A facile synthesis of N-[2-(trifluoromethyl)allyl]amides and their transformation into angularly trifluoromethylated bicyclic cyclopentenones

Article abstract of DOI:10.1246/cl.2007.22

On treatment with sec-BuLi at -105°C, 2-bromo-3,3,3-trifluoropropene undergoes rapid lithium-halogen exchange to generate thermally unstable 1-(trifluoromethyl)vinyllithium, which reacts with W-tosylimines to afford N-[2-(trifluoromethyl)allyl] amides in high yield. Propargylation of the amides, followed by the Pauson-Khand reaction, readily provides pyrrolidine ring-fused cyclopentenones with an angular trifluoromethyl group. Copyright

Full text of DOI:10.1246/cl.2007.22

22
Chemistry Letters Vol.36, No.1 (2007)
A Facile Synthesis of N-[2-(Trifluoromethyl)allyl]amides and Their Transformation
into Angularly Trifluoromethylated Bicyclic Cyclopentenones
Ryo Nadano and Junji Ichikawaꢀ
Department of Chemistry, Graduate School of Science, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033
(Received September 21, 2006; CL-061101; E-mail: Junji@chem.s.u-tokyo.ac.jp)
On treatment with sec-BuLi at ꢁ105 ꢂC, 2-bromo-3,3,3-
Table 1. Preparation of 1-(trifluoromethyl)vinyllithium 1
trifluoropropene undergoes rapid lithium–halogen exchange to
generate thermally unstable 1-(trifluoromethyl)vinyllithium,
which reacts with N-tosylimines to afford N-[2-(trifluoro-
methyl)allyl]amides in high yield. Propargylation of the amides,
followed by the Pauson–Khand reaction, readily provides
pyrrolidine ring-fused cyclopentenones with an angular tri-
fluoromethyl group.
CF3
CF3
MeOH
RLi (1.0 ma)
15 min/Et2O
CF2
•
+
Br
2
3
4
ma: molar amount
Entry
R
temp./ꢂC
3
4
recovery of 2
1
2
3
4
5
n-Bu
ꢁ78
ꢁ96
20% 54%
31% 11%
17%
58%
76%
10%
5%
ꢁ105
ꢁ105
ꢁ105
14%
1%
sec-Bu
tert-Bu
60% 26%
50% 35%
2-Trifluoromethyl-1-alkenes [1-(trifluoromethyl)vinyl com-
pounds] constitute a versatile class of building blocks1 for the se-
lective introduction of fluorine-containing carbon substituents
into bioactive molecules and molecular devices.2,3 This is due
to their electron-withdrawing CF3 groups, reactive double bonds
toward nucleophiles, and allylic fluorine atoms as potential leav-
ing groups. Along this line, we have recently developed flexible
synthetic routes to: (i) 1,1-difluoro-1-alkenes via an SN20-type
reaction;4 (ii) fluorocarbon-substituted heterocycles via intramo-
lecular nucleophilic reactions;5 and (iii) 5-trifluoromethyl-2-cy-
clopentenones via a regioselective Nazarov cyclization.6 Despite
the synthetic potential of the 1-(trifluoromethyl)vinyl moiety, its
introduction as a C3 unit remains a difficult task, because of the
thermal instability of the corresponding reactive metal species,
such as a vinyllithium reagent.7–9
In our previous paper, we reported the efficient synthesis of
[1-(trifluoromethyl)vinyl]-substituted alcohols by the ring open-
ing of cyclic ethers with 3,3,3-trifluoroprop-1-en-2-yllithium
(1).8 On treatment of 2-bromo-3,3,3-trifluoroprop-1-ene (2) with
an equimolar amount of n-BuLi at ꢁ100 ꢂC, slow lithium–
halogen exchange gave rise to a mixture of vinyllithium 1 and
n-BuLi. We succeeded, however, in the selective trapping of 1
with appropriate electrophiles, such as oxiranes and oxetanes,
yield of 3,3,3-trifluoroprop-1-ene (3), along with a 76% recovery
of 2 (Entry 3). These results indicate that incomplete conversion
is inevitable in the reaction with n-BuLi, due to the slow ex-
change rate and the thermal instability of 1. In contrast, the lithi-
um–halogen exchange reaction with sec-BuLi proceeded rapidly
even at ꢁ105 ꢂC to consume 90% of 2, which generated 1 in at
least 60% yield along with 26% of difluoroallene 4 (Entry 4).
