Kinetics of action of a two-stage pro-inhibitor of serine β-lactamases

Article abstract of DOI:10.1021/bi400873r

β-Lactamase inhibitors are important in medicine in the protection of β-lactam antibiotics from β-lactamase-catalyzed destruction. The most effective inhibitors of serine β-lactamases covalently modify the enzyme active site. We have recently studied O-acyl and O-phosphyl hydroxamates as a new class of such inhibitors. In this paper, we describe our studies of the N-acyl derivatives of a cyclic O-acyl hydroxamic acid, 3H-benzo[d][1,2]oxazine- 1,4-dione, and, in particular, the N-tert-butoxycarbonyl derivative. This compound is not a β-lactamase inhibitor itself but undergoes spontaneous hydrolysis in aqueous solution, yielding an O-phthaloyl hydroxamic acid, which is a β-lactamase inhibitor. This compound spontaneously, but reversibly, cyclizes in solution to form phthalic anhydride, which is also a β-lactamase inhibitor. Both inhibitors react to form the same transiently stable phthaloyl-enzyme complex. Thus, we have a two-step cascade, beginning with a pro-inhibitor, in which each step leads to a different inhibitor, presumably with different enzyme specificities. The kinetics of these transformations have been elucidated in detail. The phthaloyl derivatives, where the free carboxylate is important for facile reaction with the enzyme, represent a new lead for serine β-lactamase inhibitors. Analogues can be conveniently constructed in situ by reaction of nucleophiles with phthalic anhydrides and then screened for activity. Active hits may then become new leads.

Full text of DOI:10.1021/bi400873r

Article
pubs.acs.org/biochemistry
Kinetics of Action of a Two-Stage Pro-Inhibitor of Serine
β‑Lactamases
Ronak Tilvawala and R. F. Pratt*
Department of Chemistry, Wesleyan University, Lawn Avenue, Middletown, Connecticut 06459, United States
S
* Supporting Information
ABSTRACT: β-Lactamase inhibitors are important in medicine in the protection of β-lactam antibiotics from β-lactamase-
catalyzed destruction. The most effective inhibitors of serine β-lactamases covalently modify the enzyme active site. We have
recently studied O-acyl and O-phosphyl hydroxamates as a new class of such inhibitors. In this paper, we describe our studies of
the N-acyl derivatives of a cyclic O-acyl hydroxamic acid, 3H-benzo[d][1,2]oxazine-1,4-dione, and, in particular, the N-tert-
butoxycarbonyl derivative. This compound is not a β-lactamase inhibitor itself but undergoes spontaneous hydrolysis in aqueous
solution, yielding an O-phthaloyl hydroxamic acid, which is a β-lactamase inhibitor. This compound spontaneously, but
reversibly, cyclizes in solution to form phthalic anhydride, which is also a β-lactamase inhibitor. Both inhibitors react to form the
same transiently stable phthaloyl−enzyme complex. Thus, we have a two-step cascade, beginning with a pro-inhibitor, in which
each step leads to a different inhibitor, presumably with different enzyme specificities. The kinetics of these transformations have
been elucidated in detail. The phthaloyl derivatives, where the free carboxylate is important for facile reaction with the enzyme,
represent a new lead for serine β-lactamase inhibitors. Analogues can be conveniently constructed in situ by reaction of
nucleophiles with phthalic anhydrides and then screened for activity. Active hits may then become new leads.
β-Lactamases catalyze the hydrolysis of β-lactams and are
thereby the main source of bacterial resistance to these
antibiotics. Inhibitors of these enzymes have demonstrated
potential in medical practice where they are used in
combination with β-lactams.1 The β-lactamase inhibitor of
such a pair eliminates the resistance produced by these
enzymes, allowing the β-lactam to reach its DD-transpeptidase
target. The continuing evolution of β-lactamases, enforced by
the selective pressure of both β-lactams and β-lactamase
inhibitors, has provided the impetus for the ongoing search for
new β-lactamase inhibitors.2−5
The search for broad-spectrum β-lactamase inhibitors is
complicated by the fact that these are two very different groups
of β-lactamases, the serine β-lactamases (classes A, C, and D)
that belong to the β-lactam-recognizing enzyme superfamily6
and the metallo-β-lactamases (class B), members of a distinct
superfamily of metalloenzymes that exhibit a variety of
functions.7 Catalysis by the serine enzymes involves a covalent
acyl−enzyme intermediate, whereas that by the metalloen-
zymes comprises attack on the substrate by a metal ion-
coordinated water molecule. Broad-spectrum inhibitors that
cover both of these major groups of enzymes are difficult to
achieve. Some progress in this direction has been made,8 but no
candidates have yet been brought forward for development. It
has been easier to achieve potent inhibitors of each group
separately, with design focusing on acyl transfer chemistry in
the case of the serine enzymes and on the metal ion in that of
the metallo-β-lactamases.
Most effective inhibitors of the serine β-lactamases interact
covalently with the active site, either as transition state
analogues or as mechanism-based inhibitors.9 Many are β-
lactams themselves, such as clavulanic acid, sulbactam, and
tazobactam,1,2 which are currently used in medical practice.
Such inhibitors remain, however, susceptible to evasion by
mutant β-lactamases.10 Thus, inhibitors that are not β-lactams
are of considerable interest. Examples of these investigated to
date include boronic acids, and various phosphonyl, sulfonyl,
and acyl derivatives.5 Recently, we have described aryloxy-
Received: July 3, 2013
Revised: September 10, 2013
Published: September 26, 2013
© 2013 American Chemical Society
7060
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 1
carbonyl hydroxamates 1, which are general inhibitors of serine
β-lactamases.11,12 These compounds first acylate the active site
serine residue with release of the aryloxide. Irreversible
interaction then follows by intramolecular aminolysis by a
conserved active site lysine residue (2) with displacement of the
hydroxamate leaving group and formation of a cross-linked
active site (Scheme 1).
In view of these results, we have explored other types of O-
acyl hydroxamates in more detail. Recently, we described some
new acylic variants, including a series of phosphyl hydrox-
amates.13 In the work presented here, we explore the potential
of some cyclic analogues. We have thus prepared compounds
3−6 and assessed their activities as β-lactamase inhibitors.
procedure for phthalic anhydride 41 is given in the Supporting
Information.
