A high-efficiency scar-free genome-editing toolkit for Acinetobacter baumannii

Abstract

Background: The current mutagenesis tools for Acinetobacter baumannii leave selection markers or residual sequences behind, or involve tedious counterselection and screening steps. Furthermore, they are usually adapted for model strains, rather than for MDR clinical isolates.

Objectives: To develop a scar-free genome-editing tool suitable for chromosomal and plasmid modifications in MDR A. baumannii AB5075.

Methods: We prove the efficiency of our adapted genome-editing system by deleting the multidrug efflux pumps craA, cmlA5 and resistance island 2 (RI2), as well as curing plasmid p1AB5075, and combining these mutations. We then characterized the susceptibility of the mutants compared with the WT to different antibiotics (i.e. chloramphenicol, amikacin and tobramycin) by disc diffusion assays and determined the MIC for each strain.

Results: We successfully adapted the genome-editing protocol to A. baumannii AB5075, achieving a double recombination frequency close to 100% and routinely securing the construction of a mutant within 10 working days. Furthermore, we show that both CraA and p1AB5075 are involved in chloramphenicol resistance, and that RI2 and p1AB5075 play a role in resistance to amikacin and tobramycin.

Conclusions: We have developed a versatile and highly efficient genome-editing tool for A. baumannii. We have demonstrated it can be used to modify both the chromosome and native plasmids. By challenging the method, we show the role of CraA and p1AB5075 in antibiotic resistance.

Figure 1.

Schematic representation of plasmids pEMGT and pSW-Apr. All relevant features borne in each plasmid are presented and named. SceI target sites in pEMGT are circled in dotted lines. Adapted from ‘Custom Plasmid Maps 2’, by BioRender.com (2022). Retrieved from https://app.biorender.com/biorender-templates. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Figure 2.

Schematic outline of the genome-editing strategy adapted for A. baumannii AB5075 applied to the deletion of craA. Plasmid features are represented in Figure 1. When indicated, LB agar plates were supplemented with ampicillin 100 mg/L (Ap), apramycin 200 mg/L (Apr) and/or tellurite 30 mg/L (Tel). For confirmation of craA deletion, colony PCR was performed using primers craA fw seq and craA down rv. As controls, WT AB5075 (WT) and pEMGT-craA (p) were used. M, DNA molecular weight marker, with band sizes indicated (kb). For simplicity, only the events occurring if the first recombination happened in the upstream homologous region is shown. Created with BioRender.com. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Figure 3.

PCR analysis to confirm the ΔcmlA5 deletion and assess the presence of p1AB5075. Genomic DNA extracted from the respective strain was used as template. For confirming the deletion, primer pair cmlA5 up fw/cmlA5 rv seq was used, giving bands of 2.79 kb for the WT and 1.55 kb for the ΔcmlA5 deletion mutant. The presence or absence of p1AB5075 was assessed with primer pairs oSA67/oSA68 and oSA86/oSA87, which would give PCR products of 0.7 and 0.16 kb, respectively. In the case of the p1AB5075-cured strain, no amplification was observed for any of the primer pairs. Scission of RI2 from p1AB5075 (ΔRI2), supported by Figure S6, explains the absence of PCR product using the primers to detect cmlA5 and compared with the amplification with primers to confirm the presence of p1AB5075. M, DNA molecular weight marker, with band sizes (kb).

Figure 4.

Quantification of antibiotic resistance and mucoid phenotype of the multiple mutant strains compared with AB5075. Quantifications were performed by DDAs using chloramphenicol (50 µg), amikacin (30 µg) and tobramycin (10 µg) discs according to the experiment. (a) Chloramphenicol resistance was measured for all mutant strains generated and compared with that of the WT AB5075. The susceptibility phenotypes observed for the ΔcraA and Δp1AB5075ΔcraA mutants were complemented by reintroduction of the craA coding sequence as compared with the WT and the parental strains bearing the control construction (see Materials and methods). (b) In the case of the mutants affecting RI2 or the whole p1AB5075 plasmid, as well as their combinations with the ΔcraA mutation, resistance was also assessed for the aminoglycosides amikacin and tobramycin. (c) For the mutants that produced it, the mucoid zone observed around chloramphenicol discs was measured and compared with the zone of inhibition using the WT strain (no mucoid zone formed) as control. The average zone of inhibition in millimetres (mm) measured from three biological replicates ( ±SD) is shown. Statistical significance was assessed from P values obtained from a t-test (* = P ≤ 0.05, ** = P ≤ 0.01, **** = P ≤ 0.0001).