ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Jan. 2010, p. 333–340
Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Vol. 54, No. 1
Screening and Quantification of the Expression of Antibiotic Resistance
Genes in Acinetobacter baumannii with a Microarray?
Se ´bastien Coyne,1Ghislaine Guigon,2Patrice Courvalin,1* and Bruno Pe ´richon1
Institut Pasteur, Unite ´ des Agents Antibacte ´riens,1and Plateforme Ge ´notypage des Pathoge `nes et Sante ´ Publique,275724 Paris,
Cedex 15, France
Received 24 July 2009/Returned for modification 18 October 2009/Accepted 25 October 2009
An oligonucleotide-based DNA microarray was developed to evaluate expression of genes for efflux pumps in
Acinetobacter baumannii and to detect acquired antibiotic resistance determinants. The microarray contained
probes for 205 genes, including those for 47 efflux systems, 55 resistance determinants, and 35 housekeeping
genes. The microarray was validated by comparative analysis of mutants overexpressing or deficient in the
pumps relative to the parental strain. The performance of the microarray was also evaluated using in vitro
single-step mutants obtained on various antibiotics. Overexpression, confirmed by quantitative reverse trans-
criptase PCR, of RND efflux pumps AdeABC, due to a G30D substitution in AdeS in a multidrug-resistant
(MDR) strain obtained on gentamicin, and AdeIJK, in two mutants obtained on cefotaxime or tetracycline, was
detected. A new efflux pump, AdeFGH, was found to be overexpressed in a mutant obtained on chloramphen-
icol. Study of MDR clinical isolates, including the AYE strain, whose entire sequence has been determined,
indicated overexpression of AdeABC and of the chromosomally encoded cephalosporinase as well as the
presence of several acquired resistance genes. The overexpressed and acquired determinants detected by the
microarray could account for nearly the entire MDR phenotype of the isolates. The microarray is potentially
useful for detection of resistance in A. baumannii and should allow detection of new efflux systems associated
with antibiotic resistance.
Multidrug-resistant (MDR) strains of Acinetobacter bau-
mannii have emerged in recent decades. This opportunistic
pathogen is responsible for severe infections, particularly hos-
pital-acquired pneumonia and bloodstream, urinary tract, and
wound infections, and has become of worldwide concern (13).
As in other bacterial species, multidrug resistance can be
achieved by two mechanisms: (i) horizontal transfer of genetic
information and (ii) mutation of endogenous genes. Acquired
resistance determinants that are carried by plasmids (18, 28),
transposons (23, 29), and integrons (33, 43) have been de-
scribed for Acinetobacter spp. Determination of the genomic
sequence of several A. baumannii strains has improved our
knowledge of the ways in which A. baumannii can develop
antibiotic resistance (1, 21, 38, 45). An 86-kb resistance island,
AbaR1, found in strain AYE, contains as many as 25 antibiotic
and 20 antiseptic and heavy metal resistance genes (16). Vari-
ants of this island are integrated at the same chromosomal
locus in a significantly high proportion of MDR strains (37). In
addition to these acquired resistance genetic elements, alter-
ations in endogenous functions are involved in resistance, such
as overexpression of chromosomally encoded ?-lactamases
ADC and OXA-51-like; loss of porins CarO and Omp33–36
contributing to carbapenem resistance; mutation in the GyrA
and ParC fluoroquinolone targets; and overexpression of efflux
Efflux systems are components of the bacterial membrane
that are thought to play a role in homeostasis of the cell and
extrusion of toxic compounds (32). They can also be involved
in cell-to-cell communication via quorum-sensing systems (22)
and in bacterial pathogenicity (32). Efflux pumps of the resis-
tance nodulation cell division (RND) superfamily, widespread
in gram-negative bacteria, confer multidrug resistance to their
host when overexpressed (34). In most instances, the level of
resistance remains moderate; however, reduced intracellular
accumulation of antibiotics provides a delay for the selection of
high-level resistance by drug inactivation or target alteration.
Efflux thus can act in synergy with other mechanisms to achieve
high-level resistance (34).
Two RND efflux systems, AdeABC and AdeIJK, have been
characterized for A. baumannii (9, 24). They are composed of
an efflux protein (AdeB or AdeJ) which interacts with a mem-
brane fusion protein (AdeA or AdeI) and an outer membrane
factor (AdeC or AdeK) to facilitate drug export across both
the inner and the outer membranes. The adeABC operon is
cryptic in approximately 80% of A. baumannii strains (24), and
its overexpression leads to diminished susceptibility to amino-
glycosides, cefepime, fluoroquinolones, chloramphenicol, and
tetracycline-tigecycline. Consequently, the AdeABC efflux
pump constitutes a major mechanism of multiple antibiotic
resistance in A. baumannii, and its clinical significance has
been established (31). AdeIJK is present in all A. baumannii
strains and confers decreased susceptibility to ?-lactams, fluo-
roquinolones, chloramphenicol, rifampin (rifampicin), eryth-
romycin, clindamycin, and tetracycline-tigecycline. The toxicity
observed with Escherichia coli and A. baumannii when the
adeIJK operon was cloned and overexpressed suggests its effect
on resistance is minimal (9). Another pump belonging to the
MATE family, AbeM, has been characterized (39), but its role
in the antibiotic resistance of A. baumannii remains to be
* Corresponding author. Mailing address: Institut Pasteur, Unite ´
des Agents Antibacte ´riens, 75724 Paris, Cedex 15, France. Phone: 33
1 45 68 83 20. Fax: 33 1 45 68 83 19. E-mail: patrice.courvalin@pasteur
?Published ahead of print on 2 November 2009.
determined. Analysis of the genome sequence of strain AYE
revealed the presence of 41 other genes putatively associated
with efflux (16).
Microarrays are powerful tools for large-scale screening of
the genomic or transcriptomic content of a cell population.
