Total synthesis of (±)-dysibetaine

Article abstract of DOI:10.1021/ol800245m

(Chemical Equation Presented) The marine natural product dysibetaine was synthesized in racemic form from a levulinic acid derivative using a convertible isocyanide and an ammonium acetate in the Ugi 4-center-3-component condensation reaction.

Full text of DOI:10.1021/ol800245m

ORGANIC
LETTERS
2008
Vol. 10, No. 7
1461-1463
Total Synthesis of (±)-Dysibetaine
Jerry Isaacson, Mandy Loo, and Yoshihisa Kobayashi*
Department of Chemistry and Biochemistry, UniVersity of California, San Diego,
9500 Gilman DriVe Mail Code 0343, La Jolla, California 92093-0343
ykoba@chem.ucsd.edu
Received February 2, 2008
ABSTRACT
The marine natural product dysibetaine was synthesized in racemic form from a levulinic acid derivative using a convertible isocyanide and
an ammonium acetate in the Ugi 4-center-3-component condensation reaction.
(-)-Dysibetaine (1) is a member of a group of natural
products that were isolated by Sakai and co-workers in the
late 1990s from the Micronesian sponge Dysidea herbacea.1
It was first synthesized by Snider and Gu,2 who also
established the absolute stereochemistry of the natural
product, and has been totally synthesized or studied syntheti-
cally by three other groups.3 We were drawn to dysibetaine
by its potential biological activity and because of our desire
to employ the Ugi four-center three-component reaction,
which, along with a recently developed convertible iso-
cyanide 2, would allow us to quickly assemble the cyclic
pyroglutamic acid core from a linear precursor (Figure 1).
derivatives.6 Using this technology, our laboratory reported
a stereocontrolled formal total synthesis of omuralide that
employed 2 as a convertible isocyanide in a U4C-3CR.7 We
now wish to report the application of this technology to the
racemic total synthesis of dysibetaine.
Following the U4C-3CR, the amide derived from iso-
cyanide 2 is readily converted to the corresponding N-
acylindole, which is then easily hydrolyzed.8 Indole formation
dramatically alters the character of the amide by decreasing
the double-bond character of the C-N amide bond and
greatly weakens the bond. This, in turn, makes it possible
to selectively cleave this amide bond to the corresponding
ester under very mild conditions.
(1) (a) Sakai, R.; Oiwa, C.; Takaishi, K.; Kamiya, H.; Tagawa, M.
Tetrahedron Lett. 1999, 40, 6941-6944. (b) Sakai, R.; Suzuki, K.;
Shimamoto, K.; Kamiya, H. J. Org. Chem. 2004, 69, 1180-1185.
(2) Snider, B. B.; Gu, Y. Org. Lett. 2001, 3, 1761-1763.
(3) (a) Wardrop, D. J.; Burge, M. S. Chem. Commun. 2004, 1230-1231.
(b) Langlois, N.; Le Nguyen, B. K. J. Org. Chem. 2004, 69, 7558-7564.
(c) Katoh, M.; Hisa, C.; Honda, T. Tetrahedron Lett. 2007, 48, 4691-
4694.
(4) (a) Ugi, I.; Do¨mling, A. Angew. Chem., Int. Ed. 2000, 39, 3168-
3210. (b) Do¨mling, A. Chem. ReV. 2006, 106, 17-89.
(5) (a) Short, K. M.; Mjalli, A. M. M. Tetrahedron Lett. 1997, 38, 359-
362. (b) Harriman, G. C. B. Tetrahedron Lett. 1997, 38, 5591-5594. (c)
Hanusch-Kompa, C.; Ugi, I. Tetrahedron Lett. 1998, 39, 2725-2728.
(6) (a) Isaacson, J.; Gilley, C. B.; Kobayshi, Y. J. Org. Chem. 2007, 72,
3913-3916. (b) Vamos, M.; Ozboya, K.; Kobayashi, Y. Synlett 2007,
1595-1599. (c) Kreye, O.; Westermann, B.; Wessjohann, L. A. Synlett
2007, 3188-3192.
Figure 1. Structures of (-)-dysibetaine (1) and convertible
isocyanide 2.
The Ugi four-center three-component reaction (U4C-3CR)4
has long been established as a useful method to create
pyroglutamic acid amides from an amine, isocyanide and
γ-ketoacid.5 Until recently, however, it was not possible to
elaborate the products to their corresponding carboxylic acid
(7) Gilley, C. B.; Buller, M. J.; Kobayashi, Y. Org. Lett. 2007, 9, 3631-
3634.
(8) Arai, E.; Tokuyama, H.; Linsell, M. S.; Fukuyama, T. Tetrahedron
Lett. 1998, 39, 71-74.
10.1021/ol800245m CCC: $40.75
© 2008 American Chemical Society
Published on Web 03/05/2008

