1,2-Asymmetric induction in dianionic functionalization reactions of L-aspartic acid diesters 1

Article abstract of DOI:10.1039/a900796b

Functionalization of the dianions derived from N-protected aspartic acid esters is a simple route to the preparation of novel α-amino acids and peptidomimetic precursors. We have demonstrated previously that high levels of 1,2-asymmetric induction may be obtained in this reaction under the appropriate conditions (I.B. Parr, et al., J. Med. Chem., 1996, 39, 2367). The observed diastereoselectivity results from the complicated interaction of a number of experimental factors. We now report that the nature of the electrophile and the work-up protocol both influence the level of stereocontrol in the alkylation of N-benzyloxycarbonyl aspartic acid esters. In addition, the stereochemical preference for high anti selectivity in the reaction can be reversed by increasing the steric bulk of the α-ester. Application of these observations has allowed the stereocontrolled preparation of two novel, N-protected, chimeric α-amino acids and a peptidomimetic precursor. The molecular basis for 1,2-asymmetric induction in the alkylation reaction can be rationalized on the basis of the configurational preferences of the lithium ester enolate formed from the N-protected aspartate diester used in these studies.

Full text of DOI:10.1039/a900796b

View Article Online / Journal Homepage / Table of Contents for this issue
1,2-Asymmetric induction in dianionic functionalization reactions
of L-aspartic acid diesters1
Ian B. Parr,a† Anthony B. Dribben,a‡ Simon R. Norris,b§ Mark G. Hindsb¶ and
Nigel G. J. Richards*a
a Department of Chemistry, University of Florida, Gainesville, FL 32611, USA
b Department of Chemistry, Southampton University, Southampton, UK SO9 5NH
Received (in Glasgow) 14th January 1999, Accepted 23rd February 1999
Functionalization of the dianions derived from N-protected aspartic acid esters is a simple route to the preparation
of novel α-amino acids and peptidomimetic precursors. We have demonstrated previously that high levels of
1,2-asymmetric induction may be obtained in this reaction under the appropriate conditions (I.B. Parr, et al., J. Med.
Chem., 1996, 39, 2367). The observed diastereoselectivity results from the complicated interaction of a number of
experimental factors. We now report that the nature of the electrophile and the work-up protocol both influence
the level of stereocontrol in the alkylation of N-benzyloxycarbonyl aspartic acid esters. In addition, the stereo-
chemical preference for high anti selectivity in the reaction can be reversed by increasing the steric bulk of the
α-ester. Application of these observations has allowed the stereocontrolled preparation of two novel, N-protected,
chimeric α-amino acids and a peptidomimetic precursor. The molecular basis for 1,2-asymmetric induction in the
alkylation reaction can be rationalized on the basis of the configurational preferences of the lithium ester enolate
formed from the N-protected aspartate diester used in these studies.
H
N
H
Introduction
R2
CO2Me
S
OHC
Z
The preparation of non-natural amino acids by diastereo-
selective and enantioselective synthesis has been the focus of
considerable research.2 Such compounds have proven to be
important (i) in drug discovery, and (ii) as structural and mech-
anistic probes of biological systems.3 Non-natural α-amino
acids also represent key intermediates in the synthesis of con-
formationally constrained molecular scaffolds as elements in
the design of small molecule combinatorial libraries.4 In this
regard, we were interested in developing a synthetic route for
the preparation of the substituted rigid bicyclic β-turn mimetic
1 (Scheme 1). Although the synthesis of the unsubstituted
compound 2 is described in a number of reports,5 extension of
these approaches to obtain ring-functionalized mimetics such
as 1 do not appear to have been pursued.
R
O
O
N
H
N
H
H
O
CO2R1
1 R2 = CH2OR3
2 R2 = H
3
H
CO2Me
CO2But
CO2Me
CO2But
tBuO2C
Z
ZHN
N
H
H
H
We based our retrosynthetic analysis on known methods for
the construction of bicyclic frameworks by reaction of pro-
tected aldehydes with cysteine derivatives.6 We therefore envis-
aged that the aldehyde 3 represented a useful intermediate in the
synthesis of 1 (Scheme 1), and that this compound could be
obtained by regiospecific transformation of the triester 4. In
turn, functionalization of the dianion derived from deproton-
ation of the β-carbon of the aspartic acid diesters 5 with an
appropriate haloacetate was expected to give 4 in good yield.
Alkylation of suitably protected aspartic acid esters has been
shown to be a practical approach to obtaining a variety of
compounds in optically pure form, including non-natural
amino acids and alkaloids.7 Indeed, our group has employed
this reaction to prepare a series of analogs capable of defining
the bound conformation of substrate in the C-terminal domain
5
4
Scheme 1 Retrosynthetic analysis of functionalized peptidomimetic 1.
of the enzyme asparagine synthetase.8 On the other hand, while
1,2-asymmetric induction in the alkylation of aspartic acid
ester enolates is well documented, the effect of experimental
conditions on the stereochemical outcome of the reaction
appears to be complex.7,8 For example, recent work has shown
that very high anti diastereoselectivities can be achieved using
lithium as the counter-ion in the presence of additives such as
HMPA.7a The nature of the metal counter-ion also appears to
define the ester enolate geometry in the anionic intermediate.7b
There have been no systematic studies, however, of the role of
other experimental variables in defining the level of 1,2-asym-
metric induction. We now report that the diastereoselectivity in
the β-functionalization of diester 5 is dependent on the reactiv-
ity of the electrophile, the steric bulk of the C-2 ester substitu-
ent and the procedure employed to quench the reaction. The
molecular basis for 1,2-asymmetric induction in the alkyl-
ation reaction can be rationalized on the basis of the likely
configurational preferences of the lithium ester enolate formed
from diester 5. Application of these observations has allowed
† Current address: Cubist Pharmaceuticals, 24 Emily St., Cambridge,
MA 02139, USA.
‡ Current address: Department of Chemistry, Mississippi College,
Clinton, MS 39058, USA.
§ Current address: Lancing College, Lancing, West Sussex, UK BN15
0RW.
¶ Current address: Biomolecular Research Institute, 343 Royal Parade,
Parkville 3052, Australia.
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038
1029

View Article Online
Table 1 Effect of reaction conditions on 1,2-asymmetric induction in the reaction of tert-butyl chloroacetate and the dianion derived from
deprotonation of 5 by LiHMDS
Entry
T/ЊC
t/h
Salta
Ionic strength
Workup
4a:4b
1
2
3
4
5
6
7
8
9
0
Ϫ20
3
3
3
None
None
None
None
None
None
None
1 equiv. LiCl
2.5 equiv. LiCl
5 equiv. LiCl
10 equiv. LiCl
0.087
0.087
0.087
0.087
0.087
0.087
0.087
0.116
0.160
0.232
0.377
Ab
A
A
B
A
B
C
A
A
A
A
mixture
mixture
>99:1
2:1
>99:1
2:1
Ϫ30
Ϫ30
Ϫ82
Ϫ82
Ϫ82
Ϫ75
Ϫ75
Ϫ75
Ϫ75
3
18
18
18
6
6
6
2:1
>99:1
>99:1
>99:1
>99:1
10
11
6
a Salt was added to the solution of the enolate at low temperature before addition of the electrophile. b A: Cold reaction solution is poured into 1 M
HCl at rt; B: reaction solution warmed to rt before addition to 1 M HCl; C: 1 M HCl added to cold reaction solution, and then warmed to rt.
the stereocontrolled preparation of two novel, N-protected,
chimeric α-amino acids9 and a differentially protected aspartate
precursor for peptidomimetic 1.
ization of the dianion, however, represented a second import-
ant, and unanticipated, variable. For example, addition of acid
to the reaction solution at low temperature followed by warm-
ing, as reported in the literature,7 gave a mixture of diastereo-
meric products 4a and 4b (Table 1). In contrast, the triester 4
was obtained as a single diastereoisomer if the reaction was
quenched by pouring the cold solution of the monoanion into
aqueous acid.
Results and discussion
The -aspartate diester 5 was synthesized from -aspartic acid
in three steps following literature procedures.10 Our initial
experiments involved generation of the dianion derived from
We established the relative stereochemistry of the single
diastereoisomeric triester 4 by converting this product to
the cyclic anhydride 7 (Scheme 2). This was accomplished by
treatment of 4 with trifluoroacetic acid in dichloromethane to
remove the tert-butyl esters, and subsequent cyclization of the
resulting diacid 6 using acetic anhydride.12 Careful control of
the temperature in the latter reaction was necessary, however, to
5
using two equivalents of lithium hexamethyldisilazide
(LiHMDS)11 in THF at Ϫ35 ЊC. The desired triester 4 was
then obtained as a mixture of diastereoisomers (4a:4b) by over-
night reaction of the dianion with tert-butyl chloroacetate at
room temperature (Scheme 2). No N-alkylation was observed
1
avoid extensive decomposition of the product. H NMR spec-
R2
CO2Me
R1
CO2Me
CO2But
troscopy showed that the vicinal coupling constant for the C-2
and C-3 protons (J23) of the isolated anhydride 7 was 12.1 Hz,
consistent with diastereoisomer 7a. Given that epimerization
did not occur under the deprotection and cyclization condi-
tions, triester 4a must have been obtained as the sole product
under our modified conditions for the alkylation reaction.
Thus, the alkylation reaction proceeded with anti selectivity, as
seen for other electrophiles in previous studies.7a We confirmed
our stereochemical assignment by transforming a 4:1 mixture
of the diastereoisomeric triesters 4a and 4b (prepared using the
literature procedures for alkylation of aspartate derivative 5) to
i
ZHN
CO2But
ZHN
H
H
5
4a R1 = CH2CO2But, R2 = H
4b R1 = H, R2 = CH2CO2But
ii
1
the corresponding mixture of anhydrides 7a and 7b. The H
R2
NMR resonances for the C-2 protons of 7a and 7b in the mix-
ture were easily differentiated, and J23 for the minor isomer 7b
was determined to be 5.2 Hz. This value is consistent with the
stereochemistry assigned to 7b, based on previous studies of
β-chloroglutamates.13 A number of approaches were employed
to ensure that the integrity of the chiral center at C-2 in starting
diester 5 was maintained under the alkylation conditions. First,
the optical rotation of unreacted, recovered diester 5 was iden-
tical to that of the pure starting material. Second, the triester
product 4a was converted to the amine 8 using catalytic hydro-
gen transfer to remove the benzyloxycarbonyl protecting group.
Reaction of 8 with chiral isocyanates, in two separate experi-
ments,14 then gave the urea derivatives 9a and 9b (Scheme 3).
