J. Gen. Appl. Microbiol.

, 55, 191 199 (2009)

Full Paper

Identification and characterization of the AcrR/AcrAB system of

a pathogenic Edwardsiella tarda strain

Jin-hui Hou,1, 2, 3 Yong-hua Hu,1, 2 Min Zhang,1, 2 and Li Sun1, *

1 Institute of Oceanology, Chinese Academy of Sciences, Qingdao, PR China
2
Graduate University of the Chinese Academy of Sciences, Beijing, PR China
3 Xuzhou Institute of Technology, Xuzhou, PR China

(Received November 4, 2008; Accepted January 26, 2009)

Edwardsiella tarda is one of the leading marine pathogens that can infect a wide range of cul-
tured marine species. In this study, the acrR-acrAB cluster was cloned from TX1, a pathogenic E.
tarda strain isolated from diseased fish. AcrR and AcrAB were found to be involved in resistance
against acriflavine and methyl viologen, which positively regulate the expression of acrAB. AcrR
negatively regulates its own expression and the expression of the acrAB operon, most likely by
interacting with a 24-bp operator site that overlaps the putative promoter of acrA (PacrA). The
repressive effect of AcrR on PacrA could be relieved by acriflavine, methyl viologen, and ethidium
bromide, the presence of each of which enhanced transcription from PacrA. Interruption of the
regulated expression of acrR by introducing into TX1 a plasmid that overexpresses acrR affected
growth under stress conditions, AI-2 production, and bacterial virulence. In addition, mutational
analyses identified a constitutively active AcrR mutant (named N215), which exhibits full repres-
sor activity but is impaired in its ability to interact with the inducer. Overexpression of N215
produced the same kind of but moderately stronger effect on TX1 compared to that produced by
overexpression of the wild-type acrR.

Key Words—AcrAB; acriflavine; AcrR; antimicrobial resistance; methyl viologen; virulence

