Benzopyranones with retro-amide side chains as (inhibitory) β-lactamase substrates

Article abstract of DOI:10.1016/j.bmcl.2004.07.067

3-(N-Benzylcarbamoyl)-7-carboxy-3, 4-dihydro-2H-1-benzo-pyran-2-one and its 8-carboxy analogue have been synthesized and evaluated as potential (inhibitory) substrates of β-lactam-recognizing enzymes. These compounds are bicyclic δ-lactones with retro-ami

Full text of DOI:10.1016/j.bmcl.2004.07.067

Bioorganic & Medicinal Chemistry Letters 14 (2004) 5117–5120
Benzopyranones with retro-amide side chains as (inhibitory)
b-lactamase substrates
a,
S. A. Adediran,a D. Cabaret,b J.-F. Lohier,b M. Wakselmanb, and R. F. Pratt
*
*
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 31 May 2004; revised 27 July 2004; accepted 28 July 2004
Available online 25 August 2004
Abstract—3-(N-Benzylcarbamoyl)-7-carboxy-3, 4-dihydro-2H-1-benzo-pyran-2-one and its 8-carboxy analogue have been synthe-
sized and evaluated as potential (inhibitory) substrates of b-lactam-recognizing enzymes. These compounds are bicyclic d-lactones
with retro-amide (with respect to classical b-lactams) side chains. They were found to be comparably effective as substrates of typical
class A, C and D b-lactamases as analogous benzopyranones bearing ÔnormalÕ amide side chains. The new 8-carboxy derivative,
however, formed a much more (1000-fold) tightly-bound acyl-enzyme with a class C b-lactamase than did its ÔnormalÕ analogue,
and thus provides a structural lead to new inhibitors of this class of b-lactamase.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Some time ago, we showed that certain d-lactones, of
general structure 1, were b-lactamase substrates. In
some cases, these compounds produced relatively inert
acyl-enzymes and thus may be a source of inhibitory
substrates. More recently, we showed that in analogous
acyclic depsipeptide substrates of b-lactamases such as
2,2,3 replacement of the normal amide side chain,
RCONH–, with a retro-amide, RNHCO–, led to com-
pounds, 3, that, rather surprisingly, were also b-lacta-
mase substrates and, in some cases, better than the
original compounds.4 Since the retro-amide side chain
has not previously been seriously explored in b-lacta-
mase substrates or inhibitors, we decided to examine it
further. In this paper, we report the synthesis and reac-
tivity with typical b-lactamases of the benzopyranones
4m and 4o, each carrying the retro-amide side chain.
These compounds are also of interest since, as with
In view of the present-day and growing resistance of
bacteria to b-lactam antibiotics,1 it is important to
closely examine indications of novel specificity and
reactivity in substrates of the relevant enzymes. The
latter include the DD-peptidases (penicillin-binding
proteins) that are essential for bacterial cell wall biosyn-
thesis and are the b-lactam targets, and the b-lactamases
that provide protection to bacteria by catalyzing b-lac-
tam hydrolysis. The appearance of new chemistry in
substrates, of course, can be translated into new inhibi-
tors.
RCONH
RCONH
O
-
O
O
O
CO
CO
2
n
2
3
-
O
CO
2
RNHCO
O
PhCH2NHCO
O
PhCH2NHCO
O
1; n = 1,2
-
-
2
O
CO2
O
-
CO2
4o
4m
PhCH2CONH
O
PhCH2CONH
O
Keywords: Enzyme substrates; Inhibitors; b-Lactam antibiotics; Sub-
strate analogues.
-
O
CO2
O
*
Corresponding authors. Tel.: +1 8606852629; fax: +1 8606852211;
e-mail: rpratt@wesleyan.edu
-
CO2
5o
5m
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.07.067

