Asymmetric synthesis of the 1-epi aglycon of the cripowellins A and B

Article abstract of DOI:10.1002/anie.200500556

(Chemical Equation Presented) The unusual [5.3.2]bicyclic structure of the insecticidal Amaryllidaceae alkaloids cripowellins A and B has been synthesized for the first time by a sequence of Sharpless dihydroxylation, ring-closing metathesis (RCM), and in

Full text of DOI:10.1002/anie.200500556

Communications
only with regard to their strength but also their broad activity.
It is important to note that mainly the aglycon seems to be
responsible for their biological activity because it alone shows
the same activity as the two glycosides.[2] However, practically
nothing is known about the mode of action.[4] In order to
clarify that and to determine structure–activity relationships,
the synthesis of stereoisomers and derivatives of the cripo-
wellins will be necessary.
Because of their unique structures and their high biolog-
ical activity, Bayer AG has patented both cripowellins and
some of their derivatives.[2] Their exceptional position was
also recognized by experts at the Irseer conference on natural
products in 1997, where cripowellin A (1) was voted the
second most interesting new natural product.[5] Despite all
this attention, a synthesis of the cripowellins, their aglycon, or
one of their stereoisomers has not been described so far. This
seems to be highly desirable in the light of the small quantities
obtainable from natural sources.[3]
Natural Product Synthesis
Asymmetric Synthesis of the 1-epi Aglycon of the
Cripowellins A and B**
Dieter Enders,* Achim Lenzen, and Gerhard Raabe
We now report on the first asymmetric synthesis of the
skeleton of the cripowellins A and B in the form of their 1-epi
aglycon 15. Our remarkably short synthesis started with the
asymmetric dihydroxylation[6] of the benzoylated allyl alcohol
3[7,8] to give the diol 4 (Scheme 1); the 72% yield of this
reaction is quite efficient when one considers the obvious
problem of regioselectivity. In addition, the asymmetric
induction was virtually complete (ꢀ 98% ee).[9] Acetonide
formation and saponification of the ester functionality yielded
the primary alcohol 5 in nearly quantitative yield. This alcohol
was oxidized to the corresponding acid 6 in two steps, and 6
was subsequently coupled with the amine 10 (available from
the reductive amination of bromopiperonal (9) with 3-butene-
1-amine) to give the amide 7.
In 1997 researchers at Bayer AG reported on two new
Amaryllidaceae alkaloids,[1] cripowellins A (1) and B (2),
which had been isolated from the bulbs and roots of Crinum
powellii, a popular ornamental plant in Europe.[2,3] The two
compounds differ only in their glycosidic parts. The assump-
tion that both of these sugar moieties are derived biogenically
from b-d-glucose accounts for the depicted absolute stereo-
chemistry. Their common aglycon comprises a [5.3.2]bicyclic
core, a structural motif unique among the Amaryllidaceae
alkaloids. One of both bridgehead atoms is a trisubstituted
amide N atom and therefore not a stereogenic center.
According to our synthetic route a ring-closing metathesis
(RCM)[10] of this amide to give the azacyclononene lactam
derivative 8 was proposed to follow. The structure of the
RCM precursor 7 had to be designed carefully taking into
account the “problem of medium sized rings”. It had to be
restricted conformationally in a way which would favour the
metathesis. Both the dioxolane ring[11] and the tertiary
amide[12] were assumed to operate synergistically thus favor-
ing the formation of the nine-membered ring.[13] Our assump-
tions were confirmed by the successful RCM. By employing
Grubbsꢀ second-generation catalyst and by performing the
reaction under high dilution (1 mm), we were able to obtain 8
in a very good yield of 77%.
