A Wittig transform at C(19 20) further dissected the hydroxyl [MeOTf, 2,6-di-tert-butyl-4-methylpyridine (DT- 10 (1) Searle, P. A.; Molinski, T. F. J. Am. Chem. Soc. 1995, 117, 8126. BMP)], followed by ozonolysis furnished aldehyde (-)-11 in (2) (a) Searle, P. A.; Molinski, T. F.; Brzezinski, L. J.; Leahy, J. W. J. 81% yield for two steps. Although Wittig olefination of (-)-11 Am. Chem. Soc. 1996, 118, 9422. (b) Molinski, T. F. Tetrahedron Lett. 1996, with methyl alkyne 12a (R ) Me) led to a disappointing mixture 37, 7879. (3) (a) Lee, C. S.; Forsyth, C. J. Tetrahedron Lett. 1996, 37, 6449. (b) of olefins (E/Z ca. 2.2:1), condensation with the commercially Cink, R. D.; Forsyth, C. J. J. Org. Chem. 1997, 62, 5672. (c) Ahmed, F.; available phosphonate salt 12b (R ) TMS) in THF afforded enyne Forsyth, C. J. Tetrahedron Lett. 1998, 39, 183. (d) Ye, T.; Pattenden, G. (-)-13 in good yield with acceptable selectivity (97%, 5.5:1 E/Z). Tetrahedron Lett. 1998, 39, 319. (e) Pattenden, G.; Plowright, A. T.; Tornos, J. A.; Ye, T. Tetrahedron Lett. 1998, 39, 6099. (f) Paterson, I.; Arnott, E. A. The use of a PhCH3/THF (1:1) solvent system improved the E/Z Tetrahedron Lett. 1998, 39, 7185. (g) Wolbers, P.; Hoffman, H. M. R. ratio at the expense of both yield and reproducibility (72%, 7.5:1 Tetrahedron 1999, 55, 1905. (h) Misske, A. M.; Hoffman, H. M. R. E/Z). Removal of the TMS group (K2CO3), followed by Sharpless Tetrahedron 1999, 55, 4315. (i) Williams, D. R.; Clark, M. P.; Berliner, M. 11 12,13 A. Tetrahedron Lett. 1999, 40, 2287. (j) Williams, D. R.; Clark, M. P. dihydroxylation of the enyne (AD-Mix â; 7:1 dr) and Tetrahedron Lett. 1999, 40, 2291. (k) Wolbers, P.; Hoffman, H. M. R. acetonide formation then provided (+)-14. Terminal methylation Synthesis, 1999, 5, 797. (l) Evans, D. A.; Cee, V. J.; Smith, T. E.; Santiago, K. J. Org. Lett. 1999, 1, 87. (m) Wolbers, P.; Misske, A. M.; Hoffmann, H. (6) Although aldehyde 5 was the original subtarget for the central pyran, M. R. Tetrahedron Lett. 1999, 40, 4527. (n) Smith, A. B., III; Verhoest, P. revised aldehyde (+)-23 was ultimately employed (Scheme 4). R.; Minbiole, K. P.; Lim, J. J. Org. Lett. 1999, 1, 909. (o) Smith, A. B., III; (7) Brown, H. C.; Ramachandran, P. V. Pure Appl. Chem. 1991, 63, 307. Minbiole, K. P.; Verhoest, P. R.; Beauchamp, T. J. Org. Lett. 1999, 1, 913. (8) Clive, D. L. J.; Keshava Murthy, K. S.; Wee, A. G. H.; Prasad, J. S.; (p) Wolbers, P.; Hoffman, H. M. R.; Sasse, F. Synlett 1999, 11, 1808. (q) da Silva, G. V. J.; Majewski, M.; Anderson, P. C.; Haugen, R. D.; Heerze, L. Schaus, J. V.; Panek, J. S. Org. Lett. 2000, 2, 469. (r) Pattenden, G.; Plowright, D. J. Am. Chem. Soc. 1988, 110, 6914 (see Supporting Information). A. T. Tetrahedron Lett. 2000, 41, 983. (s) Rychnovsky, S. D.; Thomas, C. R. (9) The enantiomeric excess (ee) of alcohol (+)-10 was determined to be Org. Lett. 2000, 2, 1217. (t) Williams, D. R.; Clark, M. P.; Emde, U.; Berliner, 94% via Mosher ester analysis: (a) Dale, J. A.; Mosher, H. S. J. Am. Chem. M. A. Org. Lett. 2000, 2, 3023. (u) Greer, P. B.; Donaldson, W. A. Tetrahedron Soc. 1973, 95, 512. (b) Sullivan, G. R.; Dale, J. A.; Mosher, H. S. J. Org. Lett. 2000, 41, 3801. (v) Evans, D. A.; Cee, V. J.; Smith, T. E.; Fitch, D. M.; Chem. 1973, 38, 2143. (c) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, Cho, P. S. Angew. Chem., Int. Ed. 2000, 39, 2533. H. J. Am. Chem. Soc. 1991, 113, 4092. (4) Forsyth, C. J.; Ahmed, F.; Cink, R. D.; Lee, C. S. J. Am. Chem. Soc. (10) The biphenyltertbutyl silyl (BPS) moiety had a tendency to migrate 1998, 120, 5597. to the secondary hydroxyl when more standard conditions (NaH, MeI) were (5) (a) Evans, D. A.; Fitch, D. M. Angew. Chem., Int. Ed. 2000, 39, 2536. employed. (b) Evans, D. A.; Fitch, D. M.; Smith, T. E.; Cee, V. J. J. Am. Chem. Soc. (11) Jacobsen, E. N.; Marko, I.; Mungall, W. S.; Schroder, G.; Sharpless, 2000, 122, 10033. K. B. J. Am. Chem. Soc. 1988, 110, 1968.

10.1021/ja0105055 CCC: $20.00 © 2001 American Chemical Society Published on Web 05/01/2001 Communications to the Editor J. Am. Chem. Soc., Vol. 123, No. 20, 2001 4835 of the alkyne (tBuLi; MeI) followed by desilylation (TBAF; 100% Scheme 3 yield, 2 steps) next led to alcohol (+)-15, which upon TEMPO oxidation14 afforded an unstable carboxylic acid 16;15 immediate hydrolysis of the acetonide with concomitant cyclization (FeCl3‚ 6H2O) and protection of the remaining secondary hydroxyl (TIPSCl, imid) furnished (-)-17. A two-step palladium-promoted hydrostannylation/iodination16 protocol completed construction of the desired vinyl iodide (-)-8 (83% yield, 2 steps).17 Additionally, 10-15% of the internal stannane was recovered after iodination.

Scheme 2

Extensive experimentation demonstrated the necessity of early installation of the C(28) vinyl stannane; thus vinyl stannane 4 was prepared as outlined in Scheme 4. Toward this end, alcohol (+)-213o was subjected to hydroxyl protection (DMBCl, KH) and desilylation (TBAF); subsequent exposure to the tin cuprate derived from hexamethylditin (MeLi, CuCN) followed by me- thylation (MeI, DMPU) provided (+)-22 in excellent overall yield (71%; 4 steps). Desilylation (TBAF), oxidation (SO3‚pyr), and Wittig olefination of the derived aldehyde (+)-23 with (+)-6 then proceeded smoothly to afford alkene (+)-4 (20:1 E:Z). Unfortu- nately, all attempts to introduce the C(1-2) moiety, involving removal of the BPS group and oxidation to the C(3) aldehyde, proved unsuccessful due presumably to the sensitivity of the trimethyl tin moiety to the oxidative conditions. We therefore turned to the union of the side chain fragment (-)-20 with (+)- 4, exploiting the bifunctional oxazole linchpin 9. This possibility nicely demonstrated the flexibility of the overall coupling strategy.

Assembly of the Stille coupling partner (-)-7 began with Scheme 4 known TBS-glycidol (+)-18 (Scheme 3).18 Exposure to lithium 19 TMS acetylide in the presence of BF3‚OEt2, methylation (MeOTf, DTBMP), and selective removal of the TBS group in the presence of the TMS alkyne (cat. HCl, MeOH) furnished known alcohol (-)-1920 (69% yield, 3 steps). Parikh-Doering21 oxidation then provided the corresponding aldehyde without epimerization; alternate oxidation protocols (i.e., Swern) led to 22 23 epimerization at C(43). Hodgson homologation (CrCl2,Bu3- SnCHBr2, LiI, THF/DMF) of the derived aldehyde next afforded vinyl stannane (-)-7 as a single isomer (77%). The crucial Stille coupling24 of (-)-7 and vinyl iodide (-)-8 was then achieved 25 with Pd2(dba)3‚CHCl3 in the presence of Ph2PO2NBu4 (DMF, room temperature, 4 h) to furnish (-)-20 in near quantitative yield. The required oxazole 9 was prepared exploiting a method (12) For use of the Sharpless AD reaction with enynes, see: Jeong, K.-S.; developed by Sheehan in 1949 (Scheme 5) for the synthesis of Sjo, P.; Sharpless, K. B. Tetrahedron Lett. 1992, 33, 3833. Diminished 26 diastereoselectivity with homoallylic enynols has been reported: Caddick, S.; oxazolones. Bromoacetyl bromide was exposed to silver iso- Shanmugathasan, S.; Brasseur, D.; Delisser, V. M. Tetrahedron Lett. 1997, cyanate (30 min, Et2O), filtered, and then subjected to alcohol- 27 38, 5735. free diazomethane; immediate triflation (Et3N, Tf2O, THF, -78 (13) Since the Z isomer was markedly less reactive than the E isomer in °C to room temperature)3q furnished triflate 9 in 48% overall the dihydroxylation reaction, the E/Z mixture could be used directly. 28 (14) Zhao, M.; Li, J.; Mano, E.; Song, Z.; Tschaen, D. M.; Grabowski, E. yield. J. J.; Reider, P. J. J. Org. Chem. 1999, 2564. (15) Occasionally, the carboxylic acid would undergo cyclization to the Scheme 5 corresponding lactone during workup or chromatography. (16) Zhang, H. X.; Guibe, F.; Balavoine, G. J. Org. Chem. 1990, 55, 1857. (17) Exposure of alkyne (-)-17 to the Schwartz hydrozirconation and subsequent exposure to NIS, I2, or NBS failed to afford the desired vinyl halide; instead, starting material or decomposition was observed. (18) Prepared in one step from S-glycidol; see: Cywin, C. L.; Webster, F. X.; Kallmerten, J. J. Org. Chem. 1991, 56, 2953. The stage was now set for the union of (+)-4 with (-)-20 (19) Eis, M. J.; Wrobel, J. E.; Ganem, B. J. Am. Chem. Soc. 1984, 106, 3693. utilizing 9. After optimization we found that i-PrMgCl promoted (20) Alcohol (-)-19 was first prepared by Pattenden et al. from malic acid the coupling of bromide 9 with lactone (-)-20 to afford a single in 11 steps (ref 3e); Williams later prepared (-)-19 (ref 3t). hemiketal29 in excellent yield (Scheme 6). Presumably, Grignard (21) Parikh, J. R.; v. E. Doering, W. J. Am. Chem. Soc. 1967, 89, 5505. exchange generates the metalated oxazole that subsequently (22) Epimerization was determined by reduction (BH3‚THF) to alcohol (-)- 19 and comparison of optical rotations. attacks the lactone. Interestingly, premixing of the coupling (23) Hodgson, D. M.; Boulton, L. T.; Maw, G. N. Tetrahedron 1995, 51, 3713. (26) Sheehan, J. C.; Izzo, P. T. J. Am. Chem. Soc. 1949, 71, 4059. (24) (a) Farina, V.; Krishnamurthy, V.; Scott, W. J. Organic Reactions; (27) (a) DeBoer, T. J.; Backer, H. J. Org. Synth. 1956, 36, 16. (b) Aldrich Wiley: New York, 1997. (b) Stille, J. Angew. Chem., Int. Ed. Engl. 1986, Technical Bulletin AL-121. 25, 508. (28) In this three-step process, intermediates were not purified or isolated; (25) This salt was introduced by Liebeskind to remove Bu3SnI from the thus, the assembly of 9 is possible in a matter of hours. reaction mixture and thereby accelerate the Stille coupling process: Srogl, J.; (29) Presumably, the sterochemical outcome is due to anomeric effects; Allred, G. D.; Liebeskind, L. S. J. Am. Chem. Soc. 1997, 119, 12376. see: Bonner, W. A. J. Am. Chem. Soc. 1959, 81, 1448. 4836 J. Am. Chem. Soc., Vol. 123, No. 20, 2001 Communications to the Editor

Scheme 6

partners before addition of i-PrMgCl was required to minimize Scheme 7 the dimerization of 9 (arising from the electrophilic nature of unreacted 9).30 Methyl ketal formation (pTSA, MeOH, 35 °C) completed the synthesis of the C(29-41) side chain triflate (-)- 3. Stille coupling of (-)-3 with vinyl stannane (+)-4 [Pd(PPh3)4, LiCl, 100 °C, 24 h] furnished adduct (+)-24 in a 72% yield (Scheme 6). Selective removal of the BPS group (KOH, 18-cr- 6), oxidation (Dess-Martin), and removal of the DMB group (DDQ) then afforded hydroxyaldehyde (+)-25. Appendage of a C(1-2) phosphonate moiety31 at C(24), followed by a Still- modified Horner-Emmons macrocyclization32 provided (+)-26; interestingly the Z/E selectivity improved with higher tempera- ture.33 Having arrived at the complete phorboxazole skeleton, access to the terminal vinyl bromide proved to be an unexpected challenge. Radical promoted hydrostannylation (Bu3SnH, AIBN, ∆;orBu3SnH, Et3B, room temperature), although providing a mixture of the external to internal vinyl stannanes (4:1), led to isomerization at the C(2-3) olefin. Alternatively, palladium- of the synthetic venture include the use of modified Petasis- mediated hydrostannylation [PdCl2(PPh3)2,Bu3SnH; NBS] gave predominately the internal [C(45)] bromide, albeit with no C(2,3) Ferrier rearrangements for the effective assembly of both the isomerization. Success was eventually found in the three-step C(11-15) and C(22-26) cis-tetrahydropyan rings; the design, procedure of Guibe,16 which exploited the terminal alkynyl synthesis, and application of a novel bifunctional oxazole linchpin; bromide for enhanced diastereoselectivity. Global deprotection and the preparation and Stille coupling of a C(28) trimethylstan- (6% HCl, THF, 72 h) then afforded (+)-phorboxazole A (1), nane. The longest linear sequence leading to (+)-phorboxazole which displayed spectral data identical in all respects to that A(1) was 27 steps, with an overall yield of 3%. reported for the natural material [1H NMR (600 MHz), COSY, ROESY, HRMS, UV λ max, optical rotation). Acknowledgment. Support was provided by the National Institutes In summary, a highly convergent, stereocontrolled total syn- of Health (National Cancer Institute) through grant CA-19033. An thesis of (+)-phorboxazole A (1) has been achieved. Highlights American Chemical Society, Division of Organic Chemistry Fellowship (funded by DuPont Pharmaceuticals, Inc.) is also gratefully acknowledged (30) Dimerization was also observed under inverse addition conditions (i.e., (K.P.M.). slow addition of the oxazole to a -100 °C solution of tBuLi). (31) Pickering, D. A. Ph.D. Thesis, University of Minnesota, 1996. (32) (a) Stil, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405; (b) For Supporting Information Available: Spectroscopic and analytical data the use of K2CO3/18-crown-6 in Horner-Emmons reactions, see: Aristoff, P. - - A. J. Org. Chem. 1981, 46, 1954. Also see: Nicolaou, K. C.; Seitz, S. P.; for compounds 1, 4, 7 9, 20, and 22 26, and selected experimental Pavia, M. R. J. Am. Chem. Soc. 1982, 104, 2030. procedures (PDF). This material is available free of charge via the Internet (33) We suspect that higher temperatures accelerate collapse of the at http://pubs.acs.org. intermediate oxaphosphatane, thereby minimizing equilibration to the more stable E isomer. JA0105055 10942 J. Am. Chem. Soc. 2001, 123, 10942-10953

Total Synthesis of (+)-Phorboxazole A Exploiting the Petasis-Ferrier Rearrangement

Amos B. Smith, III,* Kevin P. Minbiole, Patrick R. Verhoest, and Michael Schelhaas Contribution from the Department of Chemistry, Monell Chemical Senses Center, and Laboratory for Research on the Structure of Matter, UniVersity of PennsylVania, Philadelphia, PennsylVania 19104 ReceiVed June 29, 2001

Abstract: A highly convergent, stereocontrolled total synthesis of the potent antiproliferative agent (+)- phorboxazole A (1) has been achieved. Highlights of the synthesis include: modified Petasis-Ferrier rearrangements for assembly of both the C(11-15) and C(22-26) cis-tetrahydropyran rings; extension of the Julia olefination to the synthesis of enol ethers; the design, synthesis, and application of a novel bifunctional oxazole linchpin; and Stille coupling of a C(28) trimethyl stannane with a C(29) oxazole triflate. The longest linear sequence leading to (+)-phorboxazole A (1) was 27 steps, with an overall yield of 3%.

Marine sponges comprise a rich source of architecturally the phorboxazoles to the level of premier medicinal targets. complex, biomedically important natural products; examples Bioassays against the National Cancer Institute panel of 60 include the spongistatins, discodermolide, and the tedanolides.1 human solid tumor cell lines revealed extraordinary activity 2 -9 Despite the structural complexity, the scarcity of these molecules against the entire panel; the mean GI50 value was 1.58 × 10 in conjunction with their medicinal importance continues to M for both 1 and 2.3a Some cell lines were completely inhibited prompt intense synthetic campaigns. During a recent search for at the lowest level tested.2 Particularly noteworthy, phorboxazole novel marine antifungals, Searle and Molinski2 identified a A(1) inhibited the human colon tumor cell line HCT-116 and methanolic extract from the sponge Phorbas sp. which displayed the breast cancer cell line MCF7 with GI50 values of 4.36 × significant activity against Candida albicans. Bioassay-guided 10-10 M and 5.62 × 10-10 M, respectively. These data place extraction, flash chromatography, and subsequent reverse-phase the phorboxazoles in the company of the spongistatins,1a HPLC afforded two isomeric macrolides termed (+)-phorboxa- collectively the most potent cytostatic agents discovered to date. zoles A (1)andB(2). The structures of the phorboxazoles, Although the precise biochemical mode of action remains including relative and absolute stereochemistry, were determined undefined, (+)-phorboxazole A (1) is known to arrest the cell via a combination of NMR analyses, degradation studies, and cycle in S phase but does not inhibit tubulin polymerization or synthetic correlations.3 interfere with the integrity of microtubules. Unfortunately, further biological analysis is not possible, because access to the producing sponge is currently restricted.4 Thus, the phorboxa- zoles will be only available via total synthesis. Not surprisingly, the novel architecture combined with the impressive bioactivity has attracted wide attention in the synthetic community,5 including our own interest.6 In 1998, Forsyth and co-workers4 published the first total synthesis of (+)-phorboxazole A (1); shortly thereafter, Evans and Fitch reported completion of (+)- (4) Forsyth, C. J.; Ahmed, F.; Cink, R. D.; Lee, C. S. J. Am. Chem. Soc. 1998, 120, 5597. (5) (a) Lee, C. S.; Forsyth, C. J. Tetrahedron Lett. 1996, 37, 6449. (b) Cink, R. D.; Forsyth, C. J. J. Org. Chem. 1997, 62, 5672. (c) Ahmed, F.; Forsyth, C. J. Tetrahedron Lett. 1998, 39, 183. (d) Ye, T.; Pattenden, G. Tetrahedron Lett. 1998, 39, 319. (e) Pattenden, G.; Plowright, A. T.; Tornos, The bioactivity profile of the phorboxazoles proved excep- J. A.; Ye, T. Tetrahedron Lett. 1998, 39, 6099. (f) Paterson, I.; Arnott, E. A. Tetrahedron Lett. 1998, 39, 7185. (g) Wolbers, P.; Hoffmann, H. M. R. tional. In addition to the antifungal activity, the phorboxazoles Tetrahedron 1999, 55, 1905. (h) Misske, A. M.; Hoffmann, H. M. R. displayed antibiotic activity against saccharomyces carlsber- Tetrahedron 1999, 55, 4315. (i) Williams, D. R.; Clark, M. P.; Berliner, ensis. However, it was the antiproliferative activity that elevated M. A. Tetrahedron Lett. 1999, 40, 2287. (j) Williams, D. R.; Clark, M. P. Tetrahedron Lett. 1999, 40, 2291. (k) Wolbers, P.; Hoffmann, H. M. R. (1) (a) Spongistatin: Pettit, G. R.; Cichacz, Z. A.; Gao, F.; Herald, C. Synthesis 1999, 797. (l) Evans, D. A.; Cee, V. J.; Smith, T. E.; Santiago, L.; Boyd, M. R.; Schmidt, J. M.; Hooper, J. N. A. J. Org. Chem. 1993, 58, K. J. Org. Lett. 1999, 1, 87. (m) Wolbers, P.; Misske, A. M.; Hoffmann, 1302. (b) Discodermolide: Gunasekera, S. P.; Gunasekera, M.; Longley, H. M. R. Tetrahedron Lett. 1999, 40, 4527. (n) Wolbers, P.; Hoffmann, H. R. E.; Schulte, G. K. J. Org. Chem. 1990, 55, 4912. Correction: Gunasekera, M. R.; Sasse, F. Synlett 1999, 11, 1808. (o) Pattenden, G.; Plowright, A. T. S. P.; Gunasekera, M.; Longley, R. E.; Schulte, G. K. J. Org. Chem. 1991, Tetrahedron Lett. 2000, 41, 983. (p) Schaus, J. V.; Panek, J. S. Org. Lett. 56, 1346. (c) Tedanolide: Schmitz, F. J.; Gunasekera, S. P.; Yalamanchili, 2000, 2, 469. (q) Rychnovsky, S. D.; Thomas, C. R. Org. Lett. 2000, 2, G.; Hossain, M. B.; van der Helm, D. J. Am. Chem. Soc. 1984, 106, 7251. 1217. (r) Williams, D. R.; Clark, M. P.; Emde, U.; Berliner, M. A. Org. (2) Searle, P. A.; Molinski, T. F. J. Am. Chem. Soc. 1995, 117, 8126. Lett. 2000, 2, 3023. (s) Greer, P. B.; Donaldson, W. A. Tetrahedron Lett. (3) (a) Searle, P. A.; Molinski, T. F.; Brzezinski, L. J.; Leahy, J. W. J. 2000, 41, 3801. (t) Evans, D. A.; Cee, V. J.; Smith, T. E.; Fitch, D. M.; Am. Chem. Soc. 1996, 118, 9422. (b) Molinski, T. F. Tetrahedron Lett. Cho, P. S. Angew. Chem., Int. Ed. 2000, 39, 2533. (u) Huang, H.; Panek, 1996, 37, 7879. J. S. Org. Lett. 2001, 3, 1693.

10.1021/ja011604l CCC: $20.00 © 2001 American Chemical Society Published on Web 10/13/2001 Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10943 phorboxazole B (2).7 Herein, we disclose a full account of the diene, and an E-vinyl bromide. From the retrosynthetic perspec- total synthesis of (+)-phorboxazole A (1) recently completed tive (Scheme 3), we envisioned disconnection of the phorboxa- in our laboratory.6c A central feature of this synthetic venture zoles at C(2-3), C(19-20), and C(28-29) to reveal three was the exploitation of the Petasis-Ferrier rearrangement for subtargets (10, 11, and 12) of comparable structural complexity. the construction of the two 2,6-cis-tetrahydropyran rings resident In the synthetic sense, fragments 11 and 12 would be united in the phorboxazole macrolide ring. via a Wittig reaction. Continuing with this analysis, disconnec- Petasis-Ferrier Rearrangement. In 1996, Petasis reported tion of side chain 10 at C(40-41) and C(32-33) would furnish that the acid-promoted rearrangement of enol acetals to tetra- subtargets 13, 14, and 15. In the forward sense, vinyl stannane f hydropyranones (e.g., 3 4, Scheme 1) proceeds via fragmenta- 14 and vinyl iodide 13 could be coupled via a Stille reaction. 8 tion, followed by endo cyclization onto an oxocarbenium (ii), For union of the side chain to the macrocycle, we planned to 9 a reaction closely related to the earlier Ferrier Type-II enol exploit oxazole triflate 15a,b as a novel bidirectional linchpin f ether rearrangement (e.g., 5 6) induced by mercuric ion. (vide infra). Finally, the cornerstone for construction of the - Scheme 1 central C(20 28) tetrahydropyran, 11, and bistetrahydropyran 12 would be the Petasis-Ferrier rearrangements, respectively, of vinyl acetals 16 and 17. Importantly, the overall synthetic strategy held the promise of considerable flexibility for fragment assembly, their union, endgame operations (vide infra), and the construction of analogues.

Scheme 3

Inspection of the Petasis-Ferrier rearrangement in the context of complex molecule synthesis revealed two important attributes. First, construction of the enol acetal rearrangement substrates comprises an ideal linchpin tactic for complex fragment assembly; second, the latent element of symmetry inherent in the target cis-tetrahydropyranones permits rearrangement of either enol acetal 8 or 9 (Scheme 2). Both attributes provide considerable latitude for fragment union and thereby cis- tetrahydropyranone construction. Scheme 2

Synthetic Analysis. In addition to the two 2,6-cis-fused tetrahydropyrans (vide supra), the phorboxazoles present a wide array of architectural features, including a 21-membered macro- lactone, a trans-fused tetrahydropyran, two oxazoles, and six olefinic units: one Z and two E , an exomethylene, a

(6) (a) Smith, A. B., III; Verhoest, P. R.; Minbiole, K. P.; Lim, J. J. - Org. Lett. 1999, 1, 909. (b) Smith, A. B., III; Minbiole, K. P.; Verhoest, P. Bistetrahydropyran 12: The C(3 19) Subtarget. To R.; Beauchamp, T. J. Org. Lett. 1999, 1, 913. (c) Smith, A. B., III; Verhoest, implement the first Petasis-Ferrier rearrangement, we sought P. R.; Minbiole, K. P.; Schelhaas, M. J. Am. Chem. Soc. 2001, 123, 4834. enol acetal 17. Our point of departure entailed preparation of (7) (a) Evans, D. A.; Fitch, D. M. Angew. Chem., Int. Ed. 2000, 39, 10 2536. (b) Evans, D. A.; Fitch, D. M.; Smith, T. E.; Cee, V. J. J. Am. Chem. trans-tetrahydropyran 24 from known aldehyde 18 (Scheme 4). Soc. 2000, 122, 10033. (8) Petasis, N. A.; Lu, S.-P. Tetrahedron Lett. 1996, 37, 141. (10) Evans, D. A.; Kaldor, S. W.; Jones, T. K.; Clardy, J.; Stout, T. J. J. (9) Ferrier, R. J.; Middleton, S. Chem. ReV. 1993, 93, 2779. Am. Chem. Soc. 1990, 112, 7001. 10944 J. Am. Chem. Soc., Vol. 123, No. 44, 2001 Smith et al.

Scheme 4 followed immediately by TMSOTf -promoted19 condensation with aldehyde (-)-24, afforded cis-dioxanone (-)-30 in 61% yield, along with 18% of the trans isomer, the latter readily removed by flash chromatography. Methylenation exploiting the 20 Petasis-Tebbe reagent (Cp2TiMe2) then led to rearrangement substrate (-)-17. Unfortunately, all attempts to effect the rearrangement employing the conditions prescribed by Petasis8 failed to produce the desired 2,6-cis-tetrahydropyran.

Scheme 5

Toward this end, Brown asymmetric allylation11 and protection (TBSCl) of the resultant alcohol furnished silyl ether (+)-19 (91% ee via Mosher ester analysis).12 Oxidative removal of the 13 PMB ether (DDQ, H2O) followed by oxidation (PCC) afforded aldehyde (+)-20; a second Brown allylation orchestrated the requisite 1,3-trans stereochemical relationship with excellent selectivity (96%, 10:1 diastereomeric ratio, dr). Differential hydroxyl protection (TESCl, imidazole) then furnished silyl ether (+)-21, which upon exhaustive ozonolysis generated an unstable bisaldehyde; immediate deprotection (AcOH, THF, H2O) with concomitant cyclization and acetylation yielded 22 as an inconsequential mixture (2:1 eq/ax) of acetals (65%, 3 steps). Reduction (NaBH4), protection of the resultant alcohols (BPSCl), and axial addition of silyl enol ether 2314 then led to aldehyde (-)-24 as a single isomer (72%). The relative stereochemistry of (-)-24 was established via two-dimensional NOE experi- Improved Conditions for the Petasis-Ferrier Rearrange- ments.15 ment. Failure of the prescribed Petasis conditions led us to Construction of â-hydroxyacid 29, required for elaboration explore other Lewis acids. Increasing the Lewis acidity was of of the Petasis-Ferrier substrate 17 (Scheme 5), entailed primary concern. The lack of selectivity of the subsequent condensation of oxazole aldehyde 25, prepared independently carbonyl reduction, inherent with i-Bu3Al, was also identified in both the Williams5i and our laboratories, with the known as a significant liability. Thus, promoters incapable of reducing benzyl trimethylsilylketene acetal 2616,17 exploiting the Carreira the initially derived tetrahydropyranone were sought. enantioselective aldol18 tactic to afford benzyl ester (+)-28 in To preserve valuable intermediates, we prepared model enol 84% yield with g 98% ee.12 Removal of the benzyl ester (LiOH; ethers 35 and 36 (Scheme 6). A variety of Lewis acids were ∼100%), followed by dioxanone construction (one-pot), initiated by bis-silylation of (+)-29 with hexamethyldisilazane (HMDS), Scheme 6

(11) Brown, H. C.; Ramachandran, P. V. Pure Appl. Chem. 1991, 63, 307. (12) (a) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512. (b) Sullivan, G. R.; Dale, J. A.; Mosher, H. S. J. Org. Chem. 1973, 38, 2143. (c) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc. 1991, 113, 4092. (13) Oikawa, Y.; Yushioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982, 23, 885. (14) Jung, M. E.; Blum, R. B. Tetrahedron Lett. 1977, 3791. (15) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys. 1979, 71, 4546. (16) Slougui, N.; Rousseau, G.; Conia, J.-M. Synthesis 1982, 58. (17) The condensation was first attempted using bistrimethylsilyl ketene acetal 122, to obtain the desired â-hydroxy acid 29 directly; unfortunately, this approach met with little success (30% yield, 10% ee).

(18) Carreira, E. M.; Singer, R. A.; Lee, W. J. Am. Chem. Soc. 1994, 116, 8837. Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10945 screened; best results were obtained with Me2AlCl. Importantly, Scheme 9 Me2AlCl did not reduce the derived ketone. Moreover, the tert- butylbiphenyl (BPS) ether moiety (e.g., 35) was found to tolerate the rearrangement conditions, an important requirement for application of the Petasis-Ferrier transform in complex mol- ecule synthesis. Notwithstanding the improved conditions, enol ether (-)-17 again failed to undergo rearrangement. We surmised that preferred coordination of the Me2AlCl promoter with the neighboring oxazole nitrogen precluded productive Lewis acid chelation to the requisite enol ether oxygen in (-)-17, thereby preventing rearrangement (Scheme 7).

Scheme 7

Scheme 10 Productive Chelation. To circumvent the unproductive chelation, we examined rearrangement substrate 41, obtained by transposition of the enol ether oxygen permitted by the symmetry inherent in the linchpin construction of tetrahydro- pyranones (Scheme 8). In this way, initial coordination of 21 bidentate Lewis acid Me2AlCl with the oxazole nitrogen would allow productive activation of the enol ether oxygen (i), liberating the aluminum enolate, which in turn would rearrange to the tetrahydropyranone (iv). The transposed substrate (41) possessed two additional advantages: the oxazole acetal could lead to a resonance-stabilized oxocarbenium ion (i.e., iii), and the rearrangement would proceed via a more facile 6-exo-trig ring closure,22 compared to the 6-endo closure required for the unactivated Petasis-Ferrier vinyl acetals.

Scheme 8

to â-hydroxy acid (-)-47. The previously developed two-step sequence involving initial bis-silylation (HMDS) of (-)-47 followed by TMSOTf-catalyzed19 condensation with aldehyde 25 furnished dioxanone (-)-48 in 65% yield (10:1 dr). Selective removal of the trimethylsilyl group (HF‚pyr), oxidation (Dess- 25 Our attention thus turned to rearrangement substrate 43, which Martin), and treatment with excess Cp2TiMe2 (5 equiv) was readily constructed from previously prepared aldehyde 25 installed both the C(7) exomethylene and the C(13) enol ether and â-hydroxyacid 44 (Scheme 9). The Nagao acetate aldol23 to provide rearrangement substrate (-)-43. protocol was selected to install the C(11) stereocenter in 44 To our delight, treatment of (-)-43 with Me2AlCl at ambient (Scheme 10). Alcohol (+)-46 was obtained in 85% yield temperature rapidly (2 min) furnished tetrahydropyranone (-)- (4:1 dr, unoptimized). 42 as a single isomer in 89% yield (Scheme 11). Interestingly, Hydrolytic removal of the auxiliary exploiting basic hydrogen exposure of (-)-43 to the original Petasis conditions (i-Bu3- peroxide, followed by selective desilylation (H SiF ),24 then led 8 2 6 Al) led only to recovered starting material. Failure of i-Bu3Al, (19) (a) Seebach, D.; Imwinkelried, R.; Stucky, G. HelV. Chim. Acta a monocoordinate Lewis acid, to effect rearrangement supports 1987, 70, 448. (b) Noyori, R.; Murata, S.; Suzuki, M. Tetrahedron, 1981, the bis-chelation model for the rearrangement of (-)-43 (see 37, 3899. Scheme 8). (20) Petasis, N. A.; Bzowej, E. I. J. Am. Chem. Soc. 1990, 112, 6392. (21) For a discussion of the chelating ability of Me2AlCl, see: Evans, C(3-19) Subtarget (-)-42: A Second Generation Syn- D. A.; Allison, B. D.; Yang, M. G. Tetrahedron Lett. 1999, 40, 4457. thesis. To access (-)-42 on large scale, a second-generation (22) Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734. (23) Nagao, Y.; Yamada, S.; Kumagai, T.; Ochiai, M.; Fujita, E. J. Chem. synthesis was developed (Scheme 12). Asymmetric hetero Soc., Chem. Commun. 1985, 1418. (24) Pilcher, A. S.; DeShong, P. J. Org. Chem. 1993, 58, 5130. (25) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277. 10946 J. Am. Chem. Soc., Vol. 123, No. 44, 2001 Smith et al.

Scheme 11 To arrive at the C(3-19) subtarget (-)-12, five steps were required (Scheme 13): reduction of the C(13) ketone (K- selectride; 9:1 dr),32 silylation (TBSOTf, 2,6-lutidine), oxidative removal of the PMB ether (DDQ), generation of the primary 33 chloride (PPh3, CCl4), and displacement with tributyl phos- phine. Each step proceeded in excellent yield to provide phosphonium salt (-)-12 in 86% overall yield from (-)-42.

Scheme 13

Diels-Alder reaction26 of aldehyde 4927 with Danishefsky’s 28 diene, promoted by R-(+)-Binol/Ti(Oi-Pr)4 (10 mol %), furnished enone (-)-50 in 64% yield (90% ee).29 Importantly, this reaction could be run on large scale (∼20 g). Axial addition of vinyl cuprate then furnished trans-tetrahydropyranone (30:1 trans/cis), which in turn was subjected to chemoselective hydroboration,30 Wittig olefination, and Swern oxidation to afford aldehyde (-)-51 (60% yield, 4 steps). A second Nagao aldol reaction, with the tin enolate derived from (-)-45,23 followed by hydrolysis (LiOH, H2O2) gave â-hydroxy acid (-)- 5231 in 90% yield (2 steps, 10:1 dr). Dioxanone (-)-53 was then constructed via the now-standard HMDS-promoted bis- silylation of (-)-52 and condensation with the requisite oxazole aldehyde 25 (71%, 99% BORSM, 10:1 dr). Petasis-Tebbe methylenation (Cp2TiMe2) provided enol ether (-)-43 and, thereby, intersection with the previous synthetic sequence. The - second-generation assembly of (-)-42, proceeding in 10 steps The C(22 26) Central Tetrahydropyran. Although from the outset we envisioned the Petasis-Ferrier rearrangement to (21% overall yield), constituted a significant improvement over + the initial route (20 steps, 4.5% overall yield). be the cornerstone of the ( )-phorboxazole A (1) synthetic venture, the fully substituted central tetrahydropyran ring raised Scheme 12 the level of synthetic challenge given the requirement of a Z-exo- ethylidene acetal, instead of the simpler methylidene acetal employed to construct the C(11-15) tetrahydropyran (Scheme 14). Nonetheless, we envisioned that Lewis acid complexation to the enol ether oxygen in Z-enol acetal 16 would trigger ring opening, liberating (reversibly) the aluminum enolate (i). A least motion pathway, involving rotation by 90° with intervention of a boat conformation (e.g., ii) followed by reclosure of the enolate on the oxocarbenium ion was expected to afford 55, possessing the C(23) axial methyl. The synthetic challenge in this scenario would be efficient access to Z-enol acetal 16.

Scheme 14

We began with an Oppolzer anti aldol reaction34 (Scheme 15). Addition of the boron enolate of known propionyl sultam (30) Evans, D. A.; Fu, G. C.; Hoveyda, A. H. J. Am. Chem. Soc. 1992, 114, 6671. Interestingly, when the reaction was performed at 0 °C instead (26) Keck, G. E.; Li, X.-Y.; Krishnamurthy, D. J. Org. Chem. 1995, 60, of room temperature, ketone reduction was competitive with hydroboration. 5998. (31) The absolute configuration at C(11) was secured by Mosher ester (27) Boeckman, R. K., Jr.; Charette, A. B.; Asberom, T.; Johnston, B. analysis; see ref 12. H. J. Am. Chem. Soc. 1987, 109, 7553. (32) Reduction of the C(13) ketone to the equatorial alcohol (NaBH4) (28) Danishefsky, S. Acc. Chem. Res. 1981, 14, 400. Danishefsky, S. would provide access to (+)-phorboxazole B (2). Chemtracts: Org. Chem. 1989, 2, 273. (33) The corresponding primary iodide was prone to reduction by PBu3 (29) Enantiomeric excess was determined after Nagao aldol condensation to afford the corresponding methyl oxazole; thus, the chloride was used. by 500 MHz NMR analysis of the diastereomeric ratio. (34) Oppolzer, W.; Lienard, P. Tetrahedron Lett. 1993, 34, 4321. Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10947

Scheme 15 1-D NOE experiments. This unexpected outcome led us to reexamine the proposed rearrangement scenario. Presumably, the least motion pathway of aluminum enolate (i) did not occur, because of the increase in steric demands of the corresponding boat conformation (vide infra); instead, rotation by 180° with closure via a chair conformation (ii) furnished (+)-69.

Scheme 16

Construction of the Central C(22-26) Tetrahydropyran via the Alternate Petasis-Ferrier Rearrangement Substrate. For a second time, we resorted to the pseudosymmetry available - 34 ) 35 ( )-56 to known aldehyde 57 (R TMS) in the presence in the linchpin construction of the Petasis-Ferrier rearrangement - of excess TiCl4 furnished ( )-59 as a single isomer. Oppolzer substrates, which dictates that two possible vinyl acetals, related 34 attributed the anti selectivity to an open transition state. by the transposition of the enol ether oxygen in the substrate, Removal of the sultam with concomitant alkyne desilylation can provide access to the requisite tetrahydropyranone. With + - then afforded â-hydroxy acid ( )-61 and recovered sultam ( )- this in mind, we envisioned oxygen-transposed enol ether 70 + 63, both in good yield. The two-step condensation of ( )-61 as a viable rearrangement substrate (Scheme 17). Rearrangement 27 with aldehyde 49 provided dioxanone 64 in 59% yield, albeit involving a 180° bond rotation would lead, now via a chair as a disappointing mixture of separable epimers (3:2) favoring conformation (ii), to 71 possessing the requisite axial methyl 36 the cis-dioxanone. The poor selectivity is attributed to the low substituent at C(23). Critical to this scenario would be the 37 steric bulk of the alkyne. Ethylidenation a` la Takai unfortu- availability of 70 possessing the Z-ethylidene geometry. nately proved unsuccessful, despite exploration of a variety of conditions; only decomposition occurred.38 To provide the Scheme 17 alkyne with a measure of protection, the TIPS-alkyne congener 65 was prepared (Scheme 15).39 Again, exposure to either the Takai or related carbenoid ethylidenation conditions40 failed to afford the desired product. To confirm that the alkyne indeed was prone to decomposi- tion,38 we prepared the analogous alkane (-)-67 via hydrogena- tion (Scheme 16).41 As expected, (-)-67 could be converted readily with modest stereoselectivity (5:1 Z/E)37 to 68 via the Takai ethylidenation (54%); flash chromatography provided Z-alkene (-)-68. To our surprise, however, execution of the Petasis-Ferrier rearrangement furnished the all equatorial tetrahydropyran (+)-69 (58%, unoptimized), the latter assigned via detailed NMR coupling constant analysis in conjunction with Assembly of 70 began with an Evans boron aldol condensa- tion of oxazolidinone (+)-7242 with aldehyde 49 (Scheme 18);27 (35) Kruithof, K. J. H.; Schmitz, R. F.; Klumpp, G. W. Tetrahedron 1983, 39, 3073. removal of the auxiliary (H2O2, LiOH) afforded â-hydroxy acid (36) At this point, dioxanone 64 was used as a mixture. The cis-dioxanone (+)-73 (84%, 2 steps). Silylation followed by TMSOTf- isomer was later purified by crystallization; see ref 41. promoted19 union with aldehyde 58 then furnished dioxanone (37) Okazoe, T.; Takai, K.; Oshima, K.; Utimoto, K. J. Org. Chem. 1987, + 52, 4410. ( )-74. Initial difficulties in the scale-up of this reaction (38) Alkyne reactivity with carbenoid species has been reported. See: suggested that triflic acid was the actual catalyst; large scale Takeda, T.; Shimokawa, H.; Miyachi, Y.; Fujiwara, T. Chem. Commun. reactions did not proceed until catalytic triflic acid (2-4 mol 1997, 1055. (39) Journet, M.; Cai, D.; DiMichele, L. M.; Larsen, R. D. Tetrahedron %) was added. We suspect that advantageous water, more Lett. 1998, 39, 6427. pronounced on a smaller scale, generated triflic acid in situ from 43 (40) Horikawa, Y.; Watanabe, M.; Fujiwara, T.; Takeda, T. J. Am. Chem. TMSOTf (as well as TMS2O). Yields and diastereoselectivity Soc. 1997, 119, 1127. (41) Alkyne (+)-64 was prepared in enantiomerically pure form by (42) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103, recrystallization from hexane. 2127. 10948 J. Am. Chem. Soc., Vol. 123, No. 44, 2001 Smith et al.

Scheme 18 namic control. Although we were pleased to find that treatment of sulfone (+)-80 with n-BuLi, followed by exposure to 1,1- chloroiodoethane (81)47 and i-PrMgCl (a 1:1 mixture) at -78 °C, furnished enol ether 70 in excellent yield (95%), the E/Z selectivity was nonexistent. Petasis-Ferrier Rearrangement of Enol Ether 70: A Pleasant Surprise. Notwithstanding the mixture of enol ethers 70, treatment with Me2AlCl afforded only the desired tetra- hydropyran (+)-71 in excellent yield (91%). Although the Z isomer of 70 rearranges as anticipated presumably via a chair transition state to (+)-71 (Scheme 17), formation of (+)-71 from the E isomer implies that the unfavorable 1,3-diaxial interactions in transition state ii (Scheme 20) preclude a chair conformation and instead favor a boatlike transition state (iii).

