Synthesis of (-)-neothiobinupharidine

Article abstract of DOI:10.1021/ja310778t

An eight step, asymmetric synthesis of a dimeric thiaspirane nuphar alkaloid from 3-methyl-2-cyclo-pentenone is reported. The brevity of the route relies on a useful procedure for tandem reductive allylation of cyclopentenones, as well as the minimization of redox manipulations and other functional group interconversions. The distribution of products that arise from spontaneous dimerization points to a more complex biosynthesis.

Full text of DOI:10.1021/ja310778t

Communication
pubs.acs.org/JACS
Synthesis of (−)-Neothiobinupharidine
Daniel J. Jansen and Ryan A. Shenvi*
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, United States
S
* Supporting Information
ABSTRACT: An eight step, asymmetric synthesis of a
dimeric thiaspirane nuphar alkaloid from 3-methyl-2-cyclo-
pentenone is reported. The brevity of the route relies on a
useful procedure for tandem reductive allylation of
cyclopentenones, as well as the minimization of redox
manipulations and other functional group interconver-
sions. The distribution of products that arise from
spontaneous dimerization points to a more complex
biosynthesis.
he first of the nuphar dimers (1−3, Figure 1a) was
T
isolated from the fresh water plant Nuphar lutea by
Achmatowicz,1 and neothiobinupharidine 3d was definitively
assigned in 1965 by Birnbaum via X-ray crystallographic
analysis.2 Extensive studies by LaLonde over the course of a
decade established the stereochemistry of a family of 11
topologically complex alkaloids.3 Initial bioassays were
disappointing, as these molecules demonstrated only weak
activity (80−200 μM) against pathogenic bacteria and fungi.3
However, more recent studies on the effects of nuphar dimers
on cancer cells has reinvigorated interest in these compounds.
Yoshikawa found that in vitro invasion of murine (B16)
melanoma cells across collagen fibers was inhibited by 1c, 1a,
and 2c at IC50 levels of 29, 87, and 360 nM, respectively.
Remarkably, lung tumor formation in mice was inhibited by 1c
by >90% versus control at 10 days post injection of B16
melanoma cells.4 These same compounds induced apoptosis of
human leukemia U937 cells within only 1 h.5 More recent
studies by Golan-Goldhirsh and Gopas have found that a crude
extract of N. lutea downregulates NFκB activity, and potentiates
the cytotoxicity of cisplatin toward Hodgkin’s lymphoma-
derived cells (L428) at concentrations lower than 0.1 μg/mL.6
Currently, the biological target of these alkaloids is unknown.
An insightful proposal for the biogenesis of the thiaspirane
alkaloids’ gross structure (Figure 1b) was put forth by
LaLonde7 but has never been investigated in vivo or
reproduced in vitro. Notably, this proposed biosynthesis does
not account for distribution of these stereoisomeric compounds
in nature. For instance, monomeric enamines similar to 6 have
been shown to react preferentially with electrophiles on their
convex face (Figure 1c),8−10 but this selectivity for the C−C
(Mannich-type) and C−S (sulfenylation) bond forming
reactions of 6 and 7 would lead to a stereochemical outcome
which has never been observed (4a−d) in Nuphar sp.3,11
Here we report the shortest synthesis to date of the
monomeric nuphar quinolizidine series in enantiomerically
pure form,12 and the first total synthesis of a thiaspirane nuphar
dimer. We also show that the preferred product of sulfurizing
Figure 1. The nuphar thiaspirane dimers. (a) a: X,Y = OH, b: X =
OH, Y = H, c: X = H, Y = OH, d. X,Y = H; (b) LaLonde’s biosynthetic
proposal; (c) stereochemical preferences of monomeric enamines; (d)
retrosynthetic analysis.
dimerization of 5 in bulk solvent is not 4, but the
neothiobinupharidine series 3. This outcome derives from the
opposite stereoselectivity of bond formation as the major series
Received: November 1, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja310778t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1. Synthesis of (−)-Neothiobinupharidine in 8 Steps
(1) found in Nuphar sp. and may suggest a more complex
biosynthesis mechanism than spontaneous dimerization.13
We planned to access iminium 5 via N-epi-dehydronuphar-
idine 8 in the anticipation that a Polonovsky reaction would
only eliminate the equatorial proton (Heq) syn- to the N-oxide
(Figure 1d). The stereotetrad of 8 could arise through substrate
controlled reactions from lactam 9, and so rapid access to this
stereodiad with absolute stereochemical control was explored.
