Palladium-Catalyzed DYKAT of Vinyl Epoxides: Enantioselective Total Synthesis and Assignment of the Configuration of (+)-Broussonetine G

Article abstract of DOI:10.1002/anie.200352857

A potential general approach to potent N-heterocyclic glycosidase inhibitors is illustrated by the first total synthesis of (+)-broussonetine G (1). The key transformation of the synthesis relies on two successive palladium-catalyzed asymmetric allylic al

Full text of DOI:10.1002/anie.200352857

Angewandte
Chemie
Natural Product Synthesis
Palladium-Catalyzed DYKAT of Vinyl Epoxides:
Enantioselective Total Synthesis and Assignment
of the Configuration of (+)-Broussonetine G**
Barry M. Trost,* Daniel B. Horne, and
Michael J. Woltering
The ability of polyhydroxylated pyrrolidines and pyrrolizi-
dines to inhibit numerous glycosidases continues to draw
attention to such structural types.[1] The broussonetines are a
class of polyhydroxylated alkaloids that were isolated from
the branches of the deciduous tree Broussontia kazinoki.
Thus far, G. Kusano and co-workers have isolated and
characterized 29 unique compounds from this family of
natural products.[2] These compounds show striking glycosi-
dase-inhibitory properties, and as such have enormous
therapeutic potential as antitumor and anti-HIV agents.[3]
The members of the broussonetine family all share a
polyhydroxylated pyrrolidine core and a 13-carbon-atom
side chain that contains various functional groups.
(+)-Broussonetine G (1), one of the most active inhibitors
of this class (IC50 = 3 nm, b-galactosidase; 24 nm, b-glucosi-
dase), contains a structurally interesting 5,6-spiroketal in the
side chain.[4] The synthesis of 1 presents an additional
[*] Prof. B. M. Trost, D. B. Horne, M. J. Woltering
Department of Chemistry, Stanford University
Stanford, CA 94305-5080 (USA)
Fax:(+1)650-725-0002
E-mail: bmtrost@stanford.edu
[**] DYKAT= dynamic kinetic asymmetric transformation. We thank the
National Science Foundation and the National Institutes of Health
(GM-13598), the Eli Lilly Corporation (fellowship to DBH), and the
Deutsche Forschungsgemeinschaft (fellowship to MJW) for their
generous support of our programs. We also thank Dr. Maiko
Shibano and Prof. G. Kusano (Osaka University of Pharmaceutical
Sciences) for NMR spectra and their donation of a sample of natural
broussonetine G. Mass spectra were provided by the Mass
Spectrometry Regional Center of the University of California-San
Francisco supported by the NIH Division of Research.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 5987 –5990
DOI: 10.1002/anie.200352857
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5987

