Synthesis and β-lactamase reactivity of α-substituted phenaceturates

Article abstract of DOI:10.1016/j.bmc.2006.06.023

β-Lactams with 6α (penicillins) or 7α (cephalosporins) substituents are often β-lactamase inhibitors. This paper assesses the effect of such substituents on acyclic β-lactamase substrates. Thus, a series of m-carboxyphenyl phenaceturates, substituted at the glycyl α-carbon by -OMe, -CH₂OH, -CO₂^-, and -CH₂NH₃⁺, have been prepared, and tested for their reactivity against serine β-lactamases. The latter two are novel substituents in β-lactamase substrates. The methoxy and hydroxymethyl compounds were found to be poor to moderately good substrates, depending on the enzyme. The aminomethyl compound gave rise to a transiently stable (t_1/2 = 4.6 s) complex on its reaction with a class C β-lactamase. The reactivity of the compounds against three low molecular weight dd-peptidases was also tested. Again, the methoxy and hydroxymethyl compounds proved to be quite good substrates with no sign of inhibitory complexes. The dd-peptidases reacted with one enantiomer (the compounds were prepared as racemates), presumably the d compound. The class C β-lactamase reacted with both d and l enantiomers although it preferred the latter. The structural bases of these stereo-preferences were explored by reference to the crystal structure of the enzyme by molecular modeling studies. The aminomethyl compound was unreactive with the dd-peptidases, whereas the carboxy compound did not react with any of the above-mentioned enzymes. The inhibitory effects of the -OMe and -CH₂OH substituents in β-lactams apparently require a combination of the substituent and the pendant leaving group of the β-lactam at the acyl-enzyme stage.

Full text of DOI:10.1016/j.bmc.2006.06.023

Bioorganic & Medicinal Chemistry 14 (2006) 7023–7033
Synthesis and b-lactamase reactivity of a-substituted
phenaceturates
S. A. Adediran,a D. Cabaret,b R. R. Flavell,a J. A. Sammons,a
M. Wakselmanb,* and R. F. Pratta,*
aDepartment of Chemistry, Wesleyan University, Middletown, CT 06459, USA
ˆ
SIRCOB, UMR CNRS 8086, Universite´ de Versailles, Batiment Lavoisier, 45 Avenue des Etats Unis, F-78000, Versailles, France
b
Received 24 April 2006; revised 1 June 2006; accepted 8 June 2006
Available online 27 June 2006
Abstract—b-Lactams with 6a (penicillins) or 7a (cephalosporins) substituents are often b-lactamase inhibitors. This paper
assesses the effect of such substituents on acyclic b-lactamase substrates. Thus, a series of m-carboxyphenyl phenaceturates,
substituted at the glycyl a-carbon by –OMe, –CH2OH, –CO2ꢀ, and –CH2NH3+, have been prepared, and tested for their reac-
tivity against serine b-lactamases. The latter two are novel substituents in b-lactamase substrates. The methoxy and hydroxy-
methyl compounds were found to be poor to moderately good substrates, depending on the enzyme. The aminomethyl
compound gave rise to a transiently stable (t1/2 = 4.6 s) complex on its reaction with a class C b-lactamase. The reactivity
of the compounds against three low molecular weight DD-peptidases was also tested. Again, the methoxy and hydroxymethyl
compounds proved to be quite good substrates with no sign of inhibitory complexes. The DD-peptidases reacted with one enan-
tiomer (the compounds were prepared as racemates), presumably the D compound. The class C b-lactamase reacted with both D
and L enantiomers although it preferred the latter. The structural bases of these stereo-preferences were explored by reference
to the crystal structure of the enzyme by molecular modeling studies. The aminomethyl compound was unreactive with the DD-
peptidases, whereas the carboxy compound did not react with any of the above-mentioned enzymes. The inhibitory effects of
the –OMe and –CH2OH substituents in b-lactams apparently require a combination of the substituent and the pendant leaving
group of the b-lactam at the acyl-enzyme stage.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
With few exceptions, the potent b-lactamase inhibitors
known to date interact covalently with b-lactamases,
either as mechanism-based inhibitors or as transition
state analogues.2 Examples of the former include the
commercially available b-lactamase inhibitors, clavu-
lanic acid, sulbactam, and tazobactam, which as b-lac-
tams themselves react initially with the enzymes as
normal substrates to form an acyl-enzyme intermedi-
ate. Diversion of this initially formed acyl-enzyme into
chemically and/or conformationally inert forms then
leads to enzyme inhibition. The best known transition
state analogue-generating molecules are the phospho-
nates and boronates.2 Since the chemical transition
states of the normal reaction are covalently bound
to the enzyme, so too are good transition state
analogues.
The major source of natural resistance to b-lactam
antibiotics arises from bacterial production of b-lacta-
mases, enzymes that hydrolytically destroy the antibi-
otic molecules.1 b-Lactams that are not thus destroyed
have a good chance of inhibiting D-alanyl-D-alanine
(DD-) peptidases, which leads to cessation of bacterial
cell wall biosynthesis and ultimately to bacterial cell
death. The resistance provided by b-lactamases may
be overcome by b-lactams that are poor b-lactamase
substrates or by application of b-lactamase inhibitors.
The latter molecules may also be antibiotics in their
own right or be supplied as an appropriate mixture
with an effective b-lactam.
Keywords: b-Lactamase; DD-peptidase; Bacterial enzymes; b-Lactam
antibiotics; Steady state enzyme kinetics; Depsipeptide.
A variety of b-lactams, other than those mentioned
above, are also known to form slowly hydrolyzing
acyl-enzymes with particular b-lactamases. These are
*
Corresponding authors. Tel.: +1 860 685 2629; fax: +1 860 685 2211
(R.F.P.); e-mail: rpratt@wesleyan.edu
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.06.023

7024
S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
often referred to as inhibitory substrates since they inter-
act with the enzyme as substrates but turn over slowly.
One particular group of these has methoxy or hydroxy-
methyl substituents a to the b-lactam ring at the C-6
(penicillin) or C-7 (cephalosporin) position (1a and 1b,
respectively).
X
RCONH
O
O
-
2
CO
4
R = P hCH , X = OMe
2
5
R = Ph CH , X = CH OH
2 2
X
X
S
+
S
RCONH
O
RCONH
O
6
R = P hCH , X = CH NH
2
2
3
-
7
R = P h, X = CO
N
N
2
Y
1a
1b
-
-
CO
CO
2
2
Although many of these compounds were substrates of
representative b-lactamases and DD-peptidases, no
strong inhibitory complexes were detected. The structur-
al basis of this result is discussed.
The first characterized of these were natural products,
the cephamycins (1b, X = OMe).3 A semisynthetic vari-
ant is the well-known cefoxitin 24 which is still in clinical
use as an injectable b-lactam (indicated, for example, for
treatment of lower respiratory tract, urinary tract, and
intra-abdominal infections). It has been shown to form
transiently stable complexes with class A5,6 and class
C7 b-lactamases. Temocillin, a 6-a-methoxypenicillin
(1a, X = OMe), also forms inert complexes with b-lacta-
mases.5,8 Similarly, a-hydroxyethyl b-lactams are known
to be more stable toward b-lactamases.9,10
2. Results and discussion
The synthetic routes employed to obtain 4–7 are pre-
sented in outline in Schemes 1–4. The synthesis of 4
(Scheme 1) involved condensation of phenylacetamide
with glyoxylic acid to form the a-hydroxyglycine deriv-
ative 13 which was converted to the a-methoxy ana-
logue 14, both steps by means of previously reported
procedures.14,15 Alkaline hydrolysis of 1416 yielded
the corresponding acid 15 which was condensed with
benzyl 3-hydroxybenzoate17 to form 16, the benzyl es-
ter of 4. Catalytic hydrogenation12 of 16 then afforded
the required product 4. The synthesis of 5 (Scheme 2)
required condensation of O-benzyl-N-(phenylacetyl)ser-
ine with benzyl 3-hydroxybenzoate, followed by cata-
lytic hydrogenation. Preparation of 6 involved a
similar sequence (Scheme 3). Only the final hydrogena-
tion step of and subsequent isolation of 6 proved diffi-
cult. This compound was found to decompose readily
during manipulation and could only be isolated in pure
form as a zwitterion after hydrogenation under weakly
acidic conditions.
