8615

yields to the a series analog. Reaction of 23 with

NBS-CCL afforded the dibromide 24 which was lac-
tonized to 25 (mp 175-176") with aqueous H2SOI-
monoglyme. Catalytic dehalogenation10 followed by
chromatography gave two fractions; the major prod-
uct was bromopyridone 26 (mp 169-170") and the
minor product was the debromo analog of 26. Finally, 1 2, X,Y = 0
oxidation of each as described for the a series gave the 3, X = H ; Y = O H
CDE ring analog 27 (mp 195-197") and debromo-27
(mp 175-177').
(24) National Institutes of Health Predoctoral Fellow.
"Y q N Y o
U C H p
Jacob J. P l a t t ~ ~ e rRichard
,~~ D. Gless, Henry Rapoport* 4, X = H ; Y = OCOCH,
Department of Chemistry, University of California 5, X , Y = 0
Berkeley, California 94720
Received July 31, 1972
">c.-ao
Y
WCH,OR
A Total Synthesis of dl-Camptothecin 6 , X = H; Y = OCOCH,; R = COCH,
Sir : 7,X=H;Y=OH; R = H

The initial report of potent antileukemic and anti-

tumor activity of the novel alkaloid camptothecin (1),
whose isolation and structure determination were re-
ported' in 1966, has been followed by several total
syntheses of this important compound. Recently, -COOCH, )- COOCH,
there was reported3 from this laboratory a broadly ap-
8, X = H ; Y = OCOC,H, 11
plicable synthesis of a series of analogs of camptothecin 9, X = H ; Y = OH
containing the fused pyridone-lactone DE ring system 10. X.Y = 0
of the parent alkaloid. We now wish to present an
extension of this synthetic procedure to the total syn-
thesis of dl-camptothecin. \
The previously reported syntheses2 involved forma- " \ / CH20COCHJ
tion of the pyridone D ring via cyclization followed by >COOCH, -l---i
elaborations on the pyridone ring, generally effected 0
through Michael-type additions either before or after 12 13
D ring formation. The route we are presenting has peridone acetate 4 as a mixture of isomers. This mix-
the pyridone D ring intrinsically built into the starting ture was subjected to SeOz oxidation in glacial acetic
material, pyridine-2,5-dicarboxylic acid. Subsequent acid (70", 1 hr), and chromatography gave a 58%
methylene lactam rearrangement of a nipecotic acid4 yield of the allylic diacetate 6. Hydrolysis of this di-
gives the desired piperidone. The main feature of our acetate 6 in anhydrous methanol-KzCOs (20', 30 min)
synthesis is a facile series of alternate rearrangement- to the diol 7, mje 183, was achieved quantitatively.
oxidation reactions, proceeding in good yields and Introduction of the a-butyrate side chain was accom-
culminating in a practical preparation of dl-campto- plished by Claisen rearrangement,6 utilizing diol 7 and
thecin. excess trimethyl orthobutyrate with a catalytic amount
The bicyclic keto acid 2, obtained3 in 85 % yield from of propionic acid at 145" for 3 hr. The crude reaction
pyridine-2,5-dicarboxylicacid, was reduced by sodium mixture was distributed between methylene chloride
borohydride in methanol-water ( O O , 18 hr) to the hy-
and dilute aqueous hydrochloric acid and evaporation
droxy amino acid 3,s obtained in 86% yield after puri- of the methylene chloride phase gave a 75;< yield of
fication by ion exchange. a-Methylene lactam rear- material containing the a-butyrate side chain. This
rangement* in acetic anhydride (145", 2.5 hr) gave, after material was a mixture of the free alcohol 9 and its
chromatography on silica gel, an 8 4 % yield of the pi-
butyrate ester 8. Treatment with anhydrous K2C03
(1) M. E. Wall, M. G. Wani, c. E. Cook, .(I H . Palmer, A. T. Mc- in methanol (20",1 hr) and chromatography on silica
Phail, and G. A. Sim,J. Amer. Chem. SOC.,88,3888 (1966). gel with 5 methanol-chloroform gave the alcohol 9,
(2) (a) G. Stork and A. G. Schultz, ibid., 93, 4074 (1971); (b) R.
Volkmann, S . Danishefsky, J. Eggler, and D. M. Solomon, ibid., 93, obtained as a mixture of isomers in 100% yield.
5576 (1971); (c) M. C. Wani, H. F. Campbell, G . A. Brine, J. A. Kep- To introduce the AB ring system it was now necessary
ler, and M. E. Wall, ibid., 94, 3631 (1972); (d) M. Boch, T. Korth, to oxidize the alcohol 9 to ketone 10 in preparation for
J. M. Nelke, D . Pike, H . Radunz, and E. Winterfeldt, Chem. Ber., 105,
21 26 (1972)
_.-_ ~
a Friedlander quinoline synthesis. This was effected
(3) J. J. Plattner, R. D. Gless, and H . Rapoport, J . Amer. Chem. SOC., by oxidation of the alcohol 9 with dicyclohexylcarbo-
94,8613 (1972).
(4) M. L. Rueppel and H. Rapoport, J . Amer. Chem. SOC.,94, 3877
diimide in DMSO with a catalytic amount of phos-
(1972). phoric acid7 ( Z O O , 30 hr) giving, after chromatography,
(5) All new compounds were characterized as to purity by tlc or
gc, and their ir and nmr spectra support the assigned structures. Ele-
mental compositions were established by high-resolution mass spectra, ( 6 ) W. S. Johnson, L. Werthemann, W. R . Bartlett, T. J. Brocksom,
combustion analyses, or both. A number of compounds were ob- T. Li, D. J. Faulkner, and M. R. Peterson, J . Amer. Chem. SOC.,92,
tained as isomeric mixtures, but these were not separated since they 741 (1970).
converged after compound 11. (7) I<. E. Pfitzner and J. G. Moffatt, ibid., 85,3027 (1963).

