Exploring sequence requirements for C3/C4 carboxylate recognition in the Pseudomonas aeruginosa cephalosporinase: Insights into plasticity of the AmpC β-lactamase

Article abstract of DOI:10.1002/pro.612

In Pseudomonas aeruginosa, the chromosomally encoded class C cephalosporinase (AmpC β-lactamase) is often responsible for high-level resistance to β-lactam antibiotics. Despite years of study of these important β-lactamases, knowledge regarding how amino acid sequence dictates function of the AmpC Pseudomonas-derived cephalosporinase (PDC) remains scarce. Insights into structure-function relationships are crucial to the design of both β-lactams and high-affinity inhibitors. In order to understand how PDC recognizes the C₃/C₄ carboxylate of β-lactams, we first examined a molecular model of a P. aeruginosa AmpC β-lactamase, PDC-3, in complex with a boronate inhibitor that possesses a side chain that mimics the thiazolidine/dihydrothiazine ring and the C ₃/C₄ carboxylate characteristic of β-lactam substrates. We next tested the hypothesis generated by our model, i.e. that more than one amino acid residue is involved in recognition of the C ₃/C₄ β-lactam carboxylate, and engineered alanine variants at three putative carboxylate binding amino acids. Antimicrobial susceptibility testing showed that the PDC-3 β-lactamase maintains a high level of activity despite the substitution of C₃/C₄ β-lactam carboxylate recognition residues. Enzyme kinetics were determined for a panel of nine penicillin and cephalosporin analog boronates synthesized as active site probes of the PDC-3 enzyme and the Arg349Ala variant. Our examination of the PDC-3 active site revealed that more than one residue could serve to interact with the C₃/C₄ carboxylate of the β-lactam. This functional versatility has implications for novel drug design, protein evolution, and resistance profile of this enzyme.

Full text of DOI:10.1002/pro.612

Exploring sequence requirements for
C3/C4 carboxylate recognition in the
Pseudomonas aeruginosa
cephalosporinase: Insights into plasticity
of the AmpC b-lactamase
Sarah M. Drawz,1 Magdalena Taracila,2 Emilia Caselli,3 Fabio Prati,3
and Robert A. Bonomo2,4,5,6
*
1Department of Pathology, Case Western Reserve University School of Medicine, Cleveland, OH
2Department of Medicine, Case Western Reserve University School of Medicine, Cleveland, OH
3Department of Chemistry, University of Modena and Reggio Emilia, Modena 41100, Italy
4Louis Stokes Cleveland Department of Veterans Affairs Medical Center, Cleveland, OH
5Department of Molecular Biology and Microbiology, Case Western Reserve University School of Medicine, Cleveland, OH
6Department of Pharmacology, Case Western Reserve University School of Medicine, Cleveland, OH
Received 2 December 2010; Revised 11 February 2011; Accepted 14 February 2011
DOI: 10.1002/pro.612
Published online 14 March 2011 proteinscience.org
Abstract: In Pseudomonas aeruginosa, the chromosomally encoded class C cephalosporinase
(AmpC b-lactamase) is often responsible for high-level resistance to b-lactam antibiotics. Despite
years of study of these important b-lactamases, knowledge regarding how amino acid sequence
dictates function of the AmpC Pseudomonas-derived cephalosporinase (PDC) remains scarce.
Insights into structure-function relationships are crucial to the design of both b-lactams and high-
affinity inhibitors. In order to understand how PDC recognizes the C3/C4 carboxylate of b-lactams,
we first examined a molecular model of a P. aeruginosa AmpC b-lactamase, PDC-3, in complex
with a boronate inhibitor that possesses a side chain that mimics the thiazolidine/dihydrothiazine
ring and the C3/C4 carboxylate characteristic of b-lactam substrates. We next tested the
hypothesis generated by our model, i.e. that more than one amino acid residue is involved in
recognition of the C3/C4 b-lactam carboxylate, and engineered alanine variants at three putative
carboxylate binding amino acids. Antimicrobial susceptibility testing showed that the PDC-3 b-
lactamase maintains a high level of activity despite the substitution of C3/C4 b-lactam carboxylate
recognition residues. Enzyme kinetics were determined for a panel of nine penicillin and
cephalosporin analog boronates synthesized as active site probes of the PDC-3 enzyme and the
Arg349Ala variant. Our examination of the PDC-3 active site revealed that more than one residue
could serve to interact with the C3/C4 carboxylate of the b-lactam. This functional versatility has
implications for novel drug design, protein evolution, and resistance profile of this enzyme.
Keywords: b-lactamase inhibitor resistance; AmpC cephalosporinases; Pseudomonas aeruginosa
Grant sponsor: National Institutes of Health; Grant number: 1R01 A1063517-01; Grant sponsor: Veterans Integrated Service
Network (VISN) 10 Geriatric Research, Education, and Clinical Center (GRECC); Grant sponsor: Case Medical Scientist Training
Program (NIH); Grant number: T32 GM07250; Grant sponsor: Fondazione Cassa di Risparmio di Modena.
*Correspondence to: Robert A. Bonomo, MD, Infectious Diseases Section, Louis Stokes Cleveland Department of Veterans Affairs
Medical Center, 10701 East Blvd., Cleveland, OH 44106. E-mail: robert.bonomo@med.va.gov
C
V
Published by Wiley-Blackwell. 2011 The Protein Society
PROTEIN SCIENCE 2011 VOL 20:941—958 941

studied by Rodriguez-Martinez et al. for its
‘‘extended-spectrum’’ properties.17 We examined a
particular region of the PDC-3 enzyme active site,
which has been implicated in the binding of the C3/
C4 carboxylate common to penicillins, cephalospo-
rins, and sulfone inhibitors. Previous studies in class
A b-lactamases demonstrated that recognition of this
C3/C4 carboxylate contributes directly to substrate
and inhibitor affinity through hydrogen bonding
with areas of positive charge such as Arg244,
Arg220, and Arg276.18–25
Introduction
Pseudomonas aeruginosa is a Gram-negative bacil-
lus responsible for life-threatening nosocomial infec-
tions including pneumonia, blood stream, intra-ab-
dominal, wound, eye, ear, and urinary tract
infections.1 The organism, ‘‘an opportunistic patho-
gen,’’ is often isolated from patients with multiple ill-
nesses, in-dwelling catheters, burns, and surgical
devices.2 P. aeruginosa is also one of the most com-
monly isolated Gram-negative bacteria from patients
in intensive care units in the United States, and the
incidence has been rising during the past several
decades.3 Mortality rates up to 60% are reported for
P. aeruginosa infections, particularly in patients
with immune compromise or underlying comorbid-
ities.1,4 Treatment guidelines often advise empiric
administration of a b-lactam in combination with a
fluoroquinolone or aminoglycoside, but despite this
‘‘combination approach,’’ outcomes are poor.5 The
prevalence of P. aeruginosa strains resistant to three
or more antimicrobial agents can range from 3 to
50%.6,7 Consequently, treatment of P. aeruginosa
infections is challenging, and the scarcity of new
agents being developed and released into the clinic
stresses the need to use currently available agents
in a judicious manner.8
While the amino acids that participate in C3/C4
carboxylate binding are described for class A, the
molecular and biochemical correlates of this recogni-
tion are less well understood in class C enzymes.
Comparative studies of the Enterobacter cloacae P99
and Escherichia coli AmpC enzymes suggest that
the sites such as Xaa343, Asn346, Arg349, or
Thr316 interact with the C3/C4 carboxylate of the b-
lactam [Fig. 1(b)].26–33 Our examination of the PDC-
3 b-lactamase with a boronate b-lactamase inhibitor
docked in the active site lead us to hypothesize that
AmpCs may have also have a region or ‘‘pocket,’’
which serves to guide the substrate’s catalytically
competent conformation.33 A molecular model of
PDC-3 in complex with a boronate derivative pos-
sessing the C3/C4 carboxylate informed our hypothe-
sis that multiple possible ‘‘recognition residues’’ are
present in the active site. We tested this notion by
creating single amino acid substitutions in PDC-3 at
three active site residues. The effects of these substi-
tutions were examined by b-lactam substrate and b-
lactamase inhibitor susceptibility. Detailed steady-
state kinetic assays were performed for the PDC-3
Arg349Ala variant, which showed the largest pheno-
type difference in the antimicrobial susceptibility. To
understand further the mechanistic basis for
changes in kinetic parameters, we next synthesized
a panel of boronate compounds, designed as sub-
strate analogs, allowing us to probe the PDC-3
active site topology. Our comparative analysis
reveals the complex relationship between sequence,
structure, catalysis and inhibition in class C cepha-
The major antibiotic resistance determinants in
P. aeruginosa strains are b-lactamase enzymes (e.g.,
class C of AmpC cephalosporinases and class B ver-
ona integron-encoded metallo-b-lactamase (VIM) and
b-lactamase active on imipenem (IMP) type metallo-
b-lactamases), antibiotic efflux pumps (e.g., MexAB-
OprM), and impermeability of the outer mem-
brane.9,10 Production of chromosomal AmpCs medi-
ates resistance to b-lactams, and both genetic muta-
tions and induction from certain b-lactams can
significantly increase the expression of these b-lac-
tam-hydrolyzing enzymes.11,12 Protecting the activity
of b-lactams by inhibiting b-lactamases is a strategic
approach for preserving our current antimicrobial
armamentarium. However, the commercially avail-
able b-lactamase inhibitors are less effective at inac-
tivating b-lactamases from class C than class A.13–15
Despite the immense clinical significance of this re-
sistance determinant, relatively little research has
examined the structure–function relationships of the
‘‘Pseudomonal’’ AmpC (here referred to as Pseudomo-
nas-derived cephalosporinase, PDC) for either sub-
strates or inhibitors [Fig. 1(a)].16,17 A better under-
standing of the molecular details of catalysis and
inhibition of this AmpC can help lead design of more
effective inhibitors.
losporinases and, more importantly, PDC is
a
uniquely plastic enzyme with a region of positive
charge that drives the recognition of the C3/C4
carboxylate.34–36
Results and Discussion
C3/C4 Carboxylate binding in PDC-3
cephalosporinase
Using boronate substrate analogs, we explored
the role of specific residues in the catalytic and in-
hibitory profile of the P. aeruginosa AmpC. Our tar-
get b-lactamase, Pseudomonas-derived cephalospori-
nase-3 (PDC-3), has a single amino acid change from
the reference strain P. aeruginosa PAO1, and was
To better understand the interactions between the
PDC-3 b-lactamase and both b-lactam substrates
and inhibitors, we first created a molecular model of
the cephalothin analog boronate 5 covalently bound
to Ser64 of PDC-3 (Fig. 2). This boronate’s meta-
carboxylate group is designed to mimic in distance
942 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

