Molecular mechanisms driving the in vivo development of OXA-10-mediated resistance to ceftolozane/tazobactam and ceftazidime/avibactam during treatment of XDR Pseudomonas aeruginosa infections

Abstract

Background

The development of resistance to ceftolozane/tazobactam and ceftazidime/avibactam during treatment of Pseudomonas aeruginosa infections is concerning.

Objectives

Characterization of the mechanisms leading to the development of OXA-10-mediated resistance to ceftolozane/tazobactam and ceftazidime/avibactam during treatment of XDR P. aeruginosa infections.

Methods

Four paired ceftolozane/tazobactam- and ceftazidime/avibactam-susceptible/resistant isolates were evaluated. MICs were determined by broth microdilution. STs, resistance mechanisms and genetic context of β-lactamases were determined by genotypic methods, including WGS. The OXA-10 variants were cloned in PAO1 to assess their impact on resistance. Models for the OXA-10 derivatives were constructed to evaluate the structural impact of the amino acid changes.

Results

The same XDR ST253 P. aeruginosa clone was detected in all four cases evaluated. All initial isolates showed OprD deficiency, produced an OXA-10 enzyme and were susceptible to ceftazidime, ceftolozane/tazobactam, ceftazidime/avibactam and colistin. During treatment, the isolates developed resistance to all cephalosporins. Comparative genomic analysis revealed that the evolved resistant isolates had acquired mutations in the OXA-10 enzyme: OXA-14 (Gly157Asp), OXA-794 (Trp154Cys), OXA-795 (ΔPhe153-Trp154) and OXA-824 (Asn143Lys). PAO1 transformants producing the evolved OXA-10 derivatives showed enhanced ceftolozane/tazobactam and ceftazidime/avibactam resistance but decreased meropenem MICs in a PAO1 background. Imipenem/relebactam retained activity against all strains. Homology models revealed important changes in regions adjacent to the active site of the OXA-10 enzyme. The bla_OXA-10 gene was plasmid borne and acquired due to transposition of Tn6746 in the pHUPM plasmid scaffold.

Conclusions

Modification of OXA-10 is a mechanism involved in the in vivo acquisition of resistance to cephalosporin/β-lactamase inhibitor combinations in P. aeruginosa.

Introduction

Selection of an appropriate treatment for infections caused by Pseudomonas aeruginosa is particularly challenging due to an outstanding ability to develop MDR through the selection of mutations in chromosomal genes or through the horizontal acquisition of broad-spectrum resistance mechanisms, such as ESBLs and MBLs.^1–3

However, the recent introduction into the clinical setting of the novel cephalosporin/β-lactamase inhibitor combinations ceftolozane/tazobactam and ceftazidime/avibactam has represented a step forward in treating MDR/XDR P. aeruginosa infections. Despite the fact that these novel agents are not active against MBL-producing P. aeruginosa isolates, they retain activity against isolates that owe their β-lactam resistance to the most relevant mutation-driven β-lactam resistance mechanisms or to the production of certain β-lactamases and are therefore increasingly recognized as the most suitable antimicrobials for the treatment of infections caused by MDR/XDR P. aeruginosa strains.⁴ Ceftolozane/tazobactam is a novel oxyiminoaminothiazolyl cephalosporin equipped with a large and highly functionalized pyrazole moiety at the 3-position. This moiety confers increased stability against P. aeruginosa strains showing AmpC overproduction. Moreover, active efflux and OprD inactivation appear to have no impact on ceftolozane/tazobactam MICs.⁵ The combination of ceftolozane with the β-lactamase inhibitor tazobactam expands the spectrum of this new cephalosporin against strains producing TEM- and SHV-like class A ESBLs, which are still infrequent in P. aeruginosa,⁶ but not against class D OXA-type ESBLs, since these latter also hydrolyse broad-spectrum cephalosporins, but are resistant to common β-lactamase inhibitors.⁷ Similarly, ceftazidime/avibactam is a novel combination with robust antipseudomonal activity. Addition of the diazabicyclooctane inhibitor avibactam potentiates the activity of ceftazidime against chromosomal AmpC overproduction, as well as against the majority of the class A ESBLs most frequently encountered in P. aeruginosa, such as GES-, VEB- and BEL-like enzymes.⁶

