The use of protecting groups undoubtedly permits the development of synthetic routes
toward desired organic scaffolds. Depending on the requirements in a synthesis, numerous
protecting groups for acidic functional groups can be used. The ideal protecting group
is complementary to the requisite conditions in the relevant synthetic steps, such
as a certain pH range or reagents, and it should not affect the overall desired reactivity.
In turn, the protection and deprotection steps should be selective for the particular
functionality. Owing to their ubiquitous nature and reactivity in target molecules
and intermediates, the presence of hydroxy groups often requires the employment of
protecting groups. To address the many needs, a wide range of acetal-, ether- and
ester-based protecting groups for hydroxy functionality have been developed.[1 ]
A particularly attractive O-protecting group is allyl, thanks to several factors such
as its ease of installation, use of inexpensive and readily available reagents, and
stability under various reaction conditions (Scheme [1 ]).[2 ] Beside the methods reported for the direct cleavage of allyl groups by using nucleophiles
or oxidants,[3 ] those using homogeneous metal catalysts stand out due to their mild reaction conditions
and often high selectivity. This has been impressively demonstrated in several total
syntheses of natural compounds.[4 ] These reactions usually follow two main mechanisms. The first is based on the formation
of a metal allyl intermediate by oxidative addition in the presence of a nucleophile
or reducing agent.[5 ] The underlying principle in other cases is an initial olefin isomerization (single-bond
transposition) to an enol ether, followed by hydrolysis.[6 ] Although the isomerization of allylic alcohols, ethers, and related substrates has
been studied extensively,[7 ] its application or suggested involvement in the deallylation is much less well investigated.
Furthermore, the use of nonprecious metals in allyl deprotection methods is rare but
clearly desirable.
Scheme 1 General principle and deallylation methodologies
An important feature in deprotection chemistry is the concept of orthogonality, based
on the selective removal of a protecting group through differential reactivity and
stability.[8 ] Recently, Yamada and co-workers reported an elegant method for orthogonal oxidative
deprotection of p -methylbenzyl ethers in the presence of a p -methoxybenzyl group, and vice-versa.[9 ] In this context, the chemoselective removal of an allyl group in the presence of
a benzyl protecting group is of interest.
During our investigations of nickel hydride/ Brønsted acid catalyzed tandem reactions
of allyl ethers for the generation of oxacyclic scaffolds, we observed an unexpected
loss of an allyl group under the reaction conditions in some cases.[10 ] A plausible explanation was an initial isomerization to form an alkenyl ether, as
later proven by independent experiments, and subsequent acidic hydrolysis owing to
the presence of water in the acid.
Having obtained this result and based on our interest in isomerization reactions[11 ] and nickel catalysis,[12 ] we envisaged developing a general catalytic process for O-deallylation of ethers.
To investigate this further, we chose the O-allylated phenol 1a as a substrate and the complex nickel hydride [Ni(PMe3 )4 0]N(SO2 CF3 )2 from our original study as a precatalyst, due to its structural simplicity (Table
[1 ]). The precatalyst can by synthesized in two steps from bis(cycloocta-1,5-dienyl)nickel(0),
the appropriate ligand, and bistriflimidic acid on a gram scale, and can be stored
under argon on the bench.[13 ] In line with our previous investigations, partial isomerization to the corresponding
enol ether occurred. This was followed by subsequent addition of a Brønsted acid in
an attempt to obtain 2-methoxyphenol (2a ). Initial experiments using catalytic amounts of diphenylphosphoric and triflic acid
gave only a mixture of the starting material and its isomer (Table [1 ], entries 1 and 2). Increasing the amount of precatalyst and the use of a weaker
acid gave only traces of 2a (entries 3 and 4). The use of triflic acid (10 mol%) with a prolonged reaction time
led to formation of 2a in a moderate yield of 38% (entry 5).
Table 1 Optimization of the Reaction Conditions for the Ni-Catalyzed Deallylationa
Entry
Ni-H (mol%)
Acid (mol%)
Time (h)
Yieldb (%) of 2a
1
1
(PhO)2 P(O)OH (8)
1
–c
2
1
MeSO3 H (8)
1
–c
3
5
(PhO)2 P(O)OH (120)
0.5
traces
4
5
(PhO)2 P(O)OH (200)
0.5
traces
5d
2
F3 CSO3 H (10)
15
38
6
1
TsOH·H2 O (100)
0.5
73 (79)e
7
–
TsOH·H2 O (100)
1
-f
8
0.5
TsOH·H2 O (100)
1
84
9
0.5
TsOH·H2 O (50)
15
75
10
1
TsOH·H2 O (1)
1
traces
11g
0.5
TsOH·H2 O (10)
5
traces
12h
0.5
TsOH·H2 O (100)
1
(1:1)i
13
1
CSA (100)
5
(1:4)i
a Reaction conditions: 1a (0.25 mmol), [Ni(PMe3 )4 H]N(SO2 CF3 )2 , Brønsted acid, THF (0.16 M), 30 min, RT with Ni-H, then 60 °C for the indicated
time with acid.
b Isolated yield.
c Mixture of olefin isomers (unreacted 1a along with enol ethers).
d 30 min at RT with Ni-H then 15 h at RT with F3 CSO3 H.
e Yield at 5 mmol scale.
f
1a was recovered.
g With H2 O (1 equiv).
h All reaction components were added at the start.
i Ratio 1a /2a determined by 1 H NMR analysis.