We then attempted the reaction of aromatic and aliphatic
aldehydes with vinyllithium 1, generated in situ from 2.0 molar
amounts of vinyl bromide 2 and sec-BuLi, in consideration
of the partial decomposition of 1. The desired 2-(trifluoro-
methyl)allyl alcohols 5a and 5b were obtained in 73 and 76%
yield, respectively (eq 1). Thus, the more rapid lithium–halogen
exchange allowed the reaction with reactive electrophiles.
sec-BuLi
(2.0 ma)
CF3
CF3
Br
CF3
Li
RCHO
R
(1)
−105 to
−50 °C, 2 h
−105 °C
10 min / Et2O
OH
5a 73%
R = (CH2)2Ph 5b 76%
2 (2.0 ma)
1
R = Ph
.
Allylamines have been used as useful components for the
synthesis of N-heterocycles. For the construction of potential tri-
fluoromethylated heterocycles, we pursued the reaction of vinyl-
lithium 1 with imines to prepare 2-(trifluoromethyl)allylamines.3
Although N-benzylimine 6a was not reactive towards 1,
in the presence of BF3 OEt2 by taking advantage of the subtle
difference between reactivities of the two lithium species.10
On the other hand, 1-(trifluoromethyl)vinylation of highly
reactive electrophiles has remained problematic. The vinylation
of aldehydes suffers from nonselective addition of both 1 and n-
BuLi to give a poor yield of the desired allyl alcohols.7,11 Herein,
we report an efficient generation of (trifluoromethyl)vinyllithi-
um 1 to trap with reactive electrophiles, aldehydes and imines,
and its application to the construction of angularly trifluoro-
methylated bicyclic systems, which has been a desirable goal.
We first reexamined the generation of vinyllithium 1 from
bromotrifluoropropene 2 by treatment with several alkyllithiums
for 15 min. After quenching with methanol, the product distribu-
tions were observed by 19F NMR, as shown in Table 1. When the
reaction was carried out with n-BuLi at ꢁ78 ꢂC, the decomposi-
tion of vinyllithium 1 occurred to give 1,1-difluoroallene (4) in
54% yield, via elimination of lithium fluoride (Entry 1). When
carrying out the reaction at ꢁ105 ꢂC, we observed only a 14%
.
BF3 OEt2 promoted the reaction to afford the desired amine
7a in 81% yield (Table 2, Entry 1). When N-benzoyl- and N-to-
sylimine 6b and 6c were employed as the reactive electrophiles,
the corresponding N-allylamides 7b and 7c were obtained in ex-
cellent yield (Entries 2 and 3).12
We further examined (trifluoromethyl)vinylation of several
other N-tosylimines 6d–6g in view of their availability, ease of
handling, and the synthetic applicability of the products. All
N-tosylimines 6c–6g examined provided the corresponding N-
[2-(trifluoromethyl)allyl]sulfonamides 7c–7g in good to excel-
lent yield, as summarized in Table 2. Butanimine 6e gave 7e
in good yield, even though it had acidic protons ꢀ to the imino
group (Entry 5). Whereas 1,1-diphenylmethanimine 6g showed
Copyright Ó 2007 The Chemical Society of Japan

Chemistry Letters Vol.36, No.1 (2007)
23
Table 2. Synthesis of [(trifluoromethyl)allyl]sulfonamides 7
The sequence of (i) (trifluoromethyl)vinylation of imines,
followed by (ii) propargylation and (iii) the cobalt-mediated
intramolecular Pauson–Khand reaction, successfully provides a
facile entry to the fused N-heterocycles bearing an angular tri-
fluoromethyl group.
