Toward a total synthesis of the stemofoline alkaloids: Advancement of a 1,3-dipolar cycloaddition strategy

Article abstract of DOI:10.1016/j.tetlet.2011.05.121

Novel, intramolecular 1,3-dipolar cycloadditions of azomethine ylides have been applied to the synthesis of functionalized core structures of the stemofoline alkaloids. In an effort to maximize the efficiency of this key transformation in the context of an eventual total synthesis of these complex natural products, a number of strategic modifications to the cycloaddition substrate were investigated. The collective efforts have provided useful insights into the operative, regiochemical control elements for 1,3-dipolar cycloadditions leading to stemofoline alkaloids. A potential intermediate in the synthesis of these alkaloids was prepared.

Full text of DOI:10.1016/j.tetlet.2011.05.121

Tetrahedron Letters 52 (2011) 4076–4079
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Toward a total synthesis of the stemofoline alkaloids: advancement
of a 1,3-dipolar cycloaddition strategy
⇑
Charles S. Shanahan, Nathan O. Fuller, Bjoern Ludolph, Stephen F. Martin
Department of Chemistry and Biochemistry, The University of Texas, Austin, TX 78712, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel, intramolecular 1,3-dipolar cycloadditions of azomethine ylides have been applied to the synthesis
of functionalized core structures of the stemofoline alkaloids. In an effort to maximize the efficiency of
this key transformation in the context of an eventual total synthesis of these complex natural products,
a number of strategic modifications to the cycloaddition substrate were investigated. The collective
efforts have provided useful insights into the operative, regiochemical control elements for 1,3-dipolar
cycloadditions leading to stemofoline alkaloids. A potential intermediate in the synthesis of these alka-
loids was prepared.
Received 26 April 2011
Revised 21 May 2011
Accepted 24 May 2011
Available online 2 June 2011
Keywords:
Azomethine ylide
Dipolar cycloaddition
Stemofoline alkaloids
Natural product synthesis
Ó 2011 Elsevier Ltd. All rights reserved.
The Stemonacea family contains approximately 30 different spe-
cies of flowering plants native to various regions of Southeast Asia,
from which a plethora of biologically active natural products have
been isolated.1,2 The ground-up leaves and tuberous roots of these
plants have been used for centuries in traditional Asian medicine
to prepare herbal teas for use in treating chronic cough symptoms
associated with respiratory diseases, such as bronchitis and tuber-
culosis. Additionally, these extracts can be employed as pesticidal
remedies for both human and agricultural infestations. Extensive
investigations into the active principles of these plants have
revealed a wealth of alkaloids that pose significant challenges to
chemical synthesis, the most notable of which are the stemofoline
alkaloids (Fig. 1). This complex group of alkaloids is characterized
by a caged hexacyclic architecture varying only in the oxidation
state of the C3 side chain and the geometry of the C11/C12
carbon–carbon double bond. These alkaloids exhibit powerful
activity as insect acetylcholine receptor antagonists,3,4 as well as
activity against various human carcinoma cell lines and in vivo
anti-oxytocin activity,5 with didehydrostemofoline (3), which
was the first of the stemofoline alkaloids to be isolated,6 being
the most potent. Despite numerous efforts toward the synthesis
of these alkaloids,7 only two total syntheses have been achieved.
Kende reported the synthesis of ( )-2 in 1999,8 and compounds
( )-3 and ( )-4 were synthesized by Overman in 2003.9
Figure 1.
of this family. The initial focus of our efforts centered on the
construction of the functionalized tricyclic core 6 of these alkaloids
using an intramolecular 1,3-dipolar cycloaddition reaction of an
azomethine ylide of substrates of the general type 7 (Eq. 1). We
recently reported some of our initial findings wherein we disclosed
the successful construction of the tricyclic core 10 via a novel
Because of their structural complexity and biological activity,
we became interested in the total synthesis of selected members
⇑
Corresponding author. Tel.: +1 512 471 3915; fax: +1 512 471 4180.
E-mail address: sfmartin@mail.utexas.edu (S.F. Martin).
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.05.121

