Contributions of plasmid p1AB5075-encoded antibiotic resistance genes to multidrug resistance of Acinetobacter baumannii AB5075

Abstract

Infections caused by multidrug-resistant Acinetobacter baumannii are considered a threat to human and animal health. The widely studied A. baumannii strain AB5075 displays a high degree of antibiotic resistance. In this study, we experimentally validated that antibiotic resistance is largely mediated by resistance genes located on plasmid p1AB5075. We used a p1AB5075-deficient AB5075 strain to assess individual contributions of p1AB5075-encoded antibiotic resistance genes by ectopically (over-)expressing each gene in the Δp1AB5075 background. By determining individual contributions of seven p1AB5075-encoded antibiotic resistance genes, we show individual and overlapping roles of genes for aminoglycoside resistance and uncover the importance of extended-spectrum β-lactamase bla_GES-11 for monobactam and cephalosporin resistance in A. baumannii AB5075. We discovered that aminoglycoside N-acetyltransferase aaC(6′)-Ib3 (aacA4), which was previously shown to confer resistance to tobramycin, provides broad resistance to gentamicin, kanamycin, amikacin, streptomycin and tobramycin when overexpressed in A. baumannii AB5075. Because p1AB5075 is transferable to a wide range of environmental and clinical A. baumannii strains and non-baumannii Acinetobacter species, the relevance of our findings extends beyond A. baumannii AB5075.

Introduction

Curing infections caused by multidrug-resistant (MDR) micro-organisms in humans and animals is one of the greatest medical challenges of the twenty-first century. Among the most problematic bacteria that require investment into research and development of new antibiotics or alternative treatment strategies are carbapenem-resistant Acinetobacter baumannii [1–3]. Infections with A. baumannii are increasingly difficult to clear because of a high degree of antimicrobial resistance (AMR), strain heterogeneity and reports of pan-drug-resistant strains [4–6]. Understanding the evolution of antimicrobial resistance and resistance mechanisms may inform the development of new drugs and better treatment regimes. Early isolates that have been extensively used to study A. baumannii biology were typically susceptible to antibiotics, including the widely studied strains A. baumannii ATCC17978 and ATCC19606 [7]. Contemporary A. baumannii strains, including the widely studied A. baumannii AB5075, are likely more suitable representatives of current infections and are often resistant to a large array of antibiotics [8].

In A. baumannii AB5075, many antibiotic resistance genes (ARGs) are located on plasmid p1AB5075 within ‘Resistance Island 2’ (RI-2), but some are also found on the chromosome (e.g. blaOXA-23) [8, 9]. The plasmid p1AB5075 is the largest of three plasmids in A. baumannii AB5075 (83,160 bp), which was shown to be transferable to other A. baumannii strains of environmental and clinical origin and to diverse non-baumannii Acinetobacter species [8–11]. The RI-2 locus is chiefly responsible for aminoglycoside (hetero-)resistance in A. baumannii AB5075, which is caused in part by a RecA-dependent resistance island amplification to up to 20–24 copies [12]. The RI-2 small RNA SrvS located upstream of aadB was also shown to be involved in virulent opaque (VIR-O) to avirulent translucent colony (AV-T) type switching, showing that p1AB5075-encoded genes are engaging in regulatory cross-talk between p1AB5075 and the chromosome [13]. RI-2 is mosaic and contains, among other ARGs, an extended-spectrum β-lactamase bla_GES-11, which was first described in an A. baumannii strain isolated in France [14].

Despite their biological importance, spontaneous loss of large plasmids, including p1AB5075, has been documented for A. baumannii [5]. Plasmid pAB3 has been reported to have been lost in A. baumannii ATCC17978, which caused up-regulation of the Type-6 Secretion System [15, 16], and loss of p1AB5075 in A. baumannii AB5075 resulted in susceptibility to amikacin and tobramycin and reduced resistance to chloramphenicol [13, 17]. Here, we report loss of p1AB5075 in A. baumannii AB5075 independent of the published reports [13, 17]. We used the Δp1AB5075 genetic background to characterize the antibiotic resistance profile for the p1AB5075-deficient strain and genetically dissect the contributions of seven individual p1AB5075-encoded antibiotic resistance genes for high-level multidrug resistance.

