A multistage, one-pot procedure mediated by a single catalyst: A new approach to the catalytic asymmetric synthesis of β-amino acids

Article abstract of DOI:10.1021/jo034150e

A catalytic asymmetric procedure for the preparation of β-amino acids (specifically β-substituted aspartic acid derivatives) is reported. The cinchona alkaloid catalyst benzoylquinine (BQ) mediates up to five distinct steps of a reaction pathway, all in o

Full text of DOI:10.1021/jo034150e

A Mu ltista ge, On e-P ot P r oced u r e Med ia ted by a Sin gle Ca ta lyst:
A New Ap p r oa ch to th e Ca ta lytic Asym m etr ic Syn th esis of
â-Am in o Acid s
Ahmed M. Hafez, Travis Dudding, Ty R. Wagerle, Meha H. Shah, Andrew E. Taggi, and
Thomas Lectka*
Department of Chemistry, J ohns Hopkins University, 3400 North Charles Street,
Baltimore, Maryland 21218
lectka@jhu.edu
Received February 3, 2003
A catalytic asymmetric procedure for the preparation of â-amino acids (specifically â-substituted
aspartic acid derivatives) is reported. The cinchona alkaloid catalyst benzoylquinine (BQ) mediates
up to five distinct steps of a reaction pathway, all in one reaction vessel. The products of this reaction,
highly optically enriched â-substituted aspartic acid derivatives, were prepared from N-acyl-R-
chloroglycine esters and acid chlorides in the presence of the catalyst. This approach was also
amenable to the synthesis of small polypeptides containing â-substituted aspartic acid units,
including a non-natural fragment of the antibiotic lysobactin. The addition of Lewis acids to this
system was found to accelerate the rate of specific steps in the reaction pathway. Mechanistic aspects
of this reaction, such as imine formation and Lewis acid chelation to the â-lactam intermediate,
were investigated through comparison of IR, NMR, and other physical data.
In tr od u ction
ties.5 To date, there are several preparative methods for
the catalytic asymmetric synthesis of â-amino acids.6
J acobsen7 and Miller8 have exploited the sequential
azidation/reduction of R,â-unsaturated carbonyl com-
pounds to this end. A different approach recently reported
by Boger utilizes the Sharpless asymmetric amino hy-
droxylation of cinnamate esters.9
In this paper, we discuss results that have broadened
the overall scope of our procedure for â-amino acid
synthesis.10 A noteworthy aspect of this approach in-
cludes the use of the organic catalyst benzoylquinine (BQ)
that is capable of performing multiple catalytic roles in
a single reaction flask. This catalyst carries out a series
of chemical transformations, in essence a pathway of
reactions. This method relies on the use of readily
available and stable reagents to produce optically en-
riched â-lactams or â-substituted aspartic acid deriva-
tives. We also examine mechanistic aspects of this
reaction with the study of N-acylimine formation from
R-chloroglycine derivatives and Lewis acid chelation to
Enzymes represent some of the most sophisticated and
elegant machinery known for carrying out selective and
highly specific chemical reactions. However, concurrent
with this specificity comes the associated cost of complex-
ity that can severely hamper their broader use as organic
reagents. The use of small organic molecules as highly
effective catalysts is being realized more frequently in
organic synthesis.1 However, a fundamental limitation
prevalent with the use of small organic catalysts is that
they usually perform only one chemical transformation
in the overall reaction sequence. From a synthetic
standpoint, it would be ideal if a single catalyst were
capable of mediating a series of reactions in one reaction
vessel. A biological analogy to such a process would be
the use of a single enzyme (or enzyme complex) to carry
out multiple steps in a biochemical pathway.
Substituted aspartic acid derivatives are currently of
interest medicinally for use as inhibitors of L-asparagine
synthetase,2 key constituents of several proteins involved
in the blood-clotting cascade,3 and for their ability to act
as nontransportable glutamate transporter blockers.4
Additionally, â-amino acids are also used as R-amino acid
surrogates for the construction of peptides possessing
interesting folding patterns and conformational proper-
(5) North, M. J . Peptide Sci. 2000, 6, 301-313.
(6) (a) Enantioselective Synthesis of â-Amino Acids; J uaristi, E., Ed.;
Wiley-VCH: New York, 1997. (b) Rzasa, R. M.; Shea, H. A.; Romo, D.