Figure 1. (a) Molecular representation of the Pseudomonas-derived cephalosporinase model created to study potential
substrate carboxylate binding residues. Ser64 is shown in yellow. (b) Protein sequence alignment of E. coli AmpC and P.
aeruginosa PDC-3. Highly conserved class C motifs are underlined including S-V-S-K (first serine is Ser64), Y-X-N, and
K-T-G. Amino acids Asn343, Asn346, and Arg349 with potential roles in C3/C4 substrate carboxylate binding are also
indicated. Numbers listed are based on the consensus numbering of the mature E. coli AmpC protein, which differs slightly
from the PDC-3 protein.17
and stereochemistry the conserved C3/C4 carboxylate
on b-lactam substrates and b-lactamase inhibitors.
Thus, the placement of this group offers insight into
possible carboxylate binding residues in PDC-3.
After energy minimization and equilibration,
the PDC-3 model showed us that the meta-carboxy-
late interacts with Asn343 and Arg349, both of
which are implicated in the binding the C3/C4 car-
boxylate in other AmpC structures.26,32,33 Asn346,
another residue responsible for carboxylate recogni-
informed our hypotheses regarding the locations in
which C3/C4 carboxylate could bind PDC-3.
Antimicrobial susceptibility testing
Table I summarizes the minimum inhibitory concen-
tration (MIC) values of E. coli DH10B cells express-
ing PDC-3 alanine variants at residues 343, 346,
and 349. Most significantly, the level of resistance to
ampicillin, cephalothin, cephaloridine, and cefotax-
ime conferred by PDC-3 variants in E. coli was
lower than the WT enzyme. In contrast, the resist-
ance levels did not decrease more than one dilution
for any of the PDC-3 variants for cefoxitin,
˚
tion, is also found within ꢀ5 A of the carboxylate,
and potentially plays a role in the overall electro-
static environment of this region. This analysis
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 943

Figure 2. Cephalothin analog 5 docked into the active site
of our PDC-3 model to gain insight into the binding
relationships of the meta-carboxyphenyl group. In this
model, the boronate interacts with Arg349 and Asn343
(represented by dashed green lines), and has possible
longer range interactions with Asn346. Active site residues
Ser64, Ser318, Thr316, and Tyr150 are labeled.
ceftazidime, and cefepime or the carbapenems (imi-
penem and meropenem). We note that PDC-3 var-
iants show susceptibilities similar to WT for the
substrates with C3 carboxylates, piperacillin and
the carbapenems. The exception is ampicillin, which
also has a C3 carboxylate but may rely more on rec-
ognition of this moiety with its relatively simple R1
and R2 side chains.
Substrate kinetics: PDC-3 Arg349Ala retains
catalytic efficiency for cephalosporins
PDC-3 and PDC-3 Arg349Ala b-lactamases were puri-
fied using preparative isoelectric focusing, size exclu-
sion, and cation exchange chromatography. The molecu-
lar weight of the b-lactamases was confirmed with
Electrospray Ionization-Mass Spectrometry (ESI-MS).
The observed species were within experimental error of
the predicted b-lactamase molecular weights (PDC-3 ¼
40,775 amu, PDC-3 Arg349Ala ¼ 40,693 amu, all meas-
urements have an error of 63 amu). Both b-lactamases
include an additional Met residue on the N-terminus, as
the codon for this residue is introduced with the NdeI
restriction site used to clone the gene into the vector.
Table II summarizes our kinetic data character-
izing PDC-3 and the Arg349Ala variant b-lacta-
mases, which we selected because of its compara-
tively large phenotypic effect on MICs. In general,
the catalytic profile of PDC-3 for the penicillin and
cephalosporins tested is consistent with the kinetic
behavior described for other AmpC cephalospori-
nases. Namely, kcat values are in the order of 102
944 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

Table II. PDC-3 and PDC-3 Arg349Ala b-Lactamase Substrate Kineticsa
PDC-3
PDC-3 Arg349Ala
Substrates
Km or Ki (lM)
kcat (sꢁ1
)
kcat/Km (lMꢁ1sꢁ1
)
Km or Ki (lM)
kcat (sꢁ1
)
kcat/Km (lMꢁ1sꢁ1
)
Ampicillin
Nitrocefin
Cephalothin
Cephaloridine
Ceftazidime
Cefotaxime
0.31 6 0.02
20.7 6 1.8
24.3 6 3.1
33.4 6 0.7
51.7 6 2.1
0.26 6 0.03
0.96 6 0.02
22.0 6 2.6
1.8 6 0.1
19.4 6 1.6
36.2 6 7.6
2.0 6 0.2
453 6 45
409 6 41
130 6 13
21.9 6 2.9
16.8 6 2.7
3.9 6 0.4
310 6 9
7 6 1
27 6 1
14.1 6 1.7
4.0 6 0.5
1.4 6 0.1
a
The standard deviation of each value is reported.
sꢁ1 for narrow-spectrum cephalosporins such as
Arg349 appears to be contributing to the turnover of
the cephalothin and cephaloridine.
nitrocefin, cephalothin, and cephaloridine, and kcat
/
Km is typically >1 lMꢁ1 sꢁ1 17,37–39
.
In contrast, for
Structural analysis of the reaction pathway of
cephalothin with the E. coli AmpC reveals that the
position of cephalothin’s C4 carboxylate moves more
ampicillin and the extended-spectrum cephalospo-
rins, hydrolysis is poorly measurable, suggesting kcat
is <0.5 sꢁ1. However, these b-lactams maintain low
lM Kis in competition assays with nitrocefin.
Comparison of the PDC-3 Arg349Ala variant to
PDC-3 revealed very similar Km values for nitrocefin
and cephaloridine. Interestingly, the apparent affin-
ity of PDC-3 Arg349Ala for cephalothin is more than
10-fold higher than PDC-3 (although it should be
noted that the relative affinity measurement for
cephalothin was based on a competition assay, as
the very low kcat value hindered a direct measure-
ment of Km).
˚
than 7 A from its original position in the pre-cova-
lent substrate-enzyme complex (Fig. 3).32 In the
covalent cephalothin-AmpC complex, the hydrolytic
water is observed in the position where the C4 car-
boxylate was found in the precovalent complex. If
PDC-3 and cephalothin engage in a similar reaction
pathway as the E. coli AmpC with cephalothin, the
decreased kcat values observed for the Arg349Ala
substitution could reflect a modification of the local
active site topology; the introduction of 349Ala may
disrupt the sphere of positive charge around the C4
carboxylate and impair sequestration of the hydro-
lytic water for the narrow-spectrum cephalosporins.
The enzyme–ligand interactions of PDC-3
Arg349Ala were also altered for ampicillin, ceftazi-
dime, and cefotaxime. For both ampicillin and cefo-
taxime, the apparent Ki of the Arg349Ala variant
was increased, 3- and 10-fold, respectively. Such an
The most significant difference between the ki-
netic parameters of PDC-3 and PDC-3 Arg349Ala is
the decrease in kcat observed for nitrocefin, cephalo-
thin, and cephaloridine. The reduction in kcat is larg-
est for cephalothin and cephaloridine, 56-fold and 5-
fold lower than PDC-3, respectively. Thus, instead of
a predominant effect on the initial substrate affinity,
Figure 3. Representations of the reaction pathway for cephalothin and the E. coli AmpC: (a) the pre-covalent substrate-
enzyme complex; (b) acyl-enzyme complex; and (c) the hydrolyzed product in the active site.32 The position of the C4
carboxylate is circled in dashed lines to show the movement of the substituent over the reaction course. In panel (b), the
active site residues are labeled and the red represents the approximate location of the presumed deacylation water. Images
created with Discovery Studio Visualizer 2.5 using the deposited crystal structure coordinates 1KVL and 1KVM.
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 945

Table III. Inhibitor Kinetics for PDC-3 and PDC-3 Arg349Alaa
Inhibitors
PDC-3, Ki (lM)
PDC-3 Arg349Ala, Ki (lM)
PDC-3/PDC-3 Arg349Ala, DDGb (kcal molꢁ1
)
Commercially available inhibitors
Clavulanate
Tazobactam
Aztreonam
>10,000
24.1 6 4.1
1.31 6 0.13
>10,000
138 6 10
15.6 6 1.1
NA
þ1.03
þ1.47
Boronate substrate analogs
Compound R
Compound 1
Compound 2
Compound 3
Compound 4
Compound 5
Compound 6
Compound 7
Compound 8
Compound 9
37 6 6
0.62 6 0.10
0.42 6 0.03
55 6 4
3.1 6 0.4
þ0.23
þ0.95
þ0.36
ꢁ0.09
þ0.11
þ0.78
þ0.83
þ0.41
þ2.34
þ0.86
0.77 6 0.03
0.41 6 0.03
0.012 6 0.002
0.015 6 0.003
0.89 6 0.10
0.34 6 0.03
0.21 6 0.06
0.17 6 0.04
0.48 6 0.05
0.010 6 0.001
0.004 6 0.001
0.22 6 0.03
0.17 6 0.04
0.004 6 0.001
0.040 6 0.006
All boronate compounds were incubated with the enzyme for 5 min before addition of nitrocefin.a
a
The standard deviation of each Ki value is reported.
Differential (Gibbs) free energy of binding for PDC-3 Arg349Ala compared to PDC-3, calculated at 298 K. Values are cal-
b
culated using DDG ¼ ꢁRT ln [(Ki PDC-3)/(Ki PDC-3 Arg349Ala)]. Positive values indicate decreased affinity.
increase in apparent Ki is consistent with both the
increased susceptibility of E. coli cells expressing
the Arg349Ala variant and the hypothesis that
Arg349 offers a stabilizing hydrogen bond to the con-
served C3/C4 carboxylate. The small decrease in the
apparent Ki of ceftazidime for PDC-3 Arg349Ala var-
iant is also consistent with the lack of change in
MIC susceptibility to this substrate.
effects of the substitution on the potential carboxy-
late-binding region, we tested boronic acid compounds
with penicillin- and cephalosporin-like R1 and R2
groups. We compared the inhibition data obtained
with these boronate substrate analogs to that for the
currently available class A b-lactamase inhibitors
clavulanate and tazobactam and the monobactam
aztreonam. The results are presented in Table III.
How do cefotaxime and ceftazidime interact dif-
ferently with PDC-3? Upon inspection of the chemi-
cal structures, we see that the R1 side chains of
these two extended-spectrum cephalosporins are
similar with the replacement of a methyl group with
a dimethyl acetate group in ceftazidime. The compa-
rable apparent affinity of ceftazidime for the WT
enzyme and Arg349Ala variant suggests that the di-
methyl acetate group is engaged in productive
enzyme–ligand interactions, an observation consist-
ent with previous work on ceftazidime hydrolysis in
the AmpCs from other organisms.40 The crystal
structure of the E. coli AmpC complexed with cefta-
zidime shows that both the R1 dimethyl acetate
group and the C4 carboxylate are oriented towards
residues 343, 346, and 349. Furthermore, the elec-
tron density of the dimethyl acetate group suggests
this part of the ligand may be particularly flexible
and exist in multiple conformations.40 The flexibility
of ceftazidime in the acyl-enzyme form may allow for
residues 343, 346, and 349 to interact with both R1
and C4 carboxylate groups during the reaction coor-
dinates. Thus, loss of the Arg residue has minimal
impact on the recognition of ceftazidime.
Commercially available inhibitors
Of the inhibitors examined, the commercially avail-
able b-lactamase inhibitors (clavulanate, sulbactam,
and tazobactam) demonstrate the highest apparent
Kis. We could not achieve reduction of nitrocefin hy-
drolysis for PDC-3 and the Arg349Ala variant with
up to 10 mM of the oxapenam clavulanate, illustrat-
ing the enzymes are not inhibited by the compound.
The sulfone tazobactam has a 5-fold higher Ki for
the Arg349Ala variant as compared to the WT b-lac-
tamase. While this decreased apparent affinity sug-
gests a role of the Arg349 residue in binding recogni-
tion, the kinact values for the two PDC-3 enzymes
are very similar (0.082 6 0.005 sꢁ1 for WT and
0.071 6 0.003 sꢁ1 for PDC-3 Arg349Ala).
We next tested aztreonam, which behaves as a
‘‘slow substrate,’’ or ‘‘transient inhibitor’’ of class C b-
lactamases.41 In competition assays with PDC-3 and
nitrocefin, aztreonam demonstrated a Ki in the range
of effective b-lactamase inhibition (1.31 6 0.13 lM).
In contrast, the Arg349Ala variant had a slightly
increased Ki for this inhibitor (15.6 6 1.1 lM), but
the kinact value of 0.183 6 0.011 sꢁ1 was within error
of the WT kinact, 0.180 6 0.028 sꢁ1 (kinact values not
shown in Table III). This data is consistent with the
small effect observed in the MIC substrate testing of
aztreonam. Monobactams like aztreonam have an N-
linked SOꢁ3 group that is oriented toward the pocket
b-Lactamase inhibitor kinetics: Arg349 as part of
a carboxylate recognition region
To explore further the mechanism underlying the
changes in substrate turnover of PDC-3 and the
946 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