Methods

Bacterial strains and general growth conditions

A. baumannii AB5075 and Escherichia coli TOP10 (Invitrogen) were maintained on lysogeny broth (Lennox, l-) agar plates (10 g l⁻¹ tryptone, 5 g l⁻¹ yeast extract, 5 g l⁻¹ NaCl and 15 g l⁻¹ agar) and grown overnight in liquid l-broth (10 g l⁻¹ tryptone, 5 g l⁻¹ yeast extract and 5 g l⁻¹ NaCl) [8]. Media contained tetracycline (12 µg ml⁻¹) when necessary to maintain plasmid pWH1266 [18]. Opaque (VIR-O subpopulation) colonies were used throughout the study.

Strain and plasmid constructions

ARGs were amplified from purified genomic DNA of A. baumannii AB5075 using DNA oligonucleotides listed in Table S1 (available in the online Supplementary Material) with ~20 bp overhangs complementary to the pWH1266 backbone to enable seamless ligation-independent cloning extract (SLiCE)-mediated cloning [19]. PCR amplification of DNA was performed using Verify^™ polymerase (PCR Biosystems). Plasmid isolations and PCR purifications were carried out according to the manufacturer’s protocol using EasyPure Plasmid MiniPrep and PCR Purification Kits (TransGen Biotech). SLiCE cloning procedure was used to insert ARGs into the PCR-amplified pWH1266 backbone [19]. All plasmid construction steps were carried out in E. coli TOP10 cells. To ensure efficient and equal translation of ARG mRNAs, the forward primer contained an artificial AGGAGG ribosome-binding site. The cloned genes were constitutively expressed from the β-lactamase (blaTEM-1) promoter of pWH1266 [18]. Plasmids were re-isolated from E. coli, their insert sequence verified by Sanger sequencing (Eurofins) and AB5075 was subsequently transformed with the plasmids using natural transformation or electroporation [20]. Whole plasmid sequencing of pWH1266 was conducted by Azenta/GENEWIZ (GENEWIZ Germany GmbH) using Oxford Nanopore Technology in a GridION with a FLO-MIN114 (R10.4.1) flow cell (GenBank accession no.: PV577797.1).

Disk diffusion assays

All A. baumannii AB5075 strains were cultured overnight in cation-adjusted Mueller–Hinton broth (MH2, Merck/Millipore). Overnight cultures were adjusted to 0.5 McFarland standard, and each culture sample was swabbed thoroughly onto Mueller–Hinton agar plates using a sterile cotton swab. Plates were air-dried for 10 min before the addition of antimicrobial susceptibility discs (Oxoid). Antibiotic discs contained amikacin (AK, 30 µg), tobramycin (TOB, 10 µg), kanamycin (K, 30 µg), gentamicin (CN, 10 µg), trimethoprim–sulphamethoxazole (SXT, 25 µg), streptomycin (S, 25 µg), meropenem (MEM, 10 µg), doripenem (DOR, 10 µg), imipenem (IPM, 10 µg), vancomycin (VA, 30 µg), erythromycin (E, 15 µg), nitrofurantoin (F, 100 µg), tetracycline (TE, 30 µg), tigecycline (TGC, 15 µg), ciprofloxacin (CIP, 5 µg), cefepime (FEP, 30 µg), cefoperazone (CFP, 30 µg), ceftazidime (CAZ, 30 µg), chloramphenicol (C, 30 µg), aztreonam (ATM, 30 µg), oxacillin (OX, 1 µg), ticarcillin (TIC, 75 µg), cefoxitin (FOX, 30 µg) and piperacillin/tazobactam (TZP, 110 µg; piperacillin 104.3 µg, tazobactam 10.2 µg). Plates were incubated overnight at 37 °C before measuring the diameters of inhibition zones.