In addition to their unusual structure, which contains five-,
six-, seven-, nine-, and ten-membered rings, both alkaloids
exhibit extraordinary biological properties. Their insecticidal
activity compares well to that of natural pyrethroids—not
After that the [5.3.2]bicyclic core was supposed to be built
up by means of a Heck reaction of the piperonyl moiety to the
double bond.[14] After extensive studies, we finally succeeded
in the selective synthesis of the two Heck products 11 and 12
(Scheme 2). Under neutral reaction conditions, we observed
exclusively the formation of product 11, which has a
disubstituted Z-configurated double bond. Under cationic
conditions, on the other hand, the reaction yielded selectively
the trisubstituted olefin 12.[15] The formation of this anti-Bredt
alkene is worth mentioning because the additional ring strain
in this bicyclic compound cannot be compensated by stability
from the conjugation with the aromatic moiety (both are
nearly perpendicular to each other). The double bond of this
anti-Bredt alkene is not configurationally stable under the
[*] Prof. Dr. D. Enders, Dipl.-Chem. A. Lenzen, Prof. Dr. G. Raabe
Institut fꢀr Organische Chemie, RWTH Aachen
Landoltweg 1, 52074 Aachen (Germany)
Fax: (+49)241-809-2127
E-mail: enders@rwth-aachen.de
[**] This work was supported by the Deutsche Forschungsgemeinschaft
(SFB 380) and the Fonds der Chemischen Industrie. We thank Prof.
Rosenkranz, Prof. Stetter, Dr. Lieb, and their co-workers, Bayer AG,
for valuable discussions at the beginning of this project.
3766
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200500556
Angew. Chem. Int. Ed. 2005, 44, 3766 –3769

Angewandte
Chemie
Scheme 2. Completion of the synthesis of the 1-epi aglycon 15 of the
cripowellins A (1) and B (2). a) Pd(OAc)2 (0.2 equiv), PPh3 (0.6 equiv),
Et3N (3.5 equiv), DMF, 1108C, 6 h, 59%; b) Pd(OAc)2 (0.15 equiv),
dppp (0.2 equiv), Ag2CO3 (3.0 equiv), toluene, 1248C, 4 h, 59%;
c) K2OsO4·2H2O (0.05 equiv), NMO (97%, 3.1 equiv), acetone/H2O
(10:7), 258C, 3 h, then Na2SO3 (2.3 equiv); d) CO2Cl2 (2.5 equiv),
DMSO (5.3 equiv), Et3N (10.0 equiv), CH2Cl2, ꢁ78!258C, 55% (two
steps); e) SmI2 (excess, ca. 7.2 equiv), tBuOH (3.0 equiv), THF, 258C,
12 h, 99%; f) Dowex-50, H2O, 258C, 4.25 h, 56%. DMF=dimethyl
formamide, dppp=1,3-bis(diphenylphosphanyl)propane, NMO=N-
methylmorpholine N-oxide, THF=tetrahydrofuran.
Scheme 1. Synthesis of the nine-membered-ring lactam intermediate 8.
a) AD-mix b, K2OsO4·2H2O (0.006 equiv), MeSO2NH2 (1.0 equiv),
tBuOH/H2O (1:1), 08C, 2.25 h, 72%; b) 2,2-DMP, PTSA (0.05 equiv),
258C, 1 h; c) K2CO3 (1.5 equiv), MeOH, 258C, 2 h, 97% (two steps);
d) CO2Cl2 (1.1 equiv), DMSO (2.3 equiv), Et3N (5.0 equiv), CH2Cl2,
ꢁ78!258C; e) NaClO2 (80%, 2.5 equiv), NaH2PO4·2H2O (2.0 equiv),
2-methyl-2-butene (18 equiv), acetone/H2O (1:1), 0!258C, 0.5 h;
f) FEP (1.2 equiv), amine 10 (1.1 equiv), EtiPr2N (3.2 equiv), CH2Cl2,
0!258C, 12 h, 66% (three steps); g) Grubbs’ 2nd generation catalyst
(0.1 equiv, addition in portions), CH2Cl2, reflux, 2.5 h, then DMSO
(5.0 equiv), 258C, 12 h, 77%; h) 3-butene-1-amine (1.2 equiv), MS 4 ꢁ,
CH2Cl2, 258C, 12 h; i) NaBH4 (1.0 equiv), MeOH, 258C, 2 h, 94% (two
steps). 2,2-DMP=2,2-dimethoxypropane, DMSO=dimethyl sulfoxide,
FEP=2-fluoro-1-ethyl pyridinium tetrafluoroborate, MS=molecular
sieves, PMB=p-methoxybenzoyl, PTSA=p-toluenesulfonic acid.
hydroxy ketones are normally not very good substrates for
SmI2-induced defunctionalizations.[20] Likewise, the stability
of the acetonide-protected 1,2-diol unit is worth mentioning.