Synthesis. Compounds 3−6 were prepared as described by
Zinner et al.18
tert-Butyl 1,4-Dioxo-1H-benzo[d][1,2]oxazine-3(4H)-car-
boxylate (5). The crude solid product was recrystallized from
a dioxane/ether mixture (1:1) to afford compound 5 in 60%
yield as a colorless solid: mp 226−228 °C (lit.18 226−228 °C);
1H NMR (d6-DMSO) δ 1.56 (s, 9H), 8.02 (m, 2H), 8.16 (m,
2H); FTIR (KBr) 1780, 1762, 1713 cm−1; ES(+)MS 301.00
(M + H+).
3H-Benzo[d][1,2]oxazine-1,4-dione (3). This compound
was obtained in 75% yield as an off-white solid after
recrystallization from a cyclohexane/benzene mixture (1:1):
mp 222−224 °C (lit.18 225 °C); H NMR (d6-DMSO) δ 8.04
1
(m, 2H), 8.20 (d, J = 6.0 Hz, 2H); FTIR (KBr) 1768, 1677
cm−1; ES(−)MS 162.00 (M − H+).
3-Methyl-3H-benzo[d][1,2]oxazine-1,4-dione (4). The
crude solid product was recrystallized from ethanol to afford
4 in 40% yield as a colorless solid: mp 99−100 °C (lit.18 120
°C); 1H NMR (d6-DMSO) δ 3.56 (s, 3H), 7.95−8.16 (m, 4H);
FTIR (KBr) 1762, 1657 cm−1; ES(+)MS 178.07 (M + H+).
3-Benzoyl-3H-benzo[d][1,2]oxazine-1,4-dione (6). The
crude product was recrystallized from a benzene/cyclohexane
mixture (1:1) to afford 6 in 60% yield as a colorless solid: mp
Although none of these molecules, in themselves, turned out
to be significant inhibitors of β-lactamases, N-acyl derivatives 5
and 6 generated inhibitors in situ by interesting chemical
rearrangements that we describe in this paper. Thus, 5 and 6
were found to be pro-inhibitors. Pro-inhibitors, as leads to pro-
drugs, are of considerable significance to many classes of
drugs,14 including β-lactams.15,16 In many cases, the active drug
is more efficiently delivered to the site of action in a pro-drug
form, to be released in its active form after the important
pharmaceutical, pharmacokinetic, and pharmacodynamic bar-
riers have been finessed.17 In many cases, the pro-drug is a
chemically simple derivative from which the active principle can
be released by spontaneous or enzyme-catalyzed hydrolysis. In
some cases, however, the conversion to the active form involves
a more complicated rearrangement. Both modes are illustrated
by the molecules described in this paper.
1
136−138 °C; H NMR (d6-DMSO) δ 7.49−8.23 (m, 9H);
FTIR (KBr) 1764, 1754, 1701 cm−1; ES(+)MS 268.07 (M +
H+).
N-(tert-Butoxycarbonyl)-O-(2-carboxybenzoyl)-
hydroxylamine (7) and Its 15N Isotopomer. A solution of
triethylamine (0.57 mL, 4.13 mmol) in DCM (2 mL) was
added dropwise to a solution of tert-butyl N-hydroxycarbamate
(0.5 g, 3.75 mmol) in DCM (20 mL) at 0 °C. A solution of
phthalic anhydride (0.61 g, 4.13 mmol) and DMAP (45 mg,
0.10 mmol) in DCM (5 mL) was then added dropwise over a
period of 10 min. The reaction mixture was stirred for a further
4 h to room temperature. The organic layer was washed twice
with water, after which the organic phase was dried over
Na2SO4 and evaporated under reduced pressure. The crude
solid was purified by preparative silica gel thin layer
chromatography with an ethyl acetate/hexane/ether/acetic
acid mixture (4:45:50:1) as the solvent. The resulting material
was recrystallized from a dioxane/benzene/cyclohexane
mixture (1:10:3) to afford 7 in 35% yield as a colorless
EXPERIMENTAL PROCEDURES
■
The class C P99 β-lactamase from Enterobacter cloacae P99 and
the TEM-2 β-lactamase from Escherichia coli were purchased
from the Centre for Applied Microbiology and Research
(Porton Down, Wiltshire, U.K.). The Streptomyces R61 and
Actinomadura R39 DD-peptidases were generously supplied by
J.-M. Frer
Charlier (University of Lieg
̀
e of the University of Lieg
̀ ̀
e (Liege, Belgium) and P.
̀
e).
The following hydroxamic acids and anhydrides were
purchased commercially and used as received: 16, 18, 20, 21,
23, 24, 26, and 45 (Sigma-Aldrich), 17, 27, 42−44, 46, 48, and
49 (Acros), 19, 34, and 40 (Alfa-Aesar), 22, 39, and 50 (TCI
America), 32 and 35 (Fluka), 25 (Mallinckrodt), 36
(Calbiochem), and 47 (Matheson Coleman & Bell).
1
crystalline solid: mp 96−98 °C; H NMR (d6-DMSO) δ 1.42
(s, 9H), 7.69 (s, 3H), 7.8 (m, 1H), 10.98 (s, 1H), 13.38 (s,
1H); FTIR (KBr) 1790, 1698, 1667 cm−1; HRMS (ES+)
304.0799 (M + Na+), calcd for C13H15NO6Na 304.0797.
The 15N isotopomer of this compound was prepared as
described above but using tert-butyl [15N]-N-hydroxycarba-
mate12 as the starting material: 1H NMR (d6-DMSO) δ 1.42 (s,
Hydroxamic acids 28−31, 33, 37, and 38 were synthesized
following the procedure described previously.13 The synthetic
7061
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 2
9H), 7.69 (s, 3H), 7.8 (m, 1H), 11.00 (d, J = 92.1 Hz, 1H),
13.4 (s, 1H).
