AbuO, a TolC-Like Outer Membrane Protein of Acinetobacter baumannii, Is Involved in Antimicrobial and Oxidative Stress Resistance

Abstract

Although Acinetobacter baumannii is well accepted as a nosocomial pathogen, only a few of the outer membrane proteins (OMPs) have been functionally characterized. In this study, we demonstrate the biological functions of AbuO, a homolog of TolC from Escherichia coli. Inactivation of abuO led to increased sensitivity to high osmolarity and oxidative stress challenge. The ΔabuO mutant displayed increased susceptibility to antibiotics, such as amikacin, carbenicillin, ceftriaxone, meropenem, streptomycin, and tigecycline, and hospital-based disinfectants, such as benzalkonium chloride and chlorhexidine. The reverse transcription (RT)-PCR analysis indicated increased expression of efflux pumps (resistance nodulation cell division [RND] efflux pump acrD, 8-fold; SMR-type emrE homolog, 12-fold; and major facilitator superfamily [MFS]-type ampG homolog, 2.7-fold) and two-component response regulators (baeR, 4.67-fold; ompR, 10.43-fold) in the ΔabuO mutant together with downregulation of rstA (4.22-fold) and the pilin chaperone (9-fold). The isogenic mutant displayed lower virulence in a nematode model (P < 0.01). Experimental evidence for the binding of MerR-type transcriptional regulator SoxR to radiolabeled abuO promoter suggests regulation of abuO by SoxR in A. baumannii.

INTRODUCTION

Outer membrane proteins (OMPs) are known to have a pivotal role in bacterial physiology, such as adherence, invasion, and serum resistance, maintenance of cell structure, and binding a variety of substances, including passive and active transport (1). The archetypical OMP, TolC, has been considered to be a multifunctional protein due to its involvement in cell membrane integrity, acid tolerance, expulsion of metabolites, export of siderophores that are required in iron acquisition, export of plasmid and chromosomally encoded toxins, such as hemolysin, colicin V, and microcins, and virulence, as evident from studies in Enterobacter, Borrelia, Salmonella, Vibrio, Legionella, Francisella, and Escherichia coli (2–10). In E. coli, TolC is promiscuous because it supports the functioning of multidrug resistance efflux pumps, such as AcrD, AcrEF, and MdtABC (the resistance nodulation cell division [RND] superfamily) (11–13), EmrAB and EmrKY (the major facilitator superfamily [MFS]), and MacAB (the ATP-binding cassette [ABC] superfamily) (14–17). Although the functions of TolC homologs in many Gram-negative bacteria, such as E. coli, Vibrio vulnificus, Stenotrophomonas maltophilia, Enterobacter cloacae, and Yersinia pestis (5, 18–21), have been elucidated, the biological functions of TolC in an important human pathogen, Acinetobacter baumannii, have remained enigmatic so far.

Multidrug-resistant (MDR) Acinetobacter strains kill up to 50% of infected patients, despite treatment with last resort drugs, and the resistance rates of such strains continue to escalate globally (22, 23). Significant increases in the number of A. baumannii strains that are resistant to carbapenems, cephalosporins, aminoglycosides, and fluoroquinolones with a diverse antibiotic resistome have been reported from hospitals in the United States and other countries (24–27). In our previous study, we demonstrated the role of antibiotic resistance genes and efflux pumps in mediating antimicrobial resistance in A. baumannii isolates from Ohio (28, 29). To date, few OMPs have been implicated in carbapenem resistance when their expression is reduced: CarO, Omp33 to -36, and an OprD homolog (30–32).

In a continuation of our efforts toward understanding the origin/network of multidrug resistance in Acinetobacter, in this study, using genetic and molecular approaches, we demonstrate the role of a putative OMP (a homolog of TolC, designated AbuO) in bacterial stress physiology in general and antimicrobial resistance in particular for the first time in A. baumannii.

MATERIALS AND METHODS

Bacterial strains and media.