5118
S. A. Adediran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5117–5120
bicyclic b-lactams, following their acylation of a serine
b-lactamase or DD-peptidase, the leaving group will re-
main attached, occupying the active site of the enzyme.
This is generally believed to be an important factor in
the antibiotic activity of bicyclic b-lactams.5 The kinetics
results obtained for 4m and 4o were compared to those
from the ÔnormalÕ side chain analogues 5m and 5o.6,7
the remaining benzyl protecting group furnished the tar-
get benzopyranones 4m and 4o.
The hydrolysis of 4m and 4o could be monitored spec-
trophotometrically at 290nm (De = 1760cmꢀ1 Mꢀ1
)
and 300nm (De = 950cmꢀ1 Mꢀ1), respectively. Pseudo-
first order rate constants for hydrolysis of 4m and 4o
in 0.1M MOPS buffer, pH7.5, 25ꢁC (the buffer condi-
tions employed in all kinetics studies reported in this
paper unless otherwise noted) were 3.0 · 10ꢀ4 and
1.7 · 10ꢀ4 sꢀ1, respectively, and thus similar to those ob-
tained previously6,7 for 5m and 5o, viz. 5.6 · 10ꢀ4 and
5.4 · 10ꢀ4 sꢀ1, respectively. In all cases, the rates are
higher than those for acyclic analogues;6 d-lactones are
less stable to nucleophilic cleavage than analogous acy-
clic esters.11
2. Results and discussion
For the syntheses of the target benzopyranones of type
4m and 4o (dihydrocoumarins possessing a 3-carbamoyl
substituent), we considered the alkylation of the methyl
benzylaminocarbonylacetate 68 with a substituted benzyl
bromide 7 (Scheme 1). The alkylation of malonamate
anions has, however, been seldom used in synthesis.9
The hydrolysis of 4m and 4o, again monitored spectro-
photometrically, was accelerated on addition of b-lacta-
mases. These enzymes therefore catalyze hydrolysis of
4m and 4o as they do of 5m and 5o.6,7 All four com-
pounds, as prepared, are racemic mixtures, with a chiral
centre at C3. Progress curves for the hydrolysis of 5m
and 5o in the presence of b-lactamases were biphasic,6,7
the faster phase enzyme-catalyzed, the slower not. This
was interpreted to mean that one enantiomer, presum-
ably the 3b-isomer by analogy with b-lactams, was a
b-lactamase substrate, while the other was not. In con-
trast, the spectrophotometric total progress curves for
the hydrolysis of 4m and 4o in the presence of b-lacta-
mases exhibited only one smooth phase of reaction. This
can be interpreted in terms of Scheme 2, where fast
equilibration of the 3a- and 3b-isomers would lead to
observation of only a single reaction phase.
The anion of 6 was generated in dry DMF, using sodium
hydride as a base, and reacted with different substituted
benzyl bromides derived from salicylic acid10 or 3-
hydroxybenzoic acid;6 the methoxycarbonyl substitutent
was therefore in either the ortho or meta position relative
to the acetoxy function. The reaction mixture contained
the monoalkylated product 8 as well as some dialkylated
material in the case of the meta series. After purification
by column chromatography of the monoalkylated com-
pounds and alkaline hydrolysis of the ester functions of
8, direct thermal lactonization of the diacids 9 failed,
leading to decarboxylated products. Therefore, selective
benzylation of the carboxyl groups of 9 was achieved in
DMF or DMSO using cesium carbonate and benzyl
bromide as reactants. Then lactonization of the dibenzyl
esters 10 gave monoesters 11. Finally, hydrogenolysis of
O
O
PhCH NH
2
PhCH NH
2
a,b
CO Me
Br
AcO
+
MeO
8o,8m
MeO
O
Ac
2
O
CO Me
2
O
7o,7m
6
c
O
O
O
PhCH NH
d
2
PhCH NH
2
HO
BnO
O
H
O
H
CO H
2
O
CO Bn
2
9o,9m
10o,10m
e
O
O
f
PhCH NH
PhCH NH
2
2
O
4o,4m
O
O
O
CO H
CO Bn
2
2
11o,11m
Scheme 1. Synthesis of the benzopyranones 4o and 4m. Reagents and conditions: (a) NaH/DMF/25ꢁC; (b) SiO2/cyclohexane–EtOAc (7:3); (c)
NaOH/MeOH–H2O; (d) i. Cs2CO3 (1equiv)/DMF, ii. PhCH2Br (2equiv)/DMF/25ꢁC; (e) 220ꢁC/1.33Pa; (f) H2/Pd/C/EtOAc.