Identification of the Intermediates of Spontaneous
Reactions by H NMR. MOPS buffer (20 mM, pH 7.5, 20
1
N-(Benzyloxycarbonyl)-N-(benzoyl)hydroxylamine (8) and N-
(tert-butoxycarbonyl)-O-(benzoyl)hydroxylamine (9) were pre-
pared by standard procedures,19 the former as shown in
Scheme 2. Experimental details are provided in the Supporting
Information.
mL) was added to the stirred solution of the reactant (5 mg)
dissolved in acetonitrile (2 mL). The reaction mixture was then
stirred for an appropriate time period (7 min for 7 and 3 or 90
min for 5). The reaction was quenched with 0.1 M HCl (∼3
mL) to pH <2. The resulting solution was extracted twice with
ethyl acetate. The organic layer was separated, dried over
Na2SO4, and evaporated under reduced pressure. This residue
was dried under an oil pump vacuum overnight. H NMR
spectra were obtained with d6-DMSO as the solvent. The same
sample was used for mass spectrometric analysis.
Equilibrium Experiments with Compound 7. To determine
the spontaneous reaction rate of 7, three experiments were
conducted separately: (i) reaction of 7 alone (200 and 400
μM), (ii) same as (i) in the presence of 500 μM 11, and (iii)
reaction of 10 (200 and 400 μM) in the presence of 500 μM
11. All reactions were monitored spectrophotometrically at 260
nm. All six curves were simultaneously fit to Scheme 3 using
Dynafit.20
Kinetic Methods. Absorption spectra and spectrophoto-
metric reaction rates were obtained from Hewlett-Packard
8452A and 8452E spectrophotometers. Enzyme concentrations
were determined spectrophotometrically. Kinetic experiments
were conducted at 25 °C in a 20 mM MOPS buffer solution at
pH 7.5. The inhibitors were prepared in concentrated DMF
stock solutions and diluted to 1−5% DMF in reaction mixtures.
Spontaneous Hydrolysis Reactions. The spontaneous
hydrolysis rates of the potential inhibitors were measured
spectrophotometrically in replicate at appropriate wavelengths
and concentrations (Table 1). The full progress curves were
then fit to a first-order rate equation and the rate constants thus
obtained from averaged replicates (Table 1).
1
Direct Turnover by the P99 β-Lactamase. The enzyme
(final concentrations of 1−20 μM) was added to the buffered
solution of the inhibitory substrate (final concentrations of 100
μM for 10 and 200 μM for 7) and the subsequent hydrolysis
monitored spectrophotometrically at an appropriate wavelength
(Table 1). The measured initial rates were plotted against
enzyme concentration and fit to a linear equation.
Time-Dependent Enzyme Inhibition. The P99 β-lactamase
(0.5 μM) was incubated with inhibitor (500 μM) in buffer (100
μL) containing 0.1% BSA. At appropriate times, aliquots (5 μL)
of the reaction mixture were diluted to 0.50 mL in MOPS
buffer containing cephalothin (300 μM), and hydrolysis of the
latter was monitored spectrophotometrically at 280 nm over a
suitable period of time. The measured initial rates, proportional
to the remaining free enzyme concentration, were plotted
against the time of incubation. These data were fit to
Table 1. Rate Constants for Spontaneous Hydrolysis of Acyl
Hydroxamates and Phthalic Anhydrides
compd
k0 (s−1
)
λ (nm)
Δε (cm−1 M−1
)
a
3
<1 × 10−6
250−300
300
−
2100
3125
5000
1475
635
4
(7.6 0.1) × 10−4
0.034 0.001
5
260
6
0.029 0.001
300
b
7
(4.0 0.1) × 10−3
(2.2 0.1) × 10−3
(2.0 0.2) × 10−6
0.015 0.001
260
8
300
9
250
1775
2250
40000
2275
10
39
41
300
(8.1 0.1) × 10−3
(8.6 0.1) × 10−3
260
300
a
b
No observed reaction. Cyclization to 10 (Scheme 3).
Scheme 3
7062
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
appropriate reaction schemes (see Results and Discussion)
using Dynafit.20
phthalic anhydride 10 as a transient intermediate and tert-butyl
N-hydroxycarbamate 11 and phthalate 12 as final products
(Figure S1 of the Supporting Information). Identical spectra
(Figure S1 of the Supporting Information) were obtained from
a mixture of 10 and 11, indicating the presence of a reversible
equilibrium between these species and 7. These results were
suggestive of Scheme 3. The rate constants of Scheme 3 were
determined from spectrophotometric rate measurements.
Reaction mixtures were set up in three different ways: (i)
beginning with 7 alone, (ii) beginning with 7 together with 11,
and (iii) beginning with 10 in the presence of 11. Data from
these were fit simultaneously to Scheme 3 (Figure S2 of the
Supporting Information) to obtain values of k2 and k−2, given
the independently measured rate constant for hydrolysis of 10
to 12 (viz. 0.015 s−1). Direct hydrolysis of 7 was assumed to be
slow in view of the results for 9. The rate constants thus
obtained are shown in Scheme 3. At millimolar concentrations
of 11, the equilibrium lies in favor of 7 (vs 10), and thus, under
these conditions, the hydrolysis of 10 will be impeded by the
presence of 11. The situation displayed in Scheme 3 has
previously been observed with monoaryl phthalates.26
Competitive Enzyme Inhibition Experiments. CENTA (50
μM; Km = 19.5 μM) was employed as the substrate, and its
hydrolysis was monitored at 410 nm.21 The P99 β-lactamase
concentration was 1.0 nM, and concentrations of the inhibitor
were 10−300 μM. Total progress curves were fit to an
appropriate scheme (see Results and Discussion) using
Dynafit.20 This experiment was also conducted to screen
combinations of various anhydrides (100 μM) and nucleophiles
(100 μM).