C. S. Shanahan et al. / Tetrahedron Letters 52 (2011) 4076–4079
4077
35% overall yield.10 When 16 was treated with CF3CO2H and dim-
ethoxymethane, oxazolidine 17 was formed in 75% yield. It should
be noted that reaction times longer than 7 h resulted in the forma-
tion of appreciable amounts of 19 but none of the desired tricyclic
compound 11. Thermolysis of oxazolidine 17 presumably gener-
ated the azomethine ylide 18 that underwent dipolar cycloaddition
to deliver a mixture (1:3) of cycloadducts 11 and 19 in 96% yield.13
Although the ratio of regioisomers was poor relative to the desired
cycloadduct 11, this cycloaddition did provide the nor-cyano
stemofoline core 11 in only five steps, considerably shorter than
our earlier 10-step route to 10.10
We reasoned that electronic factors might be partially responsi-
ble for the observed difference in regioselectivities in the cycload-
Scheme 1. Reagents and conditions: (i) (COCl)2, DMSO, CH2Cl2, À78 °C; NEt3; 16 h;
69%; regioisomeric ratio of 10:12 = 5:1.
ditions of 9. Accordingly, we queried whether
a dipolar
cycloaddition involving an electron deficient dipolarophile bearing
a terminal electron withdrawing group might be more regioselec-
tive because of a better match with what we presumed was the
inherent polarization of azomethine ylide. In order to explore this
hypothesis, we converted 17 to 20 by a cross metathesis reaction
with phenyl vinyl sulfone in the presence of Grubbs 2nd generation
catalyst (Scheme 3). Unfortunately, thermolysis of 20 gave a mix-
ture (1:2) of cycloadducts 22 and 23—not significantly different
from the cyclization of 18.13
process in which the azomethine ylide 9 was generated from 8
under the conditions of a Swern oxidation (Scheme 1).10 Subse-
quent intramolecular 1,3-dipolar cycloaddition via the regioiso-
meric transition states 9a and 9b gave a mixture (ca 5:1) of 10
and 12. This reaction granted access to the tricyclic core of the
stemofoline alkaloids and provided essential proof of principle
for our approach. However, we were unable to decyanate 10 to give
11 despite repeated efforts under a variety of conditions, so 10 is
not a viable intermediate in a total synthesis of these alkaloids.
Accordingly, we explored alternate avenues that would afford
intermediates lacking superfluous substitution at C(5) (see
Fig. 1). We now report some of our findings.
We were cognizant of the discoveries of Joucla, who had shown
that azomethine ylides could be generated via thermolysis of 2,2-
unsubstituted oxazolidines.11 In order to apply this method to
the problem at hand, we turned our attention to the synthesis of
the oxazolidine 17. Toward this end, a-amidosulfone 14 was pre-
pared by condensing 13 with tert-butylcarbamate and p-toluene-
sulfinic acid sodium salt (Scheme 2). Compound 14 served as a
masked acylimine that underwent a Mannich-type reaction with
the enolate of methyl diazomethylacetoacetate to give diazo-b-
ketoester 15 in 65% yield. Although the titanium enolate of methyl
diazomethylacetoacetate was known to participate in Mannich-
type reactions with sulfonimines,12 the reaction of the correspond-
Scheme 3. Reagents and conditions: (i) phenyl vinyl sulfone, Grubbs II (5 mol%),
CH2Cl2, reflux; 50%. (ii) 160 °C, PhMe; 97%.
ing lithium enolate anion with an
a-amidosulfone was unknown.
Upon exposure to catalytic amounts of rhodium acetate, 15 cy-
clized to give pyrrolidinone 16 in 96% yield. This route to 16 re-
quired only three steps and proceeded in 46% overall yield,
whereas our previous synthesis required six steps, giving 16 in
The divergent regiochemical effects of having an electron-with-
drawing group on the termini of the dipole and the olefin of the
putative intermediates 9 and 21 are not readily interpretable.
Although we considered the possibility of varying the nature and
position of acceptors on the dipolarophile, we elected instead to
turn our attention to examining steric effects that might be modu-
lated by introducing substituents in the chain linking the dipole
and dipolarophile. Such substituents would ideally correspond to
those present in late-stage intermediates leading to the natural
products themselves.
After considering various options, we decided to introduce a
side chain at the eventual C(9) position (see Fig. 1) of an interme-
diate unsaturated azomethine ylide in anticipation that such a sub-
stituent would direct the cycloaddition to favor the desired
regioisomer. This analysis led us to target the substituted aldehyde
28 as a starting material that would be processed along a path sim-
ilar to that outlined in Scheme 2. The synthesis of 28 commenced
with coupling the commercially available crotonic acid derivative
24 with the
D-phenylglycine-derived Evans auxiliary 25 in the
presence of Piv-Cl and LiClto give imide 26 (Scheme 4). We then
employed a copper-mediated conjugate addition that was devel-
oped by Bergdahl to install a methyl group at the b-stereocenter;14
Scheme 2. Reagents and conditions: (i) H2NCO2tBu, p-TolSO2Na, HCO2H, THF:H2O
(4:1); 75%. (ii) MeCOC(N2)CO2Me, LDA, THF, -78 °C; 65%. (iii) Rh2(OAc)4 (1 mol %),
CH2Cl2; 96%. (iv) (MeO)2CH2 (10 equiv), 10% CF3CO2H in CH2Cl2, rt, 7 h; 75%. (v)
160 °C, PhMe; 96%.
a-alkylation of the imide enolate thus formed in situ with allyl io-
dide provided 27 in high (dr >15:1) diastereoselectivity. The chiral
auxiliary was removed from 27 by the action of basic hydrogen