Communications t o the Editor

 8616
a 7 6 z yield of the ketone 10. Friedlander condensa-
tion between the keto ester-a-methylene lactam 10 and
N-(2-aminobenzylidene)-p-toluidine8gave a 75 % yield
of the tetracyclic a-methylene lactam l l , 9 containing
the indolizino[ 1,2-b]quinoline ring system and having
ultraviolet absorption typical for such a substituted
quinoline (319, 312, 306, 298, 293, 288, 234 nm).
The remaining tasks were aromatization of ring D I I1 (15s1, R, =OH; R2= H
and formation of the a-hydroxylactone ring E. Both II(lSR),R,=H; R! = O H
aromatization and formation of the necessary primary
The reduction of I with R = CHICO, CH2(CH&CO,
allylic alcohol were accomplished in one step by SeOa
or i-Pr(CH&Si using borohydride type reagents, e.g.,
oxidation of a-methylene lactam 11 in glacial acetic
NaBH4-C2HSOH at - 30" or Zn(BH,),-dimethoxy-
acid (SO@,30 min) to the a-acetoxymethylpyridone 12
ethane at 0" or various lithium trialkylborohydrides
(uv 370, 290, 253 nm). Hydrolysis-lactonization of
(R1R2R3BH-Li+) at <-90", affords I1 with a ratio l5S/
12 i n 2 N HsS04-glyme at 50" for 5 hr gave a 72 yield
of deoxycamptothecin (13) from 11: mp 262-264"
15R of between 50/50 and 60/40. Similarly the 11-
deoxy derivatives of I or A lo( l)-dehydro-11-deoxy-I
dec; uv 370,290, 253 nm. Oxidation of deoxycampto-
are reduced by these hydrides to equal mixtures of 1 5 s
thecin (CuCla-DMF-02)2dis accomplished in quantita-
and 15R alcohols. One reason for the difficulty in
tive yield to give dl-camptothecin (1) whose tlc, and uv, achieving selective reduction of I is the occurrence of
nmr, and high-resolution mass spectra are identical
both s-cis and s-trans conformations of the enone unit
with those of the natural product.1° Although a de-
in I as can be seen from the infrared spectra of esters of
tailed development of the individual steps has not been
structure I with varying R which manifest enone car-
made, we have thus obtained dl-camptothecin in an bonyl absorption (in CHC13 at 25") as a doublet of
overall yield of 11 % starting from pyridine-2J-dicar- comparably intense bands at 5.90 (s-cis) and 5.97 p
boxylic acid.
(s-trans).3 Even assuming that the ketone I adopts a
(8) T. K. Liao, W. H. Nyberg, and C. C. Cheng, J . Heferocycl. Chem., trans coplanar arrangement of hydrogen at C- 12 and
8,373 (1971).
(9) The AB rings in camptothecin can be incorporated at earlier C-13 as shown in 111, stereospecificity of carbonyl re-
stages. Thus, aia the Friedlander condensation, we have obtained
compound i from keto acid 2 and compound ii from the keto-ol-methyl-
enepiperidone 5 (prepared by hydrolysis of acetate 4 followed by DCC-
DMSO oxidation7). These compounds may be further elaborated to
camptothecin, and these processes will be reported in detail in the H-, ,thexyl
future.