Mass Spectrum of the EI Complex. The P99 β-lactamase
(20 μM, 100 μL) was incubated with 7 (200 μM) for 50 s. A
small aliquot was taken and assayed for enzyme activity to
confirm inactivation. The reaction was then quenched with 50%
aqueous trichloroacetic acid (final concentration of 5%). A
sample for mass spectrometry was prepared as previously
described.13
Kinetics with Other Enzymes. The competitive inhibition
experiment described above was also performed with the TEM-
2 β-lactamase (5.0 nM) and 10 (500 μM) alone as well as with
10 (500 μM) in the presence of 11 (500 μM). The hydrolysis
of the substrate nitrocefin (200 μM; Km = 86 μM) was
monitored at 600 nm.22 This experiment was also performed
with the Streptomyces R61 DD-peptidase (0.2 μM) and
Actinomadura R39 DD-peptidase (0.3 μM), employing m-
carboxyphenyl N-phenylacetylglycinate (Km = 0.76 mM at 2.0
mM and Km ≫ 1 mM at 2.0 mM, respectively, monitored at
310 nm) as the substrate.23
Spontaneous reaction of 5 in solution is complicated.
1
Quenched reaction samples gave H NMR spectra (Figure S3
of the Supporting Information) showing the presence of 7, 3,
12 and 13. This result was supported by ES(−) mass spectra of
quenched reaction mixtures that showed peaks at 162 amu (3)
and 165 amu (12). An ES(+) spectrum showed a peak at 282
amu that corresponds to 7 and/or 13. The likely presence of 13
1
as well as 7 is indicated by a H NMR peak at 10.2 ppm
[attempts to independently synthesize 13 were unsuccessful,
but the close analogue 8 was prepared; the 1H NMR spectrum
of 8 contained a peak at 10.6 ppm (OH)]. On further reaction,
the peak at 10.2 ppm did disappear with time at a rate
comparable to that of 7 (Figure S3 of the Supporting
Information). These spectra suggested that 5 reacts in solution
to form ∼10% each of 3 and 13, with the remainder converted
to 7. This information was used in fitting the spectrophoto-
metric data, as described below.
The absorption at 260 nm of a solution of 5 at pH 7.5
decreases with time in two phases (Figure S4 of the Supporting
Information). These data were fit to Scheme 4, by fixing k2, k−2,
and k3 to the values of Scheme 3 and the rate constants for
conversion of 5 to 3 and 5 to 13 to the value of 0.1k1, the latter
assignment based on the NMR data described above. The
hydrolyses of 3 and 13, thought to represent minor
contributions to the absorption changes at 260 nm (that of 3
certainly does, from the independent study of 3, described
above), were omitted from the fit. Compound 13 does react
further, as noted above, perhaps by cyclization to form 10
(Scheme 4), but probably also directly to 12 because 8
hydrolyzes quite readily to benzoate (see above). Compound
13 is kinetically less accessible from 10 (α ≪ 1) than is 7
because addition of 11 to 10 appears to immediately generate
essentially only 7, as indicated by NMR spectra (Figure S1C of
the Supporting Information).
RESULTS AND DISCUSSION
■
Spontaneous Reactions in Solution. The pseudo-first-
order rate constants of spontaneous reactions of 3−10 in
MOPS buffer at pH 7.5 were obtained from spectrophoto-
metric measurements (see Experimental Procedures) and are
listed in Table 1. Compounds 4 and 9 hydrolyzed in a simple
one-step exponential fashion. The product from hydrolysis of 4
arose from nucleophilic attack on the acyl hydroxamate
carbonyl to yield the hydroxamic acid product.24,25 Presumably,
this reaction would also occur for 3, but in this case, the
hydrolysis was extremely slow at pH 7.5 because , as indicated
by absorption spectra taken at pH 1−13, the compound
probably exists as an inert anion at that pH (cf. 112). The
hydrolysis of 9 also yielded the hydroxamic acid (11) and
carboxylate (benzoate). The N,N-diacylhydroxylamine 8 hydro-
lyzed in a simple one-step process to yield benzoate and the
hydroxamic acid (benzyl N-hydroxycarbamate), as indicated by
1H NMR spectra taken during the reaction. It appeared to be
surprisingly reactive (Table 1) compared with the N,O-diacyl
compound 9, possibly because of intramolecular general base
catalysis by the anion of 8.
The value of k1 (0.034 0.001 s−1), obtained from the fit of
the data of Figure S4 of the Supporting Information, shows that
5 is quite unstable in solution, rapidly (t1/2 = 20 s) converting
to (mainly) 7, yielding the first kinetic phase noted above and
seen in Figure S4 of the Supporting Information. Compound 7
then proceeds via 10 to 12 (Scheme 3) in the second, slower
phase.
The rapid spontaneous reaction of 7 in aqueous buffer was
first monitored by 1H NMR spectroscopy. This identified
7063
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 4
Inhibition of the P99 β-Lactamase. Preliminary experi-
ments indicated that incubation of the P99 β-lactamase with
either 5 or 6 led to transient inhibition of the enzyme. Because
of the time dependence of this phenomenon and the instability
of 5 and 6 in solution (see above), it was necessary to carefully
assess the various species of Scheme 4 for inhibitory activity.
Because the enzyme activity loss was transient and full activity
was eventually recovered, inhibitory substrates seemed to be
involved.
Phthalic anhydride, 10, is certainly a transient inhibitor of the
β-lactamase. Figure 1A shows the activity of the enzyme as a
function of time after addition of phthalic anhydride. The
enzyme activity is lost in a rapid initial phase and then slowly
recovered. This data could be fit to Scheme 5 to give a k5 value
of (9.0 2.3) × 10−3 s−1. The value of k4 was obtained by
monitoring direct turnover of 10 spectrophotometrically
(Figure 1B), which yielded a k4 value of 800 70 M−1 s−1.
In Scheme 5, we assume, on the basis of all precedent,27−29 that
the transient inactivation of the P99 enzyme occurs by covalent
modification of the active site serine residue. To support this
assertion, we quenched a reaction mixture at an appropriate
time (see Experimental Procedures). A mass spectrum of the
precipitated protein yielded a mass of 39338 amu, which is as
expected for phthaloylation of the enzyme (mass of 39190
amu).
