Michael addition reactions between chiral equivalents of a nucleophilic glycine and (S)- or (R)-3-[(E)-enoyl]-4-phenyl-1,3-oxazolidin-2-ones as a general method for efficient preparation of β-substituted pyroglutamic acids. Case of topographically control

Article abstract of DOI:10.1021/ja0535561

This paper describes a systematic study of addition reactions between the chiral Ni(II) complex of the Schiff base of glycine with (S)-o[N-(N- benzylprolyl)amino]benzophenone and (S)- or (R)-3-[(E)-enoyl]-4-phenyl-1,3- oxazolidin-2-ones as a general and s

Full text of DOI:10.1021/ja0535561

Published on Web 10/11/2005
Michael Addition Reactions between Chiral Equivalents of a
Nucleophilic Glycine and (S)- or
(R)-3-[(E)-Enoyl]-4-phenyl-1,3-oxazolidin-2-ones as a General
Method for Efficient Preparation of â-Substituted
Pyroglutamic Acids. Case of Topographically Controlled
Stereoselectivity
Vadim A. Soloshonok,*,† Chaozhong Cai,‡ Takeshi Yamada,†,§ Hisanori Ueki,†
Yasufumi Ohfune,§ and Victor J. Hruby‡
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Oklahoma,
Norman, Oklahoma 73019, Department of Chemistry, UniVersity of Arizona,
Tucson, Arizona 85721, and Graduate School of Science, Department of Material Science,
Osaka City UniVersity, Sugimoto, Sumiyoshi, Osaka 558-8585, Japan
Received May 31, 2005; E-mail: vadim@ou.edu
Abstract: This paper describes a systematic study of addition reactions between the chiral Ni(II) complex
of the Schiff base of glycine with (S)-o-[N-(N-benzylprolyl)amino]benzophenone and (S)- or (R)-3-[(E)-
enoyl]-4-phenyl-1,3-oxazolidin-2-ones as a general and synthetically efficient approach to â-substituted
pyroglutamic acids and relevant compounds. These reactions were shown to occur at room temperature
in the presence of nonchelating organic bases and, most notably, with very high (>98% diastereomeric
excess (de)) stereoselectivity at both newly formed stereogenic centers. The stereochemical outcome of
the reactions was found to be overwhelmingly controlled by the stereochemical preferences of the Michael
acceptors, and the chirality of the glycine complex influenced only the reaction rate. Thus, in the reactions
of both the (S)-configured Ni(II) complex and the Michael acceptors, the reaction rates were exceptionally
high, allowing preparation of the corresponding products with virtually quantitative (>98%) chemical and
stereochemical yields. In contrast, reactions of the (S)-configured Ni(II) complex and (R)-configured Michael
acceptors proceeded at noticeably lower rates, but the addition products were obtained in high diastereo-
and enantiomeric purity. To rationalize the remarkably high and robust stereoselectivity observed in these
reactions, we consider an enzyme-substrate-like mode of interaction involing a topographical match or
mismatch of two geometric figures. Excellent chemical and stereochemical yields, combined with the
simplicity and operational convenience of the experimental procedures, render the present method of
immediate use for preparing various â-substituted pyroglutamic acids and related compounds.
and various fused azabicyclic derivatives, 3.1,6 Another important
class of compounds available from elaboration of 1 is the
Introduction
Among the natural “chiral pool” molecules available as
starting compounds for their chemical elaboration to more
complex structures, naturally occurring (S)-pyroglutamic acid
(Scheme 1) has enjoyed a status as a scaffold that has been
used in highly efficient syntheses of numerous biologically
relevant compounds.1 Taking into account the wide range of
synthetic applications for pyroglutamic acid, many research
groups have focused on the development of synthetic routes to
it and its derivatives. In particular, asymmetric preparation of
â-substituted analogues, 1 (Scheme 1), has attracted significant
attention.2,3 For example, 1 can be efficiently transformed to
N-bridge bicyclic compounds 2,1,4 pyrrolizidines, indolizidines,1,5
(2) For some representative papers, see: (a) Ezquerra, J.; Pedregal, C.; Merino,
I.; Flo´rez, J.; Barluenga, J.; Garc´ıa-Granda, S.; Llorca, M.-A. J. Org. Chem.
1999, 64, 6554. (b) Seebach, D.; Hoffman, M. Eur. J. Org. Chem. 1998,
1337. (c) Antolini, L.; Forni, A.; Moretti, I.; Prati, F. Tetrahedron:
Asymmetry 1996, 7, 3309. (d) Gestmann, D.; Laurent, A. J.; Laurent, E.
G. J. Fluorine Chem. 1996, 80, 27. (e) Hartzoulakis, B.; Gani, D. J. Chem.
Soc., Perkin Trans. 1 1994, 2525. (f) Suzuki, K.; Seebach, D. Liebigs Ann.
Chem. 1992, 51. (g) Belokon′, Y. N.; Bulychev, A. G.; Pavlov, V. A.;
Fedorova, E. B.; Tsyryapkin, V. A.; Bakhmutov, V. I.; Belikov, V. M. J.
Chem. Soc., Perkin Trans. 1 1988, 2075. (h) El Achqar, A.; Boumzebra,
M.; Roumestant, M.-L.; Viallefont, P. Tetrahedron 1988, 44, 5319. (i)
Pettig, D.; Scho¨llkopf, U. Synthesis 1988, 173. (j) Scho¨llkopf, U.; Pettig,
D.; Schulze, D.; Klinge, M.; Egert, E.; Benecke, B.; Noltemeyer, M. Angew.
Chem., Int. Ed. Engl. 1988, 27, 1194. (k) Fitzi, R.; Seebach, D. Tetrahedron
1988, 44, 5277. (l) Hartwig, W.; Born, L. J. Org. Chem. 1987, 52, 4352.
(m) Minowa, N.; Hirayama, M.; Fukatsu, S. Bull. Chem. Soc. Jpn. 1987,
60, 1761. (n) Belokon′, Y. N.; Bulychev, A. G.; Ryzhov, M. G.; Vitt, S.