Several genomic microarrays have been developed to detect
acquired antibiotic resistance genes (5) and transcriptomic mi-
croarrays have been used to compare spontaneous mutants
with the parental strain, highlighting the role of overexpression
of efflux genes in antibiotic resistance (14).
Overexpression of efflux systems is an important factor lead-
ing to antibiotic resistance but is not easily detectable at both
the phenotypic level, because of the low level of resistance
conferred, and the genotypic level, because the regulatory mu-
tations most often remain unknown. To study the mechanisms
of multidrug resistance in A. baumannii, particularly those
involving efflux, we have developed a microarray that allows
quantification of the expression of resistance genes, including
intrinsic and acquired determinants.
(An initial report of this work was presented by S. Coyne et
al. at the 49th Interscience Conference on Antimicrobial
Agents and Chemotherapy, abstr. C1-096, 2009.)
MATERIALS AND METHODS
Bacterial strains, growth conditions, and antibiotic susceptibility testing. The
bacterial strains used in this study are listed in Table 1. Cells were grown at 37°C
in brain heart infusion broth and agar (Difco Laboratories, Detroit, MI). Anti-
biotic susceptibility was tested by disk diffusion on Mueller-Hinton agar (Bio-
Rad, Marnes-la-Coquette, France), and MICs were determined by the Etest
procedure (AB Biodisk, Solna, Sweden).
Production of MDR mutants. Spontaneous MDR mutants BM4665 and
BM4666 were obtained from wild-type strain BM4587, and mutant BM4668 was
obtained from strain BM4667, on gradient plates (40) containing either genta-
micin, cefotaxime, or tetracycline (Table 1). Colonies growing at concentrations
higher than the normal MIC were tested for antibiotic susceptibility by diffusion,
and those exhibiting resistance to several drug classes were selected for further
DNA manipulations. A. baumannii genomic DNA was extracted as described
previously (36). Amplification of DNA was performed in a GeneAmp PCR
system 9700 (Perkin-Elmer Cetus, Norwalk, CO) with Taq (MPbio, Illkirch,
France) or Phusion (Finnzymes, Espoo, Finland) DNA polymerase. Amplifica-
tion of large DNA fragments was achieved using the Expand Long Template
PCR system (Roche Diagnostic GmbH, Mannheim, Germany) according to the
manufacturer’s recommendations. PCR elongation times and temperatures were
adjusted according to the expected sizes of the PCR products and the nucleotide
sequences of the primers (Table 2), respectively. Sequence determination was
carried out with a CEQ 2000 DNA analysis system automatic sequencer (Beck-
man Instruments, Inc., Palo Alto, CA).
RNA isolation. RNA was isolated as described previously (27). Briefly, expo-
nentially growing cells (optical density at 600 nm of 0.8 to 1.2) were harvested,
and the pellet was resuspended in 12.5 mM Tris–5 mM EDTA–10% glucose.
Cells were sheared mechanically in the presence of acid phenol (pH 4.6) and
glass beads (0.2- to 0.3-mm diameter; Sigma-Aldrich, St. Louis, MO) using a
Fastprep apparatus (Bio 101). After centrifugation, Trizol reagent (Invitrogen,
Leek, The Netherlands) was added to the supernatant. Total RNA was extracted
twice with chloroform-isoamyl alcohol (24:1 [vol/vol]), precipitated in 0.7 volume
of isopropanol and washed twice with 70% ethanol. Treatment with RNase-free
DNase (Ambion, Austin, TX) was performed according to the manufacturer’s
recommendations. RNA was quantified by determining absorbance at 260 and
280 nm; its purity and integrity were analyzed on agarose gels.
Microarray construction. Oligonucleotide probes (70-mer) were designed
with ArrayOligoSelector (4) to represent each open reading frame annotated as
an A. baumannii coding sequence in the GenBank database (http://www.ncbi.nlm
.nih.gov/) using the complete genome sequence of AYE as a reference. BLAST
analysis of probes against themselves or the AYE genome enabled nonspecific
binding effects to be eliminated. The 205 probes selected were synthesized and
spotted in triplicate on UltraGAPS glass slides (Corning Life Sciences, Corning,
NY) using a Chipwriter Pro Virtek arrayer (Bio-Rad, Hercules, CA). After
printing, arrays were treated as recommended by the slide manufacturer.
Synthesis of cDNA and microarray hybridization. cDNAs were synthesized
from 10 ?g of total RNA and labeled with Cy3 or Cy5 cyanin (GE Healthcare,
Uppsala, Sweden) using the SuperScript indirect cDNA labeling system (Invitro-
gen). Labeled cDNAs of the studied and reference strains were mixed at equal
concentrations (250 pmol of cyanin) and hybridized on the microarray. Slides
were scanned (GenePix 4000 A scanner; Axon), and images were analyzed
(GenePix5 software; Axon).
Data analysis. Script analyses were developed with the software R (R Devel-
opment Core Team; http://www.R-project.org). For each spot, the fluorescence
value corresponded to the median pixel intensity within the spot and no back-
ground was subtracted. Three independent biological replicates and a dye swap
for each replicate to overcome cyanin bias were performed. Loess normalization
(46) was applied to all spots, on a slide-by-slide basis. To determine differentially
expressed genes from replicates pooling log2ratios, a paired t test was per-
formed. Low numbers of observations per spot and of spots per slide (615) did
not permit calculation of a specific variance and clusters of spots with equal
variance, respectively. For these reasons, it was assumed that the variance of the
log2ratios was the same for all spots, and spots displaying extreme specific
variance (too small or too large) were excluded from the statistical analysis (17).
A spot was considered differentially expressed when the Bonferroni method-
adjusted P value was lower than 0.05. The mean and standard deviation were
calculated for genes corresponding to significantly differentially expressed spots
and with a change of ??1.5 or ?1.5.