OMe
S
CH CONH
2
S
2
N
OCONH
2
O
-
CO
2
For several years, we have used acylic depsipeptides
such as as convenient substrates of both
3
b-lactamases and DD-peptidases.11,12 These com-
pounds combine the amido side chain of good b-lac-
tamase and DD-peptidase substrates, a hydrolytically
susceptible/torsionally flexible ester linkage,12 and a
carboxylate in the leaving group well placed to inter-
act with polar residues of the active sites of these
enzymes.13
The preparation of 7 was a considerable challenge. The
direct condensation of a malonic acid monoester with
benzyl 3-hydroxybenzoate by means of DCC, DCC/
DMAP or DCC/p-TSA/pyridine18,19 failed. This was
probably due to the poor reactivity of the phenol with
the intermediate 5-(4H)-oxazolone, which exists pre-
dominantly in the enol form.20 We therefore turned
our attention to another route to the target molecule
7: acylation of a hippurate dianion21 with the chloro-
formate derived from benzyl 3-hydroxybenzoate
(Scheme 4). To avoid the observed hippurate N-aryl-
oxycarbonylation, we prepared an N-alkylated deriva-
tive having an acid-labile 2,4-dimethoxybenzyl
substituent.22 Thus, reductive alkylation (NaBH3CN)
of benzyl glycinate with 2,4-dimethoxybenzaldehyde
gave the known23 N-2,4-dimethoxybenzyl glycinate
22. Acylation of 22 then furnished the required substi-
tuted hippurate 23. Formation of the anion of 23
(LiHMDS) and its reaction with the chloroformate
obtained by treatment of benzyl 3-hydroxybenzoate
with triphosgene24 led to the malonate 24. Selective
removal of the N-2,4-dimethoxybenzyl-protecting group
PhCH2CONH
O
3
O
-
CO2
In view of the above, we were interested to find whether
the a-substituents of 1 would have inhibitory effects on
the turnover of 3 by a variety of serine b-lactamases of
class A and C, and by some representative low molecu-
lar weight DD-peptidases. Consequently, we have pre-
pared 4–7 and studied the kinetics of their reaction
with these enzymes. b-Lactams 1a and 1b, analogous
to 6 and 7, have not been prepared, to our knowledge.
þ
ꢀ
We hoped, however, that the –CH2NH3 and –CO2
substituents in these molecules would interact with the
polar residues of the b-lactamase active site or with
hydrolytic water molecules and thereby disrupt
catalysis.

S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
OH
CO2H
7025
H+/MeOH
O
O
OMe
CO2Me
O
CHOCO2H
N
H
N
H
NH2
13
14
OMe
O
OMe
CO2H
OH-
H2/Pd/C
MeOH
RCONH
ArOH
O
4
N
DCC, DMAP
O
15
H
16
CO2Bn
R = PhCH2-, Ar = m-CO2Bn
Scheme 1.
CH2OBn
PhCH2CONH
OH
CH2OBn
O
ArOH
PhCH2CONH
H2/Pd black
MeOH
DCC, DMAP
5
18
17
O
O
CO2Bn
Scheme 2.
+
NH3
BnOCOCl
Cu(OAc)2
NHCO2Bn
NHCO2Bn
RCOCl
OH-
H3N+
H3N+
RCONH
-
-
-
20
19
CO2
CO2
CO2
NHCO2Bn
O
ArOH
CDI
H2/Pd/C
EtOAc/HOAc
PhCH2CONH
6
21
O
CO2Bn
R = PhCH2-, Ar = m-CO2Bn
Scheme 3.
Ar
O
Ar
H2N
ArCHO
OBn
PhCOCl
22
PhCON
HN
NaBH3CN
Et3N
OBn
O
OBn
23
O
Ar
CO2Bn
OAr'
CO2Bn
O
LiHMDS
Ar'OCOCl
H2/Pd/C
EtOAc
TFA
PhCON
PhCONH
7
24
25
O
O
CO2Bn
Ar = 2,4-(MeO)2C6H4, Ar' = m-CO2Bn
Scheme 4.
with trifluoroacetic acid afforded the dibenzyl ester 25.
Simultaneous hydrogenolysis of both remaining benzyl
ester protecting groups gave the target molecule 7. This
malonic half ester, however, was not very stable and
readily decarboxylated at room temperature to give
the unsubstituted hippurate. The malonate anion, how-
ever, was more stable, thus enabling the kinetics studies
described below.
a small contribution from intramolecular general base
catalysis.
The hydrolyses of 4 and 5 were obviously enhanced
and thus catalyzed by the class C P99 b-lactamase.
The total progress curve for reaction of 4 was biphasic
(Fig. 1), where both phases were faster than the back-
ground rate. The syntheses of both 4 and 5 yielded
racemic products. Thus, it was evident that the
enzyme catalyzed the hydrolysis of both the D and
L-enantiomers of 4. The rates of the two phases were
sufficiently different that initial rates of each phase
could be obtained. Thus, steady state rate parameters
for each phase were obtained (Table 1). The less reac-
tive enantiomer was assigned the D-configuration since
this was the form that reacted with DD-peptidases (see
below). Compound 5 was a poorer substrate than 4
The background hydrolysis rates of 4–7 were deter-
mined to establish the baseline against which to measure
the enzyme-catalyzed reaction rates. In 100 mM MOPS,
pH 7.5, 25 ꢁC, the pseudo-first order rate constants for
hydrolysis of 4–7 were 2.5 · 10ꢀ4
,
4.2 · 10ꢀ5
,
6.5 · 10ꢀ4
appear to reflect the expected electronic effects of the
,
and 1.7 · 10ꢀ4 sꢀ1
, respectively. These
substituents, although the value for 7 may also include

7026
S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
E + P2OH
k3
H O
P1
2
k 4
ES
E - S
E + S
E + P2OMe
MeOH
k2
Ks
k5
RNH
2
E + P2NHR
Scheme 5.
k4/k3 of L
-4 and D-4, determined from kinetics experi-
ments as described in Section 4, were comparable to that
of 3 (Table 2), indicating that methanol had access to the
acyl-enzyme comparable to that of water. The larger
nucleophile D-Phe had much less access to the acyl-en-
zyme, with respect to water, for both L-4 and D-4 than
for 3 (see k5/k3 values, Table 2). This presumably reflects
the greater steric hindrance to nucleophile attack, and/or
conformational rigidity leading to the same, in the form-
er acyl-enzymes (see below for discussion of the structur-
al aspects).
Figure 1. The hydrolysis of L-4 (faster phase) and D-4 (slower phase),
each 0.1 mM, monitored spectrophotometrically at 290 nm and
catalyzed by the P99 b-lactamase (0.2 lM). The arrow indicates the
break between phases. In the absence of enzyme, 4 would have a half-
life of 2800 s under these conditions.