ꢁ
In the case of carboxylates, the reductive cleavage of the a-C
O bond is commonly observed.[21] The cleavage of the
acetonide protecting group in the presence of Dowex-50
finally yielded the 1-epi aglycon 15 of the cripowellins A (1)
and B (2).
reaction conditions employed since the E/Z ratio was
observed to be time dependent.[16] To our surprise, both
Heck reactions were completely diastereoselective with
regard to the orientation of the ethylene bridge; we were
able to isolate only compounds 11 and 12, in which this bridge
is on the upper face of the molecule.[17] To the best of our
knowledge, this is the first example of a (highly diastereose-
lective) intramolecular Heck reaction of a highly functional-
ized (aza)cyclonene derivative.[18]
Olefin 11 proved to be extremely difficult to functionalize.
We therefore decided to continue with the E/Z mixture of
olefins 12, which seemed to be more reasonable to us anyway
because a differentiation between the olefinic C atoms was
assumed to be easier (singly vs. doubly substituted). Both
olefins 12 were first transformed into the a-hydroxy ketone 13
in a two-step sequence consisting of dihydroxylation and
Swern oxidation. Deoxygenation of 13 with SmI2 in the
presence of tBuOH proceeded smoothly to give 14.[19] The
high yield of 99% in this reaction was astonishing because a-
At this point we were very much interested in the spatial
structure of this compound—especially with regard to the
known structure of the bisacetate of cripowellin A and the
biological activity of the cripowellins.[2,3] In the search for
derivatives of the 1-epi aglycon 15 suitable for an X-ray
structure analysis, we were finally rewarded by its bisacetate,
16 (Scheme 3).[22] A comparison of its crystal structure with
the one obtained from the bisacetate of cripowellin A clearly
shows the same spatial orientation of the keto and the lactam
carbonyl groups: both are syn to each other and in spatial
proximity. This substructure had been identified as the
probable pharmacophor by researchers at Bayer AG after
extensive investigations on structure–activity relationships.[4]
Therefore it seems reasonable to expect the 1-epi aglycon 15
to be biologically active, too. Yet, this still needs to be proven
by biological tests.
In summary, we report the first access to this unique
[5.3.2]bicyclic skeleton of the cripowellins A (1) and B (2).
Key steps in our synthesis are a highly enantioselective
Angew. Chem. Int. Ed. 2005, 44, 3766 –3769
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3767

Communications
[7] Prepared by esterification of the known alcohol with para-
methoxybenzoic acid chloride (87% yield): J. A. Rao, M. P.
Cava, J. Org. Chem. 1989, 54, 2751 – 2753.
[8] The asymmetric dihydroxylation of allylic alcohols normally
proceeds with only moderate enantioselectivity, in contrast to
the reaction of the corresponding para-methoxybenzoic acid
esters: E. J. Corey, A. Guzman-Perez, M. C. Noe, J. Am. Chem.
Soc. 1995, 117, 10805 – 10816.
[9] The enantiomeric excess was determined by HPLC on a chiral
stationary phase. For this, ent-4 was prepared analogously to 4
but using AD-mix a instead.
[10] Representative reviews on ring-closing metathesis: a) A. Dei-
ters, S. F. Martin, Chem. Rev. 2004, 104, 2199 – 2238; b) R. R.
Schrock, A. H. Hoveyda, Angew. Chem. 2003, 115, 4740 – 4782;
Angew. Chem. Int. Ed. 2003, 42, 4592 – 4633; c) A. Fꢂrstner,
Angew. Chem. 2000, 112, 3140 – 3172; Angew. Chem. Int. Ed.
2000, 39, 3012 – 3043; d) M. E. Maier, Angew. Chem. 2000, 112,
2153 – 2157; Angew. Chem. Int. Ed. 2000, 39, 2073 – 2077; e) S.
Blechert, Pure Appl. Chem. 1999, 71, 1393 – 1399; f) A. Fꢂrstner,
Top. Organomet. Chem. 1998, 1, 37 – 72; g) R. H. Grubbs, S.
Chang, Tetrahedron 1998, 54, 4413 – 4450; h) M. Schuster, S.
Blechert, Angew. Chem. 1997, 109, 2124 – 2145; Angew. Chem.
