Regiochemistry in 1,3-dipolar cycloadditions of the azomethine ylide formed from diethyl aminomalonate and paraformaldehyde

Article abstract of DOI:10.1021/jo010645x

The azomethine ylide derived from the condensation of diethyl aminomalonate with paraformaldehyde undergoes 1,3-dipolar cycloadditions with acrylate and propiolate derivatives. Contrary to a previous report, these reactions yield mixtures of regioisomers

Full text of DOI:10.1021/jo010645x

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
298 J . Org. Chem. 2002, 67, 298-300
Sch em e 1
Regioch em istr y in 1,3-Dip ola r
Cycloa d d ition s of th e Azom eth in e Ylid e
F or m ed fr om Dieth yl Am in om a lon a te a n d
P a r a for m a ld eh yd e
Charles M. Blazey and Clayton H. Heathcock*
Center for New Directions in Organic Synthesis,
Department of Chemistry, University of California,
Berkeley, California 94720
chh@steroid.cchem.berkeley.edu
Received J une 25, 2001
Abstr a ct: The azomethine ylide derived from the conden-
sation of diethyl aminomalonate with paraformaldehyde
undergoes 1,3-dipolar cycloadditions with acrylate and pro-
piolate derivatives. Contrary to a previous report, these
reactions yield mixtures of regioisomers generally favoring
the 2,2,3-trisubstituted product. However, the relative quan-
tity of the 2,2,4-trisubstituted product formed increases with
an increase in the size of the activating group on the
dipolaroplile.
Sch em e 2
In 1989, we began studies directed toward the total
synthesis of the marine sponge metabolite sarain A.1 In
a key step of our synthesis of the core structure of sarain
A, an unactivated olefin undergoes an intramolecular 1,3-
dipolar cycloaddition (1,3-DC) with an azomethine ylide
to provide a fused, bicyclic pyrrolidine (Scheme 1). This
transformation is conveniently performed by stirring the
amine precursor with paraformaldehyde in toluene at
reflux.1e
The precedent for this procedure was provided by
earlier reports of intermolecular 1,3-DC employing di-
ethyl aminomalonate (1) as the amine substrate.2,3 In
1980, Amornraksa and Grigg utilized arylaldehydes to
generate azomethine ylides which were trapped with
symmetrical, activated dipolarophiles.2a Later, Husinec
and co-workers reported the successful use of paraform-
aldehyde in similar reactions.2b These authors included
unsymmetrical dipolarophiles, but claimed to observe
complete regioselectivity in product formation. We specu-
lated that an intermolecular 1,3-DC might permit us to
improve the efficiency of our synthesis by increasing the
convergence of the route.
We began by repeating a previously reported 1,3-DC
(Scheme 2) as described by Husinec.2b Freshly distilled
1 was stirred with paraformaldehyde and ethyl acrylate
in toluene at reflux for 13 h. A Dean-Stark trap was
employed to remove the water formed through the course
of the reaction. After the allotted time, the reaction was
cooled to room temperature, and the toluene was removed
in vacuo. The crude product mixture was examined by
1H NMR spectroscopy. Contrary to the earlier report, the
crude material contained a mixture of two regioisomers.
The isomers were separated by HPLC and their struc-
tures determined spectroscopically. Analysis of the 1H
NMR spectra led us to conclude that the major isomer
was 2,2,3-trisubstituted pyrrolidine 2a rather than 2,2,4-
trisubstituted pyrrolidine 2b, as previously reported.
The characteristic coupling patterns of the carbon-
bound protons on the pyrrolidine ring proved most
valuable in establishing the regiochemistry of the cy-
cloadduct isomers. In the case of 2b, four distinct dd were
expected for the four protons at C-3 and C-5. The
remaining proton bound to C-4 was expected to appear
as a more complicated multiplet. On the other hand, the
1H NMR spectrum of 2a would possess only one dd for
the proton at C-3 with greater splitting for the four other
protons, at C-4 and C-5. Indeed, exactly these patterns
are observed for the two cycloadduct isomers isolated.
However, as noted above, the major isomer was in fact
2a , not 2b.
