Rapid Broad Spectrum Detection of Carbapenemases with a Dual Fluorogenic-Colorimetric Probe

Article abstract of DOI:10.1021/jacs.1c00462

Carbapenems stand as one of the last-resort antibiotics; however, their efficacy is threatened by the rising number and rapid spread of carbapenemases. Effective antimicrobial stewardship thus calls for rapid tests for these enzymes to aid appropriate prescription and infection control. Herein, we report the first effective pan-carbapenemase reporter CARBA-H with a broad scope covering all three Ambler classes. Using a chemical biology approach, we demonstrated that the absence of the 1β-substituent in the carbapenem core is key to pan-carbapenemase recognition, which led to our rational design and probe development. CARBA-H provides a dual colorimetric-fluorogenic response upon carbapenemase-mediated hydrolysis. A clear visual readout can be obtained within 15 min when tested against a panel of carbapenemase-producing Enterobacteriaceae (CPE) clinical isolates that notably includes OXA-48 and OXA-181-producing strains. Furthermore, CARBA-H can be applied to the detection of carbapemenase activity in CPE-spiked urine samples.

Full text of DOI:10.1021/jacs.1c00462

pubs.acs.org/JACS
Article
Rapid Broad Spectrum Detection of Carbapenemases with a Dual
Fluorogenic-Colorimetric Probe
Chi-Wang Ma, Kenneth King-Hei Ng, Bill Hin-Cheung Yam, Pak-Leung Ho, Richard Yi-Tsun Kao,
and Dan Yang*
Cite This: J. Am. Chem. Soc. 2021, 143, 6886−6894
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Carbapenems stand as one of the last-resort anti-
biotics; however, their efficacy is threatened by the rising number and
rapid spread of carbapenemases. Effective antimicrobial stewardship
thus calls for rapid tests for these enzymes to aid appropriate
prescription and infection control. Herein, we report the first effective
pan-carbapenemase reporter CARBA-H with a broad scope covering
all three Ambler classes. Using a chemical biology approach, we
demonstrated that the absence of the 1β-substituent in the
carbapenem core is key to pan-carbapenemase recognition, which
led to our rational design and probe development. CARBA-H
provides a dual colorimetric-fluorogenic response upon carbapene-
mase-mediated hydrolysis. A clear visual readout can be obtained
within 15 min when tested against a panel of carbapenemase-
producing Enterobacteriaceae (CPE) clinical isolates that notably includes OXA-48 and OXA-181-producing strains. Furthermore,
CARBA-H can be applied to the detection of carbapemenase activity in CPE-spiked urine samples.
challenge for clinicians in choosing the appropriate antibiotics
from the hierarchy of therapeutic options. Rapid identification
of carbapenemases therefore provides the bedrock for both
infection control and appropriate antibiotic prescription, two
main pillars in antimicrobial stewardship.
INTRODUCTION
■
Carbapenems are one of the most effective broad-spectrum
drugs-of-last-resort against bacterial infection. However, the
emergence and dissemination of resistance toward carbape-
nems in recent decades have become a major public health
issue worldwide.1 Microbial production of carbapenemases,
enzymes that catalyze the hydrolytic degradation of carbape-
nems, in addition to virtually all other β-lactams, represents the
most clinically important resistance mechanism in carbape-
nem-resistant bacteria.2
Carbapenemases belong to the larger family of β-lactamases
and are categorized based on their amino acid sequences and
hydrolysis mechanisms under the Ambler classification.2,3
Carbapenemases in Classes A (e.g., KPC) and D (e.g., OXA-
48-like) are serine hydrolases, and those in Class B (e.g.,
NDM, IMP, VIM) are Zn-dependent metallo-β-lactamases.
Over 200 unique carbapenemases have been identified to date,
and the numbers are ever increasing.2
These carbapenemases are usually encoded on mobile
genetic elements which greatly facilitate the transmission of
the resistance, even across different bacterial species, through
horizontal gene transfer.2 Effective infection control programs
are therefore key to curtail their spread in hospitals and
dissemination into the community.1 Patients infected by
carbapenem-resistant bacteria may not show distinct symp-
toms, yet delay in effective treatment can lead to life-
threatening complications like sepsis.4,5 This poses a great
Despite their sensitivity in detecting certain carbapenemase-
encoding genes, genotypic tests generally suffer from high cost,
labor-intensive protocols and requirement of specialized
instrumentation.6 Furthermore, β-lactamases within the
OXA-family that exhibit diverse carbapenemase activities may
not be readily distinguished by polymerase chain reaction
(PCR). Immunochromatographic assays, including RESIST-4
and CARBA 5, which do not require sophisticated instruments,
have also been developed.6 However, the scope of these assays
are delimitated by recognizing particular fragment of DNA or
protein sequence, and novel mutations may evade their
detection. Together, these shortcomings urgently call for a
reliable activity-based phenotypic assay.7,8
The only rapid phenotypic test for carbapenemase detection
approved by Clinical & Laboratory Standards Institute (CLSI)
Received: January 14, 2021
Published: April 28, 2021
© 2021 The Authors. Published by
American Chemical Society
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894
6886

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Previously reported carbapenem-based turn-on fluorogenic probes and the current design.
Figure 2. (a) Structures of some common carbapenem antibiotics. The 1β-methyl substituent was highlighted with a green background. (b)
Recombinant OXA-48 (5 μM) was incubated with MERO-Alk at different concentrations, followed by CuAAC and SDS-PAGE analysis.
Coomassie blue staining was indicated with C and in-gel fluorescence visualization was indicated with F.
to date is CarbaNP.9,10 While nominally it can offer a result in
2 h, the pH-based colorimetric assay is highly insensitive.