Upon screening of additional Brønsted acids, we found that a stoichiometric amount
of p -toluenesulfonic acid monohydrate (TsOH·H2 O) gave complete conversion into the phenol (isolated yield 73%) within 30 minutes
(Table [1 ], entry 6). TsOH·H2 O is an attractive option owing to its bench stability, its ease of handling, and
its straightforward removal through aqueous workup, after which the product 2a was determined to be spectroscopically pure without the need for column chromatographic
purification. This reaction was also demonstrated at 5 mmol scale (entry 6). A control
experiment with TsOH·H2 O in the absence of the catalyst led to complete recovery of unreacted 1a (entry 7). The precatalyst loading could be decreased to 0.5 mol% without any loss
in activity (entry 8). Unfortunately, longer reaction times were deemed necessary
to achieve complete conversion when 0.5 equiv of TsOH.H2 O were used (entry 9), and only traces of product were detected when using 1 mol%
of acid (entry 10). In addition, we examined whether a catalytic amount of acid could
be used in the presence of water to promote the hydrolysis, but we found this not
to be the case (entry 11). Addition of all components at the onset of the reaction
resulted in 1:1 mixture of 1a and 2a (entry 12). Furthermore, the use of slightly weaker camphorsulfonic acid (CSA) required
longer reaction times for complete conversion (entry 13).
Therefore, a system consisting of 1 mol% of Ni-H and 1 equivalent of TsOH·H2 O was chosen for further study of the transformation. With an operationally simple
and straightforward protocol in hand, we proceeded to evaluate its generality through
deprotection of a variety of substituted O-allylated compounds, starting with aryl
allyl ethers 1b –p (Scheme [2 ]). First, the influence of substituents in the 3- and 4-positions of the phenyl ring
was evaluated. Compounds containing electron-donating groups performed much better
under the catalytic conditions than those with electron-withdrawing substituents.
Ethers 1c , 1g , 1h , 1l , and 1n were smoothly deallylated to the corresponding phenols in yields of 68–95%. Note
that no column chromatographic purification was necessary to obtain phenols 2c and 2g in adequate purity (>95%). The boronate ester 1d was converted in a satisfactory 73% yield, albeit with a longer reaction time of
20 hours. The 4-trifluoromethylsulfanyl ether 1f underwent deallylation to give 2f in 34% yield. Upon testing the halide-substituted ethers 1b , 1m , and 1o , we obtained the corresponding phenols in low to moderate yields of 56, 18, and 37%,
respectively. A competing dehalogenation was not observed; low yields are therefore
attributable to lower reactivity in both steps. Note that the isomerization to an
enol ether was more closely investigated in our previous work; although it was never
complete (usually 50%), it did not impact the overall yield if both reactions were
successful.
Scheme 2 Substrate scope of Ni-catalyzed deprotection of aryl allyl ethers. Reaction time
for the second step: a 6 h, b 15 h, c 10 h, d 22 h [camphersulfonic acid (1 equiv)], e 15 h [TsOH·H2 O (2 equiv)].
Although the aldehyde 1e displayed poor reactivity and gave 2e in only 17% yield, its protected derivative 1i underwent deallylation to give a 52% yield of 2i . The allyl group could be successfully removed from the monoprotected hydroquinone
1j . A selective allyl deprotection was observed in the presence of a tert -butyl(diphenyl)silyl group by using a weaker acid (CSA); however, the reaction was
slow, and only a 33% yield of 2k was obtained.
In the presence of an oxazole motif, deallylation to the phenol 2p occurred in 62% yield; an additional equivalent of TsOH⋅H2 O was used in this case. In contrast to other substrates with electron-withdrawing
groups, the 2-acetyl-substituted ether 1q was converted into the product in good yield (87%). Moreover, it was found that a
sulfone functionality, which is often found in biologically active compounds, can
be present in the molecule, as 2r was obtained in 41% yield.
Primary, secondary and tertiary alkyl allyl ethers 3 also underwent facile C–O bond cleavage (Scheme [3 ]). The homobenzylic alcohol 4a , citronellol (4b ), (–)-menthol (4c ), and α-terpenol (4d ) were obtained from the corresponding allyl ethers in appreciable yields of 82–95%.
Note that the allyl protecting group was removed selectively in the presence of a
benzyl group in the case of 4e . Moreover, the protected cholesterol derivative 3f gave cholesterol (4f ) without any racemization in 84% yield. The internal double bonds in substrates 3b , 4d , and 4f did not isomerize during the transformation. It is noteworthy that, apart from 4e , all aliphatic compounds were obtained in high yields without the need for chromatographic
purification.
Scheme 3 Deallylation of alkyl allyl ethers: substrate scope
In addition, we also examined the feasibility of deallylation of amides, which has
received less attention in the literature[14 ] despite its potential in, for example, the synthesis of immobilized DNA oligomers,
where allyl groups are used to protect the nucleotide bases.[15 ] Employing our O -deallylation method, the deprotection of N -allylated benzamide 5 was possible and gave the corresponding product 6 in 48% yield (Scheme [4 ]). To the best of our knowledge, there are no previous reports on the deallylation
of amides by employing a nickel catalyst.
Scheme 4 Deprotection of amide 5
In summary, we have developed a rapid and straightforward protocol for a nickel hydride
catalyzed, Brønsted acid promoted deprotection of allyl ethers.[16 ] The transformation occurs through an initial nickel-catalyzed isomerization of the
terminal double bond to the enol ether, which, in turn, is hydrolytically cleaved
by the Brønsted acid. The deprotection of the allyl ether moiety occurs chemoselectively
in the presence of internal double bonds and other protecting groups, such as the
widely used benzyl group. The low catalyst loading along with the generality of the
substrate scope, point to the robustness of the protocol, which is suitable for both
aryl and alkyl ethers. Moreover, a deprotection of an N -allylamide has been achieved.
