Synthesis of (+)-Dumetorine and Congeners
FULL PAPER
supported 2nd-generation Grubbs catalyst, prepared as de-
scribed by Buchmeiser.[22] The RCM reaction was then per-
formed by flowing the starting material through the pre-
packed column. Unfortunately poor conversion (less than
20%) was obtained even after 48 h possibly due to the limit-
ed surface contact between the solution and the catalyst
beads.
To overcome this problem, we considered the use of a
PEG-supported Ru catalyst. In this case, the RCM reaction
would benefit from the homogeneous conditions ensured by
the solubility of the catalyst in CH2Cl2, maintaining the pos-
sibility of simple catalyst recovery by solvent-promoted pre-
cipitation and filtration. Therefore, a newly synthesized
PEG-supported Hoveyda catalyst, an analogue of a PEG-
immobilised Ru catalyst described in the literature,[23] was
easily recovered by solvent evaporation in high purity and
in 95% yield. An efficient protocol for RCM under continu-
ous flow conditions was thus developed and we seek to
apply this methodology to further examples.
In the original batch synthesis, the two final steps were af-
fected by low yields. In particular, the acid-catalyzed cleav-
age of the tert-butoxycarbonyl (Boc) protecting group result-
ed in the isolation of a small amount of the desired product
14 (10% yield) along with the tricyclic byproduct 15 (60%
yield) from Michael addition of the secondary amine to the
a,b-unsaturated lactone ring (Scheme 4). Furthermore, the
inherent instability of 14 significantly reduced the yield of
the final reductive amination.
readily prepared from
a
salicylaldehyde derivative
(Scheme 3). The obtained catalyst 13 (loading 0.2 mmolgÀ1
Scheme 4. Deprotection and reductive amination in batch. a) TFA,
CH2Cl2, RT, 3 h, 10%; b) 14, CH2O, NaBH3CN, RT, 1.5 h, 35%. TFA=
trifluoroacetic acid.
We believed that the formation of the undesired 1,4 addi-
tion byproduct could be avoided if the N-deprotection and
N-methylation occurred simultaneously. In line with this
concept, the methylation of secondary amines by treatment
with formaldehyde in the presence of formic acid at high
temperature (Eschweiler–Clarke reaction)[25] appeared as an
optimal solution to block the formation of the undesired
1,4-addition byproduct. This reaction has, to the best of our
knowledge, never been reported under flow conditions and
the use of a protected amine constitutes an additional ele-
ment of novelty. The optimal reaction conditions were rap-
idly found by quickly screening the reaction solvent, resi-
dence times and temperature. The best results were ob-
tained when the solution containing 8, dissolved in acetoni-
trile, was combined with the stream of formaldehyde and
formic acid in acetonitrile in a pre-heated PTFE reactor.
The Boc removal and the reductive amination occurred at
1408C in 15 min giving complete conversion of the starting
material to (+)-dumetorine. Flash heating is known to im-
prove the overall yield and purity in many different chemi-
cal transformations and, also in our case, a rapid warming to
1408C gave a clean product. An additional advantage of the
flow protocol is the use of acetonitrile in place of DMSO,
Scheme 3. Synthesis of PEG-supported Hoveyda Ru catalyst: a) iso-
Propyl iodide, Cs2CO3, K2CO3, DMF, RT, 7 h, quantitative yield;
b) Ph3PCH3Br, KHDMS, toluene, À788C–RT, 3 h, 85%; c) LiAlH4, THF,
08C–RT, 1 h, 93%; d) succinic anhydride, DMAP, CH2Cl2, RT, 16 h,
60%; e) pivaloyl chloride, Et3N, Et2O, 0 8C–RT, 1 h, quantitative yield;
f) PEG 5000, trioctylamine, DMAP, CH2Cl2, RT, 21 h, 99%; g) Grubbs
2nd-generation catalyst, CuCl, CH2Cl2, 408C, 1 h, 99%. DMAP=4-dime-
thylaminopyridine; KHMDS=potassium hexamethyl disilazide.
1
determined by H NMR spectroscopy), stable at room tem-
perature for more than six months, showed excellent perfor-
mance in RCM reactions and ensured a simple and quanti-
tative catalyst recovery. Moreover, it was efficiently recycled
in a model reaction up to six times before a slight decrease
in reactivity was observed.[24] The streams containing 7 and
the catalyst (13, 6% moles Ru) dissolved in dry CH2Cl2
were pumped into a poly(tetrafluoroethylene) (PTFE) loop
reactor warmed at 708C (residence time 50 min). The reac-
tion output was collected into a vial containing an appropri-
ate volume of diethyl ether to quantitatively precipitate the
PEG-bound catalyst. The catalyst was filtered off with a
solid/liquid phase separator cartridge and the product 8 was
which is often employed as
a good solvent for the
Eschweiler–Clarke reaction especially when performed
under microwave conditions.[26] Finally, the flow stream con-
taining the newly formed 1 was directly purified in-line
through a column containing silica-supported sulfonic acid
(SCX). A brief washing sequence was used to elute any resi-
due prior to release of the product by passage of a metha-
nolic ammonia solution. After this ꢂcatch and releaseꢃ purifi-
cation, the solvent was finally evaporated in vacuum afford-
ing pure (+)-dumetorine in 92% yield (>98% purity by
Chem. Eur. J. 2011, 17, 6221 – 6226
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6223