Notably, false-negative results have been reported in the
detection of OXA-48 (sensitivity as low as 11%) and enzymes
possessing lower carbapenemase activities like SME and
GES.7,9,10 Indeterminate results can also arise due to the
subtlety of the color difference, thus invalidating the test in
those circumstances. Since its identification, OXA-48 has
emerged as one of the most clinically important carbapene-
mases,11−14 yet none of the existing chemical probes could
address the challenge of its detection.9,10,15−19 In addition,
certain carbapenem-derived fluorescent probes reported in the
literature only responded to particular classes of carbapene-
mases16 (Figure 1a−c), underscoring the fact that careful
consideration must be put into probe design in order to cover
the entire range of carbapenemases. Noting the above
shortcomings, we endeavor to design a fluorogenic substrate
that selectively turns on toward a broad spectrum of
carbapenemases, encompassing the OXA-48-like carbapene-
mases in particular, and remains unresponsive toward non-
carbapenemases.
RESULTS AND DISCUSSION
■
Hypothesis and Preliminary Observation. The failure
of OXA-48 detection by existing phenotypic tests points to a
poor enzyme−substrate interaction, thus, we aimed to discover
a molecular architecture that favors the binding and
recognition. We noticed from the literature that OXA-48
exhibited at least 2 orders of magnitude higher catalytic
efficiency (kcat/KM) toward hydrolysis of imipenem and
panipenem than that of meropenem, ertapenem, and
doripenem.20,21 Although these carbapenems bear different
side chains, we recognized a dichotomous feature in which a
1β-methyl substituent is present at the carbapenem core in the
latter group but not in the former (Figure 2a). While the 1β-
methyl substituent is crucial for the suppression of hydrolytic
6887
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Scheme 1. Summary of Synthesis Toward CARBA-H
degradation of carbapenems by human peptidase DHP-1 in the
renal brush border,22,23 it may account for the failure of
existing carbapenem-based sensors in detecting OXA-48.
Meanwhile, Schofield and co-workers’ molecular dynamic
simulations provided in silico evidence that the 1β-methyl
substituent in doripenem hindered the hydrolysis of the
enzyme−substrate complex in OXA-1, a non-carbapenemase
homologue of OXA-48.24 We therefore hypothesized that the
1β-methyl substituent in carbapenems stabilizes the acyl-
enzyme intermediate in OXA-48 and impedes the enzymatic
turnover.
commercially available acid-cleavable biotin-azide. The acy-
lated OXA-48 was enriched by streptavidin−biotin enrichment
followed by trypsin digestion. The modified peptide was
cleaved and analyzed by LC-MS/MS, and the fragment search
helped us identify that the modification took place in the active
site (Figure S2). This mechanistic investigation uncovered a
novel role of the 1β-methyl substituent in carbapenems and
provided a rationale for our probe design.
In addition, we hope to integrate another structural feature
in our design to achieve generality for pan-carbapenemase
detection. It has been observed that the hydrolysis kinetics of
certain β-lactamases are sensitive to substituent effects in the
proximity of the β-lactam ring.16,25−27 As different β-
lactamases operate through different hydrolytic mechanisms,
their active sites also exhibit significantly diverse structures.28
Hence, we reasoned that by placing fluorophores, usually
featuring extended π-systems that are electronically dissimilar
to the polar side chains commonly found in carbapenems,
further away from the core would allow probe recognition by a
wider range of carbapenemases.
Probe Design and Synthesis. To this end, we aspired to
construct our carbapenemase probes from three key building
blocks, namely, a fluorogenic unit, a self-immolative spacer,
and a carbapenem core. Resorufin was selected as the
fluorogenic unit, as it not only demonstrates a strong turn-on
fluorescence upon uncaging but also exhibits long excitation
and emission wavelengths, which minimize undesired interfer-
ence from autofluorescent biological molecules.29−31 In
addition, upon release, it also provides a colorimetric change
that can be captured by naked eye, thus precluding the need of
extra instrumentation. A self-immolative allylic-hydroxyben-
zylic linker was used to bridge the fluorogenic unit and the
carbapenem core, which should steer the fluorogenic unit away
from the core and enable a broader carbapenemase
recognition. To pin down the role of the 1β-methyl group as
a stereochemical determinant for the selectivity in OXA-48, we
designed two target molecules, namely, CARBA-H and
CARBA-Me, with and without the 1β-methyl substituent in
the carbapenem core, respectively (Figure 1d).
A modular synthetic strategy was adopted (Scheme 1),
which allows for flexibility in the installation of alternative
chromophores if other emission wavelengths or detection
approaches are desired. The linker-fluorophore fragment 4 was
prepared from propargyl alcohol through an 8-step sequence
that terminates with the attachment of resorufin (Scheme S1).
For the carbapenem fragment, we adopted the azidobenzyl
ester as a carboxylic acid protecting group.32,33 It provides
To provide experimental evidence, we prepared an alkyne-
appended meropenem derivative, MERO-Alk, as a representa-
tive 1β-methyl bearing substrate for preliminary investigations
(Figure 2b). If a stable covalent adduct is formed with OXA-
48, the enzyme can be fluorescently labeled upon conjugation
to an azido fluorophore using copper-catalyzed azide−alkyne
cycloaddition (CuAAC). Accordingly, recombinant OXA-48
was incubated with different concentrations of MERO-Alk for
30 min, followed by treatment with rhodamine-azide (Rh−N3)
under CuAAC conditions, and the results were evaluated using
in-gel fluorescent visualization. In line with our hypothesis, an
increase in fluorescence was observed in the presence of
MERO-Alk in a dose dependent manner, suggesting that
MERO-Alk forms a covalent adduct with OXA-48 (Figure 2b).
To validate whether the adduct formation preferentially
takes place with 1β-methyl substituted carbapenems, we
performed competition assays with meropenem and ertapenem
(both bearing a 1β-methyl group) and imipenem (without 1β-
methyl group). Recombinant OXA-48 was preincubated with
the aforementioned carbapenems (30 min) before the addition
of MERO-Alk. In-gel visualization showed that the fluo-
rescence signal can be depleted by meropenem and ertapenem
but not by imipenem, strongly indicating that the covalent
modification took place in the enzyme active site selectively
with carbapenems featuring a 1β-methyl group (Figure S1a−
c). Analogous labeling experiments were conducted with a
panel of purified recombinant carbapenemases which are
clinically relevant. Neither carbapenemases from other classes
(i.e., Classes A and B) nor heat-denatured OXA-48 but only
catalytically active OXA-48 could be fluorescently tagged,
consistent with the notion that the covalent modification
corresponds to the formation of the acyl intermediate and is
specific toward OXA-48 (Figure S1d).