(1) For our progress toward the synthesis of sarain A, see: (a)
Henke, B. R,; Kouklis, A. J .; Heathcock, C. H. J . Org. Chem. 1992, 57,
7056. (b) Sharp, M. J .; Heathcock, C. H. Tetrahedron. Lett. 1994, 35,
3651. (c) Griffith, D. A.; Heathcock, C. H. Tetrahedron. Lett. 1995, 36,
2381. (d) Heathcock, C. H.; Clasby, M.; Griffith, D. A.; Henke, B. R.;
Sharp M. J . Synlett 1995, 467. (e) Denhart, D. J .; Griffith, D. A.;
Heathcock, C. H. J . Org. Chem. 1998, 63, 9616.
(2) (a) Amornraksa, K.; Grigg, R. Tetrahedron Lett. 1980, 21, 2197.
(b) Husinec, S.; Savic, V.; Porter, A. E. A Tetrahedron Lett. 1988, 29,
6649
(3) For reviews of 1,3-dipolar cycloadditions of azomethine ylides,
see: (a) Lown, J . W. In 1,3-Dipolar Cycloaddition Chemistry; Padwa,
A, Ed.; Wiley: New York, 1984; Vol. 1 Chapter 6. (b) Vedejs, E. In
Advances in Cycloaddition; Curran, D. P., Ed.; J ai Press: Greenwich,
CT, 1988; Vol. 1, p 33. (c) Tsuge O.; Kanemasa, S. In Advances in
Heterocyclic Chemistry; Katritzky, A. R., Ed.; Academic: San Diego,
1989; Vol. 45, p 231. (d) Tsuge O.; Kanemasa, S. In Advances in
Cycloaddition; Curran, D. P., Ed.; J ai Press: Greenwich, CT, 1993;
Vol. 3, p 99. (e) Grigg, R.; Sridharan, V. In Advances in Cycloaddition;
Curran, D. P., Ed.; J ai Press: Greenwich, CT, 1993; Vol. 3, p 161.
With the results for the 1,3-DC with ethyl acrylate in
hand, we chose to optimize the reaction conditions and
explore the impact of changing the dipolarophile. We
10.1021/jo010645x CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/14/2001

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Notes
J . Org. Chem., Vol. 67, No. 1, 2002 299
Ta ble 1. 1,3-Dip ola r cycloa d d ition s: Yield s a n d
Regioselectivity
Exp er im en ta l Section
Gen er a l P r oced u r es. All solvents for reactions and extrac-
tion were reagent grade and used as received. Paraformaldehyde
was dried over P2O5 under high vacuum. Other reagents were
used as obtained, without additional purification. 1H NMR (400
MHz) and 13C NMR (100 MHz) spectra were acquired in
deuterated chloroform (CDCl3). HPLC purification was per-
formed on a Dynamax 60-A column with HPLC grade solvents.
Reactions were performed under an N2 atmosphere.
Gen er a l P r oced u r e for 1,3 Dip ola r Cycloa d d it ion s.
Commercial diethyl aminomalonate hydrochloride was converted
to its free amine by stirring in ethanol with excess potassium
carbonate for about 1 h. The solids were then filtered and the
ethanol removed in vacuo. Diethyl aminomalonate (1) was
subsequently distilled at reduced pressure (10 Torr) using a
Kugelrohr apparatus. This material was stored in a refrigerator,
and it maintained its integrity for several days as determined
by 1H NMR. The amine 1 (1.5 mmol, 263 mg) was dissolved in
toluene (7.5 mL), and paraformaldehyde (1.5 mmol, 45 mg) and
dipolarophile (1.5 mmol) were added to the mixture. The reaction
flask was fitted with a Dean-Stark apparatus (this precaution
was later determined to be unnecessary) and heated to reflux.
After 13 h, the mixture was cooled, and the toluene was removed
in vacuo. Unless otherwise noted, addition of anisole (1.5 mmol,
16.3 µL) as an internal standard permitted the determination
of % yield by 1H NMR of the crude reaction mixture. Product
ratios were also established by 1H NMR of the crude reaction
mixture. Although most of the pairs of regioisomers could be
resolved by TLC, we found purification by flash chromatography
less satisfactory than normal phase HPLC for the isolation of
the cycloadducts.
