A rhodium carbene cyclization-cycloaddition cascade strategy toward the pseudolaric acids.

Article abstract of DOI:10.1021/ol0159165

A rhodium carbene intramolecular cyclization-cycloaddition cascade was employed as the key reaction in the synthesis of the nucleus of the cytotoxic diterpenoids pseudolaric acids A and B.

Full text of DOI:10.1021/ol0159165

ORGANIC
LETTERS
2001
Vol. 3, No. 11
1721-1724
A Rhodium Carbene
Cyclization−Cycloaddition Cascade
Strategy toward the Pseudolaric Acids
Pauline Chiu,* Bin Chen, and Kin Fai Cheng
Department of Chemistry, The UniVersity of Hong Kong, Pokfulam Road,
Hong Kong, P. R. China
pchiu@hkucc.hku.hk
Received March 31, 2001
ABSTRACT
A rhodium carbene intramolecular cyclization−cycloaddition cascade was employed as the key reaction in the synthesis of the nucleus of the
cytotoxic diterpenoids pseudolaric acids A and B.
The root bark of Pseudolarix kaempferi Gordon (Pinaceae),
a tree native to China, has been a traditional medicinal herb
called tujinpi for the treatment of dermatological fungal
infections as early as the 17th century.1 From this preparation,
a family of diterpenoids called the pseudolaric acids have
been isolated. Pseudolaric acid B (1b) has been determined
to be the main antifungal component and has been evaluated
to have antimicrobial activity comparable to that of ampho-
tericin B against a number of fungal strains.2 In vitro tests
of pseudolaric acid A (1a) and pseudolaric acid B (1b) have
demonstrated that they are cytotoxic to several cancer cell
lines at sub-micromolar levels, while being relatively non-
toxic in vivo.1b,3 Characteristic of the pseudolaric acids is
the unusual perhydroazulene skeleton bearing trans-fused
acetoxy and lactone groups, which is an extremely rare
arrangement for naturally occurring hydroazulenes. The
installation of four contiguous stereocenters, of which three
are quaternary, and the compactness of the molecule add to
the synthetic challenge posed by this structure. Owing to
their intriguing architecture and promising biological activi-
ties, a number of synthetic studies on the pseudolaric acids
have appeared in the literature.4-8
Our strategy retroanalyzed the target molecules back to a
common precursor, enol triflate 2, from which both pseudola-
ric acids A and B, with methyl and carbomethoxy groups,
respectively, at the C7 position, can be derived via transition
metal chemistry (Scheme 1).9 This retrosynthetic analysis
permits flexibility of substitution at C7, as well as at C11, to
Scheme 1. Retrosynthetic Analysis of Pseudolaric Acid
A and B
(1) (a) Zhou, B. N.; Ying, B. P.; Song, G. Q.; Chen, Z. X.; Han, J.; Yan,
Y. F. Planta Med. 1983, 47, 35. (b) Hamburger, M. O.; Shieh, H. L.; Zhou,
B. N.; Pezzuto, J. M.; Cordell, G. A. Magn. Reson. Chem. 1989, 27, 1025.
(2) Li, E.; Clark, A. M.; Hufford, C. D. J. Nat. Prod. 1995, 58, 57.
(3) Pan, D. J.; Li. Z. L.; Hu, C. Q.; Chen, K.; Chang, J. J.; Lee, K. H.
Planta Med. 1990, 56, 383.
10.1021/ol0159165 CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/05/2001

