Chemiluminescence Detection of Hydrogen Sulfide Release by β-Lactamase-Catalyzed β-Lactam Biodegradation: Unprecedented Pathway for Monitoring β-Lactam Antibiotic Bacterial Resistance

β-Lactamase positive bacteria represent a growing threat to human health because of their resistance to commonly used antibiotics. Therefore, development of new diagnostic methods for identification of β-lactamase positive bacteria is of high importance for monitoring the spread of antibiotic-resistant bacteria. Here, we report the discovery of a new biodegradation metabolite (H2S), generated through β-lactamase-catalyzed hydrolysis of β-lactam antibiotics. This discovery directed us to develop a distinct molecular technique for monitoring bacterial antibiotic resistance. The technique is based on a highly efficient chemiluminescence probe, designed for detection of the metabolite, hydrogen sulfide, that is released upon biodegradation of β-lactam by β-lactamases. Such an assay can directly indicate if antibiotic bacterial resistance exists for a certain examined β-lactam. The assay was successfully demonstrated for five different β-lactam antibiotics and eight β-lactam resistant bacterial strains. Importantly, in a functional bacterial assay, our chemiluminescence probe was able to clearly distinguish between a β-lactam resistant bacterial strain and a sensitive one. As far as we know, there is no previous documentation for such a biodegradation pathway of β-lactam antibiotics. Bearing in mind the data obtained in this study, we propose that hydrogen sulfide should be considered as an emerging β-lactam metabolite for detection of bacterial resistance.


Topic
Page no. 1 General information 02 2 Synthetic schemes and experimental procedures 03 3 3a 3b General procedure for the data given in Figure 9 18 Determination of the Limit of Detection (LOD) of probe 1 for the detection of antibiotic resistant bacteria using Ceftizoxime 19 Determination of the relative rate constant of the reaction of hydrogen sulfide release by β-lactamase-catalyzed β -lactam biodegradation of various antibiotics 20 4 NMR and MS Spectra 22 5 References 42 6 Appendix 43

Synthetic Schemes and Experimental Procedures
Scheme 1. Synthesis of probe 1.

3a. Evaluation of the selectivity of probes 1-3 towards H2S and RSH.
In this assay, we studied the detection selectivity of probes 1-3 towards H2S and RSH. Both probe 1 and 2 presented high selectivity for detection H2S in comparison to RSH, while probe 3 presented much lower selectivity ( Figure S1).

3b. Determination of the Limit of Detection (LOD) of probe 1 for detection of H2S:
In this measurement, the LOD of probe 1 was determined using various concentrations of NaSH. (LOD = 2.3 × 10 -7 M or 0.0128 µg/mL).

General assay protocol:
All antibiotics stocks solutions were prepared using DMSO or water according to their solubility. All enzyme stocks solutions were prepared according to their manufacturer solubility protocol. Two-step protocol was used to detect H2S released from antibiotics 3e. General procedure for the data given in Figure 8 (Detection of antibiotic resistant bacteria using probe 1).
Eight antibiotic resistant bacterial strains were inoculated from agar plates into culture tubes with nutrient broth (5 g/L peptone, 5 g/L NaCl; 2 g/L yeast extract, 1 g/L meat extract, pH 7.4) and incubated for 19 h at 37 °C and 150 rpm. Broth cultures were diluted to an optical density (600 nm) of 0.1 with sterile saline (0.9% NaCl). In wells of a white 96-well plate, 0.1 mL culture samples were mixed with 90 µL PBS and 10 µL Ceftizoxime or sulopenem stock solution (25 mM, final concentration 1 mM) and incubated for 30 min at room temperature. Then, 50 µL PBS with 0.125 mM of probe 1 (5x working solution, pre-incubated 1 h at room temperature) was added and luminescence (relative light units, RLU) was measured for 1 h at room temperature with a plate reader (500 ms integration time). Results of assay explained in the manuscript (Figure 8).