Dip ola r Cyclod d ition w ith Eth yl Acr yla te a s th e Dip o-
la r op h ile. Employing ethyl acrylate (153 mg, 163 µL) as the
dipolarophile in the foregoing procedure, two regioisomers were
obtained as a 3:1 mixture in 88% yield as determined by crude
NMR (68% isolated yield after HPLC). The major isomer proved
to be 2a .
Eth yl (()-2,3-Bis(eth oxyca r bon yl)p yr r olid in e-2-ca r box-
yla te (2a ). Pale yellow oil: 1H NMR δ 1.22 (t, 3H, J ) 7.4 Hz),
1.23 (t, 3H, J ) 7.1 Hz), 1.24 (t, 3H, J ) 7.5 Hz), 2.02-2.10 (m,
1H), 2.19-2.28 (m, 1H), 2.97 (ddd, 1H, J ) 5.4, 7.5, 9.9 Hz),
3.25 (dt, 1H, J ) 7.2, 9.9 Hz), 3.75 (dd, 1H, J ) 5.1, 8.2), 4.04-
4.29 (m, 6H); 13C NMR δ 13.9, 13.9, 14.0, 30.6, 45.8, 49.6, 60.9,
62.0, 62.1, 75.1, 169.5, 170.0, 172.6; IR (thin film) 3356, 1736
cm-1. Anal. Calcd for C13H21NO6 C, 54.35; H, 7.37; N, 4.88.
Found: C, 54.12; H, 7.35; N, 5.16.
found that the Dean-Stark trap was unnecessary and
had no impact on reaction rate or product distribution.
Furthermore, we found that when employing ethyl acry-
late as the dipolarophile, the reaction was complete in
less than 4 h. We were disappointed to find that ethyl
crotonate yielded no pyrrolidine cycloadducts. However,
Husinec had already reported the failure of â-substituted
R,â-unsaturated ketones to undergo 1,3-DC, so the ethyl
crotonate results were not surprising.2b We then limited
our study to commercially available â-unsubstituted
dipolarophiles.
We successfully employed acrylonitrile, tert-butyl acry-
late, and ethyl propiolate as dipolarophiles. The yields
and regioselectivities of these experiments are reported
in Table 1. The structures of the acrylonitrile and the
tert-butyl acrylate cycloadducts (3a ,b and 4a ,b, respec-
tively) were determined analogously to the ethyl acrylate
case discussed above. The structures for the correspond-
ing ethyl propiolate cycloadducts (5a ,b) could not be
established unambiguously by 1-dimensional 1H NMR
spectroscopy. However, catalytic hydrogenation of these
pyrrolines provided pyrrolidines 2a and 2b, thereby
clarifying the regioselectivity of the ethyl propiolate 1,3-
DC.
Eth yl (()-2,4-Bis(eth oxyca r bon yl)p yr r olid in e-2-ca r box-
yla te (2b). Pale yellow oil: 1H NMR δ 1.23 (t, 3H, J ) 7.3 Hz),
1.25 (t, 3H, J ) 7.1 Hz), 1.26 (t, 3H, J ) 7.3 Hz), 2.51 (dd, 1H,
J ) 7.8, 13.7 Hz), 2.74 (dd, 1H, J ) 8.3, 13.7 Hz), 3.01-3.09 (m,
1H), 3.23 (dd, 1H, J ) 7.1, 10.2 Hz), 3.28 (dd, 1H, J ) 7.7, 10.2
Hz), 4.13 (q, 2H, J ) 7.3 Hz), 4.16-4.28 (m, 4H); 13C NMR δ
14.0, 14.0, 14.1, 35.7, 43.7, 50.0, 60.9, 62.0, 62.0, 72.2, 170.8,
171.0, 173.1; IR 3353, 1735 cm-1
.
Dip ola r Cyclod d ition w ith Acr ylon itr ile a s th e Dip o-
la r op h ile. Employing acrylonitrile (80 mg, 99 µL) as the
dipolarophile in the foregoing procedure, two regioisomers were
obtained as a 4.8:1 mixture in 69% yield as determined by crude
NMR (61% isolated yield after HPLC). The major isomer proved
to be 3a . The minor isomer could not be isolated in pure form.
