3802
P. D. Barman et al. / Tetrahedron Letters 54 (2013) 3801–3804
In our initial attempt (Scheme 1), reaction of the sugar-derived
O
exocyclic olefin, the required dipolarophile 1a with the non-stabi-
lized azomethine ylide, generated in situ by the condensation of
isatin 2a with the amino acid sarcosine (3) in toluene at reflux tem-
perature for 12 h furnished the bispiropyrrolidinyl-oxindole deriv-
ative 4a in 83% yield. Similar reactions of the substituted isatins
2b–d afforded 4b–d in 78%, 85% and 81% yields, respectively. The
structure of 4a–d13,14 was elucidated from the 1H and 13C NMR
spectroscopic data. The formation of the pyrrolidine moiety in 4a
was confirmed by the appearance of three-proton singlet (NCH3)
at d 2.21 and two-proton singlet (NCH2) at d 3.43 in the 1H NMR
spectrum. The observed singlet at d 3.97 was assigned for the
CHCO2Et in the pyrrolidine moiety. The presence of the signals at
d 55.0 (CH), 56.2 (CH2), 57.4 (C) and 73.1 (C) in the 13C NMR spec-
trum also indicated the creation of the pyrrolidine ring during the
cycloaddition reaction. Finally, the structural confirmation of 4a
was obtained from a single crystal X-ray crystallographic study
(the ORTEP diagram is given in Fig. 1).15
We next turned our attention to generalize the methodology
employing non-aromatic cycloalkanones, instead of isatins, for
the generation of azomethine ylide (1,3-dipoles) by the reaction
with sarcosine, although a very few reports exist in the literature
on this method.16 Thus, at the outset, we chose cyclopentanone
(5a) and sarcosine (3) as the ylide generator and reacted with 1a
(Scheme 2) under the previously optimized reaction condition.
This although produced the desired product 6a,13 the yield was
too low (22% Table 1). A careful modification of the reaction condi-
tion was, therefore, needed.
O
H
O
O
O
O
H
n
R4
O
Me
N
COOR3
O
5a-d
O
n
R4
O
Sarcosine (3)
DIPEA, toluene
reflux
6a-e
O
3
4
CO2R3
6a
6b
6c
: R = Et, R = H, n = 1
: R = Et, R = H, n = 2
: R = Et, R = Et; n = 2
4
3
4
5a
: n = 1, R = H
4
3
4
1a: R3 = Et
5b
: n = 2, R = H
5c: n = 2, R4 = Et
5d: n = 3, R4 = H
6d: R3 = Et, R4 = H, n = 3
6e: R3 = Me, R4 = H, n = 2
1b: R3 = Me
Scheme 2. Synthesis of sugar-fused bispiroheterocycles using cycloalkanones.
Table 1
Optimization of reaction condition between 1a and 5a
Entry
Base
Solvent
Time (h)
Temp (°C)
Yielda (%)
1
2
3
4
5
6
7
8
None
None
None
None
None
Et3N
Toluene
Methanol
o-Xylene
DMF
10
10
12
12
12
10
10
10
Reflux
Reflux
120
130
130
Reflux
Reflux
Reflux
22
10
20
14
12
51
53
92
DMSO
Toluene
Toluene
Toluene
DBU
DIPEA
a
Isolated yield.
We first decided to screen solvents of the reaction. Attempts
employing methanol, o-xylene, DMF, DMSO etc. under the conven-
tional solution-phase protocols were frustrating, yielding product
6a to the extent of 10–22% (entries 1–5). Thus, in a modified ap-
proach, the addition of Et3N into the reaction mixture to facilitate
the decarboxylation of sarcosine-iminium ion9b in forming an ylide
improved the yield up to 51% (entry 6). The coupling was then fur-
ther reinvestigated by replacing Et3N by DBU. This, however, failed
to show any substantial improvement in the outcome (Table 1, en-
try 7) and had to be discarded. Finally, we tested DIPEA as the base
to condense 1a with 5a, which led to maximization of the product
yield (92%) in 10 h under reflux condition in toluene (entry 8).17
The success of the above cycloaddition reaction with cyclo-
pentanone prompted us to test this newly developed reaction tool
in other higher homologues of cycloalkanones (Scheme 2). Thus,
cyclohexanones (5b,c) and cycloheptanone (5d) were reacted with
the dipolarophiles (1a,b) under the optimal set of reaction condi-
tions, producing the corresponding cycloadducts (6b–e)12 in high
yields (Table 2, entries 2–5). The structures of all the bispiropyrro-
lidinyl-cycloalkanes were confirmed by NMR analyses. They also
exhibited exact masses in their mass spectra.
Once these bispiropyrrolidines with cycloalkane rings were
constructed, we moved towards attempting the synthesis of anal-
ogous bispiropyrrolidine using sterically hindered ketone acenaph-
thoquinone 7, which also led to the cycloadduct 8 (Scheme 3) in
75% yield upon reaction with the azomethine ylide, generated from
7 and 3, the structure of which was deduced by spectral analyses.
From mechanistic considerations, the cycloaddition reaction,
we believe, proceeds with an initial iminium ion formation, which
subsequently decarboxylates (facilitated by DIPEA) to generate an
ylide (1,3-dipole) (Fig. 2). Once the ylide is formed, it behaves as a
nucleophile and attacks the dipolarophile 1a/1b to produce the
cycloadducts in high yields upon [3+2] cycloaddition reaction. It
is worthy to mention that the cycloadducts were essentially
Table 2
Synthesis of sugar-fused bispiroheterocycles using cycloalkanones 5a–da
Entry
Dipolarophile
Ketone
Time (h)
Product
Yieldb (%)
1
2
3
4
5
1a
1a
1a
1a
1b
5a
5b
5c
5d
5b
10
14
11
11
14
6a
6b
6c
6d
6e
92
88
85
82
78
a
All the reactions were performed in toluene under reflux condition using sar-
cosine and DIPEA as the base.
Figure 1. ORTEP diagram of 4a. The displacement ellipsoids are drawn at the
probability of 50%.
b
Isolated yield.