We further confirmed the acylation indeed took place at the
active site by a tandem mass spectrometry approach. OXA-48
was incubated with MERO-Alk, followed by CuAAC with a
6888
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. (a) Change in absorbance over time of CARBA-H (10 μM in PBS, pH 7.4) in the presence of IMP-1 (2 nM). (b) Fluorescence change
over time of CARBA-H (1 μM in PBS, pH 7.4) toward IMP-1 (0.5 nM). λex = 571 nm. (c) Extracted ion chromatogram of HPLC-ESI MS analysis.
CARBA-H (15 μM) hydrolyzed by IMP-1 (10 nM) in NH4OAc buffer (10 mM, pH 7.4), normalized to the largest peak. Observed peaks with
selected mass-to-charge ratio (m/z) corresponding to the expected fragments were indicated. (d) Fluorescence change over time of CARBA-H (2
μM in PBS, pH 7.4) toward selected β-lactamases. λex = 564 nm; λem = 629 nm. The error bars correspond to the standard deviation of the three
replicates. (e) Fluorescence change of CARBA-H (2 μM in PBS, pH 7.4) in the presence of selected carbapenemases and β-lactamase inhibitors
incubated for 30 min. λex = 564 nm; λem = 629 nm. The error bars correspond to the standard deviation of the three replicates.
orthogonality toward other protecting groups, tolerates a wide
range of transformations, and can be readily removed by
Staudinger reduction. For CARBA-H, a Lewis acid (ZnCl2)-
mediated Mannich reaction yielded azetidinone 2 as a single
diastereomer. Subsequent protecting group manipulations
afforded precursor 3. After rhodium-catalyzed ring closure
and subsequent derivatization of the ensuing ketone to a vinyl
triflate, fragments 3 and 4 were united utilizing a late-stage
Suzuki coupling, catalyzed by a 1:1 mixture of Pd2dba3 and
Pd(dppf)Cl2 under the optimized conditions to provide 5.
CARBA-H was successfully obtained upon global deprotec-
tion, and the control probe CARBA-Me was synthesized in an
analogous fashion (Scheme S2).
β-Lactamase Response toward CARBA-H and CARBA-
Me. With the probes in hand, we first proceeded to measure
the photophysical change of CARBA-H upon β-lactam ring
6889
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
opening by IMP-1, a typical carbapenemase. Over the course
of reaction between CARBA-H (10 μM) and IMP-1 (2 nM), a
decrease in absorbance at 480 nm was observed along with a
simultaneous increase at 571 nm (Figure 3a). The new
absorption maximum coincided with the reported value of
resorufin.31 A single isosbestic point at 507 nm implied that no
intermediate was formed during the hydrolysis (Figure 3a).
The presence of the expected fragmentation products was
confirmed using HPLC-ESI MS (Figure 3c). Expectedly, the
addition of IMP-1 resulted in the disappearance of signal
corresponding to protonated CARBA-H (9), and the
appearance of those corresponding to the protonated
pyrrolidine 6, quinonemethide 7 and resorufin 8. For the
fluorescence change, CARBA-H (1 μM) provided more than
60-fold fluorescence enhancement upon the addition of IMP-1
(0.5 nM) in only 8 min (Figure 3b). This compares very
favorably with existing reports,10,15−18 e.g., a much higher
concentration (100 nM) of IMP-1 was required (cf. 0.5 nM
IMP-1 for CARBA-H) to achieve a similar degree of
enhancement with CB-1 under otherwise identical condi-
tions.15 The pseudo-first order rate constant for the
spontaneous hydrolysis reaction (PBS, pH 7.4) of CARBA-H
was estimated to be 1.37 × 10−5 s−1 at room temperature,
similar to that of carbapenems (Figure S3).16,19,34,35
We then evaluated the selectivity of the response of
CARBA-H toward carbapenemases over other β-lactamases.
Since NDM, IMP, VIM (Ambler Class B metallo-β-
lactamases), KPC (Class A serine carbapenemases), and
OXA-48-like carbapenemases (Class D serine carbapene-
mases) account for more than 99% of the carbapenemase-
producing isolates,36 a representative member from each of
these families, namely, NDM-1, IMP-1, VIM-2, KPC-2, and
OXA-48, were recombinantly expressed, purified, and utilized
in this study to examine the carbapenemase selectivity and
coverage of our probes (Figure S4). Two common non-
carbapenemases, AmpC (Class C serine β-lactamase) and
TEM-1 (Class A serine β-lactamase), were employed as
negative controls. The fluorescence change of CARBA-H (2
μM) toward the aforementioned enzymes (at 1 nM
concentration except 30 nM for TEM-1) was monitored
over 30 min (Figure 3d). CARBA-H only gave significant
fluorescence enhancement in the presence of carbapenemases
(i.e., NDM-1, IMP-1, VIM-2, KPC-2, and OXA-48), but not
with non-carbapenemases (i.e., AmpC and TEM-1) over the
period.
the detection limits of CARBA-H and CARBA-Me toward
each carbapenemase can be improved upon prolonged
incubation, we chose a 30 min time frame for evaluation
purposes as it represents a reasonable duration for a rapid test.
The detection limits were in picomolar range for both probes
toward VIM-2 and KPC-2, and in subpicomolar range for
IMP-1 and NDM-1 (Table 1, Table S1, Figures S6, S7). To
Table 1. Detection Limit (3S/k) of CARBA-H (2 μM in
PBS, pH 7.4) Toward Selected Carbapenemases Where S Is
a
the Standard Deviation of 6 Blank Samples
Carbapenemase
3S/k
NDM-1
IMP-1
VIM-2
KPC-2
OXA-48
0.327 pM
0.333 pM
3.28 pM
4.43 pM
30.3 pM
a
λex = 564 nm; λem = 629 nm.
our delight, while OXA-48 led to the slowest turn-on
fluorescence among all the carbapenemases, the corresponding
value is still in the picomolar range.
