Role of OmpA in the Multidrug Resistance Phenotype of Acinetobacter baumannii

Abstract

Acinetobacter baumannii has emerged as a nosocomial pathogen with an increased prevalence of multidrug-resistant strains. The role of the outer membrane protein A (OmpA) in antimicrobial resistance remains poorly understood. In this report, disruption of the ompA gene led to decreased MICs of chloramphenicol, aztreonam, and nalidixic acid. We have characterized, for the first time, the contribution of OmpA in the antimicrobial resistance phenotype of A. baumannii.

TEXT

Acinetobacter baumannii is one of the most important species associated with nosocomial infections, such as ventilator-associated pneumonia, bacteremia, urinary tract infections, skin and wound infections, and meningitis (1). During the last decade, this pathogen has become increasingly resistant to most antimicrobials, including broad-spectrum cephalosporins, penicillins, carbapenems, fluoroquinolones, and aminoglycosides. Several resistance mechanisms contribute to the multidrug resistance (MDR) phenotype in A. baumannii: decreased outer membrane protein (OMP) permeability, overexpression of efflux pumps, and acquisition of genetic elements carrying resistance determinants, such as plasmids, integrons, transposons, and resistance islands.

Gram-negative bacteria typically display diverse porins in their outer membrane that participate in the modulation of cellular permeability, outer membrane protein A (OmpA) being one of the most abundant. OmpA is a β-barrel porin highly conserved among bacterial species and in A. baumannii has been associated with a variety of interesting biological properties in in vitro model systems (2). OmpA has been shown to bind host epithelia, target mitochondria, translocate to the nucleus, and induce cell death and can also bind factor H, which may allow A. baumannii to develop serum resistance (3–6). Furthermore, OmpA has also been associated with antimicrobial resistance in related pathogens (7), although only two studies have shown the involvement of OmpA as the slow porin for β-lactams (8, 9). A clear role for OmpA in antimicrobial resistance, however, has not yet been fully demonstrated.

To evaluate the involvement of OmpA in antimicrobial resistance, an internal fragment of the ompA gene (578 bp) from A. baumannii strain ATCC 17978 was amplified by PCR using the primers given in Table 1 and cloned into the pGEM-T Easy (Promega) vector by A/T cloning (10). The resulting construct was transformed in Escherichia coli strain DH5α, and plasmid was then extracted and electroporated into A. baumannii strain ATCC 17978 in order to knock out its ompA gene by insertional mutagenesis. Transformants were selected on LB agar plates containing 80 μg/ml ticarcillin. ompA gene disruption within the resulting strain, designated JPAB01, was verified by PCR using a combination of primers matching the upstream region of ompA and the pGEM-T Easy vector as well as by analyzing the outer membrane protein (OMP) profile by SDS-PAGE and Western blotting with OmpA-specific antibodies (Fig. 1).

TABLE 1.

Primers used in this study

Primer	Nucleotide sequence (5′-3′)
OmpAint up	GTTAAAGGCGACGTAGACG
OmpAint lw	CCAGTGTTATCTGTGTGACC
OmpAext up	GGAATGGCTATAACTGACATAATC
OmpAext lw	GAATCAGGAGATTTACAAATGACC
OmpA-EcoRI F	GACTAGGAATTCGGAATGGCTATAACTGACATAATC
OmpA-EcoRI R	GACTAGGAATTCGAATCAGGAGATTTACAAATGACC
M13-Fa	GTAAAACGACGGCCAGT
M13-Ra	CAGGAAACAGCTATGAC

^aPrimer designed in pGEM-T.

FIG 1.

OMP profile and immunodetection of OmpA of A. baumannii strains. OMPs were extracted from the outer membrane of ATCC 17978 wt, JPAB01, and JPAB01+pJPAB01 strains and stained with Simply Blue SafeStain (SDS-PAGE) or electrotransferred onto a nitrocellulose membrane (Western blot). OMPs were probed with mouse anti-OmpA and horseradish peroxidase-conjugated goat anti-mouse IgG. Molecular mass standards (kDa) are shown on the right.

PCR, SDS-PAGE, and Western blot analysis confirmed the disruption of the ompA gene and the absence of OmpA expression in JPAB01 (Fig. 1 and data not shown). Etest testing of the susceptibilities of the wild-type (wt) and mutant strains to quinolones (ciprofloxacin and nalidixic acid), chloramphenicol, tetracycline, tigecycline, aztreonam, imipenem, erythromycin, trimethoprim, kanamycin, and ceftazidime was performed in duplicate. The JPAB01 strain was more susceptible than the ATCC 17978 strain to chloramphenicol (>8-fold), aztreonam (8-fold), and nalidixic acid (3-fold) (Table 2), all of which are substrates of efflux pumps.

TABLE 2.

