4506
J. Luo et al. / Tetrahedron Letters 54 (2013) 4505–4508
Table 1
O
O
Model aromatization with piperidinea
OH
a,b
c
O
HO
OH
O
O
HO
OH
N
+
HN
B
3
1
2
Ph
RO
OR
2 R2 = PhB
d
7a
TBSO
O
O
3
4
R = H
OH
R = TBS
e
COOH
70%
N
b
Entry
Ketone
Desiccant
Solvent
t
t
(h)
Yield
(%)
TBSO
OTBS TBSO
9%
OTBS
(°C)
6
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
2
2
2
2
2
2
2
2
2
2
3
3
4
4
None
CaCl2
DCM
DCM
DCM
DCM
DCM
PhH
Reflux
rt
rt
rt
rt
rt
Reflux
rt
rt
rt
Reflux
rt
rt
8
8
8
8
6
6
2
6
6
76
81
82
78
99
99
99
99
99
96
12
15
Trace
35
7g
(1) Bayere-Villiger oxidation
(2) H+
MgSO4
Na2SO4
4A MS
4A MS
4A MS
4A MS
4A MS
4A MS
4A MS
4A MS
4A MS
4A MS
O
O
statins
PhH
DMF
HO
O
8
DMSO
MeCN
DCM
DMSO
DMSO
DMSO
6
Scheme 1. Unexpected aromatization of 4 observed in the attempt to prepare
chiral lactone moiety of statins. Reagents and conditions: (a) phenylboronic acid,
toluene, reflux, 5 h, >99%; (b) PCC@Al2O3, NaOAc, 4A molecule sieve, CH2Cl2, rt, 5 h,
85%; (c) pinacol, THF, BF3ÁEt2O, rt, 24 h, >99%; (d) TBSCl, imidazole, DMF, RT, 30 min,
24
24
24
48
80
99%; (e) 4-tert-butyldimethylsiloxy-L-proline (5), nitrosobenzene, DMSO, RT, 2 d.
a
Reaction conditions: ketone (0.5 mmol), piperidine (0.6 mmol), solvent (8 mL),
4A MS (4 g) or other desiccants (1 g), rt. DCM = dichloromethane, PhH = benzene.
b
Isolated yields.
O
TBSO
NO
COOH
TBSO
N
TBSO
COOH
OTBS
aromatization, a variety of secondary amines were applied and
the results are listed in Table 2. It can be seen that strong nucleo-
philes like piperidine, morpholine, diethylamine, and dipropyl-
amine gave corresponding products 7a–d with excellent yields
(Table 2, entries 1–4). However, steric hindrance affects the reac-
tion dramatically. Compared to dipropylamine, dibutylamine
showed lower reactivity (Table 2, entry 5). When diisopropylamine
was used, only trace product 7k was detected even reacting under
reflux for 24 h (Table 2, entry 11). Compound 5 reacted faster with
2 (Table 2, entry 7) than 4 (Scheme 2) and gave 7g with higher
yield but lower chemoselectivity. The reason might be that the
acidic carboxyl facilitates the conversion but, at the same time,
accelerates the dehydration of 2 to yield 5-hydroxycyclohex-2-en-
one, which was detected in the final reaction mixture. To avoid the
4
no reaction
N
H
DMSO
DMSO, RT, 9 d
7g
5
RT, 2 d
90% yield
60% conversion
Scheme 2. Aromatization of compound 4 with 5.
These results directed us to explore a new aromatization reac-
tion to prepare anilines under mild conditions in the absence of
any metal catalyst or strong base. Since 4 is relatively stable in
comparison with its two precursors 3-phenyl-2,4-dioxa-3-bora-
bicyclo[3.3.1]nonan-7-one (2) and 3,5-dihydroxylcyclohexanone
(3), we believed that the aromatization might be realized more
efficiently by using more active 3,5-dihydroxylcyclohexanone
derivatives.
dehydration side reaction,
L
-proline methyl ester was used and the
-proline though the
result shows that its selectivity is higher than
L
reactivity is lower (Table 2, entries 8 and 9). N-methylaniline
exhibited reduced reactivity and afforded N-methyldiphenylamine
7j in moderate yield (Table 2, entry 10). Furthermore, diphenyl-
amine shows no reactivity in this reaction (Table 2, entry 12).
Compared to secondary amines, primary amines usually have
both less steric hindrance and lower nucleophilicity, but the two
factors have opposite influence on nucleophilic reactions. So a vari-
ety of primary amines were then explored. The reaction was found
to be very smooth and afforded corresponding monophenylated
amines with excellent yields (Table 3). Compared to piperidine (Ta-
ble 2, entry 1), cyclohexanamine reacted obviously faster and affor-
ded the desired product N-cyclohexylaniline (9a) in quantitative
yield (Table 3, entry 1). So we can figure out that steric hindrance
is the dominant influence factor on this reaction. This conclusion is
also supported by the following facts. Butylamine reacted faster
than dibutyl amine at the same reaction conditions, and the yield
of N-butylaniline (9b, 99%, Table 3, entry 2) was much higher than
that of N,N-dibutyl aniline (7e, 78%, Table 2, entry 5). The ratio of
compound 2 to butylamine was elevated to 3:1, but the yield
had almost no change though the reaction was a little faster. Ben-
zylamine and furfurylamine also reacted smoothly with 2 and
afforded N-benzylaniline (9c) and N-(furan-2-ylmethyl)aniline
Piperidine was used for the model aromatization to optimize
the reaction conditions. The results listed in Table 1 indicate that
compound 2 has the highest reactivity. When no desiccant was
used, the yield was 76% (Table 1, entry 1). Some commonly used
inorganic desiccants such as calcium chloride (Table 1, entry 2),
magnesium sulfate (Table 1, entry 3), and sodium sulfate (Table 1,
entry 4) were applied to absorb the water formed in the reaction,
but the yields were only promoted slightly. However, when 4A
molecular sieves were used, the yields were achieved almost quan-
titatively. Both polar and nonpolar solvents gave similar results
(Table 1, entries 5–10). 3,5-Dihydroxylcyclohexanone (3) exhibits
a much lower reactivity, only 12% yield was obtained even reacting
under reflux for 24 h in dichloromethane (Table 1, entry 11). The
yield was still too low when the reaction was performed in a polar
solvent DMSO (Table 1, entry 12). The reaction were extremely
slow when 3,5-di(tert-butyldimethylsiloxy)cyclohexanone (4)
were used as the source of the phenyl group (Table 1, entries 13
and 14). The reason might be assigned to the high stability of silox-
ane in neutral and basic environments.
Based on the results achieved above, compound 2 was used for
the following experiments.19 To investigate the scope of this