append different substituents for the synthesis of pseudolaric
acid analogues. Enol triflate 2 can be obtained via a reductive
elimination from the oxatricyclic ketone 3, in which the
tertiary acetate or alcohol has been masked as an oxygen
bridge. This oxatricyclic ketone 3 was envisioned as the key
intermediate that could be constructed by a reaction cascade
initiated by the decomposition of an appropriately function-
alized acyclic diazoketone 4. The metal carbene would
undergo cyclization intramolecularly with the carbonyl group
to form a cyclic carbonyl ylide, followed by intramolecular
[3 + 2] cycloaddition with the 2,2-disubstituted olefin to
give the oxatricyclic intermediate.10
this reaction. We synthesized another model substrate that
bears a methyl substituent R to the carbonyl group and found
that the cyclization-cycloaddition sequence produced com-
pound 7 as the major diastereomer with a selectivity of about
4:1. Gratifyingly, oxatricyclic product 7 has the substituent
R) Me cis with respect to the bridgehead substituent (X )
CH2OBn), as required for the synthesis of the pseudolaric
acids (Scheme 2).12 With these preliminary results in hand,
we proceeded to the enantioselective synthesis of the
pseudolaric acids via a chiral synthesis of precursor 4.
The optically pure cycloaddition precursor 4 was as-
sembled as shown in Scheme 3. (2-Chloromethyl-allyloxy-
Previous studies in this area have established that the
tandem metal carbene cyclization-cycloaddition sequence
produces an adduct 5, in which the bridgehead substituent
(X ) H) and the oxygen bridge are trans (Scheme 2).11 Our
Scheme 3 Synthesis of Precursor 4a
Scheme 2. Preliminary Studies on the
Cyclization-Cycloaddition Cascade
a Ref 11a. b Ref 13.
preliminary studies with a model substrate confirmed that
the cyclization-cycloaddition cascade afforded cycloadduct
6, in which a substituent (X ) CH2OBn) that can be oxidized
to give the required carboxylate at C10 is also trans with
respect to the oxygen bridge. Thus the trans-arrangement
of the carboxylate at C10 and the acetoxy group at C4 in the
pseudolaric acid nucleus should be attainable via this cascade
reaction.
a Reagents and Conditions: (a) LiBr, Aliquat 336, 60 °C, 2 h,
97%; (b) IZnCH2CH2CO2Et 9, CuCN, THF, DMA, rt, 88%; (c)
NaOH, MeOH, 98%; (d) t-BuCOCl, Et3N, DMAP, (S)-4-benzyl-
2-oxazolidone, THF, -78 °C to room temperature, 80%; (e) (i)
n-Bu2BOTf, Et3N, CH2Cl2, 0 °C, (ii) CH3CHO, -78 to 0 °C, 1 h,
67%, (92% based on recovered substrate); (f) MEMCl, DIPEA,
CH2Cl2, rt, 93%; (g) (i) LiOH, H2O2, THF-H2O, 0 °C, (ii) Na2SO3,
70%, 16 also obtained in 21% yield. (h) EtSH, DCC, DMAP,
CH2Cl2, rt, 91%; (i) ClMg(CH2)3OMgCl 17, CuI, THF, 77%, (93%
based on recovered substrate); (j) PDC, DMF, H2O, 75%; (k)(i)
i-BuOCOCl, Et3N, THF, -20 °C, (ii) CH2N2, Et2O, 0 °C to room
temperature, 72%.
However, it was not clear to what extent, if any, was the
directing effect exerted by a substituent R to the carbonyl
group of acyclic compound 4 on the diastereoselectivity of
(4) Pan, B. C.; Chang, H. Y.; Cai, G. L.; Guo, Y. S. Pure Appl. Chem.
1989, 61, 389.
(5) Higuchi, R. I. Ph.D. Dissertation, Stanford University, Stanford, CA,
1995.
(6) Bonk, J. D. Ph.D. Dissertation, University of Mississippi, 1997.
(7) Chiu, P.; Chen, B.; Cheng, K. F. Tetrahedron Lett. 1998, 39, 9229.
(8) (a) Hu, Y.; Ou L.; Bai, D. Tetrahedron Lett. 1999, 40, 545. (b) Ou,
L.; Hu, Y.; Song, G.; Bai, D. Tetrahedron 1999, 55, 13999.
(9) Review: Ritter, K. Synthesis 1993, 737. Methylation of the vinyl
triflate can be achieved using Me2CuLi: McMurry, J. E.; Scott, W. J.
Tetrahedron Lett. 1980, 21, 4313. Carbomethoxylation can be accomplished
by Pd-catalyzed carbonylation: Cacchi, S.; Morera, E.; Ortar, G. Tetrahe-
dron Lett. 1985, 26, 1109.
(10) Reviews: (a) Padwa, A.; Weingarten, M. D. Chem. ReV. 1996, 96,
223. (b) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic Methods
for Organic Synthesis with Diazo Compounds; J. Wiley & Sons: New York,
1998; Chapter 7, pp 397-416.
methyl)benzene14 underwent halide exchange under phase
transfer catalysis to afford bromide 8. The zinc homoenolate
9 was then alkylated with allylic bromide 8 to give an 88%
(12) However, there is also a case in which the cyclization-cycloaddition
of an R-substituted isomu¨nchnone diazo precursor generated exclusively a
product with opposite diastereoselectivity: Maier, M. E.; Evertz, K.
Tetrahedron Lett. 1988, 29, 1677.
(13) The major isomer, 7, can be clearly identified by a NOE between
the methyl and the benzyloxymethylene protons: Ko, R. Y. Y., unpublished
results.
(11) Leading references: (a) Padwa, A.; Hornbuckle, S. F.; Fryxell, G.
E.; Stull, P. D. J. Org. Chem. 1989, 54, 817. (b) McMills, M. C.; Zhuang
L; Wright, D. L. Watt, W. Tetrahedron Lett. 1994, 35, 8311. (c) Dauben,
W. G.; Dinges, J.; Smith, T. C. J. Org. Chem. 1993, 58, 7635.
(14) van der Louw, J.; van der Baan, L. L.; de Kanter, F. J. J.;
Bickelhaupt, F.; Klumpp, G. W. Tetrahedron 1992, 48, 6087.
1722
Org. Lett., Vol. 3, No. 11, 2001