A. baumannii AYE was purchased from the American Type Culture Collection (ATCC BAA1710). Bacterial cultures were grown in Luria-Bertani (LB) broth or agar (Difco, Becton Dickinson, Sparks, MD) with 400 μg/ml hygromycin for the mutant and 200 μg/ml zeocin for complemented strains. Restriction digestion, ligation, transformation, and agarose gel electrophoresis were done according to standard protocols. Plasmid and genomic DNAs were prepared using Gene Aid kits according to the manufacturer's protocol. DNA products were sequenced to confirm their authenticity (Applied Biosystems). The primers used in the present study were custom synthesized (Eurofins MWG Operon, Germany).

Construction of the ΔabuO mutation in A. baumannii AYE.

A 693-bp internal fragment was amplified using ΔabuO-F/ΔabuO-R primers (see Table S1 in the supplemental material), cloned into the pUC4K-derived suicide vector harboring the hygromycin cassette. The obtained plasmid, pUC-abuO, was transformed into A. baumannii AYE to construct the ΔabuO mutant. The gene disruption was confirmed by Southern hybridization and PCR analysis. The zeocin cassette was amplified from pCR Blunt II-TOPO vector (Life Technologies), using zeo-NT and zeo-CT primers (see Table S1) and cloned into shuttle vector pWH1266. Furthermore, the intact abuO gene along with its promoter was amplified with FLabuO-F and FLabuO-R (see Table S1), cloned into the modified pWH1266 vector. The resulting construct was transformed into the ΔabuO mutant and selected on LB agar plates supplemented with 200 μg/ml zeocin to obtain the transcomplemented ΔabuO ΩabuO strain. Mutant and complemented strains were characterized, and their phenotypes were compared with that of the wild type (WT), A. baumannii AYE.

Bacterial phenotypic assays.

The growth profiles of the WT, ΔabuO, and ΔabuO ΩabuO strains were monitored in LB at different pHs (5.0, 6.0, 7.0, 8.0, 10.0, and 12.0) for 18 h at 37°C with shaking using Bioscreen C automated growth analysis system (Labsystems, Helsinki, Finland) at the optical density at 600 nm (OD₆₀₀). The growth inhibition assay was performed as before with slight modifications using ciprofloxacin (0.005 μg/ml), ethidium bromide (EtBr [4 μg/ml]), and chlorhexidine (1.6 μg/ml) (33). The impact of abuO inactivation on motility behavior and biofilm formation was examined as described before (34). Studies to decipher the impact of oxidative stress-inducing agent hydrogen peroxide (H₂O₂) and nitrosative stress-inducing agents (sodium nitroprusside [SNP] and acidified nitrite) on the WT, ΔabuO, and ΔabuO ΩabuO strains were performed as described before (33). The WT, ΔabuO, and ΔabuO ΩabuO strains were exposed to hostile stress agents at different concentrations, such as bile salt deoxycholate, sodium chloride (NaCl), EtBr, acridine orange, acriflavine, rhodamine, safranin, ampicillin, neomycin, ciprofloxacin, chloramphenicol, tetracycline, benzalkonium chloride, chlorhexidine, and triclosan, and survival capability was determined as described before (33).

Drug susceptibility, efflux assay, and OMP preparations.

Antibiotic susceptibility and MIC were examined using commercial discs and E-strips (Hi Media, Mumbai, India), and data were analyzed according to the interpretation criteria recommended by the CLSI (35). Accumulation assays using fluorescent substrates EtBr and ciprofloxacin and purification of OMPs from the WT, ΔabuO, and ΔabuO ΩabuO strains were done as described before (33).

RNA isolation and real-time RT-PCR.

Total RNA was extracted from log-phase cultures using the RNeasy minikit according to the manufacturer's instructions. Aliquots of 500 ng of DNase I-treated total RNA served as the template for cDNA synthesis using superscript III reverse transcriptase (Invitrogen). Gene expression levels were monitored by real-time reverse transcription (RT)-PCR using universal SYBR green supermix (Bio-Rad) in an iCycler thermal cycler (Bio-Rad), and melting curve analysis was carried out to confirm amplification of a single product. Total RNA was isolated from three independently grown cultures, and real-time RT-PCR experiments were performed six times with rpoB as an endogenous control.