Introduction (Eswaran et al., 2004). To this class of efflux pump be-

 To date several classes of multidrug exporters have
been identified in prokaryotic cells, one of which is the
resistance nodulation division (RND) family transport-
ers (Poole and Srikumar, 2001). In this type of efflux
system, RND, an inner membrane proton-drug anti-
porter, forms a tripartite construct with a channel-form-
ing outer membrane protein and an inner membrane
protein of the MFP (membrane fusion protein) family
 * Address reprint requests to: Dr. Li Sun, Institute of Oceanol-
ogy, Chinese Academy of Sciences, 7 Nanhai Road, Qingdao
266071, PR China.
 Tel and Fax: 86 532 82898834
E-mail: lsun@ms.qdio.ac.cn
longs the AcrAB-TolC complex, in which AcrB, a cyto-
plasmic protein of the RND family (Yu et al., 2003), co-
operates with the outer membrane channel protein
TolC (Fralick, 1996; Koronakis et al., 2000; Tikhonova
and Zgurskaya, 2004) and the MFP family protein AcrA
to form an efflux pump that is effective against a broad
range of antimicrobial agents and toxic compounds,
resulting in the Mar (multiple antibiotic resistance)
phenotype (Nikaido, 1996; Randall and Woodward,
2002).
 In Escherichia coli the expression of the acrAB oper-
on is controlled at multiple levels via several distinct
mechanisms. On a general scale, it is modulated by
stress conditions (Ma et al., 1996) and the XylS/AraC
family regulators MarA (Barbosa and Levy, 2000),
192 HOU et al.
SoxS (White et al., 1997), Rob (Rosenberg et al., 2003),
and SidA (Rahmati et al., 2002; Wei et al., 2001); on a
local level the expression of the acrAB is controlled by
the transcription repressor AcrR, a member of the TetR
family of transcription regulators. The TetR family of
regulators possesses two functional domains, the N
terminal helix-turn-helix DNA binding domain and the
C terminal effector-binding domain (Grkovic et al.,
2002; Orth et al., 2000). Binding of the effector (or in-
ducer) to the effector domain presumably alters the
structural topology of the regulator and thus affects its
ability to interact with DNA.
 Edwardsiella tarda is a Gram-negative pathogen
with a broad range of hosts that includes animals and
humans. As the causative agent of edwardsiellosis, a
serious systematic disease of cultured fish and other
animal species, E. tarda has in recent years been rec-
ognized as one of the major marine pathogens that
cause severe losses to aquaculture industries world-
wide. Several virulence systems and factors have been
identified in E. tarda, but only a few studies on the ge-
netic mechanisms of antimicrobial resistance in E. tar-
da have been documented (Akinbowale et al., 2006;
Alcaide et al., 2006; Stock and Wiedemann, 2001) and
no link has been established between antimicrobial re-
sistance and virulence in this species. We presented in
this report the cloning and genetic analysis of the acrR-
acrAB cluster from TX1, a pathogenic E. tarda strain
isolated from diseased fish. Our results indicated that
the TX1 AcrR is an auto-regulated transcriptional re-
pressor that controls the expression of the acrAB oper-
on and that disruption of the regulated expression of
acrR has an attenuating effect on bacterial virulence.
Materials and Methods
 Bacterial strains and growth conditions. The Esch-
erichia coli strain DH5α (TaKaRa) and the Edwardsiella
tarda strain TX1 (Zhang et al., 2008) were cultured in
Luria-Bertani lysis broth (LB) medium (Sambrook et
al., 1989) at 37 C and 28 C, respectively. Appropriate
antibiotics were supplemented at the following con-
centrations: ampicillin (Ap), 100 μg ml­1; kanamycin
(Kn), 50 μg ml­1; tetracycline (Tc), 15 μg ml­1.
 Plasmid construction. The plasmids and primers
used in this study are listed in Table 1. The 284 bp DNA
fragment immediately upstream of the translational
start of acrA and the 141 bp DNA confined between
acrA and acrR (amplified by PCR with primer pairs
Vol. 55
TF6/TR5 and TF20/TR13, respectively) were inserted
at the SwaI site of pSC11, resulting in pSC284 and
pSC141R, respectively. pSC11 and pBT carrying the
mutant PacrA, PacrR, acrO, and acrR were constructed
by site-directed mutagenesis using the method of
overlap extension PCR (Ho et al., 1989). To construct
pAI, the DNA encoding the antisense RNA of the acrAB
(amplified by PCR with the primer pair TF29/TR23) was
inserted into pBT at the SmaI site, resulting in pBTAI;
the SwaI fragment of pBTAI containing the antisense
acrAB was inserted at the EcoRV site of pJRA, yielding
pAI. pAR was generated by inserting the acrR gene
into pBT at the SmaI site. pJT48 and pJN215 were cre-
ated by inserting the acrR-containing SwaI fragments
of pAR and pARD22, respectively, into pJRA at the
EcoRV site.
 Cloning of the acrR-acrAB cluster. TX1 genomic
DNA was digested with Sau3A1 and the fragments be-
tween 4 and 6 kb were ligated into pBU at the BamHI
site. DH5α was transformed with the ligation mix and
the transformants were selected as described previ-
ously (Zhang and Sun, 2007). The complete sequence
of the acrR-acrAB cluster was obtained by genome
walking as described previously (Zhang and Sun,
2007).
 Quantitative real time PCR (qRT-PCR). Total RNA
was extracted from fish livers and from cells grown in
LB medium to OD600 of 0.8 by using the SV total RNA
isolation system (Promega). qRT-PCR was carried out
in an ABI 7300 Real-time Detection System (Applied
Biosystems) by using the SYBR ExScript RT-PCR Kit
(TaKaRa). Each assay was performed in triplicate with
16S rRNA as a control. Data analyses were performed
as described previously (Zhang et al., 2008). All data
were given in terms of relative mRNA expressed as
means plus or minus standard errors of the means
(SE). Statistical analyses were performed by using the
two-tailed t-test.
 β-Galactosidase assay. This was carried out as de-
scribed previously (Sun et al., 1998).
 Experimental infection. Japanese flounder, weigh-
ing ∼13 g each, were randomly divided into several
groups (5 fish/group). Each group was injected intra-
peritonealy (i.p.) with 1×105 CFU of TX1/pJRA or TX1/
pJT48 or TX1/pJN215 that had been cultured to OD600
of 0.5 in LB medium, washed, and resuspended in
phosphate-buffered saline (PBS). To examine bacterial
dissemination, the livers of the infected fish were re-
moved under sterile conditions and homogenized with
2009 Analysis of Edwardsiella tarda AcrR/AcrAB 193