Both enantiomers of the hydroxymethyl analogue 5
were considerably poorer substrates of the P99 enzyme
than 3 (and 4). This may reflect electronic effects to some
degree—hydroxymethyl less electron-withdrawing than
methoxy—but must also reflect steric issues since the
D-alanine analogue of 4 (X = Me) is more reactive
than 5.26
but here also two phases, corresponding to the two
enantiomers, could be distinguished (Table 1).
Comparison of the steady state parameters of 4 with
those of the unsubstituted analogue 3 shows that the
L-enantiomer of the former is comparable to 3 as a
substrate of the P99 b-lactamase. Enzyme deacyla-
tion is known to be rate-determining under substrate
saturation conditions for 3,12,25 that is, kcat = k3
(Scheme 5).
With neither 4 nor 5, however, is there evidence of a sig-
nificantly stable/inert acyl-enzyme. The –OMe and
–CH2OH substituents do not have the effect here that
they do with b-lactams.
The compounds 4 and 5 are very poor substrates of the
class A TEM and PC1 enzymes (Table 1). Incubation of
either 4 or 5 with these enzymes, followed by dilution
into a solution of a good substrate under saturating con-
ditions, gave no indication of an inert acyl-enzyme in
any particular case. Thus, even in the case of the PC1
enzyme, which is more prone than the TEM enzyme
to generate relatively stable acyl-enzymes, including
This was also found to be true for both L-4 and D-4 by
application of the alternative nucleophiles methanol and
D-phenylalanine at high substrate (>Km) conditions.25,26
Both of these nucleophiles enhanced the observed rate
of reaction under these conditions, leading to the con-
clusion that kcat = k3 for 4 also. The partition ratios
Table 1. Steady state kinetics parameters for the hydrolysis of depsipeptides catalyzed by b-lactamases
Substrate
Enzymea
P99
TEM
PC1
kcat (sꢀ1
)
Km (mM)
kcat/Km (sꢀ1 Mꢀ1
)
kcat/Km (sꢀ1 Mꢀ1
)
kcat/Km (sꢀ1 Mꢀ1
)
4b
5b
(L) 310 20
(D) 13.7 2.5
(L) P 1.0
(D) P0.5
0.15 0.01
0.81 0.12
1.3 0.4
P0.5
P0.5
0.05 0.01
nd
0.23
3.8 · 105
1.1 · 104
2.0 · 103
1.0 · 103
3.0 · 103
35
(Km P 0.4 mM)
47
104
(Km P 0.4 mM)
66
(Km P 0.5)
(Km P 0.5)
n
6
nc
7
3e
n
n
125
5.4 · 105
1.2 · 104
(Km = 2.2 mM)
P420
(Km 6 0.1 mM)
a Enzymes: P99, the class C b-lactamase of Enterobacter cloacae P99; TEM, the class A b-lactamase of the TEM-2 plasmid; PC1, the class A b-
lactamase of Staphylococcus aureus PC1.
b The separate reactivities of D and L-enantiomers were distinguished as described in the text. The activity of the TEM and PC1 enzymes against 4
and 5 was too low to allow the distinction to be made.
c n, no detectable enzyme catalysis.
d Competitive inhibition observed, Ki = 0.16 0.03 mM.
e Taken from Ref. 39.

S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
7027
Table 2. Rate Constants and partition ratios for depsipeptide methanolysis and aminolysis (D-Phe) catalyzed by the P99 b-lactamase (Scheme 1)
0 a
0 a
k5=k3
Substrate
k3 (sꢀ1
)
k4 (sꢀ1 Mꢀ1
)
k4=k3
k5 (sꢀ1 Mꢀ1
)
4 (D)
4 (L)
3b
310
13.7
125
200 17
3.0 0.2
59
36
12
3
1
(7.4 0.6) · 104
(3.3 0.4) · 103
1.8 · 105
240 18
240 26
8.0 · 104
26
a k3 ¼ k3=½H2Oꢁ ¼ k3=55:5.
0
b Taken from Ref. 26.
with 3,12 no evidence of an accumulating transient inter-
mediate was obtained.
The results with 6 and 7 were rather unexpected. It had
been anticipated that either the positive charge on 6 or
the negative charge on 7 would interact with the active
site of either a class A or a class C enzyme to generate
an inert acyl-enzyme. Only in the case of 6 with the
P99 enzyme was this hope realized. Deacylation was slow
in this case, with an acyl-enzyme half-life of 4.6 s, but the
rate
of
acylation
was
modest—kcat/Km = k2/
Ks = 3.0 · 103 sꢀ1 Mꢀ1. It is possible, however, that the
combination of an a-methylammonium substituent
and, say, a third-generation cephalosporin side chain
might generate an inhibitory substrate of class C b-lacta-
mases. In most of the present experiments, however, 6
and 7 produced no evidence of covalent interaction; re-
turn of activity experiments, described above, gave no
indication of transiently stable or inert acyl-enzymes
formed between 6 and the class A enzymes or between
7 and any enzyme. It was surprising that neither enantio-
mer of 7 showed any reactivity at all. Non-covalent inter-
action, however, probably did take place. For example, 7
was a fast reversible inhibitor of the P99 enzyme with a Ki
value, assuming competitive inhibition, of 0.16 mM.
Figure 2. The hydrolysis of 4 (1.0 mM) monitored spectrophotomet-
rically at 300 nm and catalyzed by the R61 DD-peptidase (0.21 lM).
Only one enantiomer, presumably D, is hydrolyzed by the enzyme. At
the break point, the P99 b-lactamase (0.05 lM), which catalyzes
hydrolysis of the L enantiomer, was added.
Acyclic depsipeptides have also proven to be substrates
of bacterial DD-peptidases, the target of b-lactam antibi-
otics.26,27 Consequently, the reactivity of 4–7 against
some typical low molecular mass DD-peptidases28 was
examined. The charged species 6 and 7 appeared to have
no covalent reaction with these enzymes, they were nei-
ther substrates nor inhibitors. The compounds with the
neutral substituents, 4 and 5, were more reactive. These
compounds reacted with the Streptomyces R61 DD-pep-
tidase, for example, in what appeared to be a two-
phased fashion. The slower phase, however, comprising
half of the total material, reflected the rate of the back-
ground hydrolysis. The rate of the faster phase was en-
zyme concentration dependent. This showed that one
enantiomer each of 4 and 5 was a DD-peptidase sub-
strate, while the other was not. In view of the well-
known stereochemical preference of these enzymes for
the D-configuration at the penultimate amino acid resi-
due of a substrate,29 it seemed likely that the reactive
enantiomers of 4 and 5 had the D-configuration. The
residual compound after the DD-peptidase reaction, pre-
sumably the L-enantiomer, could then be hydrolyzed
rapidly by addition of a b-lactamase (Fig. 2), thereby
demonstrating the rate of reaction of the L-enantiomer
with that enzyme (see above).
(Table 3), and in fact, in the case of the R61 DD-pepti-
dase, were more effective than 3. Previous research has
shown, it might be noted, that, as might be expected
for a D-alanyl-D-alanine peptidase, the D-methyl ana-
logue of 4 (X = Me) is also a better substrate of the
R61 DD-peptidase than 3.26 The polarity, greater size,
and hydrogen bonding potential of the a-substituents
in 5–7, therefore, did not lead to inhibitory behavior
with the DD-peptidases.
Finally, it might be noted that the a-disubstituted
species 8 and 9 (synthesis: D. Cabaret and M.
Wakselman, unpublished) were not substrates of any
of the above-mentioned enzymes except for the P99
b-lactamase where very slow turnover was observed (8:
kcat = 0.02 sꢀ1
, ;
Km = 0.83 mM, kcat/Km = 25 sꢀ1 Mꢀ1
9:
kcat = 0.009 sꢀ1
,
Km = 0.43 mM,
kcat/Km =
20 sꢀ1 Mꢀ1).