Reactivation of the transiently inactivated enzyme could
occur either by recylization to re-form 10 (dashed arrow, k−4)
or by direct hydrolysis (solid arrow, k5) (Scheme 5). The latter
is thought to dominate for reasons given below. The effective
thermodynamic dissociation constant, Ki, of phthalic anhydride
is given by k5/k4 and is thus 11 μM. We are unaware of any
previous study of the inactivation of a serine hydrolase by
phthalic anhydride. Isatoic anhydride, however, has been shown
to be a class C β-lactamase inhibitor (see below).30
The transient inhibition caused by 10 is also readily
demonstrated by competition experiments with a chromogenic
substrate. Figure 2 shows the results of such an experiment in
which the lag in turnover and the recovery of activity are clearly
visible. Fitting of the data to a reaction scheme involving a
combination of Scheme 5 with steady state turnover of the
substrate CENTA yielded a k5/k4 or Ki value (only the ratio
was available from numerical fitting of Scheme 5 to the data of
Figure 2) of 0.26 0.03 μM. This value is clearly much smaller
than that obtained from the direct experiment described above.
The difference is readily seen qualitatively in the lags prior to
Figure 1. (A) Activity of the P99 β-lactamase (0.5 μM) as a function
of time in the presence of 10 (0.50 mM). The data were fit () to
Scheme 5. (B) Initial velocity of hydrolysis of 10 (0.10 mM) as a
function of β-lactamase concentration.
reactivation in Figure 2 that are much more extended than that
observed in Figure 1A. The difference arises because turnover
of a cephalosporin substrate by the P99 enzyme generates a
different free enzyme form that reacts with acylating agents
7064
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 5
Figure 2. Turnover of CENTA (50 μM) by the P99 β-lactamase (1.0
◇
□
○
nM), alone ( ) and in the presence of 10 [0.10 ( ), 0.20 ( ), and
△
0.30 mM ( )]. The solid lines represent the fit of the data to Scheme
5.
and/or substrates in a manner different from that of the original
free enzyme.30−32 This experimental protocol is introduced
here because, as described below, it is convenient to
qualitatively demonstrate the relative effectiveness of a variety
of anhydride−nucleophile combinations by this method.
Because 7, an intermediate in Scheme 4, could be
independently synthesized (Experimental Procedures), its
inhibitory activity could be assessed directly. It is worth noting
at this point that the compound synthesized, by reaction of 10
with the hydroxamic acid 11, was shown to be the O-acyl
product 7 rather than the N-acyl analogue 13. This was proven
by synthesis of the 15N isotopomer of 7 and the observation of
1
direct 15N−H coupling (92.1 Hz) in the H NMR spectrum
(see Experimental Procedures). The high-frequency carbonyl at
1790 cm−1 is also indicative of O-acylation. Acylation of
hydroxamic acids, under conditions of mild base catalysis,
generally yields the O-acylated material, at least as a kinetic
product.12,33
Compound 7 also appeared to be a transient inhibitor of the
P99 enzyme, although, in view of the facile interconversion of 7
and 10 (Scheme 3), its activity could not be taken for granted.
Experiments to monitor enzyme activity versus time and
turnover were conducted in a manner similar to that described
for 10, with the results shown in panels A and B of Figure 3,
respectively. In addition, taking advantage of the interconver-
Figure 3. (A) Activity of the P99 β-lactamase (0.50 μM) as a function
of time in the presence of 7 (0.50 mM). The data were fit () to
Scheme 6. (B) Initial velocity of hydrolysis of 7 (0.20 mM) as a
function of β-lactamase concentration. (C) Activity of the P99 β-
lactamase (0.50 μM) as a function of time in the presence of a mixture
of 10 (0.50 mM) and 11 (0.50 mM). The data were fit () to
Scheme 6.
sion of 7 and 10, we conducted an experiment beginning with
10 and a concentration of 11 sufficient to rapidly drive most of
7065
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 6
10 (>90%) to 7. The β-lactamase inhibitory activity of this
mixture was also monitored as a function of time (Figure 3C).
The similarity of panels A and C of Figure 3, given the position
of the equilibrium between 10 and 7 [K2 = k−2/k2 = 10 μM
(discussed above)], strongly indicates that 7 is itself an
inhibitor. These data were evaluated in terms of Scheme 6.
Linear least-squares analysis of the data in Figure 3B yielded
a k6 value of 64 11 M−1 s−1. Incorporation of this value for k6
and the other previously determined rate constants shown in
Schemes 4 and 5 into the fits of panels A and C of Figure 3
yielded k−6 values of 19 17 and 14 6 M−1 s−1, respectively.
A value for the equilibrium constant K6 (= k−6/k6) of ∼0.25 for
the nucleophile-assisted dissociation of 14 was thus established.
As quite clearly indicated by Figure 3C and the analysis
presented above, therefore, 7 is directly an inhibitor of the β-
lactamase as well as indirectly via 10. Although more stable in
solution than the anhydride (k2 = 0.0040 s−1 vs k5 = 0.015 s−1),
7 is less effective as a β-lactamase inhibitor (k6 = 64 M−1 s−1 vs
k4 = 800 M−1 s−1). The combination of 10 and 11 can be more
effective than 10 alone (compare Figures 1A and 3C, for
example), however, because of the equilibrium in favor of the
more stable 7. It is noteworthy that 9, an analogue of 7 lacking
the o-carboxylate, is not detectably an inhibitor of the P99
enzyme. Apparently, as is often true with β-lactamase
substrates, the presence of a carboxylate substantially enhances
reactivity;34,35 the best β-lactam substrates all possess such a
carboxylate, which is known to interact with active site
functional groups.36,37 The relative chemical inertness of 9
(Table 1) is undoubtably also a factor. In contrast, the
chemically more reactive N,N-diacyl analogue 8 is measurably
an inhibitor, although it lacks the carboxylate [ki = 20 2 M−1
s−1 (data not shown)]. It is possible, however, that the OH or
O− group of 8 also favorably interacts with the enzyme active
site during acylation. Although the carboxylate group appears to
be advantageous for enzyme acylation, it seems less necessary
for deacylation. The deacylation rate constant [k5 = 0.0090 s−1
(Scheme 5)] is very similar to that oberved for deacylation of
the benzoyl-enzyme, generated from benzoyl phosphates.38
Presumably, the enzyme is able to directly catalyze, albeit
slowly, the hydrolysis of both of these species.
Compound 5 itself, however, is not a significant inhibitor.