Honda synthesized 9 in 9 steps from a chiral pool starting
material,12b but we thought that a Beckman rearrangement,14
followed retrosynthetically by asymmetric vicinal difunctional-
ization of 2-cyclopentenone, may offer a more efficient entry to
the synthesis. Unfortunately, there are no known methods that
are capable of achieving a conjugate methyl addition and
tandem allylation of cyclopentenones with good absolute or
relative stereocontrol. Instead, the intermediate cyclopenteno-
lates undergo proton exchange with the product ketones and
yield diastereo- and regioisomers in addition to doubly allylated
products.15 For instance, methylzinc conjugate additions to
cyclopentenone proceed with high levels of enantioselectivity,
but allylation of these intermediates is problematic (Table 1,
entries 1,16 2,17 and 318), in line with recent observations by
Cook.19
Therefore, we designed our own procedure based on
Buchwald’s conjugate reduction reaction (entry 4),20 in
combination with a modified Tsuji-Trost allylation (entries 5
and 6). These new conditions lead to ketone 10a in one pot
with high diastereoselectivity (d.r. > 10:1), high enantiose-
lectivity (e.r. = 97.5:2.5) and only minimal byproducts,
although use of precise amounts of methyl lithium is crucial,
and the isolated yield is still modest due to the volatility of 10a.
This new procedure for cyclopentenone vicinal difunctionaliza-
tion forms the basis of a short synthesis of the nuphar alkaloids.
Ketone 10a could be converted into the corresponding
oxime (42% from 11b) and engaged in a Beckman rearrange-
ment to yield lactam 9 in 69% yield. To rapidly advance this
monocycle to quinolizidine 14, we employed Vanderwal’s
allylsilane-RCM methodology21 to access the required exocyclic
methylene. Alkylation of 9 with iodide 12 proceeded cleanly, as
did ring-closing metathesis with Grubb’s second generation
catalyst.12b In situ proto-desilylation with trifluoroacetic acid
delivers 13 in good yield. This short 5-step sequence allows 13
to be procured from 3-methyl-2-cyclopentenone (0.29
$/mmol) on multigram scale (3.4 g prepared). This
Table 1. Optimization of Vicinal Difunctionalization
a
b
Determined by GC. Residual 3-methyl-cyclopentanone from
unreacted cyclopentenolate.
quinolizidone could be elaborated to quinolizidine 14 following
a variation of Fowler’s procedure for introduction of the 3-furyl
moiety.12e We found sodium triacetoxyborohydride to be
superior to previously reported reducing agents in this
reaction.22
Oxidation of 14 with m-CPBA proceeded uneventfully to
provide the trans-quinolizidine N-oxide 8 (stereochemistry
proven by NOE enhancement experiments between Hax, Heq,
and Hb; see Supporting Information). Polonovsky elimination
of naturally occurring nuphar quinolizidines like nupharidine (a
cis-quinolizidine N-oxide) proceeds over 5 days using acetic
anhydride and provides a maximum 60% yield of Δ6-
dehydrodeoxynupharidine, whereas trifluoroacetic anhydride
is sluggish and gives decreased yields.23 In contrast, we found
that elimination of 8 is rapid with trifluoroacetic anhydride, and
selectively produces α,β-unsaturated iminium ion 5 in less than
3 h. The production of 5 from 8 can be observed by 1H NMR
in CDCl3 and this iminium is stable over 24 h at 22 °C.
Dimerization of 5 was not expected to be straightforward.