3f. Detection of carbapenem-resistant bacteria with Faropenem and probe 3.
Two imipenem (a carbapenem) resistant bacterial strains, Pseudomonas aeruginosa RKI 48/09 (VIM-15) and Klebsiella pneumonia RKI 92/08 (KPC-2), and a cephalosporinase-positive strain, Escherichia coli RKI 66/09 (CMY-2, AmpC), were inoculated from agar plates into culture tubes with nutrient broth and incubated for 22 h at 37°C and 150 rpm. In parallel, the two carbapenemase-positive strains were cultivated in nutrient broth supplemented with 8 and 4 mg/L imipenem (Imp), respectively. In addition, antibiotic sensitive strains Pseudomonas aeruginosa ATCC10145, Klebsiella pneumoniae RKI 2867/81and Escherichia coli ATCC 25922 were similarly cultivated in nutrient broth. Broth cultures were diluted to an optical density (600 nm) of 0.5 with sterile saline (0.9% NaCl). In wells of a white 96-well plate, 0.19 ml cell suspension samples were mixed with 10 µL Faropenem stock solution (20 mM in dimethyl sulfoxide, final concentration 0.8 mM) and 50 µL 5x working solution of probe 3 (0.125 mM in PBS, final concentration 25 µM). Luminescence was measured for 1 h at room temperature in a plate reader (500 ms integration time). Results are shown in Figure S5. 3g. General procedure for the data given in Figure 9 (one-step assay for antibiotic resistant bacteria with probe 1 and inhibition effect-obtained by 3-Aminophenylboronic acid).
Two strains expressing serine-type beta-lactamases, Klebsiella pneumonia RKI 92/08 (KPC-2) and Escherichia coli RKI 66/09 (CMY-2, AmpC) and two antibiotic sensitive reference strains, Klebsiella pneumonia RKI 2867/81 and Escherichia coli ATCC 25922 were inoculated from agar plates into culture tubes with nutrient broth and incubated for 19 h at 37°C and 150 rpm. Broth cultures were diluted to an optical density (600 nm) of 1.0 with sterile saline (0.9% NaCl). In wells of a white 96-well plate 90 µL PBS with 50 µM of 0.125 mM probe 1 (pre-incubated in PBS for 1 h at room temperature, final concentration 25 µM) was mixed with 100 µL cell suspension OD600 1.0 (final OD600 0.4) or sterile medium and 10 µL of 25 mM Cefazolin (final concentration 1 mM). In addition, 7 mM (final concentration) of the known serine betalactamase inhibitor 3-aminophenylboronic acid was added to a subset of wells. Immediately after all reagents had been added, luminescence (relative light units, RLU) was measured at room temperature for 15 min every min with a plate reader (500 ms integration time). Results are shown in the manuscript (Figure 9).

3h. Determination of the Limit of Detection (LOD) of probe 1 for the detection of antibiotic resistant bacteria using Ceftizoxime:
Strains cultivated in nutrient broth overnight at 37°C, cell density set to 10 9 CFU/mL by optical density, dilution series prepared in sterile saline. Assay in white 96-well plate:

LCMS method description
All samples were measured on Agilent 1260 system with single quadrupole MSD featured with multimode (ESI+APCI) ionization chamber. The ionization chamber parameters were as follows: The HPLC part of the device comprised a quadruple gradient pump, an headspace autosampler, Phenomenex Luna C8 5µm, 4.6x150 mm column and a DAD detector at 254 nm. The acidic elution method was a combination of 7 min gradient from 5% to 100% MeOH in 0.1% formic acid and subsequent 8min elution with pure MeOH. The acidic elution was used for both polarities as the betalactams open spontaneously under basic conditions.

Sample preparation for the LCMS assay
All antibiotics were purchased from commercial suppliers (sulopenem, cetazidime, imipenem, ceftizoxime from Biosynth-Carbosynth EN, the rest from Merck-Aldrich). A 10 mM stock solutions of inspected antibiotics in PBS (prepared by dissolving Merck-Aldrich P3813-1PAK in RODI water and set to 1L total volume) were made. Beta lactamase enzyme (Bacillus cereus 569/H9 from Aldrich, 426205-500U) stock solution was prepared by dissolving 1mg of dry enzyme in PBS and set by PBS to total volume of 1mL. Every LCMS sample was made by mixing 0.25 mL of antibiotic stock solution with 0.25mL of lactamase stock solution in a 2mL head space vial and incubation at RT (45min on linear shaker, 60 shakes per minute). 50 µL of the resulting solution was directly injected to LCMS after a period specified later. Stock solutions of ATB were directly injected in 10 µL loads to acquire data of starting material standards.

