Total synthesis and absolute stereochemistry of polygalolides A and B

Article abstract of DOI:10.1002/anie.200602030

(Chemical Equation Presented) Carbonyl ylides to the fore: the first total synthesis of polygalolides A (1) and B (2) has been achieved by exploiting a tandem carbonyl ylide formation/1,3-dipolar cycloaddition to construct the unprecedented 5,10-dioxatric

Full text of DOI:10.1002/anie.200602030

Communications
takes advantage of a transformation consisting of a tandem
carbonyl ylide formation/1,3-dipolar cycloaddition, which has
been extensively studied, especially in the research group of
Padwa.[2–5]
Our retrosynthetic analysis of the polygalolides is illus-
trated in Scheme 1. We elected to introduce the aryl-
methylidene moiety at a late stage in the synthesis, and the
Total Synthesis
DOI: 10.1002/anie.200602030
Total Synthesis and Absolute Stereochemistry of
Polygalolides A and B**
Seiichi Nakamura, Yukihito Sugano, Fumiaki Kikuchi,
and Shunichi Hashimoto*
In 2003, Wei and co-workers reported the isolation and
structural elucidation of the polygalolides A (1) and B (2)
obtained from Polygala fallax Hemsl., a medicinal plant from
which extracts are used as tonics and antihepatitis drugs in
China.[1] The structure and relative configuration of the
polygalolides were determined on the basis of NMR spectro-
scopic data. However, the issue of the absolute stereochem-
istry of the molecules has not been resolved. The structural
complexity of these molecules, which include an unprece-
dented trioxatetracyclic ring systemand contiguous quater-
nary stereogenic centers at C2 and C8, poses a considerable
synthetic challenge. Herein, we describe the first total syn-
thesis of (ꢀ)-polygalolides A (1) and B (2). The synthesis
Scheme 1. Retrosynthetic analysis of polygalolides A and B.
PMP=p-methoxyphenyl, TBDPS=tert-butyldiphenylsilyl.
scission of the lactone ring provided 3 as a common
intermediate. The tricyclic compound 3 was assumed to
arise froman intramolecular 1,3-dipolar cycloaddition of
carbonyl ylide 5, generated from a-diazoketone 6 in the
presence of a RhII catalyst. Compound 6 arises from
homologation of the tert-butyl ester 7, which would be
elaborated fromthe known alcohol 8,[6] which in turn can
readily be obtained from d-arabinose.
We initiated the synthesis by alkylating alcohol 8 with (Z)-
1-bromo-4-(tert-butyldiphenylsilyl)oxy-2-butene[7] to provide
the allyl ether 9 in 98% yield (Scheme 2). The oxidative
hydrolysis of the dithioacetal with iodine was followed by
oxidation with NaClO2 and esterification with (Boc)2O[8] to
give tert-butyl ester 10 in 70% yield over three steps. Catalytic
hydrogenation of alkene 10 provided the TBDPS ether 11 in
95% yield, and 11 was then treated with Bu4NF in the
presence of AcOH to give alcohol 12 in 96% yield.
[*] Dr. S. Nakamura, Y. Sugano, F. Kikuchi, Prof. Dr. S. Hashimoto
Faculty of Pharmaceutical Sciences
Hokkaido University
Sapporo 060-0812 (Japan)
Fax: (+81)11-706-3236
E-mail: hsmt@pharm.hokudai.ac.jp
[**] This research was in part supported by a Grant-in-Aid for Scientific
Research on Priority Areas 17035002 from the Ministry of
=
Installation of the C2 C3 bond was accomplished by employ-
Education, Culture, Sports, Science and Technology (MEXT). We are
grateful to Dr. X. Wei for providing spectra of the natural products.