However, P. aeruginosa is a continuously moving target and the growing catalogue of reports documenting the emergence of resistance to these novel combinations during therapy is of concern.^8,9 Overproduction and structural modification of the chromosomal cephalosporinase AmpC appears to be the main mechanism driving clinical resistance to ceftolozane/tazobactam and ceftazidime/avibactam in vivo.¹⁰ In previous studies, we observed that cross-resistance to these new antimicrobials may also emerge during therapy through the selection of extended-spectrum mutations in acquired OXA-type β-lactamases when MDR P. aeruginosa strains carrying oxacillinases such as OXA-2 or OXA-10 are targeted with cephalosporins.^11,12 As observed for AmpC variants selected upon ceftolozane/tazobactam exposure, the evolved extended-spectrum OXA mutants typically exhibit classic (e.g. Gly157Asp; OXA-14) or more recently described (e.g. ΔIle159-Glu160Lys; OXA-681) amino acid replacements or deletions in the cephalosporin-binding pocket that confer enhanced cephalosporinase activity, but impaired hydrolysis of other substrates such as penicillins and carbapenems.¹³ Nevertheless, despite the fact that OXA-2 and OXA-10 enzymes have been extensively involved in the in vivo emergence of ceftazidime resistance,^14–16 the specific role of these β-lactamases in a mechanism for the development of resistance to ceftolozane/tazobactam and ceftazidime/avibactam during the treatment of MDR/XDR P. aeruginosa infections remains poorly characterized.

Between September 2018 and March 2019, four patients admitted to the ICU of a third-level teaching hospital developed severe invasive infections caused by XDR P. aeruginosa. In all cases, the P. aeruginosa strains treated with cephalosporins developed cross-resistance to ceftazidime, ceftolozane/tazobactam and ceftazidime/avibactam during therapy. Here, we performed a detailed comparative analysis of the mutational and functional profile of the susceptible/resistant isolates in order to decipher the underlying mechanisms responsible for the emergence of resistance to these last-line antimicrobials.

Materials and methods

Clinical strains

Four consecutive pairs of ceftolozane/tazobactam- and ceftazidime/avibactam-susceptible/resistant P. aeruginosa isolates recovered from clinical specimens obtained from four patients admitted to the ICU of the Puerta del Mar University Hospital (Cádiz, Spain) between September 2018 and March 2019 were evaluated. The susceptible isolates were recovered before the start of antimicrobial therapy, whereas the resistant isolates were recovered during or after therapy.

Antimicrobial susceptibility testing

MICs of ticarcillin, piperacillin, piperacillin + 4 mg/L tazobactam, ceftazidime, ceftazidime + 4 mg/L avibactam, cefepime, aztreonam, ceftolozane, ceftolozane + 4 mg/L tazobactam, imipenem, imipenem + 4 mg/L relebactam, meropenem, tobramycin, amikacin, ciprofloxacin and colistin were determined in triplicate using the reference broth microdilution method. EUCAST v.10.0 clinical breakpoints and guidelines (http://www.eucast.org/clinical_breakpoints/) were used for reference purposes.