of the active site containing the 343, 346, and 349
residues in the Citrobacter freundii AmpC/aztreonam
complex crystal structure.42 The elimination of the
N-linked substituent following hydrolysis of the acyl-
enzyme’s sulfonamide bond was described for mono-
bactams with other class C and A b-lactamases.37,43
Assuming that aztreonam adopts a similar conforma-
tion in PDC-3, elimination of the SOꢁ3 group follow-
ing acylation could decrease the impact of the
Arg349Ala substitution, explaining the similar kinact
values between the variant and WT enzymes.
PDC-3 b-lactamase, but this value increases more
than five times for the Arg349Ala variant. Binding of
the penicillin C3 carboxylate group may be particu-
larly dependent on features of the carboxylate recogni-
tion region in this cephalosporinase. The nafcillin ana-
log (compound 2) shows Kis of a similar magnitude to
the cefotaxime and cephalothin boronates without an
R2 group, and displays less than twofold difference
between PDC-3 and PDC-3 Arg349Ala. When com-
pared with the ampicillin boronate, the two rings com-
prising the nafcillin R1 may allow for additional
enzyme–ligand interactions that help offset substitu-
tion of the Arg349 residue. In fact, cloxacillin and oxa-
cillin are described as inhibitors of AmpC b-lacta-
mases, and both possess two rings in their R1 group.46
Finally, cephalothin boronates with three differ-
ent R2 side chains were tested (compounds 4, 5, and
6). Compound 4, which lacks the meta-carboxylate of
compound 5, has a very similar (low nM Ki) for both
b-lactamases. Compound 5 has a lower apparent af-
finity for PDC-3 Arg349Ala, but the Ki value is still
very close to that for PDC-3. We can compare the
binding energy of the boronates for the two enzymes
by using Ki as an equilibrium constant in the Gibbs
free energy equation:28
Boronates
To complete our analysis, we next screened boronic
acid compounds that were synthesized to resemble
the R1 substituents on either penicillin or cephalop-
sporin b-lactams. The relative Kis of the boronates
provide insight into the ‘‘molecular preferences’’ or
‘‘substrate specificities’’ of the PDC-3 active site. Chi-
ral boronate compounds incorporate an ‘‘R2’’ side
chain, an additional functional group that resembles
the thiazolidine and dihydrothiazine rings of penicil-
lins and cephalopsorins, respectively, as well as the
conserved C3/C4 b-lactam carboxylate. These reversi-
ble inhibitors form a covalent bond with Ser64 of
class C b-lactamases, and the measured Kis are inter-
preted relative to a reference compound (compound
R) which lacks these side chains.44,45 We note that
the boronate’s ‘‘R2’’ group is not equivalent to the R2
group found on the corresponding b-lactam, for exam-
ple, the pyridinium ring found on ceftazidime. While
we cannot probe the role of the b-lactam R2 group
with the boronates, crystal structures of complexed
AmpCs show minimal interactions between the ceph-
alosporin’s R2 group and the enzyme before acylation
and elimination of the group following acylation.32,40
The boronate compounds in the series without an
R2 group (compounds 3, 7, 8) display Ki values that
are at least two orders of magnitude lower than the
reference compound for both PDC-3 and PDC-3
Arg349Ala (Ki range ¼ 0.004 6 0.001 lM to 0.48 6
0.05 lM). Boronates 3 and 7 demonstrate Kis that
were similar for both the WT and variant b-lacta-
mases. The ceftazidime R1 analog, compound 8, dis-
played the lowest Ki of the panel without R2 groups
for the WT and variant enzymes. However, this Ki was
52-fold higher for the Arg349Ala variant than PDC-3.
This result is somewhat unexpected as the compound
lacks the R2 group, which is designed to introduce
additional high-affinity substrate substituents.
Therefore, retention of the positively charged Arg in
this region must contribute to the Ki of this boronate.
The mechanism of boronate recognition likely differs
from the recognition of the same side chain as part of
the substrate (where the Ki of ceftazidime is very simi-
lar for PDC-3 and the Arg349Ala variant).
DDG ¼ ꢁRT ln ½ðKi PDC-3Þ=ðKi PDC-3 Arg349AlaÞꢂ
This calculation reveals a 0.78 kcal molꢁ1 differ-
ence between the energy associated with binding com-
pound 5 in PDC-3 as compared to PDC-3 Arg349Ala,
which is lower than expected for a hydrogen bond (ꢀ 2
kcal molꢁ1). Thus, the energy from the interaction
observed between the carboxylate and Arg349 in the
model of PDC-3 may be acquired by the hydrogen
bond with Asn343, or otherwise compensated by a
local rearrangement of the residues; Asn346 may
move into closer approximation with the meta-carbox-
ylate in the Arg349Ala variant.
The cephalothin analog boronate, compound 6, dif-
fers from compound 5 by the addition of a carbon on
the linker between the boronate group and the R2 side
chain. Both the PDC-3 and Arg349Ala AmpCs show a
significant decrease in affinity for compound 6, sug-
gesting that the extended linker introduces less favor-
able interactions in the P. aeruginosa b-lactamases.
Finally, compound 9 bearing both the ceftazi-
dime R1 side chain and an R2 meta-carboxyphenyl
shows a fourfold affinity loss for the Arg349Ala vari-
ant of PDC-3 b-lactamase. This loss contrasts the 52-
fold drop observed for compound 8 possessing just
the ceftazidime R1 side chain. Loss of the Arg349 res-
idue is mitigated by the incorporation of the R2 group
with additional enzyme binding groups.
Susceptibility testing with boronate compounds
Low nM Kis in kinetic assays do not necessarily
translate to effective enzyme inhibition; multiple
In the series of boronates with R2 groups, the
ampicillin analog (compound 1) has sub lM Ki for the
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 947

Table IV. Contribution of Boronate Analogs 5 and 8 to MIC Values (lg mLꢁ1) of P. aeruginosa 18SH and E. coli
DH10B Expressing blaPDC-3 or blaPDC-3 Variants at Possible Carboxylate Binding Sites
Isolates
Cefotaxime
Cefotaxime/5
Cefotaxime/8
P. aeruginosa 18SH
E. coli DH10B
E. coli blaPDC-3
>1024
<0.125
8
>64
<0.125
2
>64
0.125
0.5
E. coli blaPDC-3 Asn343Ala
E. coli blaPDC-3 Asn346Ala
E. coli blaPDC-3 Arg349Ala
4
4
0.5
1
1
0.25
0.25
0.25
0.125
Boronates tested at a constant concentration of 4 lg mLꢁ1
.
pharmacologic, biochemical, and microbiological
characteristics affect the ability of a compound to
restore susceptibility to a partner b-lactam in vivo.47
We determined the cephalosporin susceptibility of E.
coli expressing blaPDC-3 and the blaPDC-3 variants
with the addition of either boronate compound 5 or
8. The cephalothin analog 5 was selected because of
its low Ki for both PDC-3 and PDC-3 Arg349Ala; in
contrast, the ceftazidime boronate 8 was selected
because of the 52-fold increase in Ki for PDC-3
Arg349Ala.
At 4 lg mLꢁ1, both boronates lowered the MICs
for cefotaxime in E. coli blaPDC-3 expressing cells
(Table IV). For the cephalothin analog 5, the cefotax-
ime susceptibility was increased by two dilutions for
PDC-3, the Asn343Ala, and the Asn346Ala variants,
and one dilution for the Arg349Ala variant. Thus,
the similar Ki value of compound 5 for PDC-3 and
PDC-3 Arg349Ala was reflected in whole cell assays.
For the ceftazidime analog 8, cefotaxime susceptibil-
ities were improved by four dilutions for PDC-3, the
Asn343Ala, and the Asn346Ala variants, but only
two dilutions for the Arg349Ala variant. The
decreased MIC effect for the Arg349Ala variant may
result from compound 8’s 52-fold higher Ki for this
variant as compared to WT. However, the Ki of com-
pound 8 for PDC-3 was equal to that for compound 5
(4 nM), yet a two-dilution difference exists in the
inhibitors’ abilities to lower cefotaxime MICs to E.
mases often reveals active site locations with high-
binding affinity for the C3/C4 carboxylate but
attempts to identify a recognition region in class C
b-lactamases have been more elusive.24,30,48,49 We
have performed a careful analysis of a molecular
model of the PDC-3 protein in complex with a boro-
nate inhibitor bearing a side chain that mimics the
thiazolidine/dihydrothiazine ring and the C3/C4 car-
boxylate characteristic of b-lactam substrates. Based
on this analysis, we engineered PDC-3 variants at
residues that appeared important in the recognition
of the carboxylate moiety.
From our susceptibility testing and kinetic char-
acterizations, we identify properties of the P. aerugi-
nosa AmpC that suggest a significant degree of
enzyme plasticity. First, our studies suggest that more
than one C3/C4 carboxylate binding residue may exist
in PDC-3. Herein, we examined the effects of three
possible binding residues and observed decreases in
the MICs of these variants for several specific sub-
strates. While the magnitude of the MIC decrease
was greatest for the Arg349Ala variant, Ala substitu-
tions at Asn343 and Asn346 also have an impact on
susceptibility. This data argue for more of a binding
‘‘region’’ rather than a binding ‘‘site,’’ supported by
both the Asn and Arg residues. Loss of the positive
charge at 349 may lead to the most dramatic change
in enzyme architecture and electrostatics, but clearly
343 and 346 are also implicated in the topology of the
pocket. Furthermore, the PDC-3 Arg349Ala b-lacta-
mase maintained substantial hydrolytic activity for
the tested substrates. Despite reductions in kcat for
cephalothin and cephaloridine, the apparent affinity
remains high, preserving a robust catalytic efficiency
coli expressing blaPDC-3
.
Overall, these MIC data display some parallels
with the Ki values, but also underscores the complex-
ities of whole cell assays and microbiological proper-
ties such as periplasmic b-lactamase concentration of
enzyme and synergy with the b-lactam partner.47
>1 lMꢁ1 ꢁ1. The decrease in ampicillin and cefotax-
s
ime MICs for the Arg349Ala variant is accompanied
by a reduction in apparent Kis, but again the Kis
remain in the low lM range.
Conclusions
Our data represent one of the first in-depth exami-
nations of the structure–function relationships of
the P. aeruginosa AmpC b-lactamase, an enzyme of
tremendous clinical importance. We chose to focus
on exploring the mechanism in which the C3/C4 car-
boxylate binds in the active site. The ready recogni-
tion of this substituent by the four different b-lacta-
mase classes provides a solid argument that these
enzymes are adapted and evolved to identify this
functionality. Previous work with class A b-lacta-
The homology-based protein-modeling server
used to generate the PDC-3 representation identified
a P. fluorescens class C b-lactamase to serve as the
highest similarity PDB template (at the time the
modeling was performed, the P. aeruginosa AmpC
structure was not yet reported by Blizzard et al.16
Michaux et al.50 defined and described the structure
of this enzyme (which shares 76% sequence similar-
ity to PDC-3) and identified biochemical properties
948 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

