Synthesis of nonactin and the proposed structure of trilactone

Article abstract of DOI:10.1021/ol0609661

An efficient enantioselective route to nonactin using a novel β-inversion of an Evans syn aldol to construct the THF ring is presented. Through total synthesis, the structure for trilactone proposed in the literature is shown likely to be incorrect.

Full text of DOI:10.1021/ol0609661

ORGANIC
LETTERS
2006
Vol. 8, No. 13
2831-2834
Synthesis of Nonactin and the Proposed
Structure of Trilactone
Yikang Wu* and Ya-Ping Sun
State Key Laboratory of Bioorganic and Natural Products Chemistry,
Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences,
354 Fenglin Road, Shanghai 200032, China
yikangwu@mail.sioc.ac.cn
Received April 22, 2006
ABSTRACT
An efficient enantioselective route to nonactin using a novel
â-inversion of an Evans syn aldol to construct the THF ring is presented.
Through total synthesis, the structure for trilactone proposed in the literature is shown likely to be incorrect.
Nonactin (1) is the simplest member of a large family of
ionophore antibiotics known as macrotetrolides or polynac-
tins generated by Streptomyces griseus subspecies griseus
ETH A7796 (DSM40695).1 These compounds display pro-
nounced antibacterial,2 insecticidal,3 antitumoral,4 and im-
munosuppressive5 activities and therefore have received
considerable attention from synthetic chemists. Up to now,
most synthetic studies on the polynactins (Figure 1) have
been directed toward nonactin, presumably because it has
Figure 1. General structure of polynactins. For nonactin (1),
R1-R4 ) Me.
the highest symmetry and thus requires the minimum
(1) For reviews, see: (a) Keller-Schierlein, W.; Gerlach, H. Fortschr.
Chem. Org. Naturst. 1968, 26, 161-189. (b) Nawata, Y.; Ando, K.; Iitaka,
Y. Met. Ions Biol. Syst. 1985, 19, 207-227. (c) Zizka, Z. Folia Microbiol.
1998, 43, 7-14.
(2) Nefelova, M. V.; Sverdlova, A. N. Antibiot. Med. Biotekhnol. 1985,
30, 261-264; Chem. Abstr. 1985, 102, 201065.
workload in synthesis. In this communication, we wish to
report a new approach to nonactin6,7 and the recently
proposed8 structure for trilactone 2, a natural product related
to nonactin.
(3) (a) Oishi, H.; Sugawa, T.; Okutomi, T.; Suzuki, K.; Hayashi, T.;
Sawada, M.; Ando, K. J. Antibiot. 1970, 23, 105-106. (b) Shopotova, L.
P.; Shenin, Y. D. Zh. Prikl. Khim. 1993, 66, 1334-1338; Chem. Abstr.
1994, 120, 158329.
(4) Borrel, M. N.; Pereira, E.; Fiallo, M.; Garnier-Suillerot, A. Eur. J.
Biochem. 1994, 223, 125-133.
(5) (a) Callewaert, D. M.; Radcliff, G.; Tanouchi, Y.; Shichi, H.
Immunopharmacology 1988, 16, 25-32. (b) Tanouchi, Y.; Shichi, H.
Immunology 1988, 63, 471-475.
(6) There are five total syntheses of nonactin in the literature: (a) Gerlach,
H.; Oertle, K.; Thalmann, A.; Servi, S. HelV. Chim. Acta 1975, 58, 2036-
2043. (b) Schmidt, U.; Gombos, J.; Haslinger, E.; Zak, H. Chem. Ber. 1976,
109, 2628-2644. (c) Bartlett, P. A.; Meadows, J. D.; Ottow, E. J. Am.
Chem. Soc. 1984, 106, 5304-5311. (d) Kim, B. H.; Lee, J. Y. Tetrahedron
1996, 52, 571-588. (e) Fleming, I.; Ghosh, S. K. J. Chem. Soc., Perkin
Trans. 1 1998, 2733-2747. For a comprehensive review, see: (f) Fleming,
I.; Ghosh, S. K. Stud. Nat. Prod. Chem. 1996, 18, 229-268.
10.1021/ol0609661 CCC: $33.50
© 2006 American Chemical Society
Published on Web 05/25/2006

Because nonactin consists of residues of (+)- and
(-)-nonactic acids, its synthesis usually starts from the
construction of these two building blocks. In the present
work, the stereogenic center at C-8 of the nonactic acids was
derived through a Jacobsen hydrolytic kinetic resolution
(HKR)9 of racemic 2-methyl-oxirane. The epoxide and the
diol obtained in the HKR were used in the synthesis of
(+)- and (-)-nonactic acid, respectively, and thus made full
use of the racemic starting material. The C-6 stereogenic
center was established by a Me4NBH(OAc)310 reduction. The
C-2 and C-3 were first generated by a Crimmins11 modifica-
tion of the Evans aldolization as a syn aldol. Then, in a later
step when closing the THF ring, the configuration at C-3
was inverted.
Because the â-inversion in the presence of the carbonyl
group has never been reported before and potential risks such
as R-racemization and â-elimination do exist, to be on the
safe side, we did a model study to check the feasibility of
the â-inversion strategy (Scheme 1) before starting the real
synthesis.
Table 1. Representative Results of the Attempted Cyclization
of 7
entry
conditions (concn of 7, M)
outcome
1
2
3
4
5
6
7
Et3N/CH2Cl2/rt/24 h (0.02)
NaH/THF/rt/3 h (0.08)
no reaction
9 only
9 only
9 only
9 only
NaH/THF/-78 °C to rt/1 h (0.08)
NaH/THF/-78 °C to rt/1 h (0.02)
NaH/THF/-78 °C to rt/2 h (0.005)
NaH/THF/-78 to -40 °C/4 h (0.001)
NaHDMS/THF/HMPA/-78 to
-40 °C/1 h (0.01)
no reaction
9 only
8
9
10
11
12
13
NaHDMS/Et2O/-100 °C/1 h (0.07)
t-BuOK/THF/rt/6 h (0.08)
t-BuONa/THF/rt/12 h (0.08)
pyridine (solvent)/140 °C/6 h (0.05)
2,6-lutidine (solvent)/120 °C/1 h (0.06)
2,6-lutidine (solvent)/75 °C/4.5 h (0.04)
9 only
no reaction
no reaction
8 only
8 only
no reaction
excess base, â-elimination occurred readily in most cases
(Table 1, entries 2-5, 7, and 8), giving conjugated alkene 9
as the only product. Finally, we gratifyingly found that the
expected 8 was formed cleanly (Table 1, entry 11) when
using pyridine as solvent and running the reaction temper-
ature at 140 °C (bath). With 2,6-lutidine, the temperature
could be lowered to 120 °C (Table 1, entry 12). At even
lower temperature, essentially no reaction took place (Table
1, entry 13).
Scheme 1
With a satisfactory means to close the THF ring in hand,
we turned to the synthesis of (+)-nonactic acid (Scheme 3).
Deprotonation of the dithiane 1014 with n-BuLi followed by
reaction with epoxide 119 gave alcohol 12 in 96% yield. The
sulfur protecting group was removed with I2,15 and the
carbonyl group was stereoselectively reduced with Me4NBH-
Condensation of aldehyde 312 with imide 413 under the
Crimmins conditions led to syn aldol 5 in 75% isolated yield
(Scheme 2). The â-OH was then converted to the corre-
(7) For syntheses of nonactic acid published after 1996, see: (a) Bratt,
K.; Garavelas, A.; Perlmutter, P.; Westman, G. J. Org. Chem. 1996, 61,
2109-2117. (b) Kiyota, H.; Abe, M.; Ono, Y.; Oritani, T. Synlett 1997,
1093-1095. (c) Takatori, K.; Tanaka, K.; Matsuoka, K.; Morishita, K.;
Kajiwara, M. Synlett 1997, 159-160. (d) Mandville, G.; Girard, C.; Bloch,
R. Tetrahedron 1997, 53, 17079-17088. (e) Meiners, U.; Cramer, E.;
Frohlich, R.; Wibbeling, B.; Metz, P. Eur. J. Org. Chem. 1998, 2073-
2078. (f) Ahmar, M.; Duyck, C.; Fleming, I. J. Chem. Soc., Perkin Trans.
1 1998, 2721-2732. (g) Lee, E.; Choi, S. J. Org. Lett. 1999, 1, 1127-
1128. (h) Wang, Y.; Metz, P. Tetrahedron: Asymmetry 2000, 11, 3995-
3999. (i) Fraser, B.; Perlmutter, P. J. Chem. Soc., Perkin Trans. 1 2002,
2896-2899. (j) Jeong, J. W.; Woo, B. Y.; Ha, D. C.; No, Z. Synlett 2003,
393-395. (k) Ireland, R. E.; Vevert, J. P. Can. J. Chem. 1981, 59, 572-
583. (l) Kim, B. H.; Lee, J. Y. Tetrahedron Lett. 1992, 33, 2557-2560.
For other routes published by 1996, see ref 6f above.
Scheme 2
(8) Rezanka, R.; Spizek, J.; Prikrylova, V.; Prell, A.; Dembitskyb, V.
M. Tetrahedron 2004, 60, 4781-4787.
(9) Schaus, S. E.; Brandes, B. D.; Larrow, J. F.; Tokunaga, M.; Hansen,
K. B.; Gould, A. E.; Furrow, M. E.; Jacobsen, E. N. J. Am. Chem. Soc.
2002, 124, 1307-1315 and references therein.
(10) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.
1988, 110, 3560-3578.
(11) Crimmins, M. T.; King, B. W.; Tabet, E. A. J. Am. Chem. Soc.
1997, 119, 7883-7884.
(12) Matsuda, F.; Kito, M.; Sakai, T.; Okada, N.; Miyashita, M.;
Shirahama, H. Tetrahedron 1999, 55, 14369-14380.
(13) Ager, D. J.; Allen, D. R.; Schaad, D. R. Synthesis 1996, 1283-
1285.
sponding mesylate with careful control of the quantity of
the base. The benzyl protecting group was removed by
hydrogenolysis to give the ring-closure precursor 7.
The ring closure was then examined under a variety of
conditions (Table 1). Despite the precaution to avoid using
(14) Wu, Y.-K.; Huang, J.-H.; Shen, X.; Hu, Q.; Tang, C.-J.; Li, L. Org.
Lett. 2002, 4, 2141-2144.
(15) Nicolaou, K. C.; Bunnage, M. E.; McGarry, D. G.; Shi, S.; Somers,
P. K.; Wallace, P. A.; Chu, X.-J.; Agrios, K. A.; Gunzner, J. L.; Yang, Z.
Chem. Eur. J. 1999, 5, 599-617.
2832
Org. Lett., Vol. 8, No. 13, 2006

21a were then converted to acid 23 and ester 24a, respec-
tively, and coupled to each other by treatment with DCC/
DMAP to yield dimer 25 (Scheme 4). Part of 25 was
Scheme 3
Scheme 4
(OAc)3 to create the stereogenic center at C-6. The two
diastereomers (ca. 17:1) could be cleanly separated on silica
gel. The diol 14 was then masked as an acetonide, and the
terminal hydroxyl group was freed and oxidized to give the
intermediate aldehyde 17, which was immediately treated
with imide 4 under the Crimmins conditions to give the syn
aldol 18.
The â-hydroxyl group was then converted to a mesylate.
Using the conditions developed in the model reaction, we
obtained the desired ring-closure product 20 in 90% yield
as expected. On hydrolytic removal of the chiral auxiliary
with LiOH/H2O2, (+)-nonactic acid 21 was obtained. Be-
1
cause the H NMR data for 21 in the literature were not as
well resolved as those for its methyl ester, we also trans-
formed 21 to ester 22 (94% yield) by treatment with CH2N2.
We were pleased to find that all spectroscopic data including
optical rotation of ester 22 were in full consistence with those
reported in the literature, further confirming that neither
racemization nor elimination-addition occurred during the
construction of the THF ring under the unconventional
lutidine conditions. It is interesting to note that later we found
that the ring closure of mesylate 19 (neat) could also occur
spontaneously at ambient temperature after 10 days.
The (-)-nonactic acid 21a was synthesized in a similar
fashion and comparable yield, but with (S)-2-methyl-oxirane
11a (prepared from the diol generated in the HKR) and the
antipode of imide 4 in place of epoxide 11 and 4, respec-
tively.
deprotected on the hydroxyl group end to afford alcohol 26.
The remainder was subjected to hydrogenolysis to release a
free carboxylic group. Alcohol 26 and acid 27 were coupled
to give tetramer 28. Removal of the benzyl group and the
TBS group, however, was done in one operation by extending
the hydrogenolysis time to 36 h instead of in two steps as
described6e by Fleming and Ghosh. A final Yamaguchi
lactonization gave the end product nonactin 1 in 80% yield,
with all spectroscopic data in full consistence with those
reported6e in the literature.
In 2004, Rezanka8 and co-workers reported a novel
nonactin natural analogue trilactone 2. Because this com-
pound has never been synthesized up to now, we performed
the work outlined in Scheme 5. In the endeavor, the acid 21
The remaining steps were done essentially following the
route described by Fleming and Ghosh.6e Thus, acids 21 and
Org. Lett., Vol. 8, No. 13, 2006
2833

1
Scheme 5
Table 2. Comparison of the H NMR Data
signals (δ) of 2
signals (δ) of trilactone (ref 8)
1.09 (d, J ) 7.0 Hz, 9H)
1.26 (d, J ) 6.3 Hz, 9H)
1.68-1.42 (m, 6H)
1.10 (dd, J ) 2.7, 6.7 Hz, 9H)
1.25 (d, J ) 6.4 Hz, 9H)
1.40 (m, 3H)
1.48 (ddd, J ) 14.0, 11.3, 3.4 Hz, 3H)
1.63 (m, 3H)
1.85-1.70 (m, 6H)
2.00-1.90 (m, 6H)
1.80 (ddd, J ) 14.0, 12.1, 3.5 Hz, 3H)
1.82 (m, 3H)
1.98 (m, 3H)
2.44 (quint, J ) 7.3 Hz, 3H) 2.57 (dq, J ) 9.8, 6.7 Hz, 3H)
3.88 (tt, J ) 5.2, 8.2 Hz, 3H) 3.90 (m, 3H)
3.95 (quart, J ) 6.9 Hz, 3H) 4.01 (dddq, J ) 10.0, 9.8, 7.0,
2.7 Hz, 3H)
4.94 (sext, J ) 6.3 Hz, 3H) 4.90 (ddq, J ) 11.3, 6.4, 3.5 Hz, 3H)
confirming that the structure previously proposed for trilac-
tone is likely incorrect.
In brief, a new total synthesis of nonactin has been
developed. (R)- and (S)-2-methyl-oxirane obtained from
Jacobsen HKR of the racemate were utilized to construct
the C-8 stereogenic center of (+)- and (-)-nonactic acid,
respectively. The C-6 configuration was established through
a 1,3-asymmetric reduction. The C-2/C-3 stereogenic centers
were introduced by a Crimmins modification of the Evans
aldolization followed by a â-inversion when forming the THF
ring. Such a transformation with the leaving group at the
carbon â to a carbonyl group (which made the substrate labile
to such side reactions as R-racemization and â-elimination),
to the best of our knowledge, has never been reported before.
Finally, through total synthesis, the structure previously
proposed for trilactone is shown likely to be incorrect.
was first converted to its benzyl ester 24 by treatment with
t-BuOK/BnBr, which was coupled with acid 23 to yield the
dimer 30. The TBS protecting group was then removed with
HOAc/TsOH, and another molecule of 23 was coupled onto
the dimer, giving trimer 32. Deprotection of both the
carboxylic and the hydroxyl group ends was realized by
hydrogenolysis over Pd-C. Finally, the trilactone ring was
closed under the Yamaguchi conditions to afford the target
molecule 2. However, to our surprise, neither the optical
rotation ([R]24D +1.5 (c 0.1, EtOH); cf lit.8 [R]25D +14.5 (c
0.09, EtOH)) nor the NMR data (Table 2 and Supporting
Information) of 2 were consistent with those reported8 for
trilactone. We then noticed that in that paper the data for
“methyl nonactate” obtained by degradation of trilactone
were also incompatible with those reported7 by others,
Acknowledgment. This work was supported by the
National Natural Science Foundation of China (20025207,
20272071, 20372075, 20321202), the Chinese Academy of
Sciences (Knowledge Innovation project, KGCX2-SW-209),
and the 973 Program (G2000077502).
Supporting Information Available: Experimental pro-
cedures, physical and spectroscopic data for all new com-
1
pounds, H NMR spectra of 5-9, 12a, 13a, 14a, 15a, 16a,
18, 18a, 19, 19a, 20a, 24, 30-33, and 2, and 13C NMR
spectra of 14a, 18, 32, and 2 (41 pages). This material is
available free of charge via the Internet at http://pubs.acs.org.
OL0609661
2834
Org. Lett., Vol. 8, No. 13, 2006