Scheme 20

were similar with the added triflic acid. The cis-2,6 stereo- chemistry was confirmed by two-dimensional NMR data, specifically an NOE between the C(22) and C(26) hydrogens.15 The trans dioxanone (-)-75, recovered in 19% yield, was readily recycled to â-hydroxyacid (+)-73 (LiOH, 97%). Unfortunately, all direct attempts to install the enol ether for the Petasis-Ferrier rearrangement again proved unrewarding. Julia Type-II Olefination: A New Tactic for Enol Ether Construction. Our failure to prepare enol ether 70 directly from the lactone provided an opportunity to extend the Type-II Julia olefination44 to the synthesis of enol ethers. This protocol, which was used to great advantage in our recent total synthesis of the spongistatins,45 calls for R-alkylation of a sulfone (76a; Scheme 19) with an electrophilic R-halo Grignard reagent (77); subse- - - quent elimination furnishes the alkene (79a). We reasoned that The C(1 28) Macrolide. With access to both ( )-12 and + - a similar reaction with sulfone 76b would afford 79b, contingent ( )-71, attention turned to the construction of the C(1 28) + on preferential expulsion of phenyl sulfinate over the . macrolide. Reduction of ( )-71 with NaBH4, protection of the Toward this end, DIBAL reduction of dioxanone (+)-74, resultant alcohol as the 3,4-dimethoxybenzyl (DMB) ether, ‚ followed by in situ acetylation of the alkoxide, furnished the removal of the silyl groups, and oxidation (SO3 pyr) furnished + - intermediate hemiketal acetate. A two-step sulfone installation aldehyde ( )-82 (82%, 4 steps). Wittig condensation with ( )- 12 then afforded the trans alkene (+)-83 both in excellent yield Scheme 19 (94%) and with high E/Z selectivity (12:1).5f In turn, removal of the BPS moiety in the presence of both TBS and DMB groups (KOH, 18-crown-6),48 oxidation (Dess-Martin),25 and removal of the DMB group (DDQ) then furnished hydroxyaldehyde (+)- 84 (Scheme 21). Final elaboration of the C(1-28) macrolide entailed two steps: attachment of a two-carbon ester fragment (e.g., 85)49 (EDCI‚MeI, HOBT), followed by an intramolecular Still- modified Horner-Emmons50 reaction to provide (+)-86,asa mixture of C(2-3) olefin isomers (4:1). Although pleased that the C(1-28) macrolide was in hand, we quickly discovered that subsequent installation of the C(27-28) vinyl stannane, em- (45) Smith, A. B., III; Doughty, V. A.; Lin, Q.; Zhuang, L.; McBriar, M. D.; Boldi, A. M.; Moser, W. H.; Murase, N.; Nakayama, K.; Sobukawa, M. Angew. Chem., Int. Ed. 2001, 40, 191. (46) Evans, D. A.; Trotter, B. W.; Coˆte´, B.; Coleman, P. J. Angew. Chem., Int. Ed. Engl. 1997, 36, 2741. (47) Simpson, M. Bull. Soc. Chim. Fr. 1879, 31, 411. (48) Although we believe these precise conditions are novel for the removal of a BPS group, very similar conditions exist. See: Torisawa, Y.; Shibasaki, M.; Ikegami, S. Chem. Pharm. Bull. 1983, 31, 2607. 46 (PhSTMS, ZnI2; m-CPBA) generated (+)-80 as a single isomer (49) Pickering, D. A. Ph.D. Thesis, University of Minnesota, MN, 1996. (60%, 3 steps); presumably, the initial step is under thermody- (50) (a) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405. (b) For the use of K2CO3/18-crown-6 in Horner-Emmons reactions, see: (43) Hollis, T. K.; Bosnich, B. J. Am. Chem. Soc. 1995, 117, 4570. Aristoff, P. A. J. Org. Chem. 1981, 46, 1954. Also see: Nicolaou, K. C.; (44) De Lima, C.; Julia, M.; Verpeaux, J.-N. Synlett 1992, 133. Seitz, S. P.; Pavia, M. R. J. Am. Chem. Soc. 1982, 104, 2030. Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10949

Scheme 21 iodide (DMPU) furnished trisubstituted olefin (+)-88. Elabora- tion of the C(20) aldehyde (+)-89 (TBAF; SO3‚pyr) followed by Wittig olefination with (-)-12 then led to trans olefin (+)- 90a as the sole product. Unfortunately, extensive experimenta- tion demonstrated that the labile trimethylstannane moiety was incompatible with the oxidation conditions to remove the DMB moiety, the required prelude to macrolide construction. Our attention thus turned to the assembly of the C(29-46) side chain, with the intent of attaching this unit prior to construction of the macrolide ring (Scheme 3). Phorboxazole Side Chain. As outlined earlier (Scheme 3), assembly of the C(29-46) side chain called for lactone 13, vinyl stannane 14, and the oxazole triflate linchpin, 15a,b. We began with construction of 13 (Scheme 23). Given that methylation

Scheme 23

ploying either cuprate51 or methylzirconation52 chemistry, as + 53 required to couple with the oxazole linchpin 15 (Scheme 3), of known homoallylic alcohol ( )-91 with sodium hydride ∼ - was not possible. affords significant silyl migration ( 10 15%), we resorted to - the less basic conditions of MeOTf in the presence of 2,6-di- Undaunted, we undertook introduction of the C(27 28) vinyl 54 stannane at an earlier stage, again exploiting the inherent tert-butyl-4-methylpyridine (DTBMP). Subsequent ozonolysis followed by reductive workup (PPh3) furnished aldehyde (-)- flexibility of the overall synthetic design. Toward that end, - liberation of the terminal alkyne in (+)-87 (Scheme 22), 92 (81%, 2 steps). Although Wittig condensation of ( )-92 with followed by addition of trimethylstannyl cuprate (Me Sn , MeLi, the Wittig salt 93a possessing the methyl alkyne moiety led to 6 2 ∼ 55 CuCN) and capture of the intermediate vinyl anion with methyl a disappointing mixture (E/Z) of olefins ( 2.2:1), condensation with the commercially available TMS phosphonate salt 93b afforded enyne (-)-94b with acceptable selectivity (97%, 5.5:1 Scheme 22 E/Z). Improvement in the E/Z ratio was observed employing toluene/THF (1:1) as the solvent system, albeit at the expense of yield and reproducibility (Scheme 23). Removal of the TMS group, followed by selective Sharpless dihydroxylation56 of the enyne57,58 using AD-Mix â, then afforded the corresponding

(51) Presumably, failure of cuprate addition to the alkyne arises from reaction at the existing Michael acceptors (i.e., the unsaturated lactone and vinyl oxazole). (52) This result did not take us by surprise; during the course of this work, methylzirconation was reported to fail with a similar propargyl ether. See: Barrett, A. G. M.; Bennett, A. J.; Menzer, S.; Smith, M. L.; White, A. J. P.; Williams, D. J. J. Org. Chem. 1999, 64, 162. (53) Clive, D. L. J.; Keshava Murthy, K. S.; Wee, A. G. H.; Prasad, J. S.; da Silva, G. V. J.; Majewski, M.; Anderson, P. C.; Haugen, R. D.; Heerze, L. D. J. Am. Chem. Soc. 1988, 110, 6914 (see Supporting Information). The ee of the alcohol prepared in our hands was determined by Mosher ester analysis to be 94%; see ref 12. (54) Ireland, R. E.; Gleason, J. L.; Gegnas, L. D.; Highsmith, T. K. J. Org. Chem. 1996, 61, 6856. (55) Attempts to improve this ratio using a Horner-type reaction were unsuccessful (50%, 5:3 E/Z). (56) Jacobsen, E. N.; Marko´, I.; Mungall, W. S.; Schro¨der, G.; Sharpless, K. B. J. Am. Chem. Soc. 1988, 110, 1968. (57) For use of the Sharpless AD reaction on enynes, see: Jeong, K.-S.; Sjo¨, P.; Sharpless, K. B. Tetrahedron Lett. 1992, 33, 3833. 10950 J. Am. Chem. Soc., Vol. 123, No. 44, 2001 Smith et al. diol (73%; 7:1 dr).59 Acetonide formation and alkyne methyl- followed by methylation exploiting again conditions to prevent ation completed assembly of (+)-95. silyl migration65 (MeOTf, DTBMP), and, in turn, removal of Continuing with construction of lactone 13 (Scheme 24), the TBS group in the presence of the TMS alkyne (cat. HCl, removal of the BPS group (TBAF) and oxidation via the Merck MeOH), furnished known alcohol (-)-101 (69% yield, 3 monophasic TEMPO protocol60 afforded an unstable61 acid (96) steps).66 Although several oxidation methods (e.g., Swern, in good yield; other oxidations (PDC/DMF or Dess-Martin; Dess-Martin) led to facile epimerization at the C(43) methoxy 67 NaClO2) proved less effective. Immediate exposure of the acid center, the Parikh-Doering protocol provided the aldehyde 62 68 (96) to FeCl3‚6H2O effected acetonide hydrolysis with con- as a single isomer (92% yield). Vinyl stannylation a`la 69 comitant lactonization (72%, 2 steps). More conventional acid Hodgson (CrCl2,Bu3SnCHBr2, THF/DMF) then afforded (-)- treatment (AcOH, ∆) resulted in lower yields (∼30%). Protec- 14 (77%). In the event, the critical Stille union70 of (-)-14 with tion of the secondary hydroxyl then furnished silyl ether (-)- vinyl iodide (-)-13 proceeded in excellent yield to furnish (-)- 97. To access directly vinyl iodide (-)-13 from alkyne (-)-97, 102.71 The success of this transformation is attributed to the 63 72 we explored Schwartz hydrozirconation; only recovered start- use of Ph2PO2NBu4, a salt introduced by Liebeskind to remove ing material or decomposition occurred. Fortunately, recourse Bu3SnI from the reaction mixture and thereby accelerate the to a two-step palladium-mediated sequence involving slow Stille coupling process. addition of excess Bu3SnH to (-)-97 in the presence of catalytic PdCl2(PPh3)2 yielded a mixture (5:1) of vinyl stannane regio- Scheme 25 isomers which were not readily separated. Exposure of the mixture to I2 (0 °C) provided the desired E-vinyl iodide (-)-13 (76% yield, 2 steps), with recovery of 10-15% of internal stannane (-)-99; presumably, the lack of reactivity of the internal stannane is due to steric constraints.

Scheme 24

Potential Bidirectional Linchpins: 2-Methyl and 2-Bromo- methyl 4-Trifloyloxazoles. With both the side chain lactone (-)-102 and vinyl stannane (+)-90 in hand, the stage was set for their union via an appropriate C(29-31) linchpin. We reasoned that either 2-methyl- or 2-bromomethyl-4-trifloyloxa- zole could serve this purpose.73 To construct the requisite

(65) Again, we found that standard methylation (NaH, MeI) led to a mixture of products. (66) Alcohol (-)-101 was prepared previously by Pattenden and co- workers from malic acid; see ref 5e. This alcohol was subsequently prepared by Williams; see ref 5r. (67) Parikh, J. R.; von E. Doering, W. J. Am. Chem. Soc. 1967, 89, 5505. (68) The extent of epimerization was determined by reduction (BH3‚ THF) to alcohol (-)-101 and comparison of optical rotations. (69) Hodgson, D. M.; Boulton, L. T.; Maw, G. N. Tetrahedron 1995, Assembly of vinyl stannane (-)-14 began with known TBS- 51, 3713. glycidol (+)-10064 (Scheme 25). Exposure to the lithium ion (70) (a) Farina, V.; Krishnamurthy, V.; Scott, W. J. Organic Reactions; derived from TMS acetylene in the presence of BF ‚OEt , Wiley: New York, 1997. (b) Stille, J. K. Angew. Chem., Int. Ed. Eng. 1986, 3 2 25, 508. (58) Experimentation revealed that an E/Z mixture of enynes could be (71) It is noteworthy that vinyl iodide (-)-123, prepared by exposure of used directly in the dihydroxylation; the Z isomer was markedly less reactive vinyl stannane (-)-14 to iodine (97%), did not undergo Stille coupling with than the E isomer. the previously prepared vinyl stannane 98 under identical conditions. (59) Diminished diastereoselectivity in AD reactions with homoallylic enynols has been reported: Caddick, S.; Shanmugathasan, S.; Brasseur, D.; Delisser, V. M. Tetrahedron Lett. 1997, 38, 5735. (60) Zhao, M.; Li, J.; Mano, E.; Song, Z.; Tschaen, D. M.; Grabowski, E. J. J.; Reider, P. J. J. Org. Chem. 1999, 64, 2564. (61) Carboxylic acid 96 in some cases was observed to lactonize in workup or chromatographic purification. (62) Sen, S. E.; Roach, S. L.; Boggs, J. K.; Ewing, G. J.; Magrath, J. J. Org. Chem. 1997, 62, 6684. (63) Hart, D. W.; Blackburn, T. F.; Schwartz, J. J. Am. Chem. Soc. 1975, (72) Srogl, J.; Allred, G. D.; Liebeskind, L. S. J. Am. Chem. Soc. 1997, 97, 679. 119, 12376. (64) Cywin, C. L.; Webster, F. X.; Kallmerten, J. J. Org. Chem. 1991, (73) (a) 4-Trifloyloxazoles have received only modest attention; see ref 56, 2953. 5p. (b) Kelly, T. R.; Lang, F. J. Org. Chem. 1996, 61, 4623. Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10951

Scheme 26 Application of this tactic to the (+)-phorboxazole side chain precursor (-)-102 effected efficient coupling (76%) of 15b to afford hemiketal (+)-111 as a single isomer83 (Scheme 27). Methyl ketal formation (p-TSA, MeOH) followed by reprotec- tion of the C(38) hydroxyl as the TIPS ether (TIPSCl, imid) completed construction of the side chain subtarget (-)-10.

Scheme 27

oxazoles, we turned to a 1949 publication of Sheehan,74 reporting the conversion of benzoyl isocyanate 103 to oxazolone 104 upon treatment with diazomethane (Scheme 26); enolization and capture as the triflate would furnish the desired linchpins. In the event, dropwise addition of an ethereal solution of diazomethane (alcohol-free)75 to acetyl isocyanate,76 readily generated in situ from acetyl chloride, afforded an unstable oxazolone (106),77 which without isolation was converted via conditions developed by Panek5p to triflate 15a in 48% overall yield. Importantly, assembly of 15a required only a matter of hours and a single purification. An analogous reaction sequence beginning with bromoacetylbromide furnished the corresponding Not pleased to have to reprotect the C(38) hydroxyl, we bromide 15b in identical yield.78 installed the TIPS ether at an earlier stage via an analogous Linchpin Model Studies. Initially, we explored the metala- route (Scheme 28). A modest improvement in both the stan- tion of oxazole 15a with t-BuLi. Unfortunately, only the nylation regioselectivity (6:1) and yield was observed in the undesired C(5) adduct 109 formed upon addition to δ-valero- 79 TIPS series, presumably because of the increased steric bulk of lactone (64%, Figure 1). The equilibrating conditions specif- the TIPS group.84 Both the Stille coupling and introduction of ically developed by the Evans group (e.g., Et2NLi) to convert the oxazole triflate also proceeded smoothly and in excellent the C(5) lithium anion of an oxazole to the C(2) methyl yield. substituent did not alter the reaction outcome. Presumably, the Side Chain Appendage and Macrolide Construction. The C(5) anion represents both the thermodynamic and the kinetic plan was now to effect coupling of side chain (-)-10 with vinyl anion because of the ability of the C(4) triflate (absent in the stannane (+)-90a, followed by macrolactone construction Evans substrates) to direct lithiation. We thus turned to (Scheme 29). Initially, we examined Pd2dba3‚CHCl3 as the Stille 2-bromomethyl oxazole 15b. After considerable experimenta- 5p + 80 catalyst; unfortunately, the desired product ( )-115 was tion, we discovered that premixing δ-valerolactone with 15b obtained in less than 20% yield. Exploration of related catalysts followed by addition of i-PrMgCl furnished the desired adduct 81 and solvent regimes eventually led to Pd(PPh3)4 in dioxane with 110 in 66% yield. Presumably, rapid Grignard exchange occurs excess LiCl (100 °C, sealed tube)73 as the optimal conditions to generate the metalated oxazole which attacks the lactone. to promote the Stille coupling; under these conditions, (+)-115 Premixing was necessary to minimize self-condensation of was obtained in 72% yield.85 To the best of our knowledge, 15b.82 (77) Although alternate oxazolone syntheses exist, most require aryl or alkenyl substitution at the C(2) position: (a) Rao, Y. S.; Filler, R. Chem. Commun. 1970, 1622. (b) Troxler, F. HelV. Chim. Acta 1973, 56, 1815. (c) Rodehorst, R. M.; Koch, T. H. J. Am. Chem. Soc. 1975, 97, 8. For a discussion of the limitations of such procedures, see ref 73b. (78) The utility of triflates 15a and 15b as linchpins is under further investigation in our laboratories: Smith, A. B., III; Minbiole, K. P.; Freeze, B. S. Synlett 2001, 1543. (79) Compound 109 is drawn as the open keto-alcohol because of observation of the 13C NMR resonance and infrared absorption (186.8 ppm and 1694 cm-1, respectively). Similarly, compound 110 is drawn as the hemiketal because of observation of a 13C hemiketal NMR resonance and IR hydroxyl absorption [94.7 ppm and 3420 cm-1 (br), respectively]. (80) Lithium halogen exchange (t-BuLi) in the presence of δ-valerolac- tone afforded 110 in modest yield (∼30%). Figure 1. Oxazole metalation studies. (81) “Grignard exchange” reactions have been demonstrated on vinylic and arylic substrates. See: (a) Lee, J.; Velarde-Ortiz, R.; Guijarro, A.; Wurst, J. R.; Rieke, R. D. J. Org. Chem. 2000, 65, 5428. (b) Delacroix, T.; Berillon, (74) Sheehan, J. C.; Izzo, P. T. J. Am. Chem. Soc. 1949, 71, 4059. L.; Cahiez, G.; Knochel, P. J. Org. Chem. 2000, 65, 8108. (75) (a) DeBoer, T. J.; Backer, H. J. Org. Synth. 1956, 36, 16. (b) Aldrich (82) Self-condensation arises from the electrophilic nature of unreacted Technical Bulletin AL-121. The residual ethanol in standard diazomethane 15b. Self-condensation was also observed upon inverse addition (i.e., slow reacts with the isocyanate. addition of bromomethyl oxazole 15b to a solution of t-BuLi at -100 °C). (76) For the preparation and isolation of acetyl isocyanate, see: Etienne, (83) Presumably, the sterochemical outcome is due to the anomeric effect. A.; Bonte, B.; Druet, B. Bull. Chim. Soc. Fr. 1972, 251. Also see: Scholl, See: Bonner, W. A. J. Am. Chem. Soc. 1959, 81, 1448. R. Chem. Ber. 1890, 23, 3505. (84) The corresponding internal stannane was recovered (∼5-15%). 10952 J. Am. Chem. Soc., Vol. 123, No. 44, 2001 Smith et al.

Scheme 28 Scheme 30

Scheme 29

equipotent to (+)-phorboxazole A (1);86 and second, the conversion of (+)-117 to (+)-118 served to validate the global deprotection conditions needed to arrive at the natural product. Introduction of the C(46) E-Vinyl Bromide: A Non- Trivial Task. The last major synthetic hurdle, namely conver- + (-)-10 represents the most complex oxazole triflate employed sion of alkyne ( )-117 to the C(46) E-vinyl bromide, proved in a Stille cross coupling. particularly challenging (Scheme 31). Initially, we explored a radical hydrostannylation.87 Accordingly, treatment of (+)-117 Macrolide construction followed directly from our earlier ° synthesis of (+)-86 (Scheme 21); selective desilylation,48 with Bu3SnH and AIBN at 80 C resulted in formation of the oxidation, and DMB removal afforded hydroxyaldehyde (+)- desired E-vinyl stannane with moderate selectivity (5:1) for the 116 (73%, 3 steps, Scheme 30). Macrocyclization then pro- terminal vinyl stannane. Unfortunately, almost complete isomer- + ization of the C(2-3) cis olefin occurred. Alternative radical ceeded in excellent yield to furnish ( )-117. Interestingly, the ° E/Z selectivity improved with higher temperatures. We attribute hydrostannylation conditions (e.g., Bu3SnH, Et3B, 0 C) resulted both in poor yield and selectivity. Palladium catalyzed hydro- the enhanced selectivity to an increase in the rate of oxaphos- 88 stannylation [Cl2Pd(PPh3)2,Bu3SnH], known to be unselective phatane collapse at the higher temperatures, which minimizes R oxaphosphatane equilibration and thereby formation of the trans with alkynes lacking branching, actually furnished a prepon- isomer. Exposure of (+)-117 to 6% HCl in THF resulted in (86) (a) Hansen, T. M.; Engler, M. M.; Ahmed, F.; Cink, R. D.; Lee, C. global deprotection to furnish 118, the C(45-46) alkyne S.; Forsyth, C. J. Abstract of Papers, 220th National Meeting of the American Chemical Society, Washington, DC; American Chemical Soci- congener of phorboxazole. The significance of this transforma- ety: Washington, DC, 2000; ORGN-040. (b) Uckun, F. M.; Forsyth, C. J. tion is twofold: first, alkyne (+)-118 had been reported to be Bioorg. Med. Chem. Lett. 2001, 11, 1181. (87) Leusink, A. J.; Budding, H. A.; Drenth, W. J. Organomet. Chem. (85) The reproducibility of the reaction proved highly dependent on the 1968, 11, 541 and references therein. amount of oxygen present in the system. When a “freeze pump thaw” tactic (88) Zhang, H. X.; Guibe´, F.; Balavoine, G. J. Org. Chem. 1990, 55, was employed to deoxygenate the dioxane prior to use, the reaction 1857. Also see: Boden, C. D. J.; Pattenden, G.; Ye, T. J. Chem. Soc., Perkin consistently proceeded in ∼68-72% yield. Trans. 1 1996, 2417. Total Synthesis of (+)-Phorboxazole A J. Am. Chem. Soc., Vol. 123, No. 44, 2001 10953

Scheme 31 Scheme 32

of the synthetic venture include the use of modified Petasis- derance of the internal [C(45)] stannane (2.5:1); presumably, Ferrier rearrangements for the effective linchpin assembly of chelation of the palladium species to the C(44) methyl ether both the C(11-15) and C(22-26) cis-tetrahydropyran rings; leads to the internal C(45) stannane. Faced with the failure of extension of the Julia olefination to the synthesis of enol ethers; other methods (i.e., Schwartz hydrozirconation), we were and the design, synthesis, and application of a novel bifunctional nonetheless encouraged that palladium-catalyzed hydrostannyl- oxazole linchpin. The longest linear sequence leading to (+)- ation returned the C(2-3) cis olefin geometry intact; thus, the phorboxazole A (1) was 27 steps, with an overall yield of 3%. regioselectivity remained the final issue. Completion of the (+)-Phorboxazole A (1) Synthetic Acknowledgment. Support was provided by the National Venture. A careful review of hydrostannylation literature led Institutes of Health (National Cancer Institute) through Grant us to the work of Guibe,88 who noted improved regioselectivity CA-19033. An American Chemical Society, Division of Organic for the hydrostannylation of alkynyl bromides (e.g., 120f121, Chemistry Fellowship (funded by DuPont Pharmaceuticals, Inc.) Scheme 32). To exploit this precedent, we prepared the alkynyl and a Feodor-Lynen Fellowship from the Alexander von bromide of (+)-117 (AgNO , NBS); palladium catalyzed 3 Humboldt Foundation are also gratefully acknowledged (K.P.M hydrostannylation afforded the desired E vinyl stannane with and M.S., respectively). Mr. John Lim is also acknowledged 4:1 C(46)/C(45) regioselectivity. Without separation, facile tin- for his experimental contributions. We also thank Drs. George bromine exchange (NBS, 95%) followed by treatment with 6% T. Furst and Rakesh Kohli of the University of Pennsylvania HCl (72 h) furnished a mixture of phorboxazole vinyl bromide Spectroscopic Service Center for assistance in securing and isomers [4:1, C(46)/C(45), 70%]. HPLC separation using a interpreting high-field NMR spectra and high resolution mass Zorbax C reversed-phase column (55:45 acetonitrile/H O) 18 2 spectra, respectively. provided pure, totally synthetic (+)-phorboxazole A (1), the spectral properties of which were identical in all respects [e.g., Supporting Information Available: Experimental proce- 1H NMR, ROESY, COSY (600 MHz), HRMS, and optical dures and analytical data for all compounds (74 pages, PDF). rotation] to the corresponding spectral data obtained from natural This material is available free of charge via the Internet at (+)-phorboxazole A (1). http://pubs.acs.org. Summary. A highly convergent, stereocontrolled total syn- thesis of (+)-phorboxazole A (1) has been achieved. Highlights JA011604L 624 J. Am. Chem. SOC. 1995,117, 624-633

Total Synthesis of Taxol. 1. Retrosynthesis, Degradation, and Reconstitution

K. C. Nicolaou,* P. G. Nantermet, H. Ueno, R. K. Guy, E. A. Couladouros, and E. J. Sorensen Contributionfrom the Department of Chemistry, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, Califomia 92037, and Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093 Received July 7, 1994@

Abstract: A successful strategy for the enantioselective synthesis of the natural stereoisomer of Taxol (1) has been developed. This strategy utilized the convergent assembly of Taxol's central eight-membered B ring from preformed synthons for rings A (10) and C (9) followed by late introduction of the D ring and side chain. Degradative studies confiied the viability of certain crucial manipulations including oxidation of the C13 position (35 - 3) and regioselective introduction of the C I-hydroxyl, CZbenzoyloxy moiety (29 - 31). Additionally, a convenient method for the large-scale production of 29, a derivative useful for C2 analog production, was developed.

Introduction 0 Taxol (Figure 1, 1),lq2a diterpene produced by several of the genus was isolated from the cytotoxic methanolic extract of the bark of T. brevifolia." Taxol interacts with microtubules, important cellular structural proteins,5in a manner that catalyzes their formation from tubulin and stabilizes the resulting structures.6 In cells this phenomenon leads to an altered morphology with the microtubules forming stable 1: Taxol bundles and the cell being unable to assemble a normal mitotic spindle.' Cells treated with Taxol normally arrest at the transition between interphase and mitosis and die. The elucida- tion of this unique mechanism of action during the late 1970s and early 1980s sped Taxol's development as an anticancer drug. Since that time, Taxol has revealed unusual efficacy as a clinical agent,* experiencing rapid development for the treatment of 2: 10-deacetylbaccatin111 brea~t,~ovarian,'O skin," lung,12 and head and neck13 cancers. Figure 1. Structure of Taxol (1) and 10-deacetylbaccatin III (2). * Address correspondence to this author at The Scripps Research Institute or the University of California. In 1993, Taxol was approved by the FDA for use in the U.S. @ Abstract published in Advance ACS Abstracts, December 15, 1994. (l)Nicola~u,K. C.; Dai, W.-M.; Guy, R. K. Angew. Chem., In?. Ed. for treatment of breast and ovarian cancers. Engl. 1994, 33, 15. Taxol's development as a therapeutic agent precipitated a (21 Kingston. D. G. I. Fortschr. Chem. Ora. Narurst. 1993, 61, 1. fundamental problem with its production: the original source (3) ApGndino, G. Fitoterapia 1993, 54, Sippl. NI, 5. (4) Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggen, P.; McPhail, A. of its isolation, T. brevifolia, was a slowly growing and rare T. J. Am. Chem. SOC.1971, 93, 2325. tree whose content of Taxol could not possibly meet the (5) (a) Mandelkow, E.; Mundelkow, E.-M. Curr. Opin. Struct. Biol. 1994, demand.14 The public's perception of the ecological disaster 4, 171. (b) Avila, J. Sci. 1991, 50, 327. involved in harvesting these trees from the last remaining old (6) Manfredi, J. J.; Horwitz, S. B. Pharmacal. Ther. 1984,25,83.Schiff, P. B.; Fant, J.; Horwitz, S. B. NarUre 1979, 277, 665. growth forests of the Pacific Northwest caused an ongoing (7)Schiff, P. B.; Horwitz, S. B. Proc. Natl. Acad. Sci. USA. 1980, 77, debate about the ethics of producing Tax01.l~ A wide range of 1561. research was carried out to solve this problem, including (8) Lavelle, F. Curr. Opin. Invest. Drugs 1993, 2, 627. Rowinsky, E. K.; Onetto, N.; Canetta, R. M.; Arbuck, S. G. Semin. Oncol. 1992, 19, plantation farming, cellular culture, , and total 646. synthesis. l4 A semisynthetic process utilizing 10-deacetylbac- (9) Holmes, F. A,; Waters, R. J.; Theriault, R. I.; Forman, A. D.; Newton, catin 111 (2, Figure l), derived from the common T. baccata L. K.; Raber, M. N.; Buzdar, A. U.; Frye, D. K.; Hortobagyi, G. N. J. Natl. Cancer Inst. USA. 1991, 83, 1797. shrub, as the starting material has, at least temporarily, resolved (10) McGuire, W. P.; Rowinsky, E. K.; Rosenshein, N. B.; Grumbine, this dilemma.14 Over the past two decades some 30 synthetic F. C.; Ettinger, D. S.; Armstrong, D. K.; Donehower, R. C. Ann. Intern. Med. 1989, 111, 273. Einzig, A. I.; Wiemik, P. H.; Sasloff, J.; Garl, S.; (12) Chang, A.; Kim, K.; Glick, J.; Anderson, T.; Karp, D.; Johnson, D. Runowicz, C.; O'Hanlan, K. A.; Goldberg, G. Proc. Am. Assoc. Cancer J. Natl. Cancer Inst. USA. 1993,85, 388. Murphey, W. K.; Winn, R. J.; Res. 1990, 31, 187 (Abstract 1114). Pazdur, R.; Ho, D. H.; Lassere, Y.; Fossella, F. V.; Shin, D. M.; Hynes, H. E.; Gross, H. M.; Davila, E.; Leimert, Bready, B.; Kvakoff, I. H.; Raber, M. N. Proc. Am. SOC. Clin. Oncol. 1992, J. T.; Dhinga, H. M.; Raber, M. N.; Krakoff, I. H.; Hong, W. K. Proc. Am. 11, 111 (Abstract 265). Caldas, C.; McGuire, W. P., III. Semin. Oncol. SOC.Clin. Oncol. 1993,85,384. Ettinger, D. S. Semin. Oncol. 1993,ZO(4 1993, 20 (4 Suppl. 3), 50. Suppl. 3), 46. (1 1) Einzig, A. I.; Hochster, H.; Wiemik, P. H.; Trump, D. L.; Dutcher, (13) Forastiere, A. A. Semin. Oncol. 1993, 20 (4 Suppl. 3), 56. J. P.; Garowski, E.; Sasloff, J.; Smith. T. J. Invest. New Drum 1991.9.59. (14) Borman, S. Chem. Eng. News 1991, Sept 2, 11. Legha, S. S.; Ring, S.; Papadopoulos, N.; Raber, M. N.{Benjamin, R. (15)Hartzell, H. The Yew Tree, A Thousand Whispers; Hulogosi: Cancer 1990, 65, 2478. Eugene, OR, 1991. 0002-7863/95/1517-0624$09.00/0 0 1995 American Chemical Society Total Synthesis of Taxol. 1 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 625 groups, attracted by the molecule's challenging architecture and Scheme 1. Retrosynthetic Analysis of Taxol (1)" importance in medicine, undertook the task of the total synthesis 0 of Taxo1.l~~Herein and in the following articles16-18we report the total synthesis of Taxol (l).l9

Retrosynthetic Analysis and Strategy i)H The retrosynthetic analysis and final synthetic strategy dis- cussed below emerged after considering several options and examining information gathered during preliminary studies in this program. Aspects of alternative plans originally considered will be discussed in the context of the overall story as revealed in this and the following papers in this series. In considering a strategy for the total synthesis of Taxol (l), we set the following postulate as a condition: the route should be short and flexible to allow for the eventuality of producing the natural product and a variety of its analogs in a practical X way and to deliver the target molecule in its enantiomerically 3 pure and correct form. To best fulfill this criteria, a convergent sequence was chosen in which rings A and C were to be i constructed separately and then brought together to form the 8-membered ring B. Examples already in the literature and knowledge derived from our own experience led us to conclude that we could leave for the final stages the attachment of the side chain,20,21the oxygenation of the C13 position,22 and the formation of the oxetane ring.23-25 Scheme 1 shows the retrosynthetic analysis of Taxol (1) on which the synthetic strategy was based. Thus, appropriate 6 protection, removal of the side chain, and deoxygenation transforms at C13 led, retrosynthetically, to the baccatin derivative 3. Functional group manipulation at C1 and C2 led i to the 5-membered ring derivative 4 which was envisioned as a precursor to the 1-hydroxy-2-benzoate system of Taxol. Retrosynthetic disassembly of the oxetane ring in 4 and introduction of a double bond in ring C allowed the generation of intermediate 5 as a possible precursor. The carbocyclic ABC taxoid core 5 was then retrosynthetically broken by standard \ 0R3 '0 R3 functional group manipulations and disconnection of the C9- 7 C10 bond leading to dialdehyde 6. The latter was considered (16) Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorensen, E. J.; i Claiborne, C. F.; Guy, R. K.; Hwang, C.-K.; Nakada, M.; Nantermet, P. G. J. Am. Chem. SOC. 1995, 117, 634. (17) Nicolaou, K. C.; Yang, Z.; Liu, J.-J.; Nantermet, P. G.; Claiborne, C. F.; Renaud, J.; Guy, R. K.; Shibayama, K. J. Am. Chem. SOC. 1995, I17, xxx. (18) Nicolaou, K. C.; Ueno, H.; Liu, J.-J.; Nantermet, P. G.; Yang, Z.; NNHS0,Ar Renaud, J.; Paulvannan, K.; Chadha, R. J. Am. Chem. SOC.1995,117, xxx. (19) Nicolaou, K. C.; Yang, Z.; Liu, J.-J.; Ueno, H.; Nantermet, P. G.; Guy, R. K.; Claiborne, C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan, 10 K.; Sorensen, E. J. Nature 1994, 367, 630. Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang, F. F.; Biediger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.;Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.;Zhang, a 3 P.; Murthi, K. K.; Gentile, L. N.; Liu, J. H. J.Am. Chem. SOC. 1994, 116, 1597. Holton, R. A.; Kim, H. B.; Somoza, C.; Liang, F.; Biediger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L. N.; +OAC 11 Liu, J. H. J. Am. Chem. SOC. 1994, 116, 1599. + (20) Holton, R. A. Workshop on Taxol and Taxus, 1991. Holton, R. A. HO OH Eur. Pat. Appl. EP400,971 1990; Chem. Abstr. 1990, 114, 16456817. CN 12 (21) Ojima, I.; Habus, I.; Zhao, M.; Georg, G. I.; Jayasinghe, L. R. J. =? 13 14 Org. Chem. 1991, 56, 1681. Ojima, I.; Habus, I.; Zhao, M.; Zucco, M.; Park, Y. H.; Sun, C. M.; Brigaud, T. Tetrahedron 1992,48, 6985. Ojima, a Bz = COPh; R, R1,Rz, RJ, %, Rs = protecting groups. I.; Sun, C. M.; Zucco, M.; Park, Y. M.; Duclos, 0.;Kuduk, S. Tetrahedron Lett. 1993, 34, 4149. as a good candidate to afford, in the synthetic direction, (22) Ulman Page, P. C.; McCarthy, T. J. In Comprehensive 5 via a McMurry pinacol coupling.26 Continuing Synthesis; Trost, B. M., Fleming, I., Ley, S. V., FRS, Eds.; Pergamon with the simplification of structure, intermediate 6 was traced Press: New York, 1991; Vol. 7, p 99. (23) Ettouati, L.; Ahond, A.; Poupat, C.; Potier, P. Tetrahedron 1991, back to diol 7 and then to allylic alcohol 8 as potential 47, 9823. progenitors. Finally, disconnection of 8 via a Shapiro2' (24) Magee, T. V.; Bornmann, W. G.; Isaccs, R. C. A.; Danishefsky, S. transform led to hydrazone 10 representing ring A and aldehyde J. J. Org. Chem. 1992, 57, 3274. (25) Nicolaou, K. C.; Liu, J.-J.; Hwang, C.-K.; Dai, W.-M.; Guy, R. K. (26) McMurry, J. E. Chem. Rev. 1989,89, 1513. McMurry, J. E. Acc. J. Chem. SOC., Chem. Commun. 1992, 1118. Chem. Res. 1983, 16, 405. Lenoir, D. Synthesis 1989, 883. 626 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al. Scheme 2. Preparation of 7-TES-baccatin III (17)a Scheme 3. Benzylation of the C7 Position and Oxetane 0 Ring Opening"

1 : taxol IS: baccatin 111 15: baccatin 111 IIb I H+ OBn OBn

1q!H- A0

2: 10-deacetyl baccatin Ill 16:R=H H dL17:R=Ac 1s 20 21 Reagents and conditions: (a) excess n-B@B&, CHzClz, 25 "C, Reagents and conditions: (a) 20 equiv of benzyl tsichloroacetimi- 7 h, then AcOH, 77%; (b) 30 equiv of EtsSiCl, pyridine, 25 "C, 24 h, date, 1.0 equiv of triflic acid, CHzC12.25 "C, 40 h, 50%. Bz = COPh, 85%; (c) 20 equiv of Et3SiC1, pyridine, 25 "C, 17 h, 91%; (d) 5 equiv Bn = CHZPh. of AcCl, pyridine, 0 "C, 48 h, 82%. TES = SiEt3, Bz = COPh. roacetimidateitriflic acid),32also proved too destructive: giving 9 representing ring C. The cyclohexene derivatives 10 and 9 opening of the oxetane ring30 and leading to compound 18 were then disassembled by Diels- Alder transforms to afford (Scheme 3) as a major product (C7 stereochemistry not defined, olefins 11-14 as potential starting materials. acid-catalyzed epimerization at this position has also been The synthetic strategy derived from the analysis discussed rep~rted).~~This compound was presumably formed via above included a number of sensitive and rather daring steps intermediates 19-21 as shown in Scheme 3 by a mechanism in its final stages. In order to explore these final steps and similar to that proposed by Kingston in his oxetane opening establish their viability, we embarked, in parallel with the reaction induced by Meerwein's reagent.31 forward execution of the scheme, on degradation studies starting Failing to introduce a benzyl group at the C7 hydroxyl, we with Taxol (1) and 10-deacetylbaccatin 111 (2).** Included then turned to a silyl group. In agreement with Greene,34we amongst our goals in this program were the following: deoxy- observed that a tert-butyldimethylsilyl (TBS) group could not genation of the C13 position and exploration of its allylic be efficiently introduced. Installation of a triethylsilyl (TES) oxidation, establishment of a suitable cyclic protecting group group at C7, however, was smoothly accomplished with TESCl for the C1 and C2 hydroxyl groups and its regioselective in pyridine34(85% yield) to afford 7-TES-baccatin IIX (17). The conversion to the requisite C1 hydroxy, C2 benzoate functional- same compound was obtained from 10-deacetylbaccatin I11 (2) ity, and cleavage of the C9-C10 bond in order to obtain following Greene' s procedure34 involving selective silylation intermediates suitable for exploring the McMurry pinacol at the C7 hydroxyl group followed by acetylation of the C10 coupling as a means to construct the 8-membered ring of Taxol. hydroxyl group. The latter step (AcCl, pyridine, 0 "C) proved rather capricious on a larger scale, presumably due to the oxetane Preparation of 7-TES-baccatin I11 opening and ring A skeletal rearrangements-although the Since 7-benzyl and 7-triethylsilyl (TES) baccatin 111 were byproducts were not isolated.31~33~35-37As we will see later in projected as advanced intermediates in our synthesis, one of this discussion, however, a more reliable method for this our early objectives was to prepare these compounds from the transformation was discovered and utilized. naturally occumng Taxol (1) and 10-deacetylbaccatin III (2). While the former natural product is found in the bark of the Formation of the 1,2-Carbonate Ring and Reconversion Pacific Yew tree (T. Brevifolia) in rather limited amounts, the to the 1-Hydroxy, 2-Benzoate System latter compound is readily available from the needles of the With 7-TES-baccatin III (17) in hand, we then turned our European Yew tree (T.baccata). Scheme 2 summarizes the attention to the hydrolysis of the C2-benzoate and the C10- chemistry that led to the preparation of 17 from 1 and 2. Thus acetate in order to gain access to further degradation products removal of the side chain from Taxol (1) via reduction of the (Scheme 4). Early trials using hydrolysis, methanolysis, or C13 ester linkage proceeded according to Kingston's method (~-BU~NBH~)*~to afford baccatin I11 (15) in 77% yield. Our (31) Samaranayake, G.; Magri, N. F.; Jitrangsri, C.; Kingston, D. G. I. J. Org. Chem. 1991, 56, 5114. synthetic strategy was best served by a 7-benzyl derivative and (32) Iversen, T.; Bundle, K. R.; J. Chem. Soc., Chem. Commun. 1981, we, therefore, first considered the preparation of such an 1240. White, J. D.; Reddy, G. N.; Spessard, G. 0. J. Am. Chem. SOC.1988, intermediate. Basic conditions were, however, unacceptable 110, 1624. Widmer, U. Synthesis 1987, 568. (33) Wahl, A.; GuBritte-Voegelein, F.; GuBnard, D.; Le Goff, M.-T.; because of the well-documented epimerization at C7 via a Potier, P. Tetrahedron 1992,48, 6965. retroaldoValdo1 sequen~e.~~~~~~~Acidic conditions (benzyl trichlo- (34) Denis, J. N.; Greene, A. E.; Gutnard, D.; Gutritte-Voegelein, F.; Mangatal, L.; Potier, P. J. Am. Chem. SOC. 1988, 110, 5917. (27) Chamberlin, A. R.; Bloom, S. H. Org. React. 1990, 39, 1. (35) Appendino, G.; Ozen, H. C.; Gariboldi, P.;Torregiani, E.; Gabetta, (28) Nicolaou, K. C.; Nantermet, P. G.; Ueno, H.; Guy, R. K. J. Chem. B.; Nizzola, R.; Bombardelli, E. J. Chem. SOC.Perkin Trans. 11993, 1563. SOC.,Chem. Commun. 1994, 295. (36) Kingston, D. G. I.; Samaranayake, G.; Ivey, C. A. J. Nat. Prod. (29) Magri, N. F.; Kingston, D. G. I.; Jitrangsri, C.; Piccariello, T. J. 1990,53,1. Org. Chem. 1986, 51, 3239. (37) Gutritte-Voegelein, F.; Gutnard, D.; Potier, P. J. Nar. Prod. 1987, (30) Kingston, D. G. I. Phamcol. Ther. 1991, 52, 1. 50, 9. Total Synthesis of Taxol. I J. Am. Chem. SOC., Vol. 117, No. 2, 1995 627 Scheme 4. Early Attempts at C2 and C10 Hydrolysis" Scheme 5. Preparation of Carbonate 30 from 7-TES-baccatin III (17) and Its Transformation to Enone 25"