Caenorhabditis elegans killing assay.

Bacterial virulence assays were performed with the nematode model C. elegans strain Bristol N2 as used before but with slight modifications (36). To examine the ability of the WT, ΔabuO, ΔabuO ΩabuO, and E. coli OP50 strains to kill C. elegans, bacterial lawns of A. baumannii and the E. coli control strain were prepared on nematode growth (NG) medium and incubated at 37°C for 6 h. The plates were kept at room temperature for 1 h and then seeded with L4-stage worms (25 to 30 per plate). Furthermore, the seeded plates were incubated at 25°C and examined for live worms under a stereomicroscope (Leica MS5) after every 24 h. When the worm did not react to touch, it was considered dead. At least five replicates repeated three times were performed for each selected strain.

Gene cloning, expression, purification, and EMSAs.

The genome of the A. baumannii AYE strain reveals the presence of ∼214 signal-transducing proteins (GenBank accession no. CU459141.1). The MerR type DNA-binding helix-turn-helix HTH-type transcriptional regulator designated by gene identifier no. ABAYE2390 (soxR; 453 bp, 150 amino acids [aa] and 17.01 kDa) was amplified using gene-specific primers, which contained NdeI and BamHI sites of the pET28C vector to generate an N-terminal His₆-SoxR fusion protein. The ability of SoxR to bind the abuO promoter was deciphered by electrophoretic mobility shift assays (EMSAs) as described previously (33). To confirm that the interaction between SoxR and the promoter region of abuO was specific, experiments with competitive [specific, 10-fold excess of cold promoter; nonspecific, poly(dI-dC)] and noncompetitive (bovine serum albumin [BSA]) inhibitors were also performed.

Bioinformatic analysis and statistical analysis.

The NCBI Internet server was used to perform homology searches, similarity and identity analyses, and conserved domain architecture analysis. All data are presented as means ± standard errors of the means. Statistical analysis was performed on crude data by using Student t test. P values of <0.05 were considered significant.

RESULTS

Bioinformatic analysis of A. baumannii TolC-like protein, AbuO.

The abuO gene is composed of 1,347 bp, which encodes a 448-aa-long, type I secreted OMP, AbuO (GenBank accession no. YP_001715271.1), with a predicted signal peptide sequence cleavage site at the N-terminal region between Ala₁₉ and Leu₂₀. Analysis established that AbuO is predicted to localize to the outer membrane and contains duplicate domains that belong to the outer membrane efflux protein family. The sequence alignment of AbuO with homologs (indicated by the accession numbers shown) from different bacteria exhibited conservation at the amino acid level: for example, to E. coli TolC protein P02930 (27.8% identity and 47.9% similarity), Salmonella enterica subsp. enterica serovar Enteritidis AAC43973.1 (27% identity and 46.6% similarity), Pseudomonas aeruginosa AGV57551.1 (25.8% identity and 44.7% similarity), and Vibrio cholerae Q9K2Y1.1 (27.9% identity and 49.1% similarity) (Fig. 1).

FIG 1.

Multiple-sequence alignment of abuO and its homologs. Sequence alignments were made in CLUSTAL Omega (https://www.ebi.ac.uk/Tools/msa/clustalo) and formatted using the ESPript server (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi). The GenBank accession numbers (or gene identifier no. for A. baumannii) of TolC homologs from different bacteria are as follows: Escherichia coli, P02930; A. baumannii, ABAYE3514; Yersinia pestis KIM10+, AAM87064.1; Enterobacter aerogenes, CAD13188.1; Salmonella enterica subsp. enterica serovar Enteritidis, AAC43973.1; Vibrio cholerae, Q9K2Y1.1; Erwinia amylovora, CBA19383.1; and Pseudomonas aeruginosa PAO581, AGV57551.1. The predicted secondary structural elements of A. baumannii AbuO are shown on the lines above the sequence alignment using PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/). The arrows indicate a β-sheet, the coils indicate α-helices, TT indicates β turns, and η indicates 3₁₀ helices. Residues strictly conserved have a black background and are indicated by boldface letters; residues conserved between groups are boxed.