Broth microdilution assay

Broth microdilution assays were performed as per the Clinical and Laboratory Standards Institute guidelines for Acinetobacter to determine the MIC for antibiotics. The assays were performed in 96-well plates. The standardized inoculum (0.5 McFarland) was prepared by direct colony suspension in sterile 0.9% (w/v) saline. The inoculum was then diluted in MH2 (cation-adjusted Mueller–Hinton broth, CAMHB) broth to obtain a final concentration of 5×10⁵ c.f.u. ml⁻¹ in each well except for the negative control (MH2 medium without bacteria). Plates were incubated at 37 °C statically for 22 h before the plates were analysed visually. Antibiotics were sourced from the following suppliers: Imipenem (Acros Organics), meropenem (Thermo Scientific), doripenem (Thermo Scientific), cefepime (Sigma), cefotaxime (Melford), aztreonam (Sigma), tobramycin (Sigma), gentamicin (Roth), amikacin (Cayman Chemical Company), streptomycin (Sigma), kanamycin (Sigma), ciprofloxacin (Sigma), levofloxacin (Sigma), norfloxacin (Sigma), colistin (Sigma), tetracycline (Sigma), tigecycline (Sigma), erythromycin (Sigma) and colistin (Sigma).

Growth experiments

Growth of bacterial strains was measured in a volume of 200 µl l-broth containing 12 µg ml⁻¹ tetracycline for pWH1266 plasmid maintenance in 96-well plates. A single colony of each strain was resuspended in l-broth and the absorbance at 600 nm (A₆₀₀) was normalized to 0.05 in 200 µl l-broth and incubated with orbital shaking at 37 °C and 282 cycles per minute in a BioTek Synergy HTX Multimode Reader. Every 20 min, the A₆₀₀ was recorded for 24 h.

Results

Loss of p1AB5075 results in extensive susceptibility to multiple antibiotics

A. baumannii AB5075 is considered an MDR strain with increased virulence and a contemporary representative of A. baumannii infections. Because A. baumannii, and especially A. baumannii AB5075, shows a high degree of phenotypic and genotypic heterogeneity [5, 21–26], we whole-genome sequenced a number of A. baumannii AB5075 colonies after attempting to delete genes AB5075 by the suicide-vector method of Pokhrel et al. [27]. Serendipitously, this revealed that one of the sequenced colonies, where the wild-type genotype at the position of the attempted deletion was restored, had lost plasmid p1AB5075 but was otherwise wild-type – an event which had been independently reported before [17].

As p1AB5075 carries multiple antibiotic resistance genes as part of RI-2, we set out to characterize the antibiotic resistance profile of the Δp1AB5075 strain and to investigate the individual contributions of ARGs on p1AB5075, which so far, with the exception of cmlA contributing to chloramphenicol resistance, had not been experimentally verified gene-by-gene in A. baumannii AB5075. The predicted ARGs of p1AB5075 are shown in the context of p1AB5075 in Fig. 1, and their annotation is presented in Table 1. All ARGs except one, a predicted APH(3′)-VI family aminoglycoside O-phosphotransferase (adh/ABUW_RS19420), are located within RI-2 (Fig. 1A & B). Antibiotic disc diffusion assays were performed to assess antibiotic resistance profiles of wild-type AB5075 and the Δp1AB5075 mutant strain to a suite of antibiotics.

Fig. 1. Map of A. baumannii AB5075 p1AB5075 (a), Resistance Island-2 (b) and complementation plasmid pWH1266 (c). (a) Aminoglycoside genes are highlighted in orange, chloramphenicol resistance gene cmlA in pink, sulphonamide resistance gene sul1 in light green, beta-lactamase-encoding bla_GES-11 in light blue and trimethoprim resistance gene dfrA7 in purple. The map was created with Proksee [35]. (b) Linear depiction of p1AB5075-encoded Resistance Island-2. (c) Schematic map of the pWH1266 plasmid used for complementation. The blaTEM-1 gene of pWH1266 was replaced by the ARG (blue) resulting from transcription of ARGs from the P_blaTEM-1 promoter (grey, P_bla). All ARGs are translated using the same ribosome-binding site (orange, RBS_AGGAGG). The tetA gene of pWH1266 confers resistance to tetracycline. The plasmid is not drawn to scale.