V.; Batsanov, A. S.; Struchkov, Y. T.; Bakhmutov, V. I.; Belikov, V. M.
J. Chem. Soc., Perkin Trans. 1 1986, 1865. (o) Scho¨llkopf, U.; Pettig, D.;
Busse, U. Synthesis 1986, 737.
† University of Oklahoma.
‡ University of Arizona.
§ Osaka City University.
(1) For a recent review on various synthetic transformations of pyroglutamic
acid and its derivatives, see: Najera, C.; Yus, M. Tetrahedron: Asymmetry
1999, 10, 2245.
(3) For a review on asymmetric synthesis of â-substituted glutamic/pyro-
glutamic acids, see: Soloshonok, V. A. Curr. Org. Chem. 2002, 6, 341-
364.
9
15296
J. AM. CHEM. SOC. 2005, 127, 15296-15303
10.1021/ja0535561 CCC: $30.25 © 2005 American Chemical Society

Topographically Controlled Stereoselectivity
A R T I C L E S
Scheme 1
â-substituted γ-amino acid derivatives 4, of which Baclofen and
Rolipram1,7 are well-known commercial drugs. Of special
interest to us are transformations of pyroglutamic acid to a
family of sterically constrained â-substituted amino acids 5-111
that serve as indispensable ø-(chi)-constrained8-10 scaffolds in
the de novo design of peptides and peptidomimetics with a
predetermined three-dimensional structure.10
induction at both R- and â-positions of the resultant glutamic
acid derivatives were obtained. Surprisingly, an alternative
strategy, i.e., utilization of chiral derivatives of R,â-unsaturated
carboxylic acids in reactions with achiral glycine equivalents,
remains virtually unexplored.2a
To realize the full outstanding synthetic potential of â-sub-
stituted glutamic/pyroglutamic acids, we set a goal of developing
simple, room-temperature, organic-base-catalyzed asymmetric
Michael addition reactions, featuring a virtually complete
stereochemical outcome under operationally conVenient condi-
tions.11 In literature examples, formation of highly organized
Li-chelated transition states at low temperatures was shown to
be necessary to attain high diastereoselectivity in these addition
reactions.2f,k Therefore, the development of room-temperature,
highly diastereoselective Michael additions using a nonchelating
organic base was a challenging task. Our original idea was that
high diastereoselectivity in these reactions could be obtained
by providing geometric ((E)- or (Z)-enolates only) homogeneity
of the glycine enolate and conformational (s-cis or s-trans only)
homogeneity of the Michael acceptor. This led to a method-
ologic breakthrough12 in the development of room-temperature,
organic-base-catalyzed Michael addition reactions.
Over the last several years, we have reported on the design
and synthesis of the o-aminoacetophenone- and -benzophenone-
derived Ni(II) complexes of glycine 12a,b (Scheme 2) as
chemically stable yet highly reactive nucleophilic glycine
equivalents.13 Their superior synthetic quality over the conven-
tionally used Schiff base 13 for preparation of sterically
constrained R,R-dialkyl-R-amino acids was demonstrated.14
Glycine derivatives 12a,b were found to be particularly useful
in Michael addition reactions with chiral acceptors 14, allowing
preparation of the corresponding â-substituted glutamic/pyro-
glutamic acids, via the intermediate adducts 15, with high
chemical yield and optical purity.15 For instance, use of Schiff
base 13 in addition reactions with 14 was disappointing and
Among the various methods available in the literature,
Michael addition of nucleophilic glycine equivalents to â-sub-
stituted acrylic acid derivatives offers a methodologically concise
and synthetically attractive route to the corresponding â-sub-
stituted pyroglutamic acids 1. The asymmetric version of this
reaction has been the focus of a number of research groups,
and methods to control simultaneous formation of two stereo-
genic centers have been developed.2,3 Common drawbacks of
the literature methods include incomplete (<95% enantiomeric
excess (ee)) stereochemical outcome and chemical yield, the
need for strong bases such as BuLi to generate the corresponding
enolate, and the need to perform the reactions at -78° C, which
diminish the attractiveness of these methods. Analysis of the
relevant literature1-3 reveals that, thus far, only one strategy to
control the stereochemical outcome in these reactions has been
explored. In this approach, addition reactions with chiral glycine
equivalents and R,â-unsaturated carboxylic acid derivatives were
studied, and in some cases, reasonably high levels of asymmetric
(4) (a) Somfia, P.; Ahman, J. Tetrahedron Lett. 1992, 33, 3791. (b) Ahman,
J.; Somfia, P. Tetrahedron 1992, 48, 9537. (c) Melching, K. H.; Hiemstra,
H.; Klaver, W. J.; Speckamp, W. N. Tetrahedron Lett. 1986, 27, 4799.
(5) (a) Provot, O.; Celerier, J. P.; Petit, H.; Lhommet, G. J. Org. Chem. 1992,
57, 2163. (b) Karstens, W. F.; Stol, M.; Rutjes, F. P.; Hiemstra, H. Synlett
1998, 1126.
(6) (a) Wang, W.; Yang, J.; Ying, J.; Xiong, C.; Zhang, J.; Cai, C.; Hruby, V.
J. J. Org. Chem. 2002, 67, 6353. (b) Lim, S. H.; Ma, S.; Beak, P. J. Org.
Chem. 2001, 66, 9056.
(7) (a) Garcia, A. L. L.; Carpes, M. J. S.; de Oca, A. C. B. M.; dos Santos, M.
A. G.; Santana, C. C.; Correia, C. R. D. J. Org. Chem. 2005, 70, 1050-
1053. (b) Chang, M.-Y.; Chen, C.-Y.; Tasi, M.-R.; Tseng, T.-W.; Chang,
N.-C. Synthesis 2004, 840-846. As pointed out by a referee, “according
to the Investigational Drug database (IDdb3) and as of March 24, 2004,
the development of Rolipram was discontinued by Meiji Seika (Japan) and
Schering, A. G. (Germany)”.
