Proline mediated asymmetric ketol cyclization: A template reaction

Article abstract of DOI:10.1016/S0957-4166(99)00174-3

¹H, ¹³C and ATR-FTIR Spectroscopic studies reveal that the asymmetric cyclization of prochiral triones 1 and 2 in the presence of S-proline is a template reaction.

Full text of DOI:10.1016/S0957-4166(99)00174-3

Pergamon
Tetrahedron: Asymmetry 10 (1999) 1631–1634
Proline mediated asymmetric ketol cyclization: a template
†
reaction
a
b
b
a,∗
D. Rajagopal, M. S. Moni, S. Subramanian and S. Swaminathan
a
Department of Organic Chemistry, University of Madras, Guindy Campus, Madras 600 025, India
b
Regional Sophisticated Instrumentation Centre, Indian Institute of Technology, Madras 600 036, India
Received 30 March 1999; accepted 30 April 1999
Abstract
1H, 13C and ATR-FTIR spectroscopic studies reveal that the asymmetric cyclization of prochiral triones 1 and 2
in the presence of S-proline is a template reaction. © 1999 Elsevier Science Ltd. All rights reserved.
1
The prochiral triones 1 and 2 have been reported by us earlier to give the diones S-3 and S-4,
respectively, when neat (in the absence of solvents) in the presence of S-proline.
This study suggested that the asymmetric cyclization of the above triones by the well-known
2–4
literature methods using solvents was also probably a surface reaction. In these methods, a solution of
the trione 1 or 2 in a polar aprotic solvent was stirred with catalytic amounts of S-proline which remained
mostly undissolved and was filtered before work-up. To our surprise, we found using trione 2 that there
was no reaction (no λmax at 247 nm) in the absence of the insoluble S-proline, i.e. when the stirred mixture
of proline and solvent (CH CN or DMSO) was first filtered free of undissolved proline and the filtrate
3
reacted with trione 2. Evidently the cyclization was a surface reaction and we decided to probe into the
1
13
mechanism of both cyclizations with the aid of high resolution H and C NMR and also ATR-FTIR
techniques. In spite of several studies, the mechanism of chiral transfer in the above cyclizations is still
not clear. Hajos and Parrish in their seminal paper2 proposed two mechanisms, one involving reaction of
S-proline with one of the enantiotopic ring carbonyls to give an intermediate carbinol amine 5 followed
a
∗
†
Corresponding author.
A part of this work was presented at the 9th International Symposium on Chiral Discrimination (ISCD-97) held at Nagoya,
Japan during September 1997.
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00174-3

1632
D. Rajagopal et al. / Tetrahedron: Asymmetry 10 (1999) 1631–1634
by cyclization and another involving the side chain carbonyl to give the protonated enamine 6 followed
by cyclization with one of the diastereotopic ring carbonyls.
5
Subsequent studies have mostly favoured initial reaction with the side chain carbonyl, though the
6
work of Kasthuri et al. based on some spectroscopic studies casts doubts on the enamine mechanism.
1
13
7a
We report here results based on 400 MHz H NMR and 100.40 MHz C NMR spectra taken at
intervals to follow the course of the reaction of equivalent amounts of triketones 1 (0.25 mmol) and
2
(0.25 mmol) in DMSO-d (0.5 ml) with S-proline (0.25 mmol) at 23°C: (a) there is no evidence of
6
the reaction of S-proline with the carbonyl groups in both triketones to give enamine or carbinol amine
1
13
intermediates (in both H and C spectra); (b) S-proline is hardly soluble in DMSO at room temperature
since no signals are seen for the proline carbons (weak proton signals are however seen for proline due
to moisture associated with DMSO). Since there is no reaction as mentioned above in the absence of
undissolved proline, it is obvious that the substrates 1 and 2 can react only on the surface of the insoluble
8
crystalline proline; (c) within 15 min of mixing trione 1 with S-proline, one observes carbon signals: ten
for (3aS,7aS) ketol 7, eight for unreacted trione 1 and seven for DMSO. In the proton spectrum also taken
immediately on mixing the reactants and at intervals of 15 min over a 2 h period, only starting trione 1
and ketol 7 show up with no signals for proline as such or in a combined form. The formation of cis-ketol
7
is complete in 5 to 6 h with no trace of any other product.
With a view to detect intermediates, if any, which may have escaped detection in our NMR studies,
we studied the ATR-FTIR spectra also.7b The liquid trione 1 was stirred with an equivalent amount of
powdered dry S-proline and ATR-FTIR spectra were taken at regular intervals over a period of 40 h.
−
1
−1
Initially, one observes the twin 1,3 ring C_O absorptions at 1762 cm (m) and 1700 cm (s), the latter
−1
merging with the side chain C_O absorption in addition to proline peaks at 3500 cm (NH str. br.),
1
−1
−
−1
1
557 cm− (w) (NH2+ bending) and 1618 cm (-CO ). The C_O at 1762 cm gradually disappears
2
−l
−1
and product peaks OH at 3500 cm (s), 6-membered C_O at 1701 cm and 5-membered C_O at 1741
−1
cm show up. There is no absorption for an enamine double bond.
Evidently, ketol formation and transfer of chirality take place simultaneously on the surface of the
9
bifunctional chiral catalyst as pictured in Fig. 1. This shows preferential hydrogen bonding with the pro-
R carbonyl group which attracts the substrate to the catalyst surface followed by synchronous protonation
−
10
of the same C_O and deprotonation of the side chain CH by -CO2 of proline leading to cyclization.
3
Figure 1.