Table 1. Annotation of antibiotic resistance genes of p1AB5075.

Gene	Locus ID	Annotation
adh	ABUW_RS19420	APH(3′)-VI family aminoglycoside O-phosphotransferase
sul1	ABUW_RS19260	Sulphonamide-resistant dihydropteroate synthase
qacEΔ1	ABUW_RS19265	Quaternary ammonium compound efflux SMR transporter QacE Δ1
dfrA7	ABUW_RS19270	Trimethoprim-resistant dihydrofolate reductase DfrA7
aacA4	ABUW_RS19275	Aminoglycoside N-acetyltransferase AAC(6′)-Ib3
blaGES-11	ABUW_RS19280	Extended-spectrum class A beta-lactamase GES-11
strB	ABUW_RS19295	APH(6)-I family aminoglycoside O-phosphotransferase
aph(3'')-Ib (strA)	ABUW_RS19300	Aminoglycoside O-phosphotransferase APH(3'')-Ib
aadA2	ABUW_RS19305	ANT(3″)-Ia family aminoglycoside nucleotidyltransferase AadA2
cmlA	ABUW_RS19310	CmlA family chloramphenicol efflux MFS transporter
aadB	ABUW_RS19315	Aminoglycoside nucleotidyltransferase ANT(2″)-Ia

The Δp1AB5075 mutant strain showed widespread increased susceptibility including to all tested aminoglycosides (amikacin, gentamicin, kanamycin, streptomycin and tobramycin), but also to trimethoprim/sulphamethoxazole, aztreonam (monobactam) and ceftazidime (cephalosporin) (Fig. 2, Table 2) highlighting the crucial role of p1AB5075 in multidrug resistance. No difference was observed to ciprofloxacin (fluoroquinolone), cefoperazone, cefepime (cephalosporins), doripenem, imipenem, meropenem (carbapenems), nitrofurantoin (nitrofuran), erythromycin (macrolide), vancomycin (glycopeptide), tetracycline, tigecycline (tetracyclines), cefoxitin (cephamycin), oxacillin, ticarcillin (β-lactams), chloramphenicol and the broad-spectrum β-lactam antibiotic/inhibitor combination piperacillin/tazobactam. We noted a difference in cephalosporin resistance, where wild-type and Δp1AB5075 strains were resistant to cefoperazone and cefepime but not ceftazidime, indicating that ceftazidime resistance is mediated by Δp1AB5075 (Fig. 2). Resistance to carbapenems did not change, which is mediated by a chromosomally encoded blaOXA-23 gene [28]. Resistance to chloramphenicol was not altered either, likely due to the presence of additional chloramphenicol resistance genes craA and cpxE (ABUW_0982).

Fig. 2. Antibiotic disc diffusion assays comparing A. baumannii AB5075 wild-type (WT) and Δp1AB5075 strains. MH2 agar plates were lawned with wild-type A. baumannii AB5075 and Δp1AB5075, and antibiotic-containing discs were placed on the agar surface. The plates were incubated for 24 h at 37 °C. (a) AK (amikacin 30 µg), SXT (trimethoprim–sulphamethoxazole 25 µg), TOB (tobramycin 10 µg), S (streptomycin 25 µg), CN (gentamicin 30 µg), K (kanamycin 5 µg), (b) nitrofurantoin (F 100 µg), D (doripenem 10 µg), MEM (meropenem 10 µg), VA (vancomycin 20 µg), IPM (imipenem 10 µg), E (erythromycin 15 µg), (c) TGC (tigecycline, 15 µg), TE (tetracycline, 30 µg), OX (oxacillin, 1 µg), FOX (cefoxitin, 30 µg), TZP (piperacillin/tazobactam, 110 µg), TIC (ticarcillin, 75 µg), (d) CAZ (ceftazidime, 30 µg), CFP (cefoperazone, 30 µg), FEP (cefepime, 30 µg), ATM (aztreonam, 30 µg), C (chloramphenicol, 30 µg) and CIP (ciprofloxacin, 5 µg).