Secondary structure prediction indicated that AbuO consists of three domains—a β-barrel or channel domain (with the four β-strands β1, β2, S4, and S5), an α-helical barrel or tunnel domain (comprising long H3 and H7 and shorter H2, H4, H6, and H8 α-helices), and mixed α/β domains or an equatorial domain (small β strand and α-helical structures S3 and S6 and H1, H5, and H9)—similar to the structure of the E. coli outer membrane protein TolC (Fig. 1). The MiST2 database (www.mistdb.com) contains genome sequences of ∼157 Acinetobacter strains (142 draft and 15 complete), with sizes ranging from 2.9 to 5.0 Mbp (37, 38), with the presence of putative AbuO. Multiple alignment of these putative homologs from sequenced A. baumannii genomes exhibited 99% identity with amino acid alterations at position 178 (Asn to Ser [α3 domain]) and 218 (Thr to Ser [α4 domain]) compared to AbuO (data not shown). Overall, in silico analysis established AbuO to be a TolC-like protein highly conserved in A. baumannii.

Novel contributions of AbuO, an OMP, to stress response in A. baumannii.

Analysis of growth profiles indicated that the abuO mutant exhibited slower growth at various pH values than the WT strain (Fig. 2). When the cultures were grown in LB plates with different agar concentrations, WT cells migrated all over, while ΔabuO cells exhibited affected motile behavior (see Fig. S1A in the supplemental material). On the other side, the in vitro biofilm-forming ability of the ΔabuO strain was only ∼ 0.8 ± 0.173-fold less than that of the WT strain (see Fig. S1B). Thus, this finding indicated that abuO has no direct role to play in influencing the motility and biofilm-forming phenotypes of the pathogen. When tested with various concentrations of sodium deoxycholate (bile salt), the survival ability of the ΔabuO strain was marginally affected compared to that of the WT (Fig. 3A). The ability of cells to grow in the presence of various concentrations of NaCl was tested interestingly at 0.75 M NaCl; the percentage of survival for the WT was ∼2.15 ± 0.024-fold higher than that of the ΔabuO strain regardless of the inoculum size (Fig. 3B).

FIG 2.

Bacterial growth curves: impact of inactivation of abuO in A. baumannii. The effect on bacterial growth was monitored for the WT, ΔabuO, and ΔabuO ΩabuO strains in LB medium at various pHs. The patterns of representative pHs (5.0 [A], 6.0 [B], 7.0 [C], 8.0 [D], and 10.0 [E]; P < 0.01) are shown here. At 540 min, the mutant exhibited 1.21 ± 0.057-fold (at pH 5.0), 1.32 ± 0.032-fold (at pH 6.0), 1.37 ± 0.043-fold (at pH 7.0), 1.74 ± 0.027-fold (at pH 8.0), and 1.57 ± 0.067-fold (at pH 10.0) slower growth than the WT strain. The other tested pH conditions, 3.0 and 4.0, were toxic to both cultures. Complementation restored the growth defect. The data presented are the means of triplicate measurements performed three times.

FIG 3.

Stress challenge assays: effect of loss of OMP AbuO in A. baumannii. (A) The ability of WT to survive in the presence of varied deoxycholate concentrations (16, 64, 256, 1,024, and 4,096 μg/ml) was compared with that of the ΔabuO strain (P < 0.01). (B) The survival ability of the WT was tested at various NaCl concentrations (0.075, 0.15, 0.25, 0.5, and 0.75 M) and compared with that of the ΔabuO strain (P = 0.004). (C) The survival ability of the abuO mutant was analyzed in the presence of various H₂O₂ concentrations (0.07894, 0.7894, 1.5788, 2.3682, and 3.1576 mM) (P = 0.0023). Asterisks indicate significant difference in the mutant with respect to the WT (P < 0.01).