The methyl ester resonances in 9a and 9b were clearly
differentiable and provided a method of assessing the optical
purity of the triester 4a formed in the reaction. Thus, if there
were any racemization of 5 before alkylation, derivatization of
the resulting triester 4a would yield 9a and the enantiomer of
9b, giving rise to two methyl ester singlets in the 1H NMR of the
product. Only a single methyl ester resonance was observed for
the crude material formed by reaction of 8 with either enantio-
merically pure isocyanate. This was confirmed by doping the
NMR sample of 9a with increasing amounts of 9b formed
in the other derivatization reaction to give two methyl ester
R2
R1
O
R1
CO2Me
CO2H
iii
O
ZHN
ZHN
H
H
O
6a R1 = CH2CO2H, R2 = H
6b R1 = H, R2 = CH2CO2H
7a R1 = H, R2 = CO2Me
7b R1 = CO2Me, R2 = H
Scheme
2
Z = PhCH2OC(᎐O): i) LiHMDS, THF, Ϫ75 ЊC, then
᎐
ClCH2CO2But, Ϫ75 ЊC, 62%; ii) CF3CO2H, CH2Cl2, 70%; iii) Ac2O,
60 ЊC, 80%.
in any experiment described in this paper. Although a lack of
diastereoselectivity during the alkylation of the dianion
obtained from 5 had been reported previously for a number of
other electrophiles,7 we decided to investigate the reaction con-
ditions in a systematic fashion to determine whether selectivity
could be improved. Subsequently we discovered that two
factors appeared to cause the poor stereoselectivity observed in
our initial experiments. First, reaction temperature was an
important variable. Maintaining the temperature below Ϫ50 ЊC
during alkylation of the dianion was an essential element in
obtaining high levels of 1,2-asymmetric induction. The pro-
cedure for quenching the monoanion formed after functional-
1030
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038

View Article Online
H
H
H
H
H
CO2H
CO2H
CO2Me
CO2But
CO2H
CO2H
tBuO2C
ZHN
HO2C
ZHN
HO2C
ZHN
H
14a
10:1
14b
4a
ii
R1O2C
O
i
H
H
HO2C
CO2Me
CO2R2
H
H
CO2Me
H
O
CO2Me
CO2Me
CO2But
HO2C
ZHN
iii
tBuO2C
H2N
iii
N
H
O
O
PhCH2O
N
H
CO2H
Br
iv
H
H
10 R1 = But, R2 = But
6a
15
8
11 R1 = H, R2 = But
12 R1 = H, R2 = H
i
ii
H
tBuO2C
CONH2
CO2H
H
HO2C
ZHN
CO2Me
CO2But
R2 R1
O
H
Ph
N
N
H
13
H
H
9a R1 = Me, R2 = H
9b R1 = H, R2 = Me
Scheme 4 Z = PhCH2OC(᎐O): i) NH4OH, rt, 77%; ii) LiOH, 4:1
᎐
MeOH–H2O, 73%; iii) (CH2O)n, cat. p-TsOH, 88%.
Scheme 3 Z = PhCH2OC(᎐O): i) H2, 10% Pd/C, 95%; ii) (S)-PhCH-
of asparagine and glutamate, both products of the enzyme-
catalyzed reaction.8 The preparation of chimeric amino acids
has received considerable attention.9 In related studies, hydroly-
sis of the methyl ester 6a using LiOH proceeded smoothly to
᎐
(CH3)N᎐C᎐O or (R)-PhCH(CH3)N᎐C᎐O, THF, 100%; iii) BrC6H4-
᎐
᎐
᎐ ᎐
COCl, Et3N, 58%; iv) CF3CO2H, CH2Cl2.
signals. In separate efforts to confirm the relative and absolute
stereochemistry of the triester 4a, the amine 8 was also reacted
with 4-bromobenzoyl chloride to give the triester 10. Removal
of the tert-butyl protecting groups using trifluoroacetic acid
yielded a mixture of acids 11 and 12, which could be separated
using reversed-phase HPLC. Although both of these products
were solids, recrystallization only gave needles that were unsuit-
able for X-ray structure determination. In the last set of
experiments, the dianion of 5 was generated at Ϫ75 ЊC using
2.1 equivalents of LiHMDS following our standard procedure.
Quenching using D2O and recovery of the diester 5 only yielded
material in which deuterium had been incorporated at C-3,
suggesting that any deprotonation at C-2 occurs at levels below
the detection limit of NMR. It is not yet clear how simple
modification of the work-up conditions should play such an
important role in modulating the stereochemical outcome of
the reaction. Explanations that invoke equilibration under
reported work-up conditions in studies of the β-alkylation of
diester 5 seem to be ruled out by semi-empirical calculations
that suggest triester 4a is more thermodynamically stable than
4b.15,16 Additional experiments are under way to address this
question.
Having established the stereochemical outcome of the alkyl-
ation reaction using tert-butyl chloroacetate, we examined the
conversion of diacid 6a into derivatives with potential utility
in investigating the structure and chemical mechanism of
asparagine synthetase (Scheme 4).8 Alkylation of 5 using tert-
butyl chloroacetate represents a novel route for synthesizing
β-functionalized glutamic acid analogs.17 We note that struc-
turally related compounds exhibit potent biological activity,
particularly at specific classes of receptor in the central nervous
system.18 In our initial experiments, treatment of the diacid 6a
with aqueous ammonia yielded amide 13 in reasonable yield.
No epimerization at C-3 was observed under these reaction
conditions. The highly functionalized, novel α-amino acid 13
has potential utility in the development of asparagine syn-
thetase inhibitors because it represents a protected chimera
give the triacid 14a, after HPLC purification, in good yield.19
A
small amount of the diastereoisomeric triacid 14b was also
produced under the reaction conditions, consistent with previ-
ous reports of epimerization at C-3 in aspartate analogs of
similar structure.20 Triacid 14 is another interesting structure,
being a chimera of aspartic acid and glutamic acid. Finally,
reaction of 6a with paraformaldehyde, in the presence of a
catalytic amount of toluene-p-sulfonic acid,21 gave the desired
oxazolidinone 15 in good yield. This compound is a useful pre-
cursor of aldehyde 3 given the pattern of differential protection.
In light of the high level of 1,2-asymmetric induction
observed in the preparation of triester 6a, we undertook a
systematic investigation of the role of experimental variables
in determining the stereochemical outcome of the alkylation
reaction. Thus, the reactivity of the electrophilic species was
modified using in situ halide exchange. In these experiments
tert-butyl chloroacetate was stirred with a halide salt for 2 hours
directly prior to its reaction with the dianion formed by treat-
ment of 5 with LiHMDS. Whereas the addition of LiBr or NaI
did indeed increase the overall yield of the alkylation reaction,
the triester product obtained under these conditions was a mix-
ture of diastereoisomers 4a and 4b. For example, the presence
of NaI gave 4a:4b in a 3:1 ratio, even using our modified
workup protocol. Two possible effects were hypothesized to
give rise to the observed decrease in stereochemical control. In
order to rule out modulation of diastereoselectivity from salt
effects, the dianion of 5 was reacted with tert-butyl chloro-
acetate in the presence of increasing amounts of LiCl (Table 1).
In these experiments, the salt was mixed with the dianion prior
to addition of the electrophile. No effect on the diastereo-
selectivity of the reaction was observed, even if LiBr or LiI
were present. In all cases, only triester 6a was obtained using
our modified workup conditions. This suggested that halide
exchange was occurring in our original experiments, and that
therefore the decreased 1,2-asymmetric induction resulted
from reaction of the dianion with the tert-butyl bromoacetate
formed in situ. Evidence to support this conclusion was
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038
1031

View Article Online
Table 2 Effect of the α-ester and electrophile on alkylation diastereo-
obtained in two ways. First, we confirmed that stirring tert-
butyl chloroacetate with LiBr in THF for 2 hours at ambient
temperature gave a 3:2 mixture of the chloro- and bromo-
acetates as determined by 1H NMR measurements. Second,
reaction of the dianion derived from 5 with tert-butyl bromo-
acetate gave the triester 6 as a 3:2 mixture of 6a and 6b under
our standard reaction conditions.
selectivity
Ester
Electrophile
3R:3S Ratio
>99:1
3:2
2.5:1
1:1.2
>99:1
8:1
4:1
5
5
ClCH2CO2But
BrCH2CO2But
BrCH2CO2But
BrCH2CO2But
16
17
5
The effect of the steric bulk of the C-1 ester on the alkylation
diastereoselectivity was also investigated (Scheme 5). Diester 16
BrCH2CH᎐CH2
᎐
5
ICH2CH᎐CH2
᎐
᎐
5
BrCH2CH᎐CH
᎐
R2
᎐
BrCH2C᎐N
᎐
5
2:1
CO2Me
R1
CO2Me
CO2R
>99:1a
>99:1
i
5
PhCHO
5
H2C᎐CHCHO
᎐
ZHN
CO2R
a Aldehyde addition products are likely to be a 1:1 mixture of
diastereoisomers at C-4.
ZHN
H
H
16 R = Me
18a R = Me, R1 = CH2CO2But, R2 = H
18b R = Me, R1 = H, R2 = CH2CO2But
19a R = CPh3, R1 = CH2CO2But, R2 = H
19b R = CPh3, R1 = H, R2 = CH2CO2But
17 R = CPh3
of diastereoisomeric products, obtained by alkylation of the
dianion of 17 with tert-butyl bromoacetate, was treated directly
with trifluoroacetic acid to give the diacids 6a and 6b, which
were easily separated by reversed-phase HPLC. Diacid 6b was
shown to be the major component of the reaction mixture by
co-injection with an authentic sample, implying that 19b was
the major diastereoisomer formed in the alkylation of the
dianion obtained from aspartic acid ester 17. The observed syn
diastereoselectivity was intriguing in that the increased steric
bulk of the trityl group reverses the anti selectivity observed in
β-alkylation of the diester 5 (Table 2).7
R2
R2
R1
CO2Me
R1
CO2H
CO2H
ii, iii
ZHN
CO2Me
ZHN
H
H
18a R1 = CH2CO2But, R2 = H
18b R1 = H, R2 = CH2CO2But
14a R1 = CH2CO2H, R2 = H
14b R1 = H, R2 = CH2CO2H
Having established reaction conditions for obtaining high
levels of 1,2-asymmetric induction in the functionalization of
5 using tert-butyl chloroacetate, we next examined the stereo-
chemical outcome when the dianion was reacted with other
activated electrophiles (Table 2). In our initial experiments,
β-allylation of 5 using allyl bromide gave 20 as a single
diastereoisomer (Scheme 6), a result consistent with literature
observations.7,8a The assignment of stereochemistry was accom-
plished using two approaches. First, diacid 21 was prepared in
two steps by sequential reaction of the allylated diester 20 with
trifluoroacetic acid and LiOH.8a In contrast to the behavior of
6a, however, epimerization at C-3 was not observed during the
deprotection of the methyl ester. Cyclization of 21 on treatment
with acetic anhydride proceeded smoothly to give the anhydride
22, for which the amide proton resonance was clearly evident in
the 1H NMR spectrum. Irradiation of the amide peak allowed
us to assign that of the C-2 proton, which was visible as a
doublet for which J23 was determined to be 10.2 Hz. This com-
pares to the value of J23 = 5.6 Hz reported previously for
the benzylated aspartate derivative 23.20 Thus, the allylation
product was tentatively assigned to be 22a. Subsequently, we
prepared a mixture of 20a and 20b by reaction of the dianion of
5 with allyl iodide, and this was converted into the anhydrides
22a and 22b. As expected, the 1H NMR resonances of the C-2
proton resonances in these diastereoisomeric anhydrides were
well resolved. In this manner, we were able to determine the
cognate coupling constant J23 for the minor diastereoisomer as
8.1 Hz, consistent with the predicted anti-relationship in 22b. In
the second approach, details of which have appeared else-
where,8a the allylated diester 20a was converted in several steps
to 24 (Scheme 6). The stereochemistry of 24 has been defined by
transformation to (2S,3R)-carboxyproline.7b,8a,9b
R2
R2
R1
CO2Me
R1
CO2Me
ii
ZHN
CO2CPh3
ZHN
CO2H
H
H
19a R1 = CH2CO2But, R2 = H
19b R1 = H, R2 = CH2CO2But
6a R1 = CH2CO2H, R2 = H
6b R1 = H, R2 = CH2CO2H
Scheme
5
Z = PhCH2OC(᎐O): i) LiHMDS, THF, Ϫ75 ЊC, then
᎐
BrCH2CO2But, Ϫ75 ЊC; ii) CF3CO2H, CH2Cl2; iii) LiOH, 4:1 MeOH–
H2O.
