Communications to the Editor
= ~ . ~ H z , ~ H ) , O . ~ O ( ~ , J = ~ . 1~. 1H7Z( ~, , ~3 HH) ,) 3, . 6 0
(d, J = 7 Hz, 1 H)) in 52% yield from 5b.
Treatment of the hydroxy ketone 2b with triphenylmethylenephosphorane in dimethyl sulfoxide (1 2 h, 55 "C)Io gave
the exo-methylene compound 2d6 (mp 68.0-69.0 "C (from
hexane): N M R 8veasi(CCI4) 0.74 (d, J = 7 Hz, 3 H), 0.87 (d,
J = 7 Hz, 3 H), 0.98 (s, 3 H), 3.60 (d, J = 7.5 Hz, 1 H), 4.63
(br s, 1 H), 4.75 (br s, 1 H)) in 92% yield. Reductive cleavage
of 2d with lithium in ethylamine ( I min, 16 "C) gave a single
alcohol, l b , in 90% crude yield.' I There was no evidence that
any of the 100 epimer of l b (retention of configuration in the
ring opening) was obtained in this reaction. Without purification l b was converted into the tosylate 1c6 (mp 81-82 "C
(from hexane); N M R d M q S i (CCI4) 0.75 (d, J = 6 Hz, 3 H),
0.78 (d, J = 6 Hz, 3 H), 0.86 (d, J = 6 Hz, 3 H), 1.63 (br s, 3
H ) , 2 . 4 5 ( ~ , 3 H ) , 4 . 5 2 ( d , J = S H z , IH ) , 5 . 1 3 ( b r s , l H ) , 7 . 2 3
(d, J = 8 Hz, 2 H), 7.72 (d, J = 8 Hz, 2 H)) using tosyl chloride in pyridine (96 h, 25 "C) in 60% overall yield from 2d.
We also carried out the conversion of the model tricyclodecanone 2a into (+)-spiroaxene (Id)' in a similar manner.
Methylenation of 2a as above gave 2c ( N M R d
~ (CCI4)
~
~
0.82 (d, J = 6 Hz, 6 H), 0.97 (s, 3 H), 4.64 (br s, 1 H), 4.47 (br
s, 1 H)) which upon reaction with lithium in ethylamine gave
Id6 ( N M R bh.leasi(CC14) 0.75 (d, J = 6 Hz, 3 H), 0.83 (d, J
= 6 Hz, 6 H), 1.72 (br s, 3 H), 5.28 (br s, 1 H); [c?]*jD + I 1.6"
(c 2.0, ether)."
Bose, Kistner, and FarberI* have reported the conversion
of menthyl tosylate into neomenthylamine via S Nreaction
~
with sodium azide in aqueous dimethylformamide followed
by lithium aluminum hydride reduction of the azide. However,
attempted conversion of I C to the azide l e using their procedure led primarily to the formation of elimination products as
did the use of potassium azide in acetonitrile containing 18crown-6. It was clear that it would be necessary to carry out
the tosylate displacement under conditions which would be
more favorable to an S y 2 reaction. This was accomplished by
treating IC with 3 equiv of potassium azide in benzene containing 5 equiv of 18-crown-6 (48 h, 80 "C)." The azide l e
showed N M R dve4si(CC13) 0.71 (d, J = 6 Hz, 3 H), 0.93 (d,
J = 6 Hz, 3 H), 0.97 (d, J = 6 Hz, 3 H), 1.77 (br s, 3 H), 3.47
(s, 1 H), 5.20 (br s, 1 H); mass spectrum (70 eV), no M+, m/e
219 ( M - N2, weak), 204 ( M - "3,
strong). Reduction of
l e with LiAIH4 in ether a t reflux gave the amine If ( N M R
6
~ (CC14)
~ 0.71
~ (d,~J = 6j Hz, 3 H), 0.87 (d, J = 5 Hz, 6 H),
1.73 (br s, 3 H), 2.63 (br s, 1 H), 5.27 (br s, 1 H)) in 30%
overall yieldI4 from IC.
Using a procedure analogous to that of Hertler and Corey,"
If was converted into (-)-axisonitrile-3 (la) in 85% overall
yield by treatment with a 2: I mixture of formic acid in acetic
anhydride at reflux for 2 h followed by reaction with tosyl
chloride in pyridine a t 25 "C for 1 h. (The formamide derivative, presumably (+)-axamide-3,l,lh was isolated as an intermediate in this sequence.) The synthetic material6 showed mp
97-99 "C (from hexane); N M R 8ve4si(CC14) 0.75 (d, J = 6.5
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 ? I 2 j ~
-71" ( C 0.35, CHC13).
These physical properties generally agreed with those reported for (+)-a~isonitrile-3,~
the enantiomer of la, except for
the sign of the optical rotation.' However, there were small
discrepancies between the observed N M R chemical shifts,
particularly for the methyl groups of the isopropyl group, and
those reported by Sica and co-workers.' Therefore, verification
of the structure of the synthetic material was desirable. Unfortunately, a direct comparison of the synthetic material with
the natural product could not be made since neither a pure
authentic sample nor copies of the original spectral data were
available to us. In order to confirm the structure of the syn0002-7863/78/1500-8031$01.00/0
803 1
thetic compound, a single crystal X-ray study was carried out
using a Syntex P21 four-circle diffractometer. A complete data
set was collected; the published coordinates' were refined using
Sheldrick's S H E L X - 7 6 least-squares program. The refinement
converged, with a residual of 0.13, using isotropic thermal
parameters and without the hydrogen atoms being included.