This can be seen, for example, in Figure 4A, where reaction of 5
in the presence of the substrate CENTA and the P99 β-
lactamase shows burst kinetics. Fitting of this data to Scheme 7,
employing only the rate constants determined above, indicates
that 5 has little or no inhibitory activity, and the latter arises
only after generation of 7 and 10. In support of this
interpretation, the data of Figure 4B, which shows the recovery
of enzyme activity in the presence of 5, can also be fit by
Scheme 7, using the previously determined rate constants, i.e.,
also not requiring active inhibition by 5. Compound 5 is
therefore a pro-inhibitor, giving rise spontaneously in solution
to the inhibitors 7 and, subsequently, 10.
The experiments described above show that new inhibitors of
the P99 β-lactamase can be generated by addition of
nucleophiles to phthalic anhydride (Scheme 8). The effective-
ness of the generalized inhibitor 15 can be semiquantitatively
assessed by an experiment similar to that shown in Figure 2.
Thus, enzyme was added to a mixture of phthalic anhydride,
nucleophile, and the substrate CENTA. The enzyme-catalyzed
hydrolysis of CENTA was followed yielding data like those
shown in Figure 5.
Nucleophiles yielding traces identical to that of phthalic
anhydride alone probably did not react with the anhydride and
thus did not perturb its level of inhibition. Nucleophiles such as
17 that add to form an adduct that was not inhibitory, or less so
than 10 itself, yielded traces to the left of that for 10 alone
(Figure 5), while those that produced more inhibitory adducts,
such as the original 11 and 34, yielded traces to the right of 10
(Figure 5). A graph showing the results of such experiments
with the nucleophiles 16−38 is shown in Figure 6. The relative
effectiveness of the various compounds derives from a
combination of K2 (and the nucleophile concentration, of
course) and k6/k4. Figure 6 suggests that, in general, amines
and hydroxylamines, phenols, phosphates, and carboxylates
generate no or inactive adducts, whereas hydroxamates, as
originally discovered, generate adducts that are active inhibitors.
Those more effective than 11 are mainly aromatic hydroxamic
acids, although 36 is an interesting outlier. A good inhibitor 15
would therefore be generated by a good nucleophile (K2),
yielding an effective acylator (k6) of the active site, where the
intrinsic effectiveness of 15 would be a function of the
interactions between all of the components of 15, including the
carboxylate group, with the enzyme active site.
The thermodynamic cycle of Scheme 6 (7 ⇔ 10 ⇔ 14 ⇔ 7)
also permits calculation of k−4, the rate constant for breakdown
of 14 by recyclization to 10 (see Scheme 5). Thus, k−4
=
(K6k2k4)/k−2 = 0.0020 s−1. This indicates that some breakdown
of 14 does occur by recyclization to 10, although the major
path is by direct hydrolysis (k5).
It is possible that nucleophiles 16−19, interpreted above as
yielding inactive adducts, may alternatively act by nucleophilic
deacylation of 14. This seems less likely than the original
interpretation because nonspecific nucleophiles do not
Finally, as noted above, the original compound 5 also appears
to cause transient inhibition of the P99 β-lactamase.
7066
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
containing a mixture of CENTA and the P99 β-lactamase (see
above). It is likely that hydrolytic activation of 6 occurs, as with
5. A quantitative analysis of the activity of 6 has not yet been
performed.
One particularly important result obtained from the studies
described above is that compound 7, but not 9, is a significant
inhibitor of the P99 β-lactamase, giving rise to the acyl-enzyme
14. The carboxylate is thus important, relating, presumably, to
the general presence of carboxylate groups on β-lactamase
substrates and inhibitors.32−34 The transition state analogue
motif 51 may, presumably, also be approached in a number of
classical ways.40
Figure 4. (A) Initial part of the progress curve for hydrolysis of
CENTA (50 μM) by the P99 β-lactamase (1.0 nM) in the presence of
5 (25 μM). The data were fit () to Scheme 7. (B) Activity of the
P99 β-lactamase (0.50 μM) as a function of time in the presence of 5
(0.50 mM). The data were fit () to Scheme 7.
generally cleave acyl derivatives of the P99 enzyme.39 At any
event, the product 15 thus obtained must still be less active
than 10.
Isatoic anhydride (43) is reported to be a class C β-lactamase
inhibitor.28 In our hands, it appeared to be less effective than
10, and its activity was not enhanced by 11. Presumably, 52, the
adduct likely derived from 43, is also not a strong inactivator,
emphasizing again the importance of the o-carboxylate group in
7. Finally, the extremely reactive 50 was an irreversible
inhibitor. An o-sulfonate rather than a carboxylate must impede
hydrolysis of 14 by the enzyme. The sulfonates 53 may,
therefore, have interesting inhibitory activities.
The class A TEM-2 β-lactamase and the Streptomyces R61
and Actinomadura R39 DD-peptidases were also transiently
inhibited by 10, but more slowly than the P99 β-lactamase, and
the inhibition was weakened by the presence of 11. Presumably,
7 is not an effective inhibitor of these enzymes, although other
variants of 15 might be.
Variation of the phthalic anhydride moiety is also possible,
and thus, compounds 39−50 were screened in a manner similar
to that used for the combinations described above (Figure S5 of
the Supporting Information). Of the compounds tested, only
the substituted phthalic anhydrides 39 and 41 had significant
activity, alone and in the presence of 11. The nitro compound
40 was too unstable in aqueous solution to produce measurable
inhibition. The hydrophobic 39 was more effective in
combination with 11 than 10 itself. It seems likely that even
more effective inhibition (faster acylation and slower
deacylation) could be achieved from structure−activity studies
of a further variety of anhydride−nucleophile combinations.