The abuO mutant exhibited >4 ± 0.05-fold-stunted growth compared to the WT in LB broth when tested in the presence of various concentrations of H₂O₂ (see Fig. S1C in the supplemental material). Upon performance of the oxidative survival assay, the abuO mutant exhibited 4.5 ± 0.058-fold-reduced survival compared to the WT when treated with 3.1576 mM H₂O₂ (Fig. 3C). The role of abuO in nitrosative stress response was elucidated by comparing the growth profiles and survival rates of the WT, ΔabuO, and ΔabuO ΩabuO strains in LB broth at different concentrations of SNP (see Fig. S1D) or acidified nitrite (see Fig. S1E), and apparently no significant change was observed. Overall the results strongly suggest the involvement of AbuO, an OMP, in protecting against high osmotic and oxidative stress challenges in A. baumannii.

Role of AbuO in conferring broad-spectrum antimicrobial resistance in A. baumannii.

Analysis of MIC values for the ΔabuO strain revealed increased susceptibility to amikacin, carbenicillin, ceftriaxone, meropenem, streptomycin, tigecycline compared to the WT (Table 1). The survival of the WT, ΔabuO, and ΔabuO ΩabuO strains was monitored in the presence of antibiotics representing different classes for, e.g., ampicillin (Fig. 4A), neomycin (Fig. 4B), ciprofloxacin (Fig. 4C), chloramphenicol (Fig. 4D), and tetracycline (Fig. 4E). The total CFU of the WT at 256 μg/ml of neomycin, 512 μg/ml of ampicillin, and 16 μg/ml of tetracycline were 1.4 ± 0.018-fold, 1.7 ± 0.089-fold, and 1.25 ± 0.056-fold higher than those in ΔabuO cells, respectively. The overall results convincingly suggest that AbuO is a novel MDR determinant in A. baumannii.

TABLE 1.

Determination of MICs for the WT, ΔabuO mutant, and complemented ΔabuO ΩabuO strains in A. baumannii

Antibiotic	MIC (μg/ml) fora:		Fold changeb	MIC (μg/ml) for ΔabuO ΩabuO straina
	WT	ΔabuO mutant
Amikacin	256	64	4	128
Amoxicillin	>240	120	>2	>240
Ampicillin	>1,024	512	>2	>256
Carbenicillin	>512	128	>4	>256
Cefepime	256	256	1	256
Ceftazidime	256	256	1	256
Ceftriaxone	256	64	4	128
Ciprofloxacin	60	30	2	60
Chloramphenicol	512	256	2	>256
Clindamycin	5	2.5	2	5
Colistin	0.01	0.01	1	0.01
Co-trimoxazole	240	240	1	240
Doripenem	24	16	1.5	24
Ertapenem	>32	32	>1	32
Gentamicin	128	128	1	128
Kanamycin	240	240	1	240
Meropenem	6	2	3	4
Nalidixic acid	>240	120	>2	>240
Neomycin	256	>128	2	256
Ofloxacin	8	4	2	8
Rifampin	4	2	2	4
Sparfloxacin	1	1	1	1
Streptomycin	10	2.5	4	10
Tetracycline	16	16	1	16
Ticarcillin	10	10	1	10
Tigecycline	2	0.75	2.6	2
Vancomycin	>8	4	>2	>8

^aE-strips were used to determine the precise MICs for different groups of antibiotics, such as amikacin, amoxicillin, ampicillin, carbenicillin, cefepime, ceftazidime, ceftriaxone, ciprofloxacin, chloramphenicol, clindamycin, colistin, co-trimoxazole, doripenem, ertapenem, gentamicin, kanamycin, meropenem, nalidixic acid, neomycin, ofloxacin, rifampin, sparfloxacin, streptomycin, tetracycline, ticarcillin, tigecycline, and vancomycin following the CLSI guidelines (35). Complementation restored the MICs.

^bThe fold change is the ratio of MICs for the WT and ΔabuO mutant.

FIG 4.

Contributions of AbuO to antibiotic resistance in A. baumannii. The survival ability of the WT in the presence of ampicillin (P = 0.001) (A), neomycin (P = 0.001) (B), ciprofloxacin (P = 0.003) (C), chloramphenicol (P = 0.018) (D), and tetracycline (P = 0.003) (E) at different concentrations (0.5, 4, 16, 64, 256, and 1,024 μg/ml) was compared to that of the ΔabuO mutant. All data presented here are the means of independent measurements performed three times. Asterisks indicate significant difference in the mutant with respect to the WT (P < 0.01).