Quantitative reverse transcriptase (qRT) PCR. cDNA synthesis was per-
formed for 60 min at 42°C using 0.1 to 0.2 ?g of RNA, 0.8 ?l of avian myalo-
blastosis virus reverse transcriptase (Roche Diagnostic GmbH, Mannheim, Ger-
many), and 2 ?l of random primer according to the manufacturer’s instructions.
Real-time PCR was performed with a 20-?l reaction volume containing 2 ?l of
cDNA, 1? SYBR PCR master mix (Roche), and gene-specific primers, 0.5 mM
each (Table 2). Amplification and detection of specific products were performed
using the LightCyler sequence detection system (Roche) with the following cycle
profile: 1 cycle at 95°C for 10 min, 40 cycles at 95°C for 15 s, 58°C for 10 s, and
72°C for 30 s.
Microarray data accession number. The microarray data obtained in this study
were deposited in ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) according
to MIAME standards under accession number E-MEXP-2254.
RESULTS AND DISCUSSION
Design of the microarray. An oligonucleotide-based DNA
microarray was developed to compare the expression levels
of efflux genes of an isogenic MDR mutant and its parental
strain or those of clinical isolates and a reference strain. The
TABLE 1. Bacterial strains used in this study
Susceptible reference strain
CIP 70-10 AdeST153Mspontaneous
CIP 70-10 AdeRP116Lspontaneous
Clinical strain overexpressing AdeABC
BM4454 ?adeABC ?adeIJK
Susceptible clinical strain
BM4587 spontaneous mutant obtained on
BM4587 spontaneous mutant obtained on
Susceptible clinical strain
BM4667 spontaneous mutant obtained on
MDR clinical strain
MDR clinical strain
MDR clinical strain
334COYNE ET AL.ANTIMICROB. AGENTS CHEMOTHER.
GenBank database was screened for A. baumannii gene
sequences since no annotated genome was available at the
beginning of the study. A total of 205 genes was selected for
the microarray. The list of genes and corresponding probes is
available at ArrayExpress (http://www.ebi.ac.uk/arrayexpress/)
under accession number E-MEXP-2254. The probes were de-
signed to detect the 47 efflux-related genes of strain AYE (16),
which comprise six RND (including adeABC and adeIJK),
seven MF, and two MATE systems; 10 genes for outer mem-
brane proteins, e.g., those encoding the pore-forming protein
HMP-AB (19), the 33- to 36-kDa protein (10), and CarO (30);
and 8 genes involved in biofilm formation, notably the csu
operon (41) and its two-component regulatory system (42).
Fifty-five antibiotic resistance genes from gram-negative bac-
teria were added, several of them not yet described for A.
baumannii. They represent the most clinically relevant resis-
tance mechanisms, i.e., the chromosomal cephalosporinase,
class A ?-lactamases, metallo-?-lactamases and oxacillinases,
including carbapenemases; aminoglycoside modifying enzymes
(11 acetyl-, 4 nucleotidyl-, and 3 phosphotransferases), 16S
RNA methylases, tetracycline efflux pumps and ribosomal pro-
tection proteins, rifampin ADP ribosyltransferase (Arr-2),
chloramphenicol acetyltransferases and efflux pumps, and plas-
mid-borne quinolone resistance gene qnr. Genes conferring
resistance to arsenic (ars) and mercury (mer) heavy metals
were also included. Finally, genes involved in mobility or trans-
fer of genetic elements, such as integrases, insertion sequences,
or transposases, and several genes present in the 86-kb resis-
tance island of strain AYE, such as trxB, lspA, and uspA, were
added. These sequences should enable detection of the back-
bone structures of mobile genetic elements carrying antibiotic
TABLE 2. Oligonucleotides used in PCR and qRT-PCR
GenePrimer Sequence (5?33?)Positiona
adeA adeA For
adeC adeC For
adeK adeK For
tniA tniA For
aCoordinates refer to the first base of each gene.
bFrom reference 20.
cFrom reference 8.
VOL. 54, 2010EXPRESSION OF RESISTANCE GENES IN A. BAUMANNII335
Thirty-five housekeeping genes, such as gyrA, parC and rpoB,
as well as genes used for an MLST scheme (2) and a multilocus
PCR typing method (15), and 20 genes linked to metabolism
functions were added to avoid normalization bias due to the
low density of the chip.
Validation of the microarray. Microarray validation was first
performed by comparing strain BM4454, which overexpresses
the AdeABC efflux system (24), with derivative mutants
BM4579, BM4651, and BM4652, in which adeIJK, adeABC,
and both systems, respectively, had been inactivated by inser-
tion-deletion (9). Comparative analysis of the data obtained
with BM4454 and BM4579 (BM4454 ?adeIJK) revealed that,
as expected, the adeI and adeJ genes were not expressed in the
mutant (Table 3). Since integration of the cassette partially
deletes the adeI and adeJ genes but not adeK, no differences
were observed for adeK expression. These results were con-
firmed by qRT-PCR using specific primers (Table 2). Compar-
ison of the BM4454 and BM4651 (BM4454 ?adeABC) tran-
scriptomes confirmed inactivation of adeA and adeB in the
latter. The aac(3)-IVa cassette that was inserted to inactivate
this system was confirmed to be expressed in BM4651. Simi-
larly, adeC was overexpressed due to a transcriptional fusion
with aac(3?)-IVa in the mutant (Table 3) (9). The expression of
other efflux genes was not affected by inactivation of adeABC.
Comparative analysis of the transcriptomes of BM4454 and
BM4652 (BM4454 ?adeABC ?adeIJK) confirmed the inactiva-
tion of adeA, adeB, adeI, and adeJ and overexpression of
aac(3)-IVa and adeC. Differential expression of adeA and adeC
was not found to be statistically significant by the microarray
but was confirmed by qRT-PCR (Table 3).