A difference Fourier synthesis showed no peaks of electron
density greater than 0.5e/A3. This conclusively demonstrated
that the synthetic material was in fact (-)-axisonitrile-3
( l a ) .17
References and Notes
(1) B. DiBlasio, E. Fattorusso, S. Magno, L. Mayol, C. Pedone, C. Santacroce,
and D. Sica, Tetrahedron, 32, 473 (1976).
(2) (a) W. G. Dauben and E. J. Deviny. J. Org. Chem., 31, 3794 (1966);(b) E.
Piers and P. M.Worster, J, Am. Chem. Soc., 94, 2895 (1972);(c) D. Caine.
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);
J. A . Marshall, G. L. Bundy, and W. I. Fanta, ibid., 33, 3913 (1968).
(4) SMC Corp., Glidden Organics, Jacksonville, Fla.
(5) D. N. Kirk and J. M. Wiles, Chem. Commun., 518 (1970).
(6) A correct combustion analysis has been obtained for this compound.
(7) A . F. Kluge, K. G. Untch, and J. H. Fried, J. Am. Chem. SOC.,94, 7827
(1972).
~ this
i We suggest the abbreviation MIP (i,e., rnethoxyisopropylidine)for
derivative.
(8) (a) H. J. Reich, J. M. Renga, and I. L. Reich, J. Am. Chem. Soc., 97, 5434
(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)
For the specific application of this sequence to conversion of an octalone
to a cross-conjugated dienone, see D. Caine, A. A. Boucugnani, and W.
R. Pennington. J. Org. Chem., 41, 3632 (1976).
(9) (a) For reviews covering the photochemical conversion of bicyclic crossconjugated dienones into tricyclodecenones, see P. J. Kropp, Org. Photochem.. 1, 1 (1967);K. Schaffner, Adv. Photochem., 4, 81 (1966). (b) A
Hanau NK-20 low-pressure mercury lamp was used as the light source.
(10) R. Greenwald, M. Chaykovsky, and E. J. Corey, J. Org. Chem., 28, 1128
(19631.
~ - - ,
In general, the cyclopropane ring of simple vinylcyclopropanes is not
cleaved by metals in liquid ammonia (see S. W. Staley, Sei. Org. Transform.,
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
by lithium in ethylamine and extreme care had to be exercised to prevent
overreduction. To our knowledge the conversions of 2c and 2d into l d and
lb, respectively, represent the first examples of reductive cleavages of
vinylcyclopropanes with lithium in ethylamine. However, cleavages of
strained divinylcyclopropanes such as octamethylsemibullavene with
lithium in liquid ammonia have been reported (W. T. Borden, A. Gold, and
S. D. Young, J. Org. Chem., 43, 486 (1978)).
A. J. Bose, J. F. Kishner, and L. Farber, J, Org. Chem., 27, 2925 (1962).
We are indebted to Professor Charles L. Liotta for discussions on this point
and for a gift of 18-crown-6.
The yield of the amine I f obtained in the reduction of the azide l e with
LiAIH4 in ether was surprisingly low. Other methods for accomplishing this
conversion are being explored.
W. R. Hertler and E. J. Corey, J. Org. Chem., 23, 1221 (1958).
Spectral data for (-)-axamide-3 were not reported in ref 1 ,
We are very grateful to Dr. Donald G. VanDerveer for his assistance in
carrying out the X-ray work.
\
Drur? Caine,* Howard Deutsch
School of Cheniisfrj, Georgia I m ? i ? u ~ofr Technologj
Atlanta, Georgia 30332
Received June 2, I978
Stereospecific Total Synthesis of Gibberellic Acid.
A Key Tricyclic Intermediate
Sir:
Since the recognition of the central biological role of gibberellic acid (gibberellin A3, GA3) (1) in the plant kingdom,'
the clarification of its chemical structure,2 and commercial
production on a large scale from the fungus Gibberellafujikuroi, this substance has occupied a major position in the field
of natural product^.^ The biosynthesis of gibberellic acid from
prenyl units, though long and involved, is known in considerable detai1.'$3,4Despite extensive efforts (some 150 published
papers from about 25 different laboratories), the total chemical
synthesis of gibberellic acid has not previously been a ~ h i e v e d , ~
largely because the combination of overall molecular complexity, centers of high sensitivity toward many reagents, and
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
serves to thwart all but the most sobhisticated of approaches.6
We report in this and the following publication the first total
synthesis of gibberellic acid by a route which is structurally
unambiguous and stereospecific, and which employs a number
of crucial new synthetic method^.^
The plan of synthesis, which was derived by extensive antithetic analysis, is outlined briefly in Scheme I.8 Key features
of the approach as revealed in this summary include (mentioned in the order of synthetic execution) (1) the stereospecific
generation of the cis-fused B/C ring unit by Diels-Alder addition, (2) formation of the D ring by internal pinacol cyclization, (3) position-specific ring contraction of ring B from six
to five members, (4) formation of ring A by internal DielsAlder reaction and stereospecific methylation to form a pentacyclic lactone with all carbons in place, (5) oxidation and
isomerization at C(6) and C(7) of the pentacyclic lactone, and
(6) stereospecific elaboration of the complete A / B ring unit
of gibberellic acid.9
The phenolic ether 2 was prepared in two steps and -75%
overall yield from 2-allyloxyanisole by ( 1 ) Claisen rearrangement at 230 OC for 2 h (89% of distilled product, bp
87-97 OC a t 0.6 mm, containing 93% ortho-allyllic and 7%
para-allylic phenol),I0 followed by (2) etherification of the
resulting phenol using sequentially 1.3 equiv of sodium hydride
in tetrahydrofuran (THF) a t 0 OC and 1.6 equiv of chloromethyl 2-methoxyethyl ether (MEM chloride"), first at 0 O C
and then at 25 OC for 2 h.I2 Conversion of 2 to the phenol 3 was
accomplished by the sequence ( 1 ) oxidation of 2 with 3 equiv
of sodium periodate and 0.2 mole % of osmium tetroxide in 3: 1
THF-H20 at 0 "C for 0.5 h and 23 "C for 1.5 h, (2) immediate
R
O
/
C H
OMEM
,
reduction of the noraldehyde so obtained with sodium borohydride in absolute ethanol at 0 O C for 20 min, (3) benzylation
of the resulting primary alcohol via the sodium salt (excess
N a H in T H F ) in T H F with excess benzyl bromide a t reflux
for 15 h, and (4) selective cleavage of the MEM group using
1.5 equiv of trifluoroacetic acid in methylene chloride at 23 OC
for 18 h. After column chromatography pure phenol 3 was
obtained in 74% yield overall from 2 as a colorless oil. Oxidation of the phenol 3 to the yellow quinone 4 (Scheme I), mp
7 1-72 OC, was effected by stirring in dimethylformamide solution a t 23 OC for 4 days with molecular oxygen in the presence of 0.08 equiv of bis(salicylidene)ethylenediiminocobalt( 11) (salcomine)' 3,14 (>75% yield).