(8) For recent reviews on ø-constrained amino acids, see: (a) Gibson, S. E.;
Guillo, N.; Tozer, M. J. Tetrahedron 1999, 55, 585. (b) Hruby, V. J.; Li,
G.; Haskell-Luevano, C.; Shenderovich, M. D. Biopolymers 1997, 43, 219.
(9) For the recent collection of leading papers, see: Asymmetric Synthesis of
Novel Sterically Constrained Amino Acids. Tetrahedron Symposia-in-Print,
#88. Guest Editors: Hruby, V. J., Soloshonok, V. A. Tetrahedron 2001,
57, no. 30.
(11) In the current literature, one can notice a trend for a paradigm of the
synthetic methodology of the future, which is simplicity of experimental
conditions or, as we call it, operationally convenient reaction conditions.
(12) (a) Soloshonok, V. A.; Cai, C.; Hruby, V. J. Tetrahedron Lett. 2000, 41,
135. (b) Soloshonok, V. A.; Cai, C.; Hruby, V. J.; Meervelt, L. V.;
Yamazaki, T. J. Org. Chem. 2000, 20, 6688.
(13) For large-scale synthesis of Ni(II) complexes 12a,b, see: Ueki, H.; Ellis,
T. K.; Martin, C. H.; Soloshonok, V. A. Eur. J. Org. Chem. 2003, 1954.
(14) (a) Ellis, T. K.; Martin, C. H.; Ueki, H.; Soloshonok, V. A. Tetrahedron
Lett. 2003, 4, 1063-1066. (b) Ellis, T. K.; Martin, C. H.; Tsai, G. M.;
Ueki, H.; Soloshonok, V. A. J. Org. Chem. 2003, 68, 6208-6214. (c) Ellis,
T. K.; Hochla, V. M.; Soloshonok, V. A. J. Org. Chem. 2003, 68, 4973-
4976.
(10) (a) Hruby, V. J. Life Sci. 1982, 31, 189. (b) Hruby, V. J.; Al-Obeidi, F.;
Kazmierski, W. M. Biochem. J. 1990, 268, 249. (c) Hruby, V. J.
Biopolymers 1993, 33, 1073. (d) Cai, M.; Cai, C.; Mayorov, A. V.; Xiong,
C.; Cabello, C. M.; Soloshonok, V. A.; Swift, J. R.; Trivedi, D.; Hruby, V.
J. J. Pept. Res. 2004, 63, 116.
9
J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005 15297

A R T I C L E S
Scheme 2
Soloshonok et al.
lead to three out of four possible diastereomeric products in
low chemical yield.16 A systematic study of the addition
reactions between glycine derivatives 12a,b and chiral Michael
acceptors 14 revealed some limitations. For example, reaction
of isopropyl-containing acceptor 14 with compounds 12a,b
proceeded at very low reaction rates, allowing for less than 30%
conversion of the starting materials in 4 h reaction time. It should
be noted, however, that the corresponding addition product 15
was obtained with virtually complete diastereoselectivity.
In the search for an improved and more general method,
addition reactions between chiral Ni(II) complex (S)-16,17
introduced by Belokon,18 and achiral Michael acceptors 17 were
also studied.19,20 Though successful, these reactions showed
relatively low diastereoselectivity. The biggest problem was poor
Re/Si face stereocontrol of the complex 16-derived enolate,
whereas the face selectivity of the N-[(E)-enoyl]oxazolidin-2-
ones 17 was perfect, giving rise only to the products of like
relative topicity.21 On the other hand, exceptionally high reaction
rates were observed for these addition reactions. For instance,
reaction of complex 16 with isopropyl-derived 17 was complete
in less than 45 min, giving rise to a pair of the corresponding
diastereomers, 18 and 19, in excellent chemical yield.20 Com-
paring the results obtained in reactions of isopropyl-derived
Michael acceptors 14 and 17 with complexes 12a,b and 16,
respectively, we found that the combination of 16, a more
reactive glycine derivative, and chiral 4-substituted N-[(E)-enoyl]-
oxazolidin-2-ones 14, as stereocontrolling reagents, may result
in an improved and synthetically efficient procedure.
In this paper,22 we report a systematic study of Michael
addition reactions between enantiomerically pure N-[(E)-enoyl]-
4-phenyl-1,3-oxazolidin-2-ones 1423 and chiral Ni(II) complexes
of glycine Schiff base (S)-16. The complete chemical and optical
yields achieved under the operationally conVenient conditions,
combined with the quantitative recovery of both chiral auxil-
iaries, render this strategy practical, economical, and syntheti-
cally superior to literature methods.2,3
Results and Discussion
Michael Addition Reactions between (S)-Ni(II) Complex
16 and (S)- (R)-N-[(E)-Enoyl]-4-phenyl-1,3-oxazolidin-2-ones,
14. Taking into account the stereochemical preferences of
complex 16 and the Michael acceptors 14, with respect to the
absolute configuration of the R-stereogenic center in the
corresponding Michael adduct,24 we anticipated that the com-
bination of (S)-configured complex 16 with (S)-14 might be a
case of matched stereochemistry. The mismatched case was
expected in reactions of (S)-16 with acceptors, 14, of (R)-
(15) (a) Soloshonok, V. A.; Cai, C.; Hruby, V. J. Org. Lett. 2000, 2, 747. (b)
Soloshonok, V. A.; Cai, C.; Hruby, V. J. Tetrahedron Lett. 2000, 41, 9645.
(c) Soloshonok, V. A.; Ueki, H.; Tiwari, R.; Cai, C.; Hruby, V. J. J. Org.
Chem. 2004, 69, 4984.
(16) Soloshonok, V. A.; Ueki, H.; Ellis, T. K.; Yamada, T.; Ohfune, Y.
Tetrahedron Lett. 2005, 46, 1107.
(17) For large-scale synthesis of Ni(II) complex (S)-16, see: Ueki, H.; Ellis, T.