was simply obtained by refluxing β-methyl N-benzyloxycarb-
onylaspartate,10 an intermediate in the synthesis of 5, with
HCl in MeOH.22 Similarly, the trityl derivative 17 could be
synthesized in 98% yield by reaction of the caesium salt of
β-methyl N-benzyloxycarbonylaspartate with trityl bromide in
dry THF.23 Diesters 16 and 17 were then alkylated by reaction
of their respective lithium dianions with tert-butyl bromo-
acetate at Ϫ75 ЊC for 17 hours. As usual, each reaction was
worked up by pouring the cold mixture into 1 M HCl at room
temperature. The electrophile was specifically chosen as we
knew that the triester products 18 and 19, respectively, would be
obtained as a mixture of diastereoisomers at C-3. Given that
control of relative stereochemistry was the issue in these
experiments, no effort was made to establish whether epimeriz-
ation at C-2 had occurred under the reaction conditions. It is
known, however, that deprotonation at C-2 can be an issue in
the alkylation of aspartate diesters in which the α-proton is
not sterically hindered, leading to partial racemization of the
products.24 As anticipated, H NMR spectroscopy confirmed
1
that the alkylation products 18 and 19 were obtained as a mix-
ture of diastereoisomers. Stereochemical assignment of the
major product 18a was accomplished by conversion of triester
18 in two steps to a mixture of the triacids 14a and 14b. The
triacid 14a had been prepared previously and its reversed-phase
HPLC retention time was known. Co-injection of an authentic
sample of 14a with the mixture obtained from 18 therefore
allowed unambiguous assignment of the original product mix-
ture. In the case of the triester 19, purification of the desired
product was complicated by partial deprotection of the trityl
group under the acidic workup conditions. Hence the mixture
Reaction of the dianion generated from 5 by treatment with
LiHMDS also proceeded smoothly with propargyl bromide
(prop-2-ynyl bromide) or bromoacetonitrile, to give the novel
diesters 25 and 26, respectively (Fig. 1). Unfortunately, given
the functionality can be generated by reaction of acetylenes
and nitriles, both compounds were formed as a mixture of
diastereoisomers. Although detailed stereochemical analysis of
the products 25 and 26 was not performed, it is likely that the
major isomer formed represents that formed by anti addition
of the electrophile.7a,8a The dianion obtained from 5 was also
1032
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038

View Article Online
HO
R2
H
H
H
OLi
R1
CO2Me
CO2But
CO2Me
CO2But
CO2Me
CO2But
R
OMe
CO2But
ZHN
ZHN
N
N
Ph
H
Ph
H
Bn
Bn
27 R = Ph
25a R1 = CH2CN, R2 = H
25b R1 = H, R2 = CH2CN
28 R = CH=CH2
29
30
26a R1
=
, R2 = H
CH2C CH
26b R1 = H, R2
=
CH
CH2C
Z
OLi
OMe
Li
N
H
Fig. 1 Products obtained by reaction of various electrophiles with the
aspartic ester enolate derived from 5.
O
Li
OMe
tBuO2C
N
Li
H
O
ButO
Z
R2
O
R2
R1
31a
31b
R1
CO2H
CO2H
iv
O
ButO
ZHN
OLi
OMe
ZHN
H
O
H
O
22a R1 = CH2CH=CH2, R2 = H
22b R1 = H, R2 = CH2CH=CH2
21a R1 = CH2CH=CH2, R2 = H
21b R1 = H, R2 = CH2CH=CH2
H
Li
N
Z
31c
ii, iii
Fig. 2 Putative models 31a–31c to rationalize the observed 1,2-
asymmetric induction in the reaction of the lithium ester of 5 with
electrophiles. Z = PhCH2OC(᎐O).
᎐
H
CO2Me
R2
ref. 7
R1
CO2Me
CO2But
favorable structure for the resulting dianion would be 31a in
which the metal is chelated by the α-ester group rather than the
amide nitrogen (Fig. 2). Attack of the electrophile from the
least hindered face of the double bond, would then yield the
product of syn alkylation. This model is not therefore consist-
ent with the diastereoselection seen in our experiments, unless
the second lithium ion plays a role in directing electrophilic
attack in the transition state. Two structures, 31b and 31c, can
be envisaged in which lithium is chelated in the (Z)-lithium
ester enolate formed by deprotonation of 5 (Fig. 2). Attack
of the electrophile from the least hindered face of the double
bond in 31b would give the observed preference for anti alkyl-
ation, and such a model has been proposed previously to
explain the stereochemical outcome of alkylations involving
chiral β-hydroxyesters.27 On the other hand, if chelation is play-
ing a role in controlling the stereochemical outcome of the reac-
tion, it seems likely that increasing the steric bulk of the α-ester
should enhance anti selectivity in the alkylation reaction. Our
experiments using the trityl derivative 17 are therefore difficult
to reconcile with this model. In addition, increasing concentra-
tions of LiCl might be expected to disrupt this structure and
result in decreased diastereoselectivity. This was not observed
under our reaction conditions (Table 1).
Intermediate 31c represents an alternate structure for the
dianion in which a (Z)-lithium ester enolate is present (Fig. 2).
In this structure, the α-hydrogen is syn-periplanar to the enolate
and the second lithium ion is chelated by the remaining ester.
Electrophilic attack from the top face of the double bond
would then yield the product arising from anti alkylation, and
this model is consistent with that proposed to rationalize the
high stereoselectivity observed in the allylation of monoanion
30.7b This model also suggests that the top face of the double
bond might be sterically hindered by a trityl ester favoring
electrophilic attack from the lower face, giving the product
of syn alkylation as the major isomer. While this simple
model rationalizes our observations, it ignores the complexity
of supramolecular enolate structures in non-polar solvents.28
Further, the relevance of ground state conformational effects to
the activation energy of competing reaction pathways is not
always evident. Current efforts are therefore centered on valid-
ating this model using a combination of experiments, and ab
initio calculations on the energetics of electrophilic attack on
N
Z
CO2But
H
ZHN
H
24
20a R1 = CH2CH=CH2, R2 = H
20b R1 = H, R2 = CH2CH=CH2
i
Ph
H
O
O
CO2Me
PhFl
Ph
CO2But
N
ZHN
H
H
O
23
5
Scheme
6
Z = PhCH2OC(᎐O): i) LiHMDS, THF, Ϫ75 ЊC, then
᎐
BrCH2CH᎐CH2, Ϫ75 ЊC, 60%; ii) CF3CO2H, CH2Cl2, 89%; iii) NaOH,
᎐
1:1 MeOH–H2O, 58%; iv) Ac2O, 75 ЊC, 100%.
reacted with acrolein and benzaldehyde to yield the the alcohols
27 and 28 as 1:1 mixtures of only two diastereoisomers,
respectively. We currently assume that the diastereoisomers
formed in these reactions represent of epimers at C-4, and that
significant levels of 1,2-asymmetric induction has occurred at
C-3. This is consistent with earlier observations using benzal-
dehyde as the electrophilic component in the β-alkylation of
aspartate diesters.25
The observed preference for anti alkylation of the aspartic
acid ester dianion can be rationalized by considering models of
the intermediate in which there is chelation of the lithium
counter-ion (Fig. 2). Although definitive proof is lacking, it is
likely that deprotonation of the ester 5 using LiHMDS pro-
ceeds to yield the (Z)-lithium ester enolate. This assumption is
contrary to previous reports that deprotonation of esters by
LiHMDS in THF usually gives (E)-ester enolates,26 but is pre-
cedented by trapping experiments which clearly show that
monoanion 30, formed by reaction of aspartic acid diester 29
with LiHMDS, is the (Z)-ester enolate (Fig. 2).7b Further, in the
event that the (E)-lithium ester enolates were produced by
deprotonation of 5 under our reaction conditions, the most
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038
1033

View Article Online
models of ester enolates 31a–31c. Our results will be reported in
due course.
was a yellow oil that solidified on standing: mp 56–58 ЊC (from
EtOAc) (Found: C, 60.9; H, 7.3; N, 3.0; C23H33NO8 requires C,
60.9; H, 7.3; N, 3.0%); [α]D20 ϩ5.5 (c 0.2 in CDCl3); νmax(neat)/
cmϪ1 3358, 2979, 1732, 1506, 1456, 1223 and 1155; δH(360
MHz; CDCl3; Me4Si) 1.50 (18 H, s, (CH3)3C), 2.45 (1 H, dd,
J 5.9 and 17.0, CH2CH), 2.66 (1 H, dd, J 5.9 and 17.0, CH2CH),
3.50–3.70 (1 H, m, NCHCH), 3.70 (3 H, s, OCH3), 4.60 (1 H,
dd, J 3.6 and 8.9, NCHCH), 5.12 (2 H, s, OCH2), 5.55 (1 H, d,
J 8.9, NH), 7.35 (5 H, s, Ph); δC(75.4 MHz; CDCl3; CDCl3)
27.80 (q), 27.97 (q), 33.91 (t), 43.57 (d), 51.98 (q), 54.88 (d),
67.10 (t), 81.12 (s), 82.80 (s), 128.08 (d), 128.12 (d), 128.47 (d),
136.15 (s), 156.29 (s), 169.09 (s), 170.35 (s), 172.19 (s); m/z (CI:
NH4Cl) 469 (MNH4ϩ, 4), 452.2283 (MHϩ, 22%. C23H34NO8
requires 452.2284), 340 (100), 269 (78), 91 (73).
In summary, the work outlined here supports the utility of
aspartic acid diester β-functionalization as a rapid route for the
stereocontrolled preparation of novel α-amino acids, especially
if careful attention is paid to optimizing the appropriate
experimental variables. Work from other groups suggests
that syn selectivity in 1,2-asymmetric induction can be obtained
if potassium is employed as the counter-ion in the reaction.
Appreciation of the role of other factors in defining the
diastereoselectivity of functionalization at C-3 of aspartate, as
outlined in this paper, may enhance our ability to control the
level of 1,2-asymmetric induction when KHMDS is employed
as a base. It is also likely that the availability of β-functionalized
glutamates, such as 15, will allow the construction of novel
molecular scaffolds for use in creating functionally diverse
combinatorial libraries.