that allow this cold-adapted enzyme to optimize cata-
lytic activity at low temperatures. Specifically, psy-
chrophilic enzymes enjoy increased structural flexibil-
ity at the price of decreased stability.51,52 P.
aeruginosa is not a psychrophilic organism, but the
similarity between the two class C b-lactamases helps
generate hypotheses about the biochemical properties
of PDC-3. In general, cold-adapted enzymes have
decreased numbers of ion pairs and hydrogen bonds,
low Arg content, and increased solvent accessibility to
the active site.50,51 If PDC-3 shares attributes with its
‘‘cold-loving’’ counterpart, we may expect it to be a
more ‘‘flexible’’ enzyme with binding regions comprised
of residues capable of serving multiple recognition
roles. These notions await testing of the thermal sta-
bility and conformational flexibility of PDC-3 through
circular dichroism and thermal denaturation
experiments.
mase AmpCs).17,53 This unique perspective introdu-
ces a new appreciation of how b-lactams are hydro-
lyzed in class C b-lactamases and helps explain, in
part, the broad substrate profile of these enzymes.
Such plasticity in substrate recognition has clear
structural, mechanistic, and evolutionary implica-
tions for novel drug design.54
Materials and Methods
Antibiotics and commercially available inhibitors
The chemical structures of the substrates and com-
mercially available inhibitors studied are shown in
Figure 4. Nitrocefin was purchased from BD Bio-
sciences (San Jose, CA), imipenem from U.S. Phar-
macopeia (Rockville, MD) and meropenem from
AstraZeneca Pharmaceuticals (Wilmington, DE).
The remaining antibiotics were purchased from
Sigma (St. Louis, MO). Lithium clavulanate was a
kind gift from GlaxoSmithKline (Surrey, United
Kingdom); tazobactam was obtained from Chem-
Impex International (Wood Dale, IL).
Second, by probing the PDC-3 Arg349Ala active
site with the boronate substrate analogs, we reveal
the complexities of ligand binding in this b-lacta-
mase. Compound 5 bearing the R1 of cephalothin
and an R2 group designed to mimic the conserved
cephalosporin dihydrothiazine ring and C4 carboxy-
late demonstrates very similar Ki values for PDC-3
and PDC-3 Arg349Ala. However, the incorporation
of an R2 group on the ceftazidime boronate (com-
pound 9) leads to a fourfold decrease in Ki for the
Arg349Ala variant when compared with WT enzyme.
Paradoxically, removing this R2 group from the cef-
tazidime boronate (compound 8) serves to increase
the affinity for the PDC-3 WT b-lactamase. Consid-
ered alongside the PDC-3 Arg349Ala’s perplexing
increase in apparent affinity for the substrate cefta-
zidime, these data suggest important differences in
how PDC-3 recognizes these b-lactams and their bor-
onate analogs. b-Lactam C3/C4 carboxylate regions
may be more substrate and inhibitor specific for this
class of enzymes, that is, the same functional groups
presented on different scaffolds can use unique areas
of the enzyme for recognition.
Synthesis of boronates
Figure 5 shows all tested boronates. All reactions
were performed under argon using oven-dried glass-
ware and dry solvents. Dry tetrahydrofuran (THF)
and diethyl ether were obtained by standard meth-
ods and distilled freshly under argon from sodium
benzophenone ketyl before use. The –100ꢃC bath was
prepared by addition of liquid nitrogen to a precooled
(–78ꢃC) mixture of ethanol/methanol (1:1). Reactions
were monitored by thin layer chromatography (TLC),
which were visualized by UV fluorescence and by
Hanessian’s cerium molybdate stain. Chromato-
graphic purification of the compounds was performed
on silica gel (particle size 0.05–0.20 mm). Melting
points were measured on a Bu¨chi 510 apparatus. Op-
tical rotations were recorded at þ20ꢃC on a Perkin-
Elmer 241 polarimeter and are expressed in 10ꢁ1 deg
cm2 gꢁ1
.
1H NMR and 13C NMR spectra were
In summary, our data lead us to assert that
multiple ‘‘ancillary’’ interactions contribute to sub-
strate and inhibitor binding in the P. aeruginosa
AmpC. The individual residues in the PDC-3 active
site exhibit a high degree of versatility, plasticity,
and flexibility for recognition of both substrates and
inhibitors. These notions of enzyme plasticity are
consistent with theses developed by Todd et al.35
that key functional residues can exist in multiple
active site locations, and this residue ‘‘hopping’’ may
be a important property of class C and other b-lacta-
mases. From an evolutionary and mechanistic stand-
point, these dynamic features may lead to more
opportunities for AmpC variants with increasing
resistance profiles and novel catalytic properties,
posing and enormous challenge to the medicinal
chemist (e.g., extended-spectrum and carbapene-
recorded on a Bruker DPX-200 or Avance-400 spec-
trometer; chemical shifts (d) are reported in ppm
downfield from tetramethylsilane (TMS) as internal
standard (s singlet, d doublet, t triplet, q quartet, m
multiplet, and br broad signal); coupling constants
(J) are given in Hz. Two-dimensional NMR techni-
ques (correlation spectroscopy (COSY), heteronuclear
multiple bond correlation (HMBC), heteronuclear
single quantum coherence (HSQC)) were used to aid
in the assignment of signals in 1H and 13C spectra.
Mass spectra were determined on a gas chromatogra-
phy HP 5890 associated with mass spectrometer de-
tector HP 5972 (EI, 70 eV) or on an agilent technolo-
gies liquid chormatography-mass spectrometry (LC-
MS)(n) Ion Trap 6310A (electrospray ionization
(ESI), 70 eV). Elemental analyses were performed on
a Carlo Erba Elemental Analyzer 1110. Compounds
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 949

Figure 4. Substrates and commercially available inhibitors used in this study.
3, 7, and 8 were synthesized as in 55, compounds 4
28
13C NMR (200 MHz, CDCl3): dC 19.3 (br, C),
24.0, 26.5, 27.1, 28.2, 28.6, 35.4, 38.2, 39.5, 51.3,
78.0, 80.7, 86.0, 126.2, 128.1, 129.9, 132.0, 133.2,
139.0, 166.0. EI-MS: m/z 370 (M, 18%), 314 (71%),
297 (59%), 135 (55%), 57 (100%).
and 5 as in and compound 9 as in (Miriagou et al.,
in revision, JBC). Compounds 1, 2, and 6 were syn-
thesized starting from bromomethaneboronate of
pinanaediol in a four-step synthesis described in
detail below (Fig. 6).
(1)-Pinanediol (1S)-1-chloro-2-(3-tert.butoxycar-
bonylphenyl)ethaneboronate (III). n-BuLi (2.44
mmol of a 1.6M solution in hexane) was added drop-
wise to a solution of CH2Cl2 (0.21 mL, 3.25 mmol) in
THF (4 mL) with stirring at ꢁ100ꢃC under argon.
At the end of the BuLi addition, a white microcrys-
talline precipitate (LiCHCl2) became evident. After
30 min, a solution of II (0.752 g, 2.03 mmol) in THF
(5 mL) was slowly added at the same temperature.
The white precipitate disappeared and the mixture
was allowed to reach RT overnight. The reaction
mixture was concentrated under vacuum, and the
brownish residue was diluted with light petroleum
(50 mL). The precipitate was filtered on an MgSO4
pad and washed with the same solvent (100 mL);
the solution was then concentrated to give pale yel-
low oil (0.7 g, 82%). [a]D ¼ þ12.0 (c 2.44, CHCl3), de
ꢄ 98%, used as such in the following step.
(1)-Pinanediol
(3-tert.butoxycarbonylphenyl)-
methaneboronate (II). n-BuLi (7.23 mmol of a
2.5M solution in hexane) was added dropwise with
stirring to a solution of tert-butyl-3-bromobenzoate
(1.77 mg, 6.88 mmol) in THF (2 mL) at ꢁ100ꢃC
under argon atmosphere. After 30 min, a solution of
bromomethaneboronate of (þ)-pinanaediol I (1.31 g,
4.82 mmol) in THF (10 mL) was added, and the
stirred mixture was allowed to reach room tempera-
ture (RT) overnight.56 The reaction mixture was con-
centrated under vacuum to give a thick yellow oil,
which was purified by chromatography (light petro-
leum/ethyl acetate 98:2), affording II (0.77 g, 43%
yield) as a yellow oil, [a]D ¼ þ8.7 (c 1.42, CHCl3).
1H NMR (400 MHz, CDCl3): dH 0.87 (3H, s),
1.11 (1H, d, J ¼ 11), 1.32 (3H, s), 1.43 (3H, s), 1.63
(9H, s), 1.84–2.39 (5H, m), 2.42 (2H, s, H), 4.32 (1H,
dd, J ¼ 8.8, 2.0), 7.33 (1H, t, J ¼ 7.7), 7.40 (1H, brd,
J ¼ 7.7), 7.80 (1H, brd, J ¼ 7.7), 7.86 (1H, brs).
1H NMR (200 MHz, CDCl3): dH 0.84 (3H, s),
1.03 (1H, d, J ¼ 10.8), 1.29 (3H, s), 1.37 (3H, s), 1.61
950 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