The plan for our synthesis of dysibetaine is shown in
Scheme 1. From the natural product, a few functional group
Preparation of the γ-ketoacid 8 was completed by TBS-
protection of the primary alcohol to give 7 (not shown). This
was followed by selective hydrogenolysis of the benzyl ester,
cleanly yielding the free acid (Scheme 3).
Scheme 1. Retrosynthetic Analysis of Dysibetaine (1)
Scheme 3. Ugi Reaction of Levulinic Acid Derivative 8 to
Afford Pyroglutamic Acid Amide 10
modifications lead retrosynthetically to the N-acylindole A.
This activated intermediate, which is readily transformed to
the corresponding ester, would be derived from a linear
γ-ketoacid precursor B using the U4C-3CR and 2. B, in turn,
could be synthesized from 2-bromoallyl ester C via a
sequence of Ireland-Claisen rearrangement followed by
oxidation of the resulting exo-bromoolefin.
The stage was now set for the U4C-3CR. The amine and
isocyanide components were carefully chosen to ensure
success later in the synthesis. We necessarily chose 2 as the
isocyanide because the Ugi product amide derived from 2
is readily cleaved (vide infra). Although primary amines are
most commonly used in the U4C-3CR, we have found that
ammonium acetate works well in this reaction as an ammonia
equivalent.6 This eliminates the need for a later deprotection
step that would likely require harsh conditions.11 We believe
that the intramolecular relationship between the ketone and
acid in the γ-ketoacid prevents acetic acid from being
incorporated as the acid component in the Ugi reaction, and
no such product is ever observed. The U4C-3CR also
proceeds smoothly using methanol as solvent, but our
experience has shown that TFE gives better yields.
When the U4C-3CR was tried with the primary alcohol
unprotected (obtained by hydrogenation of 6), the yield was
greatly diminished (26%). When the reaction was tried with
the secondary alcohol unprotected (obtained by thorough
hydrogenation of 7), a similar decrease in yield was seen.
We therefore decided on the fully protected γ-ketoacid 8
and were satisfied to obtain the Ugi product 9 in 86% yield
as a mixture of diastereomers.
Scheme 2. Preparation of a Functionalized Levulinic Acid by
Ireland-Claisen Rearrangement
Scheme 2 shows the plan in action as we commenced by
constructing the necessary γ-ketoacid precursor. Coupling
commercially available 2,3-dibromopropene and benzyloxy-
acetic acid (not shown) using potassium carbonate in DMF
provided ester 3 in 86% yield. Treating 3 with LHMDS in
the presence of TMSCl affected the Ireland-Claisen re-
arrangement to yield acid 4 as a racemic mixture.9 Benzyl
protection of the carboxylic acid afforded 5, which was then
oxidized using the Upjohn method.10 Dihydroxylation of the
bromo-olefin allowed regiocontrolled installation of the
ketone and alcohol functionalities in a single step, affording
6 in excellent yield.
Moving forward, the TBS group of the Ugi product 9 was
removed using TBAF in THF and at this stage the dia-
stereomers were readily separated by silica gel chromatog-
raphy to give 10 as a single racemic diastereomer. X-ray
analysis revealed that the minor isomer, 10, which is less
polar by TLC, was the isomer needed to reach dysibetaine.
The results of the X-ray analysis of 10 are included in
Figure 2.12
(9) (a) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem. Soc.
1976, 98, 2868-2877. (b) Koch, G.; Janser, P.; Kottirsch, G.; Romero-
Giron, E. Tetrahedron Lett. 2002, 43, 4837-4840.
(10) VanRheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
17, 1973-1976.
(11) Corey, E. J.; Li, W.; Nagamitsu, T. Angew. Chem., Int. Ed. 1998,
37, 1676-1679.
(12) Please see Supporting Information for complete X-ray crystal data.
1462
Org. Lett., Vol. 10, No. 7, 2008