We thank H. Matsumoto, A. Maeda, S. Oka, M. Kiuchi, and T. Hirose
of the Center for Instrumental Analysis, Hokkaido University, for
technical assistance in obtaining MS and elemental analyses.
ing a modification of the procedure reported by Ogasawara
and co-workers,[9] and subsequent reduction with NaBH4 gave
alcohol 13 in 81% yield over both steps. A two-step sequence
involving mesylation and nucleophilic substitution with p-
methoxyphenol was used to convert the allyl alcohol 13 into
the PMP ether 14 in 78% overall yield. The acetonide was
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
6532
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6532 –6535

Angewandte
Chemie
Scheme 2. Synthesis of cyclization precursor 6. a) (Z)-1-bromo-4-(tert-
butyldiphenylsilyl)oxy-2-butene, NaH, THF/DMF (10:1), 1.5 h, 98%;
b) I2, NaHCO3, aq acetone, ꢀ258C, 5 min, 88%; c) NaClO2, NaH2PO4,
2-methyl-2-butene, aq tBuOH, 6 h; d) (Boc)2O, DMAP, tBuOH, 1 h,
80% (2 steps); e) H2, 10% Pd/C, EtOAc, 13 h, 95%; f) Bu4NF, AcOH,
THF, 15 h, 96%; g) (COCl)2, DMSO, Et3N, CH2Cl2, ꢀ78!08C, then
=
Me2N CH2I, DBU, 20 h, 85%; h) NaBH4, EtOH, 08C, 30 min, 95%;
i) MsCl, Et3N, CH2Cl2, 08C, 10 min; j) 4-MeOC6H4OH, K2CO3, MeCN,
reflux, 6 h, 78% (2 steps); k) AcOH/THF/H2O (3:1:1), 408C, 4 h, 95%;
l) TBDPSCl, imidazole, CH2Cl2, 1 h, 92%; m) Dess–Martin periodinane,
CH2Cl2, 1 h, 93%; n) TFA/CH2Cl2 (1:10), 08C, 1 h; o) ClCO2iBu, Et3N,
Et2O, 08C, 10 min, then CH2N2, 90 min, 55% (2 steps). Boc=tert-
butoxycarbonyl, DMAP=4-dimethylaminopyridine, DBU=1,8-
diazabicyclo[5.4.0]undec-7-ene, Ms=methanesulfonyl, TFA=trifluoro-
acetic acid.
Scheme 3. Total synthesis of polygalolides A (1) and B (2). a) Rh2-
(OAc)4 (5 mol%), PhCF3, 1008C, 5 min, 73%; b) (NH4)2Ce(NO3)6,
pyridine, aq MeCN, 08C, 2 h, 91%; c) Dess–Martin periodinane,
CH2Cl2, 2 h, 92%; d) NaClO2, NaH2PO4, 2-methyl-2-butene, aq tBuOH,
6 h; e) CH2N2, Et2O, 08C, 30 min, 85% (2 steps); f) Bu4NF, AcOH,
THF, 20 h, 93%; g) LDA, ZnCl2, THF, ꢀ788C, 1 h, aldehyde 4a,
ꢀ788C, 30 min, then Ac2O, 08C, 1 h; h) SiO2, CH2Cl2, 24 h; i) NaHCO3,
aq MeOH, 18% (3 steps); j) TMSOTf, Et3N, CH2Cl2, 08C, 5 h, 90%;
k) 23a or 23b (5 equiv), TMSOTf, molecular sieves (3 Š), CH2Cl2,
ꢀ788C, 30 min, 58% (for 24a) or 57% (for 24b); l) DBU, CH2Cl2,
10 min; m) NaHCO3, aq MeOH, 85% (for 1) or 72% (for 2).