Compound 6 behaved very much like 5, showing transient
enzyme inhibition, for example, when added to solutions
7067
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
Scheme 7
Scheme 8
Figure 6. Relative activity (estimated as t1/2, the time to half-
consumption of the substrate) of a series of nucleophiles (0.10 mM
each) added to 10 (0.1 mM) in the presence of the substrate CENTA
(50 μM) and the P99 β-lactamase (1.0 nM). The filled circle
represents the absence of both 10 and nucleophile, the filled square 10
alone, and the filled triangle 10 with the nucleophile 11. The register
of nucleophiles is as follows: 1−11, 16−26; 13−25, 27−38.
enzyme (t1/2 ∼ 1.25 min), which is almost certainly a covalent
phthaloyl-enzyme. Phthalic acid monoderivatives 15 may
represent new leads to β-lactamase inhibitors; the carboxylate
group of 7 was important for its activity, presumably providing
the negative charge found in most β-lactamase substrates and
inhibitors. In a second stage of reaction, the initially formed 7
cyclized to form phthalic anhydride 10, which is also a β-
lactamase inhibitor, forming the same transiently stable
(phthaloyl) intermediate that 7 did. Compound 6 was also a
transient inhibitor, presumably through similar chemistry.
Although 5 and 6 appear to be too spontaneously reactive
themselves to become practical inhibitors, it is possible that
suitable chemically deactivated analogues may have interesting
pro-drug properties. The two-stage activation of the new pro-
Figure 5. Turnover of CENTA (50 μM) by the P99 β-lactamase (1.0
◇
nM) in the presence of 10 (0.10 mM) alone ( ) and upon addition of
□
○
△
nucleophiles 17 ( ), 11 ( ), or 34 ( ), each at 0.10 mM.
Conclusions. Although the cyclic hydroxamates 3 and 4
were not β-lactamase inhibitors, their N-acyl derivatives, 5 and
6, respectively, had interesting inhibitory properties. Com-
pound 5 was not itself inhibitory, but upon spontaneous
reaction in aqueous solution, it yielded, first, the phthaloyl
monoderivative 7, which was inhibitory. Compound 7 reacted
with the P99 β-lactamase yielding a transiently inhibited
7068
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
(12) Pelto, R. B., and Pratt, R. F. (2008) Kinetics and mechanism of
inhibition of a serine β-lactamase by O-aryloxycarbonyl hydroxamates.
Biochemistry 47, 12037−12046.
(13) Tilvawala, R., and Pratt, R. F. (2013) Covalent inhibition of
serine β-lactamases by novel hydroxamic acid derivatives. Biochemistry
52, 3712−3720.
(14) Huttunen, K. M., Raunio, H., and Rautio, J. (2011) Prodrugs:
From serendipity to rational design. Pharmacol. Rev. 63, 750−771.
(15) Ferres, H. (1983) Pro-drugs of β-lactam antibiotics. Drugs Today
19, 499−519.
(16) Bryskier, A. (2005) β-Lactam prodrugs. In Antimicrobial Agents
(Bryskier, A., Ed.) Chapter 11, American Society for Microbiology
Press, Washington, DC.
(17) Testa, B. (2009) Prodrugs: Bridging pharmacodynamics/
pharmacokinetic gaps. Curr. Opin. Chem. Biol. 13, 338−344.
(18) Zinner, G., Ruthe, M., and Vollrath, R. (1971) Synthesis of 1,4-
dioxo-3,4-dihydro-benzo[d]-1,2-oxazin (O,N)-phthaloyl-hydroxyl-
amine. Synthesis, 148−149.
(19) Guimond, N., Gorelsky, S. I., and Fagnou, K. (2011)
Rhodium(III)-catalyzed heterocycle synthesis using an internal
oxidant: Improved reactivity and mechanistic studies. J. Am. Chem.
Soc. 133, 9449−9457.
inhibitors leads to two different inhibitors which, presumably,
would have different enzyme specificity. Incorporation of a β-
lactam into Nu of 15 would be a nice final touch.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic details for the preparation of compounds 8 and 9 and
Figures S1−S5 illustrating the products and kinetics of
spontaneous reactions of 5 and 7. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: rpratt@wesleyan.edu. Telephone: (860) 685-2629.
Fax: (860) 685-2211.
Funding
This research was supported by National Institutes of Health
Grant AI-17986 (R.F.P.) and by Wesleyan University.
Notes
(20) Kuzmic, P. (1996) Program DYNAFIT for the analysis of
enzyme kinetic data: Application to HIV proteinase. Anal. Biochem.
237, 260−273.
The authors declare no competing financial interest.
ABBREVIATIONS
■
(21) Bebrone, C., Moali, C., Mahy, F., Rival, S., Docquier, J. D.,
BSA, bovine serum albumin; CENTA, 7β-[(thien-2-yl)-
acetamido]-3-[(4-nitro-3-carboxyphenylthio)methyl]-3-ce-
phem-4-carboxylic acid; DCM, dichloromethane; DMF,
dimethylformamide; DMSO, dimethyl sulfoxide; ESMS,
electrospray mass spectrometry; FTIR, Fourier transform
infrared; MOPS, 3-(N-morpholino)propanesulfonic acid; mp,
melting point; NMR, nuclear magnetic resonance; TEA,
triethylamine; TFA, trifluoroacetic acid.
̀
Rossolini, G. M., Fastrez, J., Pratt, R. F., Frere, J.-M., and Galleni, M.
(2001) CENTA as a chromogenic substrate for studying β-lactamases.
Antimicrob. Agents Chemother. 45, 1868−1871.
(22) O’Callaghan, C. H., Morris, A., Kirby, S. M., and Shingler, A. H.
(1972) Novel method of detecting β-lactamases by using a
chromogenic cephalosporin substrate. Antimicrob. Agents Chemother.
1, 283−288.
(23) Cabaret, D., Adediran, S. A., Pratt, R. F., and Wakselman, M.
(2003) New substrates for β-lactam-recognizing enzymes: Aryl
malonamates. Biochemistry 42, 6719−6725.
(24) Kunitake, T., Okahata, Y., Ando, R., and Hirotsu, S. (1976)
Multifunctional hydrolytic catalysis. V. The reaction of o-substituted
benzohydroxamic acids and p-nitrophenyl acetate. Bull. Chem. Soc. Jpn.
49, 2547−2552.
REFERENCES
■
(1) Drawz, S. M., and Bonomo, R. A. (2010) Three decades of β-
lactamase inhibitors. Clin. Microbiol. Rev. 23, 160−201.
(2) Worthington, R. J., and Melander, C. (2013) Overcoming
resistance to β-lactamase inhibitors. J. Org. Chem. 78, 4207−4213.