Reaction of the quinone 4 with trans-2,4-pentadien-l-ol
(5)15 occurred upon heating in benzene solution a t reflux for
30 h to afford a single crystalline adduct ( 6 ) in 91% yield.I6
The next phase of the synthesis, transformation of the adduct
6 to the keto aldehyde 7, was originally attempted using the
well-known Woodward p r 0 ~ e d u r e . However,
l~~
because this
direct method failed completely, alternative routes were examined.17b The most satisfactory from the standpoint of reproducibility and ease of scale-up consisted of the following
sequence: (1) reaction of 6 with 1.1 equiv of dihydropyran and
0.12 mol % p-toluenesulfonic acid in methylene chloride ( 10
mL/g of 6 ) at 0 "C for 18 h to form quantitatively the tetrahydropyranyl (THP) ether; ( 2 ) reduction of the THP ether
with 1 mol equiv of sodium borohydride in absolute ethanol at
0 OC for 100 min to produce the hydroxy enone 8 (100%);(3)
conversion of 8 to the a-methoxymethylenoxy ketone 9 by
reaction with 8 equiv of N,N-diisopropylethylamine and 4
equiv of chloromethyl methyl ether in methylene chloride a t
OCH,
OH
PhCH,O
OCH,Ph
-
2
3
:q
1
3
OCH,Ph
R=H
R:CH,0CH3
uF1
THPO'
QH
THPO~
OH
o,
CH,
OCH,Ph
QOcHs
-0
THPO'QOH
OCO,CH,
H
Communications to the Editor
8033
reflux for 9 h (100%);(4)reduction of the keto group in 9 with
lithium aluminum hydride in ether at -10 OC for 1 h, isolation,
and immediate mesylation of the resulting alcohol at -58 OC
in T H F with 2 equiv each of methanesulfonyl chloride and
triethylamine, followed by slow addition of saturated aqueous
potassium bicarbonate and gradual warming to 0 OC over 45
min to effect solvolysis, and finally chromatography of the
product on silica gel to give 10 in 77% overall yield from 6; ( 5 )
selective hydrogenation18 of the enone double bond of 10 using
1 .O equiv of hydrogen and 5% rhodium-on-carbon catalyst at
23 OC in T H F ; and (6) addition of a solution of the hydrogenation product and tert-butyl alcohol (10 equiv) in THF to 12
equiv of lithium in liquid ammonia a t -78 "C over 7 min,
stirring a t reflux for 6 h, quenching with ammonium chloride,
and isolation by chromatography on silica gel to yield the THP
ether diol 11 in 63% overall yield from 10. Oxidation of 11 with
dipyridinechromium(V1) oxide (Collins reagent) (excess) in
methylene chloride a t -45 "C for 2 h a n d -25 OC for 1 h with
stirring in the presence of dry, acid-washed Celite gave after
treatment with powdered sodium bisulfate monohydrate
(stirring a t -20 O C for 30 min), dilution with dry ether, filtration, and concentration the sensitive keto aldehyde 7 in 84%
yield as a pale yellow oil which was used as such in the next step
(intermediate storage at -78 O C for a minimal period).
The persistence of a cis-ring fusion in the various intermediates derived from the Diels-Alder adduct 6 could be demonstrated chemically. Thus, reaction of 8 with N-bromosuccinimide in THF a t 23 "C afforded quantitatively the bridged
bromo ether 12. Further, reaction of the diol 11 with 1 equiv
of methyl chloroformate-pyridine gave selectively the carbonate of the primary alcohol which was oxidized (Collins
reagent) to the corresponding cyclohexanone and treated with
p-toluenesulfonic acid-methanol to produce in high overall
yield the bridged ketal 13.