Validation was also performed by studying AdeABC-over-
expressing mutants BM4546 and BM4547 obtained from strain
CIP 70-10 following point mutations in adeS and adeR, respec-
tively (Table 1) (25). The two-component system AdeRS con-
trols the expression of adeABC, AdeR being a transcriptional
regulator and AdeS a histidine kinase sensor (25). The regu-
latory operon is located upstream from adeABC and tran-
scribed in the opposite direction. Overexpression of adeA and
adeB in BM4546, but not in BM4547, was observed (Table 3).
qRT-PCR showed 10- and 4-fold increases in adeB expression
in BM4546 and BM4547, respectively. These results are in
agreement with BM4546 having an amount of adeB mRNA
larger than that in BM4547 as detected by Northern hybrid-
ization (25), despite the fact that the two mutants exhibit sim-
ilar resistance phenotypes. No changes in the expression of the
adeR and adeS genes were detected, indicating that the two-
component system was not self-regulated. Expression of sev-
eral housekeeping genes, such as cpn60 and gltA in BM4546
and recA in BM4547, was modified, suggesting collateral effects
due to AdeRS mutations or AdeABC overexpression. How-
ever, no differences in growth levels were observed between
the strains (data not shown).
Detection of adeABC and adeIJK overexpression. BM4665, a
spontaneous resistant mutant of susceptible clinical isolate
BM4587 obtained on gentamicin (Table 1), was found to be
less susceptible to aminoglycosides (12-fold increase in MICs),
fluoroquinolones, chloramphenicol, and tetracycline-tigecy-
cline (Table 4). No changes in ?-lactam susceptibility were
observed, except for cefepime. Strain BM4665 overexpressed
the adeABC operon (1.8- to 3-fold increase) (Table 3) but no
other efflux genes, and the MDR phenotype of BM4665 was in
agreement with overexpression of the pump (24). A 10- to
30-fold increase in expression of adeABC was found by qRT-
PCR. Overexpression of the AdeABC system can be due either
to mutations in the adeRS two-component system (25) or to
the insertion of ISAba1 upstream from the adeABC operon
(35). In order to elucidate the overexpression of adeABC in
BM4665, the sequence of adeRS was determined, revealing a
Gly-to-Asp substitution at position 30 of AdeS. This residue is
located in the periplasmic input domain of the sensor. Muta-
tions in this domain of several histidine protein kinases, such as
NarX (7), BvgS (26), and VanSD(11), have been described and
lead to constitutive expression of the regulated genes. Thus,
the G30D substitution is most likely responsible for overex-
pression of AdeABC in BM4665.
Spontaneous MDR mutants BM4666 and BM4668 obtained
from BM4587 and BM4667 on cefotaxime and tetracycline,
respectively (Table 1), had similar resistance profiles charac-
terized by decreased susceptibility to tetracycline-tigecycline,
fluoroquinolones, chloramphenicol, co-trimoxazole, rifampin,
and to a lesser extent, ?-lactams, whereas aminoglycosides
were not affected (Table 4). This phenotype was distinct from
that associated with overexpression of AdeABC, suggesting
another mechanism. Comparative microarray experiments re-
TABLE 3. Differential quantification of gene expression in isogenic strains of A. baumannii
Strain Change (fold)a
BM4667BM4668 NSNSNS 1.532.42 1.04b
aPositive and negative values correspond to genes with, respectively, increased or decreased expression in the mutant strain in comparison with the parental strain.
NS, not statistically significant. NA, not applicable.
bNot statistically significant by the microarray but confirmed by qRT-PCR.
336 COYNE ET AL.ANTIMICROB. AGENTS CHEMOTHER.
vealed overexpression of adeI and adeJ (1.5- and 2.4-fold in-
creases, respectively) in the MDR mutants (Table 3), which
was confirmed by qRT-PCR, indicating a 2.5- to 5-fold increase
in the expression of the adeIJK gene set in the two mutants.
Determination of the sequence of the adeIJK operon and of
the upstream flanking region containing its putative promoter
did not reveal any mutation, and regulation of expression of
adeIJK remains unknown. The level of overexpression of
adeIJK was systematically lower than that observed for adeABC
in microarray as well as in qRT-PCR experiments. trans-
complementation of BM4579 (BM4454 ?adeIJK) with plas-
mid-borne adeIJK indicated that overexpression of AdeIJK
was toxic for the host (9). Taken together, these observations
suggest that AdeIJK is tightly regulated and can be overex-
pressed only within narrow limits. The fact that the BM4666
and BM4668 spontaneous mutants overexpress adeIJK at a low
level suggests that these types of mutations can occur in clinical
strains. The role of this efflux system in antibiotic resistance of
clinical isolates remains to be determined.
Another mutant from BM4652 (BM4454 ?adeABC ?adeIJK)
obtained on chloramphenicol exhibited an MDR phenotype.
Comparison of the transcriptome of the mutant with that of
the parental strain using the microarray revealed overexpres-
sion (2.9- to 3.8-fold) of a new RND efflux system, AdeFGH,
that could be responsible for this phenotype and is currently
Detection of antibiotic resistance genes in clinical isolates.