PhCH CONH
PhCH2CONH
2
O
O
8
9
O
O
Depsipeptides 4 and 5 (D-enantiomers) were quite effec-
tive substrates of the representative DD-peptidases
-
-
CO
CO
2
2

7028
S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
Table 3. Steady state kinetics parameters for the hydrolysis of depsipeptides catalyzed by DD-peptidasesa
Enzymea
R61
Substrateb
4
5
3c
kcat (sꢀ1
Km (mM)
)
44
0.45 0.03
9.8 · 104
114
0.20 0.03
5.7 · 105
106
0.20 0.03
5.3 · 105
9
2.2 0.3
0.35 0.08
6.3 · 103
1.51
0.76
kcat/Km (sꢀ1 Mꢀ1
)
)
)
2.0 · 103
R39
kcat (sꢀ1
Km (mM)
)
7
0.65 0.03
0.85 0.06
7.7 · 102
NDd
kcat/Km (sꢀ1 Mꢀ1
PBP3
kcat (sꢀ1
Km (mM)
)
6
22
1
ND
1.8 0.1
1.2 · 104
kcat/Km (sꢀ1 Mꢀ1
a Enzymes: R61, the soluble DD-peptidase of Streptomyces R61; R39, the soluble DD-peptidase of Actinomadura R39; PBP3, penicillin binding
protein 3 of Neisseria gonorrhoeae.
b Only one enantiomer, assumed to be D, was a substrate—see text.
c Taken from Ref. 39.
d ND, not determined.
2.1. Structural considerations
the previous studies, models were made of
D versus
L methyl substituents in the tetrahedral intermediates
D-12 and L-12, which would be derived from enzymatic
attack on the peptides D-11 and L-11, respectively.
Previous studies have shown that D-alanyl depsipeptides
such as 10 and analogous D-alanyl peptides such as 11
are poorer substrates of the P99 b-lactamase than the
analogous glycyl derivatives; the reverse preference is
shown by the R61 DD-peptidase.26,30 It was suggested
that these observations reflect the natural substrate
preference of the latter enzyme, the evolution of a class
C b-lactamase from an enzyme resembling the R61
DD-peptidase, and the subsequent role of the former
as a b-lactamase. Structural models demonstrated the
presence of a D-methyl binding pocket at the active site
of the DD-peptidase and its absence in the b-lacta-
mase;26,30 the crystal structure of a complex of the
R61 DD-peptidase with a specific D-alanyl-D-alanine
substrate subsequently confirmed the existence of this
site.31 Models suggested that the phenylacetylglycyl/
alanyl moiety of these substrates may bind to the active
sites of these enzymes in two general conformations, A
or B. In conformation A, characterized, in particular,
by a dihedral angle a (Ser Cb–Ser Oc–Ctet–Ca) of ca.
ꢀ70ꢁ (see 12), the a-carbon is directed into the protein,
while in B, characterized by the dihedral angle a of ca.
ꢀ170ꢁ, it is directed outwards. D-alanyl derivatives bind
to the R61 DD-peptidase in conformation A. This
conformation is also favored by glycyl substrates of
the P99 b-lactamase but is precluded for D-alanyl
substrates which thus must bind to the b-lactamase in
conformation B.
PhCH CONH
D or L
2
N
H
OSer
D
-
O
-
a
CO
12
2
A combination of molecular dynamics and mechanics
computations, performed as described in Section 4, led
to the following results. The structure L-12 is not
dynamically stable in the A conformation at the R61
DD-peptidase active site because of unfavorable steric
interaction between the L-methyl group and the side
chain of Asn 161 and Tyr 159. In conformation B (an
energy-minimized structure is shown in Fig. 3), the steric
problems are lessened but the ligand interacts with the
active site much more weakly than does D-12 in the A
conformation26 (the Eint value32 for the latter is
28.5 kcal/mole more negative). Further, as demonstrat-
ed by 200 ns molecular dynamics runs, L-12 in the B
conformation is much more mobile than D-12 in the A
conformation. As noted previously33, the mobility of a
substrate at the R61 DD-peptidase active site appears
to inversely correlate with its reactivity. These models,
therefore, do appear in accord with the observed sub-
strate preferences of the R61 DD-peptidase. One would
expect that similar considerations would lead to the
same preferences in other DD-peptidases, such as the
R39 enzyme and pbp3 (Table 3).
PhCH2CONH
D
PhCH2CONH
D
O
N
H
D
CO2-
10
11
O
O
Eint values of energy-minimized complexes of D-12
and L-12 with the P99 b-lactamase were also comput-
ed. L-12 in conformation B (Fig. 4) was more stable
than that in conformation A by 21.2 kcal/mole and
CO2-
Similar modeling methods have now been employed to
understand the structural basis of the results described
above for the preferences displayed by the P99 b-lacta-
mase andthe R61 DD-peptidase for D versus L substituents
at the penultimate residue of a substrate. In order to re-
duce the conformational complexity, and to complement
more stable than D-12 in conformation
B
by
23.0 kcal/mole. These results are in agreement with
the experimental observations that L-4 and L-5 are
better substrates of the P99 b-lactamase than the
corresponding D-enantiomers. Frere et al.34,35 have

S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
7029
Figure 3. Stereoview of an energy-minimized tetrahedral intermediate structure 11 formed on reaction of the peptide L-10 with the P99 b-lactamase.
The ligand is shown in the (reactive) B conformation (see text); only heavy atoms are displayed.
Figure 4. Stereoview of an energy-minimized tetrahedral intermediate structure 11 formed on reaction of the peptide L-10 with the R61 DD-peptidase.
The ligand is shown in the B conformation (see text); only heavy atoms are displayed.
previously observed that L-stereochemistry at the pen-
ultimate position may produce a substrate of class C
b-lactamase of comparable or greater reactivity than
the D-analogue. Thus, with both the P99 b-lactamase
and the R61 DD-peptidase, the L enantiomers of 4,
5, and 10 probably react preferentially in the B con-
formation when bound to the enzyme (Figs. 3 and 4).
tuent and the pendant leaving group of a b-lactam is
needed to extend the acyl-enzyme lifetime. The acyl en-
zyme intermediates derived from acyclic substrates are
presumably more conformationally mobile than those
from b-lactams and can thus access reactive conforma-
tions such as B (Fig. 4). Very similar conclusions were
previously reached with respect to third-generation ceph-
alosporin side chains where again both the side chain and
the pendant leaving group were required to achieve an
inert acyl-enzyme.36,37 The aminomethyl and carboxylate
substituents of 6 and 7, respectively, did not interact pro-
ductively with the enzymes examined except with the class
C b-lactamase where reaction with 6 generated a
transiently stable acyl-enzyme.
Inspection of the models suggested that attempts at pro-
ductive binding by the a-disubstituted analogues 8 and 9
to these active sites would likely lead to disruption of the
catalytic AsnLysTyrSer combination by the additional
methyl group.
3. Conclusions
4. Experimental
The methoxy and hydroxymethyl compounds 4 and 5,
respectively, generally were substrates of representative
class A and C b-lactamases and of the low MW DD-pepti-
dases tested. No sign of transiently stable acyl-enzyme
intermediates was seen, in contrast with the b-lactams
1a and 1b. It seems likely that a combination of the substi-
4.1. Enzyme kinetics
The Enterobacter cloacae P99, TEM-2, and Staphylococcus
aureus PC1 b-lactamases were purchased from the
Centre for Applied Microbiology and Research, Porton

7030
S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
Down, Wilts., UK, and used as received. Samples of the
Streptomyces R61 and R39 DD-peptidases and penicillin
binding protein 3 of Neisseria gonorrhoeae were gener-
161, Arg 285, His 298, Thr 299, Gly 300, and Thr
301 of the R61 DD-peptidase.