D. Rajagopal et al. / Tetrahedron: Asymmetry 10 (1999) 1631–1634
1633
In 8 (Fig. 1), the trione is in the plane of the paper, the -CO2− and -NH2+ of proline are pictured to
project from the crystal surface perpendicular to the plane of the paper, suitably aligned for reaction.
After the ketol is formed, it is removed from the site by the solvent while the amino acid resumes
its zwitterionic form and is ready to ketolize a second molecule of substrate. This process can repeat
itself at a particular site a number of times. The crystal surface must function like a template as it
contains a number of such active sites where ketolization can take place simultaneously and repeatedly,
accounting for the high efficiency of the catalyst. The conformation of cis-ketol 7 with an axial -CH3
and an equatorial -OH has been established by Hajos and Parrish.2a With catalytic amounts of proline,
even after 10 h, unreacted trione 1 is still present and ketol formation is incomplete, due, no doubt, to the
reduced number of active sites. In the heterogeneous system containing insoluble proline and triketone
1
in DMSO solution, the zwitterionic nature of proline must hold all the individual molecules of proline
14
together in the crystal lattice. Only surface reactions such as hydrogen bonding may be expected to lead
to the one-step conversion of 8 to 7 (Fig. 1) with excellent enantioselectivity2 rather than a multi-step
conversion involving intermediates 5 or 6 or their transformation products for none of which we have
spectral evidence. The reaction of trione 2 with S-proline is more complicated and takes a longer time
to complete than in the case of trione 1. The immediate product on mixing the reactants (within the first
a
11
1
5 min) is the dienamine 12 easily identified by strong signals for the two double bonds (δ 95.23 (C ); δ
4
1
40.37 (C ); δ 111.07 (C ) and δ 142.26 (C ) ppm) besides unreacted trione 2. A number of other very
4a
5
6
weak signals, four in the saturated C_O region and three in the -C–OH region, are seen indicating the
2
c
presence of the known ketol 9 and other ketols (possibly 10 and 11, Scheme 1).
Scheme 1.
Within the next 15 min the diketone S-4 shows up besides dienamine 12 which at this stage has stronger
signals than dione S-4. During the next 22 h, the signals for S-4 become progressively stronger than those
of 12 and also of starting trione 2. On continued standing (a week) only signals for S-4 are seen. The
proton spectra also taken immediately on mixing the reactants and at intervals of 15 min over a period of
2.5 h and finally after 24 h, confirm the initial formation of dienamine 12 and S-4 whose signals become
progressively stronger. The dienamine has absorption at δ 5.2 (t, C vinyl H), δ 4.7 (s, C , vinyl H)
4
5
and δ 4.1 (t, C proline protons); the formation of S-4 is followed by its absorption at δ 5.8 (s, vinyl
2
13
H). After standing for a week, only signals for S-4 are seen as in the C NMR. These data indicate:
i) that unlike in the case of trione 1 where the ketol 7 is the sole product, the ketol 9 is very reactive
(
getting transformed to S-4 via dienamine 12 formed via transformations a -→ b -→ c in Scheme 1; (ii)
that the conversion of trione 2 to diketone S-4 involves transformations partly on the surface of proline
12
crystals and partly in solution. Scheme 1 is based on Spencer’s mechanism of amine-catalyzed β-ketol
dehydration. The intermediates 10, 11 and 12 must be formed on the surface of solid proline and held
13
there because of their zwitterionic nature, limiting their detection in spectra. The C NMR signals for 12
are stronger than those of 9, 10 and 11, probably because a second molecule of water becomes available
after step c to dissolve the bonding of 12 to the proline surface before hydrolysis (step d).
The marked difference in reactivity of ketols 7 and 9 towards S-proline may be due to a steroid