LDA=lithium diisopropylamide; TMS=trimethylsilyl; Tf=trifluoro-
methanesulfonyl.
removed using AcOH in aqueous THF at 408C to afford the
diol 15 in 95% yield. Selective silylation of the C9 hydroxy
group in 15 furnished the TBDPS ether 16 in 92% yield, and
16 was oxidized with Dess–Martin periodinane to give ketone
7
[10] in 93% yield. A carboxylic acid function at C5, available
for a one-carbon homologation, was liberated upon exposure
of the tert-butyl ester 7 to TFA. Although the carboxyl group
in 17 could not be converted into the acid chloride because of
the inevitable formation of the corresponding pseudo-
chloride,[11] activation of the acid through the intermediacy
of its mixed anhydride and subsequent treatment with CH2N2
afforded a-diazoketone 6 in 55% yield over the two steps.
With the cyclization precursor 6 in hand, we set out to
investigate the critical carbonyl ylide formation/intramolec-
ular 1,3-dipolar cycloaddition reaction sequence (Scheme 3).
The reaction involved the addition of a-diazoketone 6 (ca.
30 mg) over five minutes to a solution of a RhII catalyst.
Although 6 was completely consumed even at room temper-
ature in the presence of Rh2(OAc)4, the reaction gave a
complex mixture of products.[12] After some experimentation,
we found that cycloadduct 3[10] was produced as a single
isomer in 36% yield at 608C in toluene; the product yield was
improved to 68% by an increase in reaction temperature to
1108C. We screened the solvents and RhII catalysts[13] and
found Rh2(OAc)4 in benzotrifluoride[14] to be optimal, afford-
ing cycloadduct 3 in 73% yield.[15] After oxidative removal of
the PMP group at C1 with (NH4)2Ce(NO3)6,[16] two successive
oxidations and an esterification with CH2N2 gave the methyl
ester 19 in 78% yield over three steps. Desilylation and
concomitant lactone formation was effected with Bu4NF in
the presence of AcOH to provide the tetracyclic lactone 20.
With the construction of the common tetracyclic core
accomplished, we addressed the late-stage installation of the
arylmethylidene moiety. Our initial attempts focused on the
aldol reaction of ketone 20 with aldehyde 4a under basic
Angew. Chem. Int. Ed. 2006, 45, 6532 –6535
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6533

Communications
conditions. The formation of the aldol adduct was detected by
Keywords: aldol reaction · carbonyl ylides · cycloaddition ·
diazo compounds · total synthesis
.
1H NMR spectroscopy of the crude mixture when the zinc
enolate derived from 20 was reacted with 4a, but the desired
product readily underwent retroaldol reaction upon exposure
to silica gel. The coupling product 21 could be obtained along
with considerable amounts of starting material 20 when the
reaction mixture was treated with Ac2O. Although the total
synthesis of polygalolide A (1) could be completed through
exposure of the mixture to silica gel followed by deacetyla-
tion, the low efficiency of this sequence (18% yield over three
steps without intervening purification) left roomfor improve-
ment. Accordingly, the decision was made to employ the
Mukaiyama aldol-type reaction using dimethyl acetal 23a.[17]
Ketone 20 was easily converted into the silyl enol ether 22.
Gratifyingly, the TMSOTf-promoted coupling of 22 with 23a
in CH2Cl2 in the presence of molecular sieves (3 Š) at ꢀ788C
proceeded to provide coupling product 24a in 58% yield. The
b-methoxyketone 24a was smoothly converted into poly-
galolide A (1) by treatment with DBU, followed by
deacetylation. Polygalolide B (2) was also synthesized in
[1] W. Ma, X. Wei, T. Ling, H. Xie, W. Zhou, J.Nat.Prod. 2003, 66,
441 – 443.
[2] A. Padwa, G. E. Fryxell, L. Zhi, J.Am.Chem.Soc. 1990, 112,
3100 – 3109.
[3] For recent reviews, see: a) M. C. McMills, D. Wright in Synthetic
Applications of 1,3-Dipolar Cycloaddition Chemistry toward
Heterocycles and Natural Products (Eds.: A. Padwa, W. H.
Pearson), Wiley, New York, 2002, pp. 253 – 314; b) G. Mehta, S.
Muthusamy, Tetrahedron 2002, 58, 9477 – 9504; c) A. Padwa,
Helv.Chim.Acta 2005, 88, 1357 – 1374.