The pinacol cyclization of the keto aldehyde 7 to the tricyclic
intermediate 14 proved to be a more difficult proposition than
expected from earlier studies of closely related models.9a The
most satisfactory and convenient procedure involved preparation of finely powdered metallic titanium under argon by
addition of small pieces of potassium to a mixture of titanium
trichloride in THF (8.5 equiv of Tic13 and 24 equiv of
and then heating cautiously at reflux for 3 h, cooling to 23 OC,
and gradual addition of the keto aldehyde 7 in T H F . The reaction mixture was stirred a t 23 OC for 2.5h, cooled to 0 OC,
and very cautiously treated dropwise with anhydrous methanol,
diluted with aqueous K2C03, filtered through Celite, and extracted with 4:1 ether-methylene chloride. After isolation the
product (-10-g batch size) was chromatographed on a Waters
Associates Model 500 preparative machine which separated
the three major components easily and afforded 40% cis- 14,
15% trans-14, and -10% diol 11 (which was recycled).20
Oxidation of either cis- or trans- 14 (or a mixture of the two)
to form the ketol 15 without appreciable glycol cleavage was
readily accomplished via an oxysulfonium intermediate, in
accordance with previously published results.2' The most
satisfactory conditions for the oxidation involved addition of
pinacol(s) 14 to a suspension of the complex formed from 7
equiv of dimethyl sulfoxide and 3.5 equiv of trichloroacetic
anhydride in methylene chloride at -60 "C for 1 h, stirring the
mixture of reactants at -50 "C for 45 min, treatment with 3.5
equiv of triethylamine at -50 to +23 O C over 2.5h, and finally
extractive isolation (ether-methylene chloride).22The ketol
15 (yield 75-80% if purified chromatographically) was converted directly without p ~ r i f i c a t i o nto~ the
~ M E M ether derivative 16'
by treatment with 3 equiv of M E M chloride
and I O equiv of diisopropylethylamine in methylene chloride
a t reflux for -16 h (65% yield overall from 14 after chromatographic purification on silica gel).
The next stage of the synthetic approach required the ad-
K)9a319
justment of size and pattern of functionality of the B ring of
the tricyclic intermediate 16, which was achieved as follows.
Treatment of a solution of 16 in acetone-water (2.5:l)with
1.3 equiv of N-methylmorpholine N-oxide and 0.05 equiv of
osmium tetroxide a t 23 "C for 80 h furnished a single cis diol
(17) (89% yield after c h r ~ m a t o g r a p h y which
) ~ ~ was cleaved
by reaction with 1.03equiv of lead tetraacetate in benzene a t
5 O C for 0.5 h. The sensitive dialdehyde 18, isolated from the
reaction mixture simply by addition of ether, filtration through
Celite-anhydrous sodium sulfate, and concentration in vacuo
was used directly in the next step without purification (and
with minimal delay26). Treatment of 18 with 0.2equiv of dibenzylammonium triflu~roacetate~'
in benzene at 50 O C for
1 h gave after chromatography on silica gel the desired a$unsaturated aldehyde 19 in 64% overall yield from the diol 17.
The yield of 17 from the cyclization is actually higher since 19
undergoes partial decomposition on silica gel; in practice,
therefore, 19 was used for the next step without purification.
Reaction of 19 with 5 equiv of methylenetriphenylphosphorane
in THF-hexamethylphosphoramide a t reflux for 3.5 h furnished the Wittig product 20 in good yield (44% overall from
diol 17; 80% from purified keto aldehyde 19). Exposure of 20
to acetic acid-THF-water (3:l:l)at 35 OC for 40 h produced
the diene alcohol 21 cleanly without detectible cleavage of the
M E M protecting group (32% overall for 5 steps from 16).The
successful production of the critical tricyclic intermediate 21
in quantity set the stage for the completion of the synthesis of
gibberellic acid as described in the following communicati0n.2~J9
References and Notes
(1) See, H. N. Krishnamurthy, Ed., "Gibberellins and Plant Growth", Wiley, New
York, 1975.
(2) (a) P. W. Brian, J. F. Grove, and J. MacMillan, Fortschr. Chem. Org. Naturst.,
18, 350 (1960); (b) J. F. Grove, 0.Rev. (London). 15, 56 (1961); (c) F. M.
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.
SOC.,85, 3414 (1963).
(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,
"Natural Products Chemistry", Vol. 1, Academic Press, New York, 1974,
p 265.
(4) See also (a) B. E. Cross in "Progress in Phytochemistry", Vol. 1. Interscience, New York, 1968; (b) B. Dockerill, R. Evans, and J. R. Hanson, J,
Chem. SOC.,Chem. Commun., 919 (1977), and papers therein cited.
(5) A number of simpler gibberellins and degradation products of gibberellic
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.
Mori, M. Shiozaki, N. Itaya, M. Matsui, and Y. Sumiki, Tetrahedron, 25, 1293
(1969), for a 52-stage synthesis of gibberellin Aq; (b) W. Nagata, T. Wakabayashi, Y. Nayase. M. Narisada, and S. Kamata. J. Am. Chem. SOC.,93,
5740 (1971), and 92, 3202 (1970), for a 49-step route to gibberellin AIS;
(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 epialiogibberic acid.
(6) For a discussion see R. L. Danheiser, Ph.D. Dissertation, Harvard University,
1978.
(7) The problem of the synthesis of gibberellic acid has provided the impetus
for the development of many new synthetic methods,6 for example from
the laboratories of Stork (see G. Stork, D. F. Taber, and M. Marx, Tetrahedron Left., 2445 (19781, and references therein cited), Loewenthai (see
H. J. E. Loewenthal and S. Schatzmiller, J. Chem. Soc., Perkin Trans. 7,
2149 (1975), and references cited), S. Masamune (see S. Masamune, J.
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.
Beames, J. A. Halleday, andL. N. Mander, Aust. J. Chem., 25, 137 (1972)).