Comparison by the microarray of AYE with susceptible refer-
ence strain CIP 70-10 indicated that genes adc for the chro-
mosomal cephalosporinase and adeABC for the efflux system
were overexpressed and that, among the 70 genes of the
AbaR1 resistance island, 19 out of the 33 printed on the mi-
croarray were expressed, e.g., blaVEB, ant(2??)-Ia, cat, arr-2,
aac(3)-Ia, floR, merA, IS1999, intI1, and sul1 (Table 5). Se-
quencing of the PCR product obtained with the ISAba1b and
amp Rev primers (Table 2) confirmed the presence of a copy
of ISAba1 upstream from adc that likely provides a strong
promoter (20) for enhanced production of cephalosporinase in
AYE. Expression of the chromosomally located blaOXA-51-type
was not detected, and analysis of the AYE genome sequence
did not show the presence of ISAba1 upstream from the struc-
tural gene for the ?-lactamase that is required for overexpres-
sion leading to carbapenem resistance (44). Fourteen genes of
the resistance island were not detected by the microarray prob-
ably due to lack of, or weak, expression. All the resistance
genes present in AYE except tet(A) and blaOXA-10were de-
tected; tet(A) transcription is repressed by TetR in the absence
of tetracycline (6), while blaOXA-10, part of a class 1 integron
that also contains blaVEB, ant(2??)-Ia, arr-2, cmlA5, and
ant(3??)Ia, was weakly expressed, although it possesses its own
promoter (12). Within an integron, the expression of promot-
erless, cassette-associated resistance genes is influenced mark-
edly by their position in the cassette array (12). A constant
decrease in the expression of genes blaVEB, ant(2??)-Ia, and
arr-2, depending on the distance from the promoter borne by
IS1999, was found by the microarray (Fig. 1). The microarray
detected the expression of other genes in AbaR1, such as the
mercury resistance operon and intI1 and qacE?1-sul1, that
comprise the 5? and 3? conserved sequences of the class 1
integron, respectively (Table 5). Combination of (i) overex-
pression of the cephalosporinase and AdeABC efflux system
and (ii) the presence of acquired resistance genes, most of
them carried by the 86-kb island, resulted in multiple antibiotic
resistance of the AYE host.
BM4675 and BM4676, two other unrelated MDR clinical
isolates, were analyzed. BM4675 was highly resistant to fluo-
roquinolones, tetracyclines, rifampin, chloramphenicol, co-tri-
moxazole, and ?-lactams and moderately resistant to aminogly-
cosides and exhibited decreased susceptibility to tigecycline
(Table 4). The only antimicrobial agent tested that conserved
in vitro activity was imipenem (Table 4). Comparison of the
transcriptome of BM4675 with that of reference strain CIP
70-10 indicated that genes adc for the chromosomal cephalos-
porinase and adeABC for the efflux system were overexpressed
(Table 5). As in AYE, a copy of ISAba1 was found upstream
from adc, which probably provides the promoter for its expres-
sion. The experiment enabled detection of the ?-lactam
blaTEM-likeand tetracycline tet(B) resistance genes. As with
tet(A), regulation by the tetR repressor does not allow expres-
sion of tet(B) in the absence of tetracycline. Thus, constitutive
expression of tet(B) in BM4675 could result from a lack of or
a nonfunctional tetR. Overexpression of the ADC cephalospo-
rinase and of AdeABC, combined with production of TEM
and Tet(B), accounts for part of the resistance phenotype of
BM4675 (Table 4). Other alterations, such as mutations in type
II topoisomerases, diminished permeability, or acquired genes
not spotted on the microarray, probably combined to produce
the multidrug resistance phenotype.
TABLE 4. Antibiotic susceptibility of A. baumannii
TICTIM CAZCTXFEP ATMIPM CHLTETMIN TIGRIF NORCIPSXTGENTOBAMK NET
?256 ?256 ?256 ?256 ?256
?256 ?256 ?256 ?256
?256 ?256 ?256
96 8 32
aAMK, amikacin; ATM, aztreonam; CAZ, ceftazidime; CHL, chloramphenicol; CIP, ciprofloxacin; CLI, clindamycin; CTX, cefotaxime; ERY, erythromycin; FEP,
cefepime; GEN, gentamicin; IPM, imipenem; MIN, minocycline; NET, netilmicin; NOR, norfloxacin; RIF, rifampin; SXT, co-trimoxazole; TET, tetracycline; TIC,
ticarcillin; TIG, tigecycline; TIM, ticarcillin-clavulanic acid; TOB, tobramycin.
VOL. 54, 2010EXPRESSION OF RESISTANCE GENES IN A. BAUMANNII 337
TABLE 5. Differential quantification of gene expression in clinical isolates compared to CIP 70-10
Function or gene Description
D-Serine/D-alanine/glycine transport protein
Outer membrane protein
Putative outer membrane protein
33- to 36-kDa outer membrane protein
34-kDa outer membrane protein
Outer membrane protein A
Housekeeping and metabolism
bfmS Sensor kinase component of a two-component regulatory
Putative pilus subunit, biofilm operon
Translation elongation factor EF-G
DNA-directed RNA polymerase beta subunit
Class D carbapenemase
Rifampin ADP-ribosylating transferase
Chloramphenicol efflux pump
Chloramphenicol efflux pump
Tetracycline efflux pump
Heavy metal resistance
Heavy metal detoxification
Mercuric ion reductase
Mercury resistance operon regulatory protein
Gene mobility and island
Lipoprotein signal peptidase
Quaternary ammonium compound resistance protein
Universal stress protein
aPositive and negative values correspond to genes with increased and decreased expression, respectively, in the clinical isolate in comparison with CIP 70-10. NS,
not statistically significant. NA, not applicable.
bStatistically significant but not confirmed by qRT-PCR.
338COYNE ET AL.ANTIMICROB. AGENTS CHEMOTHER.
BM4676 had high-level resistance to imipenem that was
not restored by EDTA, suggesting the presence of a class D
carbapenemase. The strain was also resistant to ?-lactams,
chloramphenicol, fluoroquinolones, and tetracycline and
had decreased susceptibility to minocycline and tigecycline.
Moderate-level resistance was observed for netilmicin,
whereas the strain was highly resistant to the other amino-
glycosides tested (Table 4). Comparative transcriptome
analysis with reference strain CIP 70-10 indicated overex-
pression of chromosomal genes adc for the cephalospori-
nase and adeABC for the efflux system (Table 5). Weak
overexpression of adeI and adeJ, confirmed by qRT-PCR,
was observed (Table 5). As in BM4675, a copy of ISAba1
was found upstream from adc. The microarray also enabled
detection of production of OXA-23 carbapenemase and of
ANT(2??)-Ia and APH(3?)-VIa aminoglycoside nucleotidyl-
and phosphotransferase (Table 5). The presence of a copy
of ISAba4 upstream from blaOXA-23, located by PCR using
ISAba4b and oxa23- primers (Table 2), was presumably re-
sponsible for enhanced expression of this gene, as previously
demonstrated (8). As observed for BM4675, the overex-
pressed and acquired determinants detected by the microar-
ray could account for nearly the entire MDR phenotype of
BM4676 (Table 4).