-a

17: 7-TES baccatin 111 22 17: 7-TES baccatin Ill 25 lb

23 24 Reagents and conditions: (a) excess LiAlK, THF, -78 "C or -30 "C, 1-5 h; (b) excess K2CO3, MeOH, H20, 0 OC or 25 OC, 1-5 h. TES = SEt3, BZ = COPh. metal hydride reductions gave poor yields of tetrol 22, in agreement with previous observation^.^^^^^^^^ The principal byproducts seemed to result from deacetylation at C4 and various intramolecular reactions such as the opening of the oxetane ring by the newly liberated C2 hydroxyl group to form compound 23 (Scheme 4). The intramolecular engagement of the C4 acetate and C13 hydroxyl in a hydrogen-bonding 28 29:R=H arrangement (structure 24, Scheme 4) is presumably responsible 30 : R = AC for the ease of deacetylation of the C4 oxygen. Similar structures have previously been invoked to explain deacetylation of C4 in analogous sit~ations.~,~~,~~It was, therefore, decided to remove any possible interference from the C13 hydroxyl group by either oxidizing it to the enone or removing it 9 altogether. Such manipulations would also further our explora- tion of degradative and synthetic chemistry. Oxidation of Clp was projected not only as a means to remove the troublesome hydroxyl group but also as a way to 31 25 change the conformation of the molecule to the extent that might a Reagents and conditions: (a) 1.5 equiv of 4-methylmorpholine affect the rate of hydrolysis of the C4 acetate and prevent attack N-oxide (NMO), 0.05 equiv of tetrapropylammonium permthenate of the C2 alkoxide on the oxetane ring. This 0peration~9~~(17 (TPAP), CH3CN, 25 "C, 1.5 h, 98%; (b) excess KzC03, MeOH, HzO, 0 "C, 4 h, 91%; (c) 0.05 equiv of camphorsulfonic acid (CSA), 1.0 -+ 25, Scheme 5) was smoothly carried out in 98% yield using Ley's TPAPNMO system.42 As hoped, enone 25 was readily equiv of benzaldehyde dimethyl acetal or excess 2,2-dimethoxypropane, CHzC12, 25 "C, 20 h; (d) 10 equiv of phosgene, pyridine, 0 "C, 0.5 h, hydrolyzed in basic conditions (K2CO3, MeOH, HzO, 0 "C) to 85% or 6 equiv of carbonyldiimidazole, THF,40 "C, 0.5 h, then 1 N provide triol 26 in 91% yield. Contrary to the previously aqueous HC1, THF, 25 "C, 15 min, 93%; (e) 4.5 equiv of AczO, 9 accepted order of ester reactivity in taxoids (C9, C10 > C2),2 equiv of 4-(dimethylamino)pyridine (DMAP), CHzClZ, 25 "C, 0.5 h, it was observed that, if so desired, the C10-acetate of compound 95%; (f) 10 equiv of PhLi, THF, -78 "C, 0.5 h, 85%; (g) 10 equiv of 26 could be partially retained, as it reacts more slowly than the AczO, 5 equiv of DMAP, CH&, 25 "C, 2.5 h, 95%; (h) 5 equiv of C2-benzoate under the above conditions. PhLi, THF,-78 "C, 15 min, 70% plus 10% of 31. TES = SiEt3, Bz = COPh. Initial attempts to introduce a ben~ylidene,"~a potential precursor to the C1-hydroxy, C2-benzoate system,44 or an as the C2-hydroxyl group opened the oxetane ring under the a~etonide~~protecting group at the Cl-C2 site met with failure, acidic conditions used. In both instances the resulting product (38) Klein, L. L. Tetrahedron Lett. 1993, 34, 2047. was the tetrahydrofuran derivative 28 (Scheme 5).33.38,46At- (39) Chen, S.-H.; Wei, J.-M.; Farina, V. Tetrahedron Len. 1993,34,3205. tention then focused on constructing a carbonate ring at the C1- C2 site, an operation that required basic rather than acidic (40)Harrison, J. W.; Scrowsten, R. M.; Lythgoe, B. J. Chem. Soc. C 1966, 1932. conditions. Despite the scarcity of reports of nucleophilic (41) Senilh, V.; Gutritte, F.; GuCnard, D.; Colin, M.; Potier, P. C. R. additions to carbonates to form we entertained the Acad. Sei. Pans 1984, 299, 1039. possibility of converting such functionality directly to the desired (42) Griffith, W. P.; Ley, S. V. Aldrichimica Acta 1990, 23, 13. (43) Albert, R.; Dax, K.;Pleschko, R.; Stutz, K. Carbohydr. Res. 1985, 1,2-hydroxybenzoate of Taxol(1) in the synthetic direction, by 137, 282. Yamanoi, T.;Akiyama, E.; Inazu, T. Chem. Lett. 1989, 335. addition of nucleophilic phenyl species (Scheme 5). Even Crimmins, M. T.; Hollis, W. G., Jr.; Lever, G. J. Tetrahedron Lett. 1987, though the regiospecificity of such an opening was questionable, 28, 3647. (44) Binkley, R. W.; Goewey, G. S.; Johnston, J. C. J. Org. Chem. 1984, (46) Farina, V.; Huang, S. Tetrahedron Lett. 1992, 33, 3979. 49, 992. (47) Satyanarayana, G.;Sivaram, S. Synth. Commun. 1990, 20, 3273. (45) Evans, M. E.; Parrish, F. W.; Long, L., Jr. Carbohydr. Res. 1967, (48) Wender, P. A.; Kogen, H.; Lee, H. Y.;Munger, J. D.; Wilhelm, R. 3, 453. Lipshutz, B. H.; Barton, J. C. J. Org. Chem. 1988, 53, 4495. S.; Williams, P. D. J. Am. Chem. Soc. 1989, Ill, 8957. 628 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al. we expected the distinctly different steric environment of the Scheme 6. Selective Oxidation of the C13 Hydroxyl Group two positions to favor the less crowded C2 regioisomer. and Preparation of Enone 26" Treatment of triol 26 with phosgene in pyridine provided the desired carbonate 29 in 85% yield.49 It was later discovered that the carbonate 29 could be obtained in 93% yield by using carb~nyldiimidazole~~and 4-(dimethy1amino)pyridine (DMAP) in THF followed by acidic hydrolysis of the imidazole carbamate at C10. With a practical preparation of carbonate 29 secured, we then 0 0 29 32: R = H. Me proceeded to investigate the anhydrous nucleophilic opening of the carbonate ring with organometallic species-a rather daring proposition considering the presence of four additional carbonyl groups within the molecule. To our pleasant surprise, exposure of 29 to excess phenyllithium in THF at -78 "C for 0.5 h resulted in the regioselective formation of the CZbenzoate 31 in 85% yield. This product was readily acetylated (95% yield) at the C10 hydroxyl position to afford compound 25. The carbonate ring opening was also performed on the C10-acetate derivative 30, resulting in the formation of a mixture of 25 (70%) 1s 33 and the corresponding 10-deacetylderivative 31 (10%). Acety- lation of the crude reaction mixture under standard conditions followed by chromatographic purification afforded 25 in 80% IC overall yield from 30. The resistance of the other carbonyl moieties in these substrates to phenyllithium attack is, presum- ably, due to their steric shielding by the surrounding groups. In addition to providing a clear path for some of the final steps in the projected synthesis of Taxol (l), this chemistry was exploited to deliver a variety of C2 analogs of the natural produ~t.~~,~~ 31 2s "Reagents and conditions: (a) 15 equiv of Pb(OAc)4, MeOH, Attempts To Cleave the C9-C10 Bond of the Taxol benzene, 0 - 50 OC; or excess of NaI04, MeOH, H20, 25 'C; or 2 Skeleton. Preparation of Enone 26 equiv of HzOz, 8 equiv of NaOH, MeOH, HzO, 0 "C, 1.25 h; (b) 15 equiv of Pb(OAc)d, MeOH, benzene, 50 "C, 24 h; (c) 1.0 equiv of With the Cl-C2 diol system protected and the C9-C10 site 4-methylmorpholine N-oxide (NMO), 0.05 equiv of tetrapropylammo- free as a hydroxy ketone, as in compound 29 (Scheme 6), we nium permthenate (TPAP),CHzClz, 25 "C, 2 h, 96%; (d) 10 equiv of attempted the cleavage of the C9-C10 bond under a variety of KzCO3, MeOH, H20,O OC, 2.5 h, 93% based on 81% conversion. TES = SiEt3, BZ = COPh. oxidative conditions. Unfortunately, however, none of these methods (including Pb(OAc)4, Na104,53and Baeyer-Villiger/ hydr~lysis~~)led to the expected aldehyde 32 (Scheme 6) or C13 Deoxygenation, Reoxygenation, and Side-Chain Attachment any other cleavage product. Steric crowding is presumably again responsible for this inertness. This phenomenon also In order to delve further into our planned synthetic strategy, manifested itself in the reluctance of 7-TES-10-deacetylbaccatin we focused our efforts on the deoxygenation of the C13 position III (16) to enter in any cleavage process to afford 33 (Scheme and on its subsequent reoxygenation. The fust objective proved 6) under similar conditions. In the reaction of 16 with rather problematic as initial attempts of Wolf-Kishner reduc- Pb(OAc)4, it was surprising to observe a 20% yield of the C13- tionJ5 of enone 25 (Scheme 5) and thioacetal formation/ oxidized product, namely enone 31 (Scheme 6), in addition to reduction56 of the same compound failed. A Barton deoxygen- recovered starting material (60%). This selective oxidation (16 ation5' was then considered. Although a C13 xanthate could - 31) could be carried out more efficiently with TPAP-NM@* not be produced, strenuous conditions (excess (thiocarbony1)- in methylene chloride (96% yield). Subsequent hydrolysis (K2- diimidazole and DMAP, 75 "C, 18 h) allowed the conversion C03, MeOH, HzO, 0 "C) of the C2-benzoate from 31 provided of 7-TES-baccatin 111 (17) to thiocarbamate 34 (Scheme 7) in triol 26 in 93% yield. This sequence allows the conversion of 86% yield. Treatment of 34 with excess n-Bu3SnH in toluene naturally occurring 10-deacetylbaccatinIII (2) to compound 26 at 85 "C in the presence of a catalytic amount of AIBN provided in three steps and in 81% overall yield, avoiding the problematic the desired C13 deoxy derivative 35 in 59% yield, together with acetylation at C10. its A12J3regioisomer 36 (17% yield). Increasing the concentra- tion of n-Bu3SnH in an attempt to trap the initially formed C13 (49) Haworth, W. N.; Porter, C. R. J. Chem. Soc. 1930, 151. radical before it rearranges to its C11 isomer, responsible for (50) Kutney, J. P.; Ratcliffe, A. H. Synth. Commun. 1975, 5, 47. the formation of the byproduct 36, did not change the ratio of (51) Nicolaou, K. C.; Couladouros, E. A.; Nantermet, P. G.; Renaud, J.; the two products. Guy, R. K.; Wrasidlo, Angew. Chem., Int. Ed. Engl. 1994, 33, 1581. (52) Nicolaou, K. C.; Renaud, J.; Nantermet, P. G.; Guy, R. K.; The desired oxidation of the C13 allylic position22 to a Couladouros, E. A.; Wrasidlo, W. Submitted. carbonyl function was demonstrated on intermediate 35 by (53) Shing, T. K. M. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Ley, S. V., FRS, Eds.; Pergamon Press: New York, 1991; (55) Todd. Org. React. 1962, 12, 356. Vol. 7, p 708. Crimmins, M. T.; Jung, D. K.; Gray, J. L. J. Am. Chem. (56) Van Tamelen. Org. React. 1948,4, 378. Dailey, 0. D., Jr. J. Org. Soc. 1993, 115, 3146. Chem. 1987,52, 1984. Corey, E. J.; Shimoji. Tetrahedron Lett. 1983, 24, (54) Krow, G. R. In Comprehensive Organic Synthesis; Trost, B. M., 169. Fleming, I., Ley, S. V., FRS, Eds.; Pergamon Press: New York, 1991; (57) Barton, D. H. R.; McCombie, S. W. J. Chem. SOC., Perkin Trans. Vol. 7, p 671. Ogata, Y.; Sawaki, Y.; Shiroyama, M. J. Org. Chem. 1977, 11975, 1574. Barton, D. H. R.; Dorchak, J.; Jaszberenyi Tetrahedron 1992, 42, 4061. 48, 7435. Hartwig Tetrahedron 1983, 39, 2609. Total Synthesis of Taxol. 1 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 629

Scheme 7. C13 Deoxygenation and Oxygenation" Scheme 8. Taxol's Side-Chain Attachment"

17 7-TES baccatin 111 34 Ib 40: R = TES 41:R=EE

0

36 35 IC

42: R = TES 43: R = EE

0 t 0

37

1 : Taxd Reagents and conditions: (a) excess NaBK, MeOH, 25 OC, 3 h, 94% based on 88% conversion; (b) for 42, 3 equiv of NaN- (SiMe3)2, 3.5 equiv of B-lactam 40, THF,0 "C, 0.5 h, 86% based on 89% conversion; for 43, 2.5 equiv of NaN(SiMe&, 1.2 equiv of B-lactam 41, THF, 0 "C, 20 min, 80% based on 54% conversion; (c) for 42, HF-pyridine, THF, "C, h, for 43, EtOH, 35 25 25 1.25 80%; 0.5% aqueous HC1, 0 "C, 72 h, 80%. TES = SiEt3, Bz = COPh, EE = Reagents and conditions: (a) 20 equiv of (thiocarbonyl)diimidazole, ethoxyethyl. 30 equiv of 4-(dimethylamino)pyridine (DMAP), THF,75 "C, sealed tube, 18 h, 86%; (b) 10 equiv of n-BusSnH, 0.1 equiv of azobis(isobu- tyronitile) (AIBN), toluene, 85 "C, 2 h, 59% of 35 plus 17% of 36; (c) established toward Taxol (1) as follows: (a) silylation with 20 equiv of KzC03, MeOH, THF, H20,O "C, 6 h, then -20 "C, 10 h, TESCl under standard condition^^^,^ to afford 38 (85% yield); 94% based on 62% conversion; (d) 10 equiv of phosgene, pyridine, 25 (b) carbonate ring opening with phenyllithium, as described "C, 15 min, 86%; (e) m-pyridine, THF, 25 "C, 2 h, 88%; (f) 50 equiv above, to convert 38 to 35 (80%yield, Scheme 7); (c) allylic of Et3SiC1, pyridine, 25 "C, 24 h, 85%; (g) 5 equiv of PhLi, THF,-78 oxidation (75%); (d) stereoselective reduction of the enone OC, 15 min, 80%; (h) 30 equiv of pyridinium chlorochromate (PCC), carbonyl of 25 with NaBb according to Potier's meth~d~~v~l 30 equiv of NaOAc, Celite, benzene reflux, 1 h, 75%. TES = SiEt3, Bz = COPh. to provide 7-TES-baccatin 111 (17, Scheme 8) in 94% yield, based on 88% conversion; and finally (e) attachment of the side exposure to pyridinium chlorochromate (PCC)58in the presence chain onto 17 using the Ojima-Holton p-lactam method20q21 of NaOAc and Celite in refluxing benzene to afford, in 75% (Scheme 8). To the latter end, both optically active p-lactams yield, enone 25 (Scheme 7). 40 and 41 were prepared according to Ojima's procedure and The C13 deoxy intermediate 35 was converted to the coupled to 17 using NaN(SiMe3)z to provide the 2',7-diprotected corresponding diol 37 (Scheme 7) via selective benzoate Taxol derivatives 42 and 43, respectively. Deprotection of 42 hydrolysis (K2CO3, MeOH, H20, THF, 0 "C, 94% yield based with HF~yridine~~in THF furnished Taxol (1) in 80% yield, on 62% conversion). The carbonate ring was installed at the whereas exposure of 43 to dilute HCl in EtOH61 led to the same Cl-C2 positions of the latter compound by the phosgene- target (1) in a similar fashion (80%). pyridine method,49 furnishing intermediate 38 in 86% yield. Desilylation of the C7 hydroxyl group by exposure to Conclusion HF~yridine~~led to 39, in 88% yield, a compound that was The chemistry described in this article shed light on the projected as an advanced intermediate in our synthetic scheme. chemical properties of Taxol (1) and its derivatives and opened Using the key intermediate 39 (Scheme 7), obtained from 10-deacetylbaccatin III(2) as described above, a sequence was (60) Hart, T. W.; Metcalfe, D. A.; Scheinmann, F. J. Chem. SOC., Chem. Commun. 1979, 156. Roush, W. R.; Russo-Rodrigez, S. J. Org. Chem. (58) Parish, E. J.; Wei, T. Y. Synrh. Commun. 1987.17, 1227. Rathore, 1987, 52, 598. R.; Saxena, N.; Chadrasekaran Synfh. Commun. 1986, 16, 1493. (61) Ogilvie, K. K.; Thompson, E. A.; Quiliam, M. A,;Westmore, J. B. (59) Nicolaou, K. C.; Webber, S. E. Synthesis 1986, 453. Tetrahedron Left. 1974, 2865. 630 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al.

access to a number of valuable taxoid intermediates. Specifi- 12%) and 7-TES-baccatin IJJ (17, 8.3 mg, 94% based on 88% cally, it allowed the definition of a series of key intermediates conversion) as a white powder: Rj = 0.43 (silica, 50% EtOAc in and of a track along which our total synthesis was to follow hexanes); [aIZZ~-49 (c 0.4, MeOH); IR (thin film) vm 3518, 2914, (39 - 38 - 35 - 25 - 17 - 1). Furthermore, the easy access 1723, 1448, 1237 cm-'; 'H NMR (500 MHz, CDCl3) 6 8.08 (d, J = to the 5-membered ring carbonate intermediate 29 developed 7.5 Hz, 2 H, Bz), 7.58 (t, J = 7.4 Hz, 1 H, Bz), 7.46 (t, J = 7.4 Hz, 2 H, Bz), 6.44 (s, 1 H, 10-H), 5.61 (d, = 7.0 1 H, 2-H), 4.94 (d, in this program was crucial to providing a practical entry into J Hz, J = 9.5 Hz, 1 H, 5-H), 4.82 (m, 1 H, 13-H), 4.47 (dd, J = 10.5, 6.8 a plethora of C2 analogs of Taxol (1)via nucleophilic opening Hz,lH,7-H),4.28(AofAB,d,J=8.3H~,lH,20-H),4.12(Bof of the carbonate ring with a variety of reagents. The following AB, d, J = 8.3 Hz, 1 H, 20-H), 3.86 (d, J = 7.0 Hz, 1 H, 3-H), 2.51 papers in this series describe the total synthesid6-18 of Taxol (m, 1 H, 6-H), 2.27 (s, 3 H, OAc), 2.25 (m, 1 H, 14-H), 2.17 (s, 3 H, (1) and a variety of its anal~gs.~l*~~ OAc), 2.16 (s, 3 H, 18-CH3), 2.05 (m, 1 H, 14-H), 1.85 (m, 1 H, 6-H), 1.55 (s, 3 H, Ig-CHs), 1.17 (s, 3 H, 16-CH3), 1.02 (s, 3 H, 17-CH3), Experimental Section 0.90 (t, J = 8.0 Hz,9 H, Si(CHzCH3)3). 0.62-0.50 (band, 6 H, Si(CH2- CH3)3); I3C NMR (125 MHz, CDCl3) 6 202.2, 171.0, 169.4, 167.1, General Techniques. All reactions were canied out under an argon 143.9, 133.6, 132.6, 130.1, 129.4, 128.6, 84.2, 80.8, 78.7, 76.5, 75.8, atmosphere with freshly distilled solvents under anhydrous dry, 74.7, 72.3, 67.9, 58.6, 47.2, 42.7, 38.2, 37.2, 26.8, 22.7, 21.0, 20.1, conditions, unless otherwise noted. Tetrahydrofuran (THF) and ethyl 15.0, 9.9, 6.7, 5.2; FAB HRMS (NBNCsI) de833.2339, M + Csf ether (EtZO) were distilled from sodium-benzophenone, and methylene calcd for C37HszOllSi 833.2333. chloride (CHzClz), benzene (PhH), and toluene from calcium hydride. Enone 25. A. Oxidation of Alcohol 17 to 25. To a solution of Yields refer to chromatographically and spectroscopically ('H NMR) 7-TES-baccatin (17,30 mg, 0.043 "01) and 4-methylmorpholine homogeneous materials, unless otherwise stated. All solutions used III N-oxide (NMO, 7.5 mg, 0.064 "01) in acetonitrile (5 mL) was added in workup procedures are saturated unless otherwise noted. reagents All 4-A molecular sieves (20 mg), and the suspension was stirred at 25 "C were purchased at highest commercial quality and used without further for 5 min. A catalytic amount of tetrapropylammonium permthenate purification unless otherwise stated. (TPAP) was added, and the reaction mixture was stirred at 25 "C for reactions were monitored by thin-layer chromatography canied All 1.5 h. The reaction mixture was concentrated, suspended in CHzClz out on 0.25 mm E. Merck silica gel plates (6OF-254) using W light (20 mL), and filtered through silica gel. Elution with CHzClz (20 mL) as visualizing agent and 7% ethanolic phosphomolybdic acid or and 50% EtOAc in hexanes (20 mL), followed by concentration, gave p-anisaldehyde solution and heat as developing agents. E. Merck silica enone 25 (29 mg, 98%) as a white solid. gel (60, particle size 0.040-0.063 mm) was used for flash column chromatography.62 Preparative thin-layer chromatography (F'TLC) B. Conversion of Carbonate 30 to 25. A solution of carbonate separations were carried out on 0.25 or 0.50 mm E. Merck silica gel 30 (17.6 mg, 0.028 "01) in THF (2 mL) at -78 "C was treated with plates (6OF-254). PhLi (0.070 mL, 2 M in cyclohexane, 0.14 "01) and stirred at -78 NMR spectra were recorded on Brucker AMX-500 or AM-300 "C for 15 min. The reaction was quenched with aqueous NH&l (1 instruments and calibrated using residual undeuterated solvent as an mL), and the resulting mixture was allowed to warm to 25 "C. After intemal reference. The following abbreviations were used to explain dilution with Et20 (10 mL), the organic layer was separated, dried (MgSOd), and concentrated to give hydroxy benzoate 25 containing the multiplicities: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; ca. 10% of the 10-deacetylated compound ('H band, several overlapping signals; b, broad. The carbon numbering of NMR). The crude Taxol (1) was used to assign protons. IR spectra were recorded on a mixture was dissolved in CHzClz (1.5 mL), treated with 4-(dimethy- Perkin-Elmer 1600 series FT-IR spectrometer. Optical rotations were 1amino)pyridine (DMAP, 61.0 mg, 0.50 mmol) and acetic anhydride recorded on a Perkin-Elmer 241 polarimeter. High-resolution mass (0.024 mL, 0.25 mmol), and stirred at 25 "C for 1 h. The reaction spectra (HRMS) were recorded on a VG ZAB-ZSE mass spectrometer mixture was diluted with Et20 (10 mL), washed with 10% aqueous HCl (5 (5 (5 under fast atom bombardment (FAB) conditions. Melting points (mp) mL), 10% aqueous NaOH mL), and brine mL), dried are uncorrected, recorded on a Thomas Hoover capillary melting point (MgS04), concentrated, and purified by flash chromatography (silica, 25 50% EtOAc in petroleum ether) to give hydroxy benzoate 25 apparatus. - (15.9 mg, 80%) as a white solid. Experimental techniques and data for compounds 15, 16, 18, and 28 may be found in the supplementary material. C. Acetylation of Alcohol 31 to 25. To a solution of alcohol 31 7-TES-baccatinIII (17). A. Silylation of 15 to 17. To a solution (650 mg, 0.989 mmol) in CH2C12 (50 mL) were added 4-(dimethyl- of baccatin III (15,165 mg, 0.28 "01) in pyridine (14 mL) was added amino)pyridine (DMAP, 600 mg, 4.9 "01) and acetic anhydride (0.9 chlorotriethylsilane (1.42 mL, 8.45 mol) dropwise. The solution was mL, 9.89 "01). The solution was stirred at 25 "C for 2.5 h, the stirred at 25 "C for 24 h. After dilution with Et20 (100 mL), the reaction was quenched with aqueous NaHC03 (10 mL), and the solution was washed with aqueous CuSO4 (3 x 20 mL) and brine (20 resulting mixture was diluted with Et20 (100 mL), washed with 10% aqueous HCl (50 mL), 10% aqueous NaOH (50 mL), and brine (30 mL). The organic extract was dried (MgSOd), concentrated, and purified by flash chromatography (silica, 35 - 50% EtOAc in mL), dried (MgSOd), concentrated, and purified by flash chromatog- petroleum ether) to give 17 (168 mg, 85%) as a white solid. raphy (silica, 35% EtOAc in petroleum ether) to give acetate 25 (657 B. Acetylation of 16 to 17. To a solution of 7-TES-10-deacetyl- mg, 95%) as a white solid. baccatin ID (16, 0.21 g, 0.318 mol) in pyridine (8 mL) at 0 "C was D. Allylic Oxidation of 35 to 25. A solution of 35 (1.3 mg, 0.0019 added acetyl chloride (0.1 13 mL, 1.59 "01) dropwise. The solution "01) in benzene (0.5 mL) was treated with anhydrous NaOAc (4.7 was stirred at 0 "C for 48 h. After dilution with Et20 (20 mL), the mg, 0.057 mmol), anhydrous Celite (12.0 mg), and pyridinium reaction was quenched with aqueous NaHCO3 (10 mL). The organic chlorochromate (12.0 mg, 0.056 mol) and stirred at reflux for 1 h. layer was separated, washed with aqueous CuSO4 (2 x 10 mL) and The reaction mixture was filtered through silica gel, eluted with Et20 brine (5 mL), dried (MgSOd), concentrated, and purified by flash (20 mL), concentrated, and purified by preparative TLC (silica, 30% chromatography (silica, 25 - 50% EtOAc in petroleum ether) to give Et20 in benzene) to give enone 25 (1.0 mg, 75%) as a film: Rj = 0.5 7-TES-baccatin LII (17, 183 mg, 82%) as a white solid. (silica, 50% EtOAc in hexanes); [aIz2~- 19.8 (c 0.85, CHC13); IR C. Reduction of Enone 25 to 17. A solution of enone 25 (10 mg, (thin film) vmax3499, 2956, 1758, 1732, 1673, 1657, 1604 cm-'; 'H 0.014 mmol) in MeOH (2 mL) was treated with an excess of NaBa NMR(~~~MHZ,CDCI~)~~.O~(~,J=~.~H~,~H,BZ),~.~~(~,J= for 3 h at 25 "C. The reaction was quenched with aqueous NH&l(l 7.5 Hz, 1 H, Bz), 7.47 (t, J= 7.8 Hz, 2 H, Bz), 6.57 (s, 1 H, 10-H), mL), and the resulting mixture was stirred at 25 "C for 15 min. After 5.67 (d, J = 6.7 Hz, 1 H, 2-H), 4.90 (d, J = 8.4 Hz, 1 H, 5-H), 4.46 dilution with water (5 mL), the reaction mixture was extracted with (dd, J = 10.4, 6.8 Hz, 1 H, 7-H), 4.31 (A of AB, d, J = 8.5 Hz, 1 H, CHzClz (3 x 10 mL). The combined organic layer was dried (Naz- 20-H), 4.09 (B of AB, d, J= 8.5 Hz, 1 H, 20-H), 3.89 (d, J= 6.7 Hz, sod), concentrated, and purified by flash chromatography (silica, 25 1 H, 3-H), 2.92 (A' Of A'B', d, J = 19.9 Hz, 1 H, 14-H), 2.63 (B' of 50% EtOAc in petroleum ether) to give starting enone 25 (1.2 mg, A'B', d, J = 19.9 Hz, 1 H, 14-H), 2.50 (m, 1 H, 6-H), 2.21 (s, 3 H, - OAc), 2.17 (s, 3 H, OAc), 2.16 (s, 3 H, 18-CH3), 1.82 (m, 1 H, 6-H), (62) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923. 1.65 (s, 3 H, 19-CH3), 1.25 (s, 3 H, 16-CH3), 1.17 (s, 3 H, 17-CH3), Total Synthesis of Taxol. 1 J. Am. Chem. Soc., Vol. 117, No. 2, 1995 631

0.90 (t, J = 7.9 Hz,9 H, Si(CHzCH3)3), 0.65-0.45 (band, 6 H, Si(CH2- Acetate 30. To a solution of carbonate 29 (86.0 mg, 0.159 mmol) CH3)3); 13C NMR (125 MHz, CDC13) 6 200.2, 198.3, 170.1, 168.9, in CHzClz (2 mL) were added 4-(dimethy1amino)pyridine (DMAP, 166.8, 153.0, 140.2, 133.9, 130.0, 128.8, 128.7, 83.9, 80.5,78.4, 76.1, 177.0 mg, 1.45 mmol) and acetic anhydride (0.069 mL, 0.723 mmol). 76.0, 72.8, 72.2, 59.4, 46.2, 43.4, 42.4, 37.1, 33.0, 21.7, 21.0, 18.2, The solution was stirred at 25 "C for 0.5 h, diluted with Et20 (100 13.5, 9.5, 6.7, 5.1; FAB HRMS (NBA) m/e 699.3220, M + Hf calcd mL), washed with 10% aqueous HCl(5 mL), 10% aqueous NaOH (5 for C37H50011Si 699.3201. mL) and brine (5 mL), dried (MgSOd), concentrated, and purified by Diol 26. A. Hydrolysis of 25 to 26. To a solution of enone 25 flash chromatography (silica, 10 - 50% EtOAc in petroleum ether) to (124 mg, 0.034 mmol) in MeOH (29 mL) at 0 "C was added an aqueous give carbonate 30 (94 mg, 95%) as an amorphous solid Rj = 0.50 solution of KzC03 (291 mg in 7.3 mL HzO). The solution was stirred (silica, 35% EtOAc in hexanes); [alZz~+14 (c 0.5, CHC13); IR (thin at 0 "C for 4 h. The reaction was quenched with aqueous NH&1(30 film) vmax2926, 1823, 1754, 1731, 1689 cm-'; 'H NMR (500 MHz, mL), and the resulting mixture was extracted with CHC13 (2 x 50 mL). CDCl3) 6 6.52 (s, 1 H, IO-H), 4.89 (d, J = 9.0 Hz, 1 H, 5-H), 4.60 (A The organic layer was dried (MgSOd), concentrated, and purified by of AB, d, J = 9.0 Hz, 1 H, 20-H), 4.48 (d, J = 5.5 Hz, 1 H, 2-H), 4.45 flash chromatography (silica, 25 - 50% EtOAc in petroleum ether) to (B of AB, d, J = 9.0 Hz, 1 H, 20-H), 4.42 (dd, J = 9.5, 7.0 Hz, 1 H, give triol 26 (96 mg, 91%) containing a small amount of the 7-H), 3.49 (d, J = 5.5 Hz, 1 H, 3-H), 2.90 (A' of AB', d, J = 20.0 Hz, 10-acetylated product ('H NMR). 1 H, 14-H), 2.78 (B' of A'B', d, J = 20.0 Hz, 1 H, 14-H), 2.55 (m, 1 B. Hydrolysis of 31 to 26. To a solution of enone 31 (1.44 g, H, 6-H), 2.19 (s, 3 H, OAC), 2.16 (s, 3 H, OAC), 2.07 (s, 3 H, 18- 2.19 mmol) in MeOH (300 mL) at 0 "C was slowly added an aqueous CH3), 1.87 (m, 1 H, 6-H), 1.71 (s, 3 H, 19-CH3), 1.28 (s, 3 H, 16- solution of KzC03 (3.0 g in 32 mL of HzO). The solution was stirred CH3), 1.26 (s, 3 H, 17-CH3), 0.89 (t, J = 8.0 Hz, 9 H, Si(CHzCH&), at 0 "C for 2.5 h. The reaction was quenched with aqueous NH&l 0.60-0.50 (band, 6 H, Si(CHzCH3)3); 13C NMR (125 MHz, CDCl,) 6 (150 mL), and the resulting mixture was extracted with CHzClz (2 x 200.2, 195.7, 170.5, 168.7, 152.0, 150.4, 142.5, 88.2, 83.9, 79.8, 79.2, 200 mL). The organic layer was dried (NazSOd), concentrated, and 76.6, 75.7, 71.5, 61.0, 43.1, 39.8, 37.7, 31.6, 21.5, 20.7, 18.4, 14.4, purified by flash chromatography (silica, 35 - 50% EtOAc in 9.7, 6.7, 5.1; FAB HRMS (NBA) m/e 621.2745, M + Hf calcd for petroleum ether) to give enone 31 (270 mg, 19%) and triol 26 (912 C~lH44011Si621.273 1. mg, 93% based on 81% conversion): Rj = 0.24 (silica, 50% EtOAc in Enone 31. A. Oxidation of 16 to 31. To a solution of 7-TES- hexanes); [a]*% +38 (c 0.15, CHC13); IR (thin film) vmax3414, 2957, deacetylbaccatin 111 (16, 1.5 g, 2.28 mmol) and 4-methylmorpholine 2881, 1727, 1664, 1370 cm-'; IH NMR (500 MHz, CDC13) 6 5.23 (d, N-oxide (NMO, 240 mg, 2.05 mmol) in CHzClz (5 mL) was added J = 9.5 Hz, 1 H, 10-H), 4.89 (d, J = 9.5 Hz, 1 H, 5-H), 4.63 (A of 4-A molecular sieves (200 mg), and the suspension was stirred at 25 AB, d, J = 9.5 Hz, 1 H, 20-H), 4.56 (B of AB, d, J = 9.5 Hz, 1 H, "C for 10 min. A catalytic amount of tetrapropylammonium permth- 20-H), 4.32 (dd, J = 11.0, 7.0 Hz, 1 H, 7-H), 4.28 (d, J = 2.5 Hz, 1 enate (TPAP, 40 mg, 0.11 mmol) was added by portions, and the H, 10-OH), 3.89 (dd, J = 6.5, 4.0 Hz, 1 H, 2-H), 3.57 (d, J = 6.5 Hz, reaction mixture was stirred at 25 "C for 0.5 h. Small amounts of 1 H, 3-H), 2.78 (A' of A'B', d, J = 19.5 Hz, 1 H, 14-H), 2.58 (d, 4.0 4-methylmorpholine N-oxide and TPAP were added altematively at Hz, 1 H, 2-OH), 2.52 (B' of A'B', d, J = 19.5 Hz, 1 H, 14-H), 2.46 0.5 h intervals until the starting material was consumed to the extent (m. 1 H, 6-H), 2.03 (s, 3 H, OAc), 1.88 (m, 1 H, 6-H), 1.68 (s, 3 H, of ca. 95% by TLC. The reaction mixture was filtered through silica 18-CH3), 1.21 (s, 3 H, 16-CH3), 1.04 (s, 3 H, I"-CHs), 0.90 (t, J= 8.0 gel, eluted with CHzClz (100 mL), and concentrated to give enone 31 Hz,9 H, Si(CHzCH&), 0.60-0.40 (band, 6 H, Si(CHzCH&); 13C NMR (1.44 g, 96%) as a white solid. (125 MHz, CDC13) 6 208.9, 198.5, 170.1, 156.7, 138.8, 83.8, 81.2, B. Conversion of Carbonate 29 to 31. A solution of carbonate 77.6, 75.7, 72.8, 72.5, 58.8, 45.8, 43.1, 42.8, 37.3, 32.7, 21.6, 17.5, 29 (1.5 mg, 0.0026 mmol) in THF (0.3 mL) at -78 "C was treated 13.6, 9.7, 6.7, 5.1; FAB HRMS (NBA/NaI) m/e 575.2648, M + Na+ with PhLi (0.013 mL, 0.026 mmol) and stirred at -78 "C for 0.5 h. calcd for CzgH4409Si 575.2652. The reaction was quenched with aqueous WCl (10 mL). After Carbonate 29. Method A. To a solution of diol 26 (96.0 mg, dilution with Et20 (20 mL), the organic layer was separated, washed 0.187 mmol) in pyridine (10 mL) at 0 "C was added phosgene (0.97 with brine (10 mL), dried (NazSOd), and purified by flash chromatog- mL of a 1.93 M solution in toluene, 1.87 mmol). The solution was raphy (silica, 25 - 35% EtOAc in petroleum ether) to give hydroxy stirred at 0 "C for 0.5 h and poured onto ice (10 mL). After dilution benzoate 31 (1.4 mg, 85%) as a film: Rf = 0.5 (silica, 50% EtOAc in with Et20 (25 mL), the organic layer was separated, washed with hexanes); [a]*'D +11 (c 0.56, CHC13); IR (thin film) vm 3446, 2957, aqueous CuSO4 (2 x 15 mL) and aqueous NaHC03 (20 mL), dried 2882, 1726, 1672, 1456, 1367, 1243, 1106 cm-'; IH NMR (500 MHz, (MgS04), and concentrated to give carbonate 29 (86 mg, 85%) as an CDC13) 6 8.05 (dd, J = 8.0, 1.0 Hz, 2 H, Bz), 7.61 (t, J = 7.5 Hz, 1 amorphous solid. H, Bz), 7.45 (t, J = 7.5 Hz, 2 H, Bz), 5.63 (d, J = 7.5 Hz, 1 H, 2-H), Method B. A solution of diol 26 (60.0 mg, 0.109 mmol) in THF 5.30 (d, J = 2.0 Hz, 1 H, 10-H), 4.90 (d, J = 8.0 Hz, 1 H, 5-H), 4.36 (2 mL) was treated with carbonyldiimidazole (1 10.0 mg, 0.678 mol) (dd, J = 10.5, 7.0 Hz, 1 H, 7-H), 4.31 (A of AB, d, J = 8.5 Hz, 1 H, and stirred at 40 "C for 0.5 h. The reaction mixture was concentrated 20-H), 4.30 (d, J = 2.0 Hz, 1 H, 10-OH), 4.11 (B of AB, d, J = 8.5 and redissolved in THF (5 mL). TLC analysis confmed total Hz, 1 H, 20-H), 3.93 (d, J = 7.5 Hz, 1 H, 3-H), 2.92 (A' of A'B', d, consumption of starting material. Then 1 N aqueous HCl(5 mL) was J= 19.5 Hz, 1 H, 14-H), 2.62 (B'of A'B', d, J= 19.5 Hz, 1 H, 14-H), added, and the resulting solution was allowed to stir for 15 min at 25 2.46 (m, 1 H, 6-H), 2.17 (s, 3 H, OAc), 2.08 (s, 3 H, 18-CH3), 1.87 OC. Et20 (25 mL) was added, and the organic layer was separated, (m, 1 H, 6-H), 1.77 (s, 1 H, 1-OH), 1.70 (s, 3 H, 19-CH3), 1.21 (s, 3 washed with aqueous NaHCO3 (10 mL) and brine (10 mL), dried H, 16-CH3), 1.14 (s, 3 H, 17-CH3), 0.90 (t, J = 8.0 Hz, 9 H, (MgS04), and concentrated to give carbonate 29 (58 mg, 93%) as a Si(CHzCH3)3), 0.60-0.42 (band, 6 H, Si(CHzCH3)3); 13C NMR (125 white foam: Rf = 0.50 (silica, 35% EtOAc in hexanes); [alZZ~+48 (c MHz, CDC13) 6 208.2, 198.1, 170.2, 166.8, 156.6, 139.1, 134.0, 130.0, 0.5, CHCh); IR (thin film) vmax3438, 2957, 2882, 1820, 1731, 1685, 128.8, 128.8, 84.0, 80.4, 78.5, 76.2, 75.7, 72.9, 72.8, 58.8, 45.9, 43.4, 1370, 1236 cm-'; 'H NMR (500 MHz, CDC13) 6 5.27 (d, J = 2.5 Hz, 42.5, 37.2, 33.0, 21.7, 17.5, 13.6, 9.6, 6.7, 5.1; FAB HRMS (NBA/ 1 H, 10-H), 4.89 (d, J = 9.0 Hz, 1 H, 5-H), 4.60 (A of AB, d, J = 9.0 NaI) de657.3070, M + Na+ calcd for C35&g010Si 657.3095. Hz, 1 H, 20-H), 4.45 (B of AB, d, J = 9.0 Hz, 1 H, 20-H), 4.43 (d, J Thiocarbamate 34. A solution of 7-TES-baccatin I11 (17,48 mg, = 6.0 Hz, 1 H, 2-H), 4.33 (dd, J = 10.0, 7.5 Hz, 1 H, 7-H), 4.28 (d, 0.069 mmol) in THF (1 mL) was treated with 4-(dimethylamino)- J = 2.5 Hz, 1 H, lO-OH), 3.54 (d, J = 6.0 Hz, 1 H, 3-H), 2.88 (A' of pyridine (DMAP, 251 mg, 2.05 mmol) and (thiocarbony1)diimidazole AB', d, J = 20.0 Hz, 1 H, 14-H), 2.75 (B' of A'B', d, J = 20.0 Hz, 1 (244 mg, 1.37 mmol) and stirred at 75 "C in a sealed flask for 18 h. H, 14-H), 2.50 (m, 1 H, 6-H), 2.08 (s, 3 H, OAc), 2.06 (s, 3 H, 18- The reaction mixture was diluted with EtOAc (15 mL), washed with CH,), 1.88 (m, 1 H, 6-H), 1.77 (s, 3 H, 19-C&), 1.31 (s, 3 H, 16- 10% aqueous HC1 (5 mL) and aqueous NaHC03 (10 mL), dried CH3), 1.15 (s, 3 H, 17-CH3), 0.88 (t. J = 8.5 Hz, 9 H, Si(CHzCH&), (MgS04), concentrated, and purified by flash chromatography (silica, 0.55-0.45 (band, 6 H, Si(CHzCH&); 13C NMR (125 MHz, CDC13) 6 20 - 50% EtOAc in petroleum ether) to give thiocarbamate 34 (48 208.4, 195.5, 170.5, 154.0, 152.0, 141.2, 88.4, 83.9, 79.8, 79.0, 76.7, ing, 86%) as a white solid Rf = 0.27 (silica, 25% EtOAc in benzene); 75.7, 71.9, 60.3, 43.0,41.6, 39.8, 37.7, 31.6, 21.5, 17.8, 14.4,9.7, 6.6, -59 (C 0.17, CHCls); IR (thin film) v,, 3478,2954,1726, 1465, 5.0; FAB HRMS (NBA) m/e 579.2652, M + H+ calcd for CZ~&ZO~O- 1388, 1284, 1238, 1104 cm-'; 'H NMR (500 MHz, CDC13) 6 8.50 (s, Si 579.2626. 1 H, imid.), 8.01 (d, J = 7.5 Hz, 2 H, Bz), 7.79 (s, 1 H, imid.), 7.56 632 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al. (t, J = 7.5 Hz, 1 H, Bz), 7.42 (t. J = 7.5 Hz, 2 H, Bz), 7.10 (s, 1 H, 1237 cm-I; 'H NMR (500 MHz, CDC13) 6 6.38 (s, 1 H, 10-H), 4.95 imid.), 6.53 (t, J = 9.0 Hz, 1 H, 13-H), 6.46 (s, 1 H, 10-H), 5.66 (d, (dd, J = 9.5, 1.5 Hz, 1 H, 5-H), 4.64 (A of AB, d, J = 9.0 Hz, 1 H, J = 7.0 Hz, 1 H, 2-H), 4.89 (d, J = 8.5 Hz, 1 H, 5-H), 4.46 (dd, J = 20-H), 4.55 (B of AB, d, J = 9.0 Hz, 1 H, 20-H), 4.40 (dd, J = 10.5, 10.5, 7.0 Hz, 1 H, 7-H), 4.25 (A of AB, d, J = 8.5 Hz, 1 H, 20-H), 7.0 Hz, 1 H, 7-H), 3.83 (dd, J= 6.5, 4.5 Hz, 1 H, 2-H), 3.38 (d, J= 4.13 (B of AB, d, J = 8.5 Hz, 1 H, 20-H), 3.85 (d, J = 7.0 Hz, 1 H, 6.5 Hz, 1 H, 3-H), 2.69-2.58 (band, 1 H, 13-H), 2.54 (d, 4.5 Hz, 1 H, 3-H), 2.72 (dd, J = 15.0, 9.0 Hz, 1 H, 14-H), 2.53 (m, 1 H, 6-H), 2.21 2-OH), 2.51 (m, 1 H, 6-H), 2.16 (s, 3 H, OAc), 2.14 (s, 3 H, OAc), (s, 3 H, OAc), 2.17 (s, 3 H, OAc), 2.12 (dd, J = 15.0, 7.5 Hz, 1 H, 2.13-2.01 (band, 1 H, 13-H), 2.01 (s, 3 H, 18-CH3), 1.92-1.83 (band, 14-H), 1.91 (s, 3 H, 18-CH3), 1.88 (m, 1 H, 6-H), 1.65 (s, 3 H, 19- 2 H, 14-CH2), 1.78 (m, 1 H, 6-H), 1.62 (s, 3 H, 19-C&), 1.07 (s, 3 H, CH3), 1.25 (s, 3 H, 16-CH3), 1.17 (s, 3 H, 17-CH3), 0.91 (t, J = 8.0 16-CH3), 1.05 (s, 3 H, 17-C&), 0.88 (t, J= 7.5 Hz,9 H, Si(CHzCH&), Hz,9 H, Si(CHzCH3)3), 0.62-0.51 (band, 6 H, Si(CHzCH,)3); 13C NMR 0.61-0.48 (band, 6 H, Si(CH2CHs)s); FAB HRMS (NBA/NaI) m/e (125 MHz, CDCl3) 6 201.5, 183.5, 170.0, 169.3, 166.9, 138.7, 137.7, 603.2970, M -t Naf calcd for C3&809Si 603.2965. 134.8, 133.8, 131.3, 130.0, 128.9, 128.6, 118.3, 84.1, 81.3, 80.1, 78.9, Carbonate 38. A. Conversion of Diol 37 to Carbonate 38. To 76.6, 75.0, 74.3, 72.5, 58.9, 46.8, 43.3, 37.4, 35.0, 29.7, 26.9, 21.9, a solution of diol 37 (16 mg, 0.028 "01) in pyridine (2 mL) at 25 "C 20.9, 20.3, 15.4, 9.9, 6.7, 5.2; FAB HRMS (NBA/NaI) de833.3110, was added phosgene (0.143 mL of a 1.93 M solution in toluene, 0.28 M + Naf calcd for C41H5401lN~SSi833.3115. mol). The solution was stirred at 25 "C for 15 min. After dilution Benzoate 35. A. Deoxygenation of 34 to 35. To a solution of with Et20 (20 mL), the organic layer was separated, washed with thiocarbamate 34 (960 mg, 1.18 mmol) in degased toluene (250 mL) aqueous CuSO4 (3 x 10 mL) and aqueous NaHC03 (10 mL), dried stirred at 85 "C were added tributyltin hydride (3.2 mL, 11.8 "01) (MgSOd), concentrated, and purified by flash chromatography (silica, and azobis(isobutyronitri1e) (AIBN, 16.4 mg in 1 mL of toluene, 0.1 10 - 35% EtOAc in petroleum ether) to give carbonate 38 (14.4 mg, mmol). The reaction mixture was stirred at 85 "C for 2 h, concentrated, 86%) as a white foam. and purified by flash chromatography (silica, 15 - 25% EtOAc in B. Silylation of 39 to 38. A solution of alcohol 39 (1.0 mg, 0.002 petroleum ether) to give a mixture of alcohol 35 and isomer 36 (620 mol) in pyridine (0.5 mL) was treated with chlorotriethylsilane mg, 76%) as one single fraction containing 77% of 35 (59% yield) (TESCl, 0.017 mL, 0.1 mol) and stirred at 25 "C for 24 h. After and 23% of 36 (17% yield). Analytical samples of both isomers were dilution with Et20 (10 mL), the organic layer was separated, washed obtained by preparative TLC (silica, 30% EtOAc in benzene). with aqueous CuSO4 (3 x 5 mL), dried (MgSOd), concentrated, and Isomer 35: Rf = 0.47 (silica, 25% EtOAc in benzene); [alZz~-50.6 purified by flash chromatography (silica, 10 - 35% EtOAc in (c 0.5, CHC13); IR (thin film) vmax3517,2922, 1728, 1456, 1371, 1242, petroleum ether) to give carbonate 38 (1.0 mg, 85%) as a colorless 1109 cm-'; 'H NMR (500 MHz, CDC13) 6 8.06 (d, J = 8.0 Hz, 2 H, film: Rf= 0.82 (silica, 50% EtOAc in hexanes); [a]"~-49.4 (c 0.93, Bz), 7.58 (t, J = 7.5 Hz, 1 H, Bz), 7.45 (t, J = 7.5 Hz, 2 H, Bz), 6.45 CHCb); IR (thin film) v, 2924, 1814, 1728,1461, 1372,1238 cm-I; (s, 1 H, 10-H), 5.59 (d, J = 7.0 Hz, 1 H, 2-H), 4.95 (d, J = 9.0 Hz, 1 'H NMR (500 MHz, cDc13) 6 6.40 (s, 1 H, 10-H), 4.95 (d, J = 9.0 H, 5-H), 4.45 (dd, J = 10.5, 7.0 Hz, 1 H, 7-H), 4.30 (A of AB, d, J = Hz, 1 H, 5-H), 4.60 (A of AB, d, J = 9.0 Hz, 1 H, 20-H), 4.47 (B of 8.5 Hz, 1 H, 20-H), 4.14 (B of AB, d, J= 8.5 Hz, 1 H, 20-H), 3.75 (d, AB, d, J= 9.0 Hz, 1 H, 20-H), 4.43 (dd, J= 10.0,7.5 Hz, 1 H, 7-H), J = 7.0 Hz, 1 H, 3-H), 2.65 (m, 1 H, 13-H), 2.53 (m, 1 H, 13-H), 2.30 4.39 (d, J = 5.5 Hz, 1 H, 2-H), 3.36 (d, J = 5.5 Hz, 1 H, 3-H), 2.71 (s, 3 H, OAC), 2.30-2.17 (band, 2 H, 6-CH2), 2.16 (s, 3 H, OAc), 2.07 (m, 1 H, 13-H), 2.56 (m, 1 H, 13-H), 2.17 (s, 3 H, OAc), 2.15 (s, 3 H, (s, 3 H, 18-CH3), 1.93-1.81 (band, 2 H, 14-CH2), 1.64 (s, 3 H, 19- OAc), 2.12 (m. 1 H), 2.07 (s, 3 H, 18-CH3), 1.97 (m, 1 H), 1.88 (m, 2 CH3). 1.18 (s, 3 H, 16-CH3), 1.04 (s, 3 H, 17-C&), 0.89 (t, J = 8.0 H), 1.78 (s, 3 ,H, 19-CH3), 1.23 (s, 3 H, 16-CH3), 1.17 (s, 3 H, 17- Hz, 9 H, Si(CHtCH&), 0.65-0.49 (band, 6 H, Si(CHzCH3)3); 13CNMR CH,), 0.88 (t, J = 7.5 Hz, 9 H, Si(CH2CH3)3), 0.60-0.50 (band, 6 H, (125 MHz, CDCl3) 6 202.5, 169.8, 169.5, 167.0, 141.6, 133.6, 132.3, Si(CHzCH&); 13C NMR (125 MHz, CDC13) 6 202.6, 170.3, 169.2, 130.0, 129.3, 128.6, 84.0, 81.4, 80.6, 76.5,75.9,73.8,72.4,58.8,46.8, 153.1, 144.0, 130.7, 92.8, 84.0, 80.3, 80.0,76.4,76.1,60.3,43.5,38.0, 42.2, 37.4, 30.0, 26.7, 25.2, 22.1, 21.1, 20.5, 19.0, 9.6, 6.7, 5.3; FAB 29.7,29.4,25.5,23.1,21.9,21.1,19.1,9.8,6.7,5.2;FABHRMS(NBN HRMS (NBNCsI) m/e 817.2380, M + Csf calcd for C37H52010Si CsI) de739.1929, M + Cs+ calcd for C31&010Si 739.1915. 817.2384. Alcohol 39. A solution of silyl ether 38 (3.0 mg, 0.0049 "01) in Isomer 36: Ry = 0.48 (silica, 25% EtOAc in benzene); 'H NMR THF (1.5 mL) was treated with HF-pyridine (0.5 mL) and stirred for 2 (500 MHz, CDC13) 6 8.04 (d, J = 7.0 Hz, 2 H, Bz), 7.58 (t, J = 7.5 h at 25 "C. The reaction mixture was diluted with EtOAc (10 mL), Hz, 1 H, Bz), 7.47 (t, J= 7.5 Hz, 2 H, Bz), 5.96 (s, 1 H, 10-H), 5.48 and the reaction was quenched with aqueous NaHC03 (10 mL). The (dd, J = 5.0, 1.5 Hz, 1 H, 2-H), 5.45 (m, 1 H, 13-H), 4.98 (dd, J = organic layer was separated, washed with 10% aqueous NaOH (10 mL) 8.3, 1.9Hz, lH,5-H),4.39(AofAB,d,J=8.5Hz,1H,20-H),4.35 and brine (10 mL), dried (MgSOd), and purified by preparative TLC (dd, J = 10.4, 6.5 Hz, 1 H, 7-H), 4.25 (B of AB, d, J = 8.5 Hz, 1 H, (silica, 30% EtOAc in petroleum ether) to give alcohol 39 (2.1 mg, 20-H), 4.00 (d, J = 5.0 Hz, 1 H, 3-H), 2.72 (m, 1 H, 14-H), 2.48 (m, 88%) as a colorless film: Rf = 0.22 (silica, 50% EtOAc in petroleum 1 H, 6-H), 2.29 (s, 3 H, OAC), 2.16 (s, 3 H, OAC), 2.05-1.93 (band, ether); [a12%-23 (c 1.0, CHC13); IR (thin film) vmax2923,2854, 1809, 2 H, 6-H and 14-H), 1.89 (s, 3 H, 18-CH3), 1.60 (s, 3 H, 19-C&), 1723, 1460, 1374, 1238, 1018 cm-'; 'H NMR (500 MHz, CDC13) 6 1.23 (s, 3 H, 16-CH3),1.07 (s, 3 H, 17-C&), 0.88 (t, J = 8.0 Hz,9 H, 6.25 (s, 1 H, 10-H), 4.94 (d, J = 8.0 Hz, 1 H, 5-H), 4.56 (A of AB, d, Si(CH2CH3)3), 0.65-0.49 (band, 6 H, Si(CH&H&). J = 9.0 Hz, 1 H, 20-H), 4.41 (B of AB, dd, J = 9.0, 0.5 Hz, 1 H, B. Conversion of Carbonate 38 to 35. A solution of carbonate 20-H), 4.38 (m, 1 H, 7-H), 4.31 (d, J = 5.5 Hz, 1 H, 2-H), 3.33 (d, J 38 (1 mg, 0.0016 "01) in THF (1 mL) at -78 "C was treated with = 5.5 Hz, 1 H, 3-H), 2.73 (m, 1 H, 13-H), 2.57 (m, 1 H, 6-H), 2.28 (d, PhLi (0.016 mL, 2 M in cyclohexane, 0.008 mmol) and stirred at -78 J 4.0 Hz, 1 H, OH), 2.15 (s, 3 H, OAc), 2.13 (s, 3 H, OAc), 2.12- "C for 15 min. The reaction was quenched with aqueous WCl (2 2.02 (band, 2 H, 14-CH2), 1.94 (d, J = 1.0 Hz, 3 H, 18-CH3), 1.95- mL). After dilution with Et20 (10 mL), the organic layer was separated, 1.80 (band, 2 H, 13-H and 6-H), 1.65 (s, 3 H, 19-CH3), 1.18 (s, 3 H, dried (MgSOd), concentrated, and purified by preparative TLC (silica, 16-CH3), 1.08 (s, 3 H, 17-CH3); 13C NMR (125 MHz, CDCl3) 6 204.3, 25% Et20 in benzene) to give benzoate 35 (0.9 mg, 80%) as a colorless 170.9, 170.2, 153.0, 146.4,92.8,84.2, 80.4,76.7,75.9,71.5,60.3,43.0, film. 36.4, 31.0, 29.7, 29.6, 25.5, 23.2, 21.9, 21.6, 20.9, 19.0, 9.2. Diol 37. To a mixture of benzoates 35 and 36 (71.8 mg, 0.105 mmol, 2',7-diTES-Taxol (42). To a solution of 7-TES-baccatin III (17, ca. 77:23) in MeOH (13.5 mL) and THF (3.6 mL) at 0 "C was added 20.0 mg, 0.0285 "01) and B-lactam 40 (38 mg, 0.0998 mmol) in an aqueous solution of KzCO3 (270 mg in 3.5 mL of H20). The solution THF (1.5 mL) at 0 "C was added NaN(SiMe& (0.086 mL of a 1.0 M was stirred at 0 "C for 6 h and at -20 OC for 10 h. The reaction was solution in THF, 0.086 "01) dropwise. The reaction mixture was quenched with aqueous WCl(20 mL), and the resulting mixture was stirred at 0 "C for 0.5 h, and the reaction was quenched with aqueous extracted with CHC13 (2 x 100 mL). The organic layer was dried mC1(2 mL). After dilution with Et20 (15 mL), the organic layer (MgSOd), concentrated, and purified by flash chromatography (silica, was separated, washed with brine (5 mL), dried (MgSOd), concentrated, 20 - 40% EtOAc in petroleum ether) to give the benzoate mixture and purified by flash chromatography (silica, 10 - 50% EtOAc in 35/36 (27 mg, 38%) and diol 37 (27 mg, 94% based on 62% petroleum ether) to give starting material 17 (2.2 mg, 11%) and 2',7- conversion): Rf= 0.18 (silica, 50% EtOAc in hexanes); -43.6 diTES-Taxol(42) (23.7 mg, 86% based on 89% conversion) as a white (c 0.28, CHCl3); IR (thin film) vmiu 3479, 2923, 1721, 1461, 1372, solid: Rf = 0.59 (silica, 50% EtOAc in hexanes); [a]22D-48 (c 0.4, Total Synthesis of Taxol. I J. Am. Chem. Soc., Vol. 117, No. 2, 1995 633