(3) Shlaes, D. M. (2013) New β-lactam−β-lactamase inhibitor
combinations in clinical development. Ann. N.Y. Acad. Sci. 1277, 105−
114.
(25) Khan, M. N. (1996) Kinetic evidence for the occurrence of
stable intermediates in the cleavage of N-hydroxyphthalimide under
N-methylhydroxylamine buffer. Indian J. Chem. 358, 1275−1281.
́
(26) Andres, G. O., Granados, A. M., and de Ross, R. H. (2001)
(4) Livermore, D. M., and Mushtaq, S. (2013) Activity of biopenem
(RFX 2003) combined with the boronate β-lactamase inhibitor RPX
7009 against carbapenem-resistant Enterobacteriaceae. J. Antibiot.
Chemother. 68, 1825−1831.
Kinetic study of the hydrolysis of phthalic anhydride and aryl hydrogen
phthalates. J. Org. Chem. 66, 7653−7657.
̀
(27) Matagne, A., Dubus, A., Galleni, M., and Frere, J.-M. (1999)
The β-lactamase cycle: A tale of selective pressure and bacterial
ingenuity. Nat. Prod. Rep. 16, 1−19.
(5) Pratt, R. F. (2012) β-Lactamase inhibitors: Non β-lactams. In β-
̀
Lactamases (Frere, J.-M., Ed.) Chapter 10, Nova Science, New York.
(28) Fink, A. L., and Page, M. I. (2012) The mechanisms of catalysis
(6) Ghuysen, J.-M. (1991) Serine β-lactamases and penicillin-binding
proteins. Annu. Rev. Microbiol. 45, 37−67.
by β-lactamases. In β-Lactamases (Frer
Science, New York.
(29) Buynak, J. D. (2012) β-Lactams as inhibitors of β-lactamases. In
β-Lactamases (Frere, J.-M., Ed.) Chapter 10, Nova Science, New York.
̀
e, J.-M., Ed.) Chapter 7, Nova
(7) Bebrone, C. (2007) Metallo-β-lactamases (classification, activity,
genetic organization, structure, zinc coordination) and their super-
family. Biochem. Pharmacol. 74, 1686−1701.
̀
(30) Page, M. G. P. (1993) The kinetics of non-stoichiometric bursts
of β-lactam hydrolysis catalyzed by class C β-lactamases. Biochem. J.
295, 295−304.
(8) Ganta, S. R., Perumal, S., Pagadala, S. R. R., Samuelson, D.,
Spencer, J., Pratt, R. F., and Buynak, J. D. (2009) Approaches to the
simultaneous inactivation of metallo- and serine β-lactamases. Bioorg.
Med. Chem. Lett. 19, 1618−1622.
(9) Pratt, R. F. (1992) β-Lactamase inhibition. In Chemistry of β-
Lactams (Page, M. I., Ed.) Chapter 7, Chapman and Hall, London.
(10) Papp-Wallace, K. M., Bethel, C. R., Distler, A. M., Kasuboski, C.,
Taracila, M., and Bonomo, R. A. (2010) Inhibitor resistance in the
KPC-2 β-lactamase, a preeminent property of this β-lactamase.
Antimicrob. Agents Chemother. 54, 890−897.
(11) Wyrembak, P. N., Babaoglu, K., Pelto, R. B., Shoichet, B. K., and
Pratt, R. F. (2007) O-Aryloxycarbonyl hydroxamates: New β-
lactamase inhibitors that cross-link the active site. J. Am. Chem. Soc.
129, 9548−9549.
(31) Dubus, A., Normark, S., Kania, M., and Page, M. G. P. (1995)
Role of asparagine 152 in catalysis of β-lactam hydrolysis by Escherichia
coli. AMP C β-lactamase studied by site-directed mutagenesis.
Biochemistry 34, 7757−7764.
(32) Kumar, S., Adediran, S. A., Nukaga, M., and Pratt, R. F. (2004)
Kinetics of turnover of cefotaxime by the Enterobacter cloacae P99 and
GCI β-lactamases: Two free enzyme forms of the P99 β-lactamase
detected by a combination of pre- and post-steady state kinetics.
Biochemistry 43, 2664−2672.
(33) Bauer, L., and Exner, O. (1974) The chemistry of hydroxamic
acids and N-hydroxyimides. Angew. Chem., Int. Ed. 13, 376−384.
7069
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070

Biochemistry
Article
(34) Govardhan, C. P., and Pratt, R. F. (1987) Kinetics and
mechanism of the serine β-lactamase catalyzed hydrolysis of
depsipeptides. Biochemistry 26, 3385−3395.
(35) Varetto, L., de Meester, F., Monniae, D., Marchard-Brynaert, J.,
̀
Dive, G., Jacob, F., and Frere, J.-M. (1991) The importance of the
negative charge of β-lactam compounds in the interactions with active
site serine DD-peptidases and β-lactamases. Biochem. J. 278, 801−807.
(36) Ahn, Y.-M., and Pratt, R. F. (2004) Kinetic and structural
consequences of the leaving group in substrates of a class C β-
lactamase. Bioorg. Med. Chem. 12, 1537−1542.
(37) Beadle, B. M., Trehan, J., Focia, P. C., and Shoichet, B. K.
(2002) Structural milestones in the reaction pathway of an amide
hydrolase: Substrate, acyl, and product complexes of cephalothin with
AmpC β-lactamase. Structure 10, 413−424.
(38) Majumdar, S., and Pratt, R. F. (2009) Inhibition of class A and
C β-lactamases by diaroyl phosphates. Biochemistry 48, 8285−8292.
(39) Pazhanisamy, S., Govardhan, C. G., and Pratt, R. F. (1989) β-
Lactamase-catalyzed aminolysis of depsipeptides: Amine specificity
and steady state kinetics. Biochemistry 28, 6863−6870.
(40) Kraut, J. (1977) Serine proteases: Structure and mechanism of
catalysis. Annu. Rev. Biochem. 46, 331−358.
7070
dx.doi.org/10.1021/bi400873r | Biochemistry 2013, 52, 7060−7070