The microarray also detected the expressed transposase
tnpM and tniA genes in BM4675 and the tniA, lspA, and uspA
genes, coding for a transposase, a lipoprotein signal peptide,
and a universal stress protein in BM4676, respectively (Table
5). These genes are part of the AbaR1 resistance island of
AYE (16). PCR experiments using the ATPase3? and tniA Rev
primers (Table 2) indicated that, in both strains, tniA was
inserted at the 3? end of an ATPase gene which has been
proposed to be a potential “hot spot” for integration of
genomic islands in A. baumannii (37). No amplification prod-
ucts were obtained when long-range PCR was performed be-
tween tniA and the 5? end of the ATPase gene using the tniA
For and ATPase5? primers (Table 2), possibly because either
the 5? end of the ATPase gene had been deleted or the primers
were too distant. None of the antibiotic resistance genes de-
scribed for the original AbaR1 island were detected with
BM4675 and BM4676.
Differential expression of several genes coding for outer
membrane proteins was observed with the three clinical iso-
lates relative to CIP 70-10 (Table 5). These changes could
impact membrane permeability and account for susceptibility
decrease, in particular with imipenem, either alone or in com-
bination with efflux systems and ?-lactamase production.
Other genes, in particular housekeeping genes cpn60, ppa,
and rpoC in AYE and fusA and gltA in BM4676, were found to
be differentially expressed (Table 5). Modification of expres-
sion of these genes could reflect the cost of antimicrobial
resistance for host fitness, as already proposed (14).
Microarray technology is a powerful tool that can be de-
signed for comparative genomics (presence or absence of a
gene) or for transcriptomics (level of gene expression). The
microarray described in this study is a combination of both
approaches since it enabled us to quantify expression of chro-
mosomal genes and to detect acquired determinants that are
expressed. It represents a useful tool to screen for resistance, in
particular by overexpression of efflux systems, which is difficult
to detect phenotypically and genotypically. In addition, the
microarray allowed detection of a new pump, AdeFGH, in-
volved in the antibiotic resistance of A. baumannii. The large
panel of antibiotic resistance genes spotted on the microarray,
including some that have not yet been detected in A. bauman-
nii, enabled detection of most of the mechanisms that are
responsible for the multidrug resistance phenotype of numer-
ous clinical isolates of this species.
We thank T. Lambert for providing clinical isolates and helpful
discussions; C. Gouyette for synthesis of oligonucleotides; R. Lavenir
and I. Iteman, Plateforme Ge ´notypage des Pathoge `nes et Sante ´ Pub-
lique, Institut Pasteur, for technical assistance; the Plateforme Puces a `
ADN for spotting the biochips; and P. E. Reynolds for reading the
This work was supported by Institut Pasteur and Institut de Veille
Sanitaire (Saint-Maurice, France).
1. Adams, M. D., K. Goglin, N. Molyneaux, K. M. Hujer, H. Lavender, J. J.
Jamison, I. J. MacDonald, K. M. Martin, T. Russo, A. A. Campagnari, A. M.
Hujer, R. A. Bonomo, and S. R. Gill. 2008. Comparative genome sequence
analysis of multidrug-resistant Acinetobacter baumannii. J. Bacteriol. 190:
2. Bartual, S. G., H. Seifert, C. Hippler, M. A. Luzon, H. Wisplinghoff, and F.
Rodriguez-Valera. 2005. Development of a multilocus sequence typing
scheme for characterization of clinical isolates of Acinetobacter baumannii.
J. Clin. Microbiol. 43:4382–4390.
3. Bouvet, P. J. M., and P. Grimont. 1986. Taxonomy of the genus Acinetobacter
FIG. 1. Differential gene expression in the class 1 integron of the AbaR1 island from strain AYE. The ratio was determined by comparing the
expression levels of every gene between strains AYE and CIP 70-10. 5?CS and 3?CS, 5? and 3? conserved sequences of class 1 integron; ?,
statistically significant; ND, not determined. Arrows indicate coding sequences and sense of transcription. Bent arrows represent the promoters
for the gene cassettes, with IS1999 bringing a strong promoter. Decreased intensity in gray indicates decreased overexpression from blaVEB-1to
VOL. 54, 2010EXPRESSION OF RESISTANCE GENES IN A. BAUMANNII339
with the recognition of Acinetobacter baumannii sp. nov., Acinetobacter hae-
molyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp.
nov. and emended descriptions of Acinetobacter calcoaceticus and Acineto-
bacter lwoffii. Int. J. Syst. Bacteriol. 36:228–240.
4. Bozdech, Z., J. Zhu, M. P. Joachimiak, F. E. Cohen, B. Pulliam, and J. L.
DeRisi. 2003. Expression profiling of the schizont and trophozoite stages of
Plasmodium falciparum with a long-oligonucleotide microarray. Genome
5. Call, D. R., M. K. Bakko, M. J. Krug, and M. C. Roberts. 2003. Identifying
antimicrobial resistance genes with DNA microarrays. Antimicrob. Agents
6. Chopra, I., and M. Roberts. 2001. Tetracycline antibiotics: mode of action,
applications, molecular biology, and epidemiology of bacterial resistance.
Microbiol. Mol. Biol. Rev. 65:232–260.
7. Collins, L. A., S. M. Egan, and V. Stewart. 1992. Mutational analysis reveals
functional similarity between NARX, a nitrate sensor in Escherichia coli
K-12, and the methyl-accepting chemotaxis proteins. J. Bacteriol. 174:3667–
8. Corvec, S., L. Poirel, T. Naas, H. Drugeon, and P. Nordmann. 2007. Genetics
and expression of the carbapenem-hydrolyzing oxacillinase gene blaOXA-23in
Acinetobacter baumannii. Antimicrob. Agents Chemother. 51:1530–1533.