The abuO mutant cells exhibited reduced survival when exposed to different concentrations of efflux pump substrates, such as EtBr (see Fig. S2A in the supplemental material), acridine orange (see Fig. S2B), acriflavine (see Fig. S2C), rhodamine (see Fig. S2D), and safranin (see Fig. S2E). A growth inhibition assay using carbonyl cyanide m-chlorophenylhydrazone (CCCP) with such substrates as, e.g., EtBr (4 μg/ml) (see Fig. S2F) or the antibiotic ciprofloxacin (0.005 μg/ml) (see Fig. S2G) indicated stunted growth by the mutant, reflecting the loss of drug extrusion capacity in the isogenic mutant. Results so far corroborate AbuO to be an OMP mediating MDR via active efflux.

Further whole-cell EtBr accumulation assays were performed to authenticate the observation. As the mutant lacks AbuO in its functional form, the fluorescence intensity was higher in the abuO mutant than in the WT (Fig. 5A and B). Addition of CCCP further increased the fluorescence signal in the mutant as the inhibitor dissipated the proton electrochemical gradient, diminishing active efflux. The study with ciprofloxacin yielded a similar conclusion on loss of efflux capability by the ΔabuO strain (Fig. 5C and D). Alterations in the OMP profile of the mutant with expression of additional bands indicate the pathogen's alternative strategy to combat MDR stress (data not shown). Hence, we surmise that inactivation of abuO does distort the active efflux capability in A. baumannii.

FIG 5.

Fluorimetric accumulation assay. When treated with various concentrations of fluorescent substrates in independent experiments in the presence of 0.4% glucose at 37°C, the fluorescence intensity (arbitrary fluorescence units [AU]) was relatively lower in the WT than that in the ΔabuO mutant. At 40.1 min, the WT exhibited lower EtBr accumulation than the mutant: 1.21-fold at 0.01 μg/ml, 0.85-fold at 0.05 μg/ml, 1.04-fold at 0.5 μg/ml, 1.07-fold at 1.0 μg/ml, 1.04-fold at 2.0 μg/ml, 1.21-fold at 4.0 μg/ml, 1.04-fold at 6.0 μg/ml, and 1.06-fold at 8.0 μg/ml (A and B). At 50.1 min, the WT exhibited lower ciprofloxacin accumulation than the mutant: 1.0-fold at 0.01 μg/ml, 1.02-fold at 0.05 μg/ml, 1.44-fold at 0.5 μg/ml, 1.67-fold at 1.0 μg/ml, 1.72-fold at 2.0 μg/ml, 1.43-fold at 4.0 μg/ml, 1.98-fold at 6.0 μg/ml, and 1.35-fold at 8.0 μg/ml (C and D). The fluorescence was monitored in a Hitachi spectrofluorometer at 37°C.

Survival assays of the WT, ΔabuO, and ΔabuO ΩabuO strains using different concentrations of benzalkonium chloride (Fig. 6A), chlorhexidine (Fig. 6B), and triclosan (Fig. 6C) and the growth inhibition assay (see Fig. S2H in the supplemental material) confirmed the ability of AbuO to confer disinfectant resistance in A. baumannii. The results in this section demonstrate that AbuO, an OMP, confers broad-spectrum antimicrobial resistance via active efflux in A. baumannii.

FIG 6.

Biocide challenge assays: result of loss of functional AbuO in A. baumannii. The survival ability of the WT in the presence of benzalkonium chloride (1.8 ± 0.091-fold at 6.4 μg/ml; P = 0.005) (A), chlorhexidine (1.2 ± 0.045-fold at 3.2 μg/ml; P = 0.001) (B), and triclosan (1.28 ± 0.034-fold at 0.001 μg/ml; P = 0.001) (C) was higher than that of the ΔabuO mutant, and complementation restored the phenotype. All data presented here are the means of independent measurements performed three times. Asterisks indicate significant difference in the mutant with respect to the WT (P < 0.01).