Scheme 5. Completion of Total Synthesis of (()-Dysibetaine
(1)
Figure 2. ORTEP projection of X-ray crystal structure of 10.
With 10 in hand, we were ready for the final functional
group manipulations needed to reach (()-dysibetaine (1).
The primary alcohol in 10 (Scheme 4) was protected as the
ether and reduction of the azide to the corresponding amine
15, which matched the corresponding synthetic intermediate
reported by the Langlois group.3b 15 was elaborated to the
final product using the three step procedure devised for the
corresponding ethyl ester by Snider,2 consisting of Clarke-
Eschweiler dimethylation of the amine followed by quater-
nization with iodomethane and cleaveage of the methyl ester
using basic resin.3b We thus arrived at (()-dysibetaine (1)
in 16 steps and 9% overall yield from commercially available
Scheme 4. Methanolysis of the Anilide 10 via N-Acylindole
12
1
materials. The H and 13C NMR spectra of the synthetic
material corresponded closely to those reported for the natural
product.1
We have completed a total synthesis of the biologically
interesting molecule (()-dysibetaine (1) using the Ugi four-
center three-component reaction and a recently developed
cleavable isocyanide 2. The present synthesis also revealed
that a more efficient synthesis could be achieved by
incorporating the amine into the γ-ketoacid before the Ugi
reaction and by finding a more efficient synthesis of the
necessary γ-ketoacid precursor. Using these lessons, an
asymmetric synthesis of dysibetaine is underway and will
be reported in due course.
mesylate using MsCl and Et3N in dichloromethane to give
11. Hydrolysis of the external amide of 11 is a formidable
task due to the presence of another amide in the molecule.
It is further complicated by the presence of an adjacent, fully
substituted carbon center. Fortunately, the inclusion of 2 as
the isocyanide in the U4C-3CR provides a facile solution to
the problem. Heating 11 with CSA in benzene for less than
an hour affords the N-acylindole 12. Methanolysis to 13 was
achieved using only a catalytic amount of aqueous sodium
hydroxide. The amide hydrolysis is notable for the mild,
selective conditions, compared to the very strong acid or
base/refluxing conditions that would normally be used.
Scheme 5 illustrates the final steps toward the synthesis
of (()-dysibetaine (1). Mesylate 13 was heated in DMF with
sodium azide to afford azide 14 in 64% yield (plus 21%
recovered starting material). Palladium-charcoal hydrogena-
tion of 14 allowed concomitant deprotection of the benzyl
Acknowledgment. We acknowledge Dr. Yongxuan Su
of the UCSD Mass Spectrometry Facility as well as Dr.
Antonio DiPasquale and Professor Arnold Rheingold of the
UCSD Small Molecule X-Ray Facility. We also thank
Cynthia Gilley (UCSD) for her experimental help at the
preliminary stage.
Supporting Information Available: Detailed experi-
mental prodcedure, copies of 1H and 13C NMR spectra, and
CIF file for compound 10. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL800245M
Org. Lett., Vol. 10, No. 7, 2008
1463