J . Am. Chem. Soc. 1998, 120, 591-592. (c) Cole, D. C. Tetrahedron
1994, 50, 9517-9582. (d) Kobayashi, S.; Nagayama, S. J . Am. Chem.
Soc. 1997, 119, 10049-10053. (e) Ferraris, D.; Young, B.; Dudding,
T.; Lectka, T. J . Am. Chem. Soc. 1998, 120, 4548-4549. (f) Sibi, M. P.;
Shay, J . J .; Liu, M.; J asperse, C. P. J . Am. Chem. Soc. 1998, 120, 6615-
6616. (g) Liu, M.; Sibi, M. P. Tetrahedron 2002, 58, 7991-8035.
(7) Myers, J . K.; J acobsen, E. N. J . Am. Chem. Soc. 1999, 121, 8959-
8960.
(1) For a comprehensive review, see: Dalko, P. I.; Moisan, L. Angew.
Chem., Int. Ed. 2001, 40, 3726-3748.
(2) Mokotoff, M.; Bagaglio, J . F.; Parikh, B. S. J . Med. Chem. 1975,
18, 354-358.
(3) (a) Hansson, T. G.; Kihlberg, J . O. J . Org. Chem. 1986, 51, 4490-
4492. (b) Wagner, R.; Tilley, J . W. J . Org. Chem. 1990, 55, 6289-6291.
(c) Wagner, R.; Tilley, J . W.; Lovey, K. Synthesis 1990, 785-786.
(4) Shimamoto, K.; Shigeri, Y.; Yasuda-Kamatani, Y.; Lebrun, B.;
Yumoto, N.; Nakajima, T. Bioorg. Med. Chem. Lett. 2000, 10, 2407-
2410.
(8) (a) Horstmann, T. E.; Guerin, D. J .; Miller, S. J . Angew. Chem.,
Int. Ed. 2000, 39, 3635-3638. (b) Guerin, D. J .; Miller, S. J . J . Am.
Chem. Soc. 2002, 124, 2134-2136.
(9) Boger, D. L.; Lee, R. J .; Bounaud, P.-Y.; Meier, P. J . Org. Chem.
2000, 65, 6770-6772.
(10) Dudding, T.; Hafez, A. M.; Taggi, A. E.; Wagerle, T. R.; Lectka,
T. Org. Lett. 2002, 4, 387-390.
10.1021/jo034150e CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/21/2003
J . Org. Chem. 2003, 68, 5819-5825
5819

Hafez et al.
SCHEME 1. Rea ction P a th w a y for â-Am in o Acid
Syn th esis
SCHEME 2. BQ-Ca ta lyzed Deh yd r oh a logen a tion
of r-Ch lor oglycin e 2a
metric catalyst for the [2 + 2] cycloaddition between the
ketene and imine (role 3, step 3), a nucleophilic catalyst
for ring opening (role 4, step 4), and a transesterification
catalyst for ester exchange (role 5, step 5).
Initial experiments established that addition of R-chlo-
roglycine 2a to a solution of Proton Sponge 4 produces
no reaction at room temperature. However, we found that
in the presence of a catalytic amount of BQ (or other
tertiary amines), 2a underwent dehydrohalogenation to
afford the intermediate imine 8a at room temperature
(Scheme 2). We believe that after dehydrohalogenation
by BQ, the proton is relayed to the stoichiometric base,
Proton Sponge. We have previously documented an
analogous “shuttle” deprotonation system in the synthe-
sis of reactive ketenes.12 We also noted that simply
mixing 1 equiv of Proton Sponge and acid chloride at low
temperature does not produce detectable amounts of
ketene. BQ serves as an excellent shuttle base when
added to various solutions of acid chlorides in toluene at
low temperature, allowing for ketene formation.13
â-lactams to facilitate ring opening. Furthermore, we
utilize our methodology during the course of the synthesis
of a fragment of the antibiotic lysobactin.
Resu lts a n d Discu ssion
Th e F ive Ca t a lyt ic R oles of Ben zoylq u in in e.
