N. Rabasso, A. Fadel / Tetrahedron Letters 55 (2014) 6068–6071
6069
Table 1
O
S
O
H3C
H3C
CO2R1
Preparation of nitrocyclopropylphosphonate 6a from dibromophosphonate 4a
O
P
P(OR2)2
Tolp
CO2R1
N
N
Cl
O
P
O
P
Hanessian
path a
Midura
path b
H
OEt
OEt
OEt
OEt
S
K2CO3 (3 equiv)
H
+
Br
O
O
P
DMSO-THF (5:1)
82%
O2N
5a
OH
OH
R
O2N
Br
4
( )-6a
O
H2N
O
O
N
1
O
P(OR2)2
P(OR2)2
Entry
Time (h)
[c] mol/L
6a (yield %)
Br
Couthon
path d
CO2R1
(R2O)2P
O
CN
1
2
3
4
5
24
24
20
24
5
1.00
0.20
0.14
0.10
0.14
30
Charette
path c
73
this
N2
R
NO2
81b
81
work
O
82c
P(OR2)2
Br
a
Reaction conditions: dibromo 4, methane 5a (1.2 equiv) K2CO3 (3 equiv),
Br
DMSO–THF (5:1), rt.
b
Isolated yield for reaction in DMSO.
Reaction carried out under sonication.
c
Figure 2. Synthetic routes toward b-aminocyclopropylphosphonic acids.
O
P
O
P
O
P
5a
K2CO3 (3 equiv)
O
H
NO2
O
P
OH
OH
OEt
OEt
OEt
OEt
OEt
OEt
OEt
OEt
H
K2CO3 (2 equiv)
H
10 h, 35°C
DMSO-THF
93%
P
5a
4
Br
DMSO-THF
H2N
O2N
Br
O2N
Br
90%
7
trans-( )-6a
4
6a
3a
Scheme 2. Formation of 6a from bromovinylphosphonate 7.
Scheme 1. Synthetic sequence of b-aminocyclopropylphosphonic acid.
Continuing our efforts in this field we extended our interest to
the preparation of heterocyclic
O
P
O
P
1) TMSI (3 equiv)
CH2Cl2, 3 h, rt
a
-aminophosphonic acids19–22
OH
OH
3
(
OEt
OEt
H
H
and their incorporation into peptides.23 In the present work, we
now turn our attention to the synthesis of non-substituted b-
aminophosphonic acid 1, which to date remains unknown. Exten-
sion to b-alkylated counterparts is also reported. Our approach for
the synthesis of the racemic ACPP 3a is depicted in Scheme 1. The
ACPP 3a would be prepared from the nitrophosphonate 6a through
reduction and hydrolysis. The latter would be obtained from the
cyclopropanation24 of the (1,2-dibromoethyl)phosphonate 4 with
nitromethane 5a (Scheme 1).
2
1
2) EtOH,
O
O2N
O2N
)-6a
(
)-8a
88%
Scheme 3. Synthesis of nitrocyclopropylphosphonic acid 8a.
observed. Indeed, treatment of dibromophosphonate 4, with
K2CO3 in DMSO–THF in the absence of nitromethane, furnished
the bromovinylphosphonate 7 in excellent yield (93%). Then, the
latter underwent an attack by nitromethane 5a through
a
sequence involving a Michael addition, followed by an intramolec-
ular SN2 alkylation to form the expected aminophosphonate
6a in 90% yield (Scheme 2). As a consequence, we postulate that
the formation of the vinylphosphonate 7 precedes the possible
competitive nucleophilic reaction of nitromethane on the dib-
romophosphonate 4. Then the Michael addition of nitromethane
to the vinylphosphonate 7 gives the intermediate 1-bromo-3-
nitropropylphosphonate, which is not observed. The latter
Results and discussion
The (1,2-dibromoethyl)phosphonate 4 was easily prepared
from the bromination of the commercially available diethyl vinyl-
phosphonate according to the procedure reported by Sainz-Diaz
and co-workers.25
The one-pot double alkylation was performed at room temper-
ature by a slow addition of nitromethane 5a to a stirred solution
of dibromophosphonate 4 and 3 equiv of potassium carbonate in
a mixture of DMSO and THF (5:1), to give the desired nitro-
cyclopropylphosphonate 6a in good yields (Table 1, entries 1–4).
We observed that when the concentration of the reaction med-
ium was ranging from 0.14 M to 0.10 M, the yield of the reaction
was deeply increased (compare entries 3 and 4 to entry 1). In
addition under the same conditions, the use of sonication
decreased the reaction time from 24 h to 5 h (entry 5, 82% yield).
By comparison, similar double alkylation was already reported,
under basic conditions starting from 1,2-dibromoethylcarbzoxy-
late, yielding nitrocyclopropylcarboxylate in only 59% after
48 h.26 On the other hand, in order to facilitate the work-up pro-
cedure for this reaction, we tried to lower the amount of DMSO
used, by introducing a co-solvent. We found that in DMSO or in
a mixture of DMSO/THF (5:1) the reaction occurred in comparable
yields. However, in THF only the vinylphosphonate intermediate
7 was formed.
undergoes
a fast intramolecular cyclization to provide the
nitrocyclopropylphosphonate 6a.
Next, the hydrolysis of phosphonates 6a was accomplished by
treatment with 3 equiv of a freshly opened bottle of trimethylsilyl
iodide in dichloromethane at room temperature,16 followed by the
addition of propylene oxide in ethanol to provide the b-nitrocyclo-
propylphosphonic acid 8a in 88% (Scheme 3). It is noteworthy that
lowering the amount of TMSI for the reaction, or using instead
TMSBr, resulted in the formation of about 10–30% of the monopho-
sphonate intermediate.
Next, we investigated the reduction of the nitro group to the free
amine. Precedents in the literature highlighted, during reduction of
nitrocyclopropylcarboxylate, the ability of the resulting aminocy-
clopropylcarboxylate to undergo a ring opening, due to the push–
pull effect, leading to c
-aminobutyric acid (GABA).27–29 Taking into
account this phenomena, several mild methods for the reduction of
the nitro group to the free amine were investigated. Known
reagents such as Zn/AcOH, Zn/HClaq,
30 or NiCl2/NaBH4 did not give
It is noteworthy that in this reaction the formation of a small
amount of diethyl 1-bromovinylphosphonate 7 resulting from
the dehydrobromination of dibromophosphonate 4 was always
satisfactory results: no reaction, many TLC spots, or 5% yield of
undesired product 10a were obtained, respectively. Moreover,