yield of the homologated ester 10. The alkylation of
homoenolate 9 also proceeded with 2-chloromethyl-allyl-
oxymethyl)benzene as the electrophile, albeit with a 50%
yield.
to give, in one step, the oxatricyclic product 3 in 66% yield
as two diastereomers 3a and 3b, in a ratio of 1.25:1 (Scheme
4).
The absolute stereochemistry of the R substituent in 4 was
installed by an enantioselective aldol reaction with acetal-
dehyde via Evan’s auxiliary methodology.15 Thus ester 10
was hydrolyzed and activated to couple with (S)-4-benzyl-
2-oxazolidone. The acylated oxazolidinone 11 was converted
to the boron enolate, which underwent the subsequent aldol
condensation at -78 °C with acetaldehyde to provide aldol
adduct 12 as one diastereomer in 67% yield, with recovered
11 in 27% yield.16 The hydroxyl group of the aldol adduct
was efficiently protected with the MEM group to give
compound 13.
Scheme 4. Rhodium-Catalyzed Cyclization-Cycloaddition of 4
The subsequent cleavage of the chiral auxiliary encoun-
tered some unanticipated hurdles. Our initial intention was
to extend the carbonyl terminus of compound 13 by the
addition of a three carbon nucleophile to its Weinreb amide
derivative. Although conditions have been developed for the
direct transamination of the acylated oxazolidone to the
Weinreb amide via treatment with alkylaluminum amide
reagents,17 these methodologies failed with the hindered
substrates 12 and 13.18 The products resulted predominantly
from the attack of the amide at the oxazolidone carbonyl
group. Fortunately, the hydrolysis of acylated oxazolidone
13 via lithium hydroperoxide was more successful and gave
a 70% yield of acid 14, although a 21% yield of alcohol 16
derived from oxazolidone carbonyl attack was still obtained.19
The plan to homologate acid 14 via its Weinreb amide
was again thwarted by the inefficient conversion of 14 to
the amide.20 The chain extension was finally accomplished
via thioester 15, which could be prepared from acid 14 in
excellent yield. The thioester proceeded to undergo chain
extension via the cuprate derived from Grignard reagent 1721
to afford alcohol 18.22 Oxidation with PDC in the presence
of water generated acid 19. The preparation of the diazoke-
tone 4 was accomplished by activation and treatment with
diazomethane.
In metal carbene reactions, the fact that the metal catalyst
is still associated with the carbene after diazo decomposition
is evidenced by the modulation of the chemoselectivity of
carbene reactions using metal catalysts of varying strengths
and by the ability of chiral catalysts to induce enantioselective
carbene reactions.23,24 Therefore, variations in the ligands of
the rhodium catalyst could influence the stereochemical
outcome of the cyclization-cycloaddition cascade. The use
of another catalyst, Rh2(cap)4, was attempted in the intramo-
lecular tandem cyclization-cycloaddition of diazoketone
compound 4. The reaction rate was noticeably slower than
with rhodium acetate as a result of the ligands being less
electron-deficient lactams. In the event, Rh2(cap)4 gave the
cyclized product 3 in a comparable yield, with the ratio of
diastereomers being slightly more in favor of the minor
diastereomer 3b than with Rh2(OAc)4.
The elucidation of the structures of the diastereomers was
hampered by the difficulty in separating the two compounds
1
of similar polarities. Furthermore, the H NMR signals of
the key protons H3 and H20 overlapped with the peaks of
other protons in the molecule (Scheme 5). Derivatization of
3a and 3b was done by deprotection of the MEM group
followed by oxidation to produce diastereomeric diketones
20a and 20b respectively. When the 2-D NOESY spectra of
these pure diketones were obtained, enhancement was
Treatment of the chiral precursor 4 with rhodium acetate
induced the decomposition of the diazoketone and initiated
the tandem intramolecular cyclization-cycloaddition cascade
(15) (a) Cowden, C. J.; Paterson, I. Org. React. 1997, 51, 1. (b) Evans,
D. A. Aldrichimica Acta 1982, 15, 23. (c) Evans, D. A.; Takacs, J. M.;
McGee, L. R.; Ennis, M. D.; Mathre, D. J.; Bartroli, J. Pure Appl. Chem.
1981, 53, 1109.
(16) Gage, J. R.; Evans, D. A. Org. Synth. 1990, 68, 83.
(17) (a) Evans, D. A.; Bender, S. L.; Morris, J. J. Am. Chem. Soc. 1988,
110, 2506. (b) Evans, D. A.; Bender, S. L. Tetrahedron Lett. 1986, 27,
799. (c) Shimizu, T.; Osako, K.; Nakata, T. Tetrahedron Lett. 1997, 38,
2685.
Scheme 5. Structure Elucidation of 3a and 3b via 20a and 20b
(18) Reactions of substrates 9 or 10 with the aluminum amides derived
from Me3Al, Me2AlCl, Et2AlCl uniformly failed to produce the desired
Weinreb amide.
(19) Compound 12 is presumably derived from the attack of the
oxazolidone carbonyl by peroxide, reduction by sufite to the acyl carbamic
acid, and decarboxylation.
(20) Using either DCC or BOP as coupling reagents, amide formation
required 3-4 days and resulted in product yields of 44% and 36%,
respectively.
(21) Normant, J. F.; Cahiez, G.; Alexakis, A. Tetrahedron Lett. 1978,
33, 3013.
(22) Anderson, R. J.; Henrick, C. A.; Rosenblum, L. D. J. Am. Chem.
Soc. 1974, 96, 3654.
Org. Lett., Vol. 3, No. 11, 2001
1723