CHCl3); IR (thin film)Y,, 3440, 2958, 1719, 1664 cm-'; 'H NMR lH,7-H),4.31(AOfAB,d,J=8.5H~,lH,20-H),4.19(BofAB, (500 MHz, CDC13) 6 8.11 (d, J = 7.0 Hz, 2 H, Bz), 7.72 (d, J = 7.5 d, J = 8.5 Hz, 1 H, 20-H), 3.79 (d, J = 7.0 Hz, 1 H, 3-H), 3.61 (d, J Hz, 2 H, Bz), 7.60-7.25 (band, 11 H, Ar), 7.11 (d, J= 9.0 Hz, 1 H, = 5.5 Hz, 1 H, 2'-OH), 2.55 (m, 1 H, 6-H), 2.49 (d, J = 4.0 Hz, 1 H, NH), 6.43 (s, 1 H, 10-H), 6.22 (b t, J = 8.5 Hz, 1 H, 13-H), 5.69 (m, 7-OH), 2.39 (s, 3 H, OAC), 2.40-2.25 (band, 2 H, 14-CHz), 2.24 (s, 3 2 H, 3'-H and 2-H), 4.93 (b d, J = 8.0 Hz, 1 H, 5-H), 4.69 (d, J = 2.0 H, OAc), 1.88 (m, 1 H, 6-H), 1.82 (s, 1 H, 1-OH), 1.79 (s, 3 H, 18- H~,lH,2'-H),4.45(dd,J=11.0,7.0H~,lH,7-H),4.30(AofAB,CH3), 1.69 (S, 3 H, 19-CH3), 1.24 (s, 3 H, 16-CH3), 1.14 (s, 3 H, 17- d, J 8.5 Hz, 1 H, 20-H), 4.19 (B of AB, d, J = 8.5 Hz, 1 H, 20-H), CH3); I3C NMR (125 MHz, CDC13) 6 203.6, 172.7, 171.3, 170.4, 167.0, 3.82 (d, J= 7.0 Hz, 1 H, 3-H), 2.53 (s, 3 H, OAC), 2.38 (dd, J= 9.5, 167.0, 142.0, 137.9, 133.7, 133.6, 133.1, 132.0, 130.2, 129.1, 129.0, 15.0 Hz, 1 H, 14-H), 2.18 (s, 3 H, OAC), 2.12 (dd, J= 15.0, 8.0 Hz, 128.7, 128.7, 128.4, 127.1, 127.0, 84.4, 81.1, 79.0, 76.5, 75.5, 74.9, 1 H, 14-H), 2.00 (s, 3 H, 18-CH3). 1.89 (m, 2 H, 6-CHz), 1.68 (s, 3 H, 73.2, 72.3, 72.2, 58.6, 55.0, 45.6, 43.1, 35.6, 35.6, 26.8, 22.6, 21.8, 19-CH3), 1.20 (s, 3 H, 16-CH3), 1.16 (s, 3 H, 17-CH3), 0.89 (t, J= 8.0 20.9, 14.9, 9.5; FAB HFWS (NBA) mle 854.3360, M + H+ calcd for Hz, 9 H, Si(CHzCH&), 0.80 (t, J= 8.0 Hz, 9 H, Si(CHzCH&), 0.62- C47H51014N 854.3388. 0.51 (band, 6 H, Si(CHzCH3)3), 0.51-0.35 (band, 6 H, Si(CHzCH&); '3CNMR(125MH~,CDcb)6201.7,170.1, 169.3, 167.2, 167.0, 140.1, 138.3, 134.2, 133.7, 133.6, 131.8, 130.2, 130.1, 129.2, 128.7, 128.3, Acknowledgment. We thank Dr. E. Bombardelli for a 127.9, 127.0, 126.4, 84.2, 81.2, 78.7,76.6, 75.0,74.9, 74.8,72.2, 71.5, generous gift of 10-deacetylbaccatin III and Drs. Dee H. Huang 58.4, 55.7, 46.6, 43.3, 37.2, 35.5, 26.5, 23.1, 21.5, 20.9, 14.1, 10.1, and Gary Siuzdak for NMR and mass spectroscopic assistance, 6.7, 6.5, 5.3, 4.3; FAB HRMS (NBA/CsI) mle 1214.4089, M + Cs+ respectively. This work was financially supported by NIH, The calcd for C59H79014NSi~1214.4093. Scripps Research Institute, fellowships from Mitsubishi Kasei Taxol (1). A solution of silyl ether 42 (22 mg, 0.020 mol) in Corporation (H.U.), Rh8ne-Poulenc Rorer (P.G.N.), The Office THF (1 mL) was treated with HFTyridine (0.2 mL) and stirred for of Naval Research (R.K.G.), The Agricultural University of 1.25 h at 25 "C. The reaction mixture was diluted with Et20 (15 mL), Athens (E.A.C.), R.W. Johnson-ACS Division of Organic and the reaction was quenched with aqueous NaHCO3 (5 mL). The Chemistry (E.J.S.), and grants from Merck Sharp & Dohme, organic layer was separated, washed with aqueous CuS04 (2 x 5 mL) and brine (5 mL), dried (Na~S04),and purified by flash chromatography Pfizer, Inc., Schering Plough and the ALSAM Foundation. (silica, 50 - 75% EtOAc in petroleum ether) to give Taxol (1, 13.9 mg, 80%) as a white solid Rf= 0.125 (silica, 50% EtOAc in hexanes); Supplementary Material Available: Experiment techniques -49 (c 0.45, MeOH); IR (thin film) vmax3432,2937,1720, 1652, and data for compounds 15, 16, 18, and 28 (3 pages). This 1520, 1241 cm-I; 'H NMR (500 MHz, CDC13) 6 8.13 (dd, J = 8.5, material is contained in many libraries on microfiche, im- 1.2Hz,2H,Bz),7.74(dd,J=8.2, 1.2Hz,2H,Bz),7.62,(t,J=7.5 mediately follows this article in the microfilm version of the Hz, 1 H, Bz), 7.52-7.32 (band, 7 H, Ar), 7.02 (d, J = 9.0 Hz, 1 H, journal, and can be ordered from the ACS: See any current NH), 6.27 (s, 1 H, 10-H), 6.23 (b t, J = 9.0 Hz, 1 H, 13-H), 5.79 (dd, masthead page for ordering information. J = 9.0, 2.5 Hz, 1 H, 3'-H), 5.67 (d, J= 7.0 Hz, 1 H, 2-H), 4.95 (b d, J = 8.0 Hz, 1 H, 5-H), 4.79 (dd, J = 2.5, 5.5 Hz, 1 H, 2'-H), 4.40 (m, JA9421922 634 J. Am. Chem. SOC. 1995,117, 634-644

Total Synthesis of Taxol. 2. Construction of A and C Ring Intermediates and Initial Attempts To Construct the ABC Ring System

K. C. Nicolaou,* J.-J. Liu, Z. Yang, H. Ueno, E. J. Sorensen, C. F. Claiborne, R. K. Guy, C.-K. Hwang, M. Nakada, and P. G. Nantermet Contribution from the Department of Chemistry, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, California 92037, and Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093 Received July 7, 1994@

Abstract: A method for the formation of Taxol's ABC ring system has been developed. General methods for the synthesis of versatile synthons for Taxol's A ring (8) and C ring (55) are presented. A model study using a simplified C ring synthon (17) confiied the viability of the sequential Shapiro-McMurry strategy for formation of Taxol's B ring. Careful exploration of the chemistry of various A-B ring conjugates allowed the development of a successful method for formation of the B ring in a more functionalized system.

Introduction The preceding paper' established a convergent strategy toward Taxol(1, Figure 1) and described a number of chemical studies that provided direction toward the appropriate intermediates and final path. In this article we describe the construction of rings

A and C and discuss the refinements to these methods that were 1:Taxol necessary to arrive at the key building blocks that were utilized in the synthesis. Figure 1. Structure and numbering of Taxol (1). Scheme 1. Construction of Ring A Key Intermediates 8-1W Construction of A Ring COzEt OR In keeping with the themes of convergency and of using the Diels-Alder reaction as a means to construct both rings A and - C of Taxol (l), we embarked on the synthesis of intermediates 9 and 10 as summarized in Scheme 1. The possibility of steric 2 3 hindrance ovemding the well-known electronic induction of regiocontrol in the Diels- Alder reaction2 warranted concem initially. This worry proved unfounded, however, as the readily prepared diene S3s4 and 1-chloroacrylonitrile (6) provided, 10 through a Diels-Alder reaction that proceeded smoothly at 130 \ rOR Ac "C in a sealed tube, an 80% yield of desired product 7 as a single regioisomer, whose structure was confiied by both spectroscopic and X-ray crystallographicanalyses. Application of the protocol of Shine$s6 (KOH, 'BuOH, 70 "C) freed the 9: R = TBS e 7 latent at C1 with concomitant acetate removal 10: R = MEM to give hydroxy ketone 8 (90%, based on 70% conversion). Reagents and conditions: (a) 1.2 equiv of MeMgBr, EtzO, 0 - 25 "C, Reprotection of the primary hydroxyl group of 8 as either a 8 h, then 0.2 equiv of p-TsOH, benzene, 65 "C, 3 h, 70%; (b) 2.2 equiv of i-BuzAIH, CHzClz, -78 - 25 "C, 12 h, 92%; (c) 1.1 equiv of AczO, 1.2 * Address correspondence to this author at the Scripps Research Institute equiv of Et,N, 0.2 equiv of 4-(dimethy1amino)pyridine (DMAP), CHZC12, or the University of California. 0 - 25 "C, 1 h, 96%; (d) 1.0 equiv of 5, 1.5 equiv of 6, 130 "C, 72 h, @Abstract published in Advance ACS Abstracts, December 15, 1994. 80%; (e) 6.0 equiv of KOH, t-BuOH, 70 "C, 4 h, 90% based on 70% (1) Nicolaou, K. C.; Nantennet, P. G.; Ueno, H.; Guy, R. K.; Coula- conversion; (f) for 9, 1.1 equiv of TBSCl, 1.2 equiv of imidazole, CHZC12, douros, E. A,; Sorensen, E. J. J. Am. Chem. SOC. 1995, 117, 624. 10, of (2) Carruthers, W. Cycloaddition Reactions in Organic Synthesis in 25 "C, 2 h, 85%; for 1.2 equiv MEMCl, 1.3 equiv of i-PrzNEt, Tetrahedron Organic Chemistry Series; Baldwin, J. E., FRS, Magnus, P. CHzClz, 25 "C, 3 h, 95%. TBS = Si-t-BuMez, MEM = (methoxy- D., FRS, Eds.; Pergamon Press: New York 1990; Vol. 8, p 1. ethoxy)methyl. (3) Kazi, M. A.; Khan, I. H.; Khan, M. H. J. Chem. SOC. 1964, 1511. (4) Alkonyi, I.; Szabo, D. Chem. Ber. 1967, 2773. tert-butyldimethylsily18 or (methoxyethoxy)methy17 ether af- (5) Shiner, C. S.; Fisher, A. M.; Yacobi, F. Tetrahedron Lett. 1983.24, 5687. forded compounds 9 (85% yield) and 10 (95% yield), respec- (6)See also: Madge, N. C.; Holmes, A. B. J. Chem. SOC., Chem. tively. Commun. 1980, 956. Evans, D. A.; Scott, W. L.; Truesdale, L. K. Tetrahedron Len. 1972, 121. Monti, S. A.; Chen, S. C.; Yang, Y. L.; Yuan, (7) Jacobson, R. M.; Clader, J. W. Synth. Commun. 1979, 9, 57. S. S.; Bourgeois, 0. P. J. Org. Chem. 1978, 43, 4062. (8) Corey, E. J.; Venkateswarlu, A. J. Am. Chem. SOC.1972, 94, 6190. 0002-7863/95/1517-0634$09.00/00 1995 American Chemical Society Total Synthesis of Taxol. 2 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 635

Scheme 2. Chemistry of A Ring Ketones 8-10 and Construction of of a hydrazone, a precursor to vinyllithium species, was then Hydrazones 15 and 16" attempted. To our surprise and delight, hydrazones 15 and 16 were both easily prepared from the corresponding ketones 9 and 10 via addition of (triisopropylsulfonyl)hydrazine.ll As will be discussed below, these hydrazones served admirably in Shapiro co~plings'~J~with appropriately functionalized ring C partners. a 11 A Feasibility Study for the Shapiro-McMurry Strategy With a suitable ring A vinyllithium precursor in hand, we were now ready to test the feasibility for the proposed Shapiro- McMurry strategy toward the taxoid skeleton. To this end the \\ \ model aldehyde 21,14representing Taxol's ring C, was prepared 0 0Tf from diester 1715 via the sequence summarized in Scheme 3. 10 12 Then, reaction of hydrazone 16 with 2.1 equiv of n-BuLi in THF at -78 "C followed by warming to 0 "C and addition of 4 aldehyde 21 furnished a mixture of diastereomeric C2 alcohols (ca.2:l) in 83% total yield. The major diastereoisomer, isolated chromatographically, was proven to be of the desired stereo- chemistry, as indicated in stmcture 22, by X-ray crystallographic KEM\ SnMe, analysis on a subsequent intermediate (vide infra). Vanadium- catalyzed epoxidation of allylic alcohol 22 according to the 14 13 Sharpless procedure16 proceeded regio- and stereoselectively to afford epoxide 23 in 91% yield. Regioselective opening of this epoxide using lithium aluminum hydride" in Et20 at 0-25 "C provided diol 24 in 96% yield. Following our tactical intention to preorganize the substrate prior to McMurry reactions,I8 we engaged the vicinal 1,Zdiol system in 24 as the acetonide 25.19 v Sequential removal of the primary alcohols' protecting groups NNHS0,Ar and oxidationz0 with TPAP-NMO furnished dialdehyde 30 in 9: R = TBS 15: R = TBS 50% overall yield from 25 (Scheme 4). An X-ray crystal- 10: R = MEM 18: R = MEM lographic analysis of compound 30 confirmed its structure and Reagents and conditions: (a) 1.1 equiv of KH,1.05 equiv of PhCHZBr, those of its precursors (see ORTEP drawing, Figure 2). THF, 0 - 25 "C, 1.5 h, 37%; (b) 1.1 equiv of LiN-i-Pr2, DME, -78 "C, Having secured dialdehyde 30 we were within sight of a 2 h, then 1.07 equiv of N-phenyltrifluoromethanesulfonimide,DME, -78 tricyclic taxoid skeleton provided the pending McMurry cou- - 0 "C, 4 h, 80%; (c) 0.90 equiv of MesSnSnMes, 6.35 equiv of LiCl, 0.02 equiv of (PhsP)Pd, THF, 60 "C, 18 h, 90%; (d) 1.0 equiv of PhCOCl, pling'* would be successful. Mindful of Kende's precedentz1 0.05 equiv of PhCHzPd(Cl)(PhsP)z, HMPA, 65 "C, 18 h, 65%; (e) for 15, which resulted in the formation of an olefin at the C9-C10 1.0 equiv of (2,4,6-t1iisopropylbenzenesulfonyl)hydrazine,THF, 25 "C, 24 site instead of the C9-C10 diol system that we desired, we h, 88%; for 16, 1 .O equiv of (2,4,6-t1iisopropylbenzenesulfonyl)hydrazine, proceeded cautiously and systematically to develop proper MeOH 25 "C, 4 h, 85%. Ar = 2,4,6-triisopropylbenzene,TBS = Si-t- BuMe2, MEM = (methoxyethoxy)methyl, HMPA = hexamethylphosphora- conditions for this ring closure. After considerable experimen- mide. tation it was found that exposure of dialdehyde 30 to Ti(0) generated from Tic13 and Zn-Cu couple in DME at 50 "C under Early attempts to engage ketones 9 or 10 in coupling with high-dilution conditions gave the desired diol 31 in 40% yield nucleophiles revealed their reluctance to enter in such reactions, (11)Cusack, N. J.; Reese, C. B.; Risius, A. C.; Roozpeikar, B. probably due to both steric hindrance and ease of enolization. Tetrahedron 1976, 32, 2157. The reaction of 8 with benzyl bromide under basic conditions (12)Shapiro, R. H. Org. React. 1976, 23, 405. Chamberlin, A. R.; evidenced the latter by giving rise to dibenzyl derivative 11 Bloom, S. H. Org. React. 1990, 39, 1-83. Martin, S. F.; Daniel, D.; Cherney, R. J.; Liras, S. J. Org. Chem. 1992, 57, 2523. (37%, Scheme 2), rather than the expected benzyl ether. (13)This strategy was later used by others to accomplish similar Having failed to induce the ring A derivatives 8-10 to couplings: Di Grandi, M. J.; Jung, D. K.; Krol, W. J.; Danishefsky, S. J. undergo nucleophilic additions at their carbonyl site, it was then J. Org. Chem. 1993, 58,4989. Masters, J. J.; Jung, D. K.; Bornmann, W. G.; Danishefsky, S. J. Tetrahedron Lett. 1993, 34, 7253. decided to umpolung the system, that is to convert it into a (14)Nicolaou, K.C.; Yang, Z.; Sorensen, E.; Nakada, M. J. Chem. Soc., nucleophilic species. Early attempts utilized the vinyltin Chem. Commun. 1993, 1024. derivative 13 (Scheme 2), prepared from ketone 10 via triflate (15)Mundy, B. P.; Theodore, J. J. J. Am. Chem. SOC. 1980, 102, 2005. (16)Sharpless, K. B.;Michaelson, R. C. J. Am. Chem. SOC.1973, 95, U9as summarized in Scheme 2, as a vinyllithium precursor or 6136. Sharpless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12,63. a nucleophillic partner in a Stille couplinglo reaction. However, Rao, A. S. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., neither reaction proved fruitful with a functionalized ring C Ley, S. V., FRS, Eds.; Pergamon Press: New York, 1991; Vol. 7, p 376. partner, even though a Stille coupling of 13 with benzoyl (17) Mwai, S.;Mwai, T.; Kato, S. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: New York, 1991;Vol. 8, chloride did afford enone 14 (65% yield, Scheme 2). With some p 871. reluctance due to the expected steric hindrance, the formation (18)McMurry, J. E. Chem. Rev. 1989, 89, 1513. McMurry, J. E. Acc. Chem. Res. 1983, 16, 405. McMurry, J. E.; Lectka, T.; Rico, J. G. J. (9)McMurry, J. E.; Scott, W. I. Tetrahedron Len. 1983,24,979. Wulff, Org. Chem. 1989,54, 3748. McMuny, J. E.; Rico, J. G. Tetrahedron Lett. W. D.; Peterson, G. A.; Bauta, W. E.; Chan, K.-S.; Faron, K. L.; Gilbertson, 1989, 30, 1169; Lenoir, D. Synthesis 1989, 883. S. R.; Kaesler, R. W.; Yang, D. C.; Murray, C. K. J. Org. Chem. 1986,51, (19)Evans, M. E.; Parrish, F. W.; Long, L., Jr. Carbohydr. Res. 1967, 277. 3, 453. Lipshutz, B. H.; Barton, J. C. J. Org. Chem. 1988, 53, 4495. (10) (a) Milstein, D.; Stille, J. K. J. Am. Chem. SOC.1978, 3636. (b) (20)Griffith, W. P.; Ley, S. V. Aldrichimica Acta 1990, 23, 13. Milstein, D.; Stille, J. K. J. Org. Chem. 1979, 44, 1613. (c) For a review, (21) Kende, A. S.;Johnson, S.; Sanfilippo, P.; Hodges, J. C.; Jungheim, see: Stille, J. K. Angew. Chem., Znt. Ed. Engl. 1986, 25, 508. L. N. J. Am. Chem. SOC. 1986, 108, 3513. 636 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al.

Scheme 3. Synthesis of the Acetonide Model System 25 by the Shapiro Scheme 4. Synthesis of the ABC Taxoid Systems 33 and 35 by the Reaction' McMuny Reaction" HO HO '". .. EtO,Cn -a 'p b EtOzC HO -PHO 17 18 10

30 OMEM BnO 1

c-

HO

21 20

el

22 23 g1

33 35 Reagents and conditions: (a) H2,20% Pd(OH)2 on C, EtOH, 25 OC, 2 h, 100%; (b) 1.2 equiv of Ac20, 1.3 equiv of 4-(dimethylamino)pyridine (DMAP), CHzClz, 0 - 25 "C, 2.5 h, 97%; (c) 1.0 equiv of TiCL, CHzC12, -78 OC, 10 min, then -20 "C, 10 min, 65%; (d) 0.1 equiv of KzC03, 25 24 MeOH, 25 "C, 4 h, 91%; (e) 0.05 equiv of tetrapropylammonium permthenate (TPAP),3.0 equiv of 4-methylmorpholine N-oxide (NMO), (I Reagents and conditions: (a) 5.0 equiv of i-BuzAlH, CHKlz, -78 - 25 "C, 10 h, 95%; (b) Hz, 0.2 equiv of PdC, EtOAc, 3 h, 100%; (c) 1.0 4-A sieves, CH2C12,25 OC, 10 min, 87%; (f) 8.0 equiv of TiCls-(DME)1.5, equiv of KH, 1.0 equiv of PhCHZBr, THF, 0 - 25 "C, 1.5 h, 85%; (d) 2.0 15 equiv of Zn-Cu, DME, 50 "C, 5 h, 40% of 31,25% of 32; (g) excess equiv of pyridinium dichromate (PDC), molecular sieves, CHzC12,O - 25 MnOz, CH2C12, 25 "C, 20 min, 90%; (h) excess AczO, excess pyridine, "C, 4 h, 90%; (e) 16, 2.1 equiv of n-BuLi, THF, -78 OC, 0.5 h, then 0 "C, CHzClz, 25 "C, 0.5 h, 98%; (i) 30 equiv of pyridinium chlorochromate 10 min, 1.3 equiv of 21, THF, 0 - 25 "C, 5 h, 83% (cu. 2:l diastereomeric (PCC), 30 equiv of NaOAc, Celite, benzene reflux, 2 h, 71%. MEM = mixture); (f) 1.1 equiv of r-BuOOH, 0.014 equiv of VO(acac)z, PhH, 25 (methoxyeth0xy)ethyl. OC, 2 h, 91%; (g) 2.0 equiv of LiAla, EtzO, 0 "C, 20 min, 25 "C, 6 h, 96%; (h) 2 equiv of 2,2-dimethoxypropane, 0.2 equiv of camphorsulfonic and high degree of oxygenation, presented a more serious acid (CSA), CHzC12, 25 OC, 12 h, 85%. MEM = (methoxyethoxy)ethyl, challenge to the Diels-Alder approach. Early approaches Bn = CHzPh, acac = acetylacetonate. examined the reaction of dienophile 40 (prepared from l-hy- as a mixture of two diastereoisomers (stereochemistry unas- droxy-2- (36) according to Scheme 5) and 3-car- signed). This reaction produced no A C9-C10 olefin, although bomethoxy-2-pyrone (43) (Scheme 6). According to previous the C9-Cl2 coupled byproduct 32 was formed (25% yield) as work by CoreyZ4and BrysonZ5with the latter compound, and also observed by Kende.21 The mechanistic aspects of this considering the substitution pattem of dienophile 40, we reaction will be discussed in a subsequent paper in this series. expected this reaction to proceed regio- and stereoselectively Oxidation of the mixture of 31 with Mn0222 gave the to afford product 45 via intermediate 44. Diene 45 was then dienediol 33 in 90% yield, and acetylation of 31 followed by expected to serve as a precursor to a fully functionalized ring PCC oxidation23 led to enone 35 via diacetate 34. The work C for coupling with ring A. In the event, however, this Diels- presented in Schemes 3 and 4 demonstrated the viability of our Alder reaction (155 "C, 24 h, 81% yield based on 51% Shapiro-McMuny strategy toward Taxol (1) and placed us in conversion) proceeded with the opposite regiochemistry from the position of facing the challenge of Taxol (1) itself. that expected, furnishing product 47, via presumed intermediate 46, the latter undergoing a facile decarboxylation under the Construction of C Ring Systems reaction conditions. A series of regio- and stereochemically controlled reactions, shown in Scheme 6, converted cyclo- a. The Diels-Alder Reaction. In contrast to the achiral as hexadiene system 47 into crystalline diol 51. X-ray crystal- ring A system, ring C of Taxol, with its numerous stereocenters lographic analysis of 51 (see ORTEP drawing, Figure 2) (22) Fatiadi, A. J. Synthesis 1976, 65. (23) Parish, E. J.; Wei, T. Y. Synrh. Commun. 1987,17,1227. Rathore, (24) Corey, D. I.; Watt, D. S. J. Am. Chem. SOC. 1973, 95, 2303. R.; Saxena, N.; Chadrasekaran Synth. Commun. 1986, 16, 1493. (25) Bryson, T. A.; Donelson, D. M. J. Org. Chem. 1977, 42, 2930. Total Synthesis of Taxol. 2 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 637

Scheme 5. Synthesis of Dienophiles 40-42"

Q:CI 7

37 40 R = TPS 41:R=THP 42: R=H zd a Reagents and conditions: (a) 1.1 equiv of TPSC1, 1.15 equiv of imidazole, DMF, 0 "C, 1 h, then 03, CH2C12, -78 "C, 1 h, then 2.2 equiv of Ph3P. -78 "C - 25 "C; (b) 2.1 equiv of dihydropyran, 0.005 equiv of 30 p-TsOH, CH2C12, 25 "C, 0.5 h, then 03, CH2C12, -78 T, 5 h, then 1.0 equiv of Ph3P, -78 "C - 25 "C, 98%; (c) for 40, 1.4 equiv of Ph3P=C(CH3)C02Et, CH2C12,25 "C, 20 h, 91% from 36; for 41, 1.03 equiv of Ph3P=CHC02Et, CH2C12, 0 "C, 4 h, then 25 "C, 18 h, 90%; (d) 0.05 equiv of p-TsOH, MeOH, 25 "C, 18 h, 92%. TPS = Si-t-BuPh2, THP = tetrahydropyranyl. HO OH Scheme 6. Early Diels-Alder Attemptsa 0

51 OTPS C02Me 40 43 46

-IC021 HO OH 1 Y TPSO

'"OTBS TPSO TPSO C02Me C02Me 60 44 45 47 bl RO OH TPSO OH TPSO

I""' HO E102C\""Po C02Me p,pHO A0 50: R = TPS 49 4a 'CSI:R=H Reagents and conditions: (a) 1.0 equiv of 40, 2.0 equiv of 43, neat, a7 155 OC, 24 h, 81% based on 51% conversion; (b) 4.0 equiv of m-CPBA, CH2C12, 25 'C, 4 h, 71% plus 19% of a epoxide; (c) excess i-Bu?AlH, EtzO, 0 "C, 2 h, 91%; (d) excess 2,2-dimethoxypropane, 0.05 equiv of camphorsulphonic acid (CSA), CH2C12, 25 "C, 1 h, 90%; (e) 1.0 equiv of n-BWNF (TBAF), THF, 25 "C, 1 h, 95%. TPS = Si-t-BuPh2.

confirmed its structure and those of its precursors and revealed the undesired regioselectivity of the Diels- Alder reaction. Faced with this unfortunate regiochemical outcome, we then focused our attention on 3-hydroxy-2-pyrone (52, Scheme 7) as a diene in the Diels-Alder reaction. Although Corey26has demonstrated that this system would give the opposite regio- chemical pathway from that required for our purposes, the pioneering work of Narasaka*' afforded us the possibility for success in this endeavor. Scheme 7 demonstrates Narasaka's (26) Corey, E. J.; Kozikowski, A. P. Tetrahedron Len. 1975, 2389. (27) Narasaka, K.; Shimada, S.; Osoka, K.; Iwasawa, N. Synthesis 1991, Figure 2. ORTEP drawings for intermediates 7, 30, 51, 60, 87, 103. 1171. 638 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al.