Mutation in abuO impacts expression of various cellular genes in A. baumannii.

Compared to the WT strain, the expression levels of the RND-type (e.g., acrD, 8-fold), ABC-type (macB, 18-fold), and SMR-type (emrE, 12-fold) efflux pump genes were increased in ΔabuO in A. baumannii. Altered expression of OMPs like OmpA, CarO, and CsuA, together with ∼9-fold-decreased expression of pilin chaperone and ∼12-fold-increased expression of pilT suggests possible involvement of abuO in influencing motility and membrane permeability in A. baumannii. Altered expression of signal-transducing protein genes baeS, baeR, and ompR with downregulation of rstA pinpoints the crucial role of abuO in maintaining the cellular physiology in A. baumannii (Table 2). Complementation of the abuO mutation almost restored expression of all the tested genes (P < 0.0001), implying the overall broader role of abuO in A. baumannii.

TABLE 2.

Real-time PCR analysis performed in the WT and abuO mutant strains

Function and gene identifier no.	Annotation or description	Avg fold changea
Transport
ABAYE1822	adeB; RND protein; K18146 multidrug efflux pump	−2.11 ± 0.38
ABAYE1823	adeC; outer membrane protein; K18147 outer membrane protein	−6.50 ± 1.94
ABAYE0747	RND protein (AdeB-like); K18138 multidrug efflux pump	−5.24 ± 0.69
ABAYE0746	Outer membrane protein (AdeC-like); K18139 outer membrane protein	−2.17 ± 0.51
ABAYE1796	Multidrug resistance efflux pump	5.30 ± 1.37
ABAYE3381	norM; multidrug ABC transporter; K03327 multidrug resistance protein, MATE family	1.37 ± 0.07
ABAYE3248	macB; macrolide ABC transporter ATP-binding/membrane protein	18.59 ± 2.27
ABAYE0728	MFS family transporter—ampG homolog	2.70 ± 0.32
ABAYE1181	Multidrug resistance efflux protein; K03297 small multidrug resistance family protein	12.13 ± 1.93
ABAYE3515	Hypothetical protein; K03298 drug/metabolite transporter, DME family	3.65 ± 0.56
ABAYE0008	RND-type efflux pump involved in aminoglycoside resistance (AdeT)	2.36 ± 0.77
ABAYE0010	adeT; RND-type efflux pump involved in aminoglycoside resistance	2.62 ± 0.89
ABAYE0827	Multidrug efflux protein	8.92 ± 1.21
ABAYE0640	Outer membrane protein OmpA-like	16.87 ± 2.40
ABAYE0924	Porin protein associated with imipenem resistance	18.74 ± 0.53
Motility
ABAYE1319	Protein CsuA/B; secreted protein related to type I pili	6.96 ± 1.45
ABAYE1857	Pilin chaperone; K07346 fimbrial chaperone protein	−9.85 ± 2.19
ABAYE0304	Fimbrial protein	6.32 ± 1.71
ABAYE2918	pilT; twitching motility protein	11.99 ± 2.28
Signaling systems
ABAYE0599	baeS; kinase sensor component of a two-component signal transduction system	14.51 ± 0.83
ABAYE0600	baeR; OmpR family transcriptional regulator	4.67 ± 0.50
ABAYE0259	ompR; osmolarity response regulator	10.43 ± 0.76
ABAYE3064	rstA; response regulator	−4.22 ± 0.54

^aGene expression was normalized to the endogenous control (rpoB). The average fold change is reported relative to the WT from at least six independent experiments along with standard deviations.

Role of abuO in virulence in A. baumannii.

The Caenorhabditis elegans-A. baumannii infection model was employed to determine the involvement of abuO in virulence (39). The WT and ΔabuO strains were examined for their abilities to kill C. elegans. The wild-type strain displayed 10% and 20% killing at 96 and 120 h, respectively. However, the ΔabuO and ΔabuO ΩabuO strains killed only 5% and 8% of the worms after 96 h (P < 0.01), respectively. E. coli strain OP50 was used as a negative control. Thus, our findings demonstrate that the abuO mutant kills C. elegans more slowly than the WT strain.