Eth yl (()-3-Cya n o-2-eth oxyca r bon ylp yr r olid in e-2-ca r -
boxyla te (3a ). Pale yellow oil: 1H NMR δ 1.27 (t, 3H, J ) 7.1
Hz), 1.33 (t, 3H, J ) 7.1 Hz), 2.17-2.24 (m, 1H), 2.30-2.39 (m,
1H), 2.84 (br s, 1H), 3.06 (ddd, 1H, J ) 5.4, 7.0, 10.1 Hz), 3.34
(dt, 1H, J ) 7.1, 10.1 Hz), 3.84 (dd, 1H, J ) 5.4, 7.7 Hz) 4.15-
4.37 (m, 4H); 13C NMR δ 13.8, 13.9, 31.3, 35.1, 45.8, 62.7, 63.1,
For entries 1-3 (Table 1), as the size of the dipolaro-
phile activating group increases, more of the less steri-
cally congested 2,2,4-trisubstituted product is obtained.
The electronics of the various dipolarophiles are not
expected to differ significantly, so the trend seems to
point to a steric effect. When ethyl propiolate serves as
the dipolarophile (entry 4), the regioselectivity reverses,
though only moderately. The source of this reversal is
unclear, and either electronics and/or sterics may come
into play.
74.7, 118.6, 168.4, 168.7; IR (thin film) 3352, 2244, 1736 cm-1
.
Anal. Calcd for C11H16N2O4 C, 54.99; H, 6.71; N, 11.66. Found:
C, 55.04; H, 6.50; N, 11.63.
We have reexamined the 1,3-DC of a few simple
commercial dipolarophiles with the azomethine ylide
generated from 1 and paraformaldehyde. In contrast to
the previous report, we have found these reactions not
to be regiospecific. Upon further study, we have uncov-
ered a small, but observable, steric effect based on the
size of the dipolarophile activating group.
Eth yl (()-4-Cya n o-2-eth oxyca r bon ylp yr r olid in e-2-ca r -
boxyla te (3b). Pale yellow oil. 1H NMR δ 1.26 (t, 3H, J ) 7.1
Hz), 1.28 (t, 3H, J ) 7.1 Hz), 2.55 (dd, 1H, J ) 8.0, 13.7 Hz),
2.85 (dd, 1H, J ) 8.2, 13.7 Hz), 3.05 (pentet, 1H, J ) 7.7 Hz),
3.26 (dd, 1H, J ) 7.6, 10.1 Hz), 3.39 (dd, 1H, J ) 7.0, 10.2 Hz),
4.18-4.31 (m, 4H); 13C NMR δ 13.9, 14.0, 28.2, 36.7, 50.7, 62.4,
62.6, 71.4, 119.8, 170.0, 170.1; IR 3345, 2243, 1731 cm-1
.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
300 J . Org. Chem., Vol. 67, No. 1, 2002
Notes
Dip ola r Cyclod d it ion w it h ter t-Bu t yl Acr yla t e a s t h e
Dip ola r op h ile. Employing tert-butyl acrylate (192 mg, 219 µL)
as the dipolarophile in the foregoing procedure, two regioisomers
were obtained as a 1:1 mixture in 83% yield as determined by
crude NMR (68% isolated yield after HPLC).