WGS

Isolates were sequenced in parallel using short-read (Illumina NovaSeq, Illumina) and long-read (MinION, Oxford Nanopore Technologies) approaches. The Illumina indexed paired-end libraries were generated from a TruSeq Nano DNA Library Prep Kit (Illumina). The Oxford Nanopore library was constructed using a commercial SQK-RBK004 Rapid Barcoding Kit (Oxford Nanopore Technologies). The resulting short and long reads from each isolate were assembled together using the Unicycler v0.4.6 hybrid assembler with default settings.¹⁷ The high-quality hybrid assemblies thus obtained were annotated using the Rapid Annotations Subsystem Technology (RAST) server.¹⁸

Resistance genomics

To examine the mutational resistome of the isolates, the reads obtained for each isolate were mapped against the genome of the reference P. aeruginosa UCBPP-PA14 strain (RefSeq accession number: NC_0084643) and sequence variation was analysed for the 146 chromosomal genes related to antimicrobial resistance, as previously described.¹⁹ The presence of horizontally acquired resistance mechanisms was explored using the Comprehensive Antibiotic Resistance Database (CARD) webtool (https://card.mcmaster.ca/home) and the ResFinder v 3.2 server (https://cge.cbs.dtu.dk//services/ResFinder/).

Expression studies

Bacterial RNA was isolated using a High Pure RNA Isolation Kit (Roche Diagnostics). RNA samples were purified using the RNeasy MinElut Cleanup Kit (QIAGEN) and quantified in a spectrophotometer (BioDrop 1LITE, Isogen Life Science). Primers targeting ampC were designed and expression was determined in triplicate experiments by RT–PCR in a LightCycler 480 RNA instrument (Roche), as previously described.²⁰ The expression level was standardized relative to the transcription level of the housekeeping gene rpoB.

Molecular typing

The clonal relatedness of the isolates was investigated by MLST. STs were deduced from WGS data according to the allelic profiles available in the MLST database (http://pubmlst.org/paeruginosa).

Cloning of OXA-10 mutants

bla_OXA-10 genes of the clinical isolates were amplified in parallel with the primer pair OXA-10-F-EcoRI (5´-CCGGAATTCCGGGTTAGGCCTCGCCGAAGC-3´) and OXA-10-R-BamHI (5´-CGGGATCCCGTTAGCCACCAATGATGCCC-3´), ligated to the pUCP24 plasmid and cloned into Escherichia coli TG1. Subsequently, recombinant pUCP24-OXA-10 plasmids were electroporated into the reference strain PAO1 and plated on 30 mg/L gentamicin/LB agar plates. The MICs of β-lactams were determined for the PAO1 transformants according to the above-described methodology.

Building of extended-spectrum OXA-10 variants and homology models

The homology models for the OXA-10-like enzymes were constructed using the SWISS-MODEL Homology Server.²¹ The following crystallographically determined structures of OXA-10 β-lactamase isolated from P. aeruginosa were used as templates for constructing the models: OXA-10 (PDB 6SKQ, chain A)²² for OXA-14 (OXA-10_Gly157Asp); OXA-10_Trp154Ala variant enzyme covalently modified by benzylpenicillin (PDB 2WGI, chain A)²³ for OXA794 (OXA-10_Trp154Cys); OXA-10_Trp154His variant enzyme (PDB 2RL3, chain A)²³ for OXA-795 (OXA-10_{ΔPhe153-Trp154}); and OXA-10 (PDB 6SKR chain A)²² for OXA-824 (OXA-10_Asn143Lys). The resulting OXA-10-like structures exhibited average nucleotide identities of 98.87%, 99.60%, 100% and 99.60%, respectively. The final figures depicting the major structural features of the new variants relative to the parental OXA-10 enzyme were elaborated using PyMOL software.²⁴

Genetic context of β-lactamases

The genetic location of β-lactamase genes was assessed from WGS using the PlasmidFinder tool,²³ the MOB typing tool²⁵ and plasmidSPAdes software. Moreover, the plasmid contig bearing the bla_OXA-10 gene was compared with other plasmid sequences available in the NCBI database and visualized using Geneious v.2020.0.5 (Biomatters).