`
ous gifts of Dr. Jean-Marie Frere, Dr. Paulette Charlier,
4.3. Syntheses
`
`
both of the University of Liege, Liege, Belgium, and Dr.
Robert A. Nicholas of the University of North Carolina,
Chapel Hill, NC, respectively.
4.3.1. 3-Carboxyphenyl N-(phenylacetyl)-a-methoxygly-
cinate (4) (Scheme 1). N-(Phenylacetyl)-a-hydroxygly-
cine (13) was prepared by the method of Zoller and
Ben-Ishai.14 Thus, phenylacetamide (26.1 g, 0.2 mol)
and glyoxylic acid (23.0 g, 0.22 mol) were heated under
reflux in acetone (150 ml) for 17 h. The solvent was then
removed by evaporation, yielding a residual oil that
crystallized on standing overnight. The colorless prod-
uct was recrystallized from chloroform/dioxane (1:1) in
Solution conditions for the kinetics runs were as follows:
0.1 M MOPS buffer, pH 7.5, at 25 ꢁC (P99, PC1, and
TEM b-lactamases), 0.1 M MOPS buffer, pH 7.5, at
37 ꢁC (R61 DD-peptidase), 0.1 M pyrophosphate buffer,
pH 8.5, at 37 ꢁC (penicillin binding protein 3), and
0.1 M Tris buffer, pH 7.7, at 37 ꢁC (R39 DD-peptidase).
Stock solutions of the depsipeptides were prepared in
DMSO. The DMSO concentration in reaction mixtures
after addition of substrate was 5% v/v. Reactions of 4–9
were monitored spectrophotometrically at 290 nm
(De = 1760 Mꢀ1 cmꢀ1) or 300 nm (De = 950 Mꢀ1 cmꢀ1).
Steady state kinetics parameters for the enzyme-cata-
lyzed reactions were obtained from initial rates by
non-linear least squares fitting of the data to the Hen-
ri–Michaelis–Menten equation. In some cases, where
the reaction progress curve was obviously biphasic, ini-
tial rates as a function of substrate concentration could
be determined separately for the fast and slow phases.
Enzyme-catalyzed methanolysis experiments were per-
formed as described previously;26 methanol concentra-
tions were varied from zero to 2.5 M. Substrate
aminolysis kinetics were also obtained as described in
a previous paper;30 D-phenylalanine concentrations
ranging up to 40 mM were employed.
1
60% yield. Mp 80–81 ꢁC. H NMR (DMSO-d6) d 3.49
(2H, s), 5.41 (1H, d, J = 8.7 Hz), 7.23–7.35 (5H, m),
8.89 (1H, d, J = 8.4 Hz).
The a-hydroxyacid (19.3 g, 92 mmol) was dissolved in
methanol (225 ml), and the solution was ice cooled.15
Concd sulfuric acid (3 ml) was added to the chilled mix-
ture, and the resulting solution was stirred for 2 days at
room temperature. The final cloudy mixture was poured
into ice/saturated sodium bicarbonate and extracted with
ethyl acetate (3· 125 ml). The combined ethyl acetate ex-
tracts were then dried over magnesium sulfate, filtered,
and evaporated. The resulting colorless solid, the methyl
ester 14, was then recrystallized from methanol/water
1
(1:1) in 80% yield. Mp 68–71 ꢁC. H NMR (CDCl3) d
3.43 (3H, s), 3.69 (2H, s), 3.81 (3H, s), 5.56 (1H, d,
J = 9.2 Hz), 6.41 (1H, d, J = 9.1 Hz), 7.29–7.43 (5H, m).
Alkaline hydrolysis of 14 then produced the corre-
sponding acid 15.16 Thus, the methyl ester 14 (10 g,
42 mmol) was dissolved in potassium hydroxide/meth-
anol (2.0 g/150 ml) and left at room temperature over-
night. The solvent was then evaporated and the
residue dissolved in water (100 ml). The solution was
extracted with diethyl ether (2· 50 ml) and then acid-
ified with 5% phosphoric acid (50 ml) to pH 2. The
resulting acidic solution of the product was extracted
with ethyl acetate (2· 100 ml), and the extracts were
dried over magnesium sulfate. The solvent was removed
by evaporation and the colorless residue recrystallized
from acetonitrile to give the product (15) in 53% yield.
4.2. Molecular modeling
The structures of Figures 1 and 2 were derived from
computational models of enzyme-substrate complexes
that were set up essentially as previously described26
and run on an SGI Octane
2 computer with
INSIGHT II 2005 (Accelrys, San Diego, CA). In these
models of the P99 b-lactamase, Lys 67 and Lys 315
were cationic, Tyr 150 was neutral, and the tetrahe-
dral intermediate 11 was dianionic. The corresponding
residues of the R61 DD-peptidase, Lys 62 and Tyr 159,
were also treated as cationic and neutral, respectively.
Lys 315 of the b-lactamase corresponds to His 298 of
the DD-peptidase. The latter residue was treated as
neutral in the simulations; there is no direct evidence
for this but a histidine residue is likely to be neutral
at pH 7.5, particularly when directly adjacent (closest
1
Mp 200–202 ꢁC. H NMR (DMSO-d6) d 3.19 (3H, s),
3.52 (2H, s), 5.25 (1H, d, J = 8.9 Hz), 7.20–7.31 (5H,
m), 9.41 (1H, d, J = 8.9 Hz).
The procedure employed for condensation of the acid 15
with benzyl 3-hydroxybenzoate was adapted from that
of Gallo and Gellman.17 The acid (2.0 g, 8.9 mmol),
benzyl 3-hydroxybenzoate (2.02 g, 8.9 mmol), and
4-dimethylaminopyridine (100 mg, 0.82 mmol) were dis-
solved in methylene chloride (28 ml). The stirred solu-
tion was then cooled to 0 ꢁC on ice, treated with
dicyclohexylcarbodiimide (2.28 g, 11.0 mmol), and
allowed to return to room temperature overnight. The
resulting slurry was filtered, and the filtrate washed
sequentially with water, 10% citric acid, water, saturated
sodium bicarbonate, and water. The filtrate was then
concentrated on a rotary evaporator, yielding a pale yel-
low-brown liquid that slowly solidified overnight under
˚
heavy atom contact 64 A) to an arginine side-chain
nitrogen atom (Arg 285). Initially, the ligands were
placed in conformations A and B26 and the stability
of these conformations explored by molecular dynam-
ics (200 ps runs, where the entire protein together with
solvating water molecules was unrestricted). Typical
snapshots of dominant conformations were selected
for energy minimization. The minimized structures
32
were then used to calculate interaction energies, Eint,
between ligand and protein. The residues included in
these calculations were Ser 64, Lys 67, Tyr 150, Asn
152, Lys 315, Thr 316, Gly 317, and Ser 318 of the
P99 b-lactamase and Ser 62, Lys 65, Tyr 159, Asn

S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
7031
oil pump vacuum (16, yield 94%). Mp 122 ꢁC. Rf 0.75
(CH2Cl2/MeOH 99:1). 1H NMR (DMSO-d6) d 3.32
(3H, s), 3.57 (2H, s), 5.38 (2H, s), 5.54 (1H, d,
J = 8.0 Hz), 7.29–7.83 (5H, m), 7.37–7.51 (6H, m), 7.62
(1H, t, J = 8.0 Hz), 7.74 (1H, s), 7.93 (1H, d,
J = 8.1 Hz), 9.41 (1H, d, J = 8.3 Hz). 13C NMR (CDCl3)
d 43.59, 56.08, 67.56, 79.48, 118–137, 151.25, 164.80,
167.22, 171.97. Anal. Calcd for C25H23NO6: C, 69.27;
H, 5.35; N, 3.23. Found: C, 69.25; H, 5.56; N, 3.12.