Studies on the regulation of abuO by SoxR in A. baumannii.

We assessed the promoter region of abuO, and analysis revealed the presence of a conserved putative SoxR binding site in the promoter (see Fig. S3A in the supplemental material). The soxR gene (gene identifier no. ABAYE2390) is a 453-bp gene that encodes a polypeptide of 150 aa (17.01 kDa). To define the possible interaction of SoxR with the promoter of abuO, we tested whether SoxR directly interacts with the promoter region of abuO. We carried out gel shift assays using the ³²P-labeled abuO promoter fragment and purified SoxR protein. Protein-DNA complexes formed upon incubation of SoxR with 200 bp of radiolabeled abuO promoter in reaction buffer, resolved by 5% PAGE, revealed a clear retardation that was directly proportional to the protein concentration (see Fig. S3B). The fact that there was no binding and the absence of DNA-protein complexes in autoradiograph when different controls were used, such as competitive [specific, 10-fold excess of cold promoter; nonspecific, poly(dI-dC)] and noncompetitive (BSA) inhibitors in independent experiments, clearly demonstrate the specific DNA binding ability of SoxR to the promoter region of abuO in A. baumannii.

DISCUSSION

In this study, we have shown the unprecedented involvement of AbuO, a TolC homolog in stress physiology and antimicrobial resistance in A. baumannii. The expression of OMP in Vibrio cholerae and Sinorhizobium meliloti has been associated with susceptibility to osmotic stress (40, 41). Bile is a substrate of AcrAB-TolC in S. enterica, E. coli, and V. cholerae (42–44). In the presence of osmotic/bile challenges, the abuO mutant exhibited >1.5- to 2.0-fold-lower survival capabilities than the WT, and its level of growth at physiological pH was found to be affected; therefore, we conclude that AbuO helps bacteria survive environmental challenges, such as high osmolarity and bile.

The abuO mutant was sensitive to oxidative stress; our observations are in agreement with the established role of TolC in S. meliloti and S. enterica (41, 45). AbuO may possibly help in efflux of reactive oxygen species and help bacterial survival inside humans. However, a study pertaining to this hypothesis is highly warranted. The TolC-like protein HgdD of the cyanobacterium Anabaena sp. strain PCC 7120 is reported to be involved in secondary metabolite export and antibiotic resistance (46). Our results demonstrated that inactivation of abuO rendered cells sensitive to various antibiotics. Subsequent assays pinpoint the crucial role of AbuO in active efflux in A. baumannii with broad substrate specificity. Resistance to quaternary ammonium compounds was dependent on the expression of OprR in Pseudomonas aeruginosa (47). Outer membrane changes in Pseudomonas stutzeri led to resistance to chlorhexidine diacetate and cetylpyridinium chloride (48). In conjunction, we found AbuO also had a role in conferring a biocide resistance phenotype in A. baumannii.

Besides, the abuO mutant displayed a lower virulence capability, suggesting that in addition to its role in multidrug efflux, this novel OMP may be involved in secretion of a toxin or virulence factor required for pathogenesis in A. baumannii. The altered protein interaction/signaling events prevailing in the mutant may be different from those in the WT, because the virulence defect is not fully restored upon complementation; however, additional experiments will be essential to explain the hypothesis. Alteration of expression of cellular genes in the abuO mutant indicates a broader regulatory role of AbuO in A. baumannii; detailed studies may help elucidate the interacting partners/cascade. The abuO mutants were constructed in different Indian clinical isolates and functionally characterized (data not shown); data analysis authenticated the gene's conserved functions. Therefore, AbuO indeed appears to be an intrinsic broad-spectrum antimicrobial resistance determinant in A. baumannii.

Conclusions.

Overall, this study reporting the wide physiological functions of AbuO in mediating stress response and antimicrobial resistance in A. baumannii for the very first time has brought us one step ahead in our efforts to understand the origin of multidrug resistance in Acinetobacter.