Determination of MICs of different antimicrobial agents for A. baumannii strains used in this study (mean values of duplicate assays)

Antimicrobial agent	MIC (μg/ml)			Difference in susceptibility (fold)
	ATCC 17978 wt	JPAB01	JPAB01+pJPAB01	ATCC 17978 wt vs JPAB01b	ATCC 17978 wt vs JPAB01+pJPAB01
Ceftazidime	2	3	4	0.67	0.5
Imipenem	0.5	0.38	0.38	1.31	1.31
Ciprofloxacin	0.25	0.38	0.25	0.66	1
Kanamycin	3	1.5	2	2	1.5
Tetracycline	6	4	32	1.5	0.19
Erythromycin	12	6	NDa	2	ND
Trimethoprim	32	32	ND	1	ND
Tigecycline	1	0.5	0.5	2	2
Colistin	0.5	0.25	1	2	0.5
Aztreonam	16	2	16	8	1
Aztreonam + PAβN	4	0.5	2	8	2
Nalidixic acid	16	6	16	2.67	1
Nalidixic acid + PAβN	2	2	2	1	1
Chloramphenicol	>256	32	>256	>8	ND
Chloramphenicol + PAβN	32	32	32	1	1

^aND, not determined.

^bValues in bold indicate a difference in susceptibility of greater than twofold.

To demonstrate that the increased chloramphenicol, aztreonam, and nalidixic acid susceptibilities observed for the JPAB01 strain were due to the lack of a functional OmpA protein, the complementation of JPAB01 has been performed. The ompA gene was amplified with the OmpA-EcoRI F and OmpA-EcoRI R primers (Table 1) from the ATCC 17978 wild-type (wt) genome and cloned into the EcoRI restriction site of the pWH1266 vector (11), yielding the pWH1266-OmpA plasmid called pJPAB01. pJPAB01 was transformed into DH5α and electroporated into JPAB01. Transformants were selected on LB agar plates containing 25 μg/ml tetracycline. The complementation of OmpA in the JPAB01 strain restored the MICs of chloramphenicol, aztreonam, and nalidixic up to >256, 16, and 16 μg/ml, respectively, demonstrating that OmpA was indeed involved in the chloramphenicol, aztreonam, and nalidixic acid resistance phenotypes. Moreover, the presence of the efflux pump inhibitor phenyl-arginine-β-naphthylamide (PAβN) reduced the MICs of chloramphenicol, aztreonam, and nalidixic acid at least >8-, 4-, and 8-fold in the wild-type strain and the complemented strain.

A similar reduction in the MICs of aztreonam and nalidixic acid, but not of chloramphenicol, was observed for the mutant JPAB01 strain in the presence of the inhibitor (Table 2). These results suggest the presence of additional mechanisms contributing to the reduced susceptibility to these antibiotics, likely related to the expression of efflux pumps.

This is the first description of the involvement of OmpA in antimicrobial resistance in A. baumannii. The exact mechanism of action is not clear, but it is possible that OmpA participates in the extrusion of compounds from the periplasmic space through the outer membrane and couples with inner membrane efflux systems, such as major facilitator superfamily (MFS) efflux pumps or RND systems lacking the OMP component. A similar situation has been described for the MexXY system in Pseudomonas aeruginosa, which, despite lacking a gene coding for an outer membrane porin, can associate with OprM and likely other OMPs, such as OpmB, OpmG, or OpmH, to form a functional tripartite efflux system (12, 13). A. baumannii contains several RND tripartite systems, such as AdeABC, AdeIJK, and AdeFGH. In AdeABC, the presence of AdeC is not required for antimicrobial resistance, suggesting that AdeAB can utilize another outer membrane constituent such as AdeK (14). Analysis of the genome of the ATCC 17978 strain suggests that the adeC gene is missing and that, thus, other OMPs have to compensate for the lack of adeC. Interestingly, a knockout mutant of the CraA MFS pump in A. baumannii displays a decrease in the MIC of chloramphenicol similar to that observed for JPAB01 (15), and no further effect is observed in the presence of efflux inhibitors.

In E. coli and Citrobacter freundii, increased expression of OmpA homologues has been associated with decreased susceptibility to tetracycline (7) and carbapenems (16), respectively. Taking into account that OmpA is the major outer membrane protein in A. baumannii, we cannot rule out the possibility that its absence in the knockout mutant disturbs membrane processes, including the transport of antimicrobial agents. To our knowledge, there are no studies regarding the overall modification of gene expression upon OmpA loss in A. baumannii, and this matter should be further investigated. In our study, only three nonrelated antimicrobials have shown a reduced MIC in the knockout strain, and thus, we do not believe such disturbance to be a major issue.

It is clear that additional studies are needed to evaluate the clinical relevance of OmpA expression in MDR, but we have shown the participation of OmpA in the antimicrobial resistance phenotype of A. baumannii.