9. Damier-Piolle, L., S. Magnet, S. Bremont, T. Lambert, and P. Courvalin.
2008. AdeIJK, a resistance-nodulation-cell division pump effluxing multiple
antibiotics in Acinetobacter baumannii. Antimicrob. Agents Chemother. 52:
10. del Mar Toma ´s, M., A. Beceiro, A. Perez, D. Velasco, R. Moure, R. Villan-
ueva, J. Martinez-Beltran, and G. Bou. 2005. Cloning and functional analysis
of the gene encoding the 33- to 36-kilodalton outer membrane protein
associated with carbapenem resistance in Acinetobacter baumannii. Antimi-
crob. Agents Chemother. 49:5172–5175.
11. Depardieu, F., M. L. Foucault, J. Bell, A. Dubouix, M. Guibert, J. P. Lavigne,
M. Levast, and P. Courvalin. 2009. New combinations of mutations in
VanD-type vancomycin resistant Enterococcus faecium, Enterococcus faeca-
lis, and Enterococcus avium. Antimicrob. Agents Chemother. 53:1952–1963.
12. Depardieu, F., I. Podglajen, R. Leclercq, E. Collatz, and P. Courvalin. 2007.
Modes and modulations of antibiotic resistance gene expression. Clin. Mi-
crobiol. Rev. 20:79–114.
13. Dijkshoorn, L., A. Nemec, and H. Seifert. 2007. An increasing threat in
hospitals: multidrug-resistant Acinetobacter baumannii. Nat. Rev. Microbiol.
14. Domenech-Sanchez, A., V. J. Benedi, L. Martinez-Martinez, and S. Alberti.
2006. Evaluation of differential gene expression in susceptible and resistant
clinical isolates of Klebsiella pneumoniae by DNA microarray analysis. Clin.
Microbiol. Infect. 12:936–940.
15. Ecker, J. A., C. Massire, T. A. Hall, R. Ranken, T. T. Pennella, C. Agasino
Ivy, L. B. Blyn, S. A. Hofstadler, T. P. Endy, P. T. Scott, L. Lindler, T.
Hamilton, C. Gaddy, K. Snow, M. Pe, J. Fishbain, D. Craft, G. Deye, S.
Riddell, E. Milstrey, B. Petruccelli, S. Brisse, V. Harpin, A. Schink, D. J.
Ecker, R. Sampath, and M. W. Eshoo. 2006. Identification of Acinetobacter
species and genotyping of Acinetobacter baumannii by multilocus PCR and
mass spectrometry. J. Clin. Microbiol. 44:2921–2932.
16. Fournier, P. E., D. Vallenet, V. Barbe, S. Audic, H. Ogata, L. Poirel, H.
Richet, C. Robert, S. Mangenot, C. Abergel, P. Nordmann, J. Weissenbach,
D. Raoult, and J. M. Claverie. 2006. Comparative genomics of multidrug
resistance in Acinetobacter baumannii. PLoS Genet. 2:e7.
17. Gagnot, S., J. P. Tamby, M. L. Martin-Magniette, F. Bitton, L. Taconnat, S.
Balzergue, S. Aubourg, J. P. Renou, A. Lecharny, and V. Brunaud. 2008.
CATdb: a public access to Arabidopsis transcriptome data from the URGV-
CATMA platform. Nucleic Acids Res. 36:986–990.
18. Goldstein, F. W., A. Labigne-Roussel, G. Gerbaud, C. Carlier, E. Collatz,
and P. Courvalin. 1983. Transferable plasmid-mediated antibiotic resistance
in Acinetobacter. Plasmid 10:138–147.
19. Gribun, A., Y. Nitzan, I. Pechatnikov, G. Hershkovits, and D. J. Katcoff.
2003. Molecular and structural characterization of the HMP-AB gene en-
coding a pore-forming protein from a clinical isolate of Acinetobacter bau-
mannii. Curr. Microbiol. 47:434–443.
20. Heritier, C., L. Poirel, and P. Nordmann. 2006. Cephalosporinase over-
expression resulting from insertion of ISAba1 in Acinetobacter baumannii.
Clin. Microbiol. Infect. 12:123–130.
21. Iacono, M., L. Villa, D. Fortini, R. Bordoni, F. Imperi, R. J. Bonnal, T.
Sicheritz-Ponten, G. De Bellis, P. Visca, A. Cassone, and A. Carattoli. 2008.
Whole-genome pyrosequencing of an epidemic multidrug-resistant Acineto-
bacter baumannii strain belonging to the European clone II group. Antimi-
crob. Agents Chemother. 52:2616–2625.
22. Ko ¨hler, T., C. van Delden, L. K. Curty, M. M. Hamzehpour, and J. C.
Pechere. 2001. Overexpression of the MexEF-OprN multidrug efflux system
affects cell-to-cell signaling in Pseudomonas aeruginosa. J. Bacteriol. 183:
23. Lambert, T., G. Gerbaud, and P. Courvalin. 1994. Characterization of trans-
poson Tn1528, which confers amikacin resistance by synthesis of aminogly-
coside 3?-O-phosphotransferase type VI. Antimicrob. Agents Chemother.
24. Magnet, S., P. Courvalin, and T. Lambert. 2001. Resistance-nodulation-cell
division-type efflux pump involved in aminoglycoside resistance in Acineto-
bacter baumannii strain BM4454. Antimicrob. Agents Chemother. 45:3375–
25. Marchand, I., L. Damier-Piolle, P. Courvalin, and T. Lambert. 2004. Ex-
pression of the RND-type efflux pump AdeABC in Acinetobacter baumannii
is regulated by the AdeRS two-component system. Antimicrob. Agents Che-
26. Miller, J. F., S. A. Johnson, W. J. Black, D. T. Beattie, J. J. Mekalanos, and
S. Falkow. 1992. Constitutive sensory transduction mutations in the Borde-
tella pertussis bvgS gene. J. Bacteriol. 174:970–979.
