Blocking Peptidoglycan Recycling in Pseudomonas aeruginosa Attenuates Intrinsic Resistance to Fosfomycin

Bacterial strains and culturing conditions, plasmids, and reagents

Bacterial strains and plasmids used in this study are listed in Table 1. P. aeruginosa strain PA01 was obtained from Prof. Friedrich Götz, University of Tübingen. Strain PA14 and the complete nonredundant transposon insertion mutant library were obtained from the Broad Institute, Cambridge, MA, ¹⁸ which included amgK, murU, and anmK insertion mutants (07780::MAR2xT7, 07790::MAR2xT7, and 08520::MAR2xT7 clone 06_3F4, respectively). P. aeruginosa and E. coli strains were cultured at 37°C in Luria-Bertani (LB) liquid or solid (plus 15% agar) medium. For antibiotic susceptibility studies, cation-adjusted Mueller-Hinton broth (Sigma-Aldrich) liquid or solid medium was applied. E-test fosfomycin strips were obtained from BioMérieux, Craponne, France. Fosfomycin disodium salt powder and Irgasan were ordered from Sigma-Aldrich. Imipenem and ceftazidime powders were from Santa Cruz Biotechnology. Phusion high-fidelity polymerase, SmaI restriction enzyme, and E. coli DH5α cells were obtained from New England Biolabs. T4 DNA ligase, genomic and plasmid DNA isolation kits, and DNA gel extraction kits were from Thermo Scientific.

Construction of cell wall recycling knockout mutants

For the generation of markerless knockout mutants in strain PA01 of genes PA0596 (MurNAc kinase), PA0597 (MurNAc 1P uridylyl transferase), PA0666 (AnhMurNAc kinase), and PA3005 (N-acetylglucosaminidase NagZ), PA01 genomic DNA was isolated according to the manufacturer's instructions. Ten nanograms of genomic DNA was used to amplify ≈500 bp regions upstream and downstream of the target genes by PCR (primer pairs 1 and 2, respectively; for the primer list, see Table 2). DNA products were purified using a polynucleotide gel extraction kit and further spliced by PCR yielding ≈1,000 bp products. Subsequently, the plasmid pEX18Gm was digested with SmaI and blunt end-ligated overnight with the ≈1,000 bp PCR products using T4 DNA ligase. Reactions were transformed in chemically competent E. coli DH5α cells. ¹³ Transformants were then selected on LB agar plates supplemented with 10 μg/ml gentamycin (Gm). Recombinant plasmids were isolated from single DH5α transformants and the DNA sequences were checked by restriction analysis and DNA sequencing (Eurofins). The suicide plasmids (pEX18Gm ΔamgK, pEX18Gm ΔmurU, and pEX18Gm ΔanmK) were further transformed in chemically competent E. coli helper strain S17-1 λ pir and selected on LB agar plates with 10 μg/ml Gm. For diparental mating, 0.8 ml of overnight cultures (incubated at 37°C) of S17-1 λ pir cells, containing the suicide plasmids, and 1.7 ml PA01 wild-type cells, grown at 43°C, were mixed, centrifuged, and resuspended in 200 μl of final volume. The cell suspension was plated for 4 hr at 37°C on membrane filters on top of the LB agar plates to allow transfer of the suicide plasmids from E. coli into Pseudomonas PA01 wild-type strains during conjugation. Filters were resuspended in 0.9% NaCl and all bacteria were plated on LB supplemented with 60 μg/ml Gm and 25 μg/ml Irgasan (to kill E. coli cells). Successful PA01 merodiploids were growing in the presence of antibiotic and further selected for double crossover events on LB plates without salt and 5% sucrose. The generation of mutants was verified by the absence of growth in the presence of Gm, colony PCR with specific primers in the flanking regions (test primers) for PA0596, PA0597, and PA0666 gene deletions (Table 2).