[4] For recent studies on the synthesis of bioactive molecules using a
carbonyl ylide cycloaddition strategy, see: a) H. Koyama, R. G.
Ball, G. D. Berger, Tetrahedron Lett. 1994, 35, 9185 – 9188; b) A.
Padwa, M. A. Brodney, J. P. Marino, Jr., S. M. Sheehan, J.Org.
Chem. 1997, 62, 78 – 87; c) C. S. Straub, A. Padwa, Org.Lett.
1999, 1, 83 – 85; d) D. M. Hodgson, C. Villalonga-Barber,
Tetrahedron Lett. 2000, 41, 5597 – 5600; e) J. L. Wood, B. D.
Thompson, N. Yusuff, D. A. Pflum, M. S. P. Matthäus, J.Am.
Chem.Soc. 2001, 123, 2097 – 2098; f) F. Liu, D. J. Austin, Org.
Lett. 2001, 3, 2273 – 2276; g) D. M. Hodgson, T. D. Avery, A. C.
Donohue, Org.Lett. 2002, 4, 1809 – 1811; h) B. Chen, R. Y. Y.
Ko, M. S. M. Yuen, K.-F. Cheng, P. Chiu, J.Org.Chem. 2003, 68,
4195 – 4205; i) S. Nakamura, Y. Hirata, T. Kurosaki, M. Anada,
O. Kataoka, S. Kitagaki, S. Hashimoto, Angew.Chem. 2003, 115,
41% yield fromintermediate
22 following an identical
reaction sequence. The spectroscopic data (1H and
13C NMR, IR, UV, and HRMS) of the products resulting
fromthese syntheses correspond to those reported for the
natural products, except for their specific rotations, which
were equal in sign, but with magnitudes inconsistent with
those previously reported: [a]D = ꢀ499.9 (c = 0.022 in
MeOH) and ꢀ505.2 (c = 0.018 in MeOH) compared with
[a]D = ꢀ14.4 (c = 0.018 in MeOH) and ꢀ21.3 (c = 0.015 in
MeOH).[1] This difference suggests that polygalolides might
be biosynthesized in near-racemic form. In this context, it is
interesting that Snider and Grabowski recently proposed in
their total syntheses of (ꢁ )-cartorimine and (ꢁ )-descurainin
that the 8-oxabicyclo[3.2.1]octenone skeleton of these mole-
5509 – 5513; Angew.Chem.Int.Ed.
2003, 42, 5351 – 5355; j) T.
Graening, V. Bette, J. Neudꢀrfl, J. Lex, H.-G. Schmalz, Org.Lett.
2005, 7, 4317 – 4320; k) H. Oguri, S. L. Schreiber, Org.Lett. 2005,
7, 47 – 50.
[5] For representative works on the related [5+2] cycloadditions
between oxypyryliumylides and alkenes, see: a) P. A. Wender,
H. Y. Lee, R. S. Wilhelm, P. D. Williams, J.Am.Chem.Soc. 1989,
111, 8954 – 8957; b) D. R. Williams, J. W. Benbow, E. E. Allen,
Tetrahedron Lett. 1990, 31, 6769 – 6772; c) W. E. Bauta et al., see
Supporting Information; d) P. Magnus, L. Shen, Tetrahedron
1999, 55, 3553 – 3560; e) B. B. Snider, J. F. Grabowski, Tetrahe-
dron Lett. 2005, 46, 823 – 825; f) B. B. Snider, J. F. Grabowski,
Tetrahedron 2006, 62, 5171 – 5177.
cules would be biosynthesized by a [5+2] cycloaddition
[5e,f]
between a oxypyryliumzwitterion and an alkene.