- $CHj Lif
I11 IV
i 11
duction would demand not only a preferred direction
(10) Obtained from young (90 day) C. acuminata plants by Dr. G. of hydride attack (axis a or b) but also a single enone
Sheriha. These plants were grown from seeds kindly provided by
Mr. R. L. Smith and Dr. R. Perdue of the U. S . Department of Agricul- conformation. Therefore, it is not sufficient merely to
ture. choose a group R of sufficient steric bulk to block ap-
Cyril Tang, Henry Rapoport*
proach along axis b, but it is also necessary to control
Department of Chemistry, Unicersity of California
the enone conformation as s-cis in order to direct the
Berkeley, California 94720 formation of 1 5 s alcohol.
In an earlier phase of this work1a the derivatives I
Receiued August 30, 1972 and I1 were prepared in which R = p-phenylbenzoyl,
and it was found that this protecting group was ad-
vantageous since (1) the intermediates in the synthesis
Efficient Generation of the 1 5 s Configuration in were nicely crystalline, (2) the 1 5 s and 15R forms of I1
Prostaglandin Synthesis. Attractive Interactions in were readily separable by chromatography, and (3) the
Stereochemical Control of Carbonyl Reduction ultraviolet chromophore simplified analytical and
Sir: chromatographic work. It was subsequently dis-
covered la that the reduction of I, p-phenylbenzoyl,
One of the most fascinating problems in the area of favored formation of 1 5 s alcohol to a larger degree
prostaglandin synthesis has been the development of than was observed with screening groups at C-11, e.g.,
synthetic approaches which control stereochemistry, R = CH,(CH?),CO. Using the reagent IV'" in tetra-
particularly at C-15. We have been interested for hydrofuran-ether-pentane at - 120 to - 130", it was
some time in devising a method to effect the conversion possible to convert I, R = p-C6HjC6H4C0,to a mix-
of I to I1 (15S), with high stereoselectivity. This com- ture of 15.5 and 15R alcohols in a ratio of 82:18.1a14,5
munication reports a rational attack on this problem.
(3) See K. Noack and R . N. Jones, Can. J . Chem., 39,2225 (1961).
( 1 ) See, for example: (a) E. J. Corey, S . M. Albonico, U. Koelliker, (4) The reagent IV is superior to all other trialkylborohydrides which
T. I<. Schaaf, and R. I(. Varma, J . Amer. Chem. Soc., 93, 1491 (1971); have been tested so far for the selective formation of (15S)-II. The
(b) C. J. Sih, R. Price, R . Sood, R. G. Salomon, G. Peruzzotti, and M. following reagents of the type R1R?R3BH-Li+have been found to give
Casey, ibid., 94, 3643 (1972); (c) E. J. Corey and T. Ravindranathan, the l5Sjl5R ratios indicated: dicyclohexyl-tert-butyl (59,'41); dlcyclo-
ibid., 94, 4013 (1972); (d) E. J. Corey and P. L. Fuchs, ibid., 94, 4014 hexyltrityl (64,'36); diisopinocamphenenylmethyl (68,'32); triphenyl
(1972); (e) J. Fried, J. C. Sih, C. H. Lin, and P. Dalven, ibid., 94, 4343 (67133); tri-exo-2-norbornyl (59/41); diisobutyl-tert-butyl (74!26);
(1972); (f) R. Pappo and P. W. Collins, TefrahedronL e u . , 2627 (1972). di-sec-butyIthexyl(80:20); tri-see-butyl(78 22). In addition, IV yielded
(2) Prostanoic acid numbering. considerably higher ratios of 15S,'15R I1 than did various cvclic boro-

Journal of the American Chemical Society / 94:24 / November 29, 1972