S. A. Adediran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5117–5120
5119
PhCH NHCO
PhCH NHCO
2
2
fast
-
-
O
O
O
O
CO
CO
2
2
H O
2
H O
2
β
-lactamase
OH
OH
PhCH NHCO
PhCH NHCO
2
2
-
-
CO
CO
-
-
2
2
CO
CO
2
2
Scheme 2.
The hydrogen atoms at the 3-position of 4m and 4o must
therefore be kinetically more acidic than those of 5m
and 5o. This would not be unexpected since 4m and 4o
are malonates with two carbonyls adjacent to C3. In-
deed, ready exchange of these hydrogens without notice-
able lactone hydrolysis was observed in NMR spectra of
these compounds in acetone-d6/D2O. A similar result
was found with benzofuranones analogous to 4 where
exchange at C3was also sufficiently facile as to lead to
apparently simple kinetics.12
ko
S1
P1
P2
ko
S2
K
S1
S2
kcat
E + S1
ES1
E + P1
Km
Scheme 3.
In view of the discussion above, the kinetics of b-lacta-
mase-catalyzed hydrolysis of 4m and 4o were interpreted
in terms of Scheme 3, where S1 and S2 represent the 3b-
and 3a-isomers of 4m and 4o, and P1 and P2 their hydro-
enzyme. The results of these experiments are presented
in Table 1, along with comparable data for the analo-
gous compounds with normal side chains, 5m and 5o,
and for a good substrate, cephalothin.
lysis products, respectively. It can be shown that in
obs
cat
terms of Scheme 3, k ¼ kcat and Kombs ¼ 2Km, the latter
assuming, reasonably, that K = 1. Steady state kinetics
parameters, determined spectrophotometrically, were
generally determined from initial rates of hydrolysis of
4m and 4o by the class A TEM-2 and Staphylococcus
aureus PC1 b-lactamases, the class C b-lactamase of
Enterobacter cloacae P99, and the class D OXA-1
Superficially, the first notable result is that, generally,
the retro-side chain compounds 4m and 4o are b-lacta-
mase substrates, and often as good as those with a nor-
mal amide side chain, 5m and 5o. This indicates that, as
Table 1. Steady state kinetics parameters for b-lactamase-catalyzed hydrolysis
Enzyme
Substrate
b-4m
b-4o
b-5ma
b-5ob
Cephalothin
g
P99
kcat (sꢀ1
Km (mM)
)
1.18 0.09
0.43 0.09
2.7 · 103
0.442 0.0035.6
(2.0 0.1) · 10ꢀ4c
2.2 · 106
5.4
0.29
200
0.62
8.7 · 103
0.009
2.2 · 107
kcat/Km (sꢀ1 Mꢀ1
)
)
)
)
1.9 · 104
TEM
PC1
kcat (sꢀ1
Km (mM)
)
>0.02
>0.2
92
25.5 0.07
0.29 0.02
8.8 · 104
>1
> 1.2
1.2 · 103
2.07
0.0530.2
3.9 · 104
120g
kcat/Km (sꢀ1 Mꢀ1
4
6 · 105
kcat (sꢀ1
Km (mM)
)
>0.005
>0.2
25.0 0.1
(1.06 0.01) · 10ꢀ2d
(1.8 0.2) · 10ꢀ3d
5.8 · 103
2.4 · 10ꢀ2
1.5 · 10ꢀ2
1.6 · 103
4.7 · 10ꢀ3
3.0 · 10ꢀ4
1.55 · 104
0.001g
60.001
P103
kcat/Km (sꢀ1 Mꢀ1
OXA-1e
kcat (sꢀ1
Km (mM)
)
>0.15
>0.2
735
0.25 0.09
(7.4 3.2) · 10ꢀ3
1.2
0.0139
8.6 · 104
kcat/Km (sꢀ1 Mꢀ1
640f
5
6170f
3.39 · 104
a Data from Ref. 6; rate constants from the first kinetics phase (see text).
b Data from Ref. 7; rate constants from the first kinetics phase (see text).
c Determined from competition experiments with cephalothin as the reporter substrate.
d Determined from total progress curves.
e Buffer employed: 20mM MOPS, 50mM NaHCO3, pH7.50.
f No reaction observed above background.
g Data from Ref. 7.