20 mmol) were dissolved in water (35.6 ml). The pH of
the solution was raised to 5 using 1 M NaOH solution.
MgO (0.712 g, 176 mmol) was added followed by benzyl
chloroformate (3.40 ml, 20 mmol). The reaction mixture
was stirred for 5 h during which time a blue precipitate
formed. The precipitate was collected by filtration on
sintered glass and washed separately with water
(50 ml), methanol, and diethyl ether. The solid was then
suspended in boiling 10% aqueous acetic acid solution
and thioacetamide (3.00 g, 20 mmol) added. The mix-
ture was heated for 10 min where upon a black solid ap-
peared. This was removed by filtration through Celite
and washed with a small amount of boiling water. Col-
orless crystals formed in the filtrate after its cooling on
ice for 3 h. The solid product (19) was collected by filtra-
tion and dried in vacuo over KOH; yield 2.25 g (49%).
1H NMR (D2O): d 3.55 (1H, dd, J = 2.5, 4.0 Hz), 3.7
(1H, dd, J = 2.2, 3.6 Hz), 4.7 (1H, t, J = 2.3 Hz), 5.1
(2H, s), 7.4 (5H, br s).
The required compound, 3-carboxyphenyl N-(pheny-
lacetyl)-a-methoxyglycinate, 4, was then obtained from
catalytic hydrogenation of 16. Compound 16 (1.5 g),
dissolved in methanol (75 ml), was hydrogenated
(35 psi) overnight over 10% Pd/C (150 mg). The result-
ing mixture was then filtered through a Celite pad,
washed with methanol, and evaporated. The resulting
colorless solid, 4, obtained in quantitative yield, was
recrystallized from benzene/ethyl acetate (3:1). Mp
1
168–171 ꢁC. H NMR (DMSO-d6) d 3.34 (3H, s), 3.59
(2H, s), 5.55 (1H, d, J = 8.3 Hz), 7.29–7.39 (6H, m),
7.59 (1H, t, J = 8.0 Hz), 7.71 (1H, s), 7.87 (1H, d,
J = 7.7 Hz), 9.42 (lH, d, J = 8.2 Hz). 13C NMR
(CD3COCD3) d 43.57, 56.06, 79.55, 124–137, 151.50,
166.27, 171.97. Anal. Calcd. for C18H17NO6: C, 62.96;
H, 4.99; N, 4.08. Found: C, 62.88; H, 4.96; N, 4.12.
Phenylacetylation of the a-amino group of 19 was
achieved as follows. Compound 19 (2.23 g, 9.4 mmol)
was dissolved in 3.5 M aqueous NaOH solution
(5.4 ml). The solution was cooled to 0 ꢁC on ice, and
phenylacetyl chloride (1.25 ml, 10 mmol) was added in
0.25 ml portions over 1 h with stirring. The mixture
was then stirred to room temperature over 2 h. The
resulting white suspension was dissolved by addition
of the minimum volume of hot water on a steam bath
and 0.12 g decolorizing charcoal added. The charcoal
was removed by filtration and the hot filtrate acidified
with concd HCl to pH 1. The precipitate formed on
cooling was recrystallized from water, yielding 20 in
77% yield. Mp 106–108 ꢁC. 1H NMR (DMSO-d6): d
3.35 (2H, m), 3.45 (2H, s), 4.30 (1H, m), 5.00 (2H, s),
7.25 (5H, br s), 7.35 (5H, br s), 8.3 (1H, d, J = 9.1 Hz).
4.3.2. 3-Carboxyphenyl N-(phenylacetyl)-serinate (5)
(Scheme 2). N-(Phenylacetyl)-D-serine(OBn) (17) was pre-
pared from the reaction of D-serine(OBn) with phenylace-
tyl chloride in aqueous base.12 It was recrystallized from
1
ethanol/water (1:1) in 36% yield. Mp 106 ꢀ108 ꢁC. H
NMR (DMSO-d6) d 3.54 (2H, s), 3.60 (1H, dd, J = 4.5,
10.0 Hz), 3.76 (1H, dd, J = 4.4, 10.0 Hz), 3.78 (1H, m),
4.48 (2H, s), 7.27–7.37 (10H, m), 8.46 (1H, d,
J = 8.1 Hz), 12.8 (1H, s). The condensation of 12 with
benzyl 3-hydroxybenzoate to obtain 18 (in 58% yield)
1
was carried out as described above for 16. H NMR
The condensation of 20 with benzyl 3-hydroxybenzoate
in the presence of carbonyldiimidazole was carried out
as previously described.12 The product, 21, was purified
by silica gel flash chromatography with ethyl acetate/
hexane (2:3) as eluent; yield 30%. Mp 74–76 ꢁC. 1H
NMR (DMSO-d6) d 3.55 (2H, s) 3.55 (2H, m), 4.55
(1H, m), 5.0 (2H, s), 5.4 (2H, s), 7.15–7.40 (15H, m),
7.45 (1H, d, J = 8.4 Hz), 7.55 (1H, t, J = 8.3 Hz), 7.7
(1H, s), 7.9 (1H, d, J = 8.2 Hz), 8.7 (1H, d, J = 8.6 Hz).
(DMSO-d6) d 3.58 (2H, s), 3.81 (1H, dd, J = 4.4,
9.9 Hz), 3.95 (1H, dd, J = 5.4, 9.9 Hz), 4.56, 4.62 (2H,
ABq, J = 15 Hz), 4.82 (1H, dt, J = 4.4, 5.4, 6.9 Hz), 5.39
(2H, s), 7.2–7.5 (16H, m), 7.61 (1H, t, J = 7.8 Hz), 7.65
(1H, s), 7.92 (1H, d, J = 7.9 Hz), 8.88 (1H, d, J = 7.3 Hz).
To complete the synthesis of 5, 18 (0.61 g) was dissolved
in methanol (10 ml) and palladium black (0.65 g) added
to the solution. The mixture was then shaken for 19 h
under 35 psi of hydrogen gas. The reaction mixture
was filtered through a Celite pad, which was washed
with methanol; solvent from the pooled filtrate was
evaporated. The resulting colorless solid was recrystal-
lized from ethyl acetate/benzene (1:1) in 64% yield. Mp
The catalytic hydrogenation of 21 to obtain 6 required
much attention to detail because of the instability of
the product. The following procedure was the most
effective of several tried. Compound 21 (100 mg,
0.18 mmol) was dissolved in ethyl acetate (15 ml) con-
taining a little acetic acid (100 ll). Palladium (10%) on
activated carbon (150 mg) was added and the reaction
mixture was shaken under 35 psi hydrogen gas for
12 h. At that stage, an additional 50 mg of palladium
catalyst was added and the mixture was hydrogenated
for an additional 6 h. The catalyst and a white precipi-
tate were then collected by filtration and washed with
TFA. Excess TFA was then removed by rotary evapora-
tion, isolating the TFA salt of the product. This was sus-
pended in chloroform (10 ml) and 2,6-lutidine (2 ml) was
added. The mixture was stirred at 20 ꢁC overnight and
1
163 ꢀ164 ꢁC. H NMR (DMSO-d6) d 3.52, 3.56 (2H,
ABq, J = 14.2 Hz), 3.77 (1H, dd, J = 4.8, 10.9 Hz),
3.93 (1H, dd, J = 4.8, 10.9 Hz), 4.52, 4.56 (1H, br),
7.2–8.0 (9H, m), 8.85 (1H, d, J = 6.5 Hz).