Communications
challenge in that the assignment of the relative stereochem-
istry of the spiroketal and the 1’-hydroxy group has not been
reported.
The synthesis of the side chain began with the addition of
1-pentyne to d-valerolactone, followed by protection of the
primary alcohol with trityl chloride to give alkynone 8
(Scheme 2). Reduction of 8 by using Noyori's enantioselec-
tive transfer hydrogenation protocol[9] gave propargyl alcohol
Despite the interesting structural features and biological
activity of these compounds, only one synthesis has been
reported: Yoda et al. described the chiral-pool-based syn-
thesis of broussonetine C (2).[5] Herein we describe an
enantioselective synthesis that involves a palladium-catalyzed
dynamic kinetic asymmetric transformation (DYKAT)[6]
process, which potentially allows access to all members of
the broussonetine family as well as many other pyrrolidine-
containing glycosidase inhibitors, including pyrrolizidine and
indolizidine types.
We envisioned that two successive palladium-catalyzed
asymmetric allylic alkylations of butadiene monoxide with a
suitable nitrogen nucleophile followed by ring-closing meta-
thesis (RCM) could provide a unique access to chiral,
substituted pyrrolidines. This strategy could potentially be
used to provide all diastereomers of the proposed structure, a
key feature for us since the full stereochemistry of our target
was not known. Furthermore, we hoped that this strategy
could also be used to access many other important polysub-
stituted pyrrolidine alkaloid natural products, such as
(2R,3R,4R,5R)-2,5-dihydroxymethyl-3,4-dihydroxypyrroli-
dine (DMDP; 3).[7]
Scheme 2. Synthesis of the spiroketal side chain 11. Reagents and condi-
tions: a) nBuLi, 1-pentyne, THF, ꢀ788C!RT, 94%; b) TrCl, NEt3, DMAP,
DMF, room temperature, 77%; c) [(h6-p-cymene)Ru{(1R,2R)-p-
TsNCH(Ph)CH(Ph)NH}] (3 mol%), iPrOH, room temperature, 95%,
97% ee; d) KH (10 equiv), 1,3-diaminopropane, THF, 79%; e) 3,4-dihydro-
2H-pyran, PPTS, CH2Cl2, room temperature, 76%; f) nBuLi, AlMe3,
BF3·OEt2, Et2O, ꢀ788C; then ethylene oxide, ꢀ788C!RT, 76%; g) HCl/
MeOH (1%), 95%; h) [PdCl2(PhCN)2] (2 mol%), CH3CN/THF (3:2), 85%,
97:3 d.r.; i) PPh3Br2, imidazole, THF, 91%. Tr=trityl=triphenylmethyl,
DMAP=4-dimethylaminopyridine, DMF=N,N-dimethylyformamide,
Ts =p-toluenesulfonyl, PPTS=pyridinium p-toluenesulfonate.
The synthesis began with the addition of phthalimide to
butadiene monoxide to give 4[8] in a highly regio- and
enantioselective fashion (94% yield, 98% ee) (Scheme 1).
After deprotection of the diimide to give a vinylglycinol,
formation of the cyclic carbamate with triphosgene gave
vinyloxazolidinone 5, the substrate for the second palladium-
catalyzed DYKAT reaction. Despite our attempts with
various other vinylglycinol surrogates in this reaction, we
found the best results were obtained with oxazolidinone 5,
which gave the desired product 6 in 91% yield with excellent
diastereoselectivity (d.r. 93:7). We observed that the diaster-
eoselectivity of the product is completely dependent on the
choice of ligand (see Table in Scheme 1).
9 (97% ee). The alkyne zipper reaction[10] (KH, 1,3-diamino-
propane) was performed to give the desired terminal alkyne.
Protection of the secondary alcohol followed by alkylation of
the terminal alkyne with ethylene oxide provided the desired
homopropargyl alcohol. After acid-catalyzed global depro-
tection, the resulting triol 10 underwent the requisite regio-
selective palladium-catalyzed spiroketalization reaction[11] to
give the desired 5,6-spiroketal with excellent diastereoselec-
tivity (d.r. 97:3). The high stereoselectivity of this cyclization
reaction is attributed to the anomeric affect.[12] Subsequent
bromination of the primary alcohol gave bromide 11.
The synthesis of the pyrrolidine core continued with the
protection of the homoallylic alcohol with benzyl bromide,
followed by RCM with a Grubbs second-generation cata-
lyst[13] 12 to give 2,5-dihydropyrrole 13 (Scheme 3). The
absolute stereochemistry was assigned based on the stereo-
chemical assignment of vinylglycinol performed previously[8]
and single-crystal X-ray analysis of oxazolidinone 13.[14]
Unfortunately, oxidation of the primary alcohol derived
from 13 after deprotection of the benzyl group led only to
decomposition products. We surmised that the oxazolidinone
moiety was probably responsible for the lability of the
aldehyde intermediate and therefore adopted an alternative
pathway. Saponification of the oxazolidinone and protection
of the resulting amine with BnOCOCl gave alcohol 14.
Subsequent oxidation with Dess–Martin periodinane deliv-
ered the aldehyde 15. Our attempts to add various alkyl
nucleophiles to aldehyde 15 resulted in very poor yields of the
desired products and always produced a 1:1 diastereomeric
ratio of alcohols.
Scheme 1. Synthesis of the chiral core subunit. Reagents and conditions:
a) [{(C3H5)PdCl}2] (0.4 mol%), (R,R)-7 (1.2 mol%), Na2CO3, CH2Cl2, room
temperature, 94%, 98% ee; b) ethylenediamine, EtOH, reflux, 84%;
c) triphosgene, NaHCO3, toluene/H2O, 08C, 85%; d) [Pd2(dba)3]·CHCl3
(0.25 mol%), (R,R)-7 (0.75 mol%), DBU(1 mol%), CH 2Cl2, room
temperature, 91%, 93:7 d.r. dba=trans,trans-dibenzylideneacetone,
DBU=1,8-diazabicyclo[5.4.0]undec-7-ene.
The failure of the coupling reaction led us to consider an
alternate strategy. Oxidation of alcohol 14 (TEMPO/bleach/
5988
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5987 –5990