K.; Martin, C. H.; Bolene, S. B.; Boettiger, T. U.; Soloshonok, V. A. J.
Org. Chem. 2003, 68, 7104.
(18) (a) Belokon, Yu. N. Janssen Chim. Acta 1992, 10, 4. (b) Belokon, Yu. N.
Pure Appl. Chem. 1992, 64, 1917.
(19) (a) Soloshonok, V. A.; Avilov, D. V.; Kukhar′, V. P.; Meervelt, L. V.;
Mischenko, N. Tetrahedron Lett. 1997, 38, 4903. (b) Soloshonok, V. A.;
Cai, C.; Hruby, V. J.; Meervelt, L. V.; Mischenko, N. Tetrahedron 1999,
55, 12031. (c) Soloshonok, V. A.; Cai, C.; Hruby, V. J.; Meervelt, L. V.
Tetrahedron 1999, 55, 12045.
(20) (a) Soloshonok, V. A.; Cai, C.; Hruby, V. J. Tetrahedron: Asymmetry 1999,
10, 4265. (b) Cai, C.; Soloshonok, V. A.; Hruby, V. J. J. Org Chem. 2001,
66, 1339.
(21) Seebach, D.; Prelog, V. Angew. Chem., Int. Ed. Engl. 1982, 21, 654.
(22) For a communication on preliminary results, see: Soloshonok, V. A.; Cai,
C.; Hruby, V. J. Angew. Chem., Int. Ed. Engl. 2000, 39, 2172.
(23) For convenient large-scale synthesis of Michael acceptors 14, see: Solos-
honok, V. A.; Ueki, H.; Jiang, C.; Cai, C.; Hruby, V. J. HelV. Chim. Acta
2002, 85, 3616-3623.
(24) Both (S)-16 and (S)-14 show a strong preference for the (S) configuration
of the R-stereogenic carbon of a newly formed amino acid.
9
15298 J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005

Topographically Controlled Stereoselectivity
A R T I C L E S
Scheme 3
preference for the R-(S) absolute configuration in alkyl halide
alkylation, Michael, and aldol addition products,17-19 this
stereochemical outcome was not anticipated. With these inter-
esting results in hand, we decided to examine next the generality
of these diastereomerically complete addition reactions. The
addition reaction of complex (S)-16 and acceptor (S)-14d,
bearing a bulky isopropyl group, occurred relatively easily (30
min), giving as the sole product (2S,3R)-20d which was isolated
in high chemical yield (entry 3). In contrast, the reaction of
(S)-16 with (R)-14d proceeded at a noticeably slower rate, but
the resulting (2R,3S)-21d25 was obtained as the only diastere-
omeric product (entry 4). Thus, in contrast to our expectations,
both the anticipated matched and mismatched stereochemical
preferences featured virtually complete diastereoselectivity. The
only difference was the reaction rate, which was more pro-
nounced in the additions of the sterically demanding isopropyl-
containing derivatives (R)- and (S)-14d. Taking into account
that the original definition of the concept of matched and
mismatched pairs given by Masamune et al.27 implies a higher
stereoselectivity in the former case and a lower stereoselectivity
in the latter case, we cannot use this terminology to describe
the stereochemical outcomes observed in this study.28
Table 1. Addition Reactions of (S)-16 with (S)- or (R)-14a-ja
reaction
time
products 20a−f and 21a−f
entry
14a
−
f
yieldb
abs configurationc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(S)-a
(R)-a
(S)-d
(R)-d
(S)-f
5 min
10 min
30 min
4 h
8 min
30 min
5 min
35 min
>2 h
30 min
2 h
99
97
97
67d
97
97
97
98
80
97
96
95
98
97
96
(2S,3S)-20a
(2R,3R)-21a
(2S,3R)-20d
(2R,3S)-21d
(2S,3R)-20f
(2R,3S)-21f
(2S,3R)-20j
(2R,3S)-21j
(2R,3S)-21h
(2S,3R)-20h
(2R,3R)-21b
(2R,3R)-21c
(2R,3R)-21e
(2R,3S)-21g
(2R,3S)-21i
(R)-f
(S)-j
Similar patterns of reactivity and stereochemical outcome
were observed in the reactions between complex (S)-16 and a
series of aromatic derivatives 14f,h,j. Thus, reactions of
cinnamic-acid-derived (S)- and (R)-oxazolidin-2-ones 14f with
(S)-16 gave, as the sole products, (2S,3R)-20f and (2R,3S)-21f,
respectively, in virtually complete chemical yields (Table 1,
entries 5 and 6). The reaction rate of the former was substantially
higher than that of the latter. Considering the difference in the
stereochemical requirements for methyl and phenyl groups, it
is interesting to note that the reactions of (S)-16 with (4S)-14a
and (4S)-14f occurred at almost the same rate. Even higher
reaction rates were observed in the addition reaction between
trifluoromethyl-containing (S)-14j and complex (S)-16; the
reaction was complete in 5 min, giving 20j in excellent yield
(entry 7). On the other hand, combination of (S)-16 and (R)-
14j resulted in a slower reaction rate (complete in 35 min) but
gave rise to diastereomerically pure 21j in quantitative yield
(entry 8). Reaction of acceptor (R)-14h, bearing an electron-
donating OMe group, with (S)-16 proceeded substantially
slower, allowing the corresponding product 21h to be obtained
in 80% yield after more than 2 h of reaction (entry 9). The
reaction of (S)-16 with (S)-14h was also slow, but went to
completion in 30 min, giving diastereomerically pure 20h in
excellent yield (entry 10). There was a striking similarity in
the rates of reactions of the isopropyl and p-methoxyphenyl-
containing Michael acceptors 14d and 14h, respectively (entries
3 and 10). Both required 30 min for completion of reaction,
suggesting that the steric bulk of an isopropyl group and the
electron-donating effect of a methoxy group have a similar
deleterious effect on the rates of these Michael addition
reactions.