4078
C. S. Shanahan et al. / Tetrahedron Letters 52 (2011) 4076–4079
Scheme 4. Reagents and conditions: (i) Piv-Cl, NEt3; then LiCl; 87%. (ii)
(CuI)3(Me2S)4, MeLi, TMS-I, Et2O, -78 °C; 91% (dr = 7:1). (iii) LiHMDS, allyl iodide;
70% (dr >15:1). (iv) H2O2, LiOH, H2O:MeOH (1:1); 83%. (v) LiAlH4, THF; 99%. (vi)
PCC/SiO2, CH2Cl2; 99%.
Scheme 6.
in 15 synthetic operations and in 7% overall yield. These experi-
ments also support our hypothesis that stereochemical and steric
factors can be successfully exploited to enhance the regioselectiv-
ity of dipolar cycloadditions leading to the formation of com-
pounds having the tricyclic core of the stemofoline alkaloids. We
are currently exploring related tactics to access such compounds
with improved selectivity, and the results of these efforts will be
reported in due course.
peroxide,15 the acid was reduced with LiAlH4, and the resultant
alcohol was oxidized to the corresponding aldehyde 28.
With enantiomerically-pure aldehyde 28 in hand, we employed
a general protocol developed by Davis for preparing 5-substituted
pyrrolidinones.16 Namely, 28 was condensed with the chiral sul-
fonamide 29 in the presence of Ti(OEt)4 to give sulfinimine 30
(Scheme 5). Compound 30 was then subjected to a tandem Man-
nich/cross-Claisen reaction using an excess of the enolate of
methyl acetate to give b-ketoester 31 in excellent yield and diaste-
reoselectivity (dr = 8:1).17 In the first step of this sequence, the res-
ident chirality of the sulfinimine controls the stereochemistry of
the Mannich addition. The sulfonamide of b-ketoester 31 was then
exchanged for a Boc-protecting group due to known incompatibil-
ity of the sulfonamide moiety to the impending NH-insertion reac-
tion conditions.18 Accordingly, treatment of b-ketoester 32 with
CF3CO2H in methanol, followed by sequential reaction with Boc2O
and p-acetamidobenzenesulfonyl azide provided the diazo-b-keto-
ester 32 in 79% yield (3 steps). Cyclization of 32 via a rhodium ace-
tate catalyzed NH-insertion, followed by reaction of the
intermediate protected pyrrolidine derivative with dimethoxyme-
thane and CF3CO2H gave oxazolidine 33 as an inconsequential mix-
ture (ca 1:1) of diastereomers.
With the key oxazolidine 33 in hand, the stage was set for
examining the pivotal dipolar cycloaddition. Although thermolysis
of 33 did provide an improved ratio (1:1) of the regioisomeric cyc-
loadducts 35 and 36 in excellent yield (Scheme 6),13 the increased
selectivity still falls short of what is synthetically optimal. It should
be noted that thermolysis of the isolated cycloadducts 22, 23, 35,
and 36 under the original conditions did not result in any isomer-
ization, suggesting that the observed regioisomeric ratios are
kinetically controlled.
Acknowledgments
We thank the National Institutes of Health (GM 25439 and
GM31077), Pfizer, Inc., Merck Research Laboratories, and the
Robert A. Welch Foundation (F-0652) for their generous support
of this research. B.L. gratefully acknowledges a Feodor-Lynen post-
doctoral fellowship from the Alexander von Humboldt Foundation.
We also thank Dr. Christian Harcken for helpful discussions and
conducting some preliminary experiments and Dr. Vince Lynch
for performing the X-ray analyses.
References and notes
1. Pilli, R. A.; Ferreira de Oliveira, M. C. Nat. Prod. Rep. 2000, 17, 117–127.
2. Greger, H. Planta Med. 2006, 72, 99–113.
3. Brem, B.; Seger, C.; Pacher, T.; Hofer, O.; Vajrodaya, S.; Greger, H. J. Agric. Food
Chem. 2002, 50, 6383–6388.
4. Kaltenegger, E.; Brem, B.; Mereiter, K.; Kalchhauser, H.; Kahlig, H.; Hofer, O.;
Vajrodaya, S.; Greger, H. Phytochemistry 2003, 63, 803–816.
5. (a) Jiwajinda, S.; Hirai, N.; Watanabe, K.; Santisopasri, V.; Chuengsamarnyart,
N.; Koshimizu, K.; Ohigashi, H. Phytochemistry 2001, 56, 693–695; (b) Sekine,
T.; Ikegami, F.; Fukasawa, N.; Kashiwagi, Y.; Aizawa, T.; Fujii, Y.; Ruangrungsi,
N.; Murakoshi, I. J. Chem. Soc., Perkin Trans. 1 1995, 391–393; Tip-Pyang, S.;
Tangpraprutgul, P.; Wiboonpun, N.; Veerachato, G.; Phuwapraisirisan, P.; Sup-
Udompol, B. ACGC Chem. Res. Commun. 2000, 12, 31–35.
6. Irie, H.; Masaki, N.; Ohno, K.; Osaki, K.; Taga, T.; Uyeo, S. J. Chem. Soc. D; Chem.
Commun. 1970, 1066.
Despite the regiochemical outcome of this dipolar cycloaddi-
tion, it is noteworthy that enantiomerically-pure 35 is accessible
7. (a) Epperson, M. T.; Gin, D. Y. Angew. Chem., Int. Ed. 2002, 41, 1778–1780; (b) Ye,
Y.; Velten, R. F. Tetrahedron Lett. 2003, 44, 7171–7173; (c) Baylis, A. M.; Davies,
M. P. H.; Thomas, E. J. Org. Biomol. Chem. 2007, 5, 3139–3155; (d) Carra, R. J.;
Epperson, M. T.; Gin, D. Y. Tetrahedron 2008, 64, 3629–3641; (e) Thomas, E. J.;
Vickers, C. F. Tetrahedron: Asymmetry 2009, 20, 970–979.
8. Kende, A. S.; Smalley, T. L., Jr.; Huang, H. J. Am. Chem. Soc. 1999, 121, 7431–
7432.
9. Brueggemann, M.; McDonald, A. I.; Overman, L. E.; Rosen, M. D.; Schwink, L.;
Scott, J. P. J. Am. Chem. Soc. 2003, 125, 15284–15285.
10. Dietz, J.; Martin, S. F. Tetrahedron Lett. 2011, 52, 2048–2050.
11. (a) Joucla, M.; Mortier, J. Tetrahedron Lett. 1987, 28, 2973–2974; (b) Joucla, M.;
Mortier, J.; Bureau, R. Tetrahedron Lett. 1987, 28, 2975–2976.
12. Dong, C.; Deng, G.; Wang, J. J. Org. Chem. 2006, 71, 5560–5564.
13. The ratio of 11 and 19 was based on a preliminary analysis of the 1H NMR
spectrum of the crude reaction mixture and confirmed by the isolated yields of
both regioisomers. The ratio based on the 1H NMR spectrum relies upon the
relative integrations of the diagnostic endo-C(1)-Ha proton of 11 and the endo-
C(1)-Ha proton of 19. These protons appear as well resolved, sharp doublets
between 1.5 and 2.2 ppm in all cycloadducts, with the signal from 11, 22 and
35 always being upfield from the regioisomers 19, 23, and 36. The C(1)-Ha
protons of 11, 22 and 35 have geminal coupling constants of 18.0 Hz
( 0.1 Hz),whereas the coupling constants for the C(1)-Ha protons of 19, 23,
and 36 are 15.6 Hz ( 0.1 Hz).
Scheme 5. Reagents and conditions: (i) Ti(OEt)4, CH2Cl2, rt; 74%. (ii) NaHMDS,
MeOAc, THF:Et2O, À78 °C to 0 °C; 95%. (dr = 8:1).(iii) TFA, MeOH; (iv) Boc2O, NEt3,
CH2Cl2; 81% (2 steps). (v) p-ABSA, NEt3, MeCN; 97%. (vi) Rh2(OAc)4 (1 mol%), CH2Cl2;
87%. (vii) (MeO)2CH2, CF3CO2H, CH2Cl2, rt, 7 h; 67% (dr = 1:1).