Figure 5. Boronate compounds used in this study.
(9H, s), 1.76–2.40 (5H, m), 3.14 (1H, dd, J ¼ 13.9,
chloroformiate (0.061 mL, 0.47 mmol) was added via
syringe to precooled solution of Boc-D-PheGly
8.4), 3.27 (1H, dd, J ¼ 13.9, 7.5), 3.67 (1H, dd,
J ¼ 8.4, 7.5), 4.37 (1H, dd, J ¼ 8.7, 2.0), 7.30–7.47
(2H, m), 7.85–7.90 (2H, m).
a
(0.118 g, 0.47 mmol) in THF (6 mL) with stirring at
–20ꢃC, followed by addition of N-Methylmorfoline
(NMM) (0.048 mL, 0.47 mmol). After 30 min at 0ꢃC,
the solution of silylamine was added dropwise via
syringe to the acylating mixture; the temperature
was then allowed to reach r.t. gradually, and the
resulting mixture stirred for additional 16 h. The
white suspension was then partitioned between
ethyl acetate (60 mL) and water (20 mL), and the or-
ganic phase was washed with saturated NaHCO3 so-
lution (30 mL). The organic phase was dried
(MgSO4), filtered and concentrated to give a grey
solid (0.305g), which was crystallized by diethylether
and n-hexane affording IVa (0.184 g, 62% yield) as a
white solid (mp 81–85ꢃC dec). [a]D ¼ –10.0 (c 1.05%;
CDCl3), de ꢄ 98%.
13C NMR (200 MHz, CDCl3): dC 25.3, 27.0, 28.4,
29.6, 29.7, 36.5, 39.6, 40.7, 41.5 (br, C-B), 44.1, 52.5,
80.0, 82.3, 88.3, 129.3, 129.6, 131.6, 133.5, 134.8,
139.9, 167.0.
(1)-Pinanediol (1R)-1-[(2-tert-butoxycarbonyla-
mino-2-phenyl)acetylamino]-2-[3(tert.Butoxycar-
bonyl)phenyl] ethane boronate (IVa). Lithium-
hexamethyldisilazane (1.6 mmol of a 1M solution in
THF) was added dropwise to a stirred solution of III
(0.618 g, 1.47 mmol) in THF (6 mL) at ꢁ78ꢃC under
argon atmosphere and the mixture was allowed to
warm gradually at RT overnight. The resulting solu-
tion was concentrated under vacuum, and light pe-
troleum (100 mL) was added. The precipitate was fil-
tered on a MgSO4 pad and washed with the same
solvent, and the solution concentrated to give pale
yellow oil (0.255 g, 32% yield ). The silylamine
(0.255 g, 0.47 mmol) was diluted with THF (6 mL)
and MeOH (0.224 mL of a 2.5M solution in THF)
was added dropwise at 0ꢃC. In the meantime, the
acylating mixture was prepared as follows: isobutyl-
1H NMR (400 MHz, CDCl3): dH 0.86 (3H, s),
1.09 (1H, d, J ¼ 10.9), 1.31 (3H, s), 1.34 (3H, s), 1.43
(9H, s), 1.62 (9H, s), 1.78–2.43 (5H, m), 2.87 (1H, dd,
J ¼ 13.9, 7.8), 2.98 (1H, dd, J ¼ 13.9, 5.3), 3.57 (1H,
m), 4.37 (1H, dd, J ¼ 8.9, 1.7), 5.11 (1H,s), 5.74 (1H,
b), 5.88 (1H, b), 6.90 (1H, d, J 6.6), 7.15 (1H, t, J ¼
7.6), 7.30–7.34 (5H, m), 7.77 (1H, s), 7.81 (1H, d, J ¼
7.6).
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 951

Figure 6. Four step synthesis for compounds 1, 2, and 6.
13C NMR (400 MHz, CDCl3): dC 24.0, 26.3, 27.1,
1H NMR (400 MHz, CDCl3): dH 0.95 (3H, s),
1.27 (3H, t, J ¼ 7.0), 1.35 (3H, si), 1.57 (3H, s), 1.62
(9H, s), 1.94–2.48 (5H, m), 3.05 (1H, dd, J ¼ 14.7,
11.3), 3.25 (1H, dd, J ¼ 14.7, 4.4), 3.50 (1H, m), 4.21
(2H, m), 4.46 (1H, dd, J ¼ 8.7, 2.2), 7.23 (1H, d, J ¼
9.2.), 7.39–7.46 (2H, m.), 7.55–7.61 (2H, m.), 7.72
(1H, s), 7.79 (1H, d, J ¼ 8.1.), 7.90 (1H, d, J ¼ 7.8),
7.95–7.99 (2H, m.), 8.64 (1H, d, J ¼ 8.9.).
28.2, 28.3, 28.4, 35.3, 36.5, 38.1, 38.8(br, C-B), 39.5,
51.3, 58.1, 78.2 79.9, 80.1, 86.4, 127.3, 127.5, 128.2,
128.3, 128.9, 129.9, 132.1, 133.3, 138.6, 139.1, 155.0,
165.7, 170.4.
LC-MS (ESI, Ion trap): m/z 655 (60%, [M þ
Na]þ), 633 (100%, [M þ H]þ, 533 (24%). MS/MS of
633 m/z: 533 (100%), 477 (5%), 381 (9%), 325 (2%).
Anal. Calcd. for C36H49BN2O7; C, 68.35; H, 7.81;
N, 4.43. Found: C, 68.06; H, 8.08; N, 4.69.
13C NMR (400 MHz, CDCl3): dC 14.7, 24.2, 26.6,
27.4, 28.2 (3C), 29.1, 36.5, 37.3, 38.3, 40.1, 43.8 (br,
C-B), 52.3, 65.6, 77.2, 81.0, 84.0, 113.3, 124.7, 125.6,
127.5, 127.1, 128.3, 128.5, 128.9, 129.7, 132.4 (2C),
133.1, 134.0, 134.5, 140.9, 156.2, 165.8, 170.9.
(1)-Pinanediol
carbonyl)amino]-2-[3-(tert.butoxycarbonyl)-
phenyl] ethaneboronate (IVb). Lithium-
(1R)-1-[(2-ethoxynaphtalene-1-
Anal. Calcd. for C36H44BNO6: C, 72.36; H, 7.42;
N, 2.34. Found: C, 72.52; H, 7.36; N, 2.18.
hexamethyldisilazane (1.71 mmol of a 1M solution in
THF) was added dropwise to a stirred solution of III
(0.622 g, 1.49 mmol) in THF (7 mL) at ꢁ78ꢃC under
argon atmosphere, and the mixture was allowed to
warm gradually at r.t. overnight. The resulting solu-
tion containing the silylamine was used as such for
the following step. A mixture of 2-ethoxynaphtoic
acid (0.387 g, 1.79 mmol) and 2-ethoxynaphtoylchlor-
ide (0.490 mL, 2.09 mmol) in THF (6 mL) was added
dropwise to the sylilamine solution at ꢁ78ꢃC and
allowed to reach RT overnight. After 16 h, the
brownish suspension was partitioned between ethyl
acetate (70 mL) and water (20mL); the organic
phase was washed with saturated NaHCO3 solution
(25 mL), and the aqueous phase extracted with
EtOAc (60, 40 mL). The organic phases were dried
(MgSO4), and concentrated to give a viscous yellow-
ish oil, which was purified by gradient chromatogra-
phy (light petroleum/ethyl acetate from 70/30)
affording IVb (0.123 g, 14% yield) as yellow viscous
oil. [a]D ¼ –17.1 (c 1.15%; CHCl3), de ꢄ 98%.
(1)-Pinanediol (1R)-1-[(2-thiophene-2-yl)acety-
lamino]-2-[3-(tert.butoxycarbonyl)phenyl]etha-
neboronate (IVc). Lithiumhexamethyldisilazane
(1.63 mmol of a 1M solution in THF) was added
dropwise to a stirred solution of III (0.568 g, 1.36
mmol) in THF (6 mL) at ꢁ78ꢃC under argon atmos-
phere and the mixture was allowed to warm gradu-
ally at RT overnight. The resulting solution contain-
ing the silylamine was used as such for the
following step. A mixture of thiopheneacetic acid
(0.232 g, 1.63 mmol) and thiopheneacetyl chloride
(0.201 mL, 1.63 mmol) in THF (3 mL) was added
dropwise to the sylilamine solution at ꢁ78ꢃC and
allowed to reach RT overnight. After 16 h, the
brownish suspension was partitioned between ethyl
acetate (70 mL) and water (25 mL); the organic
phase was washed with saturated NaHCO3 solution
(25 mL), and the aqueous phase extracted with
EtOAc (60, 40 mL). The organic phases were dried
952 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