House (see H. 0. House, D. G. Meiillo, and F. J. Sauter. J. Org. Chem., 38,
741 (1973)),and Ziegler (see F. E. Ziegler and J. A. Kloek, Tetrahedron
Lett..
~. 315 (19741)
- . ,I
Much of the work described herein has been disclosed in various lectures,
e.g., at the Nichols Award Symposium, Tarrytown, N.Y., March 1977, and
at the International Symposium on Organic Synthesis, Oxford, England,
1977.
Certain key stages of this synthetic approach had already been tested and
described in previous publications from this laboratory. These include (a)
internal inacol cyclization of a keto aldehyde (E. J. Corey and R. L. Carney,
J. Am. &em. Soc., 93, 7318(1971), and E. J. Corey, R. L. Danheiser, and
S. Chandrasekaran, J. Org. Chem., 41, 260 (1976)): (b) internal Dieis-Aider
approach to the construction of the A I 6 ring unit (E. J. Corey and R. L.
Danheiser, Tetrahedron Left., 4477 (1973)); (c) stereospecific placement
of all A ring ring substituents (E. J. Corey. T. M. Brennan, and R. L. Carney,
J. Am. Chem. SOC..93. 7316 11971\1
The mixture was used as suchwith purification conveniently effected at
a subsequent stage
.~
-
\
8034
Journal of the American Chemical Society
(11) E. J. Corey, J.-L. Gras, and P. Ulrich. Tetrahedron Lett., 809 (1976).
(12) Structural assignments for all stable synthetic intermediates are based upon
proton magnetic resonance (lH NMR), infrared, and mass spectra determined using purified, chromatographically homogeneous samples. In addition ultraviolet spectra were determined where appropriate and were also
consistent with the formulations shown herein. All reactions involving airor moisture-sensitive components were carried out in an atmosphere of
drv
-., nronn
(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.,
34. 273 11969).
(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
procedure was less convenient largely because of the labor involved in
preparing the reagent. Using the reactions outlined above the quinone 4
can be prepared reproducibly in 100-9 lots.
(15) (a) R. G. Glushov and 0. Y. Magidson, Med. Prom. SSSR, 16, 27 (1962);
(Chem. Abstr., 58, 4420 (1963)); (b) S. Oida and E. Ohki, Chem. Pharm.
Bull., 17, 1990 (1969).
(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
expectation is fully confirmed by the data which follow.
(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
application of one of the sequences developed in the course of our investigations, see J.-L. Gras, Tetrahedron Lett., 41 17 (1977).
(18) See S. K. Roy and D. M. S. Wheeler, J. Chem. SOC.,2155 (1963).
(19) See J. E. McMurry and M. P. Fleming, J. Org. Chem., 41, 896 (1976). For
a detailed account of the very extensive studies carried out on the cyclization of 7 and related substances with a wide range of reagents, see also
ref 6.
(20) The less polar of the isomeric pinacois, mp 97-99 OC, is probably cis-I4
and the more polar isomer, mp 87.5-89 O C , trans-14, based upon previous
experience with tricyclic analogues of known configuration (see, for example, ref 9a) and also on chemical data.
(21) See, E. J. Corey and C. U. Kim, J. Org. Chem., 38, 1233 (1973).The various
known chromium(Vi) reagents and many other standard oxidizing agents
for alcohols afford mainly glycol fission products with substrates such as
14.
(22) This is a useful modification of the procedure of Swern; see K. Omura, A.
K. Sharma, and D. Swern, J. Org. Chem., 41, 957 (1976), and S. L. Huang,
K. Omura, and D. Swern, ibid., 41, 3329 (1976).
(23) The keto1 15 is prone to 1,2 rearrangement of methylene at the bridgehead
to form the stereoisomeric a-ketoi upon exposure to base or acid or prolonged chromatography.This aiiogibberlc
gibberic type rearrangement
is driven by the relief of strain in going from cis-fused BIC rings (skew-boat
C ring) to trans-fused BiC rings (chair C ring). The occurrence of the same
rearrangement with 17-nor-17-oxoailogibberic acid represented an inconsistency in the originally assigned stereochemistry of gibberellic acid
which led us to propose the X-ray crystallographic study (see ref 2d) that
eventually produced the correction of the earlier configurational assignment at C(9).
(24) The fl-methoxyethoxymethyl (MEM) protecting group'l was originally developed for this specific application.
(25) For method see V. Van Rheenen, R. C. Kelly, and D. A. Cha, Tetrahedron
Lett., 1973 (1976).
(26) As might be expected the diaidehyde 18 is quite unstable (e.g., to water
or silica gel).
(27) The use of this outstanding selective reagent (a crystalline solid) for this
very demanding step (see ref 6) was arrived at by systematic experimental
variation of secondary amine and acid components based on the idea of
activating the methylene group a to the less hindered formyl group as an
enamine by means of a not-too-basic, sterically discriminating secondary
amine under almost neutral aprotic conditions. Other studies with this reagent will be published separately. Since these investigations one of the
undersigned has successfully applied a similar reagent to the direct ru
methylenation of ketones; see J.-L. Gras. Tetrahedron Lett., 21 11
( 1978).
(28) This investigation was supported financially by the US. National Science
Foundation to whom we are deeply grateful.
(29) We are pleased to acknowledge helpful information and experimental data
from the following colleagues: Drs. Sandor Barcza, Thomas M. Brennan,
Robert L. Carney, Tetsuo Hiraoka, Masayuki Narisada. George Strunz, and
Gerald L. Thompson.