LC-MS analysis of cell wall recycling intermediates

PA01 wild-type and respective recycling deletion mutants, as well as PA14 wild-type and the respective recycling transposon insertion mutants were grown overnight in 50 ml of LB medium. After 20 hr of incubation, bacterial cultures (OD_600nm ≈3.2) were centrifuged (10 min, 4,000 g) and pellets were dissolved in 10 ml of Tris buffer (10 mM, pH 7.6). After two washing steps, pellets were taken up in 400 μl of water and heated for 10 min at 95°C to cause cell breakage. Soluble cell extracts, containing the cell wall recycling products, were obtained by spinning down at maximum speed (12,000 g for 15 min). By adding 800 μl of acetone to each 200 μl of supernatant in an Eppendorf tube, protein precipitation was achieved and the remaining insoluble part was separated by centrifugation at maximum speed. The supernatant was transferred to a clean tube and dried in a speedvac at 60°C. The dried pellet was dissolved in 160 μl water before LC-MS measurement. Analysis of the accumulation of recycling products in P. aeruginosa mutants was conducted with a ESI-TOF mass spectrometer (MicrO-TOF II, Bruker) that was connected to the Ultimate 3000 HPLC system (Dionex), as previously described. ⁸ For high-performance liquid chromatography (HPLC) separation, a Gemini C18 column (Phenomenex, 150×4.6 mm, 110Å, 5 μM) was used. Samples (5 μl injection volume) were applied to the HPLC column at a flow rate of 0.2 ml/min and using a 45-min program (for 5 min, 100% buffer A: 0.1% formic acid with 0.05% ammonium formate, then 30 min of a linear gradient to 40% buffer B with 100% acetonitrile, 5 min of 40% buffer B, and 5 min of re-equilibration step). Data are presented as differential ion chromatograms (ΔTIC). These were obtained by subtracting the total ion chromatogram (TIC) of a mutant cell extract minus the TIC of the respective wild-type extract, calculated by automatically subtracting TICs using the program called “Metabolite Detect” (Bruker).

Determination of the MICs

MICs for P. aeruginosa strains PA01 and PA14 were determined by using the broth microdilution method, according to the Clinical and Laboratory Standards Institute (CLSI) recommendations or by applying E-test fosfomycin strips following the manufacturer's instructions. Briefly, the broth microdilution method was performed in U-form 96-well plates (VWR) using the Mueller-Hinton Broth (cation-adjusted) medium in a total volume of 100 μl and a shaking incubator at 35°C, 160 rpm for 20 hr in a humidified chamber. Bacteria were harvested at an optical density (OD_600nm) of 0.3 and diluted to a starter culture of 1×10⁴ cells. Appropriate dilution series of the antibiotic fosfomycin (FOS, 0.5–256 mg/ml), imipenem (IMP, 0.0625–8 mg/L), and ceftazidime (CAZ, 0.0625–8 mg/L) were previously diluted in the broth medium. The absence of growth at the lowest used antibiotic concentration was determined as the MIC value.

For E-tests, fosfomycin strips with a dilution range between 0.064 and 1,024 mg/ml and Mueller-Hinton broth (with Ca and Mg ions) agar plates (17.5 ml) were used. Pseudomonas overnight cultures were adjusted with 0.9% NaCl to 0.5 McFarland standard and bacteria were transferred with a sterile loop to the plates. Results were documented after incubation at 37°C for 20 hr.

Peptidoglycan recycling in P. aeruginosa involves a direct shortcut pathway to UDP-MurNAc

We first aimed to investigate whether peptidoglycan recycling in P. aeruginosa PA01 and PA14 involves orthologs of the newly discovered recycling genes amgK and murU of P. putida. If so, we expected that mutants for these genes will accumulate specific recycling intermediates, as was previously shown for P. putida. ⁸ Therefore, we generated markerless deletion mutants in PA01 of genes PA0596 (ΔamgK) and PA0597 (ΔmurU), as well as deletion mutants of known recycling genes, PA3005 (ΔnagZ) and PA0666 (ΔanmK). In addition, transposon insertion mutants were obtained from the PA14 mutant collection ¹⁸ of the respective genes of P. aeruginosa strain PA14 (except for PA14 nagZ mutant, which is not available in this collection). We demonstrated by LC-MS that MurNAc accumulates in the cell extract of amgK mutants (Fig. 2A) and MurNAc 1P accumulates in the cell extract of murU mutants (Fig. 2B), similar to observed accumulation products in P. putida orthologous mutants. ⁸ Moreover, AnhydroMurNAc was shown to accumulate in the anmK mutant strains (Fig. 2C). Thus, these results indicate that AmgK and MurU in P. aeruginosa constitute a shortcut recycling pathway, contributing to the UDP-MurNAc pool.

OPEN IN VIEWER

FIG. 2.

Identification of cell wall recycling products by LC-MS analysis. Total ion chromatograms (TICs) were obtained in the negative ion mode; the y axes represent ion intensity (counts per second). (A) Overlay of GlcNAc (m/z=220.09; colored red) and MurNAc (m/z=292.11), 5 mM each, and differential ion chromatograms (ΔTICs) obtained by subtracting TIC of mutant (PA01 ΔamgK or PA14 amgK::Tn) from the respective wild type. (B) Extracted ion chromatogram (EIC) for MurNAc 1P (m/z=372.07) and PA01 ΔmurU and PA14 murU::Tn ΔTICs. (C) TIC of AnhMurNAc (m/z=274.09), 1 mM, and PA01 ΔanmK and PA14 anmK::Tn ΔTICs.