To gain a better understanding of the substrate-enzyme
interactions between CARBA-H/CARBA-Me and our panel of
carbapenemases, the Michaelis−Menten kinetics model was
adopted. Our enzyme kinetics data showed that CARBA-H
served as an excellent substrate for carbapenemases, explaining
the rapid turn-on response (Table 2, Figure S8). For VIM-2,
a
Table 2. Michaelis-Menten Kinetics Data
enzyme
kcat (s−1
)
KM (μM)
kcat/KM (M−1 s−1
)
NDM-1
IMP-1
VIM-2
KPC-2
OXA-48
29.78 0.81
33.86 0.71
1.96 0.05
2.17 0.07
2.83 0.18
1.66 0.13
1.20 0.08
1.18 0.10
0.93 0.11
0.66 0.08
1.80 × 107
2.82 × 107
1.66 × 106
2.32 × 106
4.25 × 106
a
kcat and KM values for CARBA-H.
KPC-2, and OXA-48, the turnover numbers (kcat/KM) were
found to be in the order of 106 M−1 s−1, while those of NDM-1
and IMP-1 were even in the order of 107 M−1 s−1. In general,
the kcat/KM values for NDM-1, IMP-1, VIM-2, and KPC-2
toward CARBA-H and CARBA-Me are similar to that with
imipenem and meropenem (Table 2, Table S2, Figure
S9).38−42 However, the corresponding values for the reaction
between OXA-48 and CARBA-Me could not be determined,
as the initial hydrolysis rate of CARBA-Me was found to be
independent of OXA-48 concentration. The observation is
again consistent with our preliminary protein labeling study in
which 1β-methyl bearing substrates form a relatively stable acyl
intermediate with OXA-48, thus rendering the Michaelis−
Menten model unsuitable for the analysis.
Rapid Test Assays with Clinical Isolates. Encouraged by
the results, we proceeded to apply CARBA-H to detect
carbapenemase activities in clinical isolates of Klebsiella
pneumoniae, Escherichia coli, Enterobacter aerogenes (Klebsiella
aerogenes), and Citrobacter freundii, which represent the most
clinically relevant families of carbapenem-resistant Enter-
obacteriaceae.1 Thirteen carbapenemase-encoding strains
were used as positive controls (Figure 4). In addition to
producers of KPC-2 (strain 1) and NDM-1 (strain 2), of which
To validate that the response of CARBA-H was due to bona
fide carbapenemase activity, we tested the effect of phenyl-
boronic acid (PBA), a KPC-specific inhibitor, and ethyl-
enediaminetetraacetic acid (EDTA), a metallo-β-lactamase
inhibitor, on the hydrolysis of CARBA-H by carbapenemases.
In the presence of PBA, only KPC-2 activity was significantly
blunted, and the enzymatic activities of NDM-1, IMP-1, VIM-
2, and OXA-48 were not affected. Meanwhile, in the presence
of EDTA, the activities of NDM-1, IMP-1, and VIM-2 were
inhibited but remained unchanged for KPC-2 (Figure 3e).
Interestingly, OXA-48 displayed a slightly decreased activity in
the presence of EDTA, which is likely due to the increase in
ion concentration.37
We also evaluated the performance of CARBA-Me against
the same panel of carbapenemases (Figure S5). The
fluorescence turn-on profiles for NDM-1, IMP-1, KPC-2, and
VIM-2 were similar to that of CARBA-H; however, CARBA-
Me did not yield a significant response toward OXA-48. While
6890
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. (a) Colorimetric and fluorogenic (UV 254 nm) response of CARBA-H (10 μM) (b) CarbaNP solution A and solution B, toward clinical
isolates according to CLSI protocol. Images were taken after incubation for 15 min. The carbapenemases encoded in the clinical isolates were
determined by PCR. Strain (1) K. pneumoniae (KPC-2); (2) E. coli (NDM-1); (3) E. coli (NDM-5); (4) E. coli (IMP-4); (5) E. coli (OXA-181);
(6) E. coli (OXA-181); (7) K. pneumoniae (OXA-181); (8) K. pneumoniae (OXA-181); (9) E. coli (OXA-48); (10) E. aerogenes (OXA-48); (11) E.
aerogenes (OXA-48); (12) C. freundii (OXA-48); (13) K. pneumoniae (OXA-48); (14−21) K. pneumoniae (non-CPE); (22−29) E. coli (non-CPE).
the enzymatic activities toward CARBA-H were carefully
characterized, we included strains that produce NDM-5 (strain
3) and IMP-4 (strain 4), homologues of NDM-1 and IMP-1,
respectively, to demonstrate the pan-carbapenemase coverage
of our probe. Moreover, 9 strains (from 4 different organisms)
that express OXA-181 (strains 5−8) and OXA-48 (strains 9−
13) were employed to showcase the unique ability of CARBA-
H to detect the OXA-48-like family. The presence of the
carbapenemase-encoding genes in these isolates were cross
validated by PCR, and 16 non-carbapenemase-producing
Enterobacteriaceae (non-CPE) isolates (strains 14−29) were
used as negative controls.
While the detection can be carried out with live bacteria as
carbapenemases are often secreted (Figure S10a), we sought to
lyse the bacteria in order to release the carbapenemases from
the periplasmic space of the bacteria for rapid detection
purposes. We first investigated the efficiency of some
nondenaturing detergents, namely Triton X-100, Tween-20,
3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesul-fo-
nate (CHAPS), and B-PER, as additives in PBS buffer.