Nucleotide accession numbers

The WGS data for the strains included in this work have been deposited in the NCBI BioProject database under accession number PRJNA523008. The nucleotide sequences of the new OXA-type β-lactamases described in this work have been deposited in the GenBank database under accession numbers MK482719 (bla_OXA-794), MK482720 (bla_OXA-795) and MK750001 (bla_OXA-824).

Results and discussion

Clonal background, antimicrobial susceptibility and genotypic features of the clinical isolates

The demographic, antimicrobial susceptibility and genotypic data for the paired susceptible/resistant isolates included in the study are summarized in Table 1. The length of treatment ranged from 8 to 21 days, with an average duration of 15 days, and always included a cephalosporin (usually ceftazidime or ceftolozane/tazobactam) in combination with other antipseudomonals. In all cases, the susceptible/resistant isolates were clonally related and were assigned by MLST to the high-risk international clone ST253 (PA14-like), characterized by increasing dissemination in different European countries, a strong association with transferable resistance mechanisms and high virulence due to production of the ExoU cytotoxin.²⁶

Analysis of the susceptibility data revealed that in all cases the initial isolates showed an XDR susceptibility profile and were resistant to ticarcillin, piperacillin/tazobactam, cefepime, aztreonam, imipenem, meropenem, tobramycin and amikacin, but retained susceptibility to ceftazidime, ceftolozane/tazobactam, ceftazidime/avibactam, colistin and also to the recently available combination imipenem/relebactam. Analysis of putative chromosomal mutations that could explain the antimicrobial susceptibility profile of these isolates revealed a premature stop codon in the oprD gene, leading to carbapenem resistance, and the canonical quinolone-resistance mutations Thr83Ile in GyrA and Ser87Trp in ParC . Moreover, analysis of horizontally acquired resistance determinants revealed that all isolates produced the acquired oxacillinase OXA-10, which inefficiently hydrolyses ceftazidime, possesses weak carbapenemase activity against meropenem²⁷ and is relatively frequent among P. aeruginosa isolates worldwide.^28,29

On the other hand, the isolates recovered during/after treatment also had an XDR phenotype, but had additionally developed high-level resistance to ceftazidime, ceftolozane/tazobactam and ceftazidime/avibactam, although a variable decrease was noted for the MICs of the antipseudomonal penicillins, aztreonam, cefepime and, particularly, meropenem. Interestingly, the new imipenem/relebactam combination and colistin were the most active drugs and retained activity against the whole set of evolved mutants. Comparative genomic analysis of the isolates recovered before and during/after cephalosporin treatment failed to identify any mutation previously associated with cephalosporin resistance in efflux pump regulators (e.g. mexR, mexZ), ampC regulators (e.g. dacB, ampR, ampD, mpl) or the ampC gene itself, thus indicating that the resistance was not caused by the usual chromosomal mutations.³⁰ Moreover, comparative real-time quantitative reverse transcription experiments targeting the ampC gene revealed the absence of significant differences in the levels of ampC expression between the isolates obtained before and during/after treatment (Table S1 and Figure S1, both available as Supplementary data at JAC Online), indicating that overexpression of ampC was not involved in the emergence of resistance. However, a comprehensive analysis of the horizontally acquired resistance determinants revealed that in all cases the four resistant isolates had developed different substitutions or deletions in the bla_OXA-10 gene; the classic extended-spectrum OXA-14 (Gly157Asp) variant in one case,³¹ and three new derivatives in the other cases, designated OXA-794 (OXA-10_Trp154Cys), OXA-795 (OXA-10_{ΔPhe153-Trp154}) and OXA-824 (OXA-10_Asn143Lys). These new OXA-10 variants were curated by NCBI staff and named according to the NCBI bacterial antimicrobial resistance reference gene database.