observed for protons H3 and H20 in the spectrum of
compound 20a derived from the major isomer. On the other
hand, the spectrum of pure 20b showed the absence of an
NOE cross-peak between the corresponding protons, thereby
confirming the assignment of the structures of 3a and 3b to
be as shown. Hence oxatricyclic compound 3b is the
intermediate to proceed with the synthesis.
It is not clear at this point why the desired diastereose-
lectivity deteriorated in the reaction of substrate 4, compared
with a similar precursor bearing a methyl substituent that
was selective for the desired isomer 7. The factors governing
the observed diastereoselectivity in this cascade reaction will
need to be further elucidated in order to improve the
selectivity for the desired intermediate 3b. Efforts will be
directed toward completing the synthesis by the opening of
the oxabicyclic core and the addition of a diene nucleophile
to C11. These results will be reported in due course.
Acknowledgment. We thank Ms. Dianna C. Y. Poon for
the large-scale synthesis of substrate 10 and Ms. Rebecca
Y. Y. Ko for preliminary experiments. The work described
in this paper was supported by a grant from the Research
Grants Council of Hong Kong Special Administrative
Region, China (Project HKU 7098/97P). B.C. thanks the
University of Hong Kong for an exchange scholarship under
the PRC Nominated Research Students Program.
Supporting Information Available: Experimental pro-
1
cedures; H NMR, 13C NMR, IR, and MS spectral data for
(23) (a) Padwa, A.; Austin, D. J.; Hornbuckle, S. R.; Semones, M. A.;
Doyle, M. P.; Protopopova, M. N. J. Am. Chem. Soc. 1992, 114, 1874. (b)
Padwa, A.; Austin, D, J.; Hornbuckle, S. F. Tetrahedron Lett. 1992, 33,
6427.
(24) (a) Kitagaki, S.; Anada, M.; Kataoka, O.; Matsuno, K.; Umeda, C.;
Watanabe, N.; Hashimoto, S. J. Am. Chem. Soc. 1999, 121, 1417. (b)
Hodgson, D. M.; Stupple, P. A.; Johnstone, C. Chem. Commun. 1999, 2185.
1
all products; and H and 13C NMR spectra of compounds
3b, 4, 6-8, 10-16, 18-19, 20a, and 20b. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL0159165
1724
Org. Lett., Vol. 3, No. 11, 2001