-
E. J. Corey,* Rick L. Danheiser
Srinivasan Chandrasekaran, Patrice Siret
Gary E. Keck, Jean-Louis Gras
Department of Chemistry, Harvard Unicersity
Cambridge, Massachusetts 02138
Receiced September 5, I978
Stereospecific Total Synthesis of Gibberellic Acid
Sir:
This communication describes the completion of the stereospecific total synthesis of gibberellic acid (GA3) (1) from
a key tricyclic intermediate (2) which is readily accessible by
0002-7863/78/ 1500-8034$01 .OO/O
1 100:25 1 December 6, 1978
the approach detailed in the preceding publication.' In addition
we disclose a new facet of the chemistry of gibberellic acid
which allows access to derivatives in which the C(7) substituent
on ring B is in the unnatural (and generally less stable2) cy orientation and which also provided useful direct correlation of
GA3 with a number of advanced synthetic intermediates.
Deprotonation of the hydroxy diene 2 with 1 .O equiv of nbutyllithium in tetrahydrofuran (THF) a t -40 "C followed
by acylation with 1.55 equiv of trans-2-chloroacryly1 chloride3
at -40 "C for 0.5 h afforded the ester 3 in -80% yield (-62%
overall from the T H P ether of 2).4 When 3 was heated in
benzene solution containing -100 equiv of propylene oxide (as
a hydrogen chloride scavenger) in a sealed tube a t 160 "C for
45 h under argon the pure crystalline lactone 4, mp 149-1 50
"C, could be obtained in 55% yield after recrystalli~ation.~
The
stereochemistry of 4 is assigned from the supposition of concerted, a-face, "endo" internal Diels-Alder addition (there
was no evidence for the formation of an appreciable amount
of any stereoisomer of 4); it is supported by 'HN M R data and
also by subsequent transformation to GA3. Treatment of the
adduct 4 with 2.2 equiv of lithium isopropylcyclohexylamide
and 5 equiv of hexamethylphosphoramide i n T H F at -78 " C
for 50 min followed by reaction with 5 equiv of methyl iodide
a t -78 to 0 "C over 12 h afforded cleanly the methylated
lactone 5 (-75% yield). At this stage the MEM6 protecting
group was removed from 5 by stirring in dry chloroformether-nitromethane (1 5 5 : 1 by volume) with 25 equiv of finely
powdered anhydrous zinc bromide a t 23 OC for 3 h to yield
hydroxy lactone 6 (-70% after chromatography). The IR, IH
N M R , UV, and mass spectra and the T L C mobility of this
material were identical with those of a sample of optically
active 6 obtained from natural gibberellic acid as described
below.
The synthetic (&)-hydroxy lactone 6 was resolved using a
novel procedure designed to take advantage of the lone (tertiary bridgehead) hydroxyl in 6. Exposure of 6 to a large excess
of phosgene and 3 equiv of 4-dimethylaminopyridine in dry
methylene chloride at 23 "C for 36 h gave, after rapid filtration
through dry Celite and concentration in vacuo, crude chloroformate 77 which was directly treated with (-)-a-phenylethylamine ( [cY]*~D-41.7" in benzene) to provide after isolation
a mixture of two diastereomeric urethanes (8) (95% total yield)
which could be separated cleanly by chromatography on silica
gel using 1 : 1 ethyl acetate-hexane for elution (TLC R, values
in this solvent system, 0.24 and 0.20). The less polar diastereomer, [RI2'D +59" (c 0.44, CHC13), was identical spectroscopically (IR. ' H N M R , mass spectrum) and chromatographically with urethane prepared from hydroxy lactone 6
from natural G A j and (-)-a-phenylethylamine which showed
[ a ] 1 5+61"
~
(c 0.42, CHCI,). Reaction of this less polar
synthetic urethane 8 with 5 equiv of triethylamine and 3 equiv
of trichlorosilane in dry benzene at 25 "C for 60 h8 afforded
i n 95% yield resolved hydroxy lactone 6, mp 21 1-212 "C,
[.]*OD
+162" (c 0.58, CHC13), identical in all respects (IR,
' H N M R , mass spectrum, T L C high pressure liquid chromatography) with the hydroxy lactone 6 derived from natural
GA3 which showed [.I2'D +161" (c 0.49, CHC13).
The optically active lactone 6 was hydrolyzed to the corresponding hydroxy acid salt by heating at reflux with excess 1 .O
N aqueous potassium hydroxide for 45 min (argon atmosphere) and the resulting solution was treated at 23 "C with
2.07 equiv of 0.013 M sodium ruthenate9 in 1 N aqueous SOdium hydroxide for 2.5 h. Filtration through Celite. acidification to pH 3 at 0 "C, and extraction afforded upon isolation
the diacid 9, spectroscopically and chromatographically
identical with the diacid obtained from GA, (see below); the
corresponding dimethyl esters (from excess C H I N > in ether)
were also identical. The formation of diacid 9 clearly proceeds
by way of the intermediate acid aldehyde which undergoes
0 1978 American Chemical Society
8035
Communications to the Editor
2 R:H
1
I
-3 R=C!"c:c; Hco
%OH
RO
R
HO -HO'
n
o
co2cn1
L!