Int. Ed. Engl. 1997, 36, 2036 – 2055; i) H.-G. Schmalz, Angew.
Chem. 1995, 107, 1981 – 1984; Angew. Chem. Int. Ed. Engl. 1995,
34, 1833 – 1836; j) F.-X. Felpin, J. Lebreton, Eur. J. Org. Chem.
2003, 3693 – 3712.
[11] For the synthesis of nine-membered rings with similar conforma-
tional constraints similar to those of the dioxolane ring by
metathesis, see: a) P. W. R. Harris, M. A. Brimble, P. D. Gluck-
man, Org. Lett. 2003, 5, 1847 – 1850; b) J. S. Clark, F. Marlin, B.
Nay, C. Wilson, Org. Lett. 2003, 5, 89 – 92; c) K. P. Kaliappan, N.
Kumar, Tetrahedron Lett. 2003, 44, 379 – 381; d) M. Hirama, T.
Oishi, H. Uehara, M. Inoue, M. Maruyama, H. Oguri, M. Satake,
Science 2001, 294, 1904 – 1907; e) J. S. Clark, O. Hamelin, Angew.
Chem. 2000, 112, 380 – 382; Angew. Chem. Int. Ed. 2000, 39, 372 –
374; f) S. J. Bamford, K. Goubitz, H. L. van Lingen, T. Luker, H.
Schenk, H. Hiemstra, J. Chem. Soc. Perkin Trans. 1 2000, 345 –
351; g) T. Oishi, Y. Nagumo, M. Hirama, Chem. Commun. 1998,
1041 – 1042; h) M. Delgado, J. D. Martꢃn, Tetrahedron Lett. 1997,
38, 6299 – 6300.
Scheme 3. Synthesis and crystal structure of the bisacetylated 1-epi
aglycon 16. a) Sc(OTf)3 (0.3 equiv), Ac2O/CH3CN (1:1), 258C, 2 h,
65%. Tf=trifluoromethanesulfonyl.[22]
Sharpless dihydroxylation, a properly designed ring-closing
metathesis, and a highly diastereoselective intramolecular
Heck reaction. Considering the complexity of the target
molecule, this synthesis is very short (13 steps in the longest
linear sequence, 15 steps altogether; 5.6% overall yield). In
addition, the diastereo- and enantioselectivity are virtually
complete (ꢀ 98% de, ꢀ 98% ee). The crystal structure of the
bisacetate of the 1-epi aglycon 16 reveals the same spatial
orientation of the ketone and the lactam carbonyl groups as in
the cripowellins. Therefore one may be curious whether the 1-
epi derivatives exhibit the same biological activity as the
cripowellins and their aglycon. The 1-epi aglycon 15 might at
least help to further clarify their hitherto unknown mode of
action.[23]
Received: February 12, 2005
Published online: May 13, 2005
[12] For secondary and tertiary amides in ring-closing metatheses,
see: a) A. J. Brouwer, R. M. J. Liskamp, J. Org. Chem. 2004, 69,
3662 – 3668; b) L. Banfi, A. Basso, G. Guanti, R. Riva, Tetrahe-
dron Lett. 2003, 44, 7655 – 7658.
Keywords: alkaloids · asymmetric syntheses · Heck reactions ·
.
insecticides · ring-closing metatheses
[13] There are only very few examples for the synthesis of nine-
membered rings by methathesis without conformational con-
straints. They are limited almost exclusively to nine-membered
cyclic ethers for which the gauche effect can be exploited:
a) M. T. Crimmins, M. T. Powell, J. Am. Chem. Soc. 2003, 125,
7592 – 7595; b) M. T. Crimmins, K. A. Emmitte, A. L. Choy,
Tetrahedron 2002, 58, 1817 – 1834; c) Y. Baba, G. Saha, S. Nakao,
C. Iwata, T. Tanaka, T. Ibuka, H. Ohishi, Y. Takemoto, J. Org.
Chem. 2001, 66, 81 – 88; d) M. T. Crimmins, A. L. Choy, J. Org.
Chem. 1997, 62, 7548 – 7549.