4.3.3. 3-Carboxyphenyl 2-(phenylacetamido)-3-aminopro-
pionate (6) (Scheme 3). The b-amino group of 2,3-diami-
nopropionic
acid
was
first
protected
by
benzyloxycarbonylation, essentially by the procedure
of Folsch and Serck-Hanssen.38 Thus, 2,3-diaminoprop-
ionic acid (2.0 g, 19.0 mmol) and cupric acetate (1.42 g,

7032
S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
then placed in the freezer for an additional night to aid
precipitation of the product. The product was collected
by centrifugation, washed with chloroform, and dried in
vacuo, yielding 22 mg (37%) of 6 as an off-white powder.
in hexane (2 ml). The mixture was stirred at ꢀ78 ꢁC for
1 h, then a solution of 191 mg (about 0.66 mmol) of the
crude 3-(benzyloxycarbonyl)phenyl chloroformate in
THF (3 ml) was added dropwise. The reaction mixture
was stirred at room temperature for a further 3 h and
then poured into a 0.1 N KHSO4 solution. After extrac-
tion with ethyl acetate, the organic layer was washed
with water, dried over MgSO4, and the solvent evapo-
rated. The product was purified by two successive chro-
matographic steps on silica gel: the first eluent was
CH2Cl2/AcOEt 95:5, the second cyclohexane/AcOEt
1
Mp 115 ꢁC (d). H NMR (D2O): d 3.5 (1H, dd, J = 8.3,
13.8 Hz), 3.65 (1H, dd, J = 5.0, 14.9 Hz), 3.80 (2H, s),
4.95 (1H, dd, J = 5.9, 8.6 Hz), 7.09 (1H, d, J = 7.8 Hz),
7.37 (5H, br s), 7.46 (1H, s), 7.49 (1H, t, J = 7.8 Hz),
7.80 (1H, d, J = 7.8 Hz). ESMS(ꢀ) m/e 343.
4.3.4. 3-Carboxyphenyl phenylacetamidomalonate (7)
(Scheme 4).
1
70:30; the yield of 24 was 142 mg (43%). H NMR d
N-Benzoyl-N-(2,4-dimethoxybenzyl)glycine benzyl ester
(23). To a solution of benzyl glycinate (1.41 g, 7 mmol),
in acetic acid (15 ml), were added sodium acetate
(574 mg, 7 mmol) and then 2,4-dimethoxybenzaldehyde
(1.16 g, 7 mmol). The reaction mixture was stirred for
30 min at room temperature before addition in several
portions of NaBH3CN (440 mg, 7 mmol). The reaction
mixture was stirred overnight and the acetic acid was
evaporated. The residue was dissolved in dichlorometh-
ane and the solution washed with 0.5 N NaOH. The
organic layer was dried over MgSO4 and then the sol-
vent was evaporated. The crude benzyl N-2,4-dimeth-
(CDCl3): 3.67 and 3.79 (2s, 2· 3H), 4.58 (dd, 2H),
4.88 (s, 1H), 5.23 and 5.37 (2s, 2· 2H), 6.38–6.46 (m,
2H), 7.23 (d, 1H), 7.33–7.56 (m, 16H), 7.74 (s, 1H),
7.94 (m, 1H). 13C NMR d (CDCl3): 49.8, 55.3–55.6,
62.3, 67.1–68.2, 98.8–136.1, 151.8, 164.3, 165.3, 165.7
and 172.6. Anal. Calcd for C40H35NO9: C, 71.31; H,
5.24; N, 2.08. Found: C, 71.10; H, 5.32; N, 2.04.
2-Benzoylaminomalonic acid benzyl ester 3-(benzyloxy-
carbonyl)phenyl ester (25). To a solution of the N-2,4-
dimethoxybenzylhippurate 24 (119 mg, 0.23 mmol) in
dichloromethane (1 ml) was slowly added 1 ml TFA at
0 ꢁC. The mixture was stirred for 1 h at room temperature
and the solvents were then removed by evaporation to
give a residue, which was purified by silica gel chromatog-
raphy (eluent: CH2Cl2/AcOEt 95:5), to give 81 mg (87%)
of the title product 25. Rf 0.71. The product was recrystal-
lized from diethyl ether. Mp 126 ꢁC. 1H NMR d (CDCl3):
5.36 (dd, 2H), 5.38 (s, 2H), 5.66 (d, 1H), 7.18 (m, 1H), 7.27
(d, 1H), 7.34–7.57 (m, 14H), 7.77 (s, 1H), 7.85–7.89 (m,
2H), 7.98 (m, 1H). 13C NMR d (CDCl3): 57.3, 67.0,
68.9, 122.3–136.0, 150.5, 165.1, 165.9, 167.3. Anal. Calcd
for C31H25NO7: C, 71.12; H, 4.81; N, 2.68. Found: C,
70.93; H, 4.85; N, 2.62.
oxybenzylglycinate
2223
was
dissolved
in
dichloromethane (20 ml), then were added successively
triethylamine (0.974 ml, 7 mmol) and, dropwise, a solu-
tion of benzoyl chloride (0.810 ml, 7 mmol) in dichloro-
methane (20 ml). Stirring was continued for 2 h, the
mixture was poured into a KHSO4 solution, the organic
layer was washed with saturated NaHCO3, dried over
MgSO4, and the solvent evaporated. The residue was
purified by silica gel column chromatography (CH2Cl2/
AcOEt 90:10), affording 2.30 g (78%) of the title ester
1
23. Rf 0.53. Mp 83 ꢁC. H NMR d (CDCl3): 2 isomers:
3.67 and 3.72 (2s, 3H), 3.80 (s, 3H), 3.94 and 4.21 (s,
2H), 4.51 and 4.78 (2s, 2H), 5.13 and 5.20 (2s, 2H),
6.3–6.5 (m, 2H), 6.97 (d, 1H), 7.27–7.52 (m, 10H). 13C
NMR d (CDCl3): 2 isomers: 43.6 and 46.3, 49.2 and
50.4, 53.6 and 55.3, 55.4 and 55.6, 67.0–67.5, 98.6–
136.1, 158.8 and 161.0, 169.4 and 169.6, 172.3 and
172.8. Anal. Calcd for C25H25NO5: C, 71.58; H, 6.01;
N, 3.34. Found: C, 71.49; H, 6.11; N, 3.27.
3-Carboxyphenyl phenylacetamidomalonate (7). The
hydrogenolysis was performed at 0 ꢁC in a 50 ml
round-bottomed flask equipped with a three-way tap
and magnetic stirring. The recrystallized dibenzylester
25 (20 mg) was dissolved in ethyl acetate (6 ml) then
10 mg of 10% Pd/C was added. A vacuum was applied
to the flask until the solvent began to bubble. Then
the flask was filled with hydrogen. After 30 min, the
reaction was complete as shown by TLC (silica gel, elu-
ent: ethyl acetate/acetic acid 9:1. Rf 7: 0.40) and (silica
gel, eluent: CH2Cl2/ethyl acetate 95:5. Rf 25: 0.51). The
solution was then filtered and the solvent removed under
vacuum at 0 ꢁC. The NMR spectra of the crude product
showed no notable impurities. The compound, however,
is not very stable toward decarboxylation at room tem-
3-(Benzyloxycarbonyl)phenyl chloroformate. A solution
of benzyl 3-hydroxybenzoate (912 mg, 4 mmol) and of
N,N-dimethylaniline (912 mg, 4 mmol) in THF (6 ml)
was slowly added with cooling to a solution of 396 mg
(1.33 mmol) of bis (trichloromethyl) carbonate in THF
(6 ml). The reaction mixture was stirred at room temper-
ature overnight then poured into 0.1 N HCl at 0 ꢁC and
extracted with ethyl acetate. The organic layer was dried
over MgSO4 and the solvent evaporated to give a mixture
of the starting benzyl 3-hydroxybenzoate and the chloro-
formate, in the ratio 27:73, which was directly used in the
1
perature. H NMR d [(CD3)2CO]: 5.68 (d, 1H), 7.35–
7.64 (m, 5H), 7.97 (m, 1H), 7.98–8.03 (m, 2H), 8.39 (d,
1H). 13C NMR d [(CD3)2CO]: 58.0, 123.5–134.3,
151.8, 166.6, 166.7, 167.6, 167.7. ESMS(+), (CH3CN–
H2O), m/e 344 [M+H]+.