Scheme 7. Synthesis of Common C-Ring Intermediate 55 Scheme 8. Structural Confirmation of 55"

52 +Et 0 EtOzC OH - OH a

42 I 53 55: R = H 57 %:R=Ac 'I HO OH

e 55 54 "Reagents and conditions: 1.4 equiv of 52, 1.4 equiv of PhB(OH)z, PhH, reflux (Dean-Stark trap), 48 h, then 1.4 equiv of 2,2-dimethyl-1,3- OH '0 propanediol, 25 "C, 1 h, 79% based on 77% conversion of 42. 50: R E C02Et 60 L 59: R = CH2OH principle of temporarily tethering the two reaction partners in "Reagents and conditions: (a) 5.0 equiv of AczO, 2.5 equiv of order to dictate the regiochemistry of the Diels-Alder reaction. 4-(dimethylamino)pyridine (DMAP), CHZC12, 25 "C, 10 min, 100%; (b) Thus, reaction of dienophile 42 (prepared from 1,3-dihydroxy- 1.2 equiv of pyridinium dichromate (PDC), 4-A molecular sieves, CHZClz, (37), (52) 25 OC, 1 h, 81%; (c) 4.0 equiv of t-BuMezSiOTf, 4.0 equiv of 2,6-lutidine, cis-2- butene Scheme 6) with 2-hydroxy-2-pyrone in 0.1 equiv of DMAP, CHzClz, 0 "C, 4 h, 92%; (d) 1.1 equiv of LiAW, the presence of phenylboronic acid under dehydrating conditions Etz0,O - 25 "C, 0.5 h, 97%; (e) 0.05 equiv of camphorsulfonic acid (CSA), led, after decomplexation with excess 2,2-dimethyl-1,3-pro- CHzClz, MeOH, 25 "C, 1 h, 94%. TBS = Si-t-BuMez, Tf = S02CFs. panediol, to compound 55. Evidently, the initially formed Diels-Alder product 54 promptly rearranges under the reaction hydrazone. To this end, diol 55 (Scheme 9) was dibenzylated conditions via intramolecular acyl transfer from the secondary using excess KH and benzyl bromide31 to afford compound 61 to the primary hydroxyl group to afford the observed product which was then reduced with excess LAH in ether at 0 "C to in 79% yield based on 77% conversion of 42.28 Relief of strain give hydroxy lactol 62 as a 1:l mixture of diastereoisomers in going from the [2.2.2] cycloaddition product 54 to the [3.4.0] (71% yield from 55). Selective monoprotection of the primary bicyclic system 55 may be the primary reason for this facile alcohol using fert-butyldiphenylsilyl chloride (TPSCl) and rearrangement. imidazole in DMF3zfollowed by further reduction of the lactol with LAH in THF at 25 "C furnished diol in 78% yield. Scheme 8 demonstrates a number of useful transformations 64 Reaction of with pivaloyl chloride (1.05 equiv) under basic of compound 55 that led not only to confiiation of its structure 64 conditions led to a 1:3.2 mixture of the two pivaloate esters 65 but also to more advanced intermediates as required for our and 66 which were chromatographically separated. plans. Exhaustive acetylation of 55 led to diacetate 56 which The next task was the introduction of an alcohol at C5. Even exhibited significant downfield shifts in its proton NMR though a previous had shown that the primary hydroxyl spectrum (CDC13, 6 Ha, 4.59 -+ 5.84 and Hb, 3.10 3.90). - group in a similar system could be used to direct the hydrobo- Pyridinium dichromate (PDC) oxidationz9of 55 furnished enone ration of the cyclohexene double bond, the feasibility of using 57 in accord with the assigned structure (55), whereas persily- a mesylate (SOzCH3) as a possible directing group in lation of the same compound with TBSOTf3O gave the bis(sily1 this hydr~boration~~was explored. Such a method would more ether) 58 isolated as a C20 hydrate. The latter compound efficiently lead to the targeted oxetane system. Indeed hy- underwent selective reduction with LAH in Et20 at 0 25 "C - droboration of 68, prepared from 65 by standard mesylation, to afford primary alcohol 59 (97% yield) which was monode- with borane in THF (0-25 "C) followed by oxidative workup, silylated with camphorsulfonic acid (CSA) in Me0H:CHzClz led to the formation of the C5 alcohol 69 as the major product to afford the crystalline lactone diol 60 in 94% yield. X-ray and in 53% yield. Treatment of the latter compound with NaH crystallographic analysis of 60 (see ORTEP drawing, Figure 2) in THF at 45 "C resulted in the formation of oxetane 70 in confirmed its structure and those of its progenitors. An NMR 86% yield, confirming the stereochemical orientation of the experiment (500 MHz, CDCl3) confiied that neither acid newly generated alcohol in 69. Attempts to reach the targeted (CSA) nor base (DMAP) causes any skeletal rearrangement of C2 aldehyde were, however, thwarted by failure to cleanly 55, serving as a control for the reactions summarized in Scheme remove the pivaloate group from 70, presumably due to 8. interference from the oxetane ring under the reductive or basic b. The First Attempt at a CD Ring System. Oxetane Is conditions employed in these attempts. Nevertheless, this Formed but Interferes with Subsequent Chemistry. One of sequence confirmed the potential feasibility of constructing the our early plans was to construct a CD ring system with the oxetane ring by this method and rendered the aldehyde 67 oxetane ring already in place before coupling with a ring A (31) Evans, M. E.; Parrish, F. W.; Long, L., Jr. Carbohydr. Res. 1967, (28) For obvious practical reasons, large-scale reactions are performed 3, 453. Lipshutz, B. H.; Barton, J. C. J. Org. Chem. 1988, 53, 4495. with 1.0 equiv of diene and dienophile each and 0.95 equiv of PhB(0H)z; (32) Hanessian, S.; Lavallb, P. Can.J. Chem. 1975,53,2975. Hanessian, diol 55 is typically obtained in ca. 60% yield based on ca. 50% conversion. S.; Lavalike, P. Can. J. Chem. 1977, 55, 562. The crude starting material mixture is recycled in the same process. (33) Nicolaou, K. C.; Liu, J.-J.; Hwang, C.-K.; Dai, W.-M.; Guy, R. K. (29) Corey, E. J.; Boger, D. L. Tetrahedron Lett. 1978, 2461. J. Chem. SOC., Chem. Commun. 1992, 1118. (30) Corey, E. J.; Cho, H.; Rucker, C.; Hua, D. H. Tetrahedron Lett. (34) Smith, K.; Pelter, A. In Comprehensive Organic Synthesis; Trost, 1981, 22, 3455. B. M., Fleming, I., Eds.; Pergamon Press: New York, 1991; Vol. 8, p 703. Total Synthesis of Taxol. 2 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 639

Scheme 9. Synthesis of Oxetane-Containing C-ring 70" Scheme 10. Synthesis of ACD Ring System 81" 0 OR RO OBn TPSO QBn

TBSO bR

55: R= H 62:R=H 64 71:R=H 73:R=H a 661: R = Bn cC6yR=TPS "1 55 c72:R = MOM 674: R = CHPh

TPSO OBn Ph +

OH

67 66 65 RO MOM OR gl 79 R = TBS 75: R= H t R=H ' K76: R = Ts Tps'Qo TPSO@' OBn h TPSO'@ OBn i - - I-Pr OH 15 "bBn 'bBn Opiv OBn OPiv OPiv OMS Ph Ph

70 69 68 Reagents and conditions: (a) 3.5 equiv of PhCHzBr, 3.5 equiv of KH, 0.05 equiv of n-BN0 "C - 25 "C, 2 h, 75%; (b) 2.0 equiv of LiAlH4, Et20, 0 "C, 1 h, 94%; (c) 1.4 equiv of TPSC1, 1.5 equiv of imidazole, DMF, 0 "C, 2 h, 25 "C, 4 h, then excess n-BWF, THF, 10 h, 82% based on 54% conversion; (d) 1.3 equiv of LiAIh, THF, 0 "C - 25 "C, 0.5 h, 96%; (e) 1.05 equiv of PivC1, 1.5 equiv of 4-(dimethylamino)pyndine 80 81 (DMAP), CH2C12, 0 "C, 0.5 h, 55% of 66, plus 17% of 65, plus 24% of Reagents and conditions: (a) 2.0 equiv of KH, 1.2 equiv of TBSC1, C2-C20 dipivalate, based on 84% conversion; (0 0.05 equiv of tetrapro- THF, 25 OC, 0.5 h, 61%; (b) 2.0 equiv of MOMC1, 1.5 equiv of KH, CH2C12, pylammonium permthenate (TPAP), 1.5 equiv of 4-methylmorpholine 25 "C, 12 h, 92%; (c) 5.0equiv of LiAlb, THF, 25 "C, 1 h; (d) 3.8 equiv N-oxide (NMO), CH3CN, 25 'C, 1.5 h, 91%; (g) 1.5 equiv of MsCl, 2.0 of PhCH(OMe)Z, 0.05 equiv of camphorsulfonic acid (CSA), CH2C12, 25 equiv of DMAP, CHZClz, 0 "C - 25 "C, 1.5 h, 95%; (h) 10 equiv of "C, 72% from 72; (e) 3.0 equiv of BHyTHF, THF, 25 "C, 10 h, then excess BHqTHF, THF, 25 "C, 10 h, then excess H202, aqueous NaHC03, 53%; Hz02, aqueous NaHCO3, 37%; (f) 1.6 equiv of TsCl, 3.0 equiv of (i) 5.0 equiv of NaH, THF, 45 "C, 3 h, 86%. Bn = CHzPh, TPS = Si-t- 4-(dimethy1amino)pyridine(DMAP), CHzC12, 25 "C, 5 h; (g) 2.2 equiv of BuPh2, Piv = CO-t-Bu, Ms = S02CH3. NaH, THF, 45 "C, 10 h, 78% from 75; (h) excess n-BuflF, THF, 25 "C, 2 h, 95%; (i) 3.0 equiv of Dess-Martin periodinane, CH2Cl2, 25 "C, 2 h, 91%; Q) 1.2 equiv of 15, 2.4 equiv of n-BuLi, THF, -78 'C, 0.5 h, 80%. available through oxidation of 66 using the TPAP-NMO then 0 "C; 1.0 equiv of 79, THF, 0 "C, 0.5 h, 85% (ca. 5:3 mixture); (k) method.*O The latter compound was utilized in a subsequent excess t-BuOOH, 0.05 equiv of VO(acac)z, PhH, 25 "C, 2 h. MOM = attempt to construct the ABC ring skeleton of Taxol (1) as will methoxymethyl, TBS = Si-t-BuMez, Ts = S02-p-Tol. acac = acetylac- be discussed in a later section of this paper. etonate. c. A Second Attempt at the CD Ring System. Success 74 resulted in the formation of the C5 ,8-hydroxy compound but the Oxetane Ring Interferes Again after the Shapiro 75 in 37% yield. Tosylation (80%) of the latter followed by Coupling. After our first attempt to construct a suitable CD exposure to NaH in THF at 45 "C led to oxetane 77 (78%) via aldehyde failed, we quickly redesigned our approach, choosing tosylate 76. Finally, desilylation of 77 using TBAF,37followed new protecting groups and targeting aldehyde 79 as a potential by Dess-Martin ~xidation,~~furnished aldehyde 79 via 78 in electrophile for the Shapiro coupling. Scheme 10 outlines the 86% overall yield. The Shapiro reaction proceeded well in chemistry involved in this second approach. Thus, upon combining hydrazone 15 and aldehyde 79 to produce alcohol treatment with KH and TBSCl, intermediate 55 underwent 80 (85%, mixture of diastereoisomers, Scheme 10). Epoxide skeletal rearrangement involving acyl migration from the 81 could not, however, be cleanly obtained from the major primary to the secondary hydroxyl group, presumably driven isomer 80 using the vanadium-catalyzed procedure. by trapping of the primary hydroxyl as a silyl ether, to afford Due to the problems encountered in the two approaches 71. Protection of the tertiary alcohol as a methoxymethyl discussed above, the strategy of having the oxetane installed in (MOM) ether7 led to 72. Reduction of the ester and lactone the molecule prior to the coupling reactions was abandoned in functionalities in 72 using excess LAH in THF formed triol favor of schemes involving oxetane construction at a later stage. 73. Introduction of the benzylidene group35protected the C7- d. Successful Progression to the McMurry Cyclization C9 diol system in the latter compound, furnishing 74 in 72% Stage. Having just experienced the complications of the highly yield from 72. The possibility of generating a C7 benzyl, C9 (36)Takano, S.;Akiyama, M.; Sato, S.;Ogasawara, K. Chem. Lett. 1983, hydroxy derivative directly from the ben~ylidene~~dictated the 1593. Hatakeyama, S.;Sakurai, K.; Saijo, K.; Takano, S. Tetrahedron Lett. choice of this protecting group. Directed hydroboration of olefin 1985, 26, 1333. Schreiber, S. L.; Wang, Z.; Schulte, G. Tetrahedron Lett. 1988, 29, 4085. Adam, G.;Seebach, D. Synthesis 1988, 373. (35)Albert, R.; Dax, K.; Pleschko, R.; Stutz, K. Carbohydr. Res. 1985, (37)Corey, E. J.; Venkateswarlu, A. J. Am. Chem. SOC.1972, 94, 6190. 137, 282. Yamanoi, T.; Akiyama, E.; Inazu, T. Chem. Lett. 1989, 335. (38)Dess, D. B.; Martin, J. C. J. Org. Chem. 1983,48,4155.Dess, D. Crimmins, M. T.; Hollis, W. G., Jr.; Lever, J. G. Tefrahedron Left. 1987, B.; Martin, J. C. J. Am. Chem. SOC. 1991, 113, 7277. Ireland, R. E.; Liu, 28, 3647. L. J. Org. Chem. 1993, 58, 2899. 640 J. Am. Chem. Soc., Vol. I 17, No. 2, 1995 Nicolaou et al.

Scheme 11. Synthesis of A-C Ring System 87" (Re face) HO oen Nu' 1%

QB" a OH

62 82: X= 0 683:X = (OMe)*

88: Li+ chelate derived from aldehyde 86

86

TBSO

Figure 3. Stereoselectivity of the Shapiro reaction. The model was generated with Chem3d, most hydrogens are omitted for clarity.

i-Pr. Early attempts to unblock the aldehyde group of 87 under 15, Ar =*i-pr 07 acidic conditions failed due to formation of a cyclic hemiacetal i-Pr with the C2 hydroxyl group. Epoxidation of the allylic system a Reagents and conditions: (a) 3.0 equiv of Dess-Martin periodinane, in 87 proved rather slow and, therefore, the C20 hydroxyl group CH2C12, 0 "C - 25 "C, 12 h; (b) excess of HC(OMe)3, 0.05 equiv of camphorsulfonic acid (CSA), MeOH, CH2C12, 25 "C, 12 h, 81% from 62; was called upon to assist in this reaction. Treatment of 87 with (c) 1.2 equiv of LiAl&, THF, reflux, 1 h; (d) 1.5 equiv of PivC1,5.0 equiv LAH in ether resulted in the formation of diol 89 (88% yield) of 4-(dimethylamino)pyridine (DMAP), CH2C12, 0 "C, 15 min, 70% from which underwent smooth epoxidation with tBuOOH in the 83; (e) 1.7 equiv of Dess-Martin periodinane, CH2C12, 25 "C, 1.5 h, 83%; presence of VO(acac)2 catalyst to afford epoxide 90 in 82% (f) 15, 2.2 equiv of n-BuLi, THF, -78 "C, 0.5 h, then 0 "C; 1.2 equiv of 86, THF, -40 "C, 5 min, 74%. Bn = CH2Ph, Piv = CO-t-Bu, TBS = yield (Scheme 12). Si-t-BuMe2. At this point our plan involved engaging the two hydroxyl oxygenated intermediates of the previous schemes, we decided groups of our latest intermediate (90) in a cyclic system in order to minimize such problems by targeting aldehyde 86 (Scheme to both prevent the undesired hemiacetal formation and to 11). Oxidation of intermediate 62, readily available as described preorganize the substrate prior to the construction of ring B. in Scheme 10, with De~s-Martin~~reagent afforded aldehyde To this end, diol 90 was treated with phosgene in the presence lactone 82. Protection of the aldehyde as a methoxy acetal39 of pyridine41 in an attempt to produce the 7-membered ring produced compound 83 (81% yield from 62) which was then carbonate. These conditions, however, produced exclusively reduced with LAH in THF at reflux to give diol 84. Treatment the tetrahydrofuran derivative 91, presumably via nucleophilic of the latter compound with pivaloyl chloride in the presence attack by the C2 hydroxyl group on the activated C20 chloro- of DMAPOselectively protected the C20 alcohol as a pivaloate formyl intermediate. To circumvent this problem, both alcohols ester, leading to intermediate 85 in 70% yield from 83. were engaged in a cyclic lactone by exposure of diol 90 to Molecular models revealed the C2 hydroxyl group of 84 to be Dess-Martin reagent,38giving the y-lactone 92 in 61% yield. more crowded [interference from bis(methoxy) group] than the Removal of the silyl group from compound 92, followed by C20 hydroxyl group (pseudo axial position) and thus the oxidation with Dess-Martin reagent,38 afforded aldehyde 94, selectivity observed. Finally, oxidation with either TPAP- via intermediate alcohol 93, in 71% overall yield. Revealing NM020 or Dess -Martin reagent38 easily converted compound the C9 aldehyde by exposure to trifluoroacetic acid42(TFA) at 85 to aldehyde 86 (83% yield). 0 OC, produced, in addition to dialdehyde 95 (51% yield), the With the aldehyde 86 in hand, we then proceeded to the 96 (24%) (Scheme 13), presumably arising Shapiro reaction utilizing hydrazone 15 as the precursor to the from 95 via acid-induced epoxide opening. vinyllithium reagent. This coupling reaction furnished alcohol Several attempts to cyclize dialdehyde 95 using the McMurry 87 as a single diastereoisomer in 74% yield. X-ray crystal- reaction under a variety of conditions were unsuccessful. The lographic analysis allowed the assignment of the stereochemistry only detectable product was the diol 97, apparently produced of this intermediate (see ORTEP drawing, Figure 2). The by reduction of both aldehyde groups. It became clear that this stereoselectivity of this reaction can be explained by invoking particular design did not favor the required ring closure and 6-membered ring chelate intermediate 88, as shown in Figure that we had to design yet another synthetic sequence. 3. In this model, the re face of the aldehyde is more accessible e. First Attempt with the C1-CZCarbonate Approach. to nucleophilic attack than the si face due to shielding by the Aiming to enforce a different conformation in the McMuny C8 methyl and C20 pivaloyl groups. substrate, we decided to introduce a C1 -C2-carbonate ring. (39) Wenkert, E.; Goodwin, T. E. Synth. Commun. 1977, 7, 409. (41) Haworth, W. N.; Porter, C. R. J. Chem. SOC. 1930, 151. (40) Hofle, G.; Steglich, W.; Vorbriiggen, H. Angew. Chem., Znt. Ed. (42) Ellison, R. A.; Lukenbach, E. R.; Chiu, C.-W. Tetrahedron Lett. Engl. 1978, 17, 569. 1975, 499. Total Synthesis of Taxol. 2 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 641

Scheme 12. Synthesis of Lactone 93a Scheme 13. First Attempt at the McMurry Cyclization" TBSO TBSO

-b

87 R= PN aL:R= ti 94 / O5

01 02 97 96

a Reagents and conditions: (a) trifluoroacetic acid neat, 0 "C, 15 min, 51%; (b) 10 equiv of TiC13*(DME)1,5,15 equiv of Zn-Cu, DME, 60 "C, 4 h, 95 added over 5 h (syringe pump), then 55 "C, 3 h, 34% based on 43% conversion. Bn = CHZPh. l probably responsible for this relative unreactivity. The carbon- ate 104 was then desilylated with fluoride ion and oxidized with TPAP-NMOZo to afford dialdehyde 106 via the corresponding diol (105) in 80% overall yield. 04 03 With the requisite dialdehyde 106 in hand, we proceeded to Reagents and conditions: (a) 2.0 equiv of LiAlb, EtzO, -10 "C, 5 investigate its-conversion to a cyclic taxoid system through min, 88%; (b) 2.0 equiv of r-BuOOH, 0.25 equiv of VO(acac)z, PhH, 25 McMurry coupling. In traversing the temperature range from "C, 0.5 h, 82%; (c) 5.0 equiv of phosgene, pyridine, 75 "C, 2.5 h, 35%; (d) 0 to 70 "C, no cyclic coupling products were observed; at 85 10 equiv of Dess-Martin periodinane, CHzC12,50 "C, 1 h, 61%; (e) excess "C, however, a 15% yield of the cyclic olefin 107 (Scheme 14) n-B-, THF, 25 "C,2 h; (f) 5.0 equiv of Dess-Martin periodinane, CHzClz, 25 "C, 0.5 h, 71% from 92. TBS = Si-t-BuMez, Bn = CHZPh, was isolated, suggesting that the desired cyclic diol might remain Piv = CO-t-Bu, acac = acetylacetonate. elusive even with these rigid precursors. The conclusion was that further preorganization was needed in order to lower the Molecular modeling (S ybyl) indicated that this functionality activation energy to avoid deoxygenation of carbons 9 and 10 would preorganize the expected intermediate geometry by during the McMurry cyclization. bringing the two to the same face of the molecule. Learning from our previous experience with protecting groups, Conclusion we decided to utilize aldehyde 67 (prepared as described in Scheme 9, above) and hydrazone 15 as partners for the Shapiro In this paper we described the evolution of the chemistry that reaction. Thus (as shown in Scheme 14) the Shapiro reaction eventually led to a successful construction of a taxoid system produced compound 98, as a single isomer (stereochemistry containing the ABC ring framework of Taxol (1). While the confirmed by X-ray crystallographic analysis of subsequent construction of a suitable ring A fragment proceeded smoothly intermediate 103) in 82% yield. Deprotection with LAH via a Diels-Alder approach, that of a suitable ring C fragment afforded diol 99 (87% yield). Vanadium-catalyzed epoxidation presented more difficulties. Although the highly functionalized of 99 with 'BuOOH led stereoselectively to epoxide 100 in 95% and stereochemically defined ring C intermediate was easily yield. Regioselective epoxide opening with LAH gave triol 101 produced via a boron template controlled reaction, the finetuning in 78% yield based on 81% conversion. Selective protection of the functional groups for proper elaboration required con- of the primary alcohol in 101 as a MOM ether proceeded siderable experimentation. Through the process of design, smoothly under standard conditions to afford compound 102 experimentation, and redesign, however, enough knowledge was in almost quantitative yield. Diacetate 103 was prepared using gathered that made the final push toward a suitable ABC taxoid acetic anhydride and DMAP (83% yield). X-ray crystal- ring system possible. This final and successful approach is lographic analysis of the latter compound confirmed the discussed in the following paper. previously proposed stereochemistry (see ORTEP drawing, Experimental Section Figure 2). By this time both our model studies and degradation work' General Techniques. For a description of general technique, see pointed to a carbonate protecting group at Cl-C2 as the most the Fist paper in this series.' Experimental techniques and data for suitable device for our synthetic scheme. In order to install compounds 10-14, 16, 18-35, 47-51, 57, 58, 61-80, 82-87, and the latter into our intermediate (102), it was necessary to use 89-107 can be found in the supplementary material. Diene 3. A solution of ketone 2 (245.0 g, 1.44 mol) in Et20 (1500 rather strong conditions (excess KH, phosgene, ether:HMPA, mL) at 0 "C was treated with methylmagnesium bromide (576 mL of 1:1, 25 "C, 88% yield based on 57% conversion) as compared a 3.0 M solution in EtzO, 1.73 mol). The reaction mixture was allowed to those used in making the taxoid carbonate II.' The flexibility to warm to 25 OC and stirred for 8 h. After cooling to 0 "C,the reaction of the 1,2-diol 102 as compared to the rigidity of the cor- was quenched with aqueous NI&C1(600 mL). The organic layer was responding taxoid diol employed in the degradation studies is separated and washed with H20 (2 x 400 mL) and brine (400 mL). 642 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al.

Scheme 14. Formation of the ABC taxoid system 107 by a McMuny (40-45 "C, 0.05 "Hg) to give 3 (169 g, 70%) as a colorless liquid cyclizationa Rf = 0.35 (silica, 2% Et20 in petroleum ether); 'H NMR (300 MHz, TBSO CDC13) 6 5.05 (d, J= 1.0 Hz, 1 H, HC=C), 4.74 (d, J = 1.0 Hz, 1 H, HC-), 4.14 (q, J = 7.0 Hz, 2 H, C02CH2CH3), 1.97 (s, 3 H, CH~C-CHZ),1.80 (s, 3 H, CH~CSC), 1.78 (s, 3 H, CH3C%), 1.24 (t, J = 7.0 Hz, 2 H, C02CHzCH3). Alcohol 4. A solution of ester 3 (169 g, 1.01 mol) in CH2C12 (1000 mL) at -78 "C was treated with diisobutylaluminum hydride (2220 mL of a 1.0 M solution in CHzC12, 2.22 mol) and stirred at -78 "C for 0.5 h. The reaction mixture was allowed to warm to 25 "C and stirred 98: R = Piv for 12 h. The reaction mixture was slowly poured into a mixture of bCe0: R=H ice (600 mL) and glacial acetic acid (300 mL), and the resulting mixture I was stirred for 3.5 h. The aqueous layer was separated and extracted with CHzClz (2 x 500 mL). The combined organic layer was washed TBSO TBSO "1 with brine (2 x 500 mL), dried (MgSOd), concentrated (bath tempera- ture <25 "C), and purified by flash chromatography (silica, 20% Et20 in petroleum ether) to give 4 (117.4 g, 92%) as a pale yellow oil: Rf = 0.26 (silica, 20% Et20 in petroleum ether); 'H NMR (300 MHz, CDCl3) 6 5.08 (b d, J = 1.0 Hz, 2 H, C-CHz), 4.71 (b d, J = 1.0 Hz, 2 H, C=CHz), 4.16 (s, 2 H, CHzOH), 1.82 (t, J = 1.0 Hz, 3 H, (CH3)C+Hz), 1.76 (s, 3 H, Ce(CH3)z), 1.71 (s, 3 H, C=(CH3)2). Acetate 5. A solution of alcohol 4 (113.6 g, 0.9 mol) in CHzClz 101 (lo00 mL) at 0 "C was treated with Et3N (150.5 mL, 1.08 mol), 4-(dimethylamino)pyridine (DMAP, 22 g, 0.18 mol), and Ac20 (94.3 mL, 1.0 mol). The reaction mixture was allowed to warm to 25 "C and stirred for 1 h. The reaction mixture was washed with HzO (2 x 300 mL) and brine (300 mL), and the combined aqueous layer was extracted with CH2C12 (2 x 300 mL). The combined organic layer was dried (Mgsod), concentrated, and purified by flash chromatography (silica, 5% Et20 in petroleum ether) to give 5 (145.4 g, 96%) as a pale yellow oil: Rf = 0.66 (silica, 20% Et20 in petroleum ether); 'H NMR (300 MHz, CDCls) 6 4.96 (s, 1 H, C=CH2), 4.64 (s, 3 H, CsCH2 and CHZOAC),2.02 (s, 3 H, COCH3), 1.77 (s, 3 H, CH,C=C), 1.75 (s, 3 H, CHsC=C), 1.71 (s, 3 H, CHsC=C). Chloro Nitrile 7. A mixture of diene 5 (90.3 g, 537 "01) and freshly distilled 2-chloroacrylonitrile (65 mL, 806 "01, purchased from Tokyo-Kasei) was stirred at 130 "C in a sealed tube for 72 h. During the course of the reaction, the reaction mixture turned dark 0 brown. The reaction mixture was allowed to cool to 25 "C and purified by flash chromatography (silica, 10% Et20 in petroleum ether) to give -j& 7 (110 g, 80%) as clear crystals: mp 86-88 "C, from EtzO; Rf = 0.25 (silica, 10% Et20 in petroleum ether); IR (thin film) v,, 2979, 2938, 1730, 1436, 1370, 1240 cm-'; 'H NMR (300 MHz, CDC13) 6 4.63 (s, "'OBn "1- OI H OBn 2 H, CH~OAC),2.48-2.29 (band, 4 H, 13-CH2 and 14-CH2), 2.07 (s, 3 H, COCHj), 1.75 (s, 3 H, l8-CH3), 1.39 (s, 3 H, 16-CH3), 1.28 (s, 3 0 'OMOM Yo CMOM vu 0 H, 17-CH3); FAB HRMS (NBNCsI) de388.0080, M + Cs+ calcd for C13H&102N 388.0080. 107 108 Hydroxy Ketone 8. A solution of chloro nitrile 7 (15 g, 58.7 "01) a Reagents and conditions: (a) 1.3 equiv of 15, 2.6 equiv of n-BuLi, and KOH (19.8 g, 352 "01) in t-BuOH (293 mL) was heated to 70 THF, -78 "C, 0.5 h, then 0 "C, 1.0 equiv of 67, THF, -78 "C, 20 min, "C and stirred for 4 h. After being cooled to 25 "C, the reaction mixture 82%; (b) 2.0 equiv of LiAW, EtzO, 25 "C, 0.5 h, 87%; (c) 2.0 equiv of was diluted with EtOAc (lo00 mL) and washed with H20 (2 x 300 t-BuOOH, 0.05 equiv of VO(acac)z, PhH, 25 "C, 0.5 h, 95%; (d) 15 equiv mL) and brine (300 mL). The combined aqueous layer was extracted of LiAl&, EtzO, 25 "C, 3 h, 78% based on 81% conversion; (e) 10 equiv with EtOAc (4 x 200 mL), and the combined organic layer was dried of MOMCl, 12 equiv of i-PrWt, CHZClZ, 25 "C, 10 h, 99%; (f) excess n-Bu$rTF (TBAF), THF, 25 "C, 2 h, then 4.0 equiv of AczO, 6.0 equiv of (MgS04), concentrated, and purified by flash chromatography (silica, 4-(dimethylamino)pyridine(DMAP), CHZC12,25 "C, 2 h, 83%; (g) 5.0 equiv 25 - 30% EtOAc in benzene) to give 7 (4.5 g, 30%) and 8 (6.2 g, of phosgene, 5.0 equiv of KH, EtzO, HMPA, 25 "C, 1 h, 88% based on 90% based on 70% conversion) as a pale orange oil: Rj = 0.30 (silica, 57% conversion; (h) excess TBAF,THF, 25 "C, 1 h, 88%; (i) 0.05 equiv 30% EtOAc in benzene); IR (thin film) v,, 3410, 2980, 2930, 1710, of tetrapropylammonium permthenate (TPAP), 3 .O equiv of 4-methylmor- 991 cm-l; IH NMR (500 MHz, CDC13) 6 4.26 (s, 2 H, CHZOH),2.53 pholine N-oxide (NMO),CH~CN-CH~C~Z (1:l). 25 "C, 0.5 h, 91%; (j) 10 (t, J = 8.5 Hz, 2 H, 13-CHz), 2.38 (t, J= 8.5 Hz, 2 H, 14-CHz), 1.84 equiv of TiCly(DME)1.5,20 equiv of Zn-Cu, DME, reflux, 3 h, 106 added (s, 3 H, 18-CH3), 1.44 (b S, 1 H, OH), 1.19 (s, 6 H, 16-CH3 and 17- over 1 h, then 1.5 h, 15%. Piv = CO-t-Bu, TBS = Si-t-BuMez, Bn = CH3); FAB (NBA/NaI) de191.1048, M Naf calcd for CHzPh, TPS = Si-t-BuPhz, MOM = methoxymethyl. HRMS + Cl&IlaOz 191.1050. The combined aqueous layer was extracted with Et20 (2 x 200 mL). TBS Ether 9. A solution of alcohol 8 (4.00 g, 23.8 "01) and The combined organic layer was dried (MgS04) and concentrated to imidazole (1.95 g, 28.6 mol) in CHzClz (40 mL) at 0 "C was treated give the corresponding alcohol which was taken in the next step without with tert-butyldimethylsilyl chloride (TBSCl, 3.96 g, 26.2 mmol), further purification. allowed to warm to 25 "C, and stirred for 2 h. After dilution with A solution of the previous alcohol in benzene (600 mL) was treated Et20 (300 mL), the reaction mixture was washed with HzO (100 mL) with p-toluenesulfonic acid (54 g, 276 "01) and heated to 65 "C for and brine (100 mL), dried (MgSOd), concentrated, and purified by flash 3 h. After being cooled to 25 "C, the reaction mixture was treated chromatography (silica, 5 - 10% Et20 in petroleum ether) to give 9 with Et3N (39 mL, 280 mmol), diluted with Et20 (600 mL), washed (5.71 g, 85%) as a pale yellow oil: Rj = 0.48 (silica, 10% Et20 in with H20 (400 mL), aqueous NaHC03 (400 mL), and brine (400 mL), petroleum ether); IR (thin film) v,, 2929, 2857, 1716, 1462, 1377, dried (MgS04), concentrated (bath temperature <30 "C), and distilled 1253 cm-'; 'H NMR (500 MHz, CDC13) 6 4.12 (s, 2 H, lO-CHZ), 2.51 Total Synthesis of Taxol. 2 J. Am. Chem. SOC., Vol. 117, No. 2, I995 643

(t, J = 7.0 Hz, 2 H, 13-CHz), 2.36 (t, J = 7.0 Hz, 2 H, 14-CHz), 1.74 concentrated to give crude aldehyde 38 as a yellowish solid (40.7 g) (s, 3 H, 18-CH3), 1.18 (s, 6 H, 16-CH3 and 17-CH3), 0.87 (s, 9 H, which was taken to the next step without further purification. SiC(CH3)3(CH3)2), 0.05 (s, 6 H, SiC(CH3)3(CH3)2); I3C NMR (125 MHz, Aldehyde 38: Rf = 0.60 (silica, 50% Et20 in petroleum ether); 'H CDC13) 6 215.2, 135.6, 131.8, 59.2, 41.2, 35.7, 30.5, 25.8, 24.6, 22.5, NMR (300 MHz, CDC13) 6 9.72 (s, 1 H, CHO), 7.67-7.37 (band, 10 18.2, -5.5; FAB HRMS (NBNCsI) de697.3056, M + Cs+ calcd for H, Ar), 4.21 (s, 2 H, CHz), 1.09 (s, 9 H, t-Bu). C 1&002Si 697.3085. Conversion of 38 to 40. To a solution of the crude aldehyde 38 TBS Hydrazone 15. A solution of ketone 9 (2.79 g, 9.88 "01) (13.2 g) in CHzClz (200 mL) was added (carbethoxyethy1idene)- in THF (33 mL) at 25 "C was treated with (2,4,6-triisopropylbenze- triphenylphosphorane (22.3 g, 62.0 mol) in one portion. The reaction nesulfony1)hydrazine (2.95 g, 9.88 "01) and stirred for 24 h. The mixture was stirred at 25 "C for 20 h, concentrated, and purified by reaction mixture was concentrated, and the solid residue was dissolved flash chromatography (silica, 5 - 10% Et20 in petroleum ether) to in a minimum amount of Et20 (10 mL). The solution was diluted with give 40 (13.7 g, 91% from 36) as an oil: Rf = 0.80 (silica, 20% Et20 petroleum ether (50 mL) and cooled to -20 "C to induce crystallization. in petroleum ether); 'H NMR (300 MHz, CDC13) 6 7.70-7.36 (band, After removal of the mother liquor by filtration, the crystalline material lOH,Ar),6.87(t,J=5.6H~,CH=),4.36(d,J=5.6H~,2H,CHz), was washed with petroleum ether (30 mL) and dried in vacuo to give 4.20 (q, J = 7.2 Hz, 2 H, COOCHz), 1.64 (s, 3 H, Me), 1.30 (t, J = 15 (4.89 g, 88%) as colorless crystals: mp 135-137 "C, from EtzO- 7.2 Hz, 3 H, COOCHZCH~),1.05 (s, 9 H, t-Bu). petroleum ether; Rf = 0.24 (silica, 20% Et20 in petroleum ether); IR Ester 41. A solution of aldehyde 39 (159 g, 1.10 mol) in CHzClz (thin film) v,, 3250,2957, 1600 cm-I; IH NMR (500 MHz, CDC13) (600 mL) at 0 "C was treated with a solution of (carbethoxymethy1ene)- 6 7.68 (b s, 1 H, NH), 7.21 (s, 2 H, Ar), 4.28 (septet, J = 7.0 Hz, 2 H, triphenylphosphorane (408 g, 1.13 mol) in CHZC12 (1200 mL) over o-CH(CH&), 4.14 (s, 2 H, lO-CH;?), 2.95 (septet, J = 7.0 Hz, 1 H, the period of 4 h. The solution was allowed to warm to 25 "C and p-CH(CH3)2), 2.44 (t, J = 7.0 Hz, 2 H, 13-CH2), 2.21 (t, J = 7.0 Hz, stirred for 18 h. The mixture was concentrated, suspended in 30% 2 H, 14-CH2), 1.75 (s, 3 H, 18-CH3), 1.33 (d, J = 7.0 Hz, 12 H, O-CH- Et20 in hexanes, and filtered through a pad of silica gel to give 41 (cH3)2), 1.32 (d, J = 7.0 Hz, 6 H, p-CH(CH3)2), 1.13 (s, 6 H, 16-CH3 (222 g, 90%) as an oil: Rf = 0.40 (silica, 30% EtOAc in hexanes); 'H and 17-C&), 0.92 (s, 9 H, SiC(CH3)3(CH3)2), 0.10 (s, 6 H, SiC(CH3)3- NMR (500 MHz, CDC13) 6 6.78 (m, 1 H, %H), 4.59 (m, 1 H, OCHO), (CH3)2); I3C NMR (125 MHz, CDC13) 6 164.0, 152.9, 151.1, 136.0, 4.37 (m, 1 H, A of AB, CH~CHZO),4.13 (band, 3 H, =CCH20 and 131.5, 131.4, 124.0, 123.4,59.0,42.2, 34.1,30.6,29.8, 26.0,26.0,25.9, CH~CHZO),3.80 (m, 1 H, B of AB, CHZCH~O),3.47 (m, 1 H, B of 25.9, 25.8, 24.9, 24.8, 24.7, 24.6, 23.5, 21.4, 19.5, 18.3, -5.4; FAB AB, =CCHzO), 1.79 (s, 3 H, CH3), 1.76-1.47 (band, 6 H, CHz), 1.23, HRMS (NBA) de 563.3716, M + H+ calcd for C31H5403N2SSi (t, J = 7.0 Hz, 3 H, CH3CH2). 563.3703. Alcohol 42. A solution of ether 41 (222 g, 0.97 mol) in MeOH Aldehyde 39. A solution of 1,4-dihydroxy-cis-2-butene(137 g, 1.56 (2500 mL) at 25 "C was treated with p-toluenesulfonic acid (1 g) and mol) and p-toluenesulfonic acid (1.35 g, 7 "01) in CH2C12 (2000 stirred at 25 "C for 18 h. The mixture was treated with Et3N (2 mL), mL) at 25 "C was treated with dihydropyran (300 mL, 3.29 mol), concentrated, redissolved in EtOAc (1500 mL), washed with aqueous dropwise, over the period of 0.5 h. After being stirred for 10 min, the NaHC03 (2 x 100 mL), HzO (2 x 100 mL), and brine (2 x 100 mL), mixture was treated with Et3N (2.0 mL, 14 mmol), reduced in volume dried (MgS04). filtered, and concentrated to give a clear oil that was to a total of 2 L, treated with decolorizing carbon, filtered through a purified by flash chromatography (silica, 40% ethyl acetate in hexanes) pad of silica gel, and concentrated to give a yellow oil that was taken to give 42 (128 g, 92%) as a colorless oil: Rf = 0.20 (silica, 30% to the next step without further purification: Rf = 0.30 (silica, 25% EtOAc in hexanes); IR (thin film) vm, 3434,2983,2934, 1713, 1650, EtOAc in hexanes); IR (thin film) v,, 2950,2800, 1200, 1050 cm-I; 1446, 1368, 1261, 1132, 1031, 731; 'H NMR (500 MHz, CDC13) 6 'H NMR (500 MHz, CDC13) 6 5.65 (t, J = 5.0 Hz, 2 H, CH), 4.55 (b 6.62 (b S, 1 H, =CH), 4.11 (d, J= 6.0 Hz, 2 H, OCHzCH), 3.98 (q, J s, 2 H, OCHO), 4.19 (m. 2 H, CH2CH20),4.03 (m, 2 H, CHZCHZO), = 7.0 Hz, 2 H, CHZCH~),3.90 (s, 1 H, OH), 1.61 (s, 3 H, CH3C), 1.09 3.77 (m, 2 H, CHCHzO), 3.42 (m, 2 H, OCHzCH), 1.75-1.43 (band, (t, J = 7.0 Hz, 3 H, CH2CH3); 13C NMR (125 Hz, CDC13) 6 167.5, 12 H, CH2); I3C NMR (125 Hz, CDC13) 6 129.0,97.7,62.6, 61.9.30.4, 140.6, 127.5, 60.4, 58.8, 13.7, 12.0; FAB HRMS (NBNCsI) de 25.3, 19.2; FAB HRMS (NBA/NaI) de279.1572, M + Na+ calcd for 276.9841, M + Cs+ calcd for C7H1203 276.9846. Cl4H2404 279.1572. Diol 55. A. Small-Scale Procedure. A mixture of dienophile 42 A solution of the previous alkene (200 g, 0.78 mol) in CHZClz (600 (1.44 g, 10 mmol), diene 52 (1.52 g, 13.6 mmol), and PhB(OH)2 (1.7 mL) was treated with at -78 "C until the solution tumed blue. g, 13.9 "01) in benzene (30 mL) was stirred at reflux with azeotropic The reaction was quenched by the careful addition of triphenylphosphine removal of water (Dean-Stark trap) for 48 h. After the solution was (205 g, 0.78 mol) in portions. The mixture was allowed to warm to cooled to 25 "C, the reaction was quenched with 2,2-dimethyl-1,3- 25 "C over the period of 8 h, concentrated, washed with Et20 (3 x propanediol(l.45 g, 13.9 mol) and the resulting mixture was stirred 500 mL), and filtered. The combined washes were concentrated and at 25 "C for 1 h, concentrated, and purified by flash chromatography purified by flash chromatography (silica, 25% EtOAc in hexanes) to (silica, 10 - 50% EtOAc in hexanes) to give dienophile 42 (0.33 g, give aldehyde 39 (220 g, 98%) as a clear oil: Rf = 0.20 (silica, 25% 23%), diene 52 (0.51 g, 34%), and diol 55 (1.56 g, 79% based on 77% EtOAc in hexanes); IR (thin film) vmax2944,2889, 1739, 1136, 1078, conversion) as a yellow oil. 1033; 'H NMR (500 MHz, CDC13) 6 9.72 (s, 1 H, HCO), 4.63 (t, J = B. Large-Scale Procedure. A mixture of dienophile 42 (70.0 g, 4.0 Hz, 1 H, OCHO), 4.22 (d, J = 18.0 Hz, A of AB, COCHzO), 4.16 0.49 mol), diene 52 (54.4 g, 0.49 mol), and PhB(0H)z (56.3 g, 0.45 (d, J = 18.0 Hz, B of AB, COCHZO),3.83 (m, 2 H, CHZCHZO),3.50 mol) in benzene (lo00 mL) was stirred at reflux with azeotropic removal (m, 2 H, CHZCH~O),2.00 - 1.50 (band, 4 H, CHI); 13C NMR (125 of water (Dean-Stark trap) for 144 h. After the solution was cooled Hz, CDC13) 6 201.2, 99.4, 72.9, 62.5, 30.2, 25.1, 19.6; FAB HRMS to 25 "C, the reaction was quenched with 1,3-propanediol (36.8 mL, (NBADJaI) de167.0684, M + Na+ calcd for C7H1203 167.0684. 0.51 mol) and the resulting mixture was stirred at 25 OC for 2.5 h, Silyl Ether 40. Aldehyde 38 from 36. To a solution of allylic concentrated, and purified by flash chromatography (silica, 10 - 50% alcohol 36 (11.6 g, 200.0 "01) and imidazole (15.7 g, 230.9 mol) EtOAc in hexanes) to give dienophile 42 and diene 52 (64.7 g, 52%, in DMF (200 mL) was added tert-butylchlorodiphenylsilane(58.4 mL, 1:l mixture), plus diol 55 (34.88 g, 58% based on 48% conversion) as 220.0 mmol) dropwise at 0 OC. The solution was stirred at 0 OC for 1 a yellow oil: Rf = 0.13 (silica, 50% EtOAc in hexanes); IR (thin film) h. After dilution with Et20 (500 mL), the solution was washed with vmax3423,2987, 1766, 1715, 1257, 1202, 1021 cm-I; 'H NMR (500 aqueous NH4Cl (100 mL), H20 (3 x 50 mL), and brine (100 mL). MHz, CDC13) 6 6.06 (dd, J = 10.0, 4.0 Hz, 1 H, 6-H), 5.78 (b d, J = The organic layer was dried (MgS04) and concentrated to give the 10.0 Hz, 1 H, 5-H), 4.57 (dd, J = 9.5, 7.5 Hz, 1 H, 2-H), 4.57-4.55 corresponding crude silyl ether (66.2 g) which was taken to the next (band, 1 H, 7-H), 4.42 (dd, J = 9.5, 8.5 Hz, 1 H, 2-H), 4.15 (q, J = step without further purification. 7.0 Hz, 2 H, COZCH~CH~),4.18-4.12 (band, 1 H, 4-OH), 3.07 (b t, J A fraction of the crude silyl ether (13.0 g) was dissolved in CHZC12 = 8.5 Hz, 1 H, 3-H), 3.04 (b d, J = 5.0 Hz, 1 H, 7-OH), 1.25 (s, 3 H, (300 mL) and treated with 03 at -78 "C for 1 h. The reaction was 19-CH3), 1.94 (t, J = 7.0 Hz, 3 H, C02CHzCH3); I3C NMR (125 MHz, quenched with Ph3P (25.0 g, 96.0 mmol) at -78 OC, and the resulting CDC13) 6 176.5, 175.6, 133.0, 124.9,71.6, 66.8, 62.4,47.3,46.6,42.0, mixture was allowed to warm to 25 "C. After being stirred at 25 "C 15.4, 13.8; FAB HRMS (NBA/NaI) mle 279.0859, M + Na+ calcd for for 0.5 h, the reaction mixture was diluted with toluene (100 mL) and C&1606 279.0845. 644 J. Am. Chem. SOC.,Vol. 1 17, No. 2, I995 Nicolaou et al.