(R)-j
(R)-h
(S)-h
(R)-b
(R)-c
(R)-e
(R)-g
(R)-i
2 h
5 min
2 h
3 h
a All reactions were run in DMF in the presence of 15 mol % DBU at
ambient temperatures. Ratio of (S)-16/(S)-or (R)-14 is 1:1.05-1.1. b Isolated
yield of diastereo- and enantiomerically pure products. In all cases, only
one stereochemical product (de > 98%) was detected by NMR analysis of
the crude reaction mixtures. c Absolute configuration of the newly formed
stereogenic centers of the products was determined on the basis of chiroptical
properties of the Ni complexes 20a-f and 21a-f, as well as by comparison
of the optical rotation of the corresponding pyroglutamic acids isolated from
20a-f and 21a-f with literature data; see the text also. d Incomplete (70%)
conversion of the starting materials.
configuration. The addition reactions between complex (S)-16
and (S)-14 (Scheme 3, Table 1) were first studied. The reaction
between complex (S)-16 and (S)-N-crotonyl-4-phenyloxazolidin-
2-one (14a), conducted in DMF at ambient temperature (22 °C)
in the presence of 15 mol % DBU, was complete in 5 min and
gave 20a as the only product in quantitative yield (Table 1,
entry 1).
The absolute configuration of the two stereogenic centers of
the glutamic acid residue in 20a was found to be (2S,3S), in
agreement with the previously established pattern of the
diastereomeric preferences of complex (S)-16 and acceptors (S)-
14.15,20 Reactions of (S)-16 with acceptors (R)-14 were expected
to represent a mismatched case, and thus, a synthetically useful
stereochemical outcome was not anticipated. Surprisingly, with
the same reaction conditions, the addition of complex (S)-16 to
(R)-14a occurred with a reaction rate similar to that of (S)-14,
giving rise to 21a with the same yield and diastereoselectivity
(entry 2). The absolute configuration of the two newly created
stereogenic centers in 21a was unexpectedly found to be
(2R,3R). On the basis of previous knowledge of the stereocon-
trolling properties of complex (S)-16, which showed a strong
(25) The (2R,3S) absolute stereochemistry of the isopropyl and aromatic
derivatives is a consequence of the Cahn-Ingold-Prelog priorities (see
ref 26) and is stereochemically equivalent to the (2R,3R) absolute
configuration in the aliphatic series of compounds.
(26) Cahn, R. S.; Ingold, C.; Prelog, V. Angew. Chem., Int. Ed. Engl. 1966, 5,
385.
(27) Masamune, S.; Choy, W.; Petersen, J. S.; Sita, L. R. Angew. Chem., Int.
Ed. Engl. 1985, 24, 1.
(28) The authors are grateful to the referees who pointed this out.
9
J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005 15299

A R T I C L E S
Scheme 4
Soloshonok et al.
With the previous results in hand, one would assume that
the reactions of (S)-16 with (S)-14 could be extended to other
aliphatic and aromatic substrates. To demonstrate the generality
of our method, the additions of complex (S)-16 with a series of
(R)-configured Michael acceptors bearing n-alkyl groups 14b,c,
the bulky â-naphthyl group 14g, the electron-rich indolyl 14i,
and the benzyl group 14e were studied. All reactions were
conducted under the same reaction conditions and all resulted
in a virtually quantitative yield and diastereoselectivity (entries
11-15). It is important to note that previous attempts to carry
out addition reactions between achiral complexes 12a,b and
benzyl- and indolyl-containing derivatives 14e and 14i, respec-
tively, were unsuccessful.15 Therefore, the quite successful
application of chiral complex (S)-16 in the reactions with 14e
and 14i represents a significant proof of the generality of our
room-temperature, organic-base-catalyzed method. The sole
products 21b,c,e,g,i were observed in all cases in the crude
reaction mixtures by NMR analysis. Regarding the reaction
rates, the additions of ethyl and n-propyl derivatives 14b,c were
unexpectedly slow, requiring at least 2 h for completion of
reaction (entries 11 and 12). In contrast, reaction of the benzyl-
containing 14e with complex (S)-16 proceeded at a high reaction
rate, suggesting 14e is more electrophilic compared to 14a-c
(entry 13 vs entries 2, 11, and 12). The bulky â-naphthyl 14g
and the electron-rich indolyl 14i derivatives reacted with
complex (S)-16 slowly, but both gave diastereomerically pure
products 21i,g in excellent yield (entries 14 and 15).
separation was effectively achieved by precipitating (S)-24 as
the hydrochloric salt and (S)-26 remained in the chloroform
solution. Both chiral auxiliaries, (S)-24 and (S)-26, were recycled
in 87-95% yield and were reused for preparing the starting
Ni(II) complex (S)-16 and (4S)-N-[(E)-enoyl]-4-phenyloxazo-
lidin-2-ones 14, using this simple procedure. On the other hand,
the desired enantiomerically pure pyroglutamic acids 25 were
efficiently isolated (93-95% yield) from the aqueous ammonia
solution using DOWEX-50 and were recrystallized from hexane/
THF for analytical purposes. Diastereomers 21, not shown in
Scheme 4, were also converted to their corresponding pyro-
glutamic acids (R)-25, and both chiral auxiliaries, (R)-26 and
(S)-24, were recycled, using this protocol.
Although the above protocol is simple and effective for the
small scale (∼1-5 g), the preparation of amino acids 25 and
its application, on a larger scale, raised concerns regarding the
time-consuming evaporation of aqueous solutions. Thus, after
decomposition of complex 20 with aqueous HCl, the reaction
mixture was treated with 5% NaOH to cyclize intermediate 22
and generate the free base of ligand (S)-24. The precipitate of
the Ni(II) species was removed by filtration, and ligand (S)-24
and oxazolidinone (S)-26 were extracted with CHCl3 from the
filtrate. The aqueous layer was acidified using 3N-HCl and
extracted with AcOEt to give the target â-phenylpyroglutamic
acid (2S,3R)-25f in 89% yield. The modified protocol is more
efficient, but the yield of amino acid 25f is a bit lower compared
to the previous method.