C. S. Shanahan et al. / Tetrahedron Letters 52 (2011) 4076–4079
4079
15. Evans, D. A.; Sjogren, E. B.; Bartroll, J.; Dow, R. L. Tetrahedron Lett. 1986, 27,
4957–4960.
16. (a) Davis, F. A.; Yang, B.; Deng, J. J. Org. Chem. 2003, 68, 5147–5152; (b) Davis, F.
A.; Fang, T.; Goswami, R. Org. Lett. 2002, 4, 1599–1602; For a review, see: (c)
Zhou, P.; Chen, B.-C.; Davis, F. A. Tetrahedron 2004, 60, 8003–8030.
17. It is important to prepare the enolate by adding a solution (<0.5 M) of methyl
acetate to a solution of NaHMDS (<0.5 M). A 1–2 molar equivalent excess
NaHMDS should be used to suppress the deleterious homo-Claisen side
reaction that occurs during preparation of the enolate. Warming the reaction to
0 °C is also required for high conversion to the desired b-ketoester.
18. Deng, G.; Jiang, N.; Ma, K.; Wang, J. Synlett 2002, 11, 1913–1915.
14. Dambacher, J.; Anness, R.; Pollock, P.; Bergdahl, M. Tetrahedron 2004, 60, 2097–
2110.