Mechanistically, our synthesis of â-amino acids is unique
in that the chiral cinchona alkaloid derivative BQ 3a
catalyzes up to five distinct sequential steps in one
reaction flask (eq 1). To date, approaches based on the
In addition to the role of BQ in the stereoselective
cycloaddition to form â-lactams, we found that in the
presence of methanol, BQ greatly enhanced the rate of
â-lactam ring opening (Scheme 3). Even at elevated
temperatures, a large rate difference was observed
between the BQ-catalyzed and uncatalyzed methanolysis
reactions.14
To ensure that the catalyst BQ facilitated the ring
opening and transesterification steps of this reaction, a
control experiment was conducted using preformed â-lac-
tam 9a (R ) Ph). The results of this experiment con-
firmed that BQ was indeed an active catalyst for this
use of one specific catalyst capable of performing multiple
tasks in a reaction sequence remain limited.11 In this
reaction scenario (Scheme 1), BQ acts as a dehydroha-
logenating agent for the synthesis of ketene 7 (role 1, step
1) and N-acylimine 8 (role 2, step 2) from the acid chloride
1 and N-acyl-R-chloroamine 2, respectively, an asym-
(11) Some examples of catalysts that perform at least two roles in
a reaction sequence: (a) Annunziata, R.; Benaglia, M.; Cinquini, M.;
Cozzi, F. Eur. J . Org. Chem. 2001, 1045-1048. (b) Yu, H.-B.; Hu, Q.-
S.; Pu, L. J . Am. Chem. Soc. 2000, 122, 6500-6501. (c) Schenning, A.
P. H. J .; Lutje Spelberg, J . H.; Hurbert, D. H. W.; Feiters, M. C.; Nolte,
R. J . M. Chem. Eur. J . 1998, 4, 871-880. (d) Sibi, M. P.; Cook, G. R.;
Liu, P. Tetrahedron Lett. 1999, 40, 2477-2480. (e) Hamashima, Y.;
Kanai, M.; Shibasaki, M. J . Am. Chem. Soc. 2000, 122, 7412-7413.
(f) Bolm, C.; Dinter, C. L.; Seger, A.; Hocker, H.; Brozio, J . J . Org.
Chem. 1999, 64, 5730-5731. (g) Shibasaki, M. Chemtracts: Org. Chem.
1999, 12, 979-998. (h) Sanford, M. S.; Love, J . A.; Grubbs, R. H. J .
Am. Chem. Soc. 2001, 123, 6543-6554.
(12) (a) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Drury, W.
J ., III; Lectka, T. J . Am. Chem. Soc. 2000, 122, 7831-7832. (b) Taggi,
A. E.; Hafez, A. M.; Wack, H.; Young, B.; Ferraris, D.; Lectka, T. J .
Am. Chem. Soc. 2002, 124, 6626-6635.
(13) The pKa of Proton Sponge is 12, while that of BQ is not known,
but probably similar to that of quinuclidine, which is 11.
(14) The uncatalyzed methanolysis took 48 h at reflux while the
catalyzed reaction was complete in 14 h.
5820 J . Org. Chem., Vol. 68, No. 15, 2003

Catalytic Asymmetric Synthesis of â-Amino Acids
TABLE 1. On e-P ot, Mu lticom p on en t r,â-Am in o Acid
Syn th esis Usin g MeOH a s th e Rin g-Op en in g Nu cleop h ile
SCHEME 3. BQ-Ca ta lyzed Rin g-Op en in g
Meth a n olysis
process based on the notable rate depression in its
absence even at elevated temperatures (14 h with BQ
versus 48 h with no catalyst). In a second control
experiment, polymer-supported BQ 3c15 was utilized in
the reaction to determine whether Proton Sponge or
byproduct salts were responsible for the observed in-
creased rate of methanolysis and transesterification.
Removal of the polymer supported BQ prior to addition
of methanol resulted in a reduced rate of ring opening
and transesterification relative to when BQ was present.
It was also noted that the rate of methanolysis (12 h) in
the presence of polymer-supported BQ was comparable
to that of the free, unbound BQ 3a .
Ca ta lytic, Asym m etr ic Syn th esis of â-Am in o Ac-
id s. We screened a number of acid chlorides with N-acyl-
R-chloroglycine 2a (Table 1) to furnish â-amino acids in
a one-pot procedure. For example, a round-bottom flask
was charged with R-chloroglycine 2a , Proton Sponge 4,
and 10 mol % catalyst 3a in toluene. The resulting
suspension was stirred vigorously for 1 h and then diluted
with toluene and cooled to -78 °C. A solution of pheny-
lacetyl chloride 1a in toluene was then added. The
reaction was stirred for 6 h while gradually warming to
room temperature. Methanol was added, and the mixture
was heated at reflux for 14 h. Purification by column
chromatography afforded 5a in 62% yield (95% ee, 12/1
dr). Several different acid chlorides were used, including
â-O-substituted acid chlorides such as 1b and 1e. Fur-
thermore, with benzoylquinidine (BQd) 3b, the pseudo-
enantiomer of 3a , the opposite enantiomer is obtained
in similar yields and selectivities (entry 2, Table 1).