Preparation of di-tert-butyl (2S,3R)-2-benzyloxycarbonylamino-
3-methoxycarbonylglutarate (4a) and di-tert-butyl (2S,3S)-2-
benzyloxycarbonylamino-3-methoxycarbonylglutarate (4b)
A solution of the dianion of diester 5 (102 mg, 0.3 mmol) dis-
solved in dry THF (20 cm3) under nitrogen was prepared using
LiHMDS (0.75 mmol) at Ϫ80 ЊC, as described above. A solu-
tion of tert-butyl chloroacetate (1 cm3, 7 mmol) and NaI (225
mg, 1.5 mmol) dissolved in dry THF (10 cm3) was then added to
the orange dianion at such a rate so that the reaction temper-
ature did not exceed Ϫ75 ЊC. Stirring was continued for a
further 17 h at this temperature before the reaction mixture was
poured into 1 M HCl (120 cm3) at rt. This solution was then
extracted using Et2O (4 × 50 cm3), and the organic phase was
dried (MgSO4) before removal of the solvent under reduced
pressure. The resulting pale yellow oil was purified by “flash”
chromatography (eluant: 4:1 petroleum ether–EtOAc) to give
the alkylated product, as the first material eluted from the
column. The triester was obtained as a 2:1 mixture of diastereo-
Experimental
Commercially available amino acids and their derivatives were
obtained from Sigma Chemical Company. All other reagents
were purchased from Aldrich or Fisher Scientific, and were
used without further purification. Moisture sensitive reactions
were carried out under a nitrogen, or argon, atmosphere using
standard techniques. Glassware used in these reactions was
flame-dried with an inert gas sweep. THF was freshly distilled
before use from sodium benzophenone ketyl. Methanol was
dried using magnesium following standard procedures. Bis(tri-
methylsilyl)amine was distilled from CaH2 before use. All other
solvents were distilled before use. Throughout the experimental
section, petroleum ether refers to the fraction with bp 60–80 ЊC,
unless stated otherwise. Analytical TLC was performed on sil-
ica gel 60F-254 plates. Column chromatography was performed
on Kieselgel 60 (230–400 mesh) using freshly distilled solvents.
Analytical (250 mm × 4.6 mm) and preparative (50 mm × 21.4
mm) HPLC used Dynamax C18 and C8 reversed-phase columns,
with monitoring at 226 nm. Melting points are uncorrected.
Elemental analyses were recorded at the Microanalytical
Service Facility at the University of Florida. NMR spectra were
recorded on General Electric GE-300, Bruker AM-360 or
Varian Unity-500 spectrometers. 1H chemical shifts are
reported in ppm (δ) downfield of TMS as an internal reference
(δ 0.0). 13C chemical shifts are reported in ppm relative to TMS.
Splitting patterns are abbreviated as follows: br, broad, s, sing-
let, d, doublet, t, triplet, q, quartet and m, multiplet. EI, CI and
FAB mass spectra were recorded on a Finnegan MAT 25Q
spectrometer. Chemical ionization employed either ammonia or
isobutane.
1
isomers (4a:4b) (49 mg, 36%), based on H NMR analysis: 4b
δH(360 MHz; CDCl3; Me4Si) 1.50 (18 H, s, (CH3)3C), 3.22 (1 H,
dd, J 3.5 and 18.0, CH2CH), 3.32–3.37 (1 H, m, NCHCH), 3.46
(1 H, dd, J 3.5 and 18.0, CH2CH), 3.70 (3 H, s, OCH3), 4.50–
4.55 (1 H, m, NCHCH), 5.00 (2 H, s, OCH2), 5.66 (1 H, d, J 8.0,
NH), 7.35 (5 H, s, Ph).
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-methoxy-
carbonylglutaric acid (6a)
TFA (2 cm3, 12 mmol) was added to a solution of the triester 4a
(200 mg, 0.44 mmol) dissolved in CH2Cl2 (40 cm3) at 0 ЊC with
vigorous stirring. After a further 2 h at this temperature, the
reaction mixture was slowly warmed and stirred for 12 h at
15 ЊC. Removal of the volatiles under reduced pressure gave
an oil that was purified using reversed-phase HPLC with
gradient elution (C18 column; solvent flow rate 12 cm3 minϪ1
;
80:20:1 H2O–CH3CN–TFA to 20:80:1 H2O–CH3CN–TFA
over 30 min). The desired diacid 6a (105 mg, 70%) was obtained
as a white solid (retention time 13.8 min) after lyophilization:
mp 82–84 ЊC (Found: C, 53.1; H, 5.3; N, 4.2; C15H17NO8
requires C, 53.1; H, 5.05; N, 4.1%); [α]D20 ϩ32.0 (c 0.1 in CH3CN);
νmax(CHCl3)/cmϪ1 3550–2900, 2853, 1726, 1509 and 1460;
δH(360 MHz; CD3CN; Me4Si) 2.65 (1 H, dd, J 6.7 and 17.5,
CH2CH), 2.80 (1 H, dd, J 7.5 and 17.5, CH2CH), 3.55–3.65
(1 H, m, NCHCH), 3.70 (3 H, s, OCH3), 4.70 (1 H, dd, J 3.8
and 9.1, NCHCH), 5.16 (2 H, s, OCH2), 5.40–5.60 (2 H, br s,
CO2H), 6.05 (1 H, d, J 9.0, NH), 7.43 (5 H, s, Ph); δC(90.6
MHz; CD3CN; CD3CN) 31.63 (t), 43.19 (d), 51.92 (q), 54.06
(d), 66.56 (t), 127.76 (d), 128.01 (d), 128.49 (d), 137.10 (s),
156.16 (s), 170.87 (s), 171.91 (s), 172.14 (s); m/z (CI: NH4Cl)
340.1059 (MHϩ, 4%. C15H18NO8 requires 340.1032), 278 (28),
205 (13), 188 (43), 142 (34), 108 (52), 91 (100).
Preparation of di-tert-butyl (2S,3R)-2-benzyloxycarbonylamino-
3-methoxycarbonylglutarate (4a)
A solution of the diester 5 (1 g, 2.9 mmol) dissolved in dry THF
(25 cm3) was added dropwise to LiHMDS (6.5 cm3 of a 1 M
solution in THF, 6.5 mmol) at Ϫ80 ЊC, under nitrogen. After
stirring for 30 min at Ϫ78 ЊC, the pale-yellow solution was
warmed slowly to Ϫ30 ЊC and left for a further 2 h to ensure
formation of the dianion. The resulting orange solution was re-
cooled to Ϫ78 ЊC before the addition of neat tert-butyl chloro-
acetate (2.12 cm3, 14.8 mmol) at such a rate that the reaction
temperature did not exceed Ϫ75 ЊC. Stirring was continued for
a further 17 h at Ϫ78 ЊC before the reaction mixture was poured
into 1 M HCl (120 cm3) at rt. This solution was then extracted
using Et2O (4 × 50 cm3), and the organic phase was dried
(MgSO4) before removal of the solvent under reduced pressure.
The crude product was purified by “flash” chromatography29
(eluant: 4:1 petroleum ether–EtOAc) to give the reaction
product (639 mg, 48%) as the first material from the column,
followed by unreacted diester 5 (230 mg, 23%). The triester 4a
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-methoxy-
carbonylglutaric anhydride (7a)
Diacid 6a (60 mg, 0.17 mmol) was dissolved in Ac2O (5 cm3)
1034
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038

View Article Online
and the solution warmed to 60 ЊC. The reaction mixture was
stirred at this temperature for 4 h before cooling and removal
of the solvent under reduced pressure. The resulting white
solid was filtered and washed carefully with a minimum of dry
CH2Cl2 to give the desired anhydride 7a (46 mg, 80%) as a
white powder: mp 142–143 ЊC; [α]D20 ϩ3.0 (c 0.1 in CH3CN);
νmax(Nujol)/cmϪ1 3370, 1826, 1790, 1730, 1701 and 1535; δH(360
MHz; CD3CN; Me4Si) 3.05–3.20 (2 H, m, CHCH2), 3.45–3.55
(1 H, m, NCHCH), 3.72 (3 H, s, OCH3), 4.70 (1 H, dd, J 8.6
and 12.1, NCHCH), 5.16 (2 H, s, OCH2), 6.45 (1 H, d, J 8.6,
NH), 7.35–7.50 (5 H, m, Ph); δC(90.6 MHz; CD3CN; CD3CN)
32.50 (t), 43.30 (d), 52.40 (q), 54.06 (d), 66.70 (t), 127.80 (d),
128.10 (d), 128.50 (d), 145.80 (s), 156.10 (s), 164.30 (s), 170.60
(s), 172.10 (s); m/z (EI) 321 (Mϩ, 6%), 188 (14), 142 (13), 108
(20), 107 (29), 91 (100).
samples of the two ureas 9a and 9b. This method demonstrated
that the (2S,3R)-amine 8 was of >95% diastereoisomeric purity.
9a δH(500 MHz; CDCl3; Me4Si) 1.41 (9 H, s, (CH3)3C), 1.43
(9 H, s, (CH3)3C), 1.58 (3 H, d, J 7.0, CH3CH), 2.30 (1 H, m,
CH2CH), 2.51 (1 H, dd, J 9.8 and 17.1, CH2CH), 3.47 (1 H, dt,
J 4.9 and 8.8, NCHCH), 3.61 (3 H, s, OCH3), 4.08 (1 H, m,
NCHCH), 4.82 (1 H, q, J 7.0, CH3CH), 5.02 (1 H, br s,
C(᎐O)NH), 5.52 (1 H, br s, C(᎐O)NH), 7.20–7.40 (5 H, m, Ph);
᎐
᎐
δC(75.4 MHz; CDCl3; CDCl3) 22.38 (q), 27.72 (q), 27.91 (q),
33.80 (t), 44.08 (d), 50.22 (d), 51.73 (q), 54.46 (d), 80.75 (s),
82.40 (s), 125.19 (d), 126.67 (d), 128.48 (d), 142.32 (s), 156.92
(s), 170.27 (s), 170.65 (s), 172.47 (s).
9b δH(500 MHz; CDCl3; Me4Si) 1.41 (18 H, s, (CH3)3C), 1.58
(3 H, d, J 6.8, CH3CH), 2.44 (1 H, dd, J 4.9 and 16.6, CH2CH),
2.63 (1 H, dd, J 9.3 and 17.1, CH2CH), 3.49 (1 H, m, NCHCH),
3.63 (3 H, s, OCH3), 4.08 (1 H, m, NCHCH), 4.85 (1 H, q, J 6.8,
CH CH), 5.02 (1 H, br s, C(᎐O)NH), 5.48 (1 H, br s, C(᎐O)-
᎐
᎐
3
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-methoxy-
carbonylglutaric anhydride (7a) and (2S,3S)-2-benzyloxy-
carbonylamino-3-methoxycarbonylglutaric anhydride (7b)
NH), 7.28–7.38 (5 H, m, Ph); δC(75.4 MHz; CDCl3; CDCl3)
23.02 (q), 27.73 (q), 27.89 (q), 34.06 (t), 43.98 (d), 49.93 (d),
51.72 (q), 54.48 (d), 80.78 (s), 82.32 (s), 125.20 (d), 126.94 (d),
128.48 (d), 142.33 (s), 157.08 (s), 170.37 (s), 170.68 (s), 172.58 (s).