Characterization of OXA-10 mutants

A closer look at the amino acid sequence of the OXA-10 enzymes revealed that all the derivatives present in the resistant isolates exhibited amino acid replacements or deletions that mapped within or in the vicinity of a 10 amino acid region designated the Ω-loop (Figure 1), which is a hot-spot site for mutations able to modify the substrate specificity of β-lactamases from different classes.³² Remarkably, OXA-14 showed the classic glycine-to-aspartate replacement at position 157, a change shared with other natural OXA-10-like extended-spectrum variants identified in ceftazidime-resistant P. aeruginosa isolates more than 20 years ago, such as OXA-11 and OXA-16.^33,34OXA-794 displayed a mutated cysteine instead of tryptophan at position 154. Interestingly, this amino acid replacement differs from any previously found in naturally occurring OXA-10-like enzymes able to confer ceftazidime resistance. However, a cephalosporin-resistant laboratory mutant with the same position mutated, albeit differently, was obtained from OXA-10-producing transconjugants of P. aeruginosa challenged in vitro with ceftazidime.³⁵ By contrast, OXA-824 had a previously undescribed lysine instead of the asparagine at position 143. Of note, the asparagine at 143 is part of the highly conserved Tyr-Gly-Asn motif,²⁸ which is also affected in the classic extended-spectrum OXA-11, which also has the Gly157Asp replacement. Finally, the new OXA-795 variant showed a dual Phe153-Trp154 deletion, also located in the Ω-loop and including the key residue Trp154.

In order to precisely determine the impact of the OXA-10 variants on β-lactam resistance, they were cloned in parallel with the WT OXA-10 in a PAO1 background. Comparative MIC data for the PAO1 transformants expressing the parental OXA-10 and the four evolved derivatives are shown in Table 2. Compared with the parental OXA-10, the different OXA-10 mutants generated an increase in the MIC, which ranged from 32- to 256-fold for ceftazidime, from 16- to 128-fold for ceftazidime/avibactam and from 4- to 8-fold for ceftolozane/tazobactam. These results provide good evidence that the evolved resistance to ceftazidime and cephalosporin/β-lactamase inhibitor combinations was, in all cases, caused by changes in the architecture of the OXA-10 enzyme. More specifically, the greatest impact on cephalosporin MICs was noted for the transformants expressing OXA-14 and OXA-795, whereas a weaker effect was noted for the transformant expressing OXA-794. On the other hand, mitigated effects on the MICs of penicillins and carbapenems were observed for all of the mutants, thus indicating that the enhanced cephalosporinase activity is detrimental to the natural penicillinase and carbapenemase activity of the natural OXA-10 enzyme. Interestingly, the increase in cephalosporinase activity associated with decreased carbapenemase activity has previously been noted in other β-lactamase mutants selected under prolonged treatment courses with cephalosporin-containing regimens, such as Asp179Tyr mutants of KPC-3³⁶ or Glu247Lys mutants of P. aeruginosa AmpC.³⁷ Finally, the production of either OXA-10 or the extended-spectrum derivatives did not affect the MIC of the imipenem/relebactam combination, which showed the highest activity against all the transformants. As observed with clinical isolates, these findings indicate that OXA-10-like enzymes do not alter imipenem/relebactam susceptibility and highlight that potentiation of the activity of imipenem in the presence of relebactam against all isolates is probably due to the great inhibitory activity of relebactam against the intrinsic P. aeruginosa AmpC.³⁸

Homology models

Homology modelling was conducted in order to understand the structural effect of the different amino acid replacements found in the extended-spectrum OXA-10 derivatives (Figure 2). This structural approach revealed that the enhanced cephalosporinase activity is probably driven by relevant changes in the conformation of the Ω-loop and the regions adjacent to the active site. The crystal structure of the native OXA-10 from P. aeruginosa (PDB 1K55, 1.39 Å)³⁹ is shown and illustrates the following: (i) the arrangement and interactions of the carbamylated form of the catalytic Lys70 residue (KCX70) (Figure 2a),^40,41 which acts as a general base in both the activation of the catalytic serine in the acylation process and the subsequent hydrolysis of the enzyme adduct;²³ (ii) the interaction of KCX70 with Trp154, which is highly conserved among OXA enzymes, fixes the position of KCX70 in an optimal arrangement for catalysis;⁴² (iii) Phe153 and Leu155, which are all located on the same face of the Ω-loop (residues 147–156), near Trp154, enhance further this freezing effect of Trp154 by completely shielding this part of the enzyme from the environment (the overall arrangement reduces the flexibility of the Ω-loop, which controls the accessibility of the active site⁴⁰); and (iv) the conserved Tyr141–Gly142–Asn143 motif, which precedes the Ω-loop and probably stabilizes the neighbouring α-helix (α7).⁴³