Angewandte
Chemie
single diastereomer upon hydrolysis with aqueous TFA. To
complete the synthesis, removal of the protecting groups by
hydrogenolysis gave 1, whose spectra and optical rotation
matched those of an authentic sample of the natural product
in all respects. This constitutes the first synthesis of (+)-
broussonetine G.
In addition to the natural product, the remaining three
possible diastereomers of broussonetine G were synthesized
during the course of our investigations by using the proce-
dures outlined above (Schemes 1–4). Diastereomers 19 and
20 did not correlate with the spectroscopic data of the natural
1
product (Scheme 5). However, the H and 13C NMR spectra
Scheme 3. Synthesis and attempted coupling of aldehyde 15. Reagents and
conditions: a) NaH; BnBr, TBAI, THF, room temperature, 84%; b) 12
(1.2 mol%), CH2Cl2, reflux, 87%; c) 1) NaOH, EtOH/H2O (3:1), reflux;
2) BnOCOCl, NaHCO3, Na2CO3, H2O, 08C!RT, 99% from 13; d) Dess–
Martin periodinane, CH2Cl2, 08C, 95%; TBAI=tetrabutylammonium iodide;
Cbz=benzyloxycarbonyl.
KBr) led to a carboxylic acid, which was converted into
Weinreb amide 16 (Scheme 4). Gratifyingly, the subsequent
coupling of the two fragments via the alkyl magnesium
reagent derived from alkyl bromide 11 led to a ketone.
Scheme 5. [a]D Values of diastereomers 1, 19, 20, and 21.
of 21 were nearly identical to that of 1 and the natural
product.[17] Fortunately, the optical rotation was significantly
different and led to our assignment of the structure of
broussonetine G as shown (Scheme 5).
Scheme 4. Completion of the synthesis of 1. Reagents and conditions:
a) TEMPO, KBr, NaOCl, NaHCO3, acetone/H2O, 08C, 89%; b) HNMe-
(OMe)·HCl, pybop, iPr2NEt, CH2Cl2, 81%; c) 1) 11 (1.2 equiv), Mg, THF,
reflux!RT; 16, room temperature, 65%; d) DIBAL-H, Et2O, 08C, 76%, 4.3:1
d.r.; e) F3CC(O)CH3, oxone, Na2CO3, CH3CN/H2O, 08C, 68%; f) TFA, THF/
H2O, 658C, 73%; g) Pd/C, MeOH, HCl, H2 (1 atm), room temperature,
95%; TEMPO=2,2,6,6-tetramethyl-1-piperidinyloxy, free radical, pybop=
(benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate,
DIBAL-H=diisobutylaluminum hydride, TFA=trifluoroacetic acid.
Received: September 12, 2003 [Z52857]
Keywords: asymmetric synthesis · inhibitors · natural products ·
.
structure elucidation · total synthesis
[1] For reviews on the importance and use of glycosidase inhibitors,
see: a) Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond (Ed.: A. E. Stutz), Wiley-VCH, Weinheim, 1999; b) N.
Asano, R. J. Nash, R. J. Molyneux, G. W. Fleet, Tetrahedron:
Asymmetry 2000, 11, 1645.
[2] For a recent review of the isolation, biology, and chemistry of
broussonetines, see: M. Shibano, D. Tsukamoto, G. Kusano,
Heterocycles 2002, 57, 1539.
[3] a) R. Saul, J. P. Chambers, R. J. Molyneux, A. D. Elbein, Arch.
Biochem. Biophys. 1983, 221, 593; b) R. Saul, J. P. Chambers,
R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys. 1984,
230, 668; c) B. Winchester, G. W. Fleet, J. Carbohydr. Chem.
2000, 19, 471.
[4] M. Shibano, S. Nakamura, N. Akazawa, G. Kusano, Chem.
Pharm. Bull. 1998, 46, 1048.
[5] H. Yoda, T. Shimojo, K. Takabe, Tetrahedron Lett. 1999, 40,
1335.
[6] a) B. M. Trost, Chem. Pharm. Bull. 2002, 50, 1; b) B. M. Trost,
D. L. Van Vranken, C. Bingel, J. Am. Chem. Soc. 1992, 114, 9327.
[7] For a review on the synthesis and biology of DMDP, see: T. M.
Wrodnigg, Monatsh. Chem. 2002, 133, 393.
Subsequent diastereoselective reduction of the ketone with
DIBAL-H gave a 4.3:1 (R/S) diastereomeric mixture of
secondary alcohols. The major diastereomer, alcohol 17,
presumably results from a four-centered transition state.[15] In
contrast, reduction of the ketone with NaBH4 led to a
diastereomeric mixture of alcohols (R/S 1:2.6, 83%), the
major product of which was the expected Cram product. The
stereochemistry of the secondary alcohol was established by
1H NMR analysis of the (R)- and (S)-O-methylmandelates.[16]
The assignment of the stereochemistry of the C1’ center as R
also agrees with the assignment of several other broussone-
tines that contain an hydroxy group at C1’ whose config-
urations were determined by Kusano et al. by using the
Mosher method.[2] Epoxidation of dihydropyrrole 17 led to a
mixture of epoxides, both of which gave the same triol 18 as a
Angew. Chem. Int. Ed. 2003, 42, 5987 –5990
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5989

