3402
E. Convers et al. / Tetrahedron Letters 45 (2004) 3401–3404
S
S
O
S
N
N
N
N
N
H
BocHN
NHBoc
H
H
H
S
S
N
N
N
N
H
H
H
H
Figure 1.
rt using 1.2 equiv Mukaiyama reagent and 2.4 equiv
triethylamine. The expected carbodiimide was then
isolated in 91% yield.
NBoc
NHBoc
S
R
a
R
+
NH2
N
H
BocHN
NHBoc
Scheme 3. (a) PS-Mukaiyama reagent, DMF, Et3N, rt.
The conditions used for the first experiments to prepare
PS-Mukaiyama reagent were 2 equiv of 2-chloropyr-
idine, 2 equiv of potassium iodide in DMF, overnight at
60 °C in a rotating oven, with Merrifield’s resin at
3 mmol/g. In order to determine the best conditions for
this preparation, a design of experiments (DoE)
approach using the MODDE software from Umetrics
was employed.5 Experiments were designed, which
investigated the following parameters: equivalent of
2-chloropyridine, equivalent of KI, solvent (DMF, N,N-
DMA, NMP), temperature and reaction time. The ‘end-
use’ assay was then performed in order to determine the
loading of these batches of resin. The optimal conditions
for the PS-Mukaiyama reagent preparation were deter-
mined to be 2 equiv of 2-chloropyridine, 1 equiv of KI in
DMF at 40 °C for 24 h.6 These conditions gave a resin
loading of 1.33 mmol/g (72%). A lower loading sample
of Merrifield’s resin (1.5 mmol/g) was also used as
starting material in order to compare the loading of the
resulting resins. The ‘end-use’ assay performed on this
batch revealed a loading of 0.38 mmol/g (33%). It was
thus decided to use the Merrifield’s resin at 3 mmol/g.
A batch of 3 g of resin was then prepared under the
optimized conditions, overnight using the rotating oven.
The loading was then determined to be 1.02 mmol/g
(55%). As this loading was lower than expected, the
scale-up of this reaction was investigated further. It was
found that good agitation during the reaction was cru-
cial in order to get a good loading. A 10 g batch
was then prepared in a round-bottom flask using an
overhead stirrer, and the resin loading was determined
to be 1.2 mmol/g (65%).
with water, wash with a saturated bicarbonate solution
and use of carbonate resin) but in all cases a loss of
product mass was observed and the carbodiimide was
partially decomposed.
This new PS-Mukaiyama reagent was then investigated
as a reagent for the one-pot guanylation of amines with
di-Boc-thiourea. Such reactions have already been car-
ried out using Mukaiyama reagent in solution.3 It was
considered that the use of a PS-Mukaiyama reagent
would avoid any purification by chromatography as the
product would be isolated by simple filtration. At first
guanylations were accomplished with the same amines
(1a–d) as in the Lipton paper,3 using 1.2 equiv of resin at
rt overnight (Scheme 3, Table 1). Except for benzyl-
amine 1a, the conversions of these reactions were
modest. Increased amounts of resin and longer reaction
times did not improve these results. It was concluded
that PS-Mukaiyama reagent is not convenient for the
guanylations of secondary amines. A range of primary
amines (1f–j) was then investigated in order to determine
the scope of the method (Table 2). The reaction did not
work at all with ammonia 1j, and the conversion was
poor with very hindered amines such as amino-
diphenylmethane 1h. Reactions proceeded in good yield
with the other substrates.8 As a water wash was neces-
sary to remove the triethylamine salt, sometimes some
product was lost during the work-up. In order to avoid
any wash whatsoever, guanylation with benzylamine
was repeated with potassium carbonate instead of
triethylamine as a base.9 In that case, the crude product
was taken up in DCM and filtered through a plug
of silica. Clean product 2a was then isolated in 74%
yield.
Once the optimum conditions for the preparation of the
polymer-supported Mukaiyama reagent were estab-
lished, it was then decided to test this reagent for thio-
urea dehydration. During the ‘end-use’ assay,
dehydration of 1,3-di-(p-tolyl)thiourea with 1.2 equiv of
PS-Mukaiyama reagent was not complete. The reaction
was then repeated with 5 equiv of resin and the reaction
was complete after 16 h at rt. The crude material was
contaminated by a triethylamine salt: after a quick water
wash the clean carbodiimide was isolated in 84% yield.7
Dehydration of a range of thioureas was then investi-
gated (Fig. 1). Different work-up conditions were tried
in order to remove the triethylamine salt impurity (wash
3. Conclusion
In the course of our investigations we have successfully
prepared a polymer-supported Mukaiyama reagent with
a good loading. This preparation was thoroughly opti-
mized using a DoE approach and the procedure was
successfully applied to prepare the resin on a large scale
(10 g). This reagent was then successfully used for the