Table 2. Average of inhibition zones of AB5075 mutants measured in mm (n=3) determined by Kirby–Bauer disc diffusion assays.

	Aminoglycoside					Tetracyclines		β-Lactam											Others
	AK 30	CN 30	K 5	S 25	TOB 10	TE 30	TGC 15	IPM 10	MEM 10	DOR 10	CFP 30	CAZ 30	FEP 30	ATM 30	TIC 75	OX 1	FOX 30	TZP 110	CIP 5	C 30	SXT 25	F 100	E 15	VA 30
Wild-type	7	0	0	9	10	24	19	12	9	11	0	0	0	0	0	0	0	0	0	0	0	0	22	0
Δp1AB5075	21	20	17	16	18	22	17	13	10	12	0	19	0	7	0	0	0	0	0	7	25	0	20	0
Wild-type pWH1266	9.5	10	0	12	11	10	18	12	10	12	0	0	0	0	0	0	0	0	0	0	0	0	21	0
Δp1AB5075 pWH1266	24	21	19	15	18	10	19	12	9	12	0	17	18	8	0	0	0	0	0	7	26	0	19	0
Δp1AB5075 pWH1266-aadB	21	0	0	15	0	10	15	14	10	13	0	19	15	10	0	0	0	0	0	10	27	0	20	0
Δp1AB5075 pWH1266-adh	0	23	0	16	19	10	16	12	10	12	0	20	17	9	0	0	0	0	0	9	28	8	20	0
Δp1AB5075 pWH1266-aacA4	21	20	8	16	14	9	18	14	14	14	0	22	19	11	0	0	0	0	0	12	25	7	21	0
Δp1AB5075 pWH1266-aadA2	27	22	19	0	18	9	15	12	10	11	0	18	17	10	0	0	0	0	0	8	29	7	19	0
Δp1AB5075 pWH1266- strA	27	25	18	0	17	11	18	14	10	12	0	18	15	10	0	0	0	0	0	10	30	8	20	0
Δp1AB5075 pWH1266- strB	24	23	17	17	18	10	15	12	10	11	0	20	19	12	0	0	0	0	0	11	28	8	20	0
Δp1AB5075 pWH1266-blaGES-11	23	23	26	12	21	0	17	13	10	12	0	0	0	0	0	0	0	0	0	8	30	8	21	0

AK, amikacin; ATM, aztreonam; C, chloramphenicol; CAZ, ceftazidime; CFP, cefoperazone; CIP, ciprofloxacin; CN, gentamicin; DOR, doripenem; E, erythromycin; IPM, imipenem; K, kanamycin; MEM, meropenem; OX, oxacillin; S, streptomycin; SXT, trimethoprim/sulphamethoxazole; TE, tetracycline; TGC, tigecycline; TIC, ticarcillin; TOB, tobramycin; TZP, piperacillin/tazobactam; VA, vancomycin.

Plasmid-based complementation of resistance genes reveals individual contributions to AMR

As none of the antibiotic resistance genes have been genetically tested for their individual contributions towards antibiotic resistance in A. baumannii AB5075 except for the role of cmlA in resistance to chloramphenicol, we aimed to characterise seven predicted p1AB5075-encoded antibiotic resistance genes aacA4 (ABUW_RS19275), aadA2 (ABUW_RS19305), aadB (ABUW_RS19315), adh (ABUW_RS19420), strA (ABUW_RS19300), strB (ABUW_RS19295) and bla_GES-11 (ABUW_RS19280) by ectopically expressing them from a plasmid in the Δp1AB5075 strain. Plasmid-based complementation in A. baumannii routinely utilizes the shuttle plasmid (or its origin of replication) pWH1266, which has been assembled from pBR322 and pWH1277. The latter pWH1277 is only partially sequenced; therefore, we first sequenced the plasmid pWH1266, which assembled into a plasmid of 8,911 bp.