In the OXA-14 enzyme (Figure 2b),³¹ Gly157 is replaced by a negatively charged aspartate residue, which induces reorganization of the electrostatic and hydrogen bonding interactions between the residues located within the Ω-loop, mainly Lys152 and Glu156. This change leads to a conformational change in the Ω-loop that presumably alters the accessibility of the cephalosporins to the active site. A similar arrangement is seen in OXA-794 (Figure 2c), in which replacement of the Trp154 by a cysteine residue is expected to induce a change in the Ω-loop folding through additional polar contacts with the carbonyl main chain of Ala66, which is located close to the catalytic serine residue. A more pronounced effect on the Ω-loop arrangement is expected to occur in the OXA-795 variant with the dual deletion of Phe153 and Trp154, in which both the length and lipophilic character of the vicinity of the active site are expected to be reduced (Figure 2d). In addition, the deletion of these bulky residues would create a pocket in the vicinity of the active centre, altering the access of the substrates to the active site. Finally, in the OXA-824 variant, the replacement of Asn143 (neutral) in the Tyr141–Gly142–Asn143 motif by a positively charged residue (Lys143) would provoke the reorganization of the hydrogen bonding interactions in the loop, as well as the establishment of electrostatic interactions with the neighbouring Glu62 (Figure 2e). This residue is located in the loop that connects the β3 strand and the α2 helix (catalytic pocket). Both of these changes are likely to alter the accessibility of the active site.

Genetic context of β-lactamases

Analysis of the high-quality hybrid assemblies obtained by combining long- and short-read WGS data revealed that the bla_OXA-10 gene was located on a 29 634 bp plasmid, which comprised 34 ORFs and was designated pHUPM (Figure 3a). Comparative analysis of the sequence of the pHUPM plasmid and other plasmid sequences in NCBI databases revealed high similarity to the pJB12 plasmid (RefSeq accession number: KX889311), a 30 361 bp plasmid previously found to carry the gene coding for the MBL VIM-2 (i.e. bla_VIM-2) in a high-risk ST175 P. aeruginosa clone from Portugal.⁴⁴

The pHUPM and the pJB12 plasmids shared an identical plasmid backbone, which included an arsenic resistance operon and the genes involved in plasmid replication, conjugative transfer and maintenance functions. However, a closer look at the mobile region revealed that the composite Tn6352 transposon present in the pJB12 plasmid was replaced in pHUPM by a novel complex transposable element, designated Tn6746 (Figure 3b). This novel transposon also comprised ISPa17, a class 1 integron and a Tn402-like structure, but it was embedded in a different orientation in the pJB12-like plasmid backbone. The Tn6746 harboured the integron-encoded antibiotic resistance gene cassette bla_OXA-10-aacA4-qacEΔ1-sul1 and exhibited small tandem repeats flanking the ISPa17 and the Tn402-like transposon, thus suggesting that the bla_OXA-10 gene was probably acquired by transposition of Tn6746 in the pJB12-like plasmid pHUPM.