CO,R
~
o
,
,
CO,CH,
H
OQAOAO
'
0
%
rs
base-catalyzed epimerization to the more stable 6P-formyl
derivative and then further oxidation to the observed product.1°
Selective monoesterification of the diacid was accomplished
in T H F by treatment with triethylamine (1.5 equiv) and p toluenesulfonyl chloride (1 equiv) at -78 OC for 0.5 h and -50
OC for 2 h (to form the mixed sulfonic anhydride), subsequent
quenching with excess methanol at -50 OC initially, and then,
after warming to 23 OC, stirring for a further 2 h. Chromatography afforded the monoester 10 as a solid foam, [cY]*OD
-21 O (c 4.9, THF), identical in all respects with the compound
obtained from GA3 as described earlier.Il From this optically
active intermediate (10)I2the synthesis of GA3 is completed
by the previously describedi1 route which includes ( 1 ) hydroxylactonization of 10 with m-chloroperbenzoic acid to form
11;13 (2) lactone saponification and iodolactonization of 11 to
give the iodolactone 12; (3) in one flask, trifluoroacetylation
of 12 to 13, reduction with zinc to eliminate the 1-iodo and
2-trifluoroacetoxy substituents and bicarbonate treatment to
saponify the 3-trifluoroacetate forming GA3 methyl ester; and
finally (4) conversion of GA3 methyl ester to the free acid 1
using sodium n-propyl mercaptide in hexamethylphosphoramideI4 at 0 OC.
The transformation of gibberellic acid 1 to the key intermediate 6 and several other gibberellins having the C(7) substituent a-oriented a t C(6) was achieved by the use of a novel
strategy for effecting the 6 0
6 a epimerization which is
normally contrathermodynamic in this series.
Saponification of the acid ester l o i 5by heating at reflux
with excess 1 N aqueous potassium hydroxide for 40 min afforded after acidification and isolation the diacid 9 (95% yield),
[CY]"D -25.5'
(c 2.4, THF), which could be reesterified with
excess diazomethane to the same dimethyl ester obtained by
methylation of 10 (indicating that no epimerization a t C(6)
occurs in the saponification). Reaction of the diacid 9 with I O
equiv of triethylamine and 1 equiv of N,N'-dicyclohexyl carbodiimide in T H F at reflux for 7 h furnished, upon workup and
chromatography on silica gel, the anhydride 15, mp 167-1 68
OC, [ F ] * ~ D+268O (c 9.3, CHCI3) (73% yield). The stereochemistry of the anhydride, anticipated to be as shown in 15
on geometrical grounds, was shown by methanolysis
(CH~OH-C~HSN
and
) methylation of the resulting acid-ester
with diazomethane to produce a dimethyl ester stereoisomeric
with that obtained by methylation of 9 or 10.l6The success of
the 6/3
6 a epimerization involved in the formation of the
anhydride 15 depends on activation of the C(6) carboxylic acid
-
-
function under equilibrating conditions (triethylamine catalysis) and subsequent capture of the C(6) a-oriented carbonyl
by the 4a-carboxylic group. Reduction of 15 with -0.6 mol
equiv of lithium borohydride in dimethoxyethane at -25 "C
for 1.5 h yielded, upon acidification with acetic acid, workup,
and chromatography on silica gel, the lactone 6 (50%) together
with a structurally isomeric lactone (16, 16%); R, values for
6 and 16 were 0.74 and 0.85, respectively (silica gel plates,
ethyl acetate-acetic acid, 95:5). The isomeric lactone 16, mp
171 O C , [(Y]20D +133O (c 8.7, CHC13), was synthesized unambiguously from the ester acid 10 by the following sequence:
(1) reaction of the tetra-n-butylammonium salt of 10 in dry
T H F with 1 equiv of mesitylenesulfonyl chloride at -78
23
OC over 2 h to form the mixed sulfonic anhydride; (2) reduction
of the activated 4-carboxylic group to a 4-hydroxymethyl group
(without isolation) at 0 OC by addition of excess sodium
borohydride and reaction at 0 OC for 1 h: and (3) epimerization
at C(6) and concomitant lactonization of the resulting dihydroxy ester (14) by heating at reflux with sodium methoxide
in absolute methanol for 48 h .
Lithium borohydride reduction of either lactone 6 or 16
afforded the triol 17 (colorless, foam), [ ( u l Z 0 ~-17.0' (c 3.5,
T H F ) , oxidation of which with chromic acid (two phase,
ether-water) led exclusively to lactone 16 (no detectible
6)."
The research results described in this and the foregoing
paper mark the achievement of one of the more intriguing and
salient objectives i n the area of organic synthesis. They also
provide a basis for further synthetic and transformational investigations relating to gibberellic acid, and we hope to report
on the ongoing work in this area in due course.lx.Iy
-
References and Notes
( 1 ) E. J. Corey, R. L. Danheiser, S. Chandrasekaran. P. Siret, G. E. Keck. and
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.
Danheiser, Tetrahedron Lett., 4477 (1973).
(3) (a) A. N. Kurtz. W E. Billups. R. B. Greenlee. H. F. Hamil. and W. T. Pace,
J. Org. Chem., 30, 3141 (1965); (b) P. K. Freeman, B. K. Stevenson, D. M.
Balls, and D. H. Jones, ibid., 39, 546 (1974).
(4) Satisfactory infrared (IR), proton magnetic resonance ('H NMR), and mass
spectral data were obtained on purified, chromatographicallyhomogeneous
samples of the synthetic intermediates described herein. All reactions
involving air- or moisture-sensitive components were carried out in an
atmosphere of dry argon.
(5) In earlier model studies of this internal Diels-Alder step,lb satisfactory
cyclization was obtained both with trans-2-chloroacrylate and propiolate
esters. In contrast to the earlier studies the conversion of 3 to 4 failed
completely in the absence of propylene oxide. The cyclization of the
8036
Journal of the American Chemical Society
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 carboxylactivated derivatives of propiolic acid), again in contradistinctionto earlier
model studies or model experiments with other primary or secondary alcohols.