The mixture of diastereoisomeric triesters 4a and 4b was con-
verted to the corresponding mixture of cyclic anhydrides 7a and
1
7b using identical procedures to those outlined above. The H
Preparation of di-tert-butyl (2S,3R)-2-(4-bromophenyl)-
NMR resonances of the two anhydrides in the mixture could be
easily distinguished: 7b δH(360 MHz; CD3CN; Me4Si) 3.17–3.19
(2 H, m, CHCH2), 3.42–3.45 (1 H, m, NCHCH), 3.75 (3 H, s,
OCH3), 4.95 (1 H, dd, J 5.2 and 8.3, NCHCH), 5.18 (2 H, s,
OCH2), 6.35 (1 H, br s, NH), 7.35–7.50 (5 H, m, Ph).
carbonylamino-3-methoxycarbonylglutarate (10)
A suspension of 4-bromobenzoyl chloride (269 mg, 1.22 mmol)
in benzene (6 cm3) was added to a solution of the amine 8 (388
mg, 1.22 mmol) and triethylamine (0.17 cm3, 1.22 mmol) in
benzene (10 cm3). The reaction mixture was stirred vigorously
for 20 h before the solvent was removed under reduced pressure.
The residue was suspended in EtOAc and the solid removed
by filtration. The crude diester, obtained as an oil after removal
of the solvent, was then purified by column chromatography
on silica (eluant: 85:15 petroleum ether–EtOAc). The desired
diester 10 (356 mg, 58%) was isolated as a colorless oil: [α]D20
ϩ31.1 (c 3.2 in CH2Cl2); νmax(neat)/cmϪ1 3419, 2980, 1732, 1668,
1480, 1369 and 1153; δH(300 MHz; CDCl3; Me4Si) 1.43 (9 H, s,
(CH3)3C), 1.48 (9 H, s, (CH3)3C), 2.55 (1 H, dd, J 6.0 and 17.1,
CH2CH), 2.67 (1 H, dd, J 8.4 and 17.1, CH2CH), 3.65 (1 H, m,
NCHCH), 3.74 (3 H, s, OCH3), 5.02 (1 H, dd, J 3.4 and 8.5,
NCHCH), 6.96 (1 H, d, J 8.5, NH), 7.57 (2 H, m, 3Ј,5Ј-PhH2),
7.80 (2 H, m, 2Ј,6Ј-PhH2); δC(75.4 MHz; CDCl3; CDCl3) 27.82
(q), 27.95 (q), 34.32 (t), 43.35 (d), 52.01 (q), 53.37 (d), 81.25 (s),
83.03 (s), 126.48 (s), 128.73 (d), 131.76 (d), 132.63 (s), 166.36
(s), 169.02 (s), 170.28 (s), 172.61 (s); m/z (CI: NBA) 500.1150
(MHϩ, 24%. C22H31BrNO7 requires 500.1280), 444 (10), 390
(100), 388 (99), 185 (21).
Preparation of di-tert-butyl (2S,3R)-2-amino-3-methoxy-
carbonylglutarate (8)
10% Pd/C (20 mg) was added to a solution of the triester 6a
(630 mg, 1.4 mmol) dissolved in cyclohexene (20 cm3) and
MeOH (40 cm3), and the resulting heterogeneous mixture
refluxed for 28 h. The catalyst was removed by filtration
through Celite and subsequently washed well with MeOH (400
cm3). Solvent from the combined filtrate and washings was then
removed under reduced pressure to yield a thick oil that was
re-dissolved in EtOAc (30 cm3). After washing with brine, the
organic layer was dried (MgSO4) and the solvent removed to
give the desired amine 8 (419 mg, 95%) as a white solid that was
used without further purification: mp 76–78 ЊC; [α]D20 ϩ4.7 (c
0.01 in CHCl3); νmax(CHCl3)/cmϪ1 3464, 2984, 2909, 1743, 1447,
1372 and 1243; δH(360 MHz; CDCl3; Me4Si) 1.44 (9 H, s,
(CH3)3C), 1.47 (9 H, s, (CH3)3C), 2.71 (1 H, dd, J 5.5 and 16.9,
CH2CH), 2.82 (1 H, dd, J 8.3 and 16.9, CH2CH), 3.52 (1 H, m,
CHCH2), 3.72 (3 H, s, OCH3), 4.00 (1 H, d, J 2.0, NCHCH),
4.32–4.50 (2 H, br s, NH); δC(75.4 MHz; CDCl3; CDCl3) 28.31
(q), 28.43 (q), 34.33 (t), 45.58 (d), 52.27 (q), 55.90 (d), 81.44 (s),
82.47 (s), 171.54 (s), 172.78 (s), 172.86 (s); m/z (CI: NH4Cl)
318.1917 (MHϩ, 100%. C15H28NO6 requires 318.1917), 262 (19),
160 (10).
Preparation of tert-butyl (2S,3R)-2-(4-bromophenyl)carbonyl-
amino-3-methoxycarbonylglutaric acid monoester (11) and
(2S,3R)-2-(4-bromophenyl)carbonylamino-3-methoxy-
carbonylglutaric acid (12)
TFA (0.46 cm3, 6 mmol) was added dropwise to a solution of
the triester 10 (300 mg, 0.6 mmol) in CH2Cl2 (20 cm3) at rt. The
reaction was stirred for 5 days before the volatile components
were removed under reduced pressure. The resulting oily resi-
due was then dissolved in EtOAc to give a solution that was
extracted with 5% w/v aq. Na2CO3. The combined aqueous
extracts were then acidified to pH 1 using 6 M HCl, and then
extracted with EtOAc. After drying (Na2SO4) the solvent was
removed under reduced pressure to give a brown oil which was
purified using reversed-phase HPLC with gradient elution (C18
column; solvent flow rate 12 cm3 minϪ1; 80:20:1 H2O–CH3-
CN–TFA to 20:80:1 H2O–CH3CN–TFA over 30 min). The
monoacid 11 (49 mg, 19%) was obtained as a white solid after
lyophilization: mp 56–57 ЊC; νmax(CHCl3)/cmϪ1 3542–2835,
3423, 3019, 1737, 1667, 1592, 1516, 1480 and 1370; δH(300
MHz; CDCl3; Me4Si) 1.48 (9 H, s, (CH3)3C), 2.63 (1 H, dd, J 6.9
and 17.3, CH2CH), 2.76 (1 H, dd, J 7.6 and 17.3, CH2CH),
3.71 (1 H, m, NCHCH), 3.74 (3 H, s, OCH3), 5.10 (1 H, dd,
Enantiomeric purity of di-tert-butyl (2S,3R)-2-amino-3-methoxy-
carbonylglutarate (8)
Neat (S)-α-methylbenzyl isocyanate (0.12 cm3, 0.84 mmol) was
added to a solution of amine 8 (80 mg, 0.25 mmol) in dry THF
(10 cm3) under an inert atmosphere of dry N2. In a separate
reaction, neat (R)-α-methylbenzyl isocyanate (0.16 cm3, 1.1
mmol) was added to a solution of triester 6a (130 mg, 0.4
mmol) in dry THF (10 cm3) under an inert atmosphere of dry
N2. After being heated to reflux for 3 h, the volatile components
of the resulting reaction mixtures were removed under reduced
pressure. The desired urea derivatives, 9a and 9b, respectively,
were obtained as oily residues that were examined directly by 1H
1
NMR. The methyl ester singlets of each urea in the H NMR
spectrum were clearly differentiated, and the limits of detection
were determined by incremental doping experiments using pure
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038
1035

View Article Online
J 3.4 and 8.3, NCHCH), 7.01 (1 H, d, J 8.3, NH), 7.57 (2 H, m,
3Ј,5Ј-PhH2), 7.67 (2 H, m, 2Ј,6Ј-PhH2), 8.10–8.70 (1 H, br s,
CO2H); δC(75.4 MHz; CDCl3; CDCl3) 27.82 (q), 32.61 (t), 43.39
(d), 52.25 (q), 53.26 (d), 83.48 (s), 126.78 (s), 128.77 (d), 131.88
(d), 132.29 (s), 166.95 (s), 168.69 (s), 172.22 (s), 175.12 (s); m/z
(FAB: NBA) 444.0654 (MHϩ, 17%. C18H23BrNO7 requires
444.0658), 390 (100), 388 (91), 185 (48), 183 (48).
On continued elution, the diacid 12 (38 mg, 16%) was also
obtained as a white solid after lyophilization: mp 155–157 ЊC;
νmax(CHCl3)/cmϪ1 3500–3000, 3620, 3541, 1742, 1670, 1633,
and 1482; δH(300 MHz; CD3CN; Me4Si) 2.63 (1 H, dd, J 6.6
and 17.4, CH2CH), 2.81 (1 H, dd, J 7.4 and 17.4, CH2CH),
3.20–4.00 (1 H, br s, CO2H), 3.58 (1 H, m, NCHCH), 3.68 (3 H,
s, OCH3), 5.06 (1 H, dd, J 4.1 and 8.4, NCHCH), 7.30 (1 H, d,
J 7.8, NH), 7.64 (2 H, m, 3Ј,5Ј-PhH2), 7.71 (2 H, m, 2Ј,6Ј-
PhH2), 8.10–8.70 (1 H, br s, CO2H); δC(75.4 MHz; CD3CN;
CD3CN) 32.31 (t), 43.77 (d), 52.80 (q), 84.07 (d), 128.77 (s),
130.08 (d), 132.58 (d), 133.88 (s), 167.43 (s), 170.00 (s), 173.08
(s), 177.08 (s); m/z (FAB: NBA) 388.0018 (MHϩ, 15%. C14H25-
BrNO7 requires 388.0032), 289 (8), 154 (100), 136 (70).
8.80 (3 H, br s, CO2H); δC(75.4 MHz; CD3OD; CD3OD) 32.01
(t), 43.18 (d), 54.15 (d), 66.49 (t), 127.36 (d), 127.58 (d), 128.03
(d), 137.10 (s), 157.19 (s), 172.06 (s), 172.70 (s), 173.56 (s); m/z
(FAB) 326.0860 (MHϩ, 41%. C14H16NO8 requires 326.0880),
282 (29), 93 (100).
Continued elution gave a small amount of the diastereo-
isomeric triacid 14b (16 mg, 7%) as a white solid (retention time
13.7 min) after lyophilization: δH(300 MHz; CD3CN; Me4Si)
2.55 (1 H, dd, J 4.4 and 17.2, CH2CH), 2.72 (1 H, dd, J 9.4 and
17.2, CH2CH), 3.30 (1 H, m, NCHCH), 4.66 (1 H, dd, J 5.0 and
9.2, NCHCH), 5.09 (2 H, s, OCH2), 5.50–6.00 (3 H, br s,
CO2H), 6.12 (1 H, d, J 9.0, NH), 7.30-7.40 (5 H, m, Ph).