Reineke reported the unexpected dimerization of chalcone in
the presence of polysulfides to yield tetrahydrothiophanes,24
but this reaction could only be effected with chalcones, despite
attempts by LaLonde and Clardy to apply this reaction to
B
dx.doi.org/10.1021/ja310778t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
substrates relevant to the thiophane dimers.25 Furthermore,
there have been five syntheses of quinolizidine nuphar
monomers,8,12 but no reports of successful dimerizations to
sulfur-containing binuphar alkaloids. Not surprisingly, we found
that attempts to dimerize 5 according to the conditions of
Reinecke gave complex mixtures and no dimers.
After extensive optimization, we found that addition of 0.6
equivalents of aqueous sodium tetrasulfide to a concentrated
solution of 5 in water effected dimerization, albeit in low
conversion and with poor stereoselectivity (entry 1, Table 2).
sodium borohydride allowed isolation of neothiobinupharidine
in 62% yield (from N-oxide 8), which we found remarkable
given the complexity of the transformation.
1
Isomers 1−4d were easily separated and their H and 13C
NMR spectra were correlated to values reported in the
literature. Isomer 4d (named here ‘neothionuphlutine’27) has
never been isolated from N. lutea,28 so its structure was
assigned definitively by X-ray crystallographic analysis (see
Figure 2).
Table 2. Optimization of Dimerization
Figure 2. Stereoisomer 4d: ‘neothionuphlutine.’
The stereoselectivity of dimerization deserves some dis-
cussion. Nuphar quinolizidines have been shown to react
stereoselectively on the convex face.8−10 This facial selectivity is
observed for the initial Mannich-type reaction between 5 and 6,
with an average 4.5:1 preference for convex versus concave
Mannich (3d+4d:1d+2d, Table 2). However, the subsequent
concave-face enamine sulfenylation (16) that must occur to
form 3a as the major product is more difficult to rationalize
since the stereoselectivity of this process is dependent on both
solvent choice and equivalents of polysulfide. The distribution
of diastereomers does not appear to be a result of equilibration
to a thermodynamic sink, since the ratio of products does not
change over the course of the reaction.29 Convex-face
sulfenylation (15) may be disfavored due to a steric clash in
the transition state between furyl groups, which are close in
space in 4d (see Figure 2). The role of polysulfide in affecting
diastereoselectivity is challenging to explain since it can be
involved at multiple points in the complex kinetic profile
associated with enamine reactions,30 a hypothesis which we are
currently exploring.
More importantly, the distribution of dimers from
spontaneous dimerization of 5 does not reflect the ratios of
dimers isolated from N. lutea (entry 10, Table 2). Remarkably,
the major stereoisomers 3 resulting from spontaneous
dimerization derives from the opposite stereoselectivity of C−C
and C−S bond formation than the major stereoisomers 1 in
Nuphar sp. This data points to a more complex dimerization
mechanism in vivo, which may involve enzymatic mediation or
other chiral species in the cellular milieu. Thus, while
LaLonde’s insightful proposal for the origins of the nuphar
dimers’ gross structure is supported by our studies, the data also
indicates that a more complex mechanism may be necessary to
explain the predominance of 1 and 2 in nature.
In summary, a concise (8 steps from 11b), asymmetric
synthesis of (−)-neothiobinupharidine 3d has been described.
The approach relies on a new procedure for the reductive
allylation of cyclopentenones, which accomplishes a challenging
vicinal difunctionalization reaction. To the best of our
knowledge, this is the only catalytic asymmetric approach to
the nuphar quinolizidine family. Furthermore, this reaction
enables an ‘ideal’ total synthesis, without recourse to fluctuating
a
b
1
Ratios determined by HPLC and H NMR. According to HPLC.
c
d
e
1
Monomer not detected by HPLC. Determined by H NMR. 62%
isolated yield of 3d, see Scheme 1.