1
next step. H NMR d (CDCl3): 5.39 (2s, 2H), 7.34–7.48
(m, 6H), 7.52 (t, 1H), 7.93 (s, 1H), 8.05 (dd, 1H). 13C
NMR d (CDCl3): 67.5, 116.6 to 135.7, 149.7, 151.5, 165.2.
2-Benzoyl-(2,4-dimethoxybenzyl)-aminomalonic acid ben-
zyl ester 3-(benzyloxycarbonyl)phenyl ester (24). At
ꢀ78 ꢁC, a solution of 23 (276 mg, 0.66 mmol) in THF
(3 ml) was slowly added to a 1.0 N solution of LiHMDS
Acknowledgment
This research was supported by the National Institutes
of Health.

S. A. Adediran et al. / Bioorg. Med. Chem. 14 (2006) 7023–7033
7033
20. Morgan, J.; Pinhey, J. T.; Sherry, C. J. J. Chem. Soc.,
Perkin Trans. 1 1997, 613.
References and notes
21. Hardling, K. E.; Moreno, L. N.; Nace, V. M. J. Org.
Chem. 1981, 46, 2809.
`
1. Matagne, A.; Dubus, A.; Galleni, M.; Frere, J.-M. Nat.
Prod. Chem. 1999, 16, 1.
22. Weygard, F.; Steglich, W.; Bjarnason, J.; Akhtar, R.;
Khan, N. M. Tetrahedron Lett. 1966, 29, 3483.
23. Shiozaki, M.; Ishida, N.; Hiraoka, T.; Maruyama, H.
Tetrahedron 1984, 40, 1795.
24. Cabaret, D.; Garcia-Gonzales, M.; Wakselman, M.;
Adediran, S. A.; Pratt, R. F. Eur. J. Org. Chem. 2001, 141.
25. Xu, Y.; Soto, G.; Hirsch, K.; Pratt, R. F. Biochemistry
1996, 35, 3595.
2. Pratt, R. F. In The Chemistry of b-Lactams; Page, M. I.,
Ed.; Chapman and Hall: London, 1992; pp 228–271.
3. Gordon, E. M.; Sykes, R. B.. In Chemistry and Biology of
b-Lactam Antibiotics; Morin, R. B., Gorman, M., Eds.;
Academic Press: New York, 1982; Vol. 1, pp 199–370.
4. Birnbaum, J.; Stapley, E. O.; Miller, A. K.; Wallick, D.;
Hendlin, D.; Woodruff, H. B. J. Antimicrob. Chemother.
1978, 4(Suppl. B), 15.
26. Bernstein, N. J.; Pratt, R. F. Biochemistry 1999, 38, 10499.
27. Adam, M.; Damblon, C.; Plaitin, C.; Christiaens, L.;
5. Fisher, J.; Belasco, J. G.; Khosla, S.; Knowles, J. R.
Biochemistry 1980, 19, 2895.
6. Matagne, A.; Lamotte-Brasseur, J.; Dive, A.; Knox, J. R.;
`
Frere, J.-M. Biochem. J. 1990, 270, 525.
28. Ghuysen, J.-M. Annu. Rev. Microbiol. 1991, 45, 37.
`
Frere, J.-M. Biochem. J. 1993, 293, 607.
`
29. Ghuysen, J.-M.; Frere, J.-M.; Leyh-Bouille, M.; Coyette,
J.; Dusart, J.; Nguyen- Disteche, M. Annu. Rev. Biochem.
1979, 48, 73.
30. Adediran, S. A.; Zhang, Z.; Nukaga, M.; Palzkill, T.;
Pratt, R. F. Biochemistry 2005, 44, 7543.
7. Mazzella, L. J.; Pratt, R. F. Biochem. J. 1989, 259, 255.
8. Jules, K.; Neu, H. C. Antimicrob. Agents Chemother. 1982,
22, 453.
9. Dixon, R. A.; Edmondson, R. A.; Hardy, K. D.; Milner,
P. H. J. Antibiot. 1984, 37, 1729.
`
31. McDonough, M. A.; Anderson, J. W.; Silvaggi, N. R.;
Pratt, R. F.; Knox, J. R.; Kelly, J. A. J. Mol. Biol. 2002,
322, 111.
32. Curley, K.; Pratt, R. F. J. Am. Chem. Soc. 1997, 119, 1529.
33. Nagarajan, R.; Pratt, R. F. J. Org. Chem. 2004, 69, 7472.
34. Damblon, C.; Ledent, P.; Zhao, G.-H.; Jamin, M.; Dubus,
10. Dixon, R. A.; Hardy, K. D.; Kaura, A. C.; Milner, P. H.;
Taylor, A. W. J. Antibiot. 1984, 37, 1732.
11. Pratt, R. F.; Govardhan, C. P. Proc. Natl. Acad. Sci.
U.S.A. 1984, 81, 1302.
12. Govardhan, C. P.; Pratt, R. F. Biochemistry 1987, 26,
3385.
`
A.; Vanhove, M.; Raquet, X.; Christiaens, L.; Frere, J.-M.
Lett. Pept. Sci. 1995, 2, 212.
35. Damblon, C.; Zhao, G.-H.; Jamin, M.; Ledent, P.; Dubus,
13. Ahn, Y.-M.; Pratt, R. F. Bioorg. Med. Chem. 2004, 12,
1537.
14. Zoller, U.; Ben-Ishai, D. Tetrahedron 1975, 31, 863.
15. Ben-Ishai, D.; Sataty, I.; Bernstein, Z. Tetrahedron 1976,
32, 1571.
`
A.; Vanhove, M.; Raquet, X.; Christiaens, L.; Frere, J.-M.
Biochem. J. 1995, 309, 431.
36. Powers, R. A.; Caselli, E.; Focia, P. J.; Prati, F.; Shoichet,
B. K. Biochemistry 2001, 40, 9207.
16. Rabi-Barakay, A.; Ben-Ishai, D. Tetrahedron 1994, 50,
10771.
37. Kumar, S.; Adediran, S. A.; Nukaga, M.; Pratt, R. F.
Biochemistry 2004, 43, 2664.
17. Gallo, E. A.; Gellman, S. H. J. Am. Chem. Soc. 1993, 115,
9774.
38. Folsch, G.; Serck-Hanssen, K. Acta Chem. Scand. 1959,
13, 1243.
39. Adediran, S. A.; Pratt, R. F. Biochemistry 1999, 38, 1469.
18. Holmberg, K.; Hansen, B. Acta Chem. Scand. 1979, B33,
410.
19. Cabaret, D.; Liu, J.; Wakselman, M. Synthesis 1994, 480.