Decomposition of Products 20 and 21, Isolation of Pyro-
glutamic Acid 25, and Recovery of the Chiral Auxiliary 26
and the Ligand 24. With these results, we next sought to find
a simple way to isolate the target pyroglutamic acids from the
addition products 20 and 21, along with complete recovery and
recycling of chiral auxiliary 26 and ligand 24. Following our
standard procedure,15c compounds 20 were decomposed with
2N-hydrochloric acid in MeOH to afford the corresponding
glutamic acid derivatives 22, NiCl2, and the hydrochloric salt
of the ligand (S)-23 (Scheme 4). Treatment of the reaction
mixture with ammonia converted the intermediates 22 into the
target pyroglutamic acids 25 and released the oxazolidinone (S)-
26. At the same time, the hydrochloride salt of ligand (S)-23
was converted into free base 24, allowing extraction (chloro-
form) of both chiral auxiliaries (S)-24 and (S)-26 from the
aqueous ammonia solution. By taking advantage of differences
in the chemical properties of (S)-24 and (S)-26, their further
Michael Addition Reactions between (S)-16 and (S)- or
(R)-14 Using Various Bases and Solvents. The previous
experiments of the addition reactions between (S)-16 and (S)-
or (R)-14 were conducted using exclusively DBU as a base and
DMF as a solvent. Therefore, we decided to investigate these
addition reactions using different bases and solvents with the
goal of further improving and generalizing the method. To this
aim, several reactions were run in different solvents using a
number of organic base catalysts (Table 2). Attempts to use
Et3N or DABCO as a base in DMF were generally unsuccessful.
The application of highly electrophilic acceptors, such as 14j,
in some cases resulted in a reaction with the Et3N or DABCO,
but reaction rates were extremely low, suggesting that these
compounds are not basic enough to be practically useful catalysts
for the addition reactions. In contrast, the guanidine derivative
(2,3,4,6,7,8-hexahydro-1H-pyrimido[1,2-a]pyrimidine) was cata-
lytically active, providing high reaction rates even when present
9
15300 J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005

Topographically Controlled Stereoselectivity
A R T I C L E S
Table 2. Addition Reactions of (S)-16 with (S)- or (R)-14a,f,
Catalyzed by DBU and Guanidine (GUA), Conducted in DMF and
THFa
Scheme 5
reaction conditions and time
entry
14a,f
DBU/DMF
DBU/THF
GUA/DMF
GUA/THF
1
2
3
4
(S)-a
(R)-a
(S)-f
(R)-f
5 min
10 min
8 min
8 h
3 min
3 min
3 min
3 min
13 min
6 min
84 h
24 h
24 h
30 min
a All reactions were run at room temperature. Ratio of 16/14 ) 1.0:1.1,
using 15 mol % of DBU and 5 mol % of guanidine (GUA) (2,3,4,6,7,8-
hexahydro-1H-pyrimido[1,2-a]pyrimidine). In all reactions, complete con-
version of the starting compounds was observed. In all cases, only one
diastereomeric product was detected by 1H NMR of the crude reaction
mixtures. See text also.
the enolate’s oxygen, makes TS C unlikely. However, in TS
A, the only point of unfavorable steric interaction is a close
proximity of the acceptor’s R1 and the benzophenone phenyl
group of the Ni(II) complex, but this interaction can be
minimized by free rotation of the two aromatic rings. Moreover,
in TS A, the Michael acceptor’s carbonyl group and the enolate
oxygen are in close proximity to each other, allowing the
reactions to proceed with the required minimum charge separa-
tion.
In TS A, the phenyl group at C-4 of the chiral oxazolidinone
ring is pointing up, away from any possible steric interactions
with the rest of the substituents on the Ni(II) complex and the
Michael acceptor. We believe that in this position the phenyl
group does not directly control the facial diastereoselectivity
of the Michael acceptor’s CdC double bond by stereodiscrimi-
nation, but works as a topographical feature, giving a difference
in accessibility to the plane of the Michael acceptor by the plane
of the Ni(II) complex. Thus, in the case of the reaction of (S)-
16 with the Michael acceptor of opposite (S)-configuration, TS
A cannot be formed because the position of the phenyl (pointed
down) will interfere with the approaching Ni(II) complex plane.
Instead, the (S)-configured Michael acceptor will allow the
approach of the Ni complex to the opposite Michael acceptor’s
side, leading to products of R-(S) absolute configuration. This
mode of interaction, controlling the absolute configuration of
the products, represents an enzyme-substrate-like topographical
match or mismatch of two geometric figures. Our results
demonstrate that the topographically controlled face selectivity,
realized during the current study, is a much more powerful way
to achieve stereocontrol in asymmetric reactions compared with
the usual stereodiscrimination process involving interactions
between a stereocontrolling element and other substituents on
the reacting molecules. For instance, an obvious advantage of
the topographically controlled TS, evident from the present
study, is the virtually complete diastereoselectivity and the
extraordinary generality of the reactions.
Addition Reactions between Complex (S)-16 and (S)- or
(R)-Configured Michael Acceptors 27-29. To investigate the
generality of the topographically controlled stereoselectivity,
Michael acceptors 27-29 bearing several different chiral
auxiliaries were prepared and used in the addition reactions with
(S)-16, using standard reaction conditions (Scheme 5, Table 3).
Enantiomerically pure 5,5-dimethyl substituted acceptors 27
showed a noticeable difference in reaction rates depending on
their absolute configuration (entries 1 and 2). The reaction of
(S)-configured derivative 27 with (S)-16 occurred at a higher
reaction rate and gave diastereomerically pure product 30 in
excellent chemical yield (entry 1). In contrast, the reaction
between (S)-16 and (R)-configured derivative 27 was slower
(entry 1 vs 2); however, the corresponding product 31 was
Figure 1. Possible transition states for asymmetric Michael additions.
in 5 mol %. As follows from Table 2, the guanidine is a much
better catalyst for these reactions as compared with DBU,
providing higher reaction rates. Furthermore, it can be used in
smaller amounts (5 mol % vs 15 mol %). The use of protic
solvents such as alcohols always gave byproducts involving
nucleophilic ring opening or substitution of the oxazolidinone
moiety. Solvents less polar than DMF, such as THF, were found
to be suitable. THF exhibited no advantage over DMF in
combination with DBU, but did exhibit an advantage in
combination with guanidine. Therefore, the combination of
guanidine with THF is an alternative to the DMF/DBU system.