We have also applied this procedure to the preparation
of tripeptides. Using our methodology, N-acyl-R-chloro-
amine 2b was allowed to react with phenylacetyl chloride
1a and converted into the tripeptide 10 by the addition
of 2 equiv of glycine methyl ester in 44% yield with 95%
ee and 8/1 dr (eq 2). A control experiment established
that, unlike methanolysis, the aminolysis reaction by
glycine methyl ester showed no apparent rate dependence
with respect to BQ. The stronger nucleophilic character
of glycine is apparently the decisive factor in this
particular case, allowing it to compete effectively with
the BQ in the ring opening and transesterification steps
of the reaction (Table 2).
a
Isolated yield of product after column chromatography. b Enan-
tiomeric excess was determined by chiral HPLC. c Diastereomeric
ratio was determined by crude 1H NMR. Opposite enantiomer
d
was obtained by using benzoylquinidine (BQd) 3b.
An Exam in ation of Im in e For m ation fr om N-Acyl-
r-ch lor oglycin e Der iva tives. To expand the general
scope of this reaction, we have also carried out a
systematic study of alternative bases and experimental
methods for the preparation of N-acylimine 8a from the
parent R-chloroglycine 2a.16 Experimentally, N-acylimines
represent useful precursors for the preparation of N-
acylamino acids and â-lactams. Synthetic routes for the
preparation of these products commonly involve the
initial formation of a protected amine followed by sub-
sequent deprotection and acylation to afford the desired
N-acylated products. Although a more direct method
would be the use of N-acylimines, procedures for their
formation are rare due to their chemical instability.
We have found that the use of stoichiometric amounts
of sodium hydride17 offers a particularly attractive alter-
native for the formation of N-acylimine 8a from R-chlo-
roglycine 2a .18 The generation of the environmentally
benign byproduct sodium chloride as opposed to poten-
tially harmful ammonium salts, such as Proton Sponge
hydrochloride, further highlights the value of this ap-
proach.19 When sodium hydride was used as the stoichio-
metric base, comparable yields and selectivities (Table
(15) Hafez, A. M.; Taggi, A. E.; Wack, H.; Drury, W. J ., III; Lectka,
T. Org. Lett. 2000, 2, 3963-3965.
(16) N-Benzoyl-R-chloroglycine derivative 2a is easily prepared by
the condensation of benzamide with ethyl glyoxlate followed by reaction
with oxalyl chloride. See the Supporting Information.
(17) Dry NaH (95%) is utilized in the reaction.
(18) Hafez, A. M.; Taggi, A. E.; Dudding, T.; Lectka, T. J . Am. Chem.
Soc. 2001, 123, 10853-10859.
J . Org. Chem, Vol. 68, No. 15, 2003 5821

Hafez et al.
TABLE 2. On e-P ot, Mu lticom p on en t r,â-Am in o Acid Syn th esis Usin g Am in es a s th e Rin g-Op en in g Nu cleop h ile
a
b
Isolated yield of product after column chromatography. Enantiomeric excess was determined by chiral HPLC. c Diastereomeric ratio
was determined by crude 1H NMR.
TABLE 3. r,â-Am in o Acid Syn th esis Usin g Sod iu m
Hyd r id e a s a Ba se for N-Acylim in e F or m a tion
that the optimal base for dehydrohalogenation of the
parent R-chloroglycine 2a was a polymer-supported pi-
peridine. Using this polymer resin for imine formation,
followed by transfer of a filtered solution of the N-
acylimine 8a to a solution of preformed phenylketene 7a
and BQ 3a , the desired â-amino acid 5a was obtained in
52% yield with 94% ee and 12/1 dr. Previously, we have
had success with the use of the highly basic polymer
supported base BEMP as a dehydrohalogenating agent
for the generation of ketenes from acid chlorides.17 In
contrast, none of the desired dehydrohalogenated N-
acylimine product was formed upon exposure of the
R-chloroglycine 2a to BEMP resin. Instead, a series of
degradation products were isolated from this generation
method. Use of a less basic guanidine resin initially
appeared more promising, but after repeated attempts
to optimize the reaction conditions, isolated product
yields remained low (13%).