[14] Representative reviews on the Heck reaction: a) S. Brꢄse, A.
de Meijere in Metal-Catalyzed Cross-Coupling Reactions, 2nd
ed. (Eds.: A. de Meijere, F. Diederich), Wiley-VCH, Weinheim,
2004, pp. 217 – 315; b) J. T. Link, L. E. Overman in Metal-
Catalyzed Cross-Coupling Reactions (Eds.: F. Diederich, P. J.
Stang), Wiley-VCH, Weinheim, 1998, pp. 231 – 269; c) A. B.
Dounay, L. E. Overman, Chem. Rev. 2003, 103, 2945 – 2963;
d) J. T. Link in Organic Reactions, Vol. 60 (Eds.: L. E. Overman),
Wiley, New York, 2002, pp. 157 – 534; e) A. de Meijere, F. E.
Meyer, Angew. Chem. 1994, 106, 2473 – 2506; Angew. Chem. Int.
Ed. Engl. 1994, 33, 2379 – 2411.
[1] Representative reviews about the Amaryllidaceae alkaloids:
a) O. Hoshino in The Alkaloids, Vol. 51 (Ed.: G. A. Cordell),
Academic Press, New York, 1998, pp. 324 – 424; b) S. F. Martin in
The Alkaloids, Vol. 30 (Ed.: A. Brossi), Academic Press, New
York, 1987, pp. 251 – 376; c) C. Fuganti in The Alkaloids, Vol. 15
(Ed.: R. H. F. Manske), Academic Press, London, 1975, pp. 83 –
164.
[2] M. Gehling, A. Gꢁhrt, D. Gondol, J. Lenz, O. Lockhoff, H.-F.
Moeschler, R. Velten, D. Wendisch, W. Andersch, C. Erdelen, A.
Harder, N. Mencke, A. Turberg, U. Wachendorff-Neumann
(Bayer AG), DE 196 10 279A1, 1997 [Chem. Abstr. 1997, 127,
278406].
[3] R. Velten, C. Erdelen, M. Gehling, A. Gꢁhrt, D. Gondol, J. Lenz,
O. Lockhoff, U. Wachendorff, D. Wendisch, Tetrahedron Lett.
1998, 39, 1737 – 1740.
[4] Researchers at Bayer AG have determined that cripowellins A
and B are neither acetylcholin esterase inhibitors nor PP1
inhibitors (personal communication).
[5] T. Lindel, Nachr. Chem. Tech. Lab. 1997, 45, 775 – 779.
[6] Review: H. C. Kolb, M. S. VanNieuwenhze, K. B. Sharpless,
Chem. Rev. 1994, 94, 2483 – 2547.
[15] Apart from 12, minor amounts of 11 could also be isolated when
the Heck reaction was performed under cationic conditions (in
3768
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 3766 –3769

Angewandte
Chemie
the crude product the ratio 12/11 was determined to be 8.2:1 by
means of gas chromatography). However, 11 could be easily
separated from 12 by column chromatography on silica gel
because of its higher polarity.
[16] After 4 h (complete conversion): E/Z = 1:1.7; after 24 h: E/Z =
1:1 (determined by gas chromatography).
[17] The relative configurations were determined by NOE measure-
ments.
[18] For the synthesis of bicyclic systems by a sequence of RCM and
Heck reactions, see: a) M. Lautens, V. Zunic, Can. J. Chem. 2004,
82, 399 – 407; b) R. Grigg, M. York, Tetrahedron Lett. 2000, 41,
7255 – 7258; c) R. Grigg, V. Sridharan, M. York, Tetrahedron
Lett. 1998, 39, 4139 – 4142.
[19] J. D. White, T. C. Somers, J. Am. Chem. Soc. 1994, 116, 9912 –
9920.
[20] G. A. Molander, G. Hahn, J. Org. Chem. 1986, 51, 1135 – 1138.
[21] G. A. Molander in Organic Reactions, Vol. 46 (Ed.: L. A.
Paquette), Wiley, New York, 1994, pp. 211 – 367.
[22] CCDC 263133 (16) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from the Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[23] We will report separately on our alternative synthetic
approaches to the cripowellins in detail: K. Catlin, RWTH
Aachen, unpublished results; C. Janeck, Ph.D. Thesis, RWTH
Aachen, 2000; M. Backes, Ph.D. Thesis, RWTH Aachen, 2004.
Angew. Chem. Int. Ed. 2005, 44, 3766 –3769
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3769