Angewandte
Communications
Chemie
diketone 58 (prepared in 15 steps) as one of their intermedi-
ates. We hoped to forge the same compound in a more
expeditious manner using our pyrone Diels–Alder reaction to
thus achieve a formal total synthesis of the target. As shown in
Scheme 3, following an efficient preparation of 49, we
believed that subsequent use of our cascade strategy could
afford 50, which would need two subsequent changes in
oxidation state and an installation of the required N-methyl
group to complete the target diketone. As matters transpired,
this hypothesis proved true, although the final tailoring of the
structure required specific ordering of events and careful
control of reaction conditions.
proceeding smoothly on gram scale. From here, our next
goal was to reduce the olefinic portion of the vinylogous
amide. Unfortunately, as observed by others,[21] this operation
proved challenging, with several conditions such as catalytic
hydrogenation failing to deliver anything other than recov-
ered starting material.[22] Radical reduction conditions such as
[Mn(dpm)3] (dpm = dipivaloylmethane), PhSiH3, and tert-
butyl hydroperoxide (TBHP) in iPrOH,[23] by contrast, did
afford the reduced product in low conversion (ca. 5%), but
resulted in the incorrect configuration at the ring junction in
the form of 51, presumably because of the steric bulk of the
adjacent aromatic ring. Pleasingly, however, use of conditions
deployed by Rigby et al. on a related system,[21a] and originally
pioneered by Molander and McKie,[24] provided a solution.
In that key event, vinylogous amide 50 was treated with
6 equiv of SmI2, activated with 24 equiv of HMPA, and stirred
in THF at 238C for 5 h; those conditions delivered 53 as
a single, unassigned diastereomer about the starred carbon
atom with the correct ring junction configuration. In this
process we believe that 51 is initially formed, with subsequent
elimination and reattack of the amine onto the resultant
Michael system of 52 affording the needed cis-fusion in the
hydroindoline core;[25] over the time course of the reaction,
the resultant ketone is then reduced to afford 53, a material
which, for purposes of full characterization, was oxidized with
pyridinium dichromate (PDC) to afford 54 in 39% yield. In
support of this proposal, a short reaction time of only 15 min
affords isolable quantities of 51. By contrast, the use of
additional equivalents of SmI2 or adding Sm metal to the
original mixture afforded 55 after subjecting the crude
reaction products to a Salmond oxidation (CrO3, 3,5-dime-
thylpyrazole).[26] This outcome suggests that the a,b-unsatu-
rated system of putative intermediate 52 had been reduced
prior to reclosure of the ring. The identity of 55 was confirmed
by its conversion, in two further steps, into 56,[27] a side
product Gao and co-workers had also prepared and whose
1H NMR data[4b] were in full agreement with our synthetic
material.
Nevertheless, from 53 the target molecule was readily
completed by use of a Salmond oxidation[26] to effect
oxidation of both the alcohol and the lone benzylic position
to deliver 57 (20% overall from 50),[28] and a one-pot Teoc
cleavage/methylation[29] to afford 58 (in 85% yield from 57).
In total, only 10 steps were required to prepare 58 from
commercial materials, thereby affording a formal total syn-
thesis of (Æ)-gracilamine (5) in 14 steps overall.
Scheme 3. Concise, formal total synthesis of gracilamine (5) through
a 10-step synthesis of ketone 58: a) toluene, 1708C under microwave
irradiation, 12 h, then silica gel, CHCl3, 238C, 5 h, 83%; b) SmI2 (0.1m
in THF, 6.0 equiv), HMPA (24 equiv), tBuOH (2.0 equiv), THF, 238C,
5 h; c) Sm metal (6.0 equiv), SmI2 (0.1m in THF, 6.0 equiv), HMPA
(24 equiv), tBuOH (2.0 equiv), THF, 238C, 10 h; d) CrO3 (20 equiv),
3,5-DMP (20 equiv), CH2Cl2, À40 to 238C, 4 h, 15% over two steps;
e) TBAF (20.0 equiv), THF, 08C, 1 h; f) K2CO3 (20 equiv), MeI
(6.4 equiv), MeCN, 408C, 3 h, 54% over two steps; g) CrO3 (20 equiv),
3,5-DMP (20 equiv), CH2Cl2, À40 to 238C, 4 h, 20% over two steps;
similar oxidation from 54 proceeded in 51% yield; h) TFA/CH2Cl2
(1:1), 238C, 10 min, then K2CO3 (100 equiv), MeI (20 equiv), MeCN,
238C, 3 h, 85%. HMPA=hexamethylphosphoramide, DMP=dimethyl-
pyrazole, TBAF=tetra-n-butylammonium fluoride, Teoc=2-(trimethyl-
silyl)ethoxycarbonyl, TFA=trifluoroacetic acid.
In summary, a modular and procedurally simple sequence
from 4,6-dichloropyrone (7) and a range of readily prepared
amines has afforded diverse indoline and hydroindoline
frameworks. These materials are of pertinence to pharma-
ceuticals as well as natural products, with three total or formal
syntheses—including the shortest route to gracilamine in
terms of step count—being achieved as a result. Although the
core reactions rely upon the venerable Diels–Alder reac-
tion,[30] the particular variant deployed here has not, to the
best of our knowledge, been explored in any depth. Numerous
opportunities for the construction of novel heterocycles and
asymmetric variants are apparent, and are the subject of
current investigations.
Following a six-step synthesis of 49 (see the Supporting
Information),[20] the use of our cascade reaction conditions,
albeit at 1708C for 10 h, afforded polycycle 50 in 83% yield.
Its connectivities were confirmed by X-ray analysis of
a related compound. The reaction itself proved robust,
3628
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 3625 –3630