Accordingly, bacteria were incubated with CARBA-H in the
detergent-containing buffer solutions, and the efficiency of
bacterial lysis was determined by the degree of color change
after 15 min (Figure S10b-S11). CHAPS was found to provide
the most rapid color change and hence was selected for the
subsequent rapid test assay. Thus, clinical isolates were lysed in
PBS buffer containing 0.5% CHAPS for 15 min before
CARBA-H treatment (Figure S12). The carbapenemase-
encoding strains (Figure 4a; strains 1−13) induced marked
color change from pale orange to deep pink, visible under
ambient light, within 15 min. In contrast, the color of the
negative controls remained unchanged (Figure 4a; strains 14−
29). The fluorescence turn-on could also be conveniently
visualized with the aid of a hand-held UV lamp. Again, only the
CPE strains provided intense fluorescence signals (Figure 4a;
strains 1−13 vs 14−29). A consistent observation was obtained
after prolonged incubation (120 min; Figure S13a), indicating
the fluorescent probe and fluorophore are stable in the
complex biological system as a reliable rapid test. A head-to-
head comparison was made by performing the CarbaNP test in
parallel. In contrast to CARBA-H, CarbaNP did not provide
any positive color change for the bacteria expressing Class D
carbapenemases (strains 9−13) within 15 min (Figure 4a).
Even with prolonged incubation for up to 120 min, the
maximum incubation time as recommended by CLSI for
CarbaNP, CarbaNP did not provide a distinct color change,
corresponding to false-negative or inconclusive results for
those isolates (Figure S13c). Moreover, due to the high
sensitivity, even OXA-232, a known Class D carbapenemase
with notably low carbapenemase activity, can be clearly
detected (Figure S13a, strains 30−32). For completeness, a
direct comparison against CARBA-Me was also performed
(Figure S13b). As expected, after 120 min, CARBA-Me
provided the colorimetric and fluorogenic responses toward
KPC, NDM, and IMP-encoding strains but variable results
toward the strains encoding the OXA-48-like carbapenemases.
Together, our study clearly showcased the superior perform-
ance of CARBA-H over existing protocols and probe designs.
Urine Test with Clinical Isolates. To demonstrate
potential clinical applicability of CARBA-H, we set out to
detect carbapenemase activity directly in urine samples.
Urinary tract infection (UTI) was reported to be leading
healthcare-associated infection among hospitalized adults and
in critical care units, and the second or third most common
type of nosocomial infection in intensive care units.43,44 As a
proof of concept, we mimicked urine samples from UTI
patients by spiking urine from healthy subjects with known
amounts of well-characterized carbapemenase-producing clin-
ical isolates, particularly with the inclusion of OXA-48, OXA-
181, and OXA-232-producing strains. Remarkably, CARBA-H
(0.25 μM) provided significant fluorescence enhancement in
all CPE-spiked (5 × 106 cfu) urine samples, in the presence of
0.5% CHAPS, within 2 h. The turn-on response could be
readily detected by a microplate reader, with a clear
differentiation against the background from non-CPE E. coli
and K. pneumoniae isolates (Figure 5). This assay potentially
obviates the need for bacterial culture and creates new
opportunities for rapid detection of UTI by CRE, as the
6891
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(Figures S16 and Scheme S4), leading to a drastic decrease in
sensitivity. In contrast, the small allylic spacer in CARBA-H
and CARBA-Me provided a complete and immediate
fluorophore uncaging upon carbapenem ring opening (Figure
S17), resulting in maximum chromophore release suited for
rapid detection. This highlights that, in the context of small
molecule-based chemosensors, extra consideration should be
made in linker selection for future probe design to maximize
signal generation.
Through a chemical biology approach, we have uncovered a
crucial role of 1β-methyl group in carbapenems during OXA-
48 mediated hydrolysis. Recent reports by Chen47 and
Vakulenko48 on crystal structures of OXA-48 and its covalent
adducts with several carbapenems shed light on the enzymatic
mechanisms, yet these structural studies could not offer ready
explanation for the slow hydrolysis kinetics with 1β-methyl
substituted carbapenems.49 Meanwhile, Schofield and co-
workers showed that the presence of 1β-methyl group led to
a detour in the carbapenem degradation pathway with OXA-
48.24 In this study, we revealed that this substituent is crucial in
slowing down the deacylation process, and this discovery
provides new insight for antibiotic therapy and drug discovery.
Clinical research has shown that patient infected with certain
carbapenemase-producing bacteria (e.g., OXA-48-encoding
bacteria) could be effectively treated with coadministration
of two carbapenem antibiotics (i.e., double-carbapenem
therapy).50,51 To explain the observed synergistic effect, we
postulated that one of the carbapenems act as a suicidal
inhibitor for OXA-48, and the other targets the bacterial cell
wall synthesis.50,51 While there is almost no molecular evidence
to back up such treatment, our finding provided strong
rationale for this therapy from a mechanistic perspective.
Figure 5. Fluorescent response of CARBA-H (0.25 μM; λex = 564
nm; λem = 629 nm) toward carbapenemase in a panel of 71
Enterobacteriaceae clinical isolates spiked in urine (5 × 106 cfu), was
determined upon incubation at 37 °C for 2 h. Strain (1) E. coli (IMP-
4); (2) K. pneumoniae (KPC-2); (3) E. coli (NDM-1); (4) E. coli
(NDM-5); (5) C. freundii (OXA-48); (6) E. aerogenes (OXA-48); (7)
E. aerogenes (OXA-48); (8) E. coli (OXA-48); (9) K. pneumoniae
(OXA-48); (10) E. coli (OXA-181); (11) E. coli (OXA-181); (12) K.
pneumoniae (OXA-181); (13) K. pneumoniae (OXA-181); (14) E. coli
(OXA-232); (15) K. pneumoniae (OXA-232); (16) K. pneumoniae
(OXA-232); (17−24) K. pneumoniae (non-CPE); (25−71) E. coli
(non-CPE). The error bars correspond to the standard deviation of
the two replicates. Statistical significance was calculated using the
unpaired two-tailed Student’s t test with Welch’s correction (** p <
0.01, **** p < 0.0001).
CONCLUSION
■
necessary bacterial count is readily achievable by routine
centrifugation of patient urine samples as per traditional
urinalysis45 (typical UTI patients are reported to have a
bacterial count exceeding 1 × 105 cfu/mL in their urine).44,46
This could further reduce the time for culturing bacteria and
minimize delay for the golden treatment period. The
compatibility of the fluorogenic carbapenemase detection
using CARBA-H in urine samples highlights the judicial
choice of the reporting fluorophore (Figure S14) and
demonstrates the scope for extending this strategy for use in
other biological specimens.