CF3
sec-BuLi
(2.0 ma)
R1
R2
NHR3
R1R2C=NR3 6
CF3
Li
CF3
−105 °C
10 min / Et2O
−105 to
Br
2 (2.2 ma)
−50 °C, 2 h
7
1
Entry
R1
R2
R3
7
Yield/%
We thank Dr. N. Kanoh (the University of Tokyo) for the
X-ray crystal structure analysis.
1a
2
3b
4b
5
H
H
H
H
H
Ph
Ph
Ph
CH2C6H5
COC6H5
7a
7b
7c
7d
7e
7f
81
97
90
89
77
96
76 (91)c
Ts
Ts
Ts
Ts
Ts
The present work is dedicated to Professor Teruaki
Mukaiyama on the occasion of his 80th birthday.
PhCH=CH
n-Pr
t-Bu
References and Notes
6
7
H
Ph
1
2
3
M. Shimizu, T. Hiyama, Angew. Chem., Int. Ed. 2005, 44, 214, and
references therein.
Ph
7g
a
BF3 OEt2 (1.5 ma) was added. bSolution of 6 was added
.
dropwise over 20 min. Vinyl bromide 2 (3.5 ma) and sec-
BuLi (3.2 ma) were used.
For a review on the synthesis of trifluoromethylated cyclic
compounds, see: P. Lin, J. Jiang, Tetrahedron 2000, 3635.
a) H. Lebel, V. Paquet, Org. Lett. 2002, 4, 1671, and references
therein. b) V. De Matteis, F. L. van Delft, H. Jakobi, S. Lindell,
J. Tiebes, F. P. J. T. Rutjes, J. Org. Chem. 2006, 71, 7527.
J. Ichikawa, H. Fukui, Y. Ishibashi, J. Org. Chem. 2003, 68, 7800.
a) J. Ichikawa, T. Mori, Y. Iwai, Chem. Lett. 2004, 33, 1354, and
references therein. b) T. Mori, Y. Iwai, J. Ichikawa, Chem. Lett.
2005, 34, 778.
c
modest reactivity, employing an excess amount of 1, generated
from 2 (3.5 ma) and sec-BuLi (3.2 ma), improved the yield to
91% (Entry 7).
4
5
Having accomplished the synthesis of 2-(trifluoromethyl)-
allylamides 7, we turned our attention to the construction
of fused-ring systems involving an angular trifluoromethyl
group,2,13 whose framework might be constructed by an intramo-
lecular Pauson–Khand reaction.14,15 Introduction of an angular
trifluoromethyl group can be an attractive tool for the analog
synthesis of steroids and alkaloids, and remains a challenging
task. Furthermore, very few examples of 1,6-enynes bearing a
electron-withdrawing group at the C-2 vinylic carbon have been
reported in the Pauson–Khand reaction.16 These facts prompted
us to investigate the Pauson–Khand reaction of N-propargyl-N-
[2-(trifluoromethyl)allyl]amides 8, readily prepared by propar-
gylation (with a propargyl bromide and NaH in DMF at rt) of
the above-obtained amides 7 in 84–92% yield.
Enyne 8a was treated with dicobalt octacarbonyl to generate
the cobalt–yne complex. Heating the complex at 60 ꢂC in aceto-
nitrile promoted the desired Pauson–Khand reaction of a CF3-
substituted terminal alkene, to afford the pyrrolidine ring-fused
cyclopentenone 9a bearing an angular trifluoromethyl group in
81% yield with high diastereoselectivity (anti:syn = 94:6)
(Table 3, Entry 1).17,18 Substrates 8b and 8c with an internal
alkyne moiety or an alkyl group at the allylic position also read-
ily underwent the cyclization to give trifluoromethylated pyrro-
lidines 9b and 9c in 85 and 71% yield, respectively (Entries 2
and 3).