Preparation of (4S)-3-benzyloxycarbonyl-4-[(1R)-methoxy-
carbonyl-2-carboxyethyl)-5-oxooxazolidine (15)
A suspension of diacid 6a (100 mg, 0.3 mmol), paraformalde-
hyde (65 mg) and pTsOH (6 mg) in toluene (50 cm3) was
refluxed, with azeotropic removal of water using activated 4 Å
molecular sieves, for 4 h. After this time, a further portion of
paraformaldehyde (80 mg) was added and refluxing continued
for 1.5 h. After cooling, the reaction mixture was loaded onto a
silica column and eluted initially with toluene (100 cm3) and
then EtOAc (400 cm3). Removal of the solvent from the com-
bined EtOAc fractions then yielded a yellow oil that crystallized
on standing. The desired acid 15 (90 mg, 88%) was then
obtained as a white solid after recrystallization from Et2O–
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-carbamoyl-
glutaric acid (13)
Diacid 6a (200 mg, 0.59 mmol ) was dissolved in conc. NH4OH
(4 cm3) at 0 ЊC. The reaction mixture was then warmed to rt,
and stirred for 4 days before being poured into 1 M aq. HCl
(50 cm3). This solution was then extracted with EtOAc (4 × 20
cm3), and the combined organic fractions dried (MgSO4).
Removal of the solvent gave a white solid that was purified by
reversed-phase HPLC with gradient elution (C18 column; sol-
vent flow rate 10 cm3 minϪ1; 80:20:1 H2O–CH3CN–TFA to
20:80:1 H2O–CH3CN–TFA over 30 min). The desired diacid
13 (48 mg, 25%) was obtained as a white solid after lyophiliz-
ation: mp 145–147 ЊC (Found: C, 51.8; H, 4.9; N, 8.4; C14H16-
N2O7 requires C, 51.8; H, 5.0; N, 8.6%); νmax(CHCl3)/cmϪ1
3550–2900, 3423, 3340, 3019, 1739, 1693, 1650, 1531 and 1275;
δH(300 MHz; CD3OD; Me4Si) 2.57 (2 H, ddd, J 7.2, 7.4 and
17.2, CH2CH), 3.34 (1 H, m, NCHCH), 4.42 (1 H, d, J 5.0,
NCHCH), 5.12 (2 H, s, OCH2), 7.15–7.29 (5 H, s, Ph); δC(90.6
MHz; CD3OD; CD3OD) 33.27 (t), 42.39 (d), 54.59 (d), 66.50
(t), 127.37 (d), 127.61 (d), 128.05 (d), 135.55 (s), 157.19 (s),
171.90 (s), 172.10 (s), 173.22 (s); m/z (FAB: Gly, TFA) 325.1004
(MHϩ, 100%. C14H17N2O7 requires 325.1035), 281 (20), 207
(30), 186 (25).
EtOAc: mp 156–158 ЊC; [α]D20 ϩ66.5 (c 0.2 in CHCl3); νmax
-
(CHCl3)/cmϪ1 3550–2900, 1813, 1748, 1705 and 1415; δH(300
MHz; CD3CN; Me4Si) 2.75 (1 H, m, CH2CH), 2.85 (1 H, m,
CH2CH), 3.62 (1 H, dt, J 2.7 and 5.8, NCHCH), 3.72 (3 H, s,
OCH3), 4.68 (1 H, d, J 2.7, NCHCH), 5.18 (2 H, s, OCH2Ph),
5.22 (2 H, s, OCH2N), 5.58 (1 H, d, J 4.3, NH), 7.40 (5 H, s, Ph);
δC(74.5 MHz; CD3CN; CD3CN) 32.66 (t), 43.42 (d), 52.89 (q),
55.85 (d), 66.77 (t), 79.03 (t), 128.67 (d), 128.87 (d), 128.92 (d),
135.05 (s), 155.16 (s), 171.18 (s), 172.57 (s), 176.36 (s); m/z (EI)
352.1019 (MHϩ, 5%. C16H18NO8 requires 352.1032), 198
(10),170 (18), 156 (8), 91 (100).
Preparation of triphenylmethyl (2S)-2-benzyloxycarbonylamino-
3-methoxycarbonylpropionate (17)
(2S)-2-benzyloxycarbonylamino-3-methoxycarbonylpropionic
acid (750 mg, 2.67 mmol) was dissolved in aq. Cs2CO3 (435 mg
in 20 cm3). The caesium salt was then isolated as a glassy solid
after removal of the solvent by lyophilization. This solid was
redissolved in dry THF (20 cm3) together with triphenylmethyl
bromide (970 mg, 3 mmol), and the solution heated to reflux for
3.5 h. After cooling to rt and filtration, the solvent was removed
to give the diester 17 (1.35 g, 98%) as a colorless oil: [α]D20 Ϫ14.1
(c 6.7 in CHCl3); νmax(CHCl3)/cmϪ1 3427, 3060, 3052, 1732,
1494, and 1448; δH(300 MHz; CDCl3; Me4Si) 2.80–3.07 (2 H,
dd, J 4.5 and 16.7, CHCH2CO), 3.54 (3 H, s, OCH3), 4.68–4.76
(1 H, m, NCHCH2), 5.07 (2 H, m), 5.89 (1 H, d, J 9.0, NH),
7.20–7.37 (20 H, m, Ph); δC(75.4 MHz; CDCl3; CDCl3) 36.25
(t), 51.54 (d), 51.91 (q), 67.16 (t), 82.02 (s), 92.01 (s), 127.21 (d),
127.52 (d), 127.90 (d), 128.02 (d), 128.16 (d), 128.44 (d), 142.76
(s), 147.07 (s), 168.39 (s), 171.04 (s); m/z (FAB: TFA) 524
(MHϩ, 2%), 324 (17), 282 (100), 243 (85), 238 (50), 185 (85).
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-carboxy-
glutaric acid (14a) and (2S,3S)-2-benzyloxycarbonylamino-3-
carboxyglutaric acid (14b)
The N-protected diacid 6a (249 mg, 0.73 mmol) was dissolved
in a 4:1 mixture of MeOH–H2O so as to yield an 0.1 M solu-
tion. Four equivalents of LiOHؒH2O were then added and the
reaction stirred at rt until complete consumption of starting
material had occurred. After concentration under reduced pres-
sure, the residue was added to an excess of 1 M aq. HCl and the
resulting solution extraced well with EtOAc. The combined
organic extracts were then dried (MgSO4) before removal of the
solvent yielded the crude triacid as an oil. Purification of the
product was accomplished using reversed-phase HPLC with
gradient elution (C8 column; solvent flow rate 20 cm3 minϪ1
;
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-(prop-2-
enyl)succinic anhydride (22a)
80:20:1 H2O–CH3CN–TFA to 60:40:1 H2O–CH3CN–TFA
over 40 min). The desired triacid 14a (159 mg, 67%) was
obtained as a white solid (retention time 13.3 min) after
lyophilization (Found: C, 51.6; H, 4.6; N, 4.2; C14H15NO8
requires C, 51.7; H, 4.6; N, 4.3%); νmax(CHCl3)/cmϪ1 3713–
2319, 1714, 1518, 1416, 1345, 1216, 1063 and 758; δH(300 MHz;
CD3CN; Me4Si) 2.55 (1 H, dd, J 7.3 and 19.3, CH2CH), 2.72
(1 H, dd, J 7.0 and 17.5, CH2CH), 3.46–3.53 (1 H, m,
NCHCH), 4.63 (1 H, dd, J 3.6 and 9.2, NCHCH), 5.12 (2 H, s,
OCH2), 5.93 (1 H, d, J 9.0, NH), 7.30–7.40 (5 H, m, Ph), 7.80–
The N-protected diacid 21a (60 mg, 0.29 mmol) was dissolved
in Ac2O (5 cm3), and the solution heated at 75 ЊC for 3 h.
Removal of the solvent under reduced pressure then gave the
desired anhydride 22a (54 mg, 99%) as a colorless oil: [α]D20 ϩ73.5
(c 0.68 in CH3CN); νmax(neat)/cmϪ1 3366, 3036, 2931, 1872,
1790, 1728, 1520, 1455, 1227 and 1220; δH(300 MHz; CD3CN;
Me4Si) 2.20–2.30 (1 H, m, CHCH2CH), 2.45–2.55 (1 H, m,
CHCH2CH), 3.30–3.40 (1 H, m, NCHCH), 4.60 (1 H, dd, J 7.8
1036
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038

View Article Online
and 10.2, NCHCH), 5.03–5.15 (4 H, m, OCH Ph and CH ᎐
product was then purified by “flash” chromatography (eluant:
᎐
2
2
CH), 5.78–5.92 (1 H, m, CH ᎐CH), 6.52 (1 H, d, J 7.8, NH),
17:3 petroleum ether–EtOAc) to give the diester 26 as a mix-
ture of diastereoisomers (115 mg, 70%). H NMR analysis of
this oil showed a 4.5:1 ratio of 26a:26b. Careful column chrom-
atography gave the pure diester 26a (67 mg, 41%) as a clear oil:
[α]D20 ϩ31.2 (c 0.64 in CHCl3); νmax(CHCl3)/cmϪ1 3413, 3296,
᎐
2
1
7.31–7.42 (5 H, m, Ph); δC(75.4 MHz; CD3CN; CD3CN) 28.82
(t), 42.62 (d), 52.35 (d), 67.21 (t), 116.68 (t), 127.95 (d), 128.22
(d), 128.51 (d), 134.43 (d), 136.40 (s), 156.70 (s), 170.63 (s),
172.32 (s); m/z (EI) 289.0935 (Mϩ, 2%. C15H15NO5 requires
289.0950), 218 (9), 172 (12), 108 (64), 91 (100).
2979, 1732, 1713, 1455, 1337, 1156 and 1056; δH(300 MHz;
᎐
CDCl ; Me Si) 1.45 (9 H, s, (CH ) C), 2.16 (1 H, m, C᎐CH),
᎐
3
4
3 3
2.49 (1 H, ddd, J 2.6, 8.6 and 17.0, CH2CH), 2.66 (1 H, ddd,
J 2.6, 6.5 and 17.0, CH2CH), 3.32 (1 H, m, CCH), 3.71 (3 H, s,
OCH3), 4.88 (1 H, dd, J 3.2 and 9.4, NCHCH), 5.12 (2 H, s,
OCH2), 5.60 (1 H, d, J 9.4, NH), 7.30–7.45 (5 H, s, Ph); δC(75.4
MHz; CDCl3; CDCl3) 18.21 (t), 27.80 (q), 46.34 (d), 52.02 (q),
54.72 (d), 67.10 (t), 70.66 (s), 81.44 (s), 82.72 (s), 128.01 (d),
128.06 (d), 128.42 (d), 136.27 (s), 156.27 (s), 169.14 (s), 171.68
(s); m/z (CI: NBA) 376.1761 (MHϩ, 37%. C20H26NO6 requires
376.1760), 320 (100), 276 (57), 154 (42).