Isolated yield of product after column chromatography. b Enan-
a
tiomeric excess was determined by chiral HPLC. c Diastereomeric
ratio was determined by crude 1H NMR.
Previously, we reported that carbonate salts, in the
presence of a catalytic quantity of BQ, function as
effective stoichiometric bases for the dehydrohalogenation
of acid chlorides via a proton shuttle mechanism.23
Extrapolation of this work to R-chloroglycine 2a proved
to be less rewarding. This method involved stirring a
solution of R-chloroglycine 2a in toluene for 8 h with an
excess of K2CO3 and 10 mol % BQ. The carbonate salts
were filtered off and the filtrate added to a solution of
preformed phenylketene 7a at -78 °C. Subsequent
methanolysis, workup, and isolation of the desired prod-
uct afforded insignificant quantities of the substituted
â-amino acid product 5a . The ineffectiveness of this last
approach is likely due to decomposition of the N-
acylimine intermediate during the extended reaction time
with K2CO3.
3) were obtained using the one-pot procedure. Drawing
upon the work of Kobayashi,20 we have also examined
the use of polymer-bound bases as dehydrohalogenating
reagents for N-acylimine formation. Kobayashi et al. have
recently reported an asymmetric Mannich-type reaction
utilizing a chiral copper catalyst for the preparation of
N-acylated amino acids from N-acyl-R-chloroamines and
enol silanes.21
The value of polymer-supported systems resides in the
ease of product separation, a critical factor with respect
to the chemistry and preparation of reactive N-acylimines.
A survey of the commercially available resin-bound bases
BEMP,22 guanidine, and piperidine quickly established
(19) Taggi, A. E.; Wack, H.; Hafez, A. M.; France, S.; Lectka, T. Org.
Lett. 2002, 4, 627-629.
(20) Kobayashi, S.; Kitagawa, H.; Matsubara, R. J . Comb. Chem.
2001, 3, 401-403.
(22) Schwesinger, R.; Willaredt, J .; Schlemper, H.; Keller, M.;
Schmitt, D.; Fritz, H. Chem. Ber. 1994, 127, 2435-2454.
(23) Hafez, A. M.; Taggi, A. E.; Wack, H.; Esterbrook, J .; Lectka, T.
Org. Lett. 2001, 3, 2049-2051.
(21) Kobayashi, S.; Matsubara, R.; Kitagawa, H. Org. Lett. 2002, 4,
143-145.
5822 J . Org. Chem., Vol. 68, No. 15, 2003

Catalytic Asymmetric Synthesis of â-Amino Acids
To elucidate the mechanistic role of BQ 3a and Proton
Sponge 4 as dehydrohalogenating agents of R-chlorogly-
1
cine 2a the reaction was monitored by H NMR.24 The
formation of imine 8a from R-chloroglycine 2a was
carried out and monitored by the periodic addition of
aliquots of Proton Sponge and BQ to an NMR sample.
Over the course of addition, a diagnostic singlet at 7.92
ppm corresponding to the imine CH proton was observed.
This experiment verified that the formation of imine 8a
occurs under the reaction conditions. Unfortunately, the
presence of byproduct salts and BQ made a more quan-
titative assessment of imine formation difficult. To ad-
dress this issue, N-4-fluorobenzoyl-R-chloroglycine de-
rivative 2c was used in a 19F NMR study to observe the
relative rate of imine formation.25 The starting N-4-
fluorobenzoyl-R-chloroglycine 2c gave one signal at 24.42
ppm (eq 3). The addition of Proton Sponge (2 equiv) did
not result in any noticeable change in the 19F NMR
spectrum, thus ruling out the role of Proton Sponge as a
dehydrohalogenating agent. However, upon addition of
BQ (0.1 equiv) to the NMR sample, the peak at 24.42
ppm began to slowly diminish. Concurrent with the
disappearance of the singlet at 24.42 ppm was the
appearance of a new singlet at 26.68 ppm that we
attribute to the formation of N-4-fluorobenzoylimine 8c
(eq 3).26
F IGURE 1. Carbonyl stretching frequencies in the presence
of Cu(OTf)2.