6
7
J. Ichikawa, M. Fujiwara, T. Okauchi, T. Minami, Synlett 1998,
927.
a) F. G. Drakesmith, O. J. Stewart, P. Tarrant, J. Org. Chem. 1968,
33, 280. b) K. Iseki, Y. Kuroki, T. Nagai, Y. Kobayashi, J. Fluorine
Chem. 1994, 69, 5.
8
9
R. Nadano, J. Ichikawa, Synthesis 2006, 128.
For 1-(trifluoromethyl)vinylmetal species, see references in Ref. 8.
10 n-BuLi seemed less reactive than vinyllithium 1 under the reaction
conditions, presumably due to aggregation. See, Organolithiums:
Selectivity for Synthesis, ed. by J. Clayden, Elsevier, Oxford, 2002.
11 When a 1:1 mixture of 2 and n-BuLi was kept at ꢁ100 ꢂC for
15 min and then treated with benzaldehyde, two alcohols, derived
from vinyllithium 1 and n-BuLi, were obtained in 25 and 64%
yield, respectively.
12 To a solution of 2 (0.13 mL, 1.24 mmol) in Et2O (10 mL) was added
sec-butyllithium (1.07 M in cyclohexane, 1.05 mL, 1.13 mmol) in
Et2O (5 mL) at ꢁ105 ꢂC under Ar. After stirring for 10 min, 6c
(146 mg, 0.56 mmol) in Et2O (10 mL) was added over 20 min.
The reaction mixture was allowed to warm to ꢁ50 ꢂC over 2 h,
and the reaction was quenched with phosphate buffer (pH 7,
10 mL). Organic materials were extracted with EtOAc three times,
and the combined extracts were washed with brine, and dried
over MgSO4. After removal of the solvent under reduced pressure,
the residue was purified by silica-gel column chromatography (hex-
ane–EtOAc, 10:1) to give 7c (180 mg, 90%) as colorless crystals.
13 For a recent example, see: T. Barhoumi-Slimi, B. Crousse, M.
´
´
Ourevitch, M. El Gaied, J.-P. Begue, D. Bonnet-Delpon, J.
Fluorine Chem. 2002, 117, 137.
14 For reviews on the Pauson–Khand reaction, see: S. E. Gibson, N.
Mainolfi, Angew. Chem., Int. Ed. 2005, 44, 3022, and references
therein.
15 For the Pauson–Khand reaction in (ꢁ)-dendrobin synthesis, see:
J. Cassayre, S. Z. Zard, J. Organomet. Chem. 2001, 624, 316.
16 For reviews on the Pauson–Khand reaction of electron-deficient
alkenes, see: a) M. R. Rivero, J. Adrio, J. C. Carretero, Synlett
2005, 26. b) M. R. Rivero, J. Adrio, J. C. Carretero, Eur. J. Org.
Chem. 2002, 2881. See also: c) L. V. R. Bon˜aga, M. E. Krafft,
Tetrahedron 2004, 9795.
17 The configuration of the major isomer was determined to be anti by
X-ray crystallography of the cyclopentanone, derived via reduction
of 9a.
18 The Pauson–Khand reaction of a similar N-containing 1,6-enyne
with a CF3 group at C-1 was reported to be unsuccessful. M.
Ishizaki, D. Suzuki, O. Hoshino, J. Fluorine Chem. 2001, 111, 81.
Table 3. Synthesis of 3a-CF3-cyclopenta[c]pyrroles 9
R1
R1
CF3
Co2(CO)8 (1.2 ma)
TsN
CF3
TsN
O
rt, 0.5 h / CH2Cl2
then, 60 °C / CH3CN
R2
R2
8
9
Entry
R1
R2 Time/h
9
Yield/% anti:syna
1
2
3
Ph
Ph
n-Pr
H
Et
H
3
2
3
9a
9b
9c
81
85
71
94:6b
83:17c
86:14c
b
aDetermined by 19F NMR. See Ref. 17. cConfiguration was
determined in analogy with 9a by comparing 1H and
19F NMR data of each isomer.
Published on the web (Advance View) December 2, 2006; doi:10.1246/cl.2007.22