Preparation of (2S,3R)-2-benzyloxycarbonylamino-3-(prop-2-
enyl)succinic anhydride (22a) and (2S,3S)-2-benzyloxycarb-
onylamino-3-(prop-2-enyl)succinic anhydride (22b)
A mixture of the diastereoisomeric acids 21a and 21b was
treated with Ac2O as described above for the pure diastereo-
isomer 21a. The product mixture of anhydrides 22a and 22b
1
was analyzed using H NMR spectroscopy. 22b: δH(300 MHz;
CD3CN; Me4Si) 2.45–2.57 (1 H, m, CHCH2CH), 2.61–2.72
(1 H, m, CHCH2CH), 3.37–3.47 (1 H, dt, J 5.4 and 8.1,
NCHCH), 4.42 (1 H, dd, J 8.0 and 8.1, NCHCH), 5.05–5.30
26b: δH(300 MHz; CDCl3; Me4Si) 1.45 (9 H, s, (CH3)3C), 2.16
᎐
(1 H, m, C᎐CH), 2.49 (1 H, ddd, J 2.6, 8.6 and 17.0, CH CH),
᎐
2
(4 H, m, OCH Ph and CH ᎐CH), 5.75–5.90 (1 H, m, CH ᎐CH),
᎐
᎐
2
2
2
2.66 (1 H, ddd, J 2.6, 6.5 and 17.0, CH2CH), 3.32 (1 H, m,
CCH), 3.71 (3 H, s, OCH3), 4.88 (1 H, dd, J 3.2 and 9.4,
NCHCH), 5.12 (2 H, s, OCH2), 5.60 (1 H, d, J 9.4, NH), 7.30–
7.45 (5 H, s, Ph).
6.49 (1 H, d, J 8.0, NH), 7.35–7.45 (5 H, m, Ph).
Preparation of tert-butyl (2S,3R)-2-benzyloxycarbonylamino-3-
methoxycarbonyl-4-cyanobutyrate (25a) and tert-butyl (2S,3S)-
2-benzyloxycarbonylamino-3-methoxycarbonyl-4-cyanobutyrate
(25b)
Preparation of tert-butyl (2S,3R,4R)-2-benzyloxycarbonyl-
amino-3-methoxycarbonyl-4-hydroxy-4-phenylbutanoate (27a)
and (2S,3R,4S)-2-benzyloxycarbonylamino-3-methoxy-
carbonyl-4-hydroxy-4-phenylbutanoate (27b)
A solution of the diester 5 (150 mg, 0.45 mmol) dissolved in dry
THF (8 cm3) was added dropwise to LiHMDS (1.2 cm3 of a
1 M solution in THF, 1.2 mmol) at Ϫ80 ЊC, under nitrogen.
After stirring for 30 min at Ϫ78 ЊC, the pale-yellow solution
was warmed slowly to Ϫ30 ЊC and left for a further 2 h to
ensure formation of the dianion. The resulting orange solution
was re-cooled to Ϫ78 ЊC before the addition of neat bromo-
acetonitrile (0.25 cm3, 2.5 mmol) at such a rate so that the reac-
tion temperature did not exceed Ϫ75 ЊC. Stirring was continued
for a further 16 h at Ϫ78 ЊC before the reaction mixture was
poured into 1 M HCl (120 cm3) at rt. Standard workup pro-
cedures, as outlined above, followed by “flash” chromatography
(eluant: 17:3 petroleum ether–EtOAc) gave the diester 25 as an
inseparable mixture of diastereoisomers (116 mg, 69%). 1H
NMR analysis of this oil showed a 2:1 ratio of 25a:25b:
νmax(neat)/cmϪ1 3357, 2980, 2251, 1742, 1732, 1513, 1372, 1227
and 1155; δH(300 MHz; CDCl3; Me4Si) 1.47 (9 H, s, (CH3)3C),
2.60–2.94 (2 H, m, CH2CH), 3.44 (1 H, m, NCHCH), 3.75 (3 H,
s, OCH3), 4.60–4.70 (1 H, m, NCHCH), 5.13 (2 H, s, OCH2),
5.63 (1 H, d, J 8.0, NH), 7.34–7.49 (5 H, s, Ph); δC(75.4 MHz;
CDCl3; CDCl3) 16.79 (t), 27.80 (q), 44.51 (d), 52.75 (d), 54.91
(q), 67.46 (t), 83.99 (s), 117.45 (s), 128.20 (d), 128.33 (d), 128.56
(d), 135.79 (s), 156.47 (s), 167.91 (s), 167.96 (s); m/z (CI: CH4)
377.1731 (MHϩ, 10%. C19H25N2O6 requires 377.1713), 320
(MHϩ Ϫ C4H9, 8), 231 (6), 185 (4), 141 (5), 91 (100).
A solution of the diester 5 (135 mg, 0.40 mmol) dissolved in dry
THF (6 cm3) was added dropwise to LiHMDS (0.46 cm3 of a
2 M solution in THF, 0.9 mmol) at Ϫ75 ЊC, under nitrogen.
After stirring for 30 min at Ϫ78 ЊC, the pale-yellow solution
was warmed slowly to Ϫ30 ЊC and left for a further 2 h to
ensure formation of the dianion. The resulting orange solution
was re-cooled to Ϫ78 ЊC before the addition of neat benzalde-
hyde (0.25 cm3, 2.5 mmol) at such a rate that the reaction tem-
perature did not exceed Ϫ75 ЊC. Stirring was continued for a
further 27 h at Ϫ78 ЊC before the reaction mixture was worked
up using standard procedures, as outlined above. “Flash”
chromatography (eluant: 3:1 petroleum ether–EtOAc) gave the
diester 27 as an inseparable mixture of diastereoisomers (118
mg, 67%): νmax(CH2Cl2)/cmϪ1 3414, 3053, 2979, 1727, 1505 and
1264; δH(300 MHz; CDCl3; Me4Si) 1.36 (9 H, s, (CH3)3C), 3.27
(1 H, m, NCHCH), 3.57 (3 H, s, OCH3), 4.25 (1 H, m, CH-
(OH)CH), 4.93 (1 H, d, J 8.0, NCHCH), 5.05 (2 H, s, OCH2),
5.77 (1 H, d, J 8.0, NH), 7.20–7.35 (10 H, m, Ph); δC(75.4 MHz;
CDCl3; CDCl3) 27.85 (q), 52.12 (q), 53.62 (d), 54.49 (d), 67.09
(t), 72.39 (d), 83.05 (s), 126.04 (d), 126.60 (d), 128.13 (d), 128.34
(d), 128.47 (d), 128.69 (d), 136.21 (s), 140.66 (s), 156.04 (s),
169.15 (s), 172.31 (s); m/z (CI: NH4ϩ) 444.2040 (MHϩ, 12%.
C24H30NO7 requires 444.2020), 413 (100), 370 (81), 326 (39), 91
(80).
Preparation of tert-butyl (2S,3R)-2-benzyloxycarbonylamino-3-
methoxycarbonylhex-5-ynoate (26a) and tert-butyl (2S,3S)-2-
benzyloxycarbonylamino-3-methoxycarbonylhex-5-ynoate (26b)
Preparation of tert-butyl (2S,3R,4R)-2-benzyloxycarbonyl-
amino-3-methoxycarbonyl-4-hydroxyhex-5-enoate (28a) and
(2S,3R,4S)-2-benzyloxycarbonylamino-3-methoxycarbonyl-4-
hydroxyhex-5-enoate (28a)
A solution of the diester 5 (150 mg, 0.45 mmol) dissolved in dry
THF (6 cm3) was added dropwise to LiHMDS (0.94 cm3 of a
1 M solution in THF, 0.93 mmol) at Ϫ80 ЊC, under nitrogen.
After stirring for 30 min at Ϫ78 ЊC, the pale-yellow solution
was warmed slowly to Ϫ30 ЊC and left for a further 2 h to
ensure formation of the dianion. The resulting orange solution
was re-cooled to Ϫ78 ЊC before the addition of propargyl brom-
ide (prop-2-ynyl bromide), as an 80% w/v toluene solution (331
mg, 2.23 mmol) at such a rate that the reaction temperature did
not exceed Ϫ75 ЊC. Stirring was continued for a further 18 h at
Ϫ78 ЊC before the reaction mixture was poured into 1 M HCl
(120 cm3) at rt. This solution was then extracted using Et2O
(4 × 50 cm3) and the organic phase was dried (MgSO4) before
removal of the solvent under reduced pressure. The crude
A solution of the diester 5 (200 mg, 0.60 mmol) dissolved in dry
THF (6 cm3) was added dropwise to LiHMDS (1.8 cm3 of a
1 M solution in THF, 1.8 mmol) at Ϫ75 ЊC, under nitrogen.
After stirring for 30 min at Ϫ78 ЊC, the pale-yellow solution
was warmed slowly to Ϫ30 ЊC and left for a further 2 h to
ensure formation of the dianion. The resulting orange solution
was re-cooled to Ϫ78 ЊC before the addition of neat acrolein
(0.2 cm3, 3 mmol) at such a rate so that the reaction temperature
did not exceed Ϫ75 ЊC. Stirring was continued for a further 27 h
at Ϫ78 ЊC before the reaction mixture was worked up using the
standard procedures outlined above. “Flash” chromatography
(eluant: 4:1 petroleum ether–EtOAc) gave the diester 28 as an
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038
1037

View Article Online
J. E. Baldwin, R. M. Adlington, D. W. Gollins and C. R. A. Godfrey,
Tetrahedron, 1995, 51, 5169; (c) N. A. Sasaki, R. Pauly, C. Fontaine,
A. Chiaroni, C. Riche and P. Potier, Tetrahedron Lett., 1994, 35, 241;
(d) W. O. Moss, A. C. Jones, R. Wisedale, M. F. Mahon, K. C.
Molloy, R. H. Bradbury, N. J. Hales and T. Gallagher, J. Chem.
Soc., Perkin Trans. 1, 1992, 2615; (e) T. R. Webb and C. Eigenbrot,
J. Org. Chem., 1991, 56, 3009; ( f ) C. Herdeis, H. P. Hubmann and
H. Lotter, Tetrahedron: Asymmetry, 1994, 5, 351; (g) P. M. Esch,
H. Hiemstra, R. F. de Boer and W. N. Speckamp, Tetrahedron, 1992,
48, 4659.
inseparable mixture of diastereoisomers (126 mg, 54%):
νmax(neat)/cmϪ1 3418, 2954, 1728, 1512, 1455, 1369, 1227 and
1157; δH(300 MHz; CDCl3; Me4Si) 1.45 (9 H, s, (CH3)3C), 3.05
(1 H, m, NCHCH), 3.40 (1 H, m, OH), 3.70 (3 H, s, OCH3),
4.47 (1 H, m, CH(OH)CH), 4.64 (1 H, m, NCHCH), 5.12 (2 H,
s, OCH ), 5.13–5.40 (2 H, m, H C᎐CH), 5.69–5.99 (2 H, m, NH
᎐
2
2
and H C᎐CH), 7.30–7.38 (5 H, s, Ph); δ (75.4 MHz; CDCl ;
᎐
2
C
3
CDCl3) 27.84 (q), 52.07 (d), 52.47 (q), 53.29 (d), 67.20 (t), 70.05
(d), 83.01 (s), 117.03 (t), 128.10 (d), 128.20 (d), 128.49 (d),
136.55 (d), 137.63 (s), 169.23 (s), 171.70 (s), 172.08 (s); m/z (CI:
NH4ϩ) 394.1864 (MHϩ, 87%. C20H28NO7 requires 394.1863),
320 (MHϩ Ϫ C4H9, 8), 231 (6), 185 (4), 141 (5), 91 (100).