To compare resistance levels, all genes were expressed from the same pWH1266-endogenous promoter of the blaTEM-1 gene (Fig. 1C) and equipped with the same ribosome binding site (AGGAGG) to ensure an equal rate of translation. During cloning, the blaTEM-1 gene of pWH1266 is removed, leaving tetA as the sole (tetracycline) resistance gene. As controls, AB5075 WT and Δp1AB5075 were equipped with the pWH1266 ‘empty’ plasmid. None of the strains showed any severe growth defects when grown in l-broth (Fig. S1). Antibiotic disc diffusion assays were performed as before, and the phenotypes in WT and Δp1AB5075 strains remained the same in the presence of pWH1266 compared to the strains without pWH1266 (Figs 3, S2 and S3). As expected, mild resistance to tetracycline was acquired, which is mediated by tetA located on pWH1266 (Fig. S3). Typically, 12.5 µg ml⁻¹ tetracycline is used in liquid broth; therefore, full resistance is not achieved to the 30 µg of the antibiotic-containing disc.

Increased aminoglycoside susceptibility was noted upon deletion of p1AB5075 (Figs 2 and 3), and resistance could be restored by the expression of either adh (Fig. 3c), aadB (Fig. 3D), aadA2 (Fig. 3e), aacA4 (Fig. 3f) or strA (Fig. 3g). Expression of adh (APH(3′)-VI), which is not part of RI-2 and is flanked by two IS30 family transposases, restored resistance to kanamycin and further increased resistance to amikacin in comparison to the wild-type strain (Fig. 3a and c). Expression of aadB (ANT(2″)-Ia) restored resistance to kanamycin and increased resistance to gentamicin and tobramycin (Fig. 3a and d). Expression of aacA4 only partially restored resistance to kanamycin (Fig. 3F) and strA provided resistance to streptomycin (Fig. 3g). Expression of strB did not restore resistance to any of the tested antibiotics (Fig. 3h). Therefore, the disc diffusion assays detected genes responsible for p1AB5075-mediated aminoglycoside resistance showing also overlap in providing resistance, where adh, aadB and to a lower extent aacA4 restored resistance to kanamycin (Fig. 3c, d and f) and aadA2 and strA provided resistance to streptomycin (Fig. 3e and g). Previously, aacA4 was described as a potential pseudogene [9], we now show it is functional. The gene outside of RI-2, adh, was the only gene providing level resistance to amikacin, while RI-2-encoded aadB was the only gene restoring resistance to gentamicin and tobramycin. Expression of bla_GES-11 restored resistance to three antibiotics: ceftazidime, cefepime and aztreonam (Fig. 4), which was in line with previous work [14].

Fig. 3. Antibiotic disc diffusion assays comparing A. baumannii AB5075 wild-type (WT, a) and Δp1AB5075 (b) strains carrying pWH1266 or pWH1266-ARG: pWH1266-adh (c), pWH1266-aadB (d), pWH1266-aadA2 (e), pWH1266-aacA4 (f), pWH1266-strA (g) or pWH1266-strB (h). AK (amikacin 30 µg), SXT (trimethoprim–sulphamethoxazole 25 µg), TOB (tobramycin 10 µg), S (streptomycin 25 µg), CN (gentamicin 30 µg) and K (kanamycin 5 µg). MH2 agar plates containing tetracycline were lawned with wild-type A. baumannii AB5075 and Δp1AB5075 and antibiotic-containing discs were placed on the agar surface. The plates were incubated for 24 h at 37 °C.

Fig. 4. Antibiotic disc diffusion assays comparing A. baumannii AB5075 wild-type (WT, a) and Δp1AB5075 (b) strains carrying pWH1266 or pWH1266-bla_GES-11 (c). CAZ (ceftazidime, 30 µg), CFP (cefoperazone, 30 µg), FEP (cefepime, 30 µg), ATM (aztreonam, 30 µg), C (chloramphenicol, 30 µg), CIP (ciprofloxacin, 5 µg). MH2 agar plates containing tetracycline were lawned with wild-type A. baumannii AB5075 and Δp1AB5075 and antibiotic-containing discs were placed on the agar surface. The plates were incubated for 24 h at 37 °C.