Bis(sily1 ether) 58. A solution of diol 55 (28.5 g, 111 mmol), 2,6- Diol 60. A solution of alcohol 59 (43.9 g, 99 "01) in CHzClz lutidine (102 mL, 445 mmol), and 4-(dimethylamino)pyridine (DMAP, (250 mL) and MeOH (20 mL) was treated with camphorsulfonic acid 1.50 g, 12.2 "01) in CHzClz (250 mL) was treated with teri- (CSA, 0.52 g, 5 "01) and stirred at 25 OC for 1 h. After dilution butyldimethylsilyl trifluoromethanesulfonate (TBSOW, 52.0 mL, 445 with CHzClz (300 mL), the reaction was quenched with aqueous "01) and stirred at 0 "C for 4 h. The reaction mixture was added to (150 mL). The organic layer was separated, and the aqueous aqueous NaHCo3 (100 mL), extracted with Et20 (2 x 150 mL), washed layer was extracted with Et20 (2 x 200 mL). The combined organic with aqueous CuSO4 (2 x 100 mL), dried (NazSOh), concentrated, and layer was dried (NazS04), concentrated, and purified by flash chro- purified by flash chromatography (silica, 5 - 15% Et20 in petroleum matography (silica, 50% Et20 in petroleum ether) to give diol 60 (32.6 ether) to give 58 (49.6 g, 92%) as a white solid: Rf = 0.62 (silica, g, 94%) as white crystals: mp 109-1 11 "C, from EtOAc-hexanes; Rf 15% Et20 in petroleum ether); IR (thin film) vmm2960, 2936, 2857, = 0.38 (silica, EtzO); IR (thin film) vmm3433,2932,2859,1766,1469, 1746, 1256 cm-'; 'H NMR (500 MHz, C&) 6 6.19 (dd, J = 8.5, 5.0 1384, 1081, 1023; 'H NMR (500 MHz, CDCl3) 6 5.99 (ddd, J= 18.0, Hz, 1 H, 6-H), 6.10 (dd, J = 8.5, 1.0 Hz, 1 H, 5-H), 4.12 (dd, J = 8.5, 3.0, 1.5 Hz, 1 H, 5-H),5.82 (dd, J= 18.0, 1.5 Hz, 1 H, 6-H), 4.38 (A 4.5 Hz, 1 H, 2-H), 4.11 (dd, = 5.0, 1.0 Hz, 1 H, 7-H), 3.83-3.70 J of ABX, dd, J = 9.5, 7.5 Hz, 1 H, 2-H), 4.33 (B of ABX, ddd, J = (band, 2 H, CO~CHZCH~),3.40 (d, J = 8.5 Hz, 1 H, 2-H), 2.83 (d, J 9.5, 5.0, 1.0 Hz, 1 H, 2-H), 4.24 (b S, 1 H, 7-H), 3.57 (A' of A'B', d = 4.5 Hz, 1 H, 3-H), 1.22 (s, 3 H, 19-C&), 1.02 (s, 9 H, Si(C(CH3)3)- b, J = 11.0 Hz, 1 H, 9-H), 3.39 (B' of A'B', b d, J = 11.0 Hz, 1 H, (CH3)Zh 0.97 (s, 9 H, S~(C(CH~)S)(CH~)Z),0.32 (s, 3 H, Si(C(CH3M- 9-H), 2.70-2.33 (band, 2 H, 9-OH and 7-OH), 2.55 (X of dd, (cH3)2), 0.30 (s, 3 H, S~(C(CH~M(CH~)Z),0.21 (s, 3 H, Si(C(CHsh)- ABX, 7.5,S.O Hz, 1 H, 3-H), 0.88 (s, 3 H, 19-c&), 0.83 (s, 9 H, Si(C(CH3)3)- (CH3)2), 0.15 (S, 3 H, Si(C(CH&)(CH3)2); 13C NMR (125 MHz, CD.5) 6 174.0, 133.5, 132.4, 119.0, 80.0, 70.9, 62.8, 60.6, 53.0, 45.9, 26.0, (CH3)2), 0.16 (s, 6 H, Si(C(CH3)3)(CH3)2);13C NMR (125 MHz, CDC13) 25.9, 20.5, 18.4, 14.0; FAB HRMS (NBA/NaI) mle 617.1731, M + 6 175.7, 135.1, 124.4, 74.5, 68.7, 67.7, 66.4, 47.5, 41.9, 25.6, 18.1, Na+ calcd for C2&06Si2 617.1731. 12.9, -2.7, -3.1; FAB HRMS (MA) de329.1772, M + H+ calcd Alcohol 59. A solution of ester 58 (49.6 g, 102 "01) in Et20 for C1&~0~Si329.1784. (500 mL) at 0 "C was treated with LiAEb (1 10 mL of a 1 M solution, 110 mmol), allowed to warm to 25 "C, and stirred at 25 "C for 0.5 h. Acknowledgment. We thank Drs. Dee H. Huang, Gary After the solution was cooled to -78 "C, the reaction was quenched Siuzdak, and Raj Chadha for NMR, mass spectroscopic, and with EtOAc (25 mL) and aqueous NH&1 (150 mL). The reaction X-ray crystallographic assistance, respectively. This work was mixture was allowed to warm to 25 "C and stirred for 1 h. The organic financially supported by NM, The Scripps Research Institute, layer was separated, and the aqueous layer was extracted with Et20 (3 x 200 mL). The combined organic layer was dried (Na2S04), fellowships from Mitsubishi Kasei Corporation (H.U.), R.W. concentrated, and purified by flash chromatography (silica, 20 - 45% Johnson-ACS Division of Organic Chemistry (E.J.S.),The Et20 in petroleum ether) to give 59 (43.9 g, 97%) as a white solid: Rj Office of Naval Research (R. K. G.), Glaxo, Inc. (C.F.C.), Mr. = 0.22 (silica, 30% Et20 in petroleum ether); IR (thin film) vman2955, Richard Staley (C.F.C.), RhBne-Poulenc Rorer (P.G.N.), and 2931, 2857, 1471, 1280, 1253 cm-'; lH NMR (500 MHz, CDC13) 6 grants from Merck Sharp & Dohme, Pfizer, Inc., Schering 6.43 (dd, J= 8.5, 5.3 Hz, 1 H, 6-H), 6.20 (dd, J= 8.5, 1.7 Hz, 1 H, Plough, and the ALSAM Foundation. 5-H),4.10 (dd, J = 8.0, 4.1 Hz,1 H, 2-H), 3.95 (dd, J= 5.3, 1.7 Hz, 1 H, 7-H), 3.58 (d, J = 8.0 Hz, 1 H, 2-H), 3.25 (dd, J = 10.4, 4.3 Hz, Supplementary Material Available: Experimental tech- 1 H, 9-H), 3.15 (dd,J= 10.4,4.3 Hz, 1 H, 9-H), 1.60 (b t, J=4.3 Hz, 1 H, 9-OH), 1.47 (d, J = 4.1 Hz, 1 H, 3-H), 1.22 (s, 3 H, Ig-CH,), niques and data for compounds 10-14,16,18-35,47-51,57, 0.92 (s, 9 H, Si(C(CH3)3)(CH3)d70.86 (s, 9 H, S~(C(CH~)~)(CH~)Z),58, 61-80, 82-87, and 89-107 (44 pages). This material is 0.17 (s, 3 H, Si(C(CH3)3)(CH3)2). 0.15 (s, 3 H, S~(C(CH~)~)(CH~)Z),contained in many libraries on microfiche, immediately follows 0.12 (s, 3 H, Si(C(CHMCH3M, 0.10 (s, 3 H, Si(C(CH3)3)(CH3)2); this article in the microfilm version of the journal, and can be I3C NMR (125 MHz, Cas) 6 132.8, 131.7, 119.0, 80.0, 72.0, 69.6, ordered from the ACS. See any current masthead page for 63.1, 46.0,44.7, 26.0, 25.7, 18.9, 18.2, 18.0, -2.9, -3.0, -3.1, -3.2; ordering information. FAB HRMS (NBA/CsI) mle 575.1636, M + Cs+ calcd for CzzH4205- Si2 575.1625. JA942 193U J. Am. Chem. SOC.1995,117, 645-652 645

Total Synthesis of Taxol. 3. Formation of Taxol's ABC Ring Skeleton

K. C. Nicolaou,* Z. Yang, J.-J. Liu, P. G. Nantermet, C. F. Claiborne, J. Renaud, R. K. Guy, and K. Shibayama Contribution from the Department of Chemistry, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, Califomia 92037, and Department of Chemistry and Biochemistry, University of Califomia, San Diego, 9500 Gilman Drive, La Jolla, Califomia 92093 Received July 7, 1994@

Abstract: The synthesis of Taxol's ABC ring system has been achieved. The Shapiro coupling of an aldehydic C ring synthon (8) with an anionic A ring synthon derived from hydrazone 9 gave, diastereoselectively, A-B conjugate 10. Functional group manipulations and McMuny ring closure produced the highly functionalized ABC ring system 17. Extensive attempts to optimize the McMuny reaction revealed a single predominant side reaction leading to byproducts 19 and 20. Resolution of the C9,ClO-diol (f)-17 via its camphanyl esters provided the ABC ring system as its natural isomer (+)17.

Introduction In the preceding two papers1-*in this series, we described our degradation and reconstruction studies with Taxol (1, Figure l), preliminary investigations with rings A and C, and possible schemes for their elaboration to an appropriately functionalized ABC taxoid framework. Armed with the knowledge gained in these studies, we were now ready to attempt the final drive 1 : Taxol toward Taxol's ABC ring skeleton. As already discussed, the Figure 1. Structure and numbering of Taxol (1). starting materials were defined as hydrazone 92 (Scheme 2) and aldehyde 8 (Scheme l),the synthesis of which is detailed below. Scheme 1. Synthesis of C Ring Aldehyde 8" The C4-C20 five-membered acetonide group was chosen as a HO OH TPSO OR TPSO 08 means to protect the vicinal diol system of the intermediate and to introduce additional rigidity in the system prior to cyclization -C to form the 8-membered ring. )TBS Construction of Taxol's ABC Ring Skeleton 2 a. Synthesis of the C Ring Aldehyde 8. Scheme 1 summarizes the preparation of the targeted aldehyde 8 from the previously described intermediate 2.* Thus, treatment of diol 2 with tert-butyldiphenylsilyl chloride (TPSC1) and imidazole3 resulted in monosilylation of the primary alcohol, providing the C7 hydroxyl, C9 silyl ether 3 in 92% yield. Benzylation of the C7 hydroxyl group using KH and benzyl bromide4 afforded benzyl ether 4 in 88% yield. Exhaustive reduction of the lactone ring in 4, accompanied by removal of the C4 TBS group, resulted in the formation of triol 5 (80% yield). The crucial 5-membered ring acetonide was then installed using 2,2- 8 7 6 dimethoxypropane in the presence of a catalytic amount of CSA5 Reagents and conditions: (a) 1.3 equiv of TPSCl, 1.35 equiv of in methylene chloride:ether (98:2) at ambient temperature. imidazole, DMF, 25 "C, 12 h, 92%; (b) 1.2 equiv of KH, 1.2 equiv of Under these conditions, the reaction was found to be quite rapid PhCHZBr, 0.04 equiv of n-BN,EtzO, 25 OC, 1 h, 88%; (c) 3.0 equiv of LiAlH4, EtzO, 25 OC, 12 h, 80%; (d) 5.0 equiv of 2,2-dimethox- * Address correspondence to this author at The Scripps Research Institute ypropane, 0.05 equiv of camphorsulfonic acid (CSA), CHzC1Z:EtzO (98: or the University of California. 2), 25 OC, 7 h, 82%; (e) 0.05 equiv of tetrapropylammonium @ Abstract published in Advance ACS Abstracts, December 15, 1994. permthenate (PAP), 1.5 equiv 4-methylmorpholineN-oxide (1) Nicolaou, K. C.; Nantennet, P. G.; Ueno, H.; Guy, R. K.; Coula- of WO), douros, E. A.; Sorensen, E. J. J. Am. Chem. SOC.1995, 117, 624. CH3CN, 25 "C, 2 h, 97%. TBS = Si-t-BuMez, Bn = CHZPh, TPS = (2) Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorensen, E. J.; Si-t-BuPhz. Claibome, C. F.; Guy, R. K.; Hwang, C.-K.; Nakada, M.; Nantermet, P. G. J. Am. Chem. SOC. 1995, 117, 634. with the initially formed 7-membered ring acetonide 6 rear- (3) Hanessian, S.; Lavallk, P. Can. J. Chem. 1975,53,2975. Hanessian, ranging slowly and essentially completely to the desired, and S.: LavallBe. P. Can. J. Chem. 1977. 55. 562. thermodynamically more stable, 5-membered ring isomer 7 (4) Kand;, K.; Sakamoto, I.; Ogawa, S:; Suami,T. Bull. Chem. SOC.Jpn. 1987, 60, 1529. (82%). Finally, PAP-NMO oxidation6 of the remaining (5) Lipshutz, B. H.; Barton, J. C. J. Org. Chem. 1988, 53, 4495. hydroxyl group in 7 furnished the targeted aldehyde 8 in 97% 0002-786319511517-0645$09.0010 0 1995 American Chemical Society 646 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al.

(Reface) Nu'

12

16: L? chelate derived from aldehyde 8 Figure 3. Stereoselectivity of the Shapiro reaction. The model was generated with Chem3d. Most hydrogens are omitted for clarity.

Scheme 2 to afford allylic alcohol 10 as a single diastereoisomer and in 82% yield. X-ray crystallographic analysis of a subsequent intermediate confirmed the stereochemical structure of 10 (vide infra). The stereoselectivity of this reaction can be explained by invoking the chelated intermediate 16, depicted in Figure 3, in which the acetonide plays a crucial role. As seen in this model, the aldehyde group is fixed by the lithium template in a conformation in which nucleophilic attack can freely proceed from only one side, the re face, with the si face being blocked by the C8 methyl group. Directed epoxidation9 of the C1-C14 double bond in 10, although slow, proceeded smoothly to afford the single epoxide 11 in 87% yield. Regioselective openingloof the epoxide group qn0- in 11 with LiAl& resulted in the formation of diol 12 in 76% yield. The crystalline diol 12 was subjected to X-ray crystal- k" 0% lographic analysis (see ORTEP drawing, Figure 2) confirming 20 the assigned stereochemistry of all intermediates in Scheme 2. Exposure of 12 to excess KH and phosgene in ether:HMPA (3: 1) resulted in the formation of carbonate 13 (86%yield, 58% conversion). Desilylation of 13 with fluoride ion3 furnished diol 14 (80%yield), which was oxidized smoothly with TPAP- NM06 to afford the dialdehyde 15 (92% yield)- preorganized in a conformation favorable for the upcoming McMurry cyclization. c. The McMurry Cyclization and Synthesis of the ABC wo,0 Ring Skeleton 17. The search for the conditions required to yield the requisite cyclized product using the McMurry pinacol (6) Griffith, W. P.; Ley, S. V. Aldrichimica Acta 1990, 23, 13. (7) Shapiro, R. H. Org. React. 1976,23,405. Chamberlin, A. R.; Bloom, S. H. Org. React. 1990, 39, 1. Martin, S. F.; Daniel, D.; Cherney, R. J.; Liras, S. J. Org. Chem. 1992, 57, 2523. (8) This strategy was later used by others to accomplish similar O couplings: Di Grandi, M. J.; Jung, D. K.; Krol, W. J.; Danishefsky, S. J. 0Po 30 J. Org. Chem. 1993, 58, 4989. Masters, J. J.; Jung, D. K.; Bornmann, W. G.; Danishefsky, S. J. Tetrahedron Lett. 1993, 34, 7253. (9) Sharpless, K. E.; Michaelson, R. C. J. Am. Chem. SOC. 1973, 95, 6136. Sharpless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12, 63. Rao, A. S. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Figure 2. ORTEP drawings for compounds 12, 19, 20, and 30. Ley, S. V., FRS, Eds.; Pergamon Press: New York, 1991; Vol. 7, p 376. (10) Murai, S.; Murai, T.; Kato, S. In Comprehensive Organic Synthesis; Yield. Thus a rapid and efficient Pathway to key intermdate Trost, B. M., Fleming, I., Eds.; Pergamon Press: New York, 1991; Vol. 8, 8 was established. p 871. The (11) McMuny, J. E. Chem. Rev. 1989, 89, 1513. McMurry J. E. Acc. Reaction and Synthesis Of Chem. Res. 1983,16,405. McMurry, J. E.; Lectka, T.; Rico, J. G. J. Org. Dialdehyde 15* The Shapiro reacti0n7'8 Of Chem. 1989,54,3748. McMurry, J. E.; Rico, J. G. Tetrahedron Lett. 1989, 9 with aldehyde 8 proceeded under the conditions specified in 30, 1169. Lenoir, D. Synthesis 1989, 883. Total Synthesis of Taxol. 3 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 647

Scheme 2. Shapiro Coupling of 8 with 9 and Synthesis of Scheme 3. McMurry Cyclization and Synthesis of Diol 17" Dialdehyde 15" n TBSO TPSO OBn - &OTBS+NNHS02Ar OHC

i- Pr. -\ 9, Ar =ei-Pr a 10 i- Pr f TBSO TBSO

V 17 (23-25%) 18 (10%) 12 11

0

f

19 (40%) 20 (15%)

15 a Reagents and conditions: 11 equiv of TiCl3*(DME)1.5,26equiv of Zn-Cu, DME, reflux, 3.5 h, then 70 "C, then 15 added over 1 h, then a Reagents and conditions: (a) 1.1 equiv of 9,2.3 equiv of n-BuLi, 70 "C, 0.5 h. Bn = CH2Ph. THF, -78 - 0 "C, 1.0 equiv of 8, THF, -78 "C, 0.5 h, 82%; (b) 0.03 equiv of VO(acac)z, 3.0 equiv of t-BuOOH, PhH, 4 8, molecular sieves, 25 "C, 14 h, 87%; (c) 5.0 equiv of LiAlh, 25 "C, Et20,7 h, 76%; (d) 3.0 equiv of KH, Et20:HMPA (3:1), 1.6 equiv of phosgene (20% in toluene), 25 "C, 0.5 h, 86% based on 58% conversion; (e) 3.8 equiv of n-Bum (TBAF), THF, 25 OC, 14 h, 80%; (f) 0.05 equiv of tetrapropylammonium permthenate (TPAF'), 3 .O equiv of 4-methyl- morpholine N-oxide (NMO), CH3CN, CH2Cl2, (2:1), 25 "C, 2 h, 92%. Oxo0 ' 03O TBS = Si-t-BuMez, TPS = Si-t-BuPh2, Bn = CHzPh. 6 coupling methodology included varying the temperature (0 - 100 "C), solvent (e.g. THF, DME, ether) and stoichiometry, as well as the use of various bases as additives. It was finally determined that 11 equiv of TiC13*(DME)1.5 and 26 equiv of Zn-Cu couple in DME at 70 "C provided the optimum yield of diol 17 (25%, Scheme 3). In addition to diol 17, whose stereochemistry was assigned on the basis of a subsequent intermediate (vide infia), a number of other products were obtained including olefin 18 (10% yield), lactoll9 (40% yield), and formate ester 20 (15% yield). The structures of 17 and 18 were based solely upon spectroscopic evidence (except for the stereochemistry of 17 at C9 and C10 which was later confirmed, vide infra), whereas those of 19 and 20 were secured from both spectroscopic and X-ray crystallographic data (see ORTEP drawings, Figure 2). 15 Analysis of molecular models for dialdehyde indicated a 15 Figure 4. Possible ground state conformation of 15. The model was possible ground state conformation in which the two aldehyde generated with Chem3d. The C7 benzyl protecting group and all moieties of 15 are in close proximity (Figure 4), thus requiring hydrogens are omitted for clarity. Bn = CH9h. only small conformational changes to reach the geometry necessary for cyclization. Rotation around the C2-C3 carbon- structure 95) offers much higher conformational freedom via carbon bond would either bring the two aldehyde groups in very rotation around the Cl-C2 carbon-carbon bond. Analysis of close proximity, as desired, or induce strong steric interactions molecular models indicated a possible ground state conformation between ring A and the acetonide group. In contrast, dialdehyde (21) (Figure 5) for this compound in which the two aldehyde 21 (see Figure 5 and previous paper2 in this series, Scheme 13, functionalities are far apart. Failure to cyclize to such a system 648 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al.

Ph Scheme 4. Postulated Mechanism of the McMurry \ Cyclization and Formation of Products 17 and 18 Tio 0 -

21 Figure 5. Possible ground state conformation of 21. The model was generated with Chem3d. All hydrogens are omitted for clarity. 23 i 24 in the McMurry reaction may reflect the large entropic and enthalpic cost for the conformational change necessary for reaction to take place. Mechanistic rationales for the formation of products 17-28 are shown in Schemes 4 and 5. The pathways leading to 17- 19 are in accord with previous proposals by McMurryl' and Kende.12 The formation of the keto formate 20, however, requires an additional oxygen atom which may, presumably, come from molecular oxygen introduced during workup. A speculative mechanism for its formation is proposed in Schemes 17 18 4 (15 22 24) and 5 (24 25 27 28 20). ------The appearance of the chiral auxiliary on the C9 hydroxyl Attempts at masking the C11-Cl2 double bond in order to group of these esters (29 and 30) was at first surprising, avoid the formation of byproducts 19 and 20 were abandoned particularly in view of the fact that monoacetylation of diol 17 after unsuccessful early trials. Further studies along this line, leads selectively to the C10 acetate (see following paper).15 however, may prove useful in controlling product formation in Inspection of molecular models revealed rather similar steric this reaction. environments for these two positions, and therefore, predictions d. Resolution of ABC Ring System Diol 17. To secure or rationalizations were not easy to make. Apparently, the more enantiomerically pure intermediates for the synthesis of Taxol reactive allylic C10 hydroxyl group attracts the smaller acetate (l),we decided to attempt a resolution of the racemic diol 17 group, whereas only the C9 hydroxyl can accommodate the obtained from the McMurry cyclization as described above. bulkier camphanate ester functionality. Encouraged by a successful resolution of a similar taxoid13 via camphanate esters,14 we applied the sequence shown in Scheme Conclusion 6 to our system. Treatment of diol (f)-17 with an excess of (18-(-)-camphanic chloride in methylene chloride in the In this paper we describe the successful construction of a presence of Et3N resulted in the formation of two diastereomeric suitable ring C aldehyde (8) and its stereoselective coupling monoesters 29 and 30 in 36% total yield (1:l ratio). Chro- with the ring A hydrazone (9) through a Shapiro reaction. matographic separation of the mixture allowed the more polar Elaboration of the A-C-coupled product (10) led to a dialde- isomer (30, Rf= 0.21, silica, 15% EtOAc in PhH; [a]22~-133 hyde (15) which entered into a successful McMurry cyclization (c 0.49, CHC13)) to crystallize. X-ray crystallographic analysis to afford ring B with retention of the C9 and C10 . (see ORTEP drawing, Figure 2) revealed the absolute stereo- Resolution of the resulting racemic ABC taxoid diol 17 through chemistry of the latter diastereoisomer and thus allowed its diastereomeric camphanate esters (29 and 30) set the stage identification of the requisite isomer for the synthesis of Taxol for an enantioselective synthesis of Taxol(1). The final stages as the less polar diastereoisomer (29; Rf = 0.26, silica, 15% of the total synthesis of this target molecule are described in EtOAc in PhH; +117 (c 0.54, CHC13)). Hydrolysis of the following paper.15 this isomer (29) under basic conditions (K2CO3, MeOH) Experimental Section regenerated diol (+)-17 (90% yield; [o!]22D +187 (c 0.5, CHC13)), now in its enantiomerically pure form. General Techniques. For a description of general technique, see the first paper in this series.l (12) Kende, A. S.; Johnson, S.; Sanfilippo, P.; Hodges, J. C.; Jungheim, Silyl Ether 3. A solution of diol 2 (9.20 g, 28.0 mol) in DMF L. N. J. Am. Chem. SOC. 1986, 108, 3513. (50 mL) was treated with imidazole (2.58 g, 37.9 mol) and (13) Nicolaou, K. C.; Claiborne, C. F.; Nantermet, P. G.; Couladouros, E. A.; Sorensen, E. J. J. Am. Chem. SOC. 1994, 116, 1591. (15) Nicolaou, K. C.; Ueno, H.; Liu, J.-J.; Nantermet, P. G.; Yang, Z.; (14) Gerlach, H. Helv. Chim. Acta 1978, 61, 2773. Renaud, J.; Paulvannan, K.; Chadha, R. J. Am. Chem. Soc. 1995,117,653. Total Synthesis of Tmol. 3 J. Am. Chem. SOC., Vol. I 17, No. 2, 1995 649

Scheme 5. Postulated Mechanism for the Formation of Scheme 6. Resolution of Diol 17" Products 19 and 20 a O% x Ti0 -Tio /

02

(f)-l7 It I 0

25 27 "I H+ Ti 1 /I 1 0

26 28

(+)-17 1 9 1 29 [a]"~+117 (c 0.54, CHC13) OReagents and conditions: (a) 5.0 equiv of (18-(-)-camphanic chloride, 20 equiv of EtsN, 0.05 equiv of 4-(dimethy1amino)pyridine (DMAP), CHzC12, 25 "C, 1 h, 86%; (b) 7.0 equiv of K2C03, MeOH, &pli "0 25 "C, 0.5 h, 90%. Bn = CHgh. Hz, 1 H, 5-H), 4.72 (d, J= 11.5 Hz, 1 H, OCHzPh), 4.58 (d, J= 11.5 oy Hz, 1 H, OCH2Ph). 4.36 (dd, J = 2.5, 2.0 Hz, 1 H, 7-H), 4.08 (dd, J = 9.5, 7.0 Hz, 1 H, 2-H), 3.96 (dd, J = 9.5, 3.5 Hz, 1 H, 2-H), 3.69 19 20 (d, J = 10.6 Hz, 1 H, 9-H), 3.39 (d, J = 10.6 Hz, 1 H, 9-H), 2.66 (dd, tea-butylchlorodiphenylsilane(9.46 mL, 36.0 mol)and stirred at 25 J = 7.0, 3.5 Hz, 1 H, 3-H), 1.08 (s, 9 H, SiC(CH3)3Ph2), 0.78 (s, 9 H, "C for 12 h. After dilution with Et20 (400 mL), the reaction was SiC(CH3)3(CH3)2), 0.77 (s, 3 H, 19-C&), 0.12 (s, 3 H, SiC(CH3)3- quenched with aqueous NaHCO3 (100 mL). The organic layer was (cH3)2), 0.11 (s, 3 H, SiC(CHs)s(CH3)2); 13CNMR (125 MHz, CDCl3) separated, and the aqueous layer was extracted with Et20 (2 x 50 mL). 6 175.6, 138.3, 135.6, 132.9, 132.9, 132.8, 130.0, 129.8, 128.4, 127.8, The combined organic layer was washed with brine (50 mL), dried 127.7, 127.6, 127.4, 124.7,74.5,74.4,72.6,65.7,65.6,47.5,43.9,27.0, (Na2S04), concentrated, and purified by flash chromatography (silica, 25.5, 19.3, 18.0, 12.8, -2.8, -3.1; FAB HRMS (NBNCsI) mle 30% Et20 in petroleum ether) to give 3 (14.6 g, 92%) as a pale yellow 789.2395, M + Cs+ calcd for C39H5205Si~789.2408. oil: Rj = 0.41 (silica, 50% Et20 in petroleum ether); IR (thin film) Triol 5. A solution of lactone 4 (14.7 g, 22.4 mol)in Et20 (150 vmax3460,2954,2931,2857, 1770, 1471, 1110, 1086 cm-'; 'HNMR mL) was treated with LiAlH4 (66 mL of a 1 M solution in EtzO, 66.0 (500 MHz, CDCl3) 6 7.65-7.55 (band, 4 H, Ar), 7.48-7.35 (band, 6 mmol) and stirred at 25 "C for 12 h. After dilution with Et20 (200 H, Ar), 5.91 (dd, J = 10.5, 2.0 Hz, 1 H, 6-H), 5.84 (dd, J = 10.5, 2.5 mL), the reaction mixture was cooled to -78 "C, and the reaction was Hz, 1 H, 5-H), 4.58 (m, 1 H, 7-H), 4.19 (dd, J = 10.0, 6.5 Hz, 1 H, quenched with aqueous WC1 (100 mL). After the solution was 2-H), 3.95 (dd, J = 10.0, 2.0 Hz, 1 H, 2-H), 3.61 (d, J = 10.6 Hz, 1 wmed to 25 OC, the organic layer was separated, washed with brine H, 9-H), 3.41 (d, J = 10.6 Hz, 1 H, 9-H), 2.59 (dd, J = 6.5, 2 Hz, 1 (100 mL), dried (NazSOd), concentrated, and purified by flash H, 3-H), 2.05 (d, J = 5.5 Hz, 1 H, 7-OH), 1.07 (s, 9 H, SiC(CH3)s- chromatography (silica, 60% EtOAc in petroleum ether) to give 5 (9.8 Phz), 0.80 (s, 9 H, SiC(CH3)3(CH&), 0.69 (s, 3 H, 19-C&), 0.11 (s, 6 g, 80%) as a colorless oil: Rj = 0.23 (silica, 50% EtOAc in hexanes); H, SiC(CH&(CH&); 13C NMR (125 MHz, CDC13) 6 175.3, 136.1, IR (thin film) vmar3374, 2927, 2851, 1463, 1422, 1387, 1105 cm-'; 135.6, 135.5, 132.6, 132.5, 130.1, 127.9, 124.6, 74.5, 68.7, 66.6, 65.6, 'H NMR (500 MHz, CDC13) 6 7.65-7.55 (band, 4 H, Ar),7.45-7.15 47.2,44.1,26.9,25.4, 19.2, 18.0, 11.0, -2.8, -3.1; FAB HRMS (NBN (band, 11 H, Ar), 5.85 (dd, J= 10.0,2.5 Hz, 1 H, 6-H), 5.69 (dd, J= NaI) mle 589.2795, M + Na+ calcd for C3zH4605Si~589.2782. 10.0, 1.5 Hz, 1 H, 5-H), 4.55 (d, J= 11.5 Hz, 1 H, OCHzPh), 4.27 (d, Benzyl Ether 4. A solution of alcohol 3 (21.5 g, 37.9 mmol), benzyl J= 11.5 Hz, 1 H, OCHzPh), 4.01 (b S, 1 H, 7-H), 3.96-3.89 (band, 3 bromide (5.4 mL, 45.4 mmol), and n-BW (0.5 g, 1.35 mmol) in Et20 H, 20-CH2 and 2-H), 3.72 (d, J = 10.5 Hz, 1 H, 9-H), 3.70 (s, 1 H, (300 mL) was treated with KH (6 g of a 30% suspension in mineral 4-OH), 3.58 (m, 1 H, 2-H), 3.51 (d, J = 10.5 Hz, 1 H, 9-H), 3.45- oil, 44.8 mmol, prewashed with dry EtzO) and stirred at 25 "C for 1 h. 3.35 (band, 2 H, 2-OH and 20-OH), 2.15 (dd, J = 6.5, 3.5 Hz, 1 H, After the reaction was quenched with MeOH (5 mL), the reaction 3-H), 1.09 (s, 9 H, SiC(CH3)3Phz), 0.89 (s, 3 H, 19-CI-h); 13C NMR mixture was stirred at 25 "C for 15 min. After dilution with Et20 (125 MHz, CDC13) 6 138.1, 135.8, 135.7, 132.9, 131.2, 129.9, 129.8, (200 mL), the resulting solution was washed with brine (100 mL), dried 128.3, 128.2, 127.7, 127.5, 127.3, 76.2, 73.1, 71.6, 67.1, 66.7, 59.4, (Na2S04), concentrated, and purified by flash chromatography (silica, 48.0, 43.4, 27.0, 25.8, 19.3, 15.3; FAB HRMS (NBNCsI) mle 10 - 30% Et20 in petroleum ether) to give 4 (21.9 g, 88%) as a 679.1871, M + Cs+ calcd for C33H4205Si 679.1856. yellowish oil: Rj = 0.57 (silica, 25% Et20 in petroleum ether); IR Acetonide 7. A solution of triol 5 (16.2 g, 29.6 mmol) and 2,2- (thin film) vmax2956, 2925,2849, 1773, 1467, 1101 cm-I; 'H NMR dimethoxypropane (18.2 mL, 148 mmol) in CH2Cl2 (98 mL) and Et20 (500 MHz, CDC13) 6 7.65-7.55 (band, 4 H, Ar), 7.45-7.25 (band, 11 (2 mL) was treated with camphorsulfonic acid (350 mg, 1.5 mol) H, Ar), 6.04 (dd, J= 10.0, 2.5 Hz, 1 H, 6-H), 5.82 (dd, J = 10.0, 2.5 and stirred at 25 "C for 7 h. After the reaction was quenched with 650 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 Nicolaou et al.