TABLE 4. Ra tes of â-La cta m 9a Rin g Op en in g by MeOH
in th e P r esen ce of BQ a n d Lew is Acid s
catalyst
conversion timea (h)
relative rate
Cu(OTf)2 + BQ
In(OTf)3 + BQ
Zn(OTf)2 + BQ
Sc(OTf)3 + BQ
Cu(OTf)2
1
1
1.5
5
14.0
14.0
9.3
2.8
1.7
8
Sc(OTf)3
BQ
none
10
14
48
1.4
1.0
0.3
a
Conversion times were monitored by TLC and confirmed by
1H NMR.
tams.29 Intrigued by the possibility of utilizing a bifunc-
tional catalyst system for the enhancement of the BQ-
mediated alcoholysis of â-lactams, we investigated the
effect of adding metals to this reaction. As proposed, the
addition of catalytic amounts of metals dramatically
altered the rate of â-lactam ring opening. Our survey of
Lewis acids revealed that metal triflate salts greatly
enhanced the rate of ring opening. For example, addition
of In(OTf)3 (10 mol %) to a solution of â-lactam 9a and
10 mol % BQ in MeOH, at reflux, resulted in the rapid
conversion of â-lactam 9a to the substituted â-amino acid
5a in under 1 h (eq 4).
Under these experimental conditions, the N-4-fluo-
robenzoyl-R-chloroglycine 2c had been totally consumed,
and the formation of imine 8c was complete within 0.5
h. Further study established that the reaction was not
sensitive to the concentration of Proton Sponge and the
overall rate of reaction was unaffected by the addition of
excess Proton Sponge, thus supporting the role of BQ as
the dehydrohalogenating agent.
The comparable control reaction without the addition
of a Lewis acid but in the presence of BQ took over 14 h
to reach completion. Table 4 outlines a comparison of
relative rates obtained for the methanolysis of â-lactam
9a using different metal triflates. On the basis of these
results, metal salts such as In(III) and the late transition
metals Cu(II) and Zn(II) appear to be the Lewis acids of
choice for facilitating â-lactam ring opening.
Lew is Acid -Ca ta lyzed Rin g Op en in g of â-La c-
ta m s. The class B â-lactamase metalloenzymes, such as
the zinc-dependent enzyme isolated from Bacillus cereus,
are known to possess a broad substrate profile and to
readily hydrolyze most classes of â-lactams.27 Within
these enzymatic systems, a metal center serves as an
activator of the â-lactam carbonyl working in concert with
a proximate water (or metal-bound hydroxide) molecule
to facilitate ring cleavage.28
To garner further support for the catalytic role of the
Lewis acid in this reaction, an IR experiment was
conducted to determine the interaction between the
â-lactam and Lewis acid (Figure 1). Upon addition of a
stoichiometric amount of Cu(OTf)2 to a solution of N-acyl-
â-lactam 9a , a shift in the â-lactam carbonyl stretching
frequency from 1748 to 1746 cm-1 was observed. Ad-
ditional shifts attributed to the N-acyl group carbonyl
We have recently reported a bifunctional Lewis acid-
Lewis base catalyst system for the synthesis of â-lac-
(24) The 1H NMR spectra were acquired on a Varian Unity Plus
400 MHz instrument in CDCl3. The 1H (400 MHz) chemical shifts are
given in parts per million (δ) with respect to internal TMS standard.
(25) The 19F NMR spectra were acquired on a Varian Unity Plus
400 MHz instrument in C6D6. The19F (376 MHz) chemical shifts are
given in parts per million (δ) with respect to internal CFCl3 standard.
(26) We have also observed imine formation by 19F NMR using the
solid support base piperdine.
(1681 to 1684 cm-1) and ester group (1797 to 1798 cm-1
)
(28) (a) Concha, N. O.; Rasmussen, B. A.; Bush, K.; Herzberg, O.
Structure 1996, 4, 823-836. (b) Fabiane, S. M.; Sohi, M. K.; Wan, T.;
Payne, D. J .; Bateson, J . H.; Mitchell, T.; Sutton, B. J . Biochemistry
1998, 37, 12404-12411.
(29) France, S.; Wack, H.; Hafez, A. M.; Taggi, A. E.; Witsil, D. R.;
Lectka, T. Org. Lett. 2002, 4, 1603-1605.
(27) Kaminskia, N. V.; Spingler, B.; Lippard, S. J . J . Am. Chem.
Soc. 2000, 122, 6411-6422.
J . Org. Chem, Vol. 68, No. 15, 2003 5823

Hafez et al.