We have developed CARBA-H successfully as the first broad
spectrum off-on fluorogenic carbapenemase probe. Through
direct comparison on the performance of CARBA-H with
CARBA-Me and a series of mechanistic studies, we have
provided clear evidence that the 1β-methyl substituent strongly
affects the OXA-48 enzyme hydrolytic kinetics. CARBA-H
exhibit excellent sensitivity toward the 5 clinically most
relevant carbapenemases, including OXA-48, of which the
detection limit was determined to be in the pico-molar range.
We have applied our probe CARBA-H to an optimized rapid
test for carbapenemase detection in clinical isolates. The
absence of the 1β-methyl substituent is essential for pan-
carbapenemase recognition, where carbapenemase encoding
strains, including OXA-48, OXA-181, and OXA-232, could be
successfully detected. Reliable chromogenic and fluorogenic
readouts could be easily discerned by naked eye within 15 min,
demonstrating its outstanding performances compared to
currently existing assays. Our success in establishing an
operationally simple protocol for detecting carbapenemase
activities from clinical isolates should accelerate the develop-
ment of an affordable point-of-care device, for antimicrobial
resistance tests performed without specialized equipment
available only in hospitals.
DISCUSSION
■
The success in the development of CARBA-H as a fluorescent
probe for broad spectrum carbapenemase detection hinges
upon our rational design with two key features, i.e., an effective
self-immolative spacer and the lack of the 1β-methyl
substituent at the carbapenem core.
The allylic-hydroxybenzylic self-immolative spacer in our
probe design is crucial in providing high sensitivity and a broad
scope for carbapenemase detection. A CARBA-H analog
without the self-immolative spacer was also prepared as a
control (Figure S15 and Scheme S3). As expected, the
molecule displayed poor response toward KPC-2 and suffered
from poor aqueous stability. At the outset of the project, we
also briefly explored the use of a phenyl spacer, which was later
found to be utilized in a recent carbapenem-based probe
design.18 However, this strategy resulted in incomplete
fluorophore release during carbapenemase mediated hydrolysis
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c00462.
6892
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Humphries, R.; Rekasius, V.; Scangarella-Oman, N.; Westblade, L.
CLSI AST News Update 2017, 2, 1−17.
(PDF)
(8) Hemarajata, P.; Yang, S.; Hindler, J. A.; Humphries, R. M.
Development of a novel real-time PCR assay with high-resolution
melt analysis to detect and differentiate OXA-48-Like β-lactamases in
carbapenem-resistant Enterobacteriaceae. Antimicrob. Agents Chemo-
ther. 2015, 59, 5574−80.
(9) Tamma, P. D.; S, P. J. Phenotypic Detection of Carbapenemase-
Producing Organisms from Clinical Isolates. J. Clin. Microbiol. 2018,
56, e01140−18.
AUTHOR INFORMATION
Corresponding Author
Dan Yang − Morningside Laboratory for Chemical Biology,
Department of Chemistry, The University of Hong Kong,
Hong Kong, China; orcid.org/0000-0002-1726-9335;
Email: yangdan@hku.hk
■
(10) CLSI Performance Standards for Antimicrobial Susceptibility
Testing., 29thed.; Clinical and Laboratory Institute: Wayne, PA 2019.
(11) Pitout, J. D. D.; Peirano, G.; Kock, M. M.; Strydom, K. A.;
Matsumura, Y. The Global Ascendency of OXA-48-Type Carbape-
nemases. Clin. Microbiol. Rev. 2019, 33, e00102−19.
(12) Hernandez-Garcia, M.; Perez-Viso, B.; Navarro-San Francisco,
C.; Baquero, F.; Morosini, M. I.; Ruiz-Garbajosa, P.; Canton, R.
Intestinal co-colonization with different carbapenemase-producing
Enterobacterales isolates is not a rare event in an OXA-48 endemic
area. EClinicalMedicine 2019, 15, 72−79.
(13) Branas, P.; Gil, M.; Villa, J.; Orellana, M. A.; Chaves, F.
Molecular epidemiology of carbapenemase-producing Enterobacter-
iaceae infection/colonisation in a hospital in Madrid. Enferm. Infecc.
Microbiol. Clin. 2018, 36, 100−103.
(14) Cui, X.; Zhang, H.; Du, H. Carbapenemases in Enter-
obacteriaceae: Detection and Antimicrobial Therapy. Front. Microbiol.
2019, 10, 1823.
(15) Mao, W.; Xia, L.; Xie, H. Detection of Carbapenemase-
Producing Organisms with a Carbapenem-Based Fluorogenic Probe.
Angew. Chem., Int. Ed. 2017, 56, 4468−4472.
(16) Mao, W.; Wang, Y.; Qian, X.; Xia, L.; Xie, H. A Carbapenem-
Based Off-On Fluorescent Probe for Specific Detection of Metallo-β-
Lactamase Activities. ChemBioChem 2019, 20, 511−515.
(17) Kim, J.; Kim, Y.; Abdelazem, A. Z.; Kim, H. J.; Choo, H.; Kim,
H. S.; Kim, J. O.; Park, Y. J.; Min, S. J. Development of carbapenem-
based fluorogenic probes for the clinical screening of carbapenemase-
producing bacteria. Bioorg. Chem. 2020, 94, 103405.
(18) Das, S.; Ihssen, J.; Wick, L.; Spitz, U.; Shabat, D.
Chemiluminescence Carbapenem-based Molecular Probe for Detec-
tion of Carbapenemase Activity in Live Bacteria. Chem. - Eur. J. 2020,
26, 3647.
(19) Wang, J.; Xu, W.; Xue, S.; Yu, T.; Xie, H. A minor structure
modification serendipitously leads to a highly carbapenemase-specific
fluorogenic probe. Org. Biomol. Chem. 2020, 18, 4029−4033.
(20) Docquier, J. D.; Calderone, V.; De Luca, F.; Benvenuti, M.;
Giuliani, F.; Bellucci, L.; Tafi, A.; Nordmann, P.; Botta, M.; Rossolini,
G. M.; Mangani, S. Crystal structure of the OXA-48 β-lactamase
reveals mechanistic diversity among class D carbapenemases. Chem.