Communications
[8] B. M. Trost, R. C. Bunt, R. Lemoine, T. L. Calkins, J. Am. Chem.
Soc. 2000, 122, 5968.
[9] K. Matsumura, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am.
Chem. Soc. 1997, 119, 8738.
[10] C. A. Brown, A. Yamashita, J. Am. Chem. Soc. 1975, 97, 891.
[11] K. Utimoto, Pure Appl. Chem. 1983, 55, 1845.
[12] P. Deslongchamps, Stereoelectronic Effects in Organic Synthesis,
Pergamon, Elmsford, NY, 1983.
[13] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1,
953.
[14] CCDC-218668 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cam-
bridge Crystallographic Data Centre, 12, Union Road, Cam-
bridge CB21EZ, UK; fax: (+ 44)1223-336-033; or deposit@
ccdc.cam.ac.uk).
[15] a) M. M. Midland, Y. C. Kwon, J. Am. Chem. Soc. 1983, 105,
3725.b) S. Ono, S. Shuto, A. Matsuda, Tetrahedron Lett. 1996, 37,
221.
[16] B. M. Trost, J. L. Belletire, S. Godleski, P. G. McDougal, J. M.
Balkovec, J. J. Baldwin, M. E. Christy, G. S. Ponticello, S. L.
Varga, J. P. Springer, J. Org. Chem. 1986, 51, 2370.
[17] For an example of diastereoisomeric products with similar NMR
spectra, see the synthesis of mosin B: N. Maezaki, N. Kojima, A.
Sakamoto, H. Tominaga, C. Iwata, T. Tanaka, M. Monden, B.
Damdinsuren, S. Nakamori, Chem. Eur. J. 2003, 9, 389.
5990
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5987 –5990