aqueous NaHC03 (50 mL), the organic layer was separated, dried (Naz- 28.6, 26.9, 26.7, 26.1, 26.0, 24.6, 19.4, 19.3, 19.2, 18.3, -5.3; FAB concentrated, and purified by flash chromatography (silica, 50% HRMS (NBNCsI) de983.4050 M + Cs+ calcd for C52H7406Si~ Et20 in petroleum ether) to give 7 (14.25 g, 82%) as a colorless oil: 983.4078. Rf = 0.51 (silica, 50% Et20 in petroleum ether); IR (thin film) vmax Epoxide 11. A solution of allylic alcohol 10 (18.7 g, 22.0 "01) 3467,2932,2858, 1462, 1373, 1210, 1106, 1054 cm-'; 'H NMR (500 in benzene (500 mL) was treated with 4-A molecular sieves (2 g), VO- MHz, CDC13) 6 7.66-7.60 (band, 4 H, Ar), 7.45-7.20 (band, 9 H, (acac)~(175 mg, 0.66 mmol), and t-BuOOH (22 mL of a 3 M solution Ar), 7.15-7.05 (band, 2 H, Ar), 5.79 (dd, J= 10.0, 1.5 Hz, 1 H, 6-H), in decane, 66.0 "01) and stirred at 25 "C for 14 h. After the reaction 5.72 (dd, J = 10.0, 2.5 Hz, 1 H, 5-H), 4.45 (d, J = 11.5 Hz, 1 H, was quenched with MezS (5 mL) and aqueous N&Cl (300 mL), the OCHZPh), 4.16 (d, 9.0 Hz, 1 H, 20-H), 4.11 (d, J = 11.5 Hz, 1 H, reaction mixture was extracted with Et20(200 mL). The organic layer OCHzPh), 3.99 (b S, 1 H, 7-H), 3.97-3.89 (band, 2 H, 2-CH2), 3.81 was dried (Na~S04),concentrated, and purified by flash chromatography (d, 9.0 Hz, 1 H, 20-H), 3.76 (A of AB, d, J= 10.5 Hz, 1 H, 9-H), 3.73 (silica, 15% Et20 in petroleum ether) to give 11 (16.6 g, 87%) as a (B of AB, d, J = 10.5 Hz, 1 H, 9-H), 3.42 (b t, J = 6.0 Hz, 1 H, colorless oil: Rf= 0.47 (silica, 15% EtzO in petroleum ether); IR (thin 2-OH), 2.14 (t, J = 4.0 Hz, 1 H, 3-H), 1.44 (s, 3 H, C(CH3)2), 1.42 (s, film) Y- 3490, 2935, 2852, 1471, 1257, 1049 cm-'; 'H NMR (500 3 H, C(CH&), 1.09 (s, 9 H, SiC(CH&Phz), 0.87 (s, 3 H, 19-CH3); 13C MHz, CDC13) 6 7.65-7.55 (band, 4 H, Ar), 7.50-7.28 (band, 11 H, NMR (125 MHz, CDCl3) 6 138.1, 135.9, 135.8, 132.9, 132.7, 132.5, Ar), 5.82 (d, J = 10.0 Hz, 1 H, 5-H), 5.74 (dd, J = 10.0, 5.0 Hz, 1 H, 129.9, 129.8, 128.2, 127.7, 127.4, 127.2, 126.6, 107.9, 81.9,75.9,71.3, 6-H), 4.82 (d, J = 4.5 Hz, 1 H, 2-H), 4.70 (d, J = 11.5 Hz, 1 H, 70.0, 67.1, 58.8, 48.1, 44.2, 27.3, 27.0, 26.4, 19.3, 14.1; FAB HRMS OCHzPh), 4.56 (d, J = 10.0 Hz, 1 H, 20-H), 4.54 (d, J = 11.5 Hz, 1 (NBA/NaI) m/e 609.3028, M + Na+ calcd for C36H4605si 609.3012. H, OCHzPh), 4.14 (A of AB, d, J = 11.5 Hz, 1 H, 10-H), 4.1 1 (B of AB, d, J = 11.5 Hz, 1 H, 10-H), 4.06 (d, J = 10.0 Hz, 1 H, 20-H), Aldehyde 8. A solution of alcohol (9.7 g, 16.5 mol) in CH3CN 7 3.85 (d, J = 10.0 Hz, 1 H, 9-H), 3.71 (d, J = 5.0 Hz, 1 H, 7-H), 3.54 (100 mL) was treated with tetrapropylammoniumpenuthenate (TPAP, (d, J = 10.0 Hz, 1 H, 9-H), 3.35 (d, J = 4.5 Hz, 1 H, 2-OH), 2.93 (s, 290 mg, 0.83 "01) and 4-methylmorpholine N-oxide (NMO, 2.91 g, 1 H, 14-H), 2.49 (b S, 2 H, 13-CH2), 1.80 (s, 1 H, 3-H), 1.70 (s, 3 H, 24.8 "01) and stirred at 25 "C for 2 h. After dilution with CHzCl2 18-CH3), 1.41 (s, 3 H, Ig-CHs), 1.30 (s, 3 H, C(CH3)2), 1.29 (s, 3 H, (400 mL), the reaction mixture was filtered through silica gel. The C(CH3)2), 1.25 (s, 3 H, C(CH3)2), 1.24 (s, 3 H, C(CH3)2), 1.06 (s, 9 H, resulting solution was concentrated and purified by flash chromatog- SiC(CH3)3Ph2), 0.90 (s, 9 H, SiC(CH3)3(CH3)2),0.08 (s, 3 H, SiC(CH3)3- raphy (silica, 30% Et20 in petroleum ether) to give 8 (9.37 g, 97%) as (cH3)2), 0.07 (s, 3 H, SiC(CH3)3(CH3)2); 13CNMR (125 MHz, CDCl3) a white foam: Rf = 0.45 (silica, 30% Et20 in petroleum ether); IR 6 137.8, 135.9, 135.6, 135.6, 135.4, 134.1, 133.7, 129.4, 129.3, 128.3, (thin film) vmax2931, 2857, 1720, 1472, 1428, 1371, 1111 cm-'; 'H 127.7, 127.4, 127.2, 123.9, 122.0, 107.1, 79.6, 74.3, 72.3, 70.8, 69.2, NMR (500 MHz, CDCls) 6 9.98 (d, J= 3.5 Hz, 1 H, 2-H), 7.65-7.55 64.1, 58.8, 53.4, 44.9, 42.3, 39.6, 31.7, 28.3, 26.9, 26.1, 25.9, 25.9, (band, 4 H, Ar), 7.47-7.22 (band, 9 H, Ar), 7.17-7.10 (band, 2 H, 25.8, 23.2, 21.9, 19.4, 19.3, 16.8, -5.5, -5.6; FAB HRMS (NBN Ar), 5.84 (dd, J= 10.5, 1.5 Hz, 1 H, 6-H), 5.71 (dd, J= 10.5, 2.0 Hz, CsI) de999.4050, M + Cs' calcd for C52H7407Si2 999.4027. 1H,5-H),4.50(d,J~11.5H~,1H,OCH~Ph),4.22(d,J~11.5H~,Diol 12. A solution of epoxide 11 (20.06 g, 23.1 "01) in Et20 1 H, OCHzPh), 4.20 (d, 9.5 Hz, 1 H, 20-H), 4.10 (dd, J= 2.0, 1.5 Hz, (100 mL) was treated with LiA1I-b (1 15 mL of a 1 M solution in EtzO, 1 H, 7-H), 3.84 (d, 9.5 Hz, 1 H, 20-H), 3.72 (A of AB, d, J= 10.0 Hz, 115 "01) and stirred at 25 OC for 7 h. After dilution with Et20 (200 1 H, 9-H), 3.70 (B of AB, d, J = 10.0 Hz, 1 H, 9-H), 3.18 (d, J = 3.5 mL), the reaction mixture was cooled to -78 "C, and the reaction was Hz, 1 H, 3-H), 1.42 (s, 3 H, C(CH3)z). 1.39 (s, 3 H, C(CH3)z). 1.09 (s, quenched with EtOAc (25 mL) followed by aqueous NI&Cl(100 mL). 9 H, SiC(CH&Phz), 1.04 (s, 3 H, 19-CH3); 13C NMR (125 MHz, After warming to 25 "C, the organic layer was separated and the CDC13)G 202.3, 138.1, 135.8, 135.8, 135.7, 135.6, 133.0, 132.9, 131.1, aqueous layer was extracted with Et20 (2 x 100 mL). The combined 129.7, 129.7, 129.5, 128.8, 128.2, 128.2, 127.6, 127.4, 127.4, 127.2, organic layers were dried (NazSOd), concentrated, and purified by flash 127.2, 127.1, 108.6, 80.7.75.4, 71.8,70.0,65.7, 57.6,44.9, 26.9, 26.5, chromatography (silica, 30% Et20 in petroleum ether) to give 12 (15.3 19.3, 13.6; FAB HRMS (NBA/NaI) mle 607.2865, M + Na+ calcd for g, 76%) as colorless crystals: mp 115-1 17 "C, from CHzCl2-hexanes; C36Hu05Si 607.2856. Rf = 0.58 (silica, 30% Et20 in petroleum ether); IR (thin film) Y,, Alcohol 10. To a solution of hydrazone 9 (28.2 g, 50.1 "01) in 3468,2955,2857, 1471, 1367, 1254, 1052 cm-'; IH NMR (500 MHz, THF (400 mL) at -78 OC was added dropwise n-BuLi (65.5 mL of a CDC13) 6 7.65-7.61 (band, 4 H, Ar), 7.42-7.28 (band, 11 H, Ar), 1.6 M solution in hexanes, 105 "01). After the reaction mixture was 5.85(d,J=10.5H~,lH,5-H),5.67(dd,J=10.5,5.0Hz, 1H,6-H), stirred at -78 "C for 20 min, it was allowed to warm to 0 OC, resulting 4.63 (d, J = 11.0 Hz, 1 H, OCHZPh), 4.55 (d, J = 10.0 Hz, 1 H, 20- in Nz gas evolution. The resulting bright orange solution was cooled H), 4.54 (d, J = 11.0 Hz, 1 H, OCHZPh), 4.18 (d, J = 4.5 Hz, 2-H), to -78 "C, and a solution of the aldehyde 8 (26.4 g, 45.1 mmol) in 4.16 (d, J= 11.0 Hz, 1 H, 10-H), 4.07 (d, J= 10.0 Hz, 1 H, 10-H), 3.97 (d, J = 4.5 Hz, 1 H, 2-OH), 3.87 (d, J = 11.0 Hz, 1 H, 20-H), THF (100 mL) was slowly added via canula. The reaction mixture = J = was stirred at -78 "C for 0.5 h, and then the reaction was quenched 3.79 (d, J 10.0 Hz, 1 H, 9-H), 3.64 (d, 5.0 Hz, 1 H, 7-H), 3.57 (d, J = 10.0 Hz, 1 H, 9-H), 3.22 (b S, 1 H, 1-OH), 2.23-2.04 (band, with aqueous N&Cl (50 mL). After being warmed to 25 "C, the 2 H, 13-CHz), 2.15 (s, 1 H, 3-H), 1.77-1.59 (band, 2 H, 14-CH2), reaction mixture was extracted with Et20 (2 x 200 mL). The organic 1.67 (s, 3 H, 18-CH3), 1.23 (s, 6 H, C(CH3)2), 1.19 (s, 3 H, 19-CH3), layer was dried (Na2S04), concentrated, and purified by flash chro- 1.07 (s, 3 H, C(CH&), 1.06 (s, 9 H, SiC(CH&Phz), 0.98 (s, 3 H, matography (silica, 15% Et20 in petroleum ether) to give 10 (31.7 g, C(CH&), 0.92 (s, 9 H, SiC(CH3)3(CH3)2),0.09 (s, 3 H, SiC(CH3)3- 82%) as a colorless oil: Rf = 0.25 (silica, 10% Et20 in petroleum ether); (CH3)2), 0.08 (s, 3 H, SiC(CH3)3(CH3)2); I3C NMR (125 MHz, CDC13) IR (thin film) v,, 3445, 2935, 2852, 1251, 1464, 1429, 1370, 1049 6 137.5, 136.3, 135.7, 135.6, 135.0, 133.9, 133.7, 129.9, 129.4, 129.3, cm-'; 'H NMR (500 MHz, CDC13) 6 7.73-7.65 (band, 4 H, Ar), 7.48- 128.3, 127.9, 127.7, 127.3, 122.6, 107.2, 79.5, 74.5, 74.3, 72.7, 72.6, 7.25 (band, 11 H, Ar), 5.98 (b s, 1 H, 14-H), 5.97 (d, J = 10.0 Hz, 1 71.1, 68.8, 59.5, 47.2, 44.3, 43.6, 29.9, 28.5, 27.8, 26.9, 26.7, 25.9, H, 5-H), 5.79 (dd, J = 10.0, 5.0 Hz, 1 H, 6-H), 4.88 (b S, 1 H, 2-H), 20.9, 19.3, 19.1, 19.0, 18.3, -5.4, -5.5; FAB HRMS (NBNCsI) m/e 4.73 (d, J= 11.5 Hz, 1 H, OCHzPh), 4.59 (d, J= 11.5 Hz, 1 H, OCH2- 1001.4170, M cs+ calcd for C52H7607Si~1001.4184. Ph), 4.45 (d, 9.5 Hz, 1 H, 20-H), 4.33 (d, J= 10.5 Hz, 1 H, 10-H), Carbonate 13. A solution of diol 12 (9.67 g, 11.1 "01) in Et20 4.29 (d, J = 3.5 Hz, 1 H, 2-OH), 4.24 (d, J = 10.5 Hz, 1 H, 10-H), (150 mL) and hexamethylphosphoramide(HMPA, 50 mL) was treated 3.96 (d, 9.5 Hz, 1 H, 20-H), 3.79 (d, J = 10.0 Hz, 1 H, 9-H), 3.72 (d, with KH (4.41 g of a 30% suspension in mineral oil, 33.0 "01, J = 10.0 Hz, 1 H, 9-H), 3.70 (d, J = 5.0 Hz, 1 H, 7-H), 2.80-2.65 prewashed with dry Et20) and stirred at 25 OC for 20 min, after which (band, 3 H, 3-H and 13-CH2), 1.81 (s, 3 H, 18-CH3), 1.43 (s, 3 H, phosgene (10 mL of a 20% solution in toluene, 17.5 "01) was added. C(CH3)2), 1.41 (s, 3 H, C(CH3)z). 1.35 (s, 3 H, 16-CH3), 1.32 (s, 3 H, The reaction mixture was stirred at 25 OC for 0.5 h. After dilution 17-CH3), 1.25 (s, 3 H, 19-C&), 1.11 (s, 9 H, SiC(CH&Ph2), 0.98 (s, with Et20 (300 mL), the reaction mixture was added to a half saturated 9 H, SiC(CH&(CH3)2), 0.15 (s, 6 H, SiC(CH3)3(CH3)2); 13C NMR (125 solution of tartaric acid. The organic layer was separated, washed with MHz, CDC13) 6 145.1, 137.5, 137.0, 135.7, 135.7, 135.1, 133.9, 133.7, brine (150 mL), dried (NazSOd), concentrated, and purified by flash 129.4, 129.4, 129.0, 128.4, 127.8, 127.7, 127.4, 127.4, 122.6, 120.7, chromatography (silica, 2% MeOH in CH2C12) to give diol 12 (4.06 g, 106.7, 80.2, 74.1, 72.4, 71.4, 70.9, 68.4, 59.1, 46.9, 43.3, 39.2, 33.6, 42%) and carbonate 13 (4.72 g, 86% based on 58% conversion) as a Total Synthesis of Tarol. 3 J. Am. Chem. Soc., Vol. 117, No. 2, 1995 651 yellow solid Rf = 0.64 (silica, 2% MeOH in CHZC12); IR (thin film) 0.5 h. After cooling to 25 "C, the reaction mixture was added to a v,, 2932, 2857, 1800, 1472, 1254, lo00 cm-'; 'H NMR (500 MHz, saturated solution of NaHC03 (100 mL), and the resulting mixture was cDc13) 6 7.63-7.58 (band, 5 H, Ar), 7.42-7.28 (band, 10 H, Ar), stirred at 25 "C for 2 h. The organic layer was separated, and the 5.85(dd,J~10.0,5.0Hz,1H,6-H),5.79(d,J~10.0Hz,1H,5-H),aqueous phase was extracted with EtOAc (3 x 75 mL). The combined 5.32 (s, 1 H, 2-H), 4.66 (d, J= 11.5 Hz, 1 H, OCHZPh), 4.36 (d, J= organic layer was dried (NaZSOd), concentrated, and purified by flash 11.5 Hz, 1 H, OCHzPh), 4.09 (A of AB, d, J = 11.5 Hz, 1 H, 20-H), chromatography (silica, 20 -. 40% EtOAc in petroleum ether) to give 4.06(BofAB,d,J=11.5Hz,lH,20-H),4.04(d,J=9.0Hz,lH, products 17 (65.3 mg, 25%), 18 (24.6 mg, lo%), 19 (104.4 mg, 40%), lO-H), 3.97 (d, J = 9.0 Hz, 1 H, 10-H), 3.73 (d, J = 10.5 Hz, 1 H, and 20 (40.5 mg, 15%). 9-H), 3.62 (d, = 5.0 Hz, 1 H, 7-H), 3.60 (d, J = 10.5 1 H, 9-H), J Hz, Diol 17: Rf = 0.41 (silica, 50% EtOAc in hexanes); IR (thin film) 2.42-2.02 (band, 4 H, 13-CHz and 14-CH2), 2.26 (s, 1 H, 3-H), 1.65 Y- 3490,2970,1789,1456,1100 cm-1; 'H NMR (500 MHz, CDc13) (s, 3 H, 18-CH3), 1.25 (s, 3 H, C(CH3)2), 1.24 (s, 3 H, C(CH&), 1.14 6 7.42-7.31 (band, 5 H, 5.97 (dd, 10.0, 1.5 1 H, 5-H), (s, 3 H, 19-CH3), 1.09 (s, 3 H, C(CH&), 1.07 (s, 9 H, SiC(CH3)3Ph2), Ar), J= Hz, 5.63 (dd, J = 10.0, 1.5 1 H, 6-H), 5.46 (d, 5.0 Hz, 1 H, 2-H), 1.03 (s, 3 H, C(CH3)2), 0.88 (s, 9 H, SiC(CH3)3(CH3)2),0.05 (s, 3 H, Hz, J= 4.77 (d, J= 12.0 Hz, 1 H, OCHZPh), 4.49 (d, J= 8.5 Hz, 1 H, 20-H), S~C(CH~)~(CH~)Z),0.03 (s, 3 H, SiC(CH3)3(CH&); NMR (125 MHz, 4.39 (d, J = 12.0 1 H, OCHZPh), 4.29 (b t, = 6.0 1 H, CDC13) 6 154.7, 138.7, 135.7, 135.6, 134.0, 133.7, 133.5, 132.5, 130.5, Hz, J Hz, lO-H), 4.24 (dd, J= 6.0, 3.0 1 H, 9-H), 3.80 (d, J= 8.5 Hz, 1 H, 129.5, 129.4, 128.0, 127.6, 127.4, 127.3, 125.2, 107.3, 88.2,79.7,78.9, Hz, 20-H), 3.58 (b S, 1 H, 7-H), 2.87 (d, = 3.0 1 H, 9-OH), 2.70 73.1, 71.2, 71.2, 70.4, 59.4, 46.5, 44.2, 43.4, 29.3, 27.9, 27.0, 26.6, J Hz, (ddd, J = 15.0, 10.5, 3.0 1 H, 14-H), 2.54 (ddd, = 20, 12.0, 3.0 25.8, 25.2, 19.3, 19.1, -5.6; FAB (NBNCsI) mle 1027.3950, Hz, J HRMS Hz, 1 H, 13-H), 2.31 (d, J = 5.0 Hz, 1 H, 3-H), 2.18 (d, = 6.0 Hz, M 4- cs+ calcd for C53H7408Si~1027.3977. J 1 H, lO-OH), 1.93 (ddd, J = 20.0, 10.5, 3.0 Hz, 1 H, 13-H), 1.78 Diol 14. A solution of carbonate 13 (4.72 g, 5.27 mol) in THF (ddd, J = 15.0, 12.0, 3.0 Hz, 1 H, 14-H), 1.56 (s, 3 H, 18-CH3), 1.42 (20 mL) was treated with n-BW 20 mL of a 1.0 M solution (TBAF, (S, 3 H, 19-CH3), 1.39 (s, 3 H, 16-CH3), 1.38 (s, 3 H, 17-CH3), 1.16 (s, in THF,20.0 mol)and stirred at 25 "C for 14 h. After dilution with 3 H, C(CHJ)Z),1.05 (s, 3 H, C(CH3)z); 13C NMR (125 MHz, CDC13) EtzO (50 mL), HzO (50 mL) was added. The organic layer was 6 153.9, 139.4, 137.3, 136.1, 135.6, 128.7, 128.5, 128.3, 122.0, 108.2, separated, washed with brine (30 mL), dried (NaZSOd), concentrated, 93.5, 82.4, 77.9, 75.7, 74.2, 71.2, 70.4, 69.3, 46.3, 44.3, 40.0, 31.2, and purified by flash chromatography (silica, 80% Et20 in petroleum 28.9, 27.9, 26.8, 23.6, 21.7, 21.3, 16.0; FAB (NBNCsI) de ether) to give 14 (2.29 g, 80%) as a white solid: Rf = 0.49 (silica, HRMS 673.1782, M -I- Cs+ calcd for c31bo8 673.1778. EtzO); IR (thin film) v,, 3438,2980,2879, 1778, 1371, 1061 cm-'; 'H NMR (500 MHz, CDCl3) 6 7.33-7.27 (band, 5 H, Ar), 5.99 (dd, Alkene 18: Rf = 0.95 (silica, 50% EtOAc in hexanes); IR (thin J = 10.0, 3.5 Hz, 1 H, 6-H), 5.89 (d, J = 10.0 Hz, 1 H, 5-H), 5.23 (s, film) Vmax 2971,1726 cm-'; 'H NMR (500 MHz, CDC13) 6 7.35-7.27 lH,2-H),4.72(d,J=ll.OHz, IH,OCHzPh),4.42(d,J=ll.OHz, (band, 5 H, Ar), 5.93 (dd, J = 10.5, 2.5 Hz, 1 H, 6-H), 5.86 (b d, J = 12.0 Hz, 1 H, lO-H), 5.56 (dd, J = 10.5, 1.5 Hz, 1 H, 5-H), 5.48 (d, 1 H, OCHzPh), 4.27 (b d, J = 9.0 Hz, 1 H, 10-H), 4.11 (b S, 2 H, 20-CHz), 4.02 (d, J = 9.0 Hz, 1 H, 10-H), 3.69 (dd, J = 9.5, 5.0 Hz, J = 12.0 Hz, 1 H, 9-H), 4.67 (d, J = 7.0 Hz, 1 H, 2-H), 4.65 (d, J = 1 H, 9-H), 3.49 (d, J = 3.5 Hz, 1 H, 7-H), 3.25 (dd, J = 9.5, 9.0 Hz, 10.5 Hz, 1 H, OCHZPh), 4.49 (d, J= 8.0 Hz, 1 H, 20-H), 4.44 (d, J= 1 H, 9-H), 2.77 (dd, J = 9.0, 5.0 Hz, 1 H, 9-OH), 2.40-2.18 (band, 4 10.5 Hz, 1 H, OCHzPh), 3.80 (d, J = 8.0 Hz, 1 H, 20-H), 3.68 (b S, 1 H, 13-CHz and 14-CHz), 2.37 (s, 1 H, 349, 1.72 (s, 3 H, 18-CH3), H, 7-H), 2.86 (d, J = 7.0 Hz, 1 H, 3-H), 2.35-2.22 (band, 3 H, 13- 1.47 (s, 3 H, C(CH3)2), 1.44 (s, 3 H, C(CH3)2), 1.08 (s, 3 H, 19-C&), CHZand 14-H), 1.96 (m, 1 H, 14-H), 1.54 (s, 6 H, 18-CH3 and 19- 1.05 (s, 3 H, C(CH&), 1.03 (s, 3 H, C(CH3)z); 13C NMR (125 MHz, CH3), 1.45 (s, 3 H, C(CH&), 1.39 (s, 3 H, C(CH&), 1.37 (s, 3 H, CDCls) 6 154.6, 136.8, 133.9, 133.5, 132.8, 128.2, 127.8, 127.6, 126.2, C(CH&), 1.07 (s, 3 H, C(CH3)z); I3C NMR (125 MHz, cW13) 6 149.4, 106.7, 88.4, 80.5, 78.8, 74.5, 71.6, 71.2, 67.9, 58.6, 44.3, 44.2, 43.5, 143.2, 137.6, 137.3, 133.4, 128.7, 128.2, 128.1, 127.7, 125.3, 122.0, 29.2, 27.2, 26.2, 24.5, 23.7, 20.2, 19.1, 18.5; FAB HRMS (NBNCsI) 108.4, 90.6, 81.7, 75.7, 72.0, 71.0, 62.3, 47.5, 43.7, 36.3, 29.7, 29.1, mle 675.1942, M + Cs+ calcd for C31b208 675.1934. 26.8, 26.6, 26.4, 24.4, 16.1, 14.4; FAB HRMS (NBNCsI) mle Dialdehyde 15. A solution of diol 14 (0.66 g, 1.22 "01) and 639.1736, M + Cs+ calcd for C31H3806 639.1723. 4-methylmorpholine N-oxide (NMO, 0.43 g, 3.67 "01) in CH3CN Hemiacetal 19: mp 170-174 "C, 195-200 "C (corresponding (40 mL) and CHzClz (20 mL) was treated with 4-A molecular sieves aldehyde), from CHzClz-hexanes; Rf = 0.51 (silica, 50% EtOAc in (50 mg) and stirred at 25 "C for 10 min. Tetrapropylammonium hexanes ); IR (thin film) vm, 3422, 2924, 1797, 1454, 1381, 1216, permthenate ("PAP,22 mg, 0.062 "01) was added, and the reaction 1052 cm-l; 'H NMR (500 MHz, CDCl3) 6 7.35-7.30 (band, 5 H, Ar), mixture was stirred at 25 "C for 2 h. After dilution with CHzClz (100 6.05 (dd, J= 10.5, 1.0 Hz, 1 H, 5-H), 5.71 (dd, J= 10.5, 1.0 Hz, 1 H, mL), the reaction mixture was filtered through silica gel. The resulting 6-H), 5.57 (d, J = 2.0 Hz, 1 H, 10-H), 5.20 (d, J = 8.5 Hz, 1 H, 2-H), solution was concentrated to give dialdehyde 15 (0.611 g, 92%) as a 4.67 (d, J= 11.5 Hz,1 H, OCHzPh). 4.45 (d, J= 11.5 Hz, 1 H, OCHz- white solid: Rf = 0.70 (silica, 50% EtOAc in hexanes); IR (thin film) Ph), 4.27 (d, J = 8.5 Hz, 1 H, 20-H), 4.26 (s, 1 H, 9-H), 3.97 (b S, 1 vmax2919,1793, 1724, 1669,1063 cm-'; 'H NMR (500 MHz, (CD3)2- H, 7-H), 3.90 (d, J = 8.5 Hz, 1 H, 20-H), 3.19 (d, J = 8.5 Hz, 1 H, CO) 6 10.98 (S, 1 H, 10-H), 9.40 (s, 1 H, 9-H), 7.39-7.29 (band, 5 H, 3-H), 2.42 (d, J = 2.0 Hz, 1 H, 11-H), 2.30-1.85 (band, 4 H, 13-CHz Ar), 6.25 (dd, J = 10.0, 4.5 Hz, 1 H, 6-H), 5.84 (d, J = 10.0 Hz, 1 H, and 14-CHz), 1.51 (s, 3 H, 16-CH3), 1.49 (s, 3 H, I7-CH3), 1.32 (s, 3 5-H), 5.35 (d, J = 2.5 Hz, 1 H, 2-H), 4.81 (d, J = 11.0 Hz, 1 H, H, C(CH~)Z),1.24 (s, 3 H, C(CH3)2), 1.11 (s, 3 H, 18-CH3), 1.07 (s, 3 OCHzPh), 4.56 (d, J = 11.0 Hz, 1 H, OCHzPh), 4.28 (d, J = 4.5 Hz, H, NMR (125 MHz, CDC13) 6 153.2, 137.2, 134.0, 128.4, 1 H, 7-H), 3.97 (s, 2 H, 20-CHz). 2.91 (d, J = 2.5 Hz, 1 H, 3-H), 2.65 128.0, 127.9, 124.0, 108.0,98.4,89.6,82.5,77.9,74.8,71.6,69.6,62.6, (m, 1 H, 13-H), 2.52-2.46 (band, 2 H, 13-H and 14-H), 2.23 (m, 1 H, 45.3, 43.9, 42.2, 38.5, 38.1, 30.2, 29.0, 27.1, 26.4, 25.9, 20.3, 15.7; 14-H), 2.16 (s, 3 H, 18-CH3), 1.41 (s, 3 H, Ig-CHs), 1.29 (s, 3 H, FAB HRMS (NBNCsI) de673.1760, M + Cs+ calcd for c31bo8 C(CH3)z), 1.25 (s, 3 H, C(CH3)2), 1.21 (s, 3 H, C(CH3)z), 1.15 (s, 3 H, 673.1778. C(CH3)z); 13C NMR (125 MHz, (CD3)zCO)) 6 198.9, 192.2, 155.2, Formate Ester 20: mp 222-224 "C, from CHzClz-hexanes; Rf = 154.2, 139.5, 137.5, 133.3, 129.0, 128.4, 128.4, 109.3, 89.9,80.4,77.0, 0.59 (silica, 50% EtOAc in hexanes); IR (thin fi)vmax 2986, 1799, 72.6, 72.5, 72.2, 53.8, 46.4, 43.4, 32.4, 27.3, 26.8, 25.2, 24.1, 18.8, 1728, 1383, 1139, 1058, cm-l; *HNMR (500 MHz, CDCl3) 6 7.89 (s, 18.6, 17.7; FAB HRMS (NBNCsI) mle 671.1630, M + Cs+ calcd for 1 H, 9-CHO), 7.41-7.32 (band, 5 H, Ar), 6.11 (dd, J= 10.0, 1.5 Hz, C31H3808 671.1621. 1 H, 5-H), 5.71 (dd, J = 10.0, 1.0 Hz, 1 H, 6-H), 5.54 (s, 1 H, 9-H), 8-Membered Ring Intermediates 17-20. TiC13.(DME)l.5 (1.53 g, 5.16(d,J=9.0Hz,lH,2-H),4.73(d,J=11.5H~,lH,OCH2Ph), 5.3 "01) and ZdCu couple (1.66 g, 12.7 mol) were transferred to 4.52 (d, J = 11.5 Hz, 1 H, OCHzPh), 4.30 (d, J = 8.5 Hz, 1 H, 20-H), a dry flask under argon (glovebag). The mixture was further dried at 4.09 (b S, 1 H, 7-H), 3.89 (d, J = 8.5 Hz, 1 H, 20-H), 3.42 (d, J = 9.0 140 "C, under vacuum for 10 min. Freshly distilled DME (70 mL) Hz, 1 H, 3-H), 2.42-2.22 (band, 4 H, 13-CHz and 14-CHz), 1.52 (s, 3 was then added, and the suspension was stirred at reflux for 3.5 h. H, C(CH3)2), 1.50 (s, 3 H, C(CH3)2), 1.37 (s, 3 H, 16-CH3), 1.28 (s, 3 After the mixture was cooled to 70 "C, a solution of dialdehyde 15 H, 17-CH3), 0.99 (s, 3 H, 18-CH3), 0.89 (s, 3 H, 19-CH3); NMR (260 mg, 0.48 "01) in DME (25 mL) was added via syringe pump (125 MHz, CDCl3) 6 211.4, 158.4, 152.3, 136.6, 134.3, 128.6, 128.5, over 1 h. The reaction mixture was stirred at 70 "C for an additional 128.2, 123.6, 108.4, 98.5, 88.1, 82.2,77.6,77.5, 75.5, 71.5,69.5, 52.0, 652 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al.

50.6, 47.0, 43.6, 29.5, 28.9, 27.1, 25.4, 24.6, 24.6, 18.9, 15.4; FAB 1059 cm-'; 'H NMR (500 MHz, CDC13) 6 7.35-7.30 (band, 5 H, Ar), HRMS (NBNCsI) mle 687.1570, M + Cs+ calcd for C31H3809 5.96 (dd, J = 10.0, 1.5 Hz, 1 H, 6-H), 5.85 (d, J= 5.5 Hz, 1 H, 9-H), 687.1570. 5.63 (dd, J = 10.0, 1.0 Hz, 1 H, 5-H), 5.53 (d, J = 4.5 Hz, 1 H, 2-H), Camphanate Esters 29 and 30. A solution of diol 17 (42 mg, 4.71 (d, J = 12.0 Hz, 1 H, OCHZPh), 4.48 (d, J = 8.0 Hz, 1 H, 20-H), 0.077 mmol) and Et3N (0.217 mL, 1.5 "01) in CHzClz (3.5 mL) was 4.46 (d, J = 12.0 Hz, 1 H, OCH2Ph). 4.33 (dd, J = 5.5, 3.0 Hz, 1 H, treated with a catalytic amount of 4-(dimethylamino)pyridhe (DMAP, lO-H), 3.79 (d, J = 8.0 Hz, 1 H, 20-H), 3.74 (b S, 1 H, 7-H), 2.77 0.5 mg, 0.004 mmol) and (19-(-)-camphanic chloride (84 mg, 0.388 (ddd, J = 14.0, 10.5, 3.0 Hz, 1 H, 13-H), 2.68-2.55 (band, 1 H, 14- mmol) at 25 "C for 1 h. After dilution with Et20 (10 mL), the reaction H), 2.58 (d, J = 3.0 Hz, 10-OH), 2.48 (ddd, J = 13.5, 10.5, 4.0 Hz, 1 was quenched with aqueous NaHC03 (5 mL), and the resulting mixture H, CH(H)CHz camph.), 2.36 (d, J = 4.5 Hz, 1 H, 3-H), 2.15-1.92 was stirred at 25 "C for 15 min. The organic layer was separated, and (band, 3 H, 13-H and CH(H)CH(H) camph.), 1.90-1.65 (band, 2 H, the aqueous layer was extracted with CHzClz (3 x 10 mL). The 14-H and CH(H)CHz camph.), 1.72 (s, 3 H, 18-CH3), 1.57 (s, 3 H, combined organic layer was washed with brine (10 mL), dried (MgSOd), OC(O)C(CH3)), 1.44 (s, 3 H, (O)ZC(CH~)Z),1.42 (s, 3 H, (O)zC(CH3)z), concentrated, and purified by preparative TLC (silica, 20% EtOAc in 1.14 (s, 6 H, C(CH3)z cmph.), 1.1 1 (s, 3 H, 16-CH3), 1.08 (s, 3 H, benzene) to give camphanic esters 29 and 30 (23 and 25 mg, 17-CHs), 0.98 (s, 3 H, 19-CH3); 13C NMR (125 MHz, CDCl3) 6 177.8, respectively, 86% combined yield) as white solids. 166.2, 153.8, 143.6, 137.1, 135.4, 132.8, 128.6, 128.3, 128.2, 122.3, Ester 29: Rf = 0.26 (silica, 15% EtOAc in benzene); +117 108.3, 93.4, 91.5, 82.4, 77.9, 75.2, 74.1, 73.6, 71.2, 71.1, 54.8, 54.4, (c 0.54, CHC13); IR (thin film) vmax3500, 2970, 2930, 1792, 1744, 47.1, 44.7, 39.7, 31.4, 31.1, 29.0, 28.8, 27.8, 26.9, 23.5, 21.7, 21.5, 1458, 1103, 1058, 914 cm-'; IH NMR (500 MHz, CDC13) 6 7.33- 17.1, 16.8, 16.1, 9.6; FAB HRMS (NBNCsI) mle 853.2543, M + Cs+ 7.24 (band, 5 H, Ar), 5.94 (dd, J 10.5, 1.5 Hz, 1 H, 6-H), 5.74 (d, calcd for C41HSZOll 853.2564. J = 5.0 Hz, 1 H, 9-H), 5.63 (dd, J = 10.5, 1.0 Hz, 1 H, 5-H), 5.51 (d, Diol (+)-17. A solution of ester 29 (23 mg, 0.032 "01) in MeOH J = 4.5 Hz, 1 H, 2-H), 4.70 (d, J = 12.0 Hz, 1 H, OCHzPh), 4.64 (d, (3.5 mL) was treated with KzCO3 (3.0 mg, 0.22 mmol) and stirred at J = 8.5 Hz, 1 H, 20-H), 4.45 (d, J = 12.0 Hz, 1 H, OCHzPh), 4.36 25 "C for 0.5 h. After dilution with CHzClz (15 mL), the reaction was (dd, J= 5.0,3.0 Hz, 1 H, 10-H), 3.78 (d, J= 8.5 Hz, 1 H, 20-H), 3.70 quenched with aqueous mC1 (10 mL). The organic layer was (b S, 1 H, 7-H), 2.72 (ddd, J = 14.0, 10.0, 3.5 Hz, 1 H, 13-H), 2.63- separated, and the aqueous layer was extracted with CHzCl2 (3 x 10 2.53 (band, 1 H, 14-H), 2.56 (d, J = 3.0 Hz, 10-OH), 2.38 (ddd, J = mL). The combined organic layer was dried (MgSOd), concentrated, 14.0, 11.0,4.0 Hz, 1 H, CH(H)CHz camph.), 2.33 (d, J= 4.5 Hz, 1 H, and purified by flash chromatography (silica, 25 - 50% EtOAc in 3-H), 2.12-1.88 (band, 3 H, 13-H and CH(H)CH(H) Camph.), 1.81 petroleum ether) to give diol (+)-17(15.5 mg, 90%) as a white solid: (ddd, J= 14.5, 12.0, 2.5 Hz, 1 H, 14-H), 1.71 (ddd, J= 13.5,9.0, 4.0 [alzZ~+187 (c 0.5, CHC13). Hz, 1 H, CH(H)CH2 camph.), 1.62 (s, 3 H, 18-CH3), 1.57 (s, 3 H, Acknowledgment. We thank Drs. Dee H. Huang, Raj OC(O)C(CHd), 1.41 (s, 3 H, (O)zC(CH3)z), 1.40 (s, 3 H, (0)zC(CH3Mr 1.12 (s, 6 H, C(CH& camph.), 1.10 (s, 3 H, 16-CH3), 1.06 (s, 3 H, Chadha, and Gary Siuzdak for the NMR,X-ray crystallographic 17-CH3). 1.00 (s, 3 H, 19-CH3); NMR (125 MHz, CDC13) 6 178.0, analyses, and mass spectroscopy, respectively. This work was 166.2, 153.8, 143.6, 137.1, 135.5, 132.7, 128.7, 128.5, 128.3, 122.1, supported by the NIH, The Scripps Research Institute, fellow- 108.4, 93.4, 90.8, 82.5, 78.0, 74.9, 74.0, 74.0, 71.2, 70.9, 54.8, 54.3, ships from RhBne-Poulenc Rorer (P.G.N.), The Office of Naval 47.2, 44.8, 39.8, 31.5, 30.9, 29.0, 28.8, 28.0, 26.9, 23.6, 21.7, 21.7, Research (R.K.G.), Glaxo, Inc. (C.F.C.), Mr. Richard Staley 16.8, 16.8, 16.2,9.6; FAB HRMS (NBNCsI) mle 853.2545, M + Cs+ (C.F.C.), NSERC (J.R.), and grants from Merck Sharp and calcd for C~IHSZOII853.2564. Dohme, Pfizer, Inc., Schering Plough, and the ALSAM Ester 30. colorless crystals, mp 240 "C, dec, from CH2Cl~-hexanes; Foundation. Rj = 0.21 (silica, 15% EtOAc in benzene); [aIz2~133 (c 0.49, CHC13); IR (thin film) vmax3498,2976,1793,1742,1457,1378,1265, JA942194M J. Am. Chem. SOC. 1995,117, 653-659 653

Total Synthesis of Taxol. 4. The Final Stages and Completion of the Synthesis

K. C. Nicolaou,* H. Ueno, J.-J. Liu, P. G. Nantermet, Z. Yang, J. Renaud, K. Paulvannan, and R. Chadha Contribution from the Department of Chemistry, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, Califomia 92037, and Department of Chemistry and Biochemistry, University of Califomia, San Diego, 9500 Gilman Drive, La Jolla, Califomia 92093 Received July 7, I994@

Abstract: The total synthesis of (-)-Tax01 has been achieved. Functional group manipulation of diol 2 provided the ABC ring system with the correct C9-keto, C10-acetyloxy functionality. Careful optimization allowed the oxidation of the C5-C6 alkene in 4 at C5 via a hydroboration reaction. Functional group manipulation of this product, 29, provided, through two routes, the oxetane D ring as 36. Following the method developed by degradative studies provided the natural enantiomer of Taxol (1).

Introduction With a route to optically active diol 2 secured,' a total synthesis of Taxol (1, Figure 1) looked quite feasible. However, several issues still remained to be addressed before the final goal could be reached. Amongst them were the functional group adjustments at C9 and C10, the installment of an oxygen at the C5 position, oxetane construction, oxygenation at C13, and side- 1: Taxol chain attachment. Below we describe solutions to these Figure 1. Structure and numbering of Taxol (1). problems and, thus, the total synthesis of Taxol (1). Scheme 1. Functionalization of the C9 and C10 Positions of Final Stages of the Total Synthesis the Taxoid Frameworka a. Selective Functionalization at C9 and C10. Continuing the sequence from diol 2l (Scheme l), our strategy toward Taxol (1) next called for the adjustment of the functional groups at C9 and C10 to their final form. Arriving at the desired C9- keto, C10-acetate functionality required differentiating between the two hydroxyl groups of diol 2. Fortunately, the higher reactivity of the allylic C10 hydroxyl group provided high 3 selectivity in the desired direction when compound 2 was exposed to 1.5 equiv of AczO and DMAP in methylene chloride. The resulting monoacetate 3 (Scheme 1, 95% yield) was OI oxidized cleanly with TPAP-NM02 to afford, in 93% yield, the desired 9-keto, 10-acetate 4. The absence of a conjugated enone in 4 (as observed in the 13C NMR) and the detection of long-range coupling (J < 1.5 Hz) between the C10 proton (6 5.65, CDCl3, 500 MHz) and the C12 methyl group (6 1.68) of monoacetate 3 ('H NMR decoupling experiments) suggested the indicated regiochemistry of these intermediates. This assignment was confiied by X-ray crystallographic analysis 5 4 of benzoate 5, obtained by PCC oxidation3 of compound 4 (see "Reagents and conditions: (a) 1.5 equiv of AczO, 1.5 equiv of Scheme 1, and ORTEP drawing in Figure 2). This regioselec- 4-(dimethylamino)pyridine (DMAP),CH*C12,25 "C, 2 h, 95%; (b) 0.1 tivity is in contrast to that observed in the exclusive formation equiv of tetrapropylammonium perruthenate (TPAP),3.0 equiv of of the 9-camphonate ester described in the preceding paper,' in 4-methylmorpholine N-oxide (NMO),CH3CN, 25 "C, 2 h, 93%; (c) which a speculative explanation for this discrepancy is proposed. 30 equiv of pyridinium chlorochromate (PCC), 50 equiv of NaOAc, It was now time to address the introduction of an alcohol at Celite, benzene, reflux, 1 h, 50%. Bn = CH2Ph. c5. b. Early Attempts to Hydroborate the C5-C6 Double * Address correspondence to this author at The Scripps Research Institute Bond. Our experience with the hydroboration of ring C or the University of California. systems4s5 led us to adopt similar tactics for the real system. @ Abstract published in Advance ACS Abstracts, December 15, 1994. (1) Nicolaou, K. C.; Yang, Z.; Liu, J.-J.; Nantermet, P. G.; Claibome, Potential differentiation of the two faces of the double bond in C. F.; Renaud, J.; Guy, R. K.; Shibayama, K. J. Am. Chem. SOC. 1995, 117, 645. (4)Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorensen, E. J.; (2)Griffith, W. P.; Ley, S. V. Aldrichimica Acta 1990, 23 (l), 13. Claibome, C. F.; Guy, R. K.; Hwang, C.-K.; Nakada, M.; Nantermet, P. G. (3) Angyal, S. J.; James, K. Carbohydr. Res. 1970, 12, 147. J. Am. Chem. SOC. 1995, 117, 634. 0002-786319511517-0653$09.00/0 0 1995 American Chemical Society 654 J. Am. Chem. Soc., Vol. 117, No. 2, 1995 Nicolaou et al.

Scheme 2. Hydroboration Studies 1"

0

t / &I(--p 0

0 5

0 0 & 8 Figure 2. ORTEP diagram for benzoate 5. .i ring C of intermediate 4 by an incoming reagent was not obvious by inspection of molecular models. It was, therefore, decided to initially explore the utilization of the C20 hydroxyl group as a handle to direct hydroboration from the p face of the molecule and at the C5 position as in the simple C ring case. To this end the acetonide group was removed from 4 under acid conditions to afford diol 6 (Scheme 2, 88% yield based on 53% a 0 conversion). Attempts to hydroborate6 6 under a variety of 11 conditions failed, presumably due to the formation of a stable 5-membered ring borane complex involving the two hydroxyl a Reagents and conditions: (a) 3.0 equiv of p-toluenesulfonic acid, groups that is both unable to reach the internal alkene and MeOH, 25 "C, 48 h, 88% based on 53% conversion; (b) 5.0 equiv of prohibitively bulky for external hydroboration. Dess-Martin periodinane, CHZC12,25 "C, 3 h, then 20 equiv of AczO, 25 equiv of 4-(dimethylamino)pyridine (DMAP), CH2C12, 25 "C, 3 h, We next considered using the 4-acetoxy, 20-hydroxy com- then 2.0 equiv of n-BW&, THF, 25 "C, 1 h, 66% from 6; (c) 1.2 pound 7 (Scheme 2) as a possible substrate for the desired equiv of MsCl, 3.0 equiv of DMAP, CH2C12, 25 "C, 1 h, 94%; (d) 10 hydroboration reaction, but unfortunately, all attempts to prepare equiv of AczO, 15 equiv of DMAP, CH2C12, 25 "C, 5 h, 90%; (e) 10 this intermediate met with failure. Under the various conditions equiv of BHs-THF, THF, 25 "C, 2 h, then excess H202, saturated used, the acetate group migrated facilely from the C-4 to the aqueous NaHC03, 0.5 h, 67%. Bn = CH2Ph, Ms = S02CH3. C20 al~ohol,~leading to either the primary acetate 8 or the conditions, however, smoothly converted the corresponding starting diol 6 rather than the desired tertiary acetate 7. It acetate 8 (obtained conveniently by monoacetylation of diol 6) became clear that the acetate at C4 would have to be installed to the C4 benzyloxy derivative 13 (76% yield). Preparation of after oxetane formation or in an intermediate in which the C20 17 from 13 required complete deacetylation under basic hydroxy group would remain blocked until oxetane ring closure. hydrolysis conditions, followed by selective silylation at C20 We, therefore, turned to the C4 acetate, C20 mesylate 10, (triethylsilyl group), acetylation at C10, and desilylation of the prepared from diol 6 by sequential mesylation (94% yield) and C20 hydroxyl group (52% overall yield). Again, hydroboration acetylation (90% yield) as detailed in Scheme 2. Hydroboration of 17 was disappointing: the major product was the C4 deoxy of this compound (10) with borane in THF, however, resulted compound 18. Hydroboration of the C4-benzyloxy, C20- not only in hydroxylation at C5 but also in concomitant reductive mesylate 19, obtained through mesylation of 17, also failed: cleavage of the C4 acetate to afford compound 11 as the major exhibiting sluggish reactivity and undesirable products. product (67% yield) whose stereochemistry at both the C4 and The unwillingness of the C4-benzyloxy mesylate 19 to enter C5 centers was left unassigned. Similar observations have facilely into hydroboration reactions prompted us to attempt this previously been reported with simple allylic derivatives8 reaction on the sterically less demanding C4-hydroxy, C20- In order to lower the propensity of the C4 substituent toward mesylate 9 (Scheme 4). Thus, exposure of 9 to excess borane reductive elimination, the 4-benzyloxy compounds 17 and 19 in THF followed by oxidative workup resulted in the formation (Scheme 3) were chosen as the next potential candidates for of diol 20 as the major product and in 23% yield. The indicated hydroboration. Exposure to KH and benzyl bromideg failed to a stereochemistry of the newly introduced C5 hydroxyl group convert mesylate 9 to the desired benzyl ether 19, leading instead was based on 'H NMR data and was confirmed by chemical to the formation of epoxide 12 (Scheme 3). The same correlation as outlined in Scheme 4. Thus, treatment of 20 with (5)Nicolaou, K. C.; Liu, J. J.; Hwang, C.-K.; Dai, W.-M.; Guy, R. K. Et3N, DMAP, and AczO resulted in acetylation and intramo- J. Chem. SOC., Chem. Commun. 1992, 1118. lecular displacement of the mesylate group to give epoxide 21 (6)Smith, K.; Pelter, A. In Comprehensive Organic Synthesis; Trost, B. (75% yield), which was debenzylated by hydrogenolysis, leading M., Fleming, I., Eds.; Pergamon Press: New York, 1991; Vol. 8, p 703. to compound 22 (95% yield). The latter compound was (7) Samaranavake. G.: Mah.-. N. F.: Jitranesri.-.. C.: Kineston. - . D. G. I. J. Org.'Chem. 1991, 56, 5114. identical with a sample prepared from 10-deacetylbaccatin 111 (8) Brown, H. C.; Knights, E. F. J. Am. Chem. SOC.1968, 90, 4439. (23) through intermediate 24l0 by the following short sequence: Pasto. D. J.: Hickman. J. 7. Am. Chem. SOC.1968. 90, 4445. (9) Kanai, K.; Sakamoto, I.; Ogawa, S.; Suami, T. Bull. Chem. SOC.Jpn. (a) exposure of 24 to Meerwein's reagent7 leading to 25 (59%) 1987, 60, 1529. and 26 (19%); (b) mesylation of the minor product (26) to give Total Synthesis of Taxol. 4 J. Am. Chem. SOC., Vol. 11 7, No. 2, 1995 655 Scheme 3. Hydroboration Studies 2a Scheme 4. Chemical Corrolation Studiesu