10 (a) G. W. Anderson and F. M. Callahan, J. Am. Chem. Soc., 1960,
82, 3359; (b) H. Schwarz, F. M. Bumpus and I. H. Page, J. Am.
Chem. Soc., 1957, 79, 5697; (c) D. Colman, J. Chem. Soc., 1951,
2294.
11 M. W. Rathke, J. Am. Chem. Soc., 1970, 92, 3222.
12 M. Bergmann and L. Zervas, Chem. Ber., 1932, 65, 1192.
13 (a) R. B. Silverman and M. W. Holladay, J. Am. Chem. Soc., 1981,
103, 7357; (b) V. F. Bystrov, Progr. Nucl. Magn. Reson. Spectrosc.,
1976, 11, 41.
14 (a) A. M. P. Koskinen and H. Rapoport, J. Org. Chem., 1989, 54,
1859; (b) H. H. Ibrahim and W. D. Lubell, J. Org. Chem., 1993, 58,
6438.
15 A. B. Dribben and N. G. J. Richards, unpublished results.
16 (a) M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902; (b) J. J. P. Stewart, J. Comput.
Chem., 1989, 10, 209.
Acknowledgements
This work was funded by the National Cancer Institute
(CA28725), National Institutes of Health, DHHS and the
American Cancer Society, Florida Affiliate. Studentship sup-
port for our early investigations was provided by the Science
and Engineering Research Council (U.K.). Additional support
was also provided by Pfizer Central Research, U.K. Finally, one
of us (N. G. J. R.) would like to acknowledge the scientific and
intellectual debt that he owes Professor Ralph A. Raphael, to
the memory of whom this paper is dedicated.
17 (a) P. L. Ornstein, A. Melikian and M. J. Martinelli, Tetrahedron
Lett., 1994, 35, 5759; (b) A. F. Parsons, Tetrahedron, 1996, 52, 4149;
(c) S. Hanessian and S. Ninkovic, J. Org. Chem., 1996, 61, 5418;
(d) J. E. Baldwin, A. M. Fryer, M. R. Spyvee, R. C. Whitehead and
M. E. Wood, Tetrahedron, 1997, 53, 5273; (e) Z. Q. Gu, X. F. Lin
and D. P Hesson, Bioorg. Med. Chem. Lett., 1995, 5, 1973;
( f ) I. Jako, P. Uiber, A. Mann, C.-G. Wermuth, T. Boulanger,
B. Norberg, G. Evrard and F. Durant, J. Org. Chem., 1991, 56, 5729;
(g) H. Yoda, T. Shirai, T. Katagiri, K. Takabe, K. Kimata and
K. Hosoya, Chem. Lett., 1990, 2037; (h) A. Vidal-Cros, M. Gaudry
and A. Marquet, J. Org. Chem., 1989, 54, 498; (i) T. Kunieda,
T. Ishizuka, T. Higuchi and M. Hirobe, J. Org. Chem., 1988,
53, 3381; (j) U. Schöllkopf, D. Pettig, U. Busse, E. Egret and
M. Dyrbusch, Synthesis, 1986, 737.
18 (a) R. J. Bridges, M. S. Stanley, M. W. Anderson, C. W. Cotman
and A. R. Chamberlin, J. Med. Chem., 1991, 34, 717; (b) M. G.
Moloney, Nat. Prod. Rep., 1998, 15, 205; (c) I. Collado, J. Ezquerra,
A. Mazon, C. Pedregal, B. Yruretagoyena, A. E. Kingston, R.
Tomlinson, R. A. Wright, B. G. Johnson and D. D. Schoepp, Bioorg.
Med. Chem. Lett., 1998, 8, 2849; (d) K. Shimamoto and Y. Ohfune,
J. Med. Chem., 1996, 40, 407; (e) P. L. Ornstein, M. B. Arnold,
N. K. Allen, T. Bleisch, P. S. Borromeo, C. W. Lugar, J. D. Leander,
D. Lodge and D. D. Schoepp, J. Med. Chem., 1996, 39, 2219; ( f )
D. J. Chalmers, D. E. Jane, D. C. Sunter, I. C. Kilpatrick, G. A.
Thompson, P. M. Udvarhelyi and J. C. Watkins, Neuropharma-
cology, 1995, 34, 1589; (g) J. A. Monn, M. J. Valli, B. G. Johnson,
C. R. Salhoff, R. A. Wright, T. Howe, A. Bond, D. Lodge, L. A.
Spangle, J. W. Paschal, J. B. Campbell, K. Griffey, J. P. Tizzano and
D. D. Schoepp, J. Med. Chem., 1996, 39, 2990.
References
1 Presented in part at the 201st National Meeting of the American
Chemical Society, Atlanta, GA, April 1991. Abstracts of Papers,
ORGN-227.
2 (a) F. J. Sardina and H. Rapoport, Chem. Rev., 1996, 96, 1825; (b)
R. O. Duthaler, Tetrahedron, 1994, 50, 1539; (c) R. M. Williams, The
Synthesis of Optically Active α-Amino Acids, Pergamon, Oxford,
1989; (d) Y. Ohfune, Acc. Chem. Res., 1992, 25, 360.
3 (a) S. Hanessian, G. McNaughton-Smith, H.-G. Lombart and
W. D. Lubell, Tetrahedron, 1997, 53, 12789; (b) J. Gante, Angew.
Chem., Int. Ed. Engl., 1994, 33, 1699; (c) A. Giannis, Angew. Chem.,
Int. Ed. Engl., 1993, 32, 1244; (d) A. S. Ripka and D. H. Rich,
Curr. Opin. Chem. Biol., 1998, 2, 441; (e) M. G. Hinds, J. H. Welsh,
D. M. Brennand, J. Fisher, M. J. Glennie, N. G. J. Richards,
D. L. Turner and J. A. Robinson, J. Med. Chem., 1991, 34, 1777.
4 (a) M. R. Spaller, M. T. Burger, M. Fardis and P. A. Bartlett, Curr.
Opin. Chem. Biol., 1997, 1, 47; (b) M. A. Marx, A. L. Grillot,
C. T. Louer, K. A. Beaver and P. A. Bartlett, J. Am. Chem. Soc.,
1997, 119, 6153; (c) S. Booth, P. H. H. Hermkens, H. C. J.
Ottenheijm and D. C. Rees, Tetrahedron, 1998, 54, 15385.
5 (a) A. C. Bach, II, J. A. Markwalder and W. C. Ripka, Int. J. Pept.
Prot. Res., 1991, 38, 314; (b) U. Nagai and K. Sato, Tetrahedron,
1993, 49, 3577; (c) K. Sato and U. Nagai, J. Chem. Soc., Perkin
Trans. 1, 1986, 1231; (d) U. Nagai and K. Sato, Tetrahedron Lett.,
1985, 26, 47.
6 (a) N. L. Subasinghe, R. J. Bontems, E. McIntee, R. K. Mishra and
R. L. Johnson, J. Med. Chem., 1993, 36, 2356; (b) J. E. Baldwin,
R. T. Freeman, C. Lowe and C. J. Schofield, Tetrahedron, 1989, 45,
4537; (c) F. A. Etzkorn, M. A. Lipton, S. D. Goldberg and P. A.
Bartlett, J. Am. Chem. Soc., 1994, 116, 10412; (d) W. A. Slusarchyk,
J. A. Robl, P. C. Taunk, M. M. Asaad, J. E. Bird, J. DiMarco and
Y. Pan, Bioorg. Med. Chem. Lett., 1995, 5, 753; (e) J. E. Baldwin
and E. Lee, Tetrahedron, 1986, 42, 6551.
7 (a) S. Hanessian, R. Margarita, A. Hall and X. Luo, Tetrahedron
Lett., 1998, 39, 5883; (b) J. M. Humphrey, R. J. Bridges, J. A. Hart
and A. R. Chamberlin, J. Org. Chem., 1994, 59, 2467; (c) J. M.
Dener, L. H. Zhang and H. Rapoport, J. Org. Chem., 1993, 58, 1159;
(d) J. E. Baldwin, M. G. Moloney and M. North, Tetrahedron, 1989,
45, 6309; (e) P. Gmeiner, P. L. Feldman, M. Y. Chu-Moyer and
H. Rapoport, J. Org. Chem., 1990, 55, 3068; ( f ) J.-P. Wolf and
H. Rapoport, J. Org. Chem., 1989, 54, 3164.
19 E. Haslam, Tetrahedron, 1980, 36, 2409.
20 P. J. Dunn, R. Häner and H. Rapoport, J. Org. Chem., 1990, 55,
5017.
21 J. M. Scholtz and P. A. Bartlett, Synthesis, 1989, 542.
22 (a) T. Yamada, N. Isono, A. Inui, T. Miyazawa, S. Kuwata and
H. Watanabe, Bull. Chem. Soc. Jpn., 1978, 51, 1897; (b) D. Keglevic,
A. Kornhauser and S. Valentekovic, Carbohydr. Res., 1972, 22, 245.
23 K. D. Berlin, L. H. Gower, J. W. White, D. E. Gibbs and G. P. Sturm,
J. Org. Chem., 1962, 27, 3595.
24 J. E. Baldwin, M. G. Moloney and M. North, J. Chem. Soc., Perkin
Trans. 1, 1989, 833.
25 J. E. Baldwin, M. G. Moloney and M. North, Tetrahedron, 1989, 45,
6319.
26 R. E. Ireland and J. P. Daub, J. Org. Chem., 1981, 46, 479.
27 (a) D. Seebach and D. Wasmuth, Angew. Chem., Int. Ed. Engl., 1981,
20, 971; (b) D. Seebach and D. Wasmuth, Helv. Chim. Acta, 1980, 63,
197; (c) G. Fráter, Helv. Chim. Acta 1979, 62, 2825.
28 D. Seebach, Angew. Chem., Int. Ed. Engl., 1988, 27, 1624.
29 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923.
8 (a) I. B. Parr, S. K. Boehlein, A. B. Dribben, S. M. Schuster and
N. G. J. Richards, J. Med. Chem., 1996, 39, 2367; (b) I. B. Parr, S. K.
Boehlein, A. B. Dribben, S. M. Schuster and N. G. J. Richards,
J. Med. Chem., 1996, 39, 4348 (correction); (c) N. G. J. Richards and
S. M. Schuster, Adv. Enzymol. Relat. Areas Mol. Biol., 1998, 72, 145.
9 (a) R. Sharma and W. D. Lubell, J. Org. Chem., 1996, 61, 202; (b)
Paper 9/00796B
1038
J. Chem. Soc., Perkin Trans. 1, 1999, 1029–1038