(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 6, with various
alcohols to form the corresponding mixed carbonates as well as with primary amines to form the expected urethanes.
(8) Method of W. H. Pirkle and J. R. Hauske, J. Org. Chem., 42, 2781
(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 (triethylamine, Oediger's base, or chromatography on silica gel).
(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,
Ph.D, Dissertation, Harvard Universtiy, 1978.
(12) The optically active ester acid 10 may also be obtained by resolution of
racemic 10 using (-)+
1'-napthy1)ethylamine with ethyl acetate-ether
for crystallization.
(13) Hydroxy lactone 11 could also be obtained from diacid 9 by hydroxylactonization (1.05 equiv of peracetic acid in water-ethyl acetate at pH 9)
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-etherisopropyl alcohol (lOO:lO:l)).
(17) The naturally derived triol 17 was spectroscopically and chromatographically 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 Foundation.
eCN
Me
-
E. J. Corey,* Rick L. Danheiser
Srinivasan Chandrasekaran, Gary E. Keck, B. Gopalan
Samuel D. Larsen, Patrice Siret, Jean-Louis Gras
Department of Chemistry, Harcard Unicersity
Cambridge, Massachusetts 021 38
Receiced September 5, 1978
A Highly Efficient Total Synthesis
of (f)-Lycopodine
Sir:
Lycopodine ( l ) , the archetypal Lycopodium alkaloid,' has
been known since 1881,2 although its full structure was not
established until 1960.3 Intensive synthetic work during the
1 960s4 resulted in two total syntheses of the alkaloid which
were communicated in 196gS5An earlier approach resulted in
the synthesis of the unnatural diastereomer 12-epilycopodine
(2).6A recent communication reports a synthesis of racemic
anhydrolycod~line.~
Since natural anhydrolycodoline is hydrogenated to 2 and 1 in a ratio of 6.5:1,8 this work constitutes
a further formal synthesis of lycopodine. We wish to communicate a highly efficient stereospecific total synthesis of
lycopodine which is promising for application to the synthesis
of some of the many other members of this important class of
alkaloids.'
Cyanoenone 39 is converted into cyanodione 4 by stereuselective trans additionlo of lithium dimethallylcopper (ether,
-78 "C; 64%),'l followed by ozonolysis ( 0 3 , C H 3 0 H , -78
"C; 87%), or by conjugate addition of the cuprate derived from
the lithiated N,N-dimethylhydrazone of acetone, followed by
aqueous hydrolysis ((1) T H F , -78 "C, 4 h; (2) Cu2C12, T H F ,
HlO, pH 7, 25 "C, 16 h; 60%).12 Both procedures afford cyanodione 4 as a separable mixture of C2 epimers, in an approximate equimolar ratio. However, we have been unable to
detect, at this stage or any subsequent stage, C3-Cj cis dia-
0002-78631781l S00-8036$01.OO/O
/ 100:25 / December 6, 1978
&*
Me
A
4:
X
X = O ; Y = C N
Y = CN
Y = C02H
_13-
14
--
stereomers. Cyanodione 4 is converted via cyano diketal 5
( H O C H ~ C H Z O Hp-TsOH,
,
C6H6, reflux; 99%) to diketal acid
6 (KOH, H 2 0 , CzHsOH, reflux, 16 h; 90%). Treatment of
acid 6 with ethyl chloroformate in the presence of triethylamine, followed by 3-benzyloxypropylamine (THF, - I O "C;
88%),i3 affords amide 7 which is reduced to secondary amine
8 (LiAIH4, T H F , reflux, 16 h; 99%).
Treatment of amino diketal8 with HCI in methanol results
in slow intramolecular Mannich cyclization (3.2 M HCI, reflux, 14 days), affording a single tricyclic amino ketone (10)
in 65% yield. Although compound 8, like compounds 4-7, is
an equimolar mixture of C2 epimers, none of the 12-epi diastereomer (lycopodine numbering) has been found in the reaction product. This kinetic stereoselectivity was a n t i ~ i p a t e d ' ~
and is also observed in cyclization of the analogous N-benzylamine 9, which affords tricyclic amino ketone I t , uncontaminated by its diastereomer, under similar (but less stringent) conditions (2.2 equiv of HCI, CH3OH, reflux, 48 h;
66%).
Catalytic debenzylation of 10 (H20, CzHsOH, HCI, Hz,
Pd; 96%) affords crystalline alcohol 12 (mp 86-87 "C), which
undergoes Oppenauer oxidation (benzophenone, t-CdHgOK,
C6H6, reflux, 30 min)I5 with subsequent intramolecular aldolization and dehydration to afford racemic dehydrolycopodineL6(13, mp 104-105 "C; A,,
24.5 nm ( t 5000)) in 72%
yield. Catalytic hydrogenation of 13 (H2, Pt, C ~ H S O Haffords
)
racemic lycopodine (1, mp 130-131 "C (lit.5a mp 130-131
"C)) in 87% yield. The synthetic material produced in this
manner is identical with a sample of natural lycopodine by
infrared and 180-MHz ' H N M R spectroscopy.
The efficiency of the current synthesis is demonstrated by
the high overall yield (17.7% from enone 3, 1 1 . l % from dihyd r o ~ r c i n o l ' and
~ ) by the fact that no other lycopodine diastereomer may be detected in the final product, even though
isomer separations are not carried out a t any point during the
synthesis. In one continuous run, we have prepared 1.2 g of
0 1978 American Chemical Society