Faropenem decay promoted by B. Cereus metallo-lactamase, assayed by LCMS
The analysis with negative ionization was captured directly after the incubation period (injected after 45 min of incubation), positive ionization experiment started after 90 min of incubation and negative ionization experiment was repeated with identical result after 110 min of incubation.
Faropenem had hydrolysed completely during 45 min, furnishing 2 major peaks eluting at 4.08 and 4.47 min, 2 minor UV active peaks at 3.45 and 5.25 min (having only positive MS signal) and one large positive peak at 2.54 min with no UV absorption, which belongs to 2-amino-2-(hydroxymethyl)-1,3propanediol (m/z 122.1 in positive mode), the key part of TRIS buffer in which was the enzyme lyophilised by manufacturer (this peak is notable in all presented LCMS data and won´t be commented again). The two major peaks had almost identical MS spectra in both polarities. Here we list the spectra for described peaks: After inspecting all major ions detected, we have constructed major decomposition pathway for the faropenem. Major changes after opening the beta-lactam ring are a) decarboxylation, which is likely happening in the MS detector b) retro-aldol reaction. 1 Those major products are almost identical in its fragmentation and share almost the same molar weight. The proposed mechanism contains several intermediates with SH-group, capable to trigger a RSH sensing probe, and allows several ways to generate H2S from outlined thioacetal and thiohemiaminal structures.

Sulopenem decay promoted by B. Cereus metallo-lactamase, assayed by LCMS
The analysis with negative ionization was captured directly after the incubation period (injected after 45 min of incubation), positive ionization experiment started after 90 min of incubation and negative ionization experiment was repeated with identical result after 110 min of incubation. Sulopenem decomposed only partially during this period and had to incubate overnight (10h overall) to reach complete conversion. Following chromatogram shows the completed reaction mixture after 10 h.
Two major UV-absorbing peaks were detected in the hydrolysed mixture at 3.68 an 4.05 min. Negative ionization signal provided only these two peaks. Positive signal was also collected at 2.54 min, originating from the TRIS amine (m/z 122.1). Proposed decomposition pathway is roughly like the one for faropenem. The greatest difference opens the explanation of high detection rate of this antibiotic by RSH sensing probes: all the key intermediates can eliminate 2,3-dihydrothiophene 1-oxide via retro-Michael pathway and products are thiols, capable of reaction with the RSH sensing probe. 2,3-Dihydrothiophene-1-oxide signal was the strongest ion in the positive ionization MS spectrum for both major peaks, which strongly supports our explanation.

Penicillin V decay promoted by B. Cereus metallo-lactamase, assayed by LCMS
The analysis with negative ionization was captured directly after the incubation period (injected after 45 min of incubation at RT), positive ionization experiment started after 90 min of incubation. Penicillin V decomposed completely during this period. Only one major peak with higher polarity (6.51 min) was formed and remained untouched during the incubation. One minor peak was detected at 6.72 min. Fragment analysis is rather simple here: the hydrolysed beta lactam produces very stable compound, which then resists any further chemical changes. This is explaining very low luminescence of all tested beta-lactamase penicillin degradation mixtures with RSH sensing probe. The possible transformation through penillic acid, known from human metabolism is not likely in used reaction system and the key intermediate -the appropriate m/z for penillic acid -was not practically detected (the ion was present in very low concentration only). On the other hand, the ion with m/z 246 was developed in decent concentration, which speaks for good chance to produce H2S through decarboxylation/elimination pathway at elevated temperatures. This process can occur in limited scale at RT and furnish small, but detectable H2S levels, which then generated observed weak luminescent signal with H2S sensitive probe.

Meropenem decay promoted by B. Cereus metallo-lactamase, assayed by LCMS
The analysis with positive ionization was captured directly after the incubation period (injected after 45 min of incubation at RT), negative ionization experiment started 60 min after the incubation was started. Meropenem decomposed completely during this period. Only one major peak with slightly higher polarity was formed. Observed ions were corresponding to the expected hydrolytic pathway for all carbapenems: beta-lactam ring opening followed by retro-aldol elimination of acetaldehyde. The product then can undergo series of decarboxylations during the ionization process. Weak signal with RSH sensing probe can be explained by formation of the thiol intermediate with m/z 175.1 in the positive mode.