However, we found that both yield and stereoselection26 could
be modulated with solvent: water, methanol, and acetonitrile
afforded modest yields, but poor selectivity (entries 1−3);
whereas acetone induced good stereoselectivity with poor
conversion (entry 4). Eventually, we found that THF and
DMSO (entries 5 and 6) provided the highest levels of
conversion and selectivity for a single stereoisomeric series, 3.
Increased equivalents of Na2S4 (entries 7−9) led to markedly
higher levels of conversion, and in DMSO provided very high
(10:1) diastereoselectivity. Reduction of the crude mixture with
C
dx.doi.org/10.1021/ja310778t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
redox manipulation or protecting group interchange.31 Access
to 1−4 (a−d) in pure form will facilitate biological evaluation
of these molecules. Efforts are currently underway to divert the
inherent topological preferences of dimerization toward the
thiobinupharidine series of dimers.
review, see: Nicolaou, K. C.; Montagnon, T.; Snyder, S. A. Chem.
Commun. 2003, 551.
(14) Beckman rearrangement in the synthesis of a related piperidine:
LaLonde, R. T.; Muhammad, N.; Wong, C. F. J. Org. Chem. 1977, 42,
2113.
(15) See: Yoon, J.; Buchwald, S. L. Org. Lett. 2001, 3, 1129 and
references therein..
(16) Degrado, S. J.; Mizutani, H.; Hoveyda, A. H. J. Am. Chem. Soc.
2001, 123, 755.
(17) Jarugumilli, G. K.; Cook, S. P. Org. Lett. 2011, 8, 1904.
(18) Gou, S.; Yea, Z.; Shib, L.; Qinga, D.; Zhanga, W.; Wang, Y. Appl.
Organometal. Chem. 2010, 24, 517.
(19) Jarugumilli, G. K.; Zhu, C.; Cook, S. P. Eur. J. Org. Chem. 2012,
1712.
(20) Chae, J.; Yun, J.; Buchwald, S. L. Org. Lett. 2004, 6, 4809.
(21) Dowling, M. S.; Vanderwal, C. D. J. Am. Chem. Soc. 2010, 75,
6908.
(22) Use of 10 equiv of DIBAL-H as in ref. 8 led to incomplete
reduction and difficult purification.
(23) LaLonde, R. T.; Auer, E.; Wong, C. F.; Muralidharan, V. P. J.
Am. Chem. Soc. 1971, 93, 2501.
(24) Del Mazza, D.; Reinecke, M. G. J. Org. Chem. 1981, 46, 128.
(25) LaLonde, R. T.; Florence, R. A.; Horenstein, B. A.; Fritz, R. C.;
Silveira, L.; Clardy, J.; Krishnan, B. S. J. Org. Chem. 1985, 50, 85.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures and spectroscopic data for all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
rshenvi@scripps.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by an ADI grant from TSRI and a
very generous donation of equipment and supplies from
Exelixis. We thank Dr. Guojun Pan for a sample of racemic 9.
We thank the Yu lab for the use of their LCMS; the Blackmond
lab for use of their GC; the Fokin lab for use of their glovebox;
and Dr. Curtis Moore and Professor Arnold L. Rheingold for X-
ray crystallographic analysis. We especially thank Professors Avi
Golan-Goldhirsh and Jacob Gopas for a donation of N. lutea
extract.
1
(26) Determined by integration of the H NMR spectrum and LC
chromatogram before purification.
(27) Nomenclature for dimers 1−4 is nonuniform in the literature.
Compound 2d was previously designated thionuphlutine B, but we
have removed the ‘B’ since there is no unique compound
‘thionuphlutine A,’ a name that was incorrectly assigned to the
known compound thiobinupharidine (1d). Compound 4d had
therefore been given the less systematic name ‘thionuphlutine C’ in
ref. 28. The ‘neo’ prefix is more effective to differentiate series 1 from 3
and 2 from 4, and refers to the stereochemistry of the C7′ quaternary
carbon.