SCHEME 4. Lew is Acid Activa tion of â-La cta m s
for Alcoh olysis
SCHEME 5. Syn th esis of
l-th r eo-â-Hyd r oxya sp a r a gin e Der iva tive
were also identified. The ring (C(O)-lactam) carbonyl
stretch of a structurally similar N-acyl-â-lactam has been
reported by Ojima et al. to occur at 1750 cm-1, a number
matching well with our observed figure of 1748 cm-1 30
.
The (C(O)N) acyl carbonyl shift of a similar N-benzoyl-
â-lactam has been reported to occur at 1680 cm-1 31
.
It
should be noted that the relatively small magnitude of
the observed IR shifts in this experiment correlate with
other reported findings. For example, a study by Lippard
and colleagues found that the binding of a binuclear Zn
complex to cephalothin resulted in a small, yet measur-
able, shift of the â-lactam carbonyl, 1773 to 1777 cm-1
(C(O)-lactam) and nitrogen acyl group, 1680 to 1683
cm-1 (C(O)N), respectively.32 On the basis of our findings,
it is highly plausible that Lewis acid activation involves
the formation of a number of closely related chelation
modes between the Lewis acid and the N-acyl-â-lactam
9a (Scheme 4).
glycine 2a . Second, we switched from an ethyl ester to a
benzyl ester that we envisaged would be easily hydro-
genated before the final unmasking to produce L-threo-
â-hydroxyasparagine 14.
Syn th esis of th e L-th r eo-â-Hyd r oxya sp a r a gin e-
Con ta in in g Su bu n it of Lysoba ctin . To highlight the
scope of our methodology, we have prepared the un-
natural amino acid L-threo-â-hydroxyasparagine as a key
subunit in our proposed total synthesis of the antibiotic
lysobactin.33 Efforts directed toward the preparation of
lysobactin and structurally related derivatives possessing
improved therapeutic indices are of recent interest. Our
approach to the synthesis of this key subunit rested on
the disposition of the acid chloride and R-chloroglycine
that we would use in our catalytic, asymmetric reaction
to form the â-lactam intermediate. We determined that
p-methoxyphenoxyacetyl chloride 1e would provide an
ideal protecting group for the â-hydroxy functionality in
the final product. Our standard R-chloroglycine had to
be modified in two ways in order to simplify the synthe-
sis. First, the glycine nitrogen had to be protected by a
biphenyl acyl group34 that could be more easily removed
than the benzoyl group on the original N-acyl-R-chloro-
With these modifications made to our substrates, we
set forth on our preparation of the â-hydroxyasparagine
14 by utilizing our standard PS/BQ protocol with (biphe-
nyl-4-carbonyl)aminochloroacetic acid benzyl ester 2d
and p-methoxyphenoxyacetyl chloride 1e to form the cis-
â-lactam 9g in 48% yield (11/1 dr, 95% ee). Aminolysis
of the â-lactam gave the protected L-threo-â-hydroxyas-
paragine 11 in 88% yield. Hydrogenation of 11 removed
the benzyl ester giving 12 in 99% yield. Deprotection of
the biphenyl protecting group on the nitrogen by Na/Hg
amalgam proceeded in 67% yield to give the free amine
13. CAN oxidation unmasked the L-threo-â-hydroxyas-
paragine 14 in 82% yield (Scheme 5).
Similar to Boger and others, we observed that the
unprotected carboxamide of an asparagine residue un-
dergoes side reactions.5 The L-threo-â-hydroxyasparagine
14 deteriorated into several byproducts upon standing
after the CAN oxidative deprotection. Luckily, protection
of the carboxamide prior to CAN oxidation eliminates
these deleterious reactions and should provide for a
straightforward solution in our proposed total synthesis
of lysobactin that will be reported in due course.
(30) Ojima, I.; Kudik, S. D.; Slater, J . C.; Gimi, R. H.; Sun, C. M.
Tetrahedron 1996, 52, 209-224.
(31) Ojima, I.; Ivn, H.; Mangzhu, Z.; Martine, Z.; Hoon, P. Y.
Tetrahedron 1992, 48, 6985-7012.
(32) Kaminskaia, N. V.; Spingler, B.; Lippard, S. J . J . Am. Chem.
Soc. 2001, 123, 6555-6563.
(33) O’Sullivan, J .; McCullough, J . E.; Tymiak, A. A.; Kirsch, D. R.;
Trejo, W. H.; Principe, P. A. J . Antibiot. 1988, 41, 1740. (b) Tymiak,
A. A.; McCormick, T. J .; Unger, S. E. J . Org. Chem. 1989, 54, 1149-
1157.