Eth yl (()-3-ter t-Bu toxyca r bon yl-2-eth oxyca r bon ylp yr -
r olid in e-2-ca r boxyla te (4a ). Pale yellow oil. 1H NMR δ 1.18
(t, 3H, J ) 7.1 Hz), 1.35 (s, 9H), 1.92-1.99 (m, 1H), 2.15-2.24
(m, 1H), 2.91 (ddd, 1H, J ) 4.4, 7.7, 9.6 Hz), 2.98 (br s, 1H),
3.17 (q, 1H, J ) 8.2 Hz), 3.59 (dd, 1H, J ) 4.0, 8.3 Hz), 4.00-
4.10 (m, 1H), 4.11-4.27 (m, 1H); 13C NMR δ 13.8, 27.7, 30.9,
45.5, 50.3, 61.7, 61.9, 74.7, 80.9, 169.5, 170.0, 171.7; IR (thin
film) 3358, 1737 cm-1. Anal. Calcd for C15H25NO6 C, 57.13; H,
7.99; N, 4.44. Found: C, 57.24; H, 7.92; N, 4.52
Eth yl (()-4-ter t-Bu toxyca r bon yl-2-eth oxyca r bon ylp yr -
r olid in e-2-ca r boxyla te (4b). Pale yellow oil. 1H NMR δ 1.16
(t, 3H, J ) 7.1 Hz), 1.17 (t, 3H, J ) 7.1 Hz), 1.34 (s, 9H), 2.37
(dd, 1H, J ) 7.9, 13.6 Hz), 2.61 (dd, 1H, J ) 8.3, 13.6 Hz), 2.86
(br s), 2.84-2.92 (m, 1H), 3.08-3.17 (m, 2H), 4.05-4.20 (m, 4H);
13C NMR δ 13.8, 13.8, 27.7, 35.5, 44.5, 49.8, 61.7, 61.8, 72.0, 80.6,
170.7, 170.9, 172.1; IR (thin film) 3352, 1728 cm-1. Anal. Calcd
for C15H25NO6 C, 57.13; H, 7.99; N, 4.44. Found: C, 57.09; H,
8.05; N, 4.74.
Dip ola r Cyclod d ition w ith Eth yl P r op iola te a s th e
d ip ola r op h ile. Amine 1 (5.7 mmol, 1.0 g) was dissolved in
toluene (30 mL). To this mixture were added ethyl propiolate
(5.7 mmol, 559 mg, 578 µL) and paraformaldehyde (5.7 mmol,
171 mg). The reaction mixture was heated to reflux and stirred
12 h before cooling to room temperature. The toluene was
removed in vacuo and o-dinitrobenzene (0.58 mmol, 100.5 mg)
was added as an NMR standard. Two regioisomers were
obtained as a 1.2:1 mixture in 58% yield as determined by crude
NMR (43% isolated yield after HPLC). The major isomer proved
to be 5b.
Eth yl (()-2,3-Bis(eth oxyca r bon yl)-3-p yr r olin e-2-ca r box-
1
yla te (5b). Pale yellow oil. H NMR δ 1.26 (t, 6H, J ) 7.1 Hz),
1.28 (t, 3H, J ) 7.2 Hz), 3.13 (br s, 1H), 4.04 (d, 1H, J ) 2.2 Hz),
4.18-4.25 (m, 6H), 6.71 (t, 1H, J ) 2.2 Hz); 13C NMR δ 13.9,
14.1, 52.8, 60.9, 62.4, 79.8, 135.2, 138.1, 162.9, 169.3; IR 3362,
1730, 1643 cm-1. Anal. Calcd for C13H19NO6 C, 54.73; H, 6.71;
N, 4.91 Found: C, 54.46; H, 6.73; N, 5.19.
Eth yl (()-2,4-Bis(eth oxyca r bon yl)-3-p yr r olin e-2-ca r box-
1
yla te (5a ). Pale yellow oil. H NMR δ 1.27 (t, 6H, J ) 7.1 Hz),
1.28 (t, 3H, J ) 7.1 Hz), 2.97 (br s, 1H), 4.00 (d, 1H, J ) 2.1 Hz),
4.17-4.31 (m, 6H), 6.98 (t, 1H, J ) 2.1 Hz); 13C NMR δ 13.9,
14.1, 53.3, 60.7, 62.2, 77.9, 133.2, 143.6, 162.3, 170.1; IR 3362,
1733, 1639 cm-1. Anal. Calcd for C13H19NO6 C, 54.73; H, 6.71;
N, 4.91 Found: C, 54.52; H, 6.74; N, 5.16.
Ack n ow led gm en t. This work was supported by a
research grant from the United States Public Health
Service (GM46057). The Center for New Directions in
Organic Synthesis is supported by Bristol-Myers Squibb
as Sponsoring Member.
J O010645X