Biol. 2009, 16, 540−7.
(21) Papp-Wallace, K. M.; Kumar, V.; Zeiser, E. T.; Becka, S. A.; van
den Akker, F. Structural Analysis of The OXA-48 Carbapenemase
Bound to A ″Poor″ Carbapenem Substrate, Doripenem. Antibiotics
2019, 8, 145−158.
(22) El-Gamal, M. I.; Brahim, I.; Hisham, N.; Aladdin, R.;
Mohammed, H.; Bahaaeldin, A. Recent updates of carbapenem
antibiotics. Eur. J. Med. Chem. 2017, 131, 185−195.
(23) Fukasawa, M.; Sumita, Y.; Harabe, E. T.; Tanio, T.; Nouda, H.;
Kohzuki, T.; Okuda, T.; Matsumura, H.; Sunagawa, M. Stability of
meropenem and Effect of 1β-Methyl Substitution on Its Stability in
the Presence of Renal Dehydropeptidase I. Antimicrob. Agents
Chemother. 1992, 7, 1577−1579.
(24) Lohans, C. T.; van Groesen, E.; Kumar, K.; Tooke, C. L.;
Spencer, J.; Paton, R. S.; Brem, J.; Schofield, C. J. A New Mechanism
for β-Lactamases: Class D Enzymes Degrade 1β-Methyl Carbapenems
through Lactone Formation. Angew. Chem., Int. Ed. 2018, 57, 1282−
1285.
Authors
Chi-Wang Ma − Morningside Laboratory for Chemical
Biology, Department of Chemistry, The University of Hong
Kong, Hong Kong, China; orcid.org/0000-0002-6785-
5148
Kenneth King-Hei Ng − Morningside Laboratory for
Chemical Biology, Department of Chemistry, The University
of Hong Kong, Hong Kong, China
Bill Hin-Cheung Yam − Department of Microbiology and
Carol Yu Centre for Infection, The University of Hong Kong,
Hong Kong, China
Pak-Leung Ho − Department of Microbiology and Carol Yu
Centre for Infection, The University of Hong Kong, Hong
Kong, China; orcid.org/0000-0002-8811-1308
Richard Yi-Tsun Kao − Department of Microbiology and
Carol Yu Centre for Infection, The University of Hong Kong,
Hong Kong, China
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c00462
Notes
The authors declare the following competing financial
interest(s): The results reported in this manuscript were
included in a PCT Application (No. PCT/CN2019/106645)
filed by D.Y. and C.-W.M.
ACKNOWLEDGMENTS
■
We acknowledge support from The University of Hong Kong,
Morningside Foundation, small equipment grant (2019/20),
the Hong Kong Research Grants Council under the General
Research Fund Scheme (106200223), Areas of Excellence
Scheme (AoE/P-705/16), and Health and Medical Research
Fund (CID-HKU1-13).
REFERENCES
■
(1) CDC Antibiotic Resistance Threats in the United States, 2019.
Department of Health and Human Services, CDC. Atlanta, GA:U.S:
2019.
(2) Bush, K. Past and Present Perspectives on β-Lactamases.
Antimicrob. Agents Chemother. 2018, 62, e01076−18.
(3) Davies, J.; Davies, D. Origins and evolution of antibiotic
resistance. Microbiol. Mol. Biol. Rev. 2010, 74, 417−33.
(4) Funk, D. J.; Kumar, A. Antimicrobial therapy for life-threatening
infections: speed is life. Crit. Care Clin. 2011, 27, 53−76.
(5) Ibrahim, E. H.; Sherman, G.; Ward, S.; Fraser, V. J.; Kollef, M.
H. The Influence of Inadequate Antimicrobial Treatment of
Bloodstream Infections on Patient Outcomes in the ICU Setting.
Chest 2000, 118, 146−155.
(6) Meunier, D.; Freeman, R.; Hopkins, K. L.; Hopkins, S.;
Woodford, N.; Brown, C.; Chilton, C.; Davies, F.; Davies, K.; Lewis,
T.; Otter, J.; Richards, D.; Soyfoo, R.; Wilcox, M. Commercial assays
for the detection of acquired carbapenemases.; Public Health England:
2019.
(25) Palzkill, T. Structural and Mechanistic Basis for Extended-
Spectrum Drug-Resistance Mutations in Altering the Specificity of
TEM, CTX-M, and KPC β-lactamases. Front. Mol. Biosci. 2018, 5, 16.
(7) Hindler, J. A.; Schuetz, A. N.; Abbott, A.; Antonara, S.;
Bobenchik, A.; Castanheira, M.; Charnot-Katsikas, A.; Galas, M.;
6893
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(26) Neu, H. C. β-Lactam Antibiotics: Structural Relationships
Affecting in Vitro Activity and Pharmacologic Properties. Clin. Infect.
Dis. 1986, 8, S237−S259.
(27) Romero, A.; Bou, G. N.; Beceiro, A.; Santillana, E. Crystal
structure of the carbapenemase OXA-24 reveals insights into the
mechanism of carbapenem hydrolysis. Proc. Natl. Acad. Sci. U. S. A.
2007, 104, 5354−5359.
(28) Tooke, C. L.; Hinchliffe, P.; Bragginton, E. C.; Colenso, C. K.;
Hirvonen, V. H. A.; Takebayashi, Y.; Spencer, J. β-Lactamases and β-
Lactamase Inhibitors in the 21st Century. J. Mol. Biol. 2019, 431,
3472−3500.
(29) Chan, J.; Dodani, S. C.; Chang, C. J. Reaction-based small-
molecule fluorescent probes for chemoselective bioimaging. Nat.
Chem. 2012, 4, 973−84.
(30) Islam, M. S.; Honma, M.; Nakabayashi, T.; Kinjo, M.; Ohta, N.
pH dependence of the fluorescence lifetime of FAD in solution and in
cells. Int. J. Mol. Sci. 2013, 14, 1952−63.