(MgSO4), and concentrated to give viscous yellowish
oil, which was purified by gradient chromatography
(light petroleum/ethyl acetate from 1:1) affording
IVc (0.097 g, 14% yield) as white solid (mp 154ꢃC).
[a]D ¼ –25.5 (c 0.84%; CHCl3), de ꢄ 98%.
white solid (mp 79–81ꢃC), [a]D ¼ –55.8 (c 0.998%;
MeOD), de ꢄ 98%.
1H NMR (400 MHz, MeOD): dH 1.45 (3H, t, J ¼
7.0), 2.98 (1H, dd, J ¼ 14.5, 10.3), 3.27 (1H, dd, J ¼
14.5, 5.2), 3.75 (1H, dd, J ¼ 10.2, 5.3), 4.29–4.39
(2H, m), 7.42–7.50 (3H, m), 7.55–7.60 (2H, m.), 7.89
(1H, d, J ¼ 8.0.), 7.93 (1H, d, J ¼ 7.8), 8.02 (1H, br),
8.07 (1H, d, J ¼ 8.7.), 8.14 (1H, d, J ¼ 9.2.).
13C NMR (400 MHz, MeOD): d 13.7, 35.3, 51.4
(br, C-B), 65.5, 113.4, 123.3, 124.4, 126.2, 127.7,
128.2, 128.3, 128.5, 128.9, 129.7, 130.9, 131.4, 133.1,
135.4, 157.1, 167.4.
1H NMR (400 MHz, CDCl3): d 0.90 (3H, s), 1.30
(1H, d, J ¼ 10.7), 1.32 (3H, s), 1.44 (3H, s), 1.64 (9H,
s), 1.87–2.40 (5H, m), 2.82 (1H, dd, J ¼ 13.7, 10.0),
3.03 (2H, m), 3.87 (3H, m), 4.34 (1H, dd, J ¼ 8.7,
2.0), 6.04 (1H, br), 6.90 (1H, dd, J ¼ 3.5, 1.1), 6.95
(1H, dd, J ¼ 5.2, 3.5), 7.23 (1H, dd, J ¼ 5.2, 1.2),
7.24 (1H, t, J ¼ 1.5), 7.30 (1H, t, J ¼ 7.5), 7.78 (1H,
t, J ¼ 1.5), 7.84 (1H, dt, J ¼ 7.6, 1.5).
13C NMR (400 MHz, CDCl3): d 2.24, 26.4, 27.3,
28.3, 28.8, 34.7, 35.9, 37.0, 38.2, 39.8, 41.8 (C-B),
51.9, 77.3, 81.0, 84.9, 125.9, 127.5, 127.9, 128.4,
129.9, 132.3, 133.0, 134.0, 139.9, 165.7, 172.9.
EI-MS: m/z 523 (M, 4%), 371 (27%), 315 (19%),
208 (30%), 135 (61%), 97 (100%), 93 (38%), 57 (45%).
Anal. Calcd. for C29H38BNO5S: C, 66.54; H,
7.32; N, 2.68; S, 6.13. Found: C, 66.38; H, 7.42; N,
2.75; S, 6.09.
(1R)-1-(2-Thiophene-2-yl)-2-(3-carboxyphenyl)-
ethaneboronic acid (6). Compound
6
was
obtained (0.042 g, yield 70%) as white solid (mp 233ꢃC
dec.), [a]D ¼ –124.5 (c 0.22%; MeOH), de ꢄ 98%.
1H NMR (400 MHz, MeOD): d 2.62–2.73 (1H,
m), 2.89–2.99 (2H, m), 3.96 (2H, s), 6.04 (1H, s),
6.98–7.02 (2H, m), 7.40 (1H, d, J ¼ 4.0), 7.41 (1H, t,
J ¼ 7.6), 7.49 (1H, d, J ¼ 7.6), 7.88 (1H, d, J¼7.6),
7.92 (1H, s, H2).
13C NMR (400 MHz, MeOD): d 30.7, 36.4 (Cb),
47.7 (br, C-B), 125.3, 126.7, 127.2, 127.3, 128.2,
129.9, 130.6, 133.4, 133.6, 141.1, 168.7, 176.4.
Deprotection of the carboxy and boronic func-
tionalities: general procedure. Trifluoroacetic
acid (0.4 mL, 5.3 mmol) and catalytic anisol were
added to a solution of IVa–c (0.18 mmol) in anydrous
CH2Cl2 (4mL) and after 2 h, the mixture was concen-
trated under vacuum to give a yellowish solid which
was directly dissolved in CH3CN (4 mL). To this solu-
tion HCl (0.18 mmol of a 1M solution in H2O), phenyl-
boronic acid (0.022 g, 0.18 mmol) and n-exane (4 mL)
were sequentially added, and the resulting biphasic
solution was vigorously stirred. After 30 min, the n-
hexane solution (containing the pinanediol phenylbor-
onate) was removed, and n-hexane (4 mL) was added.
This last procedure was repeated several times until
a TLC analysis revealed no more presence of IVa–c.
The acetonitrile phase was concentrated affording the
desired compounds 1, 2, and 6 with overall yields of
5.2, 4.2, and 3.5%, respectively.
Molecular representations
The PDC-3 model was generated by the SWISS-
MODEL automated protein structure homology-
modeling server, available at http://swissmodel.
expasy.org.57 The determined PDC-3 protein
sequence was entered, and a model generated by the
software using the Pseudomonas fluorescens class C
b-lactamase template (Protein Data Bank entry
2QZ6).50 PDC-3 and the P. fluorescens protein share
76% sequence similarity; accurate high resolution
protein models can be generated from templates
with greater than 50% sequence similarity.58
The generated PDC-3 model was optimized by
energy minimization using Discovery Studio 2.1 soft-
ware (Accelrys, San Diego, CA). The minimization
was performed in several steps, using Steepest
Descent and Conjugate Gradient algorithms to reach
(1R)-1-(2-Amino-2-phenylacetylamino)-2-(3-car-
boxyphenyl)ethaneboronic acid (1). Compound
1 was obtained (0.048 g, yield 74%) as white solid (mp
226ꢃC dec.), [a]D ¼ –59.3 (c 0.96%; MeOD), de ꢄ 98%.
1H NMR (400 MHz, DMSO): dH 2.65 (1H, dd, J
¼ 13.7, 8.9), 2.81 (1H, dd, J ¼ 13.7, 4.3), 3.58 (1H,
m), 4.95 (1H,s), 6.83 (1H, d, J 7.4), 7.15 (1H, t, J ¼
7.6), 7.14–7.55 (5H, m), 7.63 (1H, d, J ¼ 7.6). 7.67
(1H, s), 8.28 (1H, b), 8.65 (2H, br), 12.82 (1H, br).
13C NMR (400 MHz, DMSO): dC 36.6, 41.5 (br,
C-B), 55.8, 127.2, 128.0, 128.3, 128.9, 129.2, 130.2,
130.9, 133.7, 134.7, 140.8, 167.0, 167.8.
the minimum convergence (0.02 kcal molꢁ1*A). The
˚
˚
protein was immersed in a water box, 7 A from any
face of the box, and the solvation model used was
with periodic boundary conditions. The force-field
parameters of CHARMm were used for minimization
and the Particle Mesh Ewald method addressed
long-range electrostatics. The bonds that involved
hydrogen atoms were constrained with the SHAKE
algorithm. Following equilibration, two separate 2
ps molecular dynamics simulations (heating/cooling
and production) at constant pressure and tempera-
ture (300 K) were carried out for the PDC-3 model.
The trajectories were analyzed, and the minimum
energy conformation was chosen.
(1R)-1-[(2-Ethoxynaphtalene-1-carbonyl)amino]-
2-(3-carboxyphenyl)ethaneboronic
acid
(2).
Compound 2 was obtained (0.064 g, yield 87%) as
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 953

The minimized and equilibrated PDC-3 model
was used for constructing the acylation complexes of
the PDC-3 b-lactamase and the chiral cephalothin
analog 5. The ligand structure was built using Dis-
covery Studio Fragment Builder tools. The
CHARMm force field was applied; the molecule was
solvated with periodic boundary conditions and
minimized using a Standard Dynamics Cascade pro-
tocol (one minimization using Steepest Descent algo-
rithm, followed by Adopted Basis Newton-Raphson
algorithm and three subsequent dynamics stages at
NVT and 300 K).
50 GCG GAG GGG CGG GGA AGC GCT CAT 30,
respectively).63
The PCR template was prepared from genomic
DNA isolated from overnight cultures of the P. aeru-
ginosa 18SH strain.39 The PDC upstream and PDC
downstream primers successfully amplified a 1400-
base pair amplicon from the P. aeruginosa 18SH
DNA. The amplicon was ligated into the pCR 2.1-
TOPO vector (Invitrogen, Carlsbad, CA) and trans-
formed into E. coli DH10B cells (Invitrogen). Follow-
ing sequence verification of the P. aeruginosa 18SH
blaPDC-3 gene, we performed a restriction enzyme
digest of the recombinant plasmid using the BamHI
and XbaI restriction enzymes (Promega, Madison,
WI). The digestion product was ligated to the
BamHI- and XbaI-cut pBC SK(ꢁ) vector and E. coli
DH10B cells were transformed with the recombinant
plasmid. The M13 Universal, M13 Reverse primers,
PDC upstream, PDC downstream, and primers
designed to the PAO1 active site were used to verify
the sequence of P. aeruginosa 18SH blaPDC-3 in pBC
SK(ꢁ).64
The minimized ligand was docked in the active
site of the enzyme using LibDock.59 The generated
conformations (30–40 poses) were visually inspected,
and the most favorable ones chosen based on mini-
mum energy. The complex between the ligand and
the enzyme was created, solvated, and energy mini-
mized. The acyl-enzyme complex was created by
making a bond with Ser64 of PDC-3, and the assem-
bly was further minimized using conjugate gradient
algorithm with periodic boundary conditions to 0.001
minimum derivatives. To reach the minimum equi-
librium, the complexes were equilibrated using mo-
lecular dynamic simulations.
This blaPDC-3 gene was also directionally sub-
cloned into the pET 24a(þ) vector (Novagen, Madi-
son, WI) for ‘‘large scale’’ protein expression from E.
coli BL21(DE3) RP CodonPlus cells (Stratragene, La
Jolla, CA). CodonPlus cells contain a pACYC plas-
mid with extra copies of the argU and proL tRNA
genes to help overcome codon bias encountered dur-
ing expression of P. aeruginosa protein in E. coli. We
designed PCR amplification primers to introduce
NdeI and BamH1 restriction sites into the beginning
of the gene sequence that coded for the mature pro-
tein (i.e., the b-lactamase leader sequence was not
included) and the end of the coding sequence,
respectively. PCR mutagenesis was performed on the
blaPDC-3 in the pCR 2.1-TOPO vector, and the suc-
cessful introduction of the restriction sites confirmed
by digesting the resultant plasmids with NdeI and
BamH1. The insert was purified from agarose gel,
ligated to a NdeI- and BamH1-digested plasmid prep
of the pET 24a(þ) vector, and E. coli DH10B cells
were transformed with the purified insert. Plasmids
were isolated and sequenced, and E. coli BL21(DE3)
RP CodonPlus cells were transformed with plasmids
of confirmed blaPDC-3 sequence according to the
manufacturer’s protocol.
Antimicrobial susceptibility (MICs)
Susceptibility profiles were determined by cation-
adjusted Mueller-Hinton agar dilution MICs accord-
ing to the Clinical and Laboratory Standard Insti-
tute (CLSI) standards.60 For the piperacillin/tazobac-
tam and cefotaxime/boronate combinations, the
substrate concentrations were varied, whereas the
inhibitors were tested at a constant concentration of
4 lg mLꢁ1
.
Cloning of blaPDC-3 into pBC SK(-) vector
The P. aeruginosa blaampC gene used in these studies
was cloned from the laboratory strain P. aeruginosa
18SH, a kind gift from Dr. M. G. P. Page (Basilea).
The AmpC b-lactamase from P. aeruginosa 18SH is
produced at constitutively high levels as its regula-
tion is ‘‘derepressed,’’ that is, the expression does not
increase with induction by b-lactams.61,62 The AmpC
from the PAO1 comparator strain has only one
amino acid that is different from the P. aeruginosa
18SH AmpC, and has been designated PDC-1 in the
´
´
classification system proposed by Rodrıguez-Martı-
nez and colleagues.17 By this nomenclature, the P.
aeruginosa 18SH AmpC is designated PDC-3.
Mutagenesis and sequencing
Site-directed mutagenesis primers were designed to
replace the blaPDC-3 wild-type (WT) amino acid
codon with the alanine codon at amino acid positions
343, 346, and 349 (based on sequence alignment
with P99, E. coli AmpC, and Acinteobacter-derived
cephalosporinase).65 Using the template blaPDC-3
To subclone the P. aeruginosa 18SH blaPDC-3
gene into the pBC SK(ꢁ) vector, we designed Poly-
merase Chain Reaction (PCR) primers to the con-
served regions of the blaPDC gene, including approxi-
mately 50 base pairs on either side of the start and
stop codons (PDC upstream and PDC downstream,
50 CGT CGT TTG CGG CAA ATC CTG CGC 30 and
gene in the pBC SK(ꢁ) vector, we used Stratagene’s
R
V
QuikChange Mutagenesis Kit
to introduce the
alternate codon. For the Arg349Ala variant,
954 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