(28) Compound 4d was produced in trace amounts by thermolysis of
a sulfoxide derived from 3d followed by reduction, but its structure
was not unequivocally proven. See: LaLonde, R. T.; Wong, C. F. Can.
J. Chem. 1978, 56, 56.
REFERENCES
■
(1) (a) Achmatowicz, O.; Wrobel, J. T. Tetrahedron Lett. 1964, 5,
129. (b) Achmatowlcz, O.; Banaszek, H.; Splteller, G.; Wrobel, J. T.
Tetrahedron Lett. 1964, 5, 927. (c) Achmatowicz, O.; Bellen, Z.
Tetrahedron Lett. 1962, 24, 1121.
(2) Birnbaum, G. I. Tetrahedron Lett. 1965, 6, 4149.
(3) For a review, see: LaLonde, R. T.; Wong, C. Pure Appl. Chem.
(29) Deteremined by HPLC analysis of reaction aliquots.
(30) Bures, J.; Armstrong, A.; Blackmond, D. G. J. Am. Chem. Soc.
2012, 134, 6741.
1977, 49, 169.
(4) Matsuda, H.; Morikawa, T.; Oda, M.; Asao, Y.; Yoshikawa, M.
Bioorg. Med. Chem. Lett. 2003, 13, 4445.
(31) Gaich, T.; Baran, P. S. J. Org. Chem. 2010, 75, 4657.
(5) Matsuda, H.; Yoshida, K.; Miyagawa, K.; Nemoto, Y.; Asao, Y.;
Yoshikawa, M. Bioorg. Med. Chem. Lett. 2006, 16, 1567.
(6) Ozer, J.; Eisner, N.; Ostrozhenkova, E.; Bacher, A.; Eisenreich,
W.; Benharroch, D.; Golan-Goldhirsh, A.; Gopas, J. Cancer Biol. Ther.
2009, 8, 1860.
(7) Wong, C. F.; LaLonde, R. T. Experientia 1975, 31, 15.
(8) Goodenough, K. M.; Moran, W. J.; Raubo, P.; Harrity, J. P. A. J.
Org. Chem. 2005, 70, 207.
(9) Lalonde, R. T.; Wong, C. F.; Das, K. C. J. Am. Chem. Soc. 1973,
95, 6342.
(10) LaLonde, R. T.; Eckert, T. S. Can. J. Chem. 1981, 59, 2298.
(11) Yoshikawa, M.; Murakami, T.; Wakao, S.; Ishikado, A.;
Murakami, N.; Yamahara, J.; Matsuda, H. Heterocycles 1997, 45, 1815.
(12) For chiral pool syntheses, see: (a) Katoh, M.; Mizutani, H.;
Honda, T. Tetrahedron Lett. 2005, 46, 5161. (b) Katoh, M.; Mizutani,
H.; Honda, T. Heterocycles 2006, 69, 193. (c) Ref. 8; (d) Moran, W. J.;
Goodenough, K. M.; Raubo, P.; Harrity, J. P. A. Org. Lett. 2003, 5,
3427. (e) Fowler, F. W.; Hwang, Y. C. J. Org. Chem. 1985, 50, 2719.
(13) For recent examples of spontaneous dimerizations to complex
secondary metabolites, see: (a) Nicolaou, K. C.; Vassilikogiannakis, G.;
Simonsen, K. B.; Baran, P. B.; Zhong, Y.-L.; Vidali, V. P.; Pitsinos, E.
N.; Couladouros, E. A. J. Am. Chem. Soc. 2000, 122, 3071. (b) Li, C.;
Lobkovsky, E.; Porco, J. A., Jr. J. Am. Chem. Soc. 2000, 122, 10484.
(c) Herzon, S. B.; Myers, A. G. J. Am. Chem. Soc. 2005, 127, 5342.
(d) Baran, P. S.; Hafensteiner, B. D.; Ambhaikar, N. B.; Guerrero, C.
A.; Gallagher, J. D. J. Am. Chem. Soc. 2006, 128, 8678. (e) For a
D
dx.doi.org/10.1021/ja310778t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX