the desired stereochemistry in 87% yield. Debenzylation of
16 under hydrogenolysis conditions afforded 17 in a rather
lower yield (48%), which was then transformed into (()-
goniomitine in a quantitative yield as shown in Scheme 3.
Thus the new total synthesis of (()-1wasachieved, although
it is obvious that optimization of the chemical yields, in
particular, the conversion of 16 into 17, still remains.
Our next phase was to complete the convergent total
synthesis of the natural (-)-goniomitine (1) using the
chiral lactam species instead of 13 for the reaction with
the 2-vinylindole derivative 10 (Scheme 4). Amat12
reported a reliable procedure for the preparation of some
chiral oxazolopiperidone lactams from phenylglycinol and
δ-valerolactone. Thus, the known oxazolopiperidone lac-
tam 18,12 prepared by Amat’s procedure, was successively
treated with ozone and NaBH4 to give the one-carbon
shortened 19 in 87% yield, which was then dehydrated
under Grieco’s conditions, providing the desired vinyllactam
20 in 76% yield. The aforementioned cross-metathesis
condition that was effective for the reaction between the
indole 10 and the 3-vinylpiperidin-2-one derivative 13 was
unfortunately found to be fruitless, but reexamination of
the conditions successfully provided a better chemical
yield. Namely, a solution of 10 and the vinyllactam 20 in
xylene was heated at 140 °C in the presence of 30 mol % of
the Hoveyda-Grubbs-II catalyst9 to furnish 21 in 65%
yield.13 Compound 21 was subsequently hydrogenated to
afford 22 in 92% yield. The DIBAL-mediated one-step
construction of the tetracyclic framework described in the
transformation of 15 to 16 did not work in the case of 22,
and no reaction took place, presumably due to the serious
steric congestion around the lactam. We tentatively
assumed that the nitrogen anion derived from 22 by base
treatment might coordinate with DIBAL, leading to the
formation of certain ate-complexes, which should be more
powerful hydride sources than DIBAL itself. On the basis
of this assumption, several bases were added to 22 before
treating with DIBAL. As a result, sodium hydride gave the
best result in which the tetracyclic compound 23 was
exclusively obtained in 62% yield accompanied by cleav-
age of the oxazolidine ring. The highly stereoselective
construction of the diazabicyclo[4.4.0] framework was
again achieved similar to the case of 16. Removal of the
protecting groups on both the nitrogen and oxygen atoms
of 23 produced (-)-goniomitine (1) in 61% yield. Further-
more, the synthetic route in Scheme 4 with use of ent-20
instead of 20 also afforded the unnatural (þ)-goniomitine.
With natural, unnatural, and racemic goniomitine now in
hand, we executed their preliminary bioactive assay as follows.
To determine whether goniomitine is effluxed by human
P-glycoprotein (P-gp, a gene product of MDR1), antipro-
liferative activity of goniomitine was studied in canine kidney
MDCK II with enforced expression of P-gp (MDCK/
MDR1) and Mock-transfected MDCK II (Mock) cells. As
Scheme 4. Completion of Total Synthesis of (-)-Goniomitine
(1)
shown in Table 1 and in Figure 1 in the Supporting Informa-
tion, no significant increase in 50% growth inhibitory
concentration (IC50) of (þ)-, (-)-, and (()-goniomitine was
observed in MDCK/MDR1 as compared with those in Mock
cells. However, it is noteworthy that a significant difference
between the stereoisomers was found in the antiproliferative
activity. The IC50 of (-)-goniomitine was 2.40- and 3.18-fold
lower than those in Mock and MDCK/MDR1 cells, respec-
tively, whereas the values were very close to those of
(()-goniomitine. This observation suggested that (-)-gonio-
mitine is more potent to retard cell growth than (þ)-gonio-
mitine. Although the detailed mechanism of greater antipro-
liferative activity of (-)-goniomitine needs to be investigated,
our establishment of the synthesis of (-)-goniomitine may
provide us with a clue to develop novel anticancer agents.
Table 1. Antiproliferative Activity of (þ)-, (-)-, and (()-Go-
niomitine (1)
substrate
cell lines
IC50 [μM]a
(þ)-goniomitine
Mock
160.7 ( 28.2
174.8 ( 22.1
66.8 ( 8.8
55.3 ( 6.0
36.8 ( 7.2
55.7 ( 9.1
MDR-1-MDCK
Mock
(-)-goniomitine
(()-goniomitine
MDR-1-MDCK
Mock
MDR-1-MDCK
a IC50 values were estimated by fitting proliferation assay data to Hill
plot as described in the Supporting Information. Experiments were
repeated at least twice with n = 6.
In summary, the total syntheses of the natural (-)-
goniomitine and unnatural (þ)-goniomitine were attained
(in 10 steps from the commercially available 3-butyn-1-ol
or 13 steps from the commercially available δ-valerolactone)
ꢀ
ꢀ
(12) (a) Amat, M.; Bosch, J.; Hidalgo, J.; Canto, M.; Perez, M.; Llor,
N.; Molins, E.; Miravitlles, C.; Orozco, M.; Luque, J. J. Org. Chem.
2000, 65, 3074–3084. (b) Amat, M.; Lozano, O.; Escolano, C.; Molins,
E.; Bosch, J. J. Org. Chem. 2007, 72, 4431–4439.
(13) The unreacted 20 was recovered and reusable.
1798
Org. Lett., Vol. 13, No. 7, 2011