To better quantify the contributions of the ARGs to antibiotic resistance and because there was overlap in providing resistance to several antibiotics, minimal inhibitory concentrations were determined by broth microdilution assays for selected antibiotics (Tables 3 and 4). No difference in the MIC was observed for seven clinically relevant antibiotics (Table 3). Expression of ARGs from pWH1266 not only restored resistance, but also routinely increased MICs, likely due to increased gene copy number and overexpression from the blaTEM-1 promoter of the pWH1266 plasmid (Table 3). Expression of aadB increased the MIC to tobramycin from <0.5 to 128 µg ml⁻¹ and to gentamicin from 8 to >256 µg ml⁻¹, adh increased the MIC to amikacin from <4 to 512 µg ml⁻¹, aadA2 and strA increased the MIC to streptomycin from 16 to 1,024 µg ml⁻¹ (Table 4) matching the data obtained from the disc diffusion assays (Fig. 2, Table 2). Complementation with pWH1266-aacA4 restored the resistance to tobramycin to wild-type levels (16 µg ml⁻¹) but also provided an increased low-level resistance to multiple aminoglycosides (increases of gentamicin 4-fold, kanamycin >32-fold and amikacin >2-fold) compared to Δp1AB5075 pWH1266 but did not reach wild-type resistance levels. As observed in disc diffusion assays, strB did not recover any of the antibiotic susceptibilities. Finally, the expression of bla_GES-11 increased resistance to cefepime (from 32 to 256 µg ml⁻¹) and aztreonam (from 64 µg ml⁻¹ to >256 µg ml⁻¹), as shown in the disc diffusion experiments, but also showed increased resistance from 32 to 1,024 µg ml⁻¹ to another tested cephalosporin (cefotaxime).

Table 3. Minimum inhibitory concentration of clinically relevant antibiotics for A. baumannii AB5075 determined by broth microdilution assay (n=4).

Values are in μg ml⁻¹.

Strain	CIP	NOR	LEV	TE	TGC	E	CS
WT	>32 (R)	> 256 (R)	2 (S)	0.5 (S)	<0.5 (S)	4	0.19 (S)
Δp1AB5075	>32 (R)	> 256 (R)	2 (S)	0.5 (S)	<0.5 (S)	4	0.19 (S)

*Organisms that are susceptible to tetracycline are also considered susceptible to doxycycline and minocycline.

CIP, ciprofloxacin; E, erythromycin; I, intermediate; LEV, levofloxacin; NOR, norfloxacin; R, resistant; S, susceptible; TE, tetracycline; TGC, tigecycline.

Table 4. MIC of antibiotics determined by broth microdilution assay (n=4).

Values are in μg ml⁻¹. CN (gentamicin), K (kanamycin), AK (amikacin), S (streptomycin), TOB (tobramycin), FEP (cefepime), CTX (cefotaxime), ATM (aztreonam), IPM (imipenem), MEM (meropenem), DOR (doripenem). nd (not determined). S (susceptible), I (intermediate), R (resistant).

Strain	CN	K	AK	S	TOB	FEP	CTX	ATM	IPM	MEM	DOR
WT pWH1266	>256 (R)	> 2048	128 (R)	128	16 (R)	>256 (R)	2048 (R)	>256	16 (R)	16 (R)	16
Δp1AB5075 pWH1266	8 (I)	<4	<4 (S)	16	<0.5 (S)	32 (R)	32 (I)	64	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-aadB	> 256 (R)	512	<4 (S)	64	128 (R)	32 (R)	32 (I)	32	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-adh	8 (I)	>2048	512 (R)	32	<0.5 (S)	32 (R)	32 (I)	64	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-aadA2	8 (I)	<4	<4 (S)	1024	<0.5 (S)	32 (R)	32 (I)	32	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-aacA4	32 (R)	128	8 (S)	32	16 (R)	32 (R)	32 (I)	32	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-strA	8 (I)	<4	<4 (S)	1024	<0.5 (S)	32 (R)	32 (I)	32	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-strB	8 (I)	<4	<4 (S)	32	<0.5 (S)	16 (I)	32 (I)	32	16 (R)	16 (R)	16
Δp1AB5075 pWH1266-blaGES-11	nd	nd	nd	nd	nd	256 (R)	1024 (R)	>256	16 (R)	16 (R)	16