(31) Bueno, C.; Villegas, M. L.; Bertolotti, S. G.; Previtali, C. M.;
Neumann, M. G.; Encinas, M. V. The Excited-State Interaction of
Resazurin and Resorufin with Aminesin Aqueous Solutions. Photo-
chem. Photobiol. 2002, 76, 385−390.
(32) Wang, J.; Chen, Y.; Yang, C.; Wei, T.; Han, Y.; Xia, M. An ICT-
based colorimetric and ratiometric fluorescent probe for hydrogen
sulfide and its application in live cell imaging. RSC Adv. 2016, 6,
33031−33035.
(43) Suay-Garcia, B.; Perez-Gracia, M. T. Present and Future of
Carbapenem-resistant Enterobacteriaceae (CRE) Infections. Anti-
biotics 2019, 8, 122.
(44) Chan, J. F.-W.; Yuen, K.-Y. A new ASPECT for complicated
urinary tract infections. Lancet 2015, 385, 1920−1922.
(45) Wilson, M. L.; Gaido, L. Laboratory Diagnosis of Urinary Tract
Infections in Adult Patients. Clin. Infect. Dis. 2004, 38, 1150−1158.
(46) Foxman, B. The epidemiology of urinary tract infection. Nat.
Rev. Urol. 2010, 7, 653−60.
(47) Smith, C. A.; Stewart, N. K.; Toth, M.; Vakulenkob, S. B.
Structural Insights into the Mechanism of Carbapenemase Activity of
the OXA-48 β-Lactamase. Antimicrob. Agents Chemother. 2019, 63,
e01202−19.
(48) Akhtar, A.; Pemberton, O. A.; Chen, Y. Structural Basis for
Substrate Specificity and Carbapenemase Activity of OXA-48 Class D
β-Lactamase. ACS Infect. Dis. 2020, 6, 261−271.
(49) Stojanoski, V.; Hu, L.; Sankaran, B.; Wang, F.; Tao, P.; Prasad,
B. V. V.; Palzkill, T. Mechanistic Basis of OXA-48-like β-Lactamases’
Hydrolysis of Carbapenems. ACS Infect. Dis. 2021, 7, 445−460.
(50) Galani, I.; Nafplioti, K.; Chatzikonstantinou, M.; Souli, M. In
vitro evaluation of double-carbapenem combinations against OXA-48-
producing Klebsiella pneumoniae isolates using time-kill studies. J.
Med. Microbiol. 2018, 67, 662−668.
(51) Mashni, O.; Nazer, L.; Le, J. Critical Review of Double-
Carbapenem Therapy for the Treatment of Carbapenemase-
Producing Klebsiella pneumoniae. Ann. Pharmacother. 2019, 53,
70−81.
(33) Griffin, R. J.; Evers, E.; Davison, R.; Gibson, A. E.; Layton, D.;
Irwin, W. J. The 4-azidobenzyloxycarbonyl function; application as a
novel protecting group and potential prodrug modification for amines.
J. Chem. Soc., Perkin Trans. 1 1996, 1205−1211.
(34) Dailly, E.; Bouquie, R.; Deslandes, G.; Jolliet, P.; Le Floch, R. A
liquid chromatography assay for a quantification of doripenem,
ertapenem, imipenem, meropenem concentrations in human plasma:
application to a clinical pharmacokinetic study. J. Chromatogr. B: Anal.
Technol. Biomed. Life Sci. 2011, 879, 1137−42.
(35) Viaene, E.; Chanteux, H.; Servais, H.; Mingeot-Leclercq, M. P.;
Tulkens, P. M. Comparative stability studies of antipseudomonal β-
lactams for potential administration through portable elastomeric
pumps (home therapy for cystic fibrosis patients) and motor-operated
syringes (intensive care units). Antimicrob. Agents Chemother. 2002,
46, 2327−32.
(36) English Surveillance Programme for Antimicrobial Utilisation
and Resistance (ESPAUR). Public Health England, London, 2018.
(37) Stojanoski, V.; Chow, D. C.; Fryszczyn, B.; Hu, L.; Nordmann,
P.; Poirel, L.; Sankaran, B.; Prasad, B. V.; Palzkill, T. Structural Basis
for Different Substrate Profiles of Two Closely Related Class D β-
Lactamases and Their Inhibition by Halogens. Biochemistry 2015, 54,
3370−80.
(38) Docquier, J. D.; Riccio, M. L.; Mugnaioli, C.; Luzzaro, F.;
Endimiani, A.; Toniolo, A.; Amicosante, G.; Rossolini, G. M. IMP-12,
a new plasmid-encoded metallo-β-lactamase from a Pseudomonas
putida clinical isolate. Antimicrob. Agents Chemother. 2003, 47, 1522−
8.
(39) Poirel, L.; Heritier, C.; Tolun, V.; Nordmann, P. Emergence of
oxacillinase-mediated resistance to imipenem in Klebsiella pneumo-
niae. Antimicrob. Agents Chemother. 2004, 48, 15−22.
(40) Alba, J.; Ishii, Y.; Thomson, K.; Moland, E. S.; Yamaguchi, K.
Kinetics study of KPC-3, a plasmid-encoded class A carbapenem-
hydrolyzing β-lactamase. Antimicrob. Agents Chemother. 2005, 49,
4760−2.
(41) Piccirilli, A.; Brisdelli, F.; Aschi, M.; Celenza, G.; Amicosante,
G.; Perilli, M. Kinetic Profile and Molecular Dynamic Studies Show
that Y229W Substitution in an NDM-1/L209F Variant Restores the
Hydrolytic Activity of the Enzyme toward Penicillins, Cephalosporins,
and Carbapenems. Antimicrob. Agents Chemother. 2019, 63, e02270−
18.
(42) Docquier, J. D.; Lamotte-Brasseur, J.; Galleni, M.; Amicosante,
G.; Frere, J. M.; Rossolini, G. M. On functional and structural
heterogeneity of VIM-type metallo-β-lactamases. J. Antimicrob.
Chemother. 2003, 51, 257−66.
6894
https://doi.org/10.1021/jacs.1c00462
J. Am. Chem. Soc. 2021, 143, 6886−6894