Conclusions

Production of carbapenemases, and particularly MBL production, is the most frequent cause of primary resistance to ceftolozane/tazobactam and ceftazidime/avibactam in P. aeruginosa. However, the most common pathway for the development of ceftolozane/tazobactam and ceftazidime/avibactam resistance during the treatment of MDR/XDR P. aeruginosa infections involves the stepwise accumulation of chromosomal mutations first, leading to overproduction, and then structural modification of the intrinsic cephalosporinase AmpC. In this study, we provide a molecular insight into the role of the modification of the classic OXA-10 β-lactamase, which is relatively common in P. aeruginosa, as an alternative emerging mechanism for in vivo development of cross-resistance to all of the commercially available cephalosporins. Our findings show that treatment with classic (ceftazidime) or new (ceftolozane/tazobactam) cephalosporins may facilitate the selection of mutations leading to important conformational changes in the Ω-loop and adjacent regions of the OXA-10 enzyme and resulting in a single-step high-level cephalosporin resistance response. As previously observed in other β-lactamase variants selected on cephalosporin exposure, production of these OXA-10 mutants resulted in partial restoration of meropenem MICs, even in the presence of carbapenem resistance mechanisms (OprD deficiency). Although caution is encouraged, these findings suggest that the evolved oxacillinase-mediated cephalosporin resistance is paired with the restoration of a potential carbapenem-based therapy, thus paving the way for the potential benefit of a meropenem/cephalosporin combination for the treatment of infections caused by MDR/XDR P. aeruginosa strains carrying extended-spectrum OXA-10-like enzymes.

From a therapeutic point of view, another finding probably of more interest is the increased stability of the combination of imipenem and relebactam against the entire set of ceftolozane/tazobactam- and ceftazidime/avibactam-susceptible/resistant isolates included in this study. This finding indicates that this combination is a very valuable addition to our currently available antipseudomonal arsenal. Finally, the fact that the bla_OXA-10 gene is located on a transposable mobilization platform embedded on a plasmid and carried by a virulent high-risk ST253 international clone adds further concern to our findings and alerts to the potential spread of this resistance mechanism. Altogether, our findings emphasize the need to maintain active surveillance of emerging broad-spectrum resistance in P. aeruginosa.

Acknowledgements

Professor Patrice Nordmann (University of Fribourg, Fribourg, Switzerland) is gratefully acknowledged for providing an aliquot of the pUCP24 plasmid used in the cloning experiments. We are also grateful to MSD for providing relebactam powder.

Funding

This work was supported by MSD through the Investigator Initiated Studies Program, which provided financial support and relebactam. This work was also supported by the Ministerio de Economía y Competitividad of Spain, Instituto de Salud Carlos III (ISCIII), through grants PI15/00860 and PI18/00501 to G.B., PI18/00076 to A.O. and P14/00059 and P17/01482 to A.B. Support was also provided by: Planes Nacionales de I + D+i 2013-2016 and ISCIII, Subdireción General de Redes y Centros de Investigación Cooperativa, Ministerio de Economía y Competitividad, Spanish Network for Research in Infectious Diseases (REIPI RD16/0016/006) co-financed by the European Development Regional Fund ‘A way to achieve Europe’; and the operative Intelligent Growth Program 2014-2020. C.G.-B. acknowledges financial support from the Spanish Ministry of Economy and Competiveness (SAF2016-75638-R), the Xunta de Galicia [ED431B 2018/04 and Centro singular de investigación de Galicia accreditation 2019-2022 (ED431G 2019/03)] and the European Regional Development Fund (ERDF). J.A.-S. was financially supported by the Fundación Profesor Novoa Santos through a Post-Especialización Grant (2019) and by the Rio Hortega Program (ISCIII, CM19/00219). J.C.V.-U. was financially supported by the pFIS Program (ISCIII, PI17/01482). C.L.-M. was financially supported by an IN606A-2019/029 Grant (Xunta de Galicia).

Transparency declarations

MSD provided financial support and relebactam; however, MSD did not exercise control over the conduct or reporting of the research.

Supplementary data

Table S1 and Figure S1 are available as Supplementary data at JAC Online.

References