Because THF can be easily removed by evaporation, its
application as a solvent offers new possibilities for improved
or alternative workup procedures. It is interesting to note that
the stereochemical outcome of these reactions was not influ-
enced either by the bases or by the solvents tested in this study.
Thus, in all of the reactions carried out using a variety of
conditions, only one diastereomeric product was detected in the
crude reaction mixture, suggesting the robust nature of the
stereocontrol. In general, the reactions between (S)-16 and (S)-
or (R)-14 are more stereoselective and cleaner compared to
previously described reactions of picolinic-acid-derived com-
plexes 12a,b and acceptors (S)- or (R)-14.15
Mechanistic Considerations. To account for the observed
stereochemical outcome in these addition reactions, three
possible transition states (TS) A, B, and C (Figure 1), giving
rise only to the products of like10 relative topicity, can be
considered. TS B and C are likely be ruled out because of
repulsive steric interactions and/or nonminimum charge separa-
tion. In TS B, the substituent R1 of the Michael acceptor points
directly into the Ni atom, which might be difficult to accom-
modate considering that reactions with acceptors bearing an
n-alkyl or isoalkyl group and an aromatic ring easily occur. On
the other hand, in TS B, the acceptor’s carbonyl group and the
enolate oxygen are in close proximity to each other, allowing
the reaction to proceed with the required minimum charge
separation.2f In TS C, steric interactions appear to be minimized,
but the position of the acceptor’s carbonyl group, away from
9
J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005 15301

A R T I C L E S
Soloshonok et al.
Table 3. Addition Reactions of (S)-16 with (S)- or (R)-27-29a
Scheme 6
a All reactions were run in DMF in the presence of 15 mol % of DBU
at room temperature. Molar ratio of 16/27-29 ) 1.0:1.1. b Isolated yield
of pure products. c In all cases, only one diastereomer was detected by 1H
NMR of crude reaction mixtures.
obtained with complete stereochemical outcome. Comparing
these results with the reactions of (S)-16 with unsubstituted (S)-
and (R)-14a, one may conclude that the increased steric bulk
in 27 is an unfavorable factor slowing down the reaction rates.
A similar effect was observed in the reaction of (S)-16 with
(R)-28 (entry 3), which proceeded with a substantially lower
rate but gave product 32 with high diastereoselectivity and yield.
The lowest reaction rate in this series was observed in the
addition of (S)-16 with the N-methyl substituted (4R,5S)-29
leading to the formation of a diastereomerically pure product
33 (entry 4). In this case, the slow rate of addition can be
explained considering electronic factors because substitution of
the oxazolidinone oxygen with an N-methyl group renders the
whole Michael acceptor (4S,5R)-29 less electrophilic and
therefore less reactive. In general, this study showed that
Michael acceptors 27-29 are less synthetically useful in the
additions with (S)-16, requiring longer reaction times compared
to reactions with (S)- or (R)-14. On the other hand, data
demonstrate that regardless of the structure of the chiral auxiliary
on the Michael acceptor, all reactions proceed with perfect
stereoselectivity, furnishing only one of four theoretically
possible diastereomeric products. These results directly support
our argument regarding the topographical nature of the observed
stereocontrol in these Michael addition reactions.
Michael Addition Reactions of (S)-16 with Acceptors 14a,f
in the Presence of i-PrOD. Highly Stereoselective γ-Deu-
teration. Additional support for our mechanistic rationale comes
from the reactions of complexes 12b and (S)-16 with Michael
acceptors 14a,f conducted in the presence of i-PrOD. Thus, as
suggested by TS A (Figure 1), high levels of stereoselectivity
in γ-protonation/deuteration of the corresponding intermediate
adducts should be expected because the plane of the Ni(II)
complex could completely block one face of the enolate. To
test this assumption, reactions of achiral 12b and chiral (S)-16
complexes with (S)- and (R)-configured crotonyl- and cinnamyl-
derived acceptors 14a,f, were carried out in the presence of
i-PrOD using standard conditions (Scheme 6). In agreement with
predictions, all four reactions produced diastereomerically pure
products with completely stereoselective deuteration in the R-
and γ-positions of the corresponding glutamic acid residues in
34-36. On the basis of previous results,12 complexes 12b and
(S)-16 likely first undergo deuteration of their glycine residues
and then react with the acceptors 14a,f, giving rise to the
R-deuterated products 34-36. However, the highly diastereo-
selective deuteration of the γ-position of glutamic acid residues
in 34-36 is truly remarkable and offers a unique opportunity
for preparing stereochemically defined, selectively γ-isotopically
labeled glutamic/pyroglutamic acids, compounds of critical
importance in biological/enzymatic studies.29
Conclusions
Michael addition reactions between chiral Ni(II) complex (S)-
16 and chiral (S)- or (R)-oxazolidin-2-one derivatives, 14, offer
a synthetically powerful and efficient approach to optically pure
â-substituted pyroglutamic acids. Combined with previous
results, the data reported in this study reveal some important
features regarding the reactivity and nature of the stereocontrol
in these Michael addition reactions. First, chiral Ni(II) complex
(S)-16, containing the metal coordinated to an sp3 nitrogen, has
a higher reactivity than achiral complexes 12a,b with an sp2
nitrogen. Second, combination of (S)-configured Michael ac-
ceptor 14 and (S)-16 allows for the corresponding addition
reaction to proceed with high reaction rates, whereas the
reactions of (S)-16 with (R)-14 occur with lower rates. Third,
in both cases, the stereochemical outcome was virtually
complete, indicating that the stereocontrolling power of Michael
acceptors 14 overwhelms the stereochemical preferences of
glycine derivative 16. Fourth, the perfect diastereoselectivity
obtained in these addition reactions is not influenced by the
nature of substituents on the chiral oxazolidinone moiety,
solvent, or base used, suggesting an unusual mode of stereo-
control which we call topographic by analogy to the mechanism
of stereoselective enzyme-catalyzed reactions (lock and key).
