Tandem radical and non-radical reactions mediated with thiols - A new method of cleavage of allylic amines

Article abstract of DOI:10.1039/b109781d

Thiyl radical promotes the isomerisation of allylic amines into enamines via two consecutive hydrogen atom abstraction steps, and the subsequent polar addition of the corresponding thiol to the enamine results in the cleavage of the C-N bond via a thioaminal intermediate: this reaction provides a mild, metal-free methodology for the deprotection of allylated primary and secondary amines.

Full text of DOI:10.1039/b109781d

Tandem radical and non-radical reactions mediated with thiols—a new
method of cleavage of allylic amines
Michèle P. Bertrand,a Stéphanie Escoubet,a Stéphane Gastaldia and Vitaliy I. Timokhinb
a Laboratoire de Chimie Moléculaire Organique – UMR 6517, Boite 562 – Faculté des Sciences St Jérˆome,
Université d’Aix-Marseille III, Av. Escadrille Normandie–Niemen 13397. Marseille Cedex 20, France
b Department of Physical Chemistry, Institute of Physical Chemistry, National Academy of Sciences of
Ukraine, 3A Naukova Street 79053. Lviv, Ukraine
Received (in Cambridge, UK) 26th October 2001, Accepted 3rd December 2001
First published as an Advance Article on the web 8th January 2002
Thiyl radical promotes the isomerisation of allylic amines
into enamines via two consecutive hydrogen atom abstrac-
tion steps, and the subsequent polar addition of the
corresponding thiol to the enamine results in the cleavage of
the C–N bond via a thioaminal intermediate: this reaction
provides a mild, metal-free methodology for the deprotec-
tion of allylated primary and secondary amines.
radical abstracts the allylic hydrogen atom which leads to 5.
According to thermodynamical data, this reaction should be
nearly thermoneutral, or slightly endothermic [BDE (TolS–H)
= 80–83 kcal mol21;2 BDE (C–H) = 82.6 kcal mol21 3]. The
delocalised radical transfers back a hydrogen atom from
thiocresol to give the enamine 6.4 Both reactions a and c are
formally reversible. We believe that the equilibrium is displaced
owing to the ready addition of the thiol to the enamine.5
It is to be noted that the hydrogen atom transfer between the
electron rich C–H bond in 4 and the electrophilic thiyl radical is
favoured on the grounds of polar effects. The same assumption
applies to step c where the hydrogen atom is transferred from
the thiol—acting here as a 8 protic 9 hydrogen atom donor,
according to Roberts6—to the nucleophilic carbon centered
radical.
We have tried to get an experimental probe for the formation
of the enamine, from the NMR spectrum of the crude mixture,
starting from an aniline derivative which should lead to a less
reactive enamine. When the reaction was performed on N-
methyl N-prenyl aniline, the reaction led to two products, the
unchanged substrate and the enamine in a 73+27 ratio. In this
case, the enamine was not basic enough to react with thiocresol,
only a trace amount of the cleavage product was detected. The
presence of the enamine in the crude reaction mixture was
characterised by the signals corresponding to the vinylic protons
b- to the nitrogen atom in the two isomers (5.80 ppm and 5.30
ppm respectively).
Our interest in sulfur-centered radical mediated cyclisations of
1,6-dienes1 led us to investigate the addition of thiocresol to
dienes 1a and 1b. These reactions reported in Scheme 1 led to
an unexpected result: the only products, isolated in 70 and 65%
respectively, were amines 2a and 2b resulting from the cleavage
of the primary allylic C–N bond. No trace of any cyclisation
product was detected. It is noteworthy that the amine that would
result from the cleavage of the secondary allylic C–N bond was
not formed either. Control experiments have been carried out.
When 1a was heated alone, or with thiocresol, but in the absence
of AIBN, nothing happened. When AIBN was added to the
mixture, the deprenylation occurred. Heating the substrate with
AIBN in the absence of thiol did not give rise to any
transformation either. These results point strongly to a radical
chain mechanism involving thiyl radicals. The reaction is likely
to proceed, via the migration of the double bond, through
enamines 3 that would be hydrolysed upon treatment.
Although the very first experiments were conducted with a
syringe pump monitored addition of thiol and AIBN to the
substrate, the same results could be obtained mixing all the
reagents together at once. In order to get a more accurate picture
of both the mechanism and the scope and limits of the reaction,
other substrates were submitted to the following experimental
conditions: refluxing a 0.06 M solution of the allylic amine,
with 1.2 equivalent of thiocresol and AIBN (20 mol % in two
portions), over 4 h. Starting from either 4 or 8 led cleanly to the
formation of thioaminal 7 or 9 respectively, identified in each
1
case as the single product from the H NMR spectrum of the
crude mixture [Ha signal appears as a dd at 4.34 ppm (J = 8.9
and 5.9 Hz) in 7, and as a pseudo t at 4.26 ppm (J = 7.5 Hz) in
9]; the chemical shift of the corresponding carbon is 76.6 ppm
in 7 and 82.0 ppm in 9) (Scheme 2).
The migration of the double bond can be explained on the
grounds of the mechanism proposed in Scheme 2. The thiyl
Scheme 1
Scheme 2
216
CHEM. COMMUN., 2002, 216–217
This journal is © The Royal Society of Chemistry 2002