Blocking peptidoglycan recycling in P. aeruginosa does not affect growth

Growth of Pseudomonas wild-type as well as amgK, murU, and anmK mutants was compared using the LB medium. Identical growth rates were observed for wild-type (PA01 and PA14) and respective recycling mutant strains (Fig. 3A, B), indicating that blocking cell wall recycling does not affect general fitness in a rich medium. Previous studies comparing wild-type and recycling mutants in E. coli and P. putida also revealed no difference in growth rates.^8,31 Since the peptidoglycan layer of E. coli, and presumably of P. aeruginosa as well, accounts for only about 2% of the cell mass, ²² Park and Uehara suggested that cell wall recycling is not essential for cell viability and growth. ³¹ Still, as shown by LC-MS results (Fig. 2), the recycling pathway through AnmK, AmgK, and MurU channels cell wall recycling fragments directly into the UDP-MurNAc pool and, thus, bypasses the MurA-catalyzed enzymatic reaction that is the target of fosfomycin (cf. Fig. 1). Cell wall recycling therefore contributes to the intrinsic fosfomycin resistance in P. aeruginosa.

OPEN IN VIEWER

FIG. 3.

Growth fitness comparison of P. aeruginosa strains in the LB medium. Optical densities (OD_600nm) are presented as base 10 logarithms (log10). (A) PA01 wild-type and respective deletion mutants (ΔamgK, ΔmurU, and ΔanmK). (B) PA14 wild-type and respective transposon (Tn) insertion mutants (amgK::Tn, murU::Tn, and anmK::Tn). Results are shown as a mean±SD of two independent experiments, each done in duplicate.

Pseudomonas cell wall recycling mutants are more susceptible to fosfomycin compared with wild type

The contribution of cell wall recycling to the UDP-MurNAc pool, and thus, the intrinsic resistance to fosfomycin, was shown by comparison of the MIC values for Pseudomonas wild-type and mutant strains (Table 3). By using the broth microdilution method, we observed that the MIC value for fosfomycin of 32 mg/L for the PA01 wild type was reduced to 8 mg/L in ΔamgK, ΔmurU, ΔanmK, and ΔnagZ deletion mutants. A fourfold increase in fosfomycin susceptibility in the absence of the indicated PA01 recycling genes was also seen by applying E-test strips; the MICs dropped from 64 mg/L in PA01 wild-type to 16 mg/L in the recycling mutants.

Overall, fosfomycin susceptibility of strain PA14 was two to three times higher compared with PA01. Still, fosfomycin sensitivity was similarly increased in the PA14 cell wall recycling mutants. Broth microdilution tests showed a MIC value of 16 mg/L for the PA14 wild-type strain, and eightfold decreased MICs (2 mg/L) in the PA14 anmK, amgK, and murU transposon insertion mutants. E-test strips confirmed these results. With this test, a MIC value of 24 mg/L for fosfomycin was determined in the PA14 wild type, and a value of 4 mg/L in the absence of recycling genes (a sixfold increase in fosfomycin sensitivity).

We wondered whether the fosfomycin sensitivity of cell wall recycling mutants is a specific or a general stress response. Therefore, we tested the susceptibility of PA01 wild type and ΔamgK, ΔmurU, ΔanmK, and ΔnagZ mutants toward the β-lactam antibiotics, ceftazidime (a 3rd generation cephalosporin) and imipenem (a carbapenem), both having inhibitory effects on P. aeruginosa. ²¹ We obtained identical MIC values (1 mg/L) for the two antibiotics in wild-type and mutant strains (Table 3). This indicates that blocking cell wall recycling does not generally affect susceptibility to antimicrobial drugs.

Application of fosfomycin in combinatory therapy

Spontaneous mutations in the cell wall recycling pathway are a common cause of acquired resistance to β-lactam antibiotics in P. aeruginosa clinical isolates. ⁴⁰ A specific cell wall recycling product, AnhMurNAc-pentapeptide, was shown to accumulate in these isolates, acting as an inducer of the chromosomally encoded AmpC β-lactamase.^14,21 Hyperproduction of AmpC was attenuated by blocking cell wall recycling at the level of NagZ, and thus, β-lactam susceptibility was restored.^14,40 We showed that the PA01 ΔnagZ mutant, similar to the other recycling mutants, is four times more susceptible to fosfomycin compared with wild type (Table 3). However, the sensitivity toward the β-lactams ceftazidime and imipenem was not affected in the ΔnagZ mutant (Table 3), in agreement with previous reports.^21,40 Nevertheless, in multidrug-resistant clinical isolates, inhibition of cell wall recycling may be applied in a combinatory therapy with fosfomycin to reduce its intrinsic resistance, and simultaneously, the acquired resistance to certain β-lactams.