Fifth, when these addition reactions were conducted in the
presence of i-PrOD, completely stereoselective R- and γ-deu-
teration was observed providing an opportunity for preparation
of isotopically labeled glutamic/pyroglutamic acids, which have
critical importance in biological/enzymatic studies.
(29) See, for instance: (a) Oez, G.; Berkich, D. A.; Henry, P.-G.; Xu, Y.;
LaNoue, K.; Hutson, S. M.; Gruetter, R. J. Neurosci. 2004, 24, 11273. (b)
Jongen-Relo, A. L.; Amaral, D. G. J. Neurosci. Methods 2000, 101, 9.
9
15302 J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005

Topographically Controlled Stereoselectivity
A R T I C L E S
a glass bar to initiate crystallization of the product. The crystalline
product was filtered off, thoroughly washed with water, and dried in
vacuo to afford addition products 20 or 21. Yields of the products were
given in Tables 1 and 2, and their melting points, spectral, and chiroptic
data are listed in the Supporting Information.
In summary, we have demonstrated that the new strategy for
controlling the stereochemical outcome of the asymmetric
Michael addition reactions developed in this work is method-
ologically superior to previous methods, most notably in terms
of the generality and synthetic efficiency. Excellent chemical
yields and diastereoselectivities, combined with the operational
conVenience of simple experimental procedures, render the
present method of immediate use for preparing a variety of
3-substituted pyroglutamic acids, related amino acids, and
biologically relevant compounds available via conventional
transformations of the pyroglutamic acids.
Decomposition of Complexes 20 and 21, Isolation of (2S,3S)-3-
Alkyl- and (2S,3R)-3-Arylpyroglutamic Acid or (2R,3R)-3-Alkyl-
and (2R,3S)-3-Arylpyroglutamic Acid 25, and Recovery of Ligand
(S)-24 and Starting Chiral Auxiliary (S)- or (R)-26. A solution of
diastereomerically pure complex 20 or 21 (15 mmol) in MeOH (50
mL) was slowly added with stirring to a mixture of aqueous 3N-HCl
and MeOH (90 mL, ratio of 1:1) at 70 °C. Upon disappearance of the
red color of the starting complex, the reaction mixture was evaporated
in vacuo to dryness. Water (80 mL) was added, and the resultant mixture
was treated with an excess of concentrated NH4OH and extracted with
CHCl3. The CHCl3 extracts were dried over MgSO4 and evaporated in
vacuo to afford 6.0 g of a 1:1 mixture (99%) of 25 and chiral auxiliary
(S)- or (R)-26. The aqueous solution was evaporated in vacuo, dissolved
in a minimum amount of water, and loaded on a cation-exchange resin
Dowex 50 × 2 100 column. The column was washed with water, and
the acidic fraction was collected to give, after evaporation in vacuo,
pyroglutamic acid 25. An analytically pure sample of the product was
obtained by crystallization of the compound from THF/n-hexane. Yields
and spectral and chiroptic data for compounds 25 are listed in the
Supporting Information.
Experimental Section
General. 1H, 13C, and 19F NMRs were performed on Varian Unity-
300 (299.94 MHz) and Gemini-200 (199.98 MHz) spectrometers using
TMS, CDCl3, and CCl3F as internal standards. High-resolution mass
spectra (HRMS) were recorded on a JEOL HX110A instrument. Optical
rotations were measured on a JASCO P-1010 polarimeter. Melting
points (mp) are uncorrected and were obtained in open capillaries. All
reagents and solvents, unless otherwise stated, are commercially
available and were used as received. Synthesis of the Ni(II) complex
of the Schiff base of (S)-BPB and glycine (S)-16 was accomplished by
the procedure given in ref 17. Chiral (S)- or (R)-N-[(E)-enoyl]-4-phenyl-
1,3-oxazoline-2-ones 14a-j were prepared according to the general
method given in ref 23. Unless otherwise stated, yields refer to isolated
yields of products of greater than 95% purity, as estimated by 1H, 19F,
and 13C NMR spectrometry. All new compounds were characterized
Acknowledgment. The work was supported by the Depart-
ment of Chemistry and Biochemistry, University of Oklahoma,
and by the grants from U.S. Public Health Service Grants DA
06284, DA 04248, and DK 17420. The views expressed are
those of the authors and not necessarily those of the USPHS.
V.A.S. thanks Professor P. F. Cook for proofreading the final
manuscript.
1
by H, 19F, 13C NMR, and HRMS.
General Procedure for the Reactions of the Glycine Complex
(S)-16 with (S)- or (R)-N-[(E)-enoyl]-4-phenyl-3-oxazoline-2-ones
14a-j. To a suspension of complex (S)-16 (0.25 g, 0.50 mmol) in DMF
(1.5 mL) was added (S)- or (R)-N-[(E)-enoyl]-5-phenyl-3-oxazoline-
2-ones 14 (0.53 mmol) with stirring. The mixture was stirred at room
temperature for 10-15 min to get a homogeneous solution and then
DBU (0.011 g, 0.072 mmol) was added dropwise. The course of the
reaction was monitored by thin-layer chromatography (TLC) (SiO2).
Each sample was quenched with 5% aqueous acetic acid, and the
products were extracted with chloroform before being applied to the
plate. Upon disappearance of the starting (S)-1, the reaction mixture
was poured into icy 5% aqueous acetic acid (80 mL) and stirred with
Supporting Information Available: General experimental
methods, 1H NMR, and proton-decoupled 13C NMR spectra of
new compounds (22 pages). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0535561
9
J. AM. CHEM. SOC. VOL. 127, NO. 43, 2005 15303