According to our rationalisation, the more acidic the thiol, the
more efficient the addition of the thiol to the enamine and, as a
consequence, the faster the cleavage of the C–N allylic bond.
The comparison of reactions conducted, for exactly 4 h, on
amine 10 (Fig. 1) with thiocresol, thioglycolic acid methyl ester,
and octanethiol respectively, is in agreement with this proposal.
Only the first reaction involving the most acidic thiol was
completed. The conversion to dicyclohexylamine was 100% in
the first case, 78% in the second case, and 57% in the third
one.
amine by tosyl chloride).7 When the benzylated amine 14b was
submitted to the same experimental conditions, the reaction led
to a single diastereoisomer in 60% yield. In this case, the yield
was slightly higher and the reaction mixture was very clean, no
other product could be detected. This might be explained by the
fact that the stability of the a-aminoalkyl radical increases with
the substitution at the nitrogen atom,8 this should accelerate the
abstraction of the allylic hydrogen atom compared to the
competitive abstraction of the benzylic hydrogen atom. In the
case of 15a, as with 14a, the yield in 16 was moderate (44%),
but again no epimerised product was isolated nor identified.
This is a totally surprising result since, under such conditions
one would have expected a capto-dative position to be
epimerised [BDE (capto-dative C–H) = 82–83 kcal mol21 9].
As regards to the limits, it must be noted that the cleavage
does not take place when the nitrogen atom bears an electron
withdrawing group like Ts or Boc.
These results are closely related to those reported by Roberts
on the isomerisation of allylsilyl ethers to silylenol ethers.10
Under our experimental conditions allyl- and prenyl alkyl ethers
remained unchanged. In conclusion, the abstraction of the
allylic hydrogen atom by thiyl radical promotes the cleavage of
allylic amines. The reaction is chemoselective, a primary C–N
allylic bond can be cleaved selectively in the presence of a
secondary one.11 This reaction, conducted under relatively mild
conditions, complements the methodologies already available
in the literature which mainly consist in heavy metal catalysed
isomerisations.12
Fig. 1
This reaction is general and the cleavage occurred, with
moderate to good yields, with different allylic groups like
crotyl, allyl or cinnamyl (Table 1). The phototochemical
initiation worked also, but gave a lower yield in our hands.
Table 1 Cleavage of amines 12–15 with TolSHa
Isolated yield %
Notes and references
Substrate
D
hn
1 M. P. Bertrand, S. Gastaldi and R. Nouguier, Tetrahedron Lett., 1996,
37, 1229; M. P. Bertrand, S. Gastaldi and R. Nouguier, Tetrahedron,
1998, 54, 12829.
2 E. T. Denisov, Russ. J. Phys. Chem. (Engl. Transl.), 1996, 70, 238; D.
F. Mc Millen and D. M. Golden, Ann. Rev. Phys. Chem., 1982, 33,
493.
10
11
12
13
97
63
95
70
64
a Conditions: 0.06 M solution of substrate in PhH; ToISH (1.2 equiv.);
AIBN (20 mol% in 2 portions); reflux
3 G. W. Dombrowski, J. P. Dinnocenzo, S. Farid, J. L. Goodman and I. R.
Gould, J. Org. Chem., 1996, 64, 427.
4 Examples of intramolecular H-abstraction competing with the cyclisa-
tion of thiyl radicals have been reported by Surzur and co-workers, see:
M. Kaafarani, M. P. Crozet and J.-M. Surzur, Bull. Soc. Chim. Fr., 1987,
885.
5 S.-O. Lawesson, E. H. Larsen and H. J. Jakobsen, Recl. Trav. Chim.
Pays-Bas, 1964, 83, 461.
Secondary allylic amines behaved similarly, as shown in
Scheme 3. Starting from 14a, the yield was identical (50%),
whatever the initiation (thermal or photochemical), and what-
ever the work up (acidific work up, or trapping of the primary
6 B. P. Roberts, Chem. Soc. Rev., 1999, 28, 25.
7 Amazingly, no epimerisation at the benzylic position was observed. The
analysis of the 1H NMR spectrum of the crude mixture and the moderate
yield suggest that once formed, the benzylic radical might undergo a
fragmentation leading to unidentified products, instead of being
reduced.
8 P. Renaud and L. Giraud, Synthesis, 1996, 913.
9 A. Rauk, D. Yu and D. A. Armstrong, J. Am. Chem. Soc., 1998, 120,
8848.
10 A. J. Fielding and B. P. Roberts, Tetrahedron Lett., 2001, 42, 4061.
11 This might be due to the minimisation of the allylic strain leading to a
conformation around the secondary C–N bond unfavourable to the
hydrogen abstraction, as in the case of allylic ethers (cf. ref. 10).
12 T. W. Greene and G. M. Wuts, in Protective Groups in Organic
Synthesis (3rd Ed.), Wiley, New York, 1999, pp. 574–576.
Scheme 3
CHEM. COMMUN., 2002, 216–217
217