27. Milohanic, E., P. Glaser, J. Y. Coppee, L. Frangeul, Y. Vega, J. A. Vazquez-
Boland, F. Kunst, P. Cossart, and C. Buchrieser. 2003. Transcriptome anal-
ysis of Listeria monocytogenes identifies three groups of genes differently
regulated by PrfA. Mol. Microbiol. 47:1613–1625.
28. Moubareck, C., S. Bremont, M. C. Conroy, P. Courvalin, and T. Lambert.
2009. GES-11, a novel integron-associated GES variant in Acinetobacter
baumannii. Antimicrob. Agents Chemother. 53:3579–3581.
29. Mugnier, P. D., L. Poirel, and P. Nordmann. 2009. Functional analysis of
insertion sequence ISAba1, responsible for genomic plasticity of Acineto-
bacter baumannii. J. Bacteriol. 191:2414–2418.
30. Mussi, M. A., A. S. Limansky, and A. M. Viale. 2005. Acquisition of resis-
tance to carbapenems in multidrug-resistant clinical strains of Acinetobacter
baumannii: natural insertional inactivation of a gene encoding a member of
a novel family of beta-barrel outer membrane proteins. Antimicrob. Agents
31. Peleg, A. Y., B. A. Potoski, R. Rea, J. Adams, J. Sethi, B. Capitano, S.
Husain, E. J. Kwak, S. V. Bhat, and D. L. Paterson. 2007. Acinetobacter
baumannii bloodstream infection while receiving tigecycline: a cautionary
report. J. Antimicrob. Chemother. 59:128–131.
32. Piddock, L. J. 2006. Multidrug-resistance efflux pumps—not just for resis-
tance. Nat. Rev. Microbiol. 4:629–636.
33. Ploy, M. C., F. Denis, P. Courvalin, and T. Lambert. 2000. Molecular char-
acterization of integrons in Acinetobacter baumannii: description of a hybrid
class 2 integron. Antimicrob. Agents Chemother. 44:2684–2688.
34. Poole, K. 2004. Efflux-mediated multiresistance in Gram-negative bacteria.
Clin. Microbiol. Infect. 10:12–26.
35. Ruzin, A., D. Keeney, and P. A. Bradford. 2007. AdeABC multidrug efflux
pump is associated with decreased susceptibility to tigecycline in Acineto-
bacter calcoaceticus-Acinetobacter baumannii complex. J. Antimicrob. Che-
36. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory
manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,
37. Shaikh, F., R. P. Spence, K. Levi, H. Y. Ou, Z. Deng, K. J. Towner, and K.
Rajakumar. 2009. ATPase genes of diverse multidrug-resistant Acinetobacter
baumannii isolates frequently harbour integrated DNA. J. Antimicrob. Che-
38. Smith, M. G., T. A. Gianoulis, S. Pukatzki, J. J. Mekalanos, L. N. Ornston,
M. Gerstein, and M. Snyder. 2007. New insights into Acinetobacter bauman-
nii pathogenesis revealed by high-density pyrosequencing and transposon
mutagenesis. Genes Dev. 21:601–614.
39. Su, X. Z., J. Chen, T. Mizushima, T. Kuroda, and T. Tsuchiya. 2005. AbeM, an
H?-coupled Acinetobacter baumannii multidrug efflux pump belonging to the
MATE family of transporters. Antimicrob. Agents Chemother. 49:4362–4364.
40. Szybalski, W., and V. Bryson. 1952. Genetic studies on microbial cross
resistance to toxic agents. I. Cross resistance of Escherichia coli to fifteen
antibiotics. J. Bacteriol. 64:489–499.
41. Tomaras, A. P., C. W. Dorsey, R. E. Edelmann, and L. A. Actis. 2003.
Attachment to and biofilm formation on abiotic surfaces by Acinetobacter
baumannii: involvement of a novel chaperone-usher pili assembly system.
42. Tomaras, A. P., M. J. Flagler, C. W. Dorsey, J. A. Gaddy, and L. A. Actis.
2008. Characterization of a two-component regulatory system from Acineto-
bacter baumannii that controls biofilm formation and cellular morphology.
43. Turton, J. F., M. E. Kaufmann, J. Glover, J. M. Coelho, M. Warner, R. Pike,
and T. L. Pitt. 2005. Detection and typing of integrons in epidemic strains of
Acinetobacter baumannii found in the United Kingdom. J. Clin. Microbiol.
44. Turton, J. F., M. E. Ward, N. Woodford, M. E. Kaufmann, R. Pike, D. M.
Livermore, and T. L. Pitt. 2006. The role of ISAba1 in expression of OXA
carbapenemase genes in Acinetobacter baumannii. FEMS Microbiol. Lett.
45. Vallenet, D., P. Nordmann, V. Barbe, L. Poirel, S. Mangenot, E. Bataille, C.
Dossat, S. Gas, A. Kreimeyer, P. Lenoble, S. Oztas, J. Poulain, B. Segurens,
C. Robert, C. Abergel, J. M. Claverie, D. Raoult, C. Medigue, J. Weissen-
bach, and S. Cruveiller. 2008. Comparative analysis of Acinetobacters: three
genomes for three lifestyles. PLoS One 3:e1805.
46. Yang, Y. H., S. Dudoit, P. Luu, D. M. Lin, V. Peng, J. Ngai, and T. P. Speed. 2002.
Normalization for cDNA microarray data: a robust composite method addressing
single and multiple slide systematic variation. Nucleic Acids Res. 30:e15.
340COYNE ET AL. ANTIMICROB. AGENTS CHEMOTHER.