Discussion

Multidrug resistance in A. baumannii is mediated through extensive genetic diversity supported by different resistance islands, plasmids, including p1AB5075, and other mobile genetic elements. Here, we characterized the contribution of p1AB5075 and nine p1AB5075-encoded ARGs towards antibiotic resistance. We observed that loss of p1AB5075 renders A. baumannii AB5075 susceptible to multiple antibiotics including aminoglycosides, cephalosporins, trimethoprim/sulphamethoxazole and chloramphenicol. We used this Δp1AB5075 genetic background to genetically dissect the contributions of p1AB5075-encoded antibiotic resistance genes.

As the plasmid p1AB5075 is transferable from AB5075 to many other A. baumannii and non-baumannii strains, transfer would convert any antibiotic-susceptible strain to an MDR strain [10, 11]. We determined that amikacin resistance is predominantly mediated by a gene outside of RI-2 (adh, (ABUW_RS19420, APH(3′)-VI), which is flanked by IS30 family transposases, which have been described to contain ARGs in A. baumannii including the first description of a bla_NDM-1 in Poland [29]. However, overexpression of aacA4 increased the MIC to amikacin fourfold, showing that RI-2 amplification through recombination alone can lead to substantial amikacin resistance [13]. Indeed, overexpression of aacA4 conferred low-level cross-resistance to other aminoglycosides (gentamicin, kanamycin and streptomycin) as well, expanding our knowledge of A. baumannii aminoglycoside resistance. Chloramphenicol resistance remained unchanged in the crude disc diffusion assay upon deletion of p1AB5075, which is in agreement with a previous study [17], and is due to the presence of additional chloramphenicol resistance genes craA and cpxE [30, 31].

To our knowledge, although expected upon loss of bla_GES-11, the susceptibility to cephalosporins has not been shown for AB5075 yet. We now confirm experimentally that this is due to the presence of bla_GES-11 located in RI-2. AB5075 Δp1AB5075 remains resistant to carbapenems imipenem and meropenem, though, due to the presence of a blaOXA-23 gene on the chromosome. Similar to AB5075, a clinical isolate of A. baumannii from Tunisia with a chromosomally encoded blaOXA-23 and a plasmid-encoded bla_GES-11 copy isolated from a clinical sample was reported in 2014 [32]. Conjugation of the plasmid carrying bla_GES-11 led to increased resistance to ticarcillin, ticarcillin/clavulanic acid, piperacillin, piperacillin/tazobactam, cefotaxime, ceftazidime, cefepime, aztreonam and even a slight increase in meropenem and imipenem, which might have been masked by the presence of blaOXA-23 in our strains [14]. A concerning finding was that bla_GES-11 may mutate (Gly170Ser) to bla_GES-14, resulting in a significant increase in imipenem (from two to >32 µg ml⁻¹) and meropenem (from 4 to 32 µg ml⁻¹) resistance, as shown in A. baumannii isolated from wound infections of soldiers of Ukraine [33].

Apart from the clear role of pAB5075 for AMR in AB5075, future studies may investigate the wider impact on A. baumannii AB5075 biology as the plasmid is transferable. Other examples have been reported where (loss) of plasmids had profound effects on gene regulation and pathogenesis. The plasmid pAB3 was shown to repress type-6 secretion in A. baumannii ATCC17978 [16] and a plasmid-encoded copy of H-NS located on plasmid pAB5 in A. baumannii UPAB1 was reported to regulate biofilm formation [34]. As p1AB5075 also contains a copy of an H-NS family protein, there is potential that it might mediate similar crosstalk between p1AB5075- and chromosomally encoded genes. Certainly, a more comprehensive study of plasmids in A. baumannii will be required to better understand their impact on A. baumannii biology beyond AMR.

Abbreviations

AMR

antimicrobial resistance

ARGs

antibiotic resistance genes

MDR

multidrug-resistant

RI-2

Resistance Island 2

SLiCE

seamless ligation-independent cloning extract