1634
D. Rajagopal et al. / Tetrahedron: Asymmetry 10 (1999) 1631–1634
conformation for ketol 9 in contrast to the non-steroid conformation for ketol 7. Some related ketols
with steroid-like conformation have been reported.5
In summary, spectroscopic studies reveal that the asymmetric cyclization of triones 1 and 2 in the
13
presence of S-proline is a template reaction and that the primary reaction products are the cis-ketols 7
and 9, respectively, with different reactivities.
Acknowledgements
The authors thank the Dept. of Science and Technology, Govt. of India for financial assistance.
References
1
2
. Rajagopal, D.; Rajagopalan, K.; Swaminathan, S. Tetrahedron: Asymmetry 1996, 7, 2189.
. (a) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615. (b) Hajos, Z. G.; Parrish, D. R. Org. Synth. 1985, 63, 26. (c)
Hajos, Z. G.; Parrish, D. R. US Patent 3975440, 1976.
3
4
5
6
7
. Eder, U.; Sauer, G.; Weichert, R. Angew. Chem., Int. Ed. Engl. 1971, 10, 496.
. Buchschacher, P.; Furst, A. Org. Synth. 1985, 63, 37.
. Agami, C.; Meynier, F.; Puchot, C.; Guilhem, J.; Pascard, C. Tetrahedron 1984, 40, 1031.
. Sarkar, A.; Jois, H. R. Y.; Kasthuri, T. R.; Dasgupta, D.; Proc. Indian Acad. Sci., Chem. Sci. 1982, 91, 475.
. (a) NMR spectra were recorded on a JEOL 400 MHz GSX model. (b) IR spectra were recorded on a NICOLET AVATAR
360 FTIR EST with Multi Bounce HATR.
8. Dr. Hajos has drawn our attention to US Patent 3975440, 1976 where the synthesis of S-3 without using solvents was
reported earlier.
9
. Hajos and Parrish reported2a that there was no reaction between trione 1 and 2S-trans-4-hydroxy proline and ruled out any
solid state reaction.
1
1
0. Similar deprotonation by crystalline salts of carboxylic acids occur in the Perkin reaction.
1. Also, 13C NMR studies of reaction between ketones like cyclohexanone, cyclopentanone and ethyl methyl ketone in DMSO
with S-proline did not show any evidence of reaction.
12. (a) Ferran Jr., H. E.; Drake, D. A.; Spencer, T. A. J. Org. Chem. 1975, 40, 2017. (b) Hupe, D. J.; Kendall, M. C. R.;
Spencer, T. A. J. Am. Chem. Soc. 1973, 95, 2271.
13. We can invoke this template mechanism also for the antibody catalyzed enantioselective conversion of trione 2 to S-4
reported recently by Zhong, G.; Hofmann, T.; Lenner, R. A.; Danishefsky, S.; Barbas, C. J. Am. Chem. Soc. 1997, 119,
8131.
1
4. We are thankful to the referee for drawing our attention to the following two references where a similar surface reaction
has been invoked: (a) Shvo, Y.; Gal, M.; Becker, Y.; Elgavi, A. Tetrahedron: Asymmetry 1996, 7, 911. (b) Thoen, J. C.;
Lipton, M. A. Tetrahedron: Asymmetry 1997, 8, 3947.