mutagenesis was also performed in the pET 24a(þ)
vector using the same primers. DNA sequencing of
the blaPDC genes was performed to confirm introduc-
tion of the desired mutation.
observed previously with some of the chiral boro-
nates, all boronates were preincubated with enzyme
for 5 min in phosphate-buffered saline before initiat-
ing the reaction with the addition of sub-
strate.28,49,68–70
In these competition assays, initial velocities
b-Lactamase purification
(v0) can be represented by the following equation71
:
The PDC-3 and PDC-3 Arg349Ala b-lactamases were
purified from the E. coli BL21(DE3) RP CodonPlus
cells by induction with isopropyl b-D-thiogalactopyra-
noside. Briefly, 12 mL of an overnight culture grown
in lysogeny broth were used to inoculate 500 mL of
super optimal broth and grown at 37ꢃC for 2 h. We
added isopropyl b-D-thiogalactopyranoside at a final
concentration of 1 mM, and grew the cultures for an
additional 2.5 h. For all the P. aeruginosa and E.
coli cultures, the cells were pelleted, resuspended in
50 mM Tris (pH 7.4), and b-lactamase liberated with
lysozyme and ethylenediaminetetraacetic acid per
established methods.66 Accordingly, PDC-3 enzyme
was purified by preparative isoelectric focusing and
fast protein liquid chromotography with a Sephadex
Hi Load 16/60 column and a HiTrap High Perform-
ance sulfopropyl strong cation exchanger (Pharma-
cia, Uppsala, Sweden).67 The protein was quantified
by BSA assay and purity was assessed by 5% stack-
ing, 12% resolving SDS-PAGE.39
m0 ¼ ðVmax½SꢂÞ=fKmð1 þ I=KiÞ þ ½Sꢂg
To determine Ki, the initial velocities immedi-
ately after mixing (1/v0) were plotted as a function
of inhibitor concentration:
ð1=m0Þ ¼ Km=ðVmax½SꢂÞ ꢅ ð1 þ I=KiÞ þ 1=Vmax
Rearrangement of the equation shows:
ð1=m0Þ ¼ Km=ðVmax½SꢂÞ ꢅ ðI=KiÞ þ 1=Vmaxð1 þ Km=½SꢂÞ
where I is Ki obs, the equation is solved by setting 1/
v0 to 0.
Ki obs ¼ f1=Vmaxð1 þ Km=½SꢂÞg=ðKm=Vmax½SꢂKiÞ
Ki obs ¼ Kið½Sꢂ=Km þ 1Þ þ ðKm=Vmax½SꢂÞ
Ki ¼ Ki obs=ð1 þ ½Sꢂ=KmÞ
Kinetics
Steady-state kinetics were performed on an Agilent
8453 diode array spectrophotometer (Palo Alto, CA).
Each continuous assay was performed in 10 mM
phosphate-buffered saline at pH 7.4 at RT in a quartz
cuvette with a 1-cm pathlength. Measurements were
obtained using nitrocefin (De482 ¼ 17,400 Mꢁ1 cmꢁ1),
cephalothin (De262 ¼ ꢁ7660 Mꢁ1 cmꢁ1), and cephalor-
idine (De260 ¼ ꢁ10,200 Mꢁ1 cmꢁ1).
Thus, the Ki
is the concentration of I that
obs
reduces the velocity by 50%, and can be calculated
from the x-intercept times ꢁ1.72
For tazobactam and aztreonam, the first-order
rate constant for enzyme and inhibitor complex inac-
tivation, kinact, was obtained by monitoring the reac-
tion time courses in the presence of inhibitor. Fixed
concentrations of enzyme (10 nM) and nitrocefin
(100 lM) and increasing concentrations of inhibitor
were used in each assay. The kobs was determined
The kinetic parameters, Vmax and Km, were
obtained with non-linear least squares fit of the data
(Henri Michaelis-Menten equation) using Origin
R
V
7.5 (OriginLab, Northampton, MA):
using a nonlinear least squares fit of the data using
m ¼ ðVmax ꢅ ½SꢂÞ=ðKm þ ½SꢂÞ
R
V
Origin 7.5 :
As previously reported, direct hydrolysis of the
substrates ampicillin, ceftazidime, and cefotaxime
could not be measured at this time (in assays with
up to 5 lg of protein), and thus the apparent affinity
values were obtained in competition assays with
nitrocefin.37 Similarly, competition assays were
employed to measure the relative affinity of PDC-3
Arg349Ala variant for cephalothin, whereas kcat was
estimated from direct hydrolysis assays.
A ¼ A0 þ mf t þ ðm0 þ mf Þ½1 ꢁ expðꢁkobstÞꢂ=kobs
Here, A is absorbance, v0 (expressed in variation
of absorbance per unit time) is initial velocity, vf is
final velocity, and t is time. Each kobs was plotted
versus I and fit to determine kinact
.
Electrospray Ionization Mass Spectrometry
For the boronic acid inhibitors, Ki values were
calculated by measuring the initial velocity (0–10 s)
Mass spectrometry (MS) was performed to ascertain
the molecular weight of the mature PDC-3 and
PDC-3 Arg349Ala proteins expressed in E. coli
BL21(DE3) RP CodonPlus cells. We prepared 14 lM
of each protein in phosphate-buffered saline. The
sample was acidified by the addition of 0.1%
in the presence of
a constant concentration of
enzyme (3 nM) and increasing concentrations of the
inhibitors competed against the indicator substrate,
nitrocefin. Because of the time-dependent inhibition
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 955

10. Walsh TR (2010) Emerging carbapenemases: a global
trifluoroacetic acid and immediately desalted and
concentrated using a C18 ZipTip (Millipore, Bedford,
MA) according to the manufacturer’s protocol.
Samples were then placed on ice and analyzed
within 1 h.
perspective. Int
S8–S14.
J Antimicrob Agents 36(Suppl 3):
11. Hanson ND, Sanders CC (1999) Regulation of inducible
AmpC b-lactamase expression among Enterobacteria-
ceae. Curr Pharm Des 5:881–894.
Spectra of the intact proteins were generated on
a Q-STAR XL Quadrupole-Time-of-Flight mass spec-
trometer (Applied Biosystems, Framingham, MA)
equipped with a nanospray source. Experiments
were performed by diluting the protein sample with
50% acetonitrile/0.1% trifluoroacetic acid to a con-
centration of 10 lM. This protein solution was then
infused at a rate of 0.5 ll minꢁ1 and the data were
collected for 2 min. Spectra were deconvoluted using
the Analyst program (Applied Biosystems). We
assign an error of 63 amu to each measurement.
12. Juan C, Gutierrez O, Oliver A, Ayestaran JI, Borrell N,
Perez JL (2005) Contribution of clonal dissemination
and selection of mutants during therapy to Pseudomo-
nas aeruginosa antimicrobial resistance in an intensive
care unit setting. Clin Microbiol Infect 11:887–892.
13. Bush K, Macalintal C, Rasmussen BA, Lee VJ, Yang Y
(1993) Kinetic interactions of tazobactam with b-lacta-
mases from all major structural classes. Antimicrob
Agents Chemother 37:851–858.
14. Monnaie D, Frere JM (1993) Interaction of clavulanate
with class C b-lactamases. FEBS Lett 334:269–271.
15. Kazmierczak A, Cordin X, Duez JM, Siebor E, Pechinot
A, Sirot J (1990) Differences between clavulanic acid
and sulbactam in induction and inhibition of cephalo-
sporinases in enterobacteria. J Int Med Res 18(Suppl
4):67D–77D.
Acknowledgments
The authors thank Arne Rietsch and Krisztina Papp-
Wallace for helpful suggestions concerning the purifi-
cation of the P. aeruginosa proteins and manuscript
preparation.
16. Blizzard TA, Chen H, Kim S, Wu J, Young K, Park YW,
Ogawa A, Raghoobar S, Painter RE, Hairston N, Lee
SH, Misura A, Felcetto T, Fitzgerald P, Sharma N, Lu
J, Ha S, Hickey E, Hermes J, Hammond ML (2010)
Side chain SAR of bicyclic b-lactamase inhibitors
(BLIs). 1. Discovery of a class C BLI for combination
with imipinem. Bioorg Med Chem Lett 20:918–921.
17. Rodriguez-Martinez JM, Poirel L, Nordmann P (2009)
Extended-spectrum cephalosporinases in Pseudomonas
aeruginosa. Antimicrob Agents Chemother 53:
1766–1771.
References
1. Page MG, Heim J (2009) Prospects for the next anti-
Pseudomonas drug. Curr Opin Pharmacol 9:558–565.
2. Hidron AI, Edwards JR, Patel J, Horan TC, Sievert
DM, Pollock DA, Fridkin SK (2008) NHSN annual
update: antimicrobial-resistant pathogens associated
with healthcare-associated infections: annual summary
of data reported to the National Healthcare Safety Net-
work at the Centers for Disease Control and Preven-
tion, 2006–2007. Infect Control Hosp Epidemiol 29:
996–1011.
18. Jacob-Dubuisson F, Lamotte-Brasseur J, Dideberg O,
Joris B, Frere JM (1991) Arginine 220 is a critical resi-
due for the catalytic mechanism of the Streptomyces
albus G b-lactamase. Protein Eng 4:811–819.
19. Knox JR, Moews PC (1991) b-lactamase of Bacillus
licheniformis 749/C. Refinement at 2 A resolution and
analysis of hydration. J Mol Biol 220:435–455.
3. Gaynes R, Edwards JR (2005) Overview of nosocomial
infections caused by Gram-negative bacilli. Clin Infect
Dis 41:848–854.
20. Imtiaz U, Billings EM, Knox JR, Manavathu EK,
Lerner SA, Mobashery S (1993) Inactivation of class A
b-lactamases by clavulanic acid: the role of Arginine
244 in a proposed noncerted sequence of events. J Am
Chem Soc 115:4435–4442.
4. Vidal F, Mensa J, Almela M, Martinez JA, Marco F,
Casals C, Gatell JM, Soriano E, Jimenez de Anta MT
(1996) Epidemiology and outcome of Pseudomonas aer-
uginosa bacteremia, with special emphasis on the influ-
ence of antibiotic treatment. Analysis of 189 episodes.
Arch Intern Med 156:2121–2126.
21. Zafaralla G, Manavathu EK, Lerner SA, Mobashery S
(1992) Elucidation of the role of Arginine-244 in the
turnover processes of class A b-lactamases. Biochemis-
try 31:3847–3852.
22. Thomson JM, Distler AM, Bonomo RA (2007) Overcom-
ing resistance to b-lactamase inhibitors: comparing sul-
bactam to novel inhibitors against clavulanate
resistant SHV enzymes with substitutions at Ambler
position 244. Biochemistry 46:11361–11368.
5. Rahal JJ (2006) Novel antibiotic combinations against
infections with almost completely resistant Pseudomo-
nas aeruginosa and Acinetobacter species. Clin Infect
Dis 43(Suppl 2):S95–S99.
6. Goossens H (2003) Susceptibility of multi-drug-resist-
ant Pseudomonas aeruginosa in intensive care units:
results from the European MYSTIC study group. Clin
Microbiol Infect 9:980–983.
23. Moews PC, Knox JR, Dideberg O, Charlier P, Frere JM
(1990) b-lactamase of Bacillus licheniformis 749/C at 2
A resolution. Proteins 7:156–171.
24. Marciano DC, Brown NG, Palzkill T (2009) Analysis of
the plasticity of location of the Arg244 positive charge
within the active site of the TEM-1 b-lactamase. Pro-
tein Sci 18:2080–2089.
7. Tam VH, Chang KT, Abdelraouf K, Brioso CG, Ameka
M, McCaskey LA, Weston JS, Caeiro JP, Garey KW
(2010) Prevalence, mechanism and susceptibility of
multidrug resistant bloodstream isolates of Pseudomo-
nas aeruginosa. Antimicrob Agents Chemother 54:
1160–1164.
25. Perez-Llarena FJ, Cartelle M, Mallo S, Beceiro A,
Perez A, Villanueva R, Romero A, Bonnet R, Bou G
(2008) Structure-function studies of arginine at position
276 in CTX-M b-lactamases. J Antimicrob Chemother
61:792–797.
8. Rice LB (2009) The clinical consequences of antimicro-
bial resistance. Curr Opin Microbiol 12:476–481.
9. Livermore DM (2002) Multiple mechanisms of antimi-
crobial resistance in Pseudomonas aeruginosa: our
worst nightmare? Clin Infect Dis 34:634–640.
26. Beadle BM, Shoichet BK (2002) Structural basis for
imipenem inhibition of class C b-lactamases. Antimi-
crob Agents Chemother 46:3978–3980.
956 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC

27. Lobkovsky E, Billings EM, Moews PC, Rahil J, Pratt
RF, Knox JR (1994) Crystallographic structure of a
phosphonate derivative of the Enterobacter cloacae P99
cephalosporinase: mechanistic interpretation of a b-lac-
tamase transition-state analog. Biochemistry 33:
6762–6772.
bacter freundii indicates a mechanism for b-lactam hy-
drolysis. Nature 343:284–288.
43. Mourey L, Miyashita K, Swaren P, Bulychev A,
Samama JP, Mobashery
S (1998) Inhibition of the
NMC-A b-lactamase by a penicillanic acid derivative
and the structural bases for the increase in substrate
profile of this antibiotic resistance enzyme. J Am Chem
Soc 120:9382–9383.
28. Morandi F, Caselli E, Morandi S, Focia PJ, Blazquez J,
Shoichet BK, Prati F (2003) Nanomolar inhibitors of
AmpC b-lactamase. J Am Chem Soc 125:685–695.
29. Dubus A, Wilkin JM, Raquet X, Normark S, Frere JM
(1994) Catalytic mechanism of active-site serine b-lac-
tamases: role of the conserved hydroxy group of the
Lys-Thr(Ser)-Gly triad. Biochem J 301(Pt 2):485–494.
30. Zhang Z, Yu Y, Musser JM, Palzkill T (2001) Amino
acid sequence determinants of extended spectrum ceph-
alosporin hydrolysis by the class C P99 b-lactamase. J
Biol Chem 276:46568–46574.
44. Chen CC, Rahil J, Pratt RF, Herzberg O (1993) Struc-
ture of a phosphonate-inhibited b-lactamase. An analog
of the tetrahedral transition state/intermediate of b-lac-
tam hydrolysis. J Mol Biol 234:165–178.
45. Crompton IE, Cuthbert BK, Lowe G, Waley SG (1988)
b-lactamase inhibitors. The inhibition of serine b-lacta-
mases by specific boronic acids. Biochem
453–459.
J 251:
46. Bush K, Jacoby GA, Medeiros AA (1995) A functional
classification scheme for b-lactamases and its correla-
tion with molecular structure. Antimicrob Agents Che-
mother 39:1211–1233.
31. Patera A, Blaszczak LC, Shoichet B (2000) Crystal
structures of substrate and inhibitor complexes with
AmpC b-lactamase: possible implications for substrate-
assisted catalysis. J Am Chem Soc 122:10504–10512.
32. Beadle BM, Trehan I, Focia PJ, Shoichet BK (2002)
Structural milestones in the reaction pathway of an
amide hydrolase: substrate, acyl, and product com-
plexes of cephalothin with AmpC b-lactamase. Struc-
ture 10:413–424.
47. Drawz SM, Bonomo RA (2010) Three decades of b-lac-
tamase inhibitors. Clin Microbiol Rev 23:160–201.
48. Lobkovsky E, Moews PC, Liu H, Zhao H, Frere JM,
Knox JR (1993) Evolution of an enzyme activity: crys-
tallographic structure at 2-A resolution of cephalospori-
nase from the ampC gene of Enterobacter cloacae P99
and comparison with a class A penicillinase. Proc Natl
Acad Sci USA 90:11257–11261.
33. Powers RA, Shoichet BK (2002) Structure-based
approach for binding site identification on AmpC b-lac-
tamase. J Med Chem 45:3222–3234.
49. Morandi S, Morandi F, Caselli E, Shoichet BK, Prati F
(2008) Structure-based optimization of cephalothin-ana-
logue boronic acids as b-lactamase inhibitors. Bioorg
Med Chem 16:1195–1205.
34. Tomatis PE, Fabiane SM, Simona F, Carloni P, Sutton
BJ, Vila AJ (2008) Adaptive protein evolution grants
organismal fitness by improving catalysis and flexibil-
ity. Proc Natl Acad Sci USA 105:20605–20610.
50. Michaux C, Massant J, Kerff F, Frere JM, Docquier
JD, Vandenberghe I, Samyn B, Pierrard A, Feller G,
Charlier P, Van Beeumen J, Wouters J (2008) Crystal
structure of a cold-adapted class C b-lactamase. FEBS
J 275:1687–1697.
35. Todd AE, Orengo CA, Thornton JM (2002) Plasticity of
enzyme active sites. Trends Biochem Sci 27:419–426.
36. Brown NG, Pennington JM, Huang W, Ayvaz T, Palz-
kill T (2010) Multiple global suppressors of protein sta-
bility defects facilitate the evolution of extended-
spectrum TEM b-lactamases. J Mol Biol 404:832–846.
37. Endimiani A, Doi Y, Bethel CR, Taracila M, Adams-
Haduch JM, O’Keefe A, Hujer AM, Paterson DL, Skal-
weit MJ, Page MG, Drawz SM, Bonomo RA (2010)
Enhancing resistance to cephalosporins in class C b-
lactamases: impact of Gly214Glu in CMY-2. Biochemis-
try 9:1014–1023.
51. Feller G, Gerday C (2003) Psychrophilic enzymes: hot
topics in cold adaptation. Nat Rev Microbiol 1:200–208.
52. Wang X, Minasov G, Shoichet BK (2002) Evolution of
an antibiotic resistance enzyme constrained by stability
and activity trade-offs. J Mol Biol 320:85–95.
53. Kim JY, Jung HI, An YJ, Lee JH, Kim SJ, Jeong SH,
Lee KJ, Suh PG, Lee HS, Lee SH, Cha SS (2006)
Structural basis for the extended substrate spectrum of
CMY-10, a plasmid-encoded class C b-lactamase. Mol
Microbiol 60:907–916.
38. Galleni M, Amicosante G, Frere JM (1988) A survey of
the kinetic parameters of class C b-lactamases. Cepha-
losporins and other b-lactam compounds. Biochem J
255:123–129.
54. Khersonsky O, Tawfik DS (2010) Enzyme promiscuity:
a mechanistic and evolutionary perspective. Annu Rev
Biochem 79:471–505.
39. Hujer KM, Hamza NS, Hujer AM, Perez F, Helfand
MS, Bethel CR, Thomson JM, Anderson VE, Barlow M,
Rice LB, Tenover FC, Bonomo RA (2005) Identification
of a new allelic variant of the Acinetobacter baumannii
55. Caselli E, Powers RA, Blasczcak LC, Wu CY, Prati F,
Shoichet BK (2001) Energetic, structural, and antimi-
crobial analyses of b-lactam side chain recognition by
b-lactamases. Chem Biol 8:17–31.
56. Davoli P, Fava R, Spaggiari A, Morandi S, Prati F
(2005) Enantioselective total synthesis of (-)-microcar-
palide. Tetrahedron 61:4427–4436.
cephalosporinase, ADC-7 b-lactamase: defining
a
unique family of class C enzymes. Antimicrob Agents
Chemother 49:2941–2948.
40. Powers RA, Caselli E, Focia PJ, Prati F, Shoichet BK
(2001) Structures of ceftazidime and its transition-state
analogue in complex with AmpC b-lactamase: implica-
tions for resistance mutations and inhibitor design.
Biochemistry 40:9207–9214.
57. Arnold K, Bordoli L, Kopp J, Schwede T (2006) The
SWISS-MODEL workspace: a web-based environment
for protein structure homology modelling. Bioinfor-
matics 22:195–201.
58. Kryshtafovych A, Fidelis K (2009) Protein structure
prediction and model quality assessment. Drug Discov
Today 14:386–393.
41. Bush K, Freudenberger JS, Sykes RB (1982) Interac-
tion of azthreonam and related monobactams with b-
lactamases from gram-negative bacteria. Antimicrob
Agents Chemother 22:414–420.
59. Diller DJ, Merz KM Jr (2001) High throughput docking
for library design and library prioritization. Proteins
43:113–124.
42. Oefner C, D’Arcy A, Daly JJ, Gubernator K, Charnas
RL, Heinze I, Hubschwerlen C, Winkler FK (1990)
Refined crystal structure of b-lactamase from Citro-
60. Clinical and Laboratory Standard Institute (2006)
Methods for dilution antimicrobial susceptibility tests
Drawz et al.
PROTEIN SCIENCE VOL 20:941—958 957

for bacteria that grow aerobically. Approved Standard,
CLSI document M7-A7, 7th ed., Wayne, PA.
61. Berks M, Redhead K, Abraham EP (1982) Isolation
position Gly238 in the SHV-1 b-lactamase: exploring
sequence requirements for resistance to penicillins and
cephalosporins. Antimicrob Agents Chemother 46:
3971–3977.
and properties of an inducible and
a constitutive
b-lactamase from Pseudomonas aeruginosa.
Microbiol 128:155–159.
62. Flett F, Curtis NA, Richmond MH (1976) Mutant
J
Gen
67. Pattanaik P, Bethel CR, Hujer AM, Hujer KM, Distler
AM, Taracila M, Anderson VE, Fritsche TR, Jones RN,
Pagadala SR, van den Akker F, Buynak JD, Bonomo
RA (2009) Strategic design of an effective b-lactamase
inhibitor: LN-1–255, A 6-alkylidene-20-substituted peni-
cillin sulfone. J Biol Chem 284:945–953.
of Pseudomonas aeruginosa 18S that synthesizes
type Id b-lactamase constitutively.
J Bacteriol 127:
1585–1586.
63. Lodge JM, Minchin SD, Piddock LJ, Busby JW (1990)
Cloning, sequencing and analysis of the structural
gene and regulatory region of the Pseudomonas aerugi-
nosa chromosomal ampC b-lactamase. Biochem J 272:
627–631.
68. Thomson JM, Distler AM, Prati F, Bonomo RA (2006)
Probing active site chemistry in SHV b-lactamase var-
iants at Ambler position 244. Understanding unique
properties of inhibitor resistance. J Biol Chem 281:
26734–26744.
64. Knott-Hunziker V, Petursson S, Jayatilake GS, Waley
SG, Jaurin B, Grundstrom T (1982) Active sites of b-
lactamases. The chromosomal b-lactamases of Pseudo-
monas aeruginosa and Escherichia coli. Biochem J 201:
621–627.
65. Drawz SM, Babic M, Bethel CR, Taracila M, Distler
AM, Ori C, Caselli E, Prati F, Bonomo RA (2010) Inhi-
bition of the class C b-lactamase from Acinetobacter
spp.: insights into effective inhibitor design. Biochemis-
try 49:329–340.
69. Chen Y, Shoichet B, Bonnet R (2005) Structure, func-
tion, and inhibition along the reaction coordinate of
CTX-M b-lactamases. J Am Chem Soc 127:5423–5434.
70. Wang X, Minasov G, Blazquez J, Caselli E, Prati F,
Shoichet BK (2003) Recognition and resistance in TEM
b-lactamase. Biochemistry 42:8434–8444.
71. Copeland RA (2005)Evaluation of Enzyme Inhibitors in
Drug Discovery; John Wiley & Sons, Inc.
72. Cheng Y, Prusoff WH (1973) Relationship between the
inhibition constant (Ki) and the concentration of inhibi-
tor which causes 50 per cent inhibition (IC50) of an
enzymatic reaction. Biochem Pharmacol 22:3099–3108.
66. Hujer AM, Hujer KM, Helfand MS, Anderson VE,
Bonomo RA (2002) Amino acid substitutions at Ambler
958 PROTEINSCIENCE.ORG
C3/C4 Carboxylate Recognition in P. aeruginosa AmpC