5120
S. A. Adediran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5117–5120
to the congested active site of the acyl-enzymes derived
from bicyclic substrates.
k3[H2O]
E + POH
k2
ES
E + S
ES'
Ks
Thus, the rate constant for acylation of the P99 b-lacta-
mase by 4o is an impressive 2.2 · 106 sꢀ1 Mꢀ1 and the
deacylation rate is a modest 0.44sꢀ1. These parameters
suggest that 4o may represent a lead compound to a
new series of inhibitory substrates of class C b-lacta-
mases. Compound 4o also more rapidly inactivated the
Streptomyces R61 DD-peptidase than did 5o; the second
order rate constants for such inactivation were 22 and
1.5sꢀ1 Mꢀ1,7 respectively.
E + POMe
k4[MeOH]
Scheme 4.
with the acyclic analogues previously examined,4 the retro-
amide functionality must fit into the active site without
significant difficulty, and perhaps, in some cases, with
positive interactions. The lack of side chain specificity
evident from this result may reflect the role of b-lacta-
mases as resistance enzymes; evolutionary selection of
such enzymes may reward breadth of specificity. The
most dramatic result seen in Table 1, is that of 4o with
the P99 b-lactamase where a Km value of 0.2lM indi-
cates tight binding of this molecule to the active site.
The comparable parameter for 5o is more than 1000
times larger (weaker binding). The class D OXA-1 b-lac-
tamase appears to discriminate between these substrates
more than the other enzymes: although b-5o reacts rap-
idly with the enzyme, b-4o and, more particularly, the
meta-compounds are much less reactive.
Acknowledgements
This research was supported by the U.S. National Insti-
tutes of Health (RFP).
References and notes
1. Koch, A. L. Crit. Rev. Microbiol. 2000, 26, 205.
2. Govardhan, C. P.; Pratt, R. F. Biochemistry 1987, 26,
3385.
3. Xu, Y.; Soto, G.; Hirsch, K. R.; Pratt, R. F. Biochemistry
1996, 35, 3595.
Experiments with an additional nucleophile showed that
kcat for both 4m and 4o with the P99 enzyme must rep-
resent k3, the deacylation rate constant [Scheme 4, where
ES and ES0 represent noncovalent and covalent (acyl-
enzyme) enzyme–substrate complexes, respectively].
Methanol accelerated the observed initial rates of reac-
tion of 4m and 4o to an extent corresponding to k4/k3
values of 42 and 27, respectively. These values are com-
parable to those generally observed for depsipeptide
substrates of the P99 b-lactamase2–4,6,7,12 and indicate
that methanol has unfettered access to the acyl-enzyme.
On the other hand, D-phenylalanine (0–40mM) had no
effect on the rate of disappearance of 4. This contrasts
with results with acyclic depsipeptides with normal amido
side chains,13 but is in accord with results for bicyclic
substrates, including 5 and b-lactams,6 and with results
for 3.4 More hindered nucleophiles have limited access
4. Cabaret, D.; Adediran, S. A.; Pratt, R. F.; Wakselman, M.
Biochemistry 2003, 42, 6719.
5. Pratt, R. F. J. Chem. Soc., Perkin Trans. 2 2002, 851.
6. Cabaret, D.; Adediran, S. A.; Garcia-Gonzalez, M. J.;
Pratt, R. F.; Wakselman, M. J. Org. Chem. 1999, 64, 713.
7. Adediran, S. A.; Cabaret, D.; Pratt, R. F.; Wakselman, M.
Bioorg. Med. Chem. Lett. 1999, 9, 341.
8. Smith L. H. DE 2,745,222; Chem. Abstr. 1978, 89, 42840.
9. Gardner, P. D.; Brandon, R. L. J. Org. Chem. 1957, 22,
1704.
10. Regnier, G.; Canevari, R.; Le Douarec, J. C. Bull. Soc.
Chim. Fr. 1966, 2821.
11. Bruice, T. C.; Benkovic, S. A. In Bioorganic Mechanisms;
Benjamin: N.Y., 1966; Vol. 1, pp 21–22.
12. Adediran, S. A.; Cabaret, D.; Drouillat, B.; Pratt, R. F.;
Wakselman, M. Bioorg. Med. Chem. 2001, 9, 1175.
13. Pazhanisamy, S.; Govardhan, C. P.; Pratt, R. F. Biochem-
istry 1989, 28, 6863.