(34) The choice of a biphenyl amide protecting group afforded easy
deprotection without interfering with other functional groups present
in the molecule while maintaining the high yields and selectivities of
the asparagine derivatives.
5824 J . Org. Chem., Vol. 68, No. 15, 2003

Catalytic Asymmetric Synthesis of â-Amino Acids
dried and distilled by standard procedures. All acid chlorides
were distilled by standard procedures. HPLC analysis was
performed with a Waters Millipore Model 510 head unit, a
Regis Technologies (R,R) Whelk-01 Chiral, a Waters Millipore
Lambda-Max Model 481L spectrophotometer and a Hewlett-
Packard integrator. Solution-phase IR data were recorded on
a Bruker Vector 22 FTIR spectrometer.
Con clu sion . We have demonstrated that the organic
catalyst BQ can serve as many as five distinct catalytic
roles in the synthesis of â-substituted aspartic acid
derivatives. This method involves the reaction of ketenes
and N-acylimines formed in situ through dehydrohalo-
genation of the respective acyl halides and N-acyl-R-
chloroglycines. In addition to its role as the dehydroha-
logenating agent in this reaction, BQ also serves as a
catalyst for the highly stereoselective formation of N-acyl-
â-lactams with ee >90% and dr >10/1 in reasonable to
good yields. Furthermore, the rates of â-lactam ring
opening to the aspartic acid derivatives and subsequent
transesterification were dramatically enhanced in the
presence of BQ.
The use of alternative stoichiometric bases for imine
generation was also examined. It was determined that
the less costly base sodium hydride functioned as well
as Proton Sponge to afford comparable yields of the
desired products. This process was also amenable to the
preparation of polypeptides by addition of individual
amino acids or peptide fragments during the â-lactam
ring opening stage of the reaction. Also, we established
that the addition of Lewis acids during BQ-catalyzed
alcoholysis further enhanced the overall rate of the
reaction. The late transition metals Cu(OTf)2 and In-
(OTf)3 were found to be superior in the bifunctional Lewis
acid-Lewis base-catalyzed methanolysis of N-acyl-â-
lactams. Experimental evidence supports the existence
of a cooperative interaction between the Lewis acid and
Lewis base leading to the observed rate enhancement.
Furthermore, utilization of this methodology allowed for
the preparation of the L-threo-â-hydroxyaspararagine
subunit of the antibiotic lysobactin whose total synthesis
will be reported shortly.
Gen er a l P r oced u r e for â-Su bstitu ted Asp a r tic Acid s
5a -e. A 25 mL round-bottom flask equipped with a stir bar
was loaded under nitrogen with R-chloroamine 2a (63 mg, 0.26
mmol), Proton Sponge 4 (83 mg, 0.39 mmol), and the ben-
zoylquinine catalyst 3a (6 mg, 0.013 mmol). Toluene (1 mL)
was added to the mixture and stirred for 1 h. The solution
was diluted with toluene (7 mL) and cooled to -78 °C in a dry
ice/acetone bath. Phenylacetyl chloride 1a (20 mg, 0.13 mmol)
in toluene (1 mL) was added to the reaction dropwise. The
reaction was allowed to slowly warm to room temperature
overnight. Excess methanol (6 mL) was added, and the solution
was refluxed. The reaction was monitored by TLC and stopped
when all of the â-lactam had reacted (∼14 h). The solvent was
removed in vacuo, and the residue was taken up in chloroform
(10 mL) and washed with 1 M HCl (3 × 10 mL). The organic
layer was dried with MgSO4 and filtered through Celite. The
filtrate was concentrated, and the residue was submitted to
flash column chromatography to yield 5a in 62% yield (22 mg)
and 95% ee.
Ackn owledgm en t. T.L. thanks the NIH (GM54348),
Merck, Eli Lilly, the NSF Career Program for support,
the Dreyfus Foundation for a Teacher-Scholar Award,
and the Alfred P. Sloan Foundation for a Fellowship.
A.T. thanks the Organic Division of the American
Chemical Society for a Graduate Fellowship sponsored
by Organic Reactions, Inc. (2001-2002).
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures. This material is available free of charge via the
Internet at http://pubs.acs.org.
Exp er im en ta l Section
Gen er a l Meth od s. Unless otherwise stated, all reactions
were carried out under strictly anhydrous, air-free conditions.
All reagents used are commercially available. All solvents were
J O034150E
J . Org. Chem, Vol. 68, No. 15, 2003 5825