0 0 8 12

0 OMS S:R=H 0 0 bC8:R=Ac 20 24

0 0 21 ci If

18 18 Reagents and conditions: (a) 1.3 equiv of KH, 5.0 equiv of PhCHZBr, 0.05 equiv of n-BWI, EtzO, HMPA, 25 "C, 15 min, 79%; (b) 1.2 equiv of Ac20, 1.5 equiv of 4-(dimethylamino)pyridine (DMAP), 22 28 CHzC12, 25 "C, 20 min, 95%; (c) 5.0 equiv of KH, 15 equiv of PhCHzBr, 0.05 equiv of n-BWI, EtzO, HMPA, 25 "C, 4 h, 76%; (d) Reagents and conditions: (a) 10 equiv of BH3*THF,THF, 25 "C, 10 equiv of DBU, MeOH, CH2Cl2, 25 "C, 3 h, 98%; (e) 1.2 equiv of 1.5 h, then excess H202, aqueous NaHCOs, 0.5 h, 23%; (b) 30 equiv Et3SiC1, 1.5 equiv of DMAP, DMF, 25 "C, 1 h; (f) 6.0 equiv of AczO, of Ac20, excess EtsN, 0.05 equiv of 4-(dimethylamino)pyridine 6.0 equiv of DMAP, CH2C12, 25 "C, 0.5 h; (g) HF-pyridine, THF, 25 (DMAP), CHzC12,25 "C, 12 h, 75%; (c) H2, Pd(OH)dC, EtOH, 25 "C, "C, 1 h, 52% from 14; (h) 5.0 equiv of BH3*THF, THF, 0 "C, 0.5 h, 25 1 h, 95%; (d) 2.1 equiv of EtsOBF4, CH2C12,O "C, 1 h, 59% of 25 plus "C, 4 h, then excess H202, aqueous NaHCO3, 0.5 h; (i) 3.0 equiv of 19% of 26 (e) 10 equiv of MsC1,20 equiv of Et3N, 2.0 equiv of DMAP, MsCl, 5.0 equiv of DMAP, CHzC11, 25 "C, 1 h, 94%. DBU = 1,s- CH2C12, 25 "C, 0.5 h, 90%; (f) excess KH, THF, 25 "C, 0.5 h, 92%; diazabicyclo[5.4.0]undec-7-ene,Bn = CHZPh, Ms = S02CH3, TES = (g) HFTyridine, THF, 25 "C, 1 h, 98%. Bn = CHZPh, Ms = S02CH3, SiEt3. TES = SiEt3. indicated that the a face was somewhat less hindered than the 27; (c) treatment with KH in THF to form the epoxide ring; p face, although the absence of a free hydroxy handle in the and (d) exposure to HFpyridine to remove the silyl group vicinity of the C5 position raised questions regarding the (81% overall yield from 26) to afford 22. This chemical regiochemical outcome of the intended hydroboration. In the correlation firmly established the regio- and stereoselectivity event, exposure of 4 to excess borane in THF followed by the of the hydroboration reaction of 9. A better candidate was, usual oxidative workup furnished a mixture of the C5a-alcohol however, needed to serve as a precursor to the desired oxetane 29 (42% yield based on 83% conversion) and its C6 regioisomer system. (22% yield based on 83% conversion) (Scheme 5). While the c. Final Hydroboration Route to a C5a-Hydroxy Inter- stereochemistry of the C6 regioisomer remains unassigned, that mediate. Having realized that the C5a-hydroxy compounds of the C5 a-isomer was confirmed by conversion to intermediate might be a more accessible series of precursors to the oxetane 32, previously obtained from 10-deacetylbaccatin 111 (23) via system, we decided at this point to examine the hydroboration desilylation of intermediate 25 (Schemes 4 and 5). Thus, acid- of acetonide 4 (Scheme 5). Inspection of molecular models catalyzed removal of the acetonide group from 29 afforded triol

~~~ ~ (10) Nicolaou, K. C.; Nantermet, P. G.; Ueno, H.; Guy, R. K.; 30 (80% yield based on 88% conversion). Under carefully Couladouros, E. A.; Sorensen, R. J. J. Am. Chem. SOC. 1995, 11 7, 624. controlled conditions, acetylation of the primary hydroxyl group 656 J. Am. Chem. SOC., Vol. 117, No. 2, I995 Nicolaou et al.

Scheme 5. Hydroboration Studies 3. Successful Scheme 6. Construction of the Oxetane Ring" Hydroboration of the C5-C6 Double Bond"

,632:ReH 33 4 20 25: R = TES bl

wsOMs'"OH W:OR"'OH

'Vo OAC 'k6 OTMS 0 0 0 0 37 34:R=H dG 35 : R = Tf 32:R=H 30:R=H er25 : R = TES 'L31:R=Ac a Reagents and conditions: (a) 10.0 equiv of BH3*THF,THF, 0 "C, hl 3 h, then excess HzOz, saturated aqueous NaHC03, 25 "C, 1 h, 42% plus 22% of C6-OH regioisomer, based on 83% conversion; (b) MeOH: concd HC1 (2:l) 25 "C, 5 h, 80% based on 88% conversion; (c) 1.25 equiv of AczO, 5.0 equiv of pyridine, 0.05 equiv of 4-(dimethylami- no)pyridine (DMAP), CHzC12,25 "C, 0.5 h, 95%; (d) Hz, 10% Pd(OH)2/ C, EtOAc, 25 OC, 0.5 h, 97%; (e) HFpyridine, THF, 25 "C, 2 h, 96%. Bn = CHZPh, TES = SiEt3. 0 0 in 30 proceeded selectively to afford monoacetate 31 (95% 38 36:RrH yield). Finally, hydrogenolysis of the benzyl group from 31 fL24 : R = AC furnished alcohol 32, identical to material obtained from Reagents and conditions: (a) 25 equiv of EtsSiCl, pyridine, 25 "C, 12 h, 85%; (b) 10 equiv of KzC03, MeOH-HZ0,O "C, 15 min, 97%; desilylation (HFpyridine, THF, 96% yield) of 25 in all respects (c) 10 equiv of Me3SiC1, 30 equiv of pyridine, CHZClz, 0 "C, 15 min; including absolute stereochemistry (synthetic: [aIz2~-85.2 (c (d) 15 equiv of TfzO, 30 equiv of i-PrzNEt, CHzClz, 0 "C, 0.5 h; (e) 0.115, CHC13); degradative: -85.6 (c 0.43, CHCl3). With 0.05 equiv of camphorsulfonic acid (CSA), MeOH, 25 "C, 15 min, the synthesis of 32, the road to Taxol (1) was now open. then silica gel, CHZClz, 25 "C, 1 h, 40% from 33; (f) 8.0 equiv of d. Installation of the Oxetane Ring and Completion of AczO, 15 equiv of 4-(dimethylamino)pyridine (DMAP), CHzC12, 25 the Total Synthesis. The last remaining challenge in the total "C, 4 h, 94%; (g) 10 equiv of MsCl, 20 equiv of DMAP, CHZCL, 25 synthesis of Taxol (l),namely the construction of the oxetane "C, 1 h, 73%; (h) 10 equiv of KzCO3, MeOH, HzO, 0 "C, 15 min; (i) 12 equiv of n-BWOAc, butanone, reflux, 5 h, 72% from 37. TES = ring, was accomplished following two routes which were based SiEt3, TMS = SiMe3, Tf = SOzCF3, Ms = SOzCH3. on work previously performed by Potier's group" on a taxoid skeleton and by Danishefsky's group12 on a C ring model as before, leading to diol 38 in quantitative yield. The latter system. Both sequences utilized intermediate 25 (available from compound was heated in refluxing butanone to afford hydroxy total synthesis by silylation of 32 with TESC1-pyridine (85% oxetane 36 (72% yield), which was converted to acetate 24 as yield) or from degradation of 10-deacetylbaccatin 1II)'O and described above. proceeded as outlined below. The final drive toward Taxol (1) from intermediate 24 was In the first approach, which was modeled after Danishefsky's carried out as outlined in Scheme 7 and proceeded along the work,12 the C20-acetate group was selectively removed from lines already described in paper 1 of this series.1° Synthetic 25 under mildly basic conditions (KzC03-MeOH) to afford triol Taxol (1) was identical with an authentic sample by all usual 33 in 97% yield (Scheme 6). The newly generated primary criteria, including Rf (TLC), fR (HPLC), [a]% IR, 'H and 13C alcohol was then selectively silylated with TMSCl in the NMR,HFMS, and biological assay (microtubule stabilization presence of base and exposed to triflic anhydride and base to and cytotoxicity against a panel of eight cell lines). afford the triflate silyl ether 35 via intermediate 34. The latter compound converted to oxetane 36 when exposed to mildly Conclusion acidic conditions (silica gel, CH2C12) through sequential desi- lylation of the C20-hydroxyl group followed by internal SN~ This and the accompanying paper^',^,'^ in this series describe displacement of the triflate. The resulting hydroxy oxetane 36 the studies in these laboratories which eventually culminated was acetylated to afford the targeted oxetane system 24 in 40% in the total synthesis of Taxol (1). This synthetically challenging overall yield from triol 33. molecule with its 11 stereocenters, four skeletal rings, and The second route (Scheme 6), modeled after Potier's studies," unusual steric congestion, particularly around its 8-membered featured selective mesylation of diol 25 (73% yield) to furnish ring, provided several serious obstacles and opportunities to hydroxy mesylate 37 which was selectively deacetylated at C20 create new strategies and to expand the scope and generality of known synthetic methods. New knowledge was gained on (11) Ettouati, L.; Ahond, A,; Poupat, C.; Potier, P. Tetrahedron 1991, issues of regio-, stereo-, and chemoselectivity. Of particular 47, 9823. (12) Magee, T. V.; Bornmann, W. G.;Isaccs, R. C. A.; Danishefsky, S. interest were the applications of the Diels-Alder reaction to J. J. Org. Chem. 1992, 57, 3274. form rings A and C, the Shapiro and McMuny couplings to Total Synthesis of Taxol. 4 J. Am. Chem. SOC., Val. 117, No. 2, 1995 657 Scheme 7. Completion of the Total Synthesis" organic layer was dried (NazSOd), concentrated, and purified by flash chromatography (silica, 30% Et20 in petroleum ether) to give 3 (141 mg, 95%) as a white foam: Rf = 0.55 (silica, 60% Et20 in pe- troleum ether); [aIZZD +181 (c 0.48, CHC13); IR (thin film) vmPx3406, 2385, 1792, 1733, 1654, 1457, 1234, 1018 cm-I; IH NMR (500 MHz, CDC13) 6 7.35-7.27 (band, 5 H, Ar), 5.92 (dd, J = 10.0, 2.0 Hz, 1 H, 6-H), 5.62 (d, J = 5.0 Hz, 1 H, 10-H), 5.57 (dd, J = 10.0, 1.5 Hz, 1 H, 5-H), 5.49 (d, J = 4.5 Hz, 1 H, 2-H), 4.69 (d, J = 12.0 Hz, 1 H, 6 24 30 OCH~h),4.46(d,J=8.0H~,1H,20-H),4.44(d,J=12.0Hz,lH, OCHzPh), 4.28 (b d, J= 5.0Hz, 1 H, 9-H), 3.77 (d, J= 8.0 Hz, 1 H, 20-H), 3.71 (b S, 1 H, 7-H), 2.72 (ddd, J = 14.5, 10.0, 3.5 Hz, 1 H, bl 13-H), 2.58 (ddd, J = 20.0, 11.5, 3.0 Hz, 1 H, 14-H), 2.42 (b S, 1 H, 9-OH), 2.36 (d, J= 4.5 Hz, 1 H, 3-H), 2.09 (s, 3 H, OAC),2.01 (ddd, J= 20.0, 10.0, 3.5 Hz, 1 H, 14-H), 1.80 (ddd, J= 14.5, 11.5, 3.0 Hz, 1 H, 13-H), 1.68 (s, 3 H, 18-CH3), 1.53 (s, 3 H, 19-CH3), 1.42 (s, 3 H, C(CH3)2), 1.40 (s, 3 H, C(CH3)z). 1.13 (s, 3 H, 16-CH3), 1.05 (s, 3 H, 17-CH3); 13C NMR (125 MHz, Cmh) 6 169.2, 153.9, 142.5, 137.4, 135.4, 133.1, 128.5, 128.2, 122.5, 108.2, 93.3, 82.5, 78.1, 75.3, 74.1, 41 40 72.5, 71.2, 47.0, 44.7, 39.9, 31.3, 28.9, 27.8, 26.8, 23.6, 21.7, 21.2, 16.2; FAB HRMS (NBA/NaI) de605.2720, M + Na+ calcd for c33&zog 605.2727. a dl Ketone 4. A solution of alcohol 3 (141 mg, 0.242 mol) in CH3- CN (10 mL) was treated with tetrapropylammonium perruthenate (PAP, 85.0 mg, 0.0242 mmol) and 4-methylmorpholine N-oxide (NMO, 85.0 mg, 0.726 mmol) and stirred at 25 "C for 2 h. After dilution with CHzClz (30 mL), the reaction mixture was filtered through silica gel. The resulting solution was concentrated to give 4 (131 mg, 93%) as a white solid: Rf = 0.62 (silica, 30% EtOAc in petroleum ether); [aIz2D +14 (c 0.52, CHC13); (thin film) vmax2925, 1807, e 43:R=TES 42 IR 1 : R = H, Taxol 1746, 1717, 1458, 1374, 1230 cm-'; lH NMR (500 MHz, CDC13) 6 Reagents and conditions: (a) 5.0 equiv of PhLi, THF, -78 "C, 10 7.35-7.27 (band, 5 H, Ar), 6.47 (s, 1 H, 10-H), 5.90 (dd, J= 10.5, min, then 10 equiv of AczO, 5.0 equiv of 4-(dimethylamino)pyridine 2.0 Hz, 1 H, 6-H), 5.67 (dd, J = 10.5, 1.5 Hz, 1 H, 5-H), 4.66 (d, J = (DMAP), CH2C12, 2.5 h, 80%; (b) 30 equiv of pyridinium chlorochro- 11.5 Hz, 1 H, OCHSh), 4.57 (d, J= 11.5 Hz, 1 H, OCHZPh), 4.40 (d, mate (PCC), 30 equiv of NaOAc, Celite, benzene reflux, 1 h, 75%; (c) J=8.5Hz,lH,20-H),4.32(m,lH,7-H),4.18(d,J=5.5Hz,lH, excess N&&, MeOH, 25 "C, 3 h, 94% based on 88% conversion; (d) 2-H), 3.78 (d, J = 8.5 Hz, 1 H, 20-H), 2.78 (d, J = 5.5 Hz, 1 H, 3-H), 3.0 equiv of NaN(SiMes)z, 3.5 equiv of B-lactam 42, THF, 0 "C, 0.5 h, 2.78-2.70 (band, 2 H, 13-H and 14-H), 2.23 (m, 1 H, 14-H), 2.22 (s, 86% based on 89% conversion; (e) HF-pyridine, THF, 25 "C, 1.25 h, 3 H, OAc), 1.93 (m, 1 H, 13-H), 1.90 (s, 3 H, 18-CH3), 1.44 (s, 3 H, 80%. TES = SiEt3, BZ = COPh. C(CH3)2), 1.43 (s, 3 H, C(CH3)z), 1.26 (s, 3 H, 19-CH3), 1.27 (s, 3 H, 16-CH3), 1.15 (s, 3 H, 17-CH3); 13C NMR (125 MHz, CDC13) 6 203.2, construct ring B, and the regioselective opening of carbonates 169.3, 152.6, 143.3, 137.1, 134.8, 128.9, 128.4, 128.3, 127.9, 123.9, with organometallic reagents to form hydroxy esters. 108.9, 96.5, 81.8, 80.2, 76.5, 76.2, 71.7, 71.1, 58.9, 47.5, 40.5, 29.9, The resulting convergent route to Taxol (1) was utilized for 28.7, 26.8, 26.1, 23.2, 21.8, 20.8, 18.9, 12.8; FAB HRMS (NBMCsI) the construction of several new designed taxoids. A number de713.1720, M + Cs+ calcd for C33&09 713.1727. of these compounds obtained by total synthesis13 or semisyn- Acetate 25. Conversion of Oxetane 24 to Acetates 25 and 26. the~is'~J~have demonstrated interesting properties and shed light A solution of oxetane 24 (14.0 mg, 0.023 mmol) in CHzClz (2.5 mL) on the structural requirements for Taxol's biological activity. at 0 "C was treated with Et30BF4 (Meerwein's reagent, 1.O M in CHI- Clz, 0.048 mL, 0.048 mmol) and stirred at 0 "C for 1 h. Furthermore, water-soluble taxoids that arose from these studies The reaction mixture was diluted with Et20 mL), washed with aqueous mC1 providing useful information regarding the conformation of (IO are (5 mL) and brine (5 mL), dried (MgSOd), concentrated, and purified Taxol in waterI6 and the design of prod rug^^^*^^ of this newly by preparative TLC (silica, 50% EtOAc in petroleum ether) to give established chemotherapeutic agent. acetate 25 (8.5 mg, 59%) and acetate 26 (2.8 mg, 19%), both as colorless fiis. Experimental Section Acetate 25: Rj = 0.28 (silica, 50% EtOAc in petroleum ether); [aIZZD General Techniques. For a description of general technique, see -74 (C 0.75, CHC13); IR (thin film) vman3483,2943,2884,1802,1743, the first paper in this series.1° Experimental techniques and data for 1461, 1373, 1232, 1120, 1014 cm-l; 'H NMR (500 MHz, CDC13) 6 compounds 5,6,8-22,27, and 28 may be found in the supplementary 6.53 (s, 1 H, lO-H), 4.46 (d, J = 12.0 Hz, 1 H, 20-H), 4.40 (d, J = material. 12.0 Hz, 1 H, 20-H), 4.39 (dd, J = 11.0, 3.5 Hz, 1 H, 7-H), 4.23 (d, Acetate 3. A solution of diol 2 (138 mg, 0.0256 mmol) and J=5.0Hz,lH,2-H),3.71(t,J=3.5H~,lH,5-H),3.39(d,J=5.0 4-(dimethy1amino)pyridine (DMAP, 47.0 mg, 0.0383 mmol) in CHz- Hz, 1 H, 3-H), 3.16 (s, 1 H, 4-OH), 2.82 (ddd, J = 14.0, 10.0, 3.0 Hz, Clz (10 mL) was treated with AczO (0.04 mL, 0.0383 mol)and stirred 1 H, 13-H), 2.79 (s, 1 H, SOH), 2.71 (m, 1 H, 14-H), 2.25-2.05 (band, at 25 "C for 2 h. After dilution with Et20 (50 mL), the reaction was 2 H, 6-H and 14-H), 2.14 (s, 3 H, OAC), 2.10 (s, 3 H, 18-CH3), 1.88 quenched with aqueous NH&1(50 mL), and the resulting mixture was (m, 1 H, 14-H), 1.75 (m, 1 H, 6-H), 1.20 (s, 3 H, 16-CH3), 1.18 (s, 3 stirred at 25 "C for 15 min. The organic layer was separated, and the H, 17-CH3), 1.14 (s, 3 H, 0.63 (t, J = 7.5 Hz, 9 H, aqueous layer was extracted with Et20 (3 x 20 mL). The combined Si(CHzCH&), 0.58-0.45 (band, 6 H, Si(CHzCH3)3); I3C NMR (125 MHz, CDC13) 6 202.8, 170.6, 169.2, 153.2, 144.7, 130.1, 93.4, 81.5, (13) Nicolaou, K. C.; Claibome, C. F.; Nantermet, P. G.; Couladouros, 76.0, 74.8, 70.4, 68.5, 64.8, 61.3, 43.0, 40.4, 33.8, 30.2, 26.5, 23.0, E. A.; Sorensen, E. J. J. Am. Chem. Soc. 1994, 116, 1591. 21.1, 20.9, 20.8, 18.9, 11.9, 6.7, 5.1; FAB HRMS (NBA/NaI) de (14) Nicolaou, K. C.; Couladouros, E. A.; Nantermet, P. G.; Renaud, J.; 647.2845, M Na+ calcd for C31&8011Si 647.2864. Guy, R. K.; Wrasidlo, Angew. Chem., Znt. Ed. Engl. 1994,33,1581. + (15) Nicolaou, K. C.; Renaud, J.; Guy, R. K.; Nantermet, P. G.; Acetate 26: Rf = 0.36 (silica, 50% EtOAc in petroleum ether); IH Couladouros, E. A.; Wrasidlo, W. Submitted. NMR (500 MHz, CDC13) 6 6.51 (s, 1 H, 10-H), 5.21 (t, J = 3.0 Hz, (16) Gomez Paloma, L.; Guy, R. K.; Nicolaou, K. C. Chem. Bid 1994, lH,5-H),4.30(dd,J=l1.0,4.5H~,lH,7-H),4.20(d,J=4.5Hz, I, 107. (17) Nicolaou, K. C.; Guy, R. K.; Pitsinos, E. N.; Wrasidlo, W. Angew. (18) Nicolaou, K. C.; Riemer, C.; Ken, M. A.; Rideout, D.; Wrasidlo, Chem., Znt. Ed. Engl. 1994, 33, 1583. W. Nature 1993, 364, 464. 658 J. Am. Chem. SOC., Vol. 117, No. 2, 1995 Nicolaou et al.

1H,2-H),4.03(d,J=11.0H~,1H,20-H),3.58(d,J=11.0Hz,1(0.003 mL, 0.040 mmol), and 4-(dimethylamino)pyridine (DMAP, H, 20-H), 3.34 (s, 1 H, 4-OH), 3.20 (d, J = 4.5 Hz, 1 H, 3-H), 2.94 catalytic) and stirred at 25 "C for 0.5 h. The reaction was quenched (ddd, J= 14.0, 10.0, 3.5 Hz, 1 H, 13-H), 2.75 (m, 1 H, 14-H), 2.20 (s, with aqueous NaHC03 (1 mL), and the resulting mixture was extracted 3 H, 18-CH3), 2.18 (s, 3 H, OAc), 2.16 (s, 3 H, OAc), 2.12 (m, 1 H, with Et20 (3 x 20 mL). The combined organic layer was washed 14-H), 1.96 (ddd, J = 15.0, 4.5, 4.5 Hz, 1 H, 6-H), 1.90-1.84 (band, with HzO (5 mL) and brine (5 mL), dried (MgSOd), concentrated, and 2 H, 6-H and 13-H), 1.19 (s, 3 H, 16-CH3), 1.14 (s, 3 H, 17-CH3), purified by flash chromatography (silica, 50% EtOAc in petroleum 1.04 (s, 3 H, 19-C&), 0.86 (t, J = 8.0 Hz, 9 H, Si(CH2CH,),), 0.55- ether) to give diol 8 (4.6 mg, 95%) as an amorphous solid Rr = 0.40 0.49 (band, 6 H, S~(CHZCH~)~). (silica, 50% EtOAc in petroleum ether); [alZ2~-51 (c 0.08, CHCb); Silylation of Triol 32 to 25. A solution of triol 32 (2.0 mg, 0.0039 'H NMR (500 MHz, CDC13) 6 7.40-7.20 (band, 5 H, Ar),6.51 (s, 1 "01) in pyridine (0.5 mL) was treated with chlorotriethylsilane H, lO-H), 4.55 (d, J = 11.5 Hz, 1 H, OCHzPh), 4.48 (d, J = 12.0 Hz, (TESC1, 0.017 mL, 0.098 "01) and stirred at 25 "C for 12 h. The 1 H, 20-H), 4.44 (d, J = 11.5 Hz, 1 H, OCHZPh), 4.41 (d, J = 12.0 reaction mixture was diluted with Et20 (10 mL), washed with aqueous Hz, 1 H, 20-H), 4.22 (d, J = 4.5 Hz, 1 H, 2-H), 4.06 (dd, J = 11.O, 4.5 CuSO4 (3 x 5 mL) and brine (5 mL), dried (MgSOd), concentrated, Hz, 1 H, 7-H), 3.74 (m, 1 H, 5-H), 3.41 (d, J = 4.5 Hz, 1 H, 3-H), and purified by preparative TLC (silica, 50% EtOAc in petroleum ether) 3.14 (S, 1 H, 4-OH), 2.83 (ddd, J = 14.5, 10.5, 4.0 Hz, 1 H, 13-H), to give silyl ether 25 (2.0 mg, 85%) as a colorless film. 2.75 (b s, 1 H, 5-OH), 2.72 (m, 1 H, 14-H), 2.29 (ddd, J = 14.5, 4.0, Alcohol 29. To a solution of acetate 4 (18.7 mg, 0.032 "01) in 4.0 Hz, 1 H, 6-H), 2.18 (m, 1 H, 14-H), 2.17 (s, 3 H, OAc), 2.10 (s, 3 THF (2 mL) at 0 "C was added BHyTHF (1.0 M, 0.32 mL, 0.32 mmol), H, OAc), 2.03 (s, 3 H, 18-CH3). 1.90 (m, 1 H, 13-H), 1.69 (m, 1 H, and the reaction mixture was stirred at 0 "C for 3 h. The reaction was 6-H), 1.28 (s, 3 H, Ig-CHs), 1.20 (s, 3 H, 16-CH3), 1.15 (s, 3 H, 17- quenched with aqueous NaHCO3 (0.5 mL) and HZOZ(0.5 mL), and CH3); FAB HRMS (NBNCsI) de733.1633, M + Cs+ calcd for the resulting solution was allowed to warm to 25 "C, stirred at 25 OC c32&011 733.1625. for 1 h, and extracted with Et20 (3 x 30 mL). The combined organic Triol 32. Hydrogenation of 31. A solution of diol 31 (4.6 mg, layer was washed with HzO (5 mL) and brine (5 mL), dried (MgSOd), 0.0077 mol)in EtOAc (1 mL) was treated with Pd(0H)JC (1.0 mg) concentrated, and purified by preparative TLC (silica, 10% Et20 in under an atmospheric pressure of hydrogen and stirred at 25 OC for CHzCl2) to give acetate 4 (3.1 mg, 17%), the monoalcohol 29 (6.8 mg, 0.5 h. The reaction mixture was filtered, concentrated, and purified 42% based on 83% conversion) as an amorphous solid, and the by preparative TLC (silica, EtOAc) to give triol 32 (3.8 mg, 97%) as corresponding 6-OH regioisomer (3.3 mg, 22% based on 83% conver- an amorphous solid Rf = 0.20 (silica, EtZO); [alz2~-85.2 (c 0.115, sion) as an amorphous solid. CHC13); IR (thin film) Y,, 3492,2941,1795,1737, 1714, 1457, 1370, Alcohol 29: Rf = 0.80 (silica, 10% Et20 in CH2C12); ta1220-58 (c 1230, 1032 cm-'; 'H NMR (500 MHz, CDC13) 6 6.45 (s, 1 H, 10-H), 0.45, CHC13); IR (thin film) v,, 3523,2924, 1803, 1746,1716, 1459, 4.43 (s, 2 H, ~O-CHZ),4.42 (m, 1 H, 7-H), 4.20 (d, J = 5.0 Hz, 1 H, 1372, 1230, 1064 cm-I; 'H NMR (500 MHz, CDCl3) 6 7.38-7.22 2-H), 3.77 (t, J = 3.0 Hz, 1 H, 5-H), 3.39 (d, J = 5.0 Hz, 1 H, 3-H), (band, 5 H, Ar),6.50 (s, 1 H, 10-H), 5.58 (d, J = 11.5 Hz, 1 H, OCH2- 3.21 (S, 1 H, 4-OH), 2.83 (s, 1 H, 5-OH), 2.86-2.71 (band, 2 H, 13-H Ph), 4.48 (d, J = 11.5 Hz, 1 H, OCHZPh), 4.23 (d, J = 8.5 Hz, 1 H, and 14-H), 2.27-2.12 (band, 2 H, 6-H and 14-H), 2.18 (s, 3 H, OAC), 20-H),4.16(d,J=4.0H~,1H,2-H),4.06(dd,J=11.0,4.5Hz,12.11 (s, 3 H, OAc), 2.08 (s, 3 H, 18-CH3), 1.91 (m, 1 H, 13-H), 1.80 H, 7-H), 3.87 (t, J = 3.0 Hz, 1 H, 5-H), 3.77 (d, J = 8.5 Hz, 1 H, (m, 1 H, 6-H), 1.23 (s, 6 H, 16-CH3 and 17-CH3), 1.10 (s, 3 H, 19- 20-H), 3.46 (d, J = 4.0 Hz, 1 H, 3-H), 2.81-2.68 (band, 2 H, 13-H CH3); 13C NMR (125 MHz, cDcl3) 6 204.3, 170.8, 170.7, 153.1, 146.7, and 14-H), 2.61 (b s, 1 H, 5-OH), 2.36 (m, 1 H, 14-H), 2.20 (m, 1 H, 129.2, 93.3, 81.6, 76.2, 74.8, 70.2, 68.4, 64.8, 61.1, 42.8, 40.4, 32.4, 13-H), 2.19 (s, 3 H, OAc), 2.03 (s, 3 H, 18-CH3), 1.92 (m, 1 H, 6-H), 30.4, 26.4, 22.9, 21.7, 20.9, 20.8, 18.8, 11.4; FAB HRMS (NBNCsI) 1.61 (m, 1 H, 6-H), 1.45 (s, 6 H, C(CH3)2), 1.22 (s, 3 H, 16-CH3), 1.18 de643.1175, M + Cs' calcd for C25H34011 643.1155. (s, 3 H, 17-CH3), 1.13 (s, 3 H, 19-CH3); NMR (125 MHz, CDCl3) Desilylation of 25. A solution of diol 25 (6.5 mg, 0.010 mmol) in 6 203.0, 169.2, 153.0, 144.6, 137.5, 129.8, 128.2, 127.9, 127.5, 108.8, THF (2.0 mL) at 25 "C was treated with HFTyridine (0.4 mL) and 92.7, 84.6, 80.7, 76.1, 73.8, 71.2, 70.5, 68.8, 60.3, 40.6, 30.2, 29.7, stirred for 2 h. The reaction mixture was diluted with EtOAc (10 mL), 29.7, 29.6,26.4,26.2,23.0,21.3,20.9,18.8, 11.5; FAB HRMS (NEW washed with 10% aqueous NaOH (5 mL) and brine (5 mL), dried NaI) de621.2658, M + Na+ calcd for C33b2010 647.2676. (MgS04), concentrated, and purified by preparative TLC (silica, EtOAc) Triol 30. A solution of alcohol 29 (6.8 mg, 0.01 14 "01) in MeOH to give triol 32 (5.1 mg, 96%) as a colorless film. (2 mL) was treated with concentrated HCl(1 mL) and stirred at 25 "C Triol 33. A solution of acetate 25 (96.0 mg, 0.154 mmol) in MeOH for 5 h. The reaction was quenched with aqueous NaHC03 (1 mL), (16 mL) at 0 'C was treated with a solution of K2C03 (212 mg, 1.54 and the resulting mixture was extracted with EtOAc (3 x 30 mL). The "01) in HzO (4 mL). The reaction mixture was stirred at 0 "C for combined organic layer was washed with H20 (2 mL) and brine (2 15 min, and the reaction was quenched with aqueous N&C1(5 mL). mL), dried (MgSOd), concentrated, and purified by preparative TLC The reaction mixture was extracted with CHZClz (3 x 10 mL), and the (silica, 75% EtOAc in petroleum ether) to give monoalcohol 29 (0.8 combined organic layer was washed with brine (10 mL), dried (MgSOd), mg, 12%) and triol 30 (6.8 mg, 80% based on 88% conversion) as an concentrated, and purified by flash chromatography (silica, 25 - 50% amorphous solid: Rf = 0.30 (silica, 75% EtOAc in petroleum ether); EtOAc in petroleum ether) to give triol 33 (87.0 mg, 97%) as a white -71 (c 0.16, CHCls); IR (thin film) v,, 3453,2906, 1795, 1743, foam: Rr = 0.42 (silica, 50% EtOAc in petroleum ether); [alZzD-78 1714, 1458, 1372, 1233, 1038 cm-'; 'H NMR (500 MHz, CDC13) 6 (C 0.25, CHC13); IR (thin film) vm, 3458, 2955, 1796, 1751, 1714, 7.28-7.26 (band, 5 H, Ar), 6.49 (s, 1 H, 10-H), 4.55 (d, J= 11.0 Hz, 1461, 1373, 1234 cm-l; IH NMR (500 MHz, CDC13) 6 6.51 (s, 1 H, 1 H, OCHzPh), 4.45 (d, J = 11.0 Hz, 1 H, OCHZPh), 4.20 (d, J = 4.5 10-H), 4.39 (dd, J = 11.0, 4.5 Hz, 1 H, 7-H), 4.22 (d, J = 5.0 Hz, 1 Hz, 1 H, 2-H), 4.05 (dd, J 11.0, 4.5 Hz, 1 H, 7-H), 4.03 (b dd, J = H, 2-H), 4.01 (b d, J = 9.5 Hz, 1 H, 20-H), 3.81 (s, 1 H, 4-OH), 3.70 11.0, 4.5 Hz, 1 H, 20-H), 3.87 (s, 1 H, 4-OH), 3.73 (b t, J = 2.5 Hz, (b S, 1 H, 5-H), 3.52 (d, J = 9.5 Hz, 1 H, 20-H), 3.33 (d, J = 5.0 Hz, 1 H, 5-H), 3.52 (b dd, J = 11.0, 3.0 Hz, 1 H, 20-H), 3.35 (d, J = 4.5 1 H, 3-H), 3.06 (s, 1 H, 20-OH), 2.95-2.85 (band, 2 H, 13-H and Hz, 1 H, 3-H), 3.01 (b S, 1 H, 5-OH), 2.92 (ddd, J= 14.5, 10.5, 4.0 5-0H), 2.71 (m, 1 H, 14-H), 2.23 (ddd, J = 19.5, 9.0, 3.0 Hz, 1 H, Hz, 1 H, 13-H), 2.72 (ddd, J = 20.0, 12.0, 4.0 Hz, 1 H, 14-H), 2.59 14-H), 2.14 (s, 6 H, OAc and 18-CH3), 2.12 (m, 1 H, 6-H), 1.86 (ddd, (m, 1 H, 20-OH), 2.30 (ddd, J = 14.5, 3.5, 3.0 Hz, 1 H, 6-H), 2.21 J= 14.0, 12.0, 3.0 Hz, 1 H, 6-H), 1.69 (m, 1 H, 13-H), 1.18 (s, 3 H, (ddd, J = 20.0, 10.5, 3.0 Hz, 1 H, 14-H), 2.17 (s, 3 H, OAC),2.03 (s, 16-CH3), 1.14 (s, 3 H, 17-CH3), 1.09 (s, 3 H, 19-CH3), 0.87 (t, J= 8.0 3 H, 18-CH3), 1.87 (ddd, J = 14.5, 12.0, 2.5 Hz, 1 H, 13-H), 1.61 (m, Hz,9 H, Si(CHzCH&), 0.60-0.45 (band, 6 H, Si(CHzCH&); I3C NMR 1 H, 6-H), 1.20 (s, 3 H, 19-cH3), 1.16 (s, 3 H, 16-CH3), 1.15 (s, 3 H, (125 MHz, CDCls) 6 203.3, 169.2, 153.9, 144.8, 130.2, 93.7, 81.9, 17-CH3); 13C NMR (125 MHz, CDCls) 6 203.1, 169.2, 153.4, 144.7, 76.1, 73.6, 11.7, 68.5, 62.6, 61.4, 42.7, 40.5, 33.7, 30.3, 26.4, 22.9, 137.7, 129.8, 128.2, 127.8, 127.5, 97.9, 93.4, 81.5, 76.1, 74.5, 73.7, 21.2,20.9, 18.9, 11.9,6.7,5.1;FABHRMS(NBA/NaI)de605.2735, 72.5, 62.5, 60.0, 42.8, 30.2, 29.6, 29.4, 26.3, 22.8, 21.3, 20.9, 18.8, M + Na+ calcd for C29&010Si 605.2758. 12.4; FAB HRMS (NBA/NaI) de581.2341, M + Na+ calcd for Oxetane 36. Conversion of Triol 33 to 36. A solution of triol 33 C30H38010581.2363. (10.0 mg, 0.017 "01) and pyridine (0.142 mL, 0.51 "01) in CH2- Acetate 31. A solution of triol 30 (4.5 mg, 0.008 "01) in CHzClz Clz (2.0 mL) at 0 "C was treated with chlorotrimethylsilane (TMSC1, (1 mL) was treated with Ac2O (0.0009 mL, 0.010 mmol), pyridine 0.022 mL, 0.17 mol) and stirred at 0 "C for 15 min. The reaction Total Synthesis of Tanol. 4 J. Am. Chem. SOC.,Vol. 117, No. 2, 1995 659 was quenched with aqueous NaHC03 (2.0 mL). The resulting mixture and stirred at 25 "C for 1 h. The reaction mixture was diluted with was allowed to warm to 25 "C and extracted with Et20 (3 x 5 mL). Et20 (20 mL), washed with 1 N aqueous HCl (10 mL), aqueous The combined organic layer was washed with brine (10 mL), dried NaHC03 (5 mL), and brine (5 mL), dried (MgS04), concentrated, and (MgSOd), and concentrated to give the crude silyl ether 34, which was purified by flash chromatography (silica, 10 - 20% EtOAc in taken to the next step without further purification. petroleum ether) to give mesylate 37 (37.0 mg, 73%) as a white solid: A solution of silyl ether 34 and i-PratN (0.090 mL, 0.51 "01) in Rf = 0.38 (silica, 33% EtOAc in petroleum ether); [alz2~-40 (c 0.50, CHzClz (2.0 mL) at 0 "C was treated with triflic anhydride (TfzO, 0.044 CHC13); R (thin film) Y- 3495,2925, 1804, 1746, 1461, 1365, 1232 mL, 0.26 mmol) and stirred at 0 "C for 0.5 h. The reaction was then cm-I; 'H NMR (500 MHz, CDC13) 6 6.57 (s, 1 H, 10-H), 4.71 (t, J = quenched with aqueous NaHCO3 (1.5 mL), and the resulting mixture 2.5 Hz, 1 H, 5-H), 4.53 (d, J = 12.0 Hz, 1 H, 20-H), 4.50 (d, J = 12.0 was allowed to warm to 25 "C and extracted with Et20 (3 x 5 mL). Hz, lH,20-H),4.37(dd,J=11.0,4.5H~,lH,7-H),4.26(d,J=4.5 The combined organic layer was washed with brine (10 mL), dried Hz, 1 H, 2-H), 3.37 (d, J = 4.5 Hz, 1 H, 3-H), 3.15 (s, 1 H, 4-OH), (MgSOd), and concentrated to give the crude triflate 35, which was 3.08 (s, 3 H, OMS), 2.87 (ddd, J = 14.5, 10.0, 3.5 Hz, 1 H, 13-H), taken to the next step without further purification. 2.74 (ddd, J = 19.5, 12.0, 3.5 Hz, 1 H, 14-H), 2.38 (ddd, J = 19.5, A solution of triflate 35 in MeOH (2.0 mL) was treated with 10.0, 3.0 Hz, 1 H, 14-H), 2.23 (ddd, J = 15.0, 4.5, 2.5 Hz, 1 H, 6-H), camphorsulfonic acid (CSA, 0.5 mg, 0.002 "01) and stirred at 25 "C 2.19 (s, 3 H, 18-CH3), 2.18 (s, 3 H, OAC), 2.15 (s, 3 H, OAC), 2.02 for 15 min. The reaction was quenched with aqueous NaHC03 (1.5 (ddd,J= 15.0, 11.0, 2.5 Hz, 1 H, 6-H), 1.92 (ddd, J= 14.5, 12.0, 3.0 mL), and the mixture was extracted with CHzClz (3 x 5 mL). The Hz, 1 H, 13-H), 1.27 (s, 3 H, 19-CH3), 1.22 (s, 3 H, 16-CH3), 1.17 (s, combined organic layer was washed with brine (10 mL), dried (MgS04), 3 H, 17-C&), 0.91 (t, J = 8.0 Hz,9 H, Si(CHzCH&), 0.59-0.54 (band, and concentrated. The resulting residue was dissolved in CH2Clz(2.0 6 H, Si(CH2CH3)d; NMR (125 MHz, CDC13) 6 202.0, 170.9, 169.2, mL) and treated with silica gel (E. Merck, 0.1 g) at 25 "C for 1 h. The 152.9, 145.4, 130.0, 81.1, 80.9,75.9,73.5,68.5,64.4, 61.1,44.4,40.4, reaction mixture was filtered, concentrated, and purified by preparative 38.9, 34.7, 30.0, 29.7, 26.5, 23.1, 21.0, 20.9, 20.7, 18.9, 12.3, 6.7, 5.0; TLC (silica, 50% EtOAc in petroleum ether) to give oxetane 36 (3.9 FAB HRMS (NBNCsI) de835.1811, M + Csf calcd for C32H50013- mg, 40% from 33) as a colorless film: Rj = 0.35 (silica, 33% EtOAc SiS 835.1796. in petroleum ether); [a]22~-47 (c 0.42, CHC13); IR (thin film) vmax Diol 38. A solution of acetate 37 (24.0 mg, 0.034 mmol) in MeOH 3462,2927, 1805,1747, 1716,1595,1460,1372,1237 cm-l; 'H NMR (3.0 mL) at 0 "C was treated with a solution of K2CO3 (60.0 mg, 0.34 (500 MHz, CDC13) 6 6.39 (s, 1 H, 10-H), 4.82 (dd, J = 9.5, 2.0 Hz, 1 "01 in 0.5 mL of HzO) and stirred at 0 "C for 15 min. The reaction H, 5-H), 4.66 (d, J = 9.0 Hz, 1 H, 20-H), 4.42 (d, J = 9.0 Hz, 1 H, was quenched with aqueous WCl (2 mL), and the resulting mix- 20-H), 4.37 (d, J = 5.5 Hz, 1 H, 2-H), 4.12 (dd, J = 10.5, 7.0 Hz, 1 ture was extracted with CHzClz (3 x 5 mL). The organic layer was H, 7-H), 2.71 (m, 1 H, 14-H), 2.63 (d, J= 5.5 Hz, 1 H, 3-H), 2.62 (m, washed with brine (5 mL), dried (MgS04), and concentrated to give 1 H, 13-H), 2.48 (ddd, J = 15.0, 9.5, 7.0 Hz, 1 H, 6-H), 2.43 (s, 1 H, crude diol 38, which was taken to the next step without further 4-OH), 2.19 (m, 1 H, 14-H), 2.15 (s, 3 H, OAc), 2.06 (s, 3 H, 18- purification. CH3), 1.93 (ddd, J = 15.0, 10.5, 2.0 Hz, 1 H, 6-H), 1.89 (ddd, J = Diol 38: Rf = 0.51 (silica, 50% EtOAc in petroleum ether); [a]zzD 14.5, 12.0, 2.5 Hz, 1 H, 13-H), 1.62 (s, 3 H, 19-C&), 1.19 (s, 3 H, -35 (c 0.63, CHC13); IR (thin film) vmax3742, 2925, 1800, 1749, 16-CH3), 1.18 (s, 3 H, 17-CH3), 0.87 (t, J= 8.0 Hz,9 H, Si(CHzCH3)3), 1716, 1461, 1363, 1234 cm-l; 'H NMR (500 MHz, CDC13) 6 6.56 (s, 0.54 (q, J = 8.0 Hz, 6 H, Si(CHzCH&); 13C NMR (125 MHz, CDC13) lH, lO-H),4.74(t,J=3.0Hz, lH,5-H),4.38(dd,J=11.0,4.OH~, 6 203.0, 169.3, 153.4, 143.9, 131.1, 93.2, 87.5, 80.7, 80.5, 76.5, 73.8, 1 H, 7-H), 4.24 (d, J = 4.5 Hz, 1 H, 2-H), 4.01 (b d, J = 11.0 Hz, 1 71.9, 59.7, 51.6, 47.1, 37.8, 30.0, 26.2, 22.9, 21.7, 20.9, 19.0, 9.8, 6.7, H, 20-H), 3.83 (s, 1 H, 4-OH), 3.61 (b d, J = 11.0 Hz, 1 H, 20- 5.1; FAB HRMS (NBNCsI) de 697.1790, M + Cs+ calcd for H), 3.35 (d, J = 4.5 Hz, 1 H, 3-H), 3.11 (s, 3 H, OMS), 2.96 (ddd, J C29hOgSi 697.1809. = 14.5, 10.0, 4.0 Hz, 1 H, 13-H), 2.74 (m, 1 H, 14-H), 2.36 (ddd, J = Conversion of Mesylate 38 to Oxetane 36. A solution of crude 19.5, 10.0, 3.0 Hz, 1 H, 14-H), 2.26 (ddd, J = 15.0, 4.0, 3.0 Hz, 1 H, diol 38 (11.0 mg, 0.017 mmol) in butanone (1.0 mL) was treated with 6-H), 2.20 (s, 3 H, 18-CH3), 2.18 (s, 3 H, OAC), 1.98-1.90 (band, 2 n-Bu4NOAc (60.0 mg, 0.20 "01) and stirred at reflux for 5 h. The H, 6-H and 13-H), 1.22 (s, 3 H, 1%CH3), 1.17 (s, 3 H, 16-CH3), 1.16 reaction mixture was allowed to cool to 25 "C and partitioned between (s, 3 H, 17-CHs), 0.90 (t, J = 8.0 Hz, 9 H, Si(CHzCH&), 0.59-0.53 Et20 (10 mL) and H2O (5 mL). The organic layer was washed with (band, 6 H, Si(CHzCH3)3); NMR (125 MHz, CDC13) 6 202.4, brine (5 mL), dried (MgS04). concentrated, and purified by flash 169.1, 153.6, 145.4, 130.0,82.3,81.3,76.0,72.7,68.4,62.4,53.2,43.8, chromatography (silica, 10 - 20% EtOAc in petroleum ether) to give 40.4, 38.5, 34.6, 30.1, 29.6, 26.4, 22.9, 21.0, 20.8, 18.8, 12.1, 6.6,4.9; oxetane 36 (6.8 mg, 72% from 37) as a colorless film. FAB MS (NBA/NaI) de683, M + Na+ calcd for C3&8012SiS Acetate 24. A solution of oxetane 36 (4.0 mg, 0.0091 "01) and 683. 4-(dimethy1amino)pyridine (DMAP, 17.0 mg, 0.14 "01) in CHzCl2 For the conversion of carbonate 24 to Tax01 (1) and physi- (2.0 mL) was treated with acetic anhydride (0.0067 mL, 0.071 "01) cal data for compounds 1, 39-41, and 43, see the fust paper in this and stirred at 25 OC for 4 h. The reaction mixture was diluted with series.'O Et20 (10 mL), washed with 1 N aqueous HCl (5 mL) and aqueous NaHC03 (5 mL), dried (MgSOb), concentrated, and purified by Acknowledgment. We thank Drs. Dee H. Huang, Gary preparative TLC (silica, 33% EtOAc in petroleum ether) to give acetate Siuzdak, Raj Chadha, and Wolfgang Wrasidlo for NMR, mass 24 (4.0 mg, 94%) as a colorless film: Rf = 0.82 (silica, 50% EtOAc in spectroscopic, X-ray crystallographic, and biological assays hexanes); [al"~-49.4 (c 0.93, CHCl,); IR (thin film) v,, 2924, 1814, assistance, respectively. We also thank Dr. Luigi Gomez- 1728, 1461, 1372, 1238 cm-l; lH NMR (500 MHz, CDC13) 6.40 (s, 6 Paloma for helpful discussions regarding and stereochem- 1 H, 10-H), 4.95 (d, J = 9.0 Hz, 1 H, 5-H), 4.60 (A of d, J = 9.0 NMR AB, ical issues. This work was financially supported by NIH, The Hz, 1 H, 20-H), 4.47 (B of AB,d, J = 9.0 Hz, 1 H, 20-H), 4.43 (dd, J = 10.0. 7.5 Hz, 1 H, 7-H), 4.39 (d, J = 5.5 Hz, 1 H, 2-H), 3.36 (d, Scripps Research Institute, fellowships from Mitsubishi Kasei J = 5.5 Hz, 1 H, 3-H), 2.71 (m, 1 H, 13-H), 2.56 (m, 1 H, 13-H), 2.17 Corporation (H.U.), Rh6ne-Poulenc Rorer (P.G.N.), NSERC (s, 3 H, OAc), 2.15 (s, 3 H, OAc), 2.12 (m, 1 H), 2.07 (s, 3 H, 18- (J.R.), and grants from Merck Sharp & Dohme, Pfizer, Inc., CHd, 1.97 (m, 1 H), 1.88 (m, 2 H), 1.78 (s, 3 H, 19-CH3), 1.23 (s, 3 Schering Plough, and the ALSAM Foundation. H, 16-CH3), 1.17 (s, 3 H, 17-CH3), 0.88 (t, J = 7.5 Hz, 9 H, Si(CHzCH&), 0.60-0.50 (band, 6 H, Si(CHzCH3)s); I3C NMR (125 Supplementary Material Available: Experimental tech- MHz, CDC13) 6 202.6, 170.3, 169.2, 153.1, 144.0, 130.7, 92.8, 84.0, niques and data for compounds 5, 6, 8-22, 27, and 28 (12 80.3, 80.0, 76.4, 76.1, 60.3, 43.5, 38.0, 29.7, 29.4, 25.5, 23.1, 21.9, 21.1, 19.1, 9.8, 6.7, 5.2; FAB HRMS (NBNCs1)de 739.1929, M + pages). This material is contained in many libraries on Cs+ calcd for C31II46010Si 739.1915. microfiche, immediately follows this article in the microfilm version of the journal, and can be ordered from the ACS. See Mesylate A solution of alcohol (46.0 mg, 0.074 mmol) and 37. 25 any current masthead page for ordering information. 4-(dimethylamino)pyridine (DMAP, 180 mg, 1.48 "01) in CHzClz (6.0 mL) was treated with mesyl chloride (MsC1.0.056 mL, 0.72 mol) JA942195E