They
speculated that [5+2] cycloadditions in a chiral environment
could lead to optically enriched products as the very small
[a]D values of these natural products indicate that they are not
completely racemic. We surmise that polygalolides would also
be synthesized in the medicinal plant through [5+2] cyclo-
addition of the fructose-derived oxypyryliumzwitterion 25
with an isoprene derivative, followed by an intramolecular
hetero-Michael addition, a lactone formation, and an aldol
reaction with vanillin derivatives.[18] Our result would provide
an experimental proof for the intriguing speculation by the
Snider group.
In conclusion, we have accomplished the first total
synthesis of polygalolides A (1) and B (2) fromalcohol 8,
with a longest linear sequence of 25 steps and with overall
yields of 3.8% and 3.2%, respectively. This total synthesis
serves to confirmthe absolute stereochemistry of these
natural products. The synthesis illustrates the power of the
carbonyl ylide cycloaddition methodology for the rapid
assembly of the unusual dioxatricyclic ring system, which is
difficult to construct by other means.
[6] M. Y. H. Wong, G. R. Gray, J.Am.Chem.Soc. 1978, 100, 3548 –
3553.
[7] R. M. Lemieux, P. N. Devine, M. F. Mechelke, A. I. Meyers, J.
Org.Chem. 1999, 64, 3585 – 3591.
[8] K. Takeda, A. Akiyama, H. Nakamura, S. Takizawa, Y. Mizuno,
H. Takayanagi, Y. Harigaya, Synthesis 1994, 1063 – 1066.
[9] S. Takano, K. Inomata, K. Samizu, S. Tomita, M. Yanase, M.
Suzuki, Y. Iwabuchi, T. Sugihara, K. Ogasawara, Chem.Lett.
1989, 1283 – 1284.
[10] The homochirality of ketone 7 and cycloadduct 3 was confirmed
by HPLC on a chiral stationary phase.
[11] W. J. Elliott, J. Fried, J.Org.Chem. 1978, 43, 2708 – 2710. We
thank a referee for bringing the reference to our attention.
[12] The intermolecular reaction of diazoketone 6 with dimethyl
acetylenedicarboxylate occurred in the presence of Rh2(OAc)4
at room temperature to provide a 1:1 mixture of diastereomeric
cycloadducts in 35% yield, indicating that carbonyl ylide 5 was
formed under these conditions. Heating is required to promote
the intramolecular cycloaddition with a nonactivated alkene.
[13] Product yields obtained with Rh2(OAc)4 in other solvents:
benzene (57%), (CH2Cl)2 (31%). Other RhII catalysts afforded
lower product yields: Rh2(OCOC7H15)4 (trace), Rh2(OCOC3F7)4
(complex mixture), Rh2(OCOCPh3)4 (24%), Rh2(NHAc)4
(41%).
Received: May 22, 2006
Published online: September 5, 2006
[14] For the use of benzotrifluoride as an effective solvent in RhII-
catalyzed carbonyl ylide formation/1,3-dipolar cycloaddition
6534
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6532 –6535

Angewandte
Chemie
reactions, see: S. Kitagaki, M. Anada, O. Kataoka, K. Matsuno,
C. Umeda, N. Watanabe, S. Hashimoto, J.Am.Chem.Soc. 1999,
121, 1417 – 1418.
[15] The cycloaddition is not scale dependent and the cycloadduct 3
was obtained in 70% yield when the reaction was performed on
a 500-mg scale.
[16] T. Fukuyama, A. A. Laird, L. M. Hotchkiss, Tetrahedron Lett.
1985, 26, 6291 – 6292.
[17] a) T. Mukaiyama, M. Hayashi, Chem.Lett. 1974, 15 – 16; b) S.
Murata, M. Suzuki, R. Noyori, J.Am.Chem.Soc.
1980, 102,
3248 – 3249; for a review, see: T. Mukaiyama, M. Murakami,
Synthesis 1987, 1043 – 1054.
[18] A plausible biogenesis of polygalolides is illustrated schemati-
cally below.
Angew. Chem. Int. Ed. 2006, 45, 6532 –6535
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6535