Cefotaxime decay promoted by B. Cereus metallo-lactamase, assayed by LCMS
The analysis with positive ionization was captured directly after the incubation period (injected after 50 min of incubation at RT), negative ionization experiment started 70 min after the incubation was started. Meropenem decomposed completely during this period. Only one broad major peak with slightly lower polarity was formed (5.22 min). MS signals of both polarities were very weak, which is given by zwitterion character of most molecules. Observed ions are corresponding with expected hydrolytic pathway: the acetoxy group eliminates along with the hydrolysis of the beta lactam ring. The gamma-lactone intermediate is easy detectable by its decarboxylated fragment ion (m/z 368 in acquired negative MS spectra). Under the positive ionization, m/z 201 and 101 are most remarkable, corresponding to the acylamide abstracted from the 7-position and its 2-aminothiazolyl fragment. Clearly, no formation of RSH or even H2S is allowed during this hydrolytic decomposition, which agrees with poor signal measured by thiol sensing luminescent probes. The analysis with positive ionization was captured directly after the incubation period (injected after 70 min of incubation at RT), negative ionization experiment started after 90 min after the incubation was started. Imipenem decomposed completely during this period. Only one major peak with slightly higher polarity was formed (2.94 min). According to MS signals, imipenem undergoes lactamase mediated transformation quite in accord with other penems: first the lactamase opens the lactam ring and then catalyses the retro-aldol reaction. Formed product is then fragmented in MS ionization chamber through NH3 eliminating steps forming isocyanide fragments, which in parallel decarboxylate. No product nor fragment possessing SH group was detected, which explains low signal after treatment with the RSH sensing probes. The analysis with positive ionization was captured directly after the incubation period (injected after 80 min of incubation at RT), negative ionization experiment started 100 min after the incubation was started. Ceftizoxime decomposed completely during this period. One major peak with lower polarity was formed (5.25 min) and one minor peak with slightly higher polarity (4.93 min). As the tested sample was the sodium salt of ceftizoxime, M+Na adducts were observed in higher intensity. Major decomposition pathway follows the cefotaxime scheme with one important exception: detected ions with m/z 243 and m/z 134 in the positive mode signal can be explained by the opening of the dihydrothiazine ring, when an aldehyde structure and corresponding enamine-thiol is formed. We believe this process can occur at lower temperatures as well and explains high signal strength with RSH sensing probes.
Ceftazidime decay promoted by B. Cereus metallo-lactamase, assayed by LCMS The analysis with positive ionization was captured directly after the incubation period (injected after 80 min of incubation at RT), negative ionization experiment started 100 min after the incubation was started. Ceftazidime decomposed completely during this period. One major peak with higher polarity was formed (3.92 min) and one minor peak with even higher polarity (3.18 min). MS signal of both polarities was weak, which was caused by zwitterion character of relevant structures, and heavily fragmented, which is resulting from fragile structure. Pyridinium ring eliminates readily after the lactam ring was opened and the oxime side chain is also steady for fast decarboxylation in the ionization chamber. One tentative thiol structure with m/z 146 was identified, which could eventually explain weak luminescent signal observed in this case.  Cefazoline decay promoted by B. Cereus metallo-lactamase, assayed by LCMS 50 min of incubation at RT), negative ionization experiment started 70 min after the incubation was started. Cefazoline decomposed completely during this period. Two small peaks were formed at The analysis with positive ionization was captured directly after the incubation period (injected after 50 min of incubation at RT), negative ionization experiment started 70 min after the incubation was started. Cefazoline decomposed completely during this period. Two small peaks were formed at 5.14 and 5.51 min. In this case the captured ions clearly display the mechanism of H2S release through hydrolytic opening of the dihydrothiazine ring, producing positive ions with m/z 263, 279 and 295. 5-Methyl-1,3,4-thiadiazole-2-thiol is another RSH component generated quickly after the lactam ring was