Table 1. Plasmids and primers used in this study.

Plasmid or primer Relevant characteristics Source or reference

Plasmids
 pBT ApR; cloning vector Zhang et al., 2008
 pBU KnR; signal sequence trap Zhang and Sun, 2007
 pAI ApR; expressing acrAB antisense RNA This study
 pAR ApR; expressing acrR This study
 pJRA ApR; broad host range vector Zhang et al., 2008
 pJN215 ApR; pJRA expressing N215 This study
 pJT48 ApR; pJRA expressing acrR This study
 pARD5 ApR; expressing acrR bearing deletion of the last 5 residues This study
 pARD12 ApR; expressing acrR bearing deletion of the last 12 residues This study
 pARD22 ApR; expressing acrR bearing deletion of the last 22 residues This study
 pARD32 ApR; expressing acrR bearing deletion of the last 32 residues This study
 pSC11 KnR; promoter probe plasmid Zhang et al., 2008
 pSC284 KnR; carrying PacrA-lacZ fusion This study
 pSC284M1 KnR; pSC284 with mutated ­10 This study
 pSC284M3 KnR; pSC284 with mutated acrO This study
 pSC141R KnR; carrying PacrR-lacZ fusion This study
 pSC141RM1 KnR; pSC141R with mutated ­10 This study
Primers Sequences (5 →3 )a
 TF6 ATTTAAATGGCCCCGCGG (SwaI)
 TF20 GATATCGGCGTTGGTTTTTTAAAG (EcoRV)
 TF29 ATTTAAATAGCGCAGCAATGGGT (SwaI)
 TR5 ATTTAAATACCTCAAGAGTCCGAGT (SwaI)
 TR13 GATATCATTATACCTCAAGAGTCC (EcoRV)
 TR23 ATTTAAATCTATTTATTGTGCGGTTCAAC (SwaI)

 aUnderlined nucleotides are restriction sites of the enzymes indicated in parentheses at the ends of the sequences.

glass homogenizers. The homogenates were plated
on LB plates supplemented with ampicillin and tetra-
cycline. After incubation at 28 C for 48 h, the colonies
that appeared on the plates were enumerated. The na-
ture of these colonies was verified by PCR analysis of
20 colonies randomly selected from each plate using
primers specific to TX1 and the plasmid; one of the 20
PCR products was subsequently analyzed by DNA se-
quencing. Statistical analysis was performed by using
the t-test.
 AI-2 assay. This was carried out as described pre-
viously (Zhang et al., 2008). Briefly, overnight cultures
of bacteria grown in LB medium at 28 C were diluted
1：100 in fresh LB medium; 2 ml of cell culture was
taken every 30 min, from which the cell-free superna-
tant was obtained by centrifugation and then filtering
through a 0.22-μm filter (Millipore). For measurement
of bioluminescence induction, overnight culture of the
V. harveyi strain BB170 (American Type Culture Collec-
tion (ATCC)) grown in AB medium at 28 C was diluted
1：5,000 in fresh AB medium supplemented with 10%
cell-free supernatant (as prepared above) of the tested
strains or with the growth medium (as the control). The
growth was continued and light production was mea-
sured by using a Glomax luminometer (Promega).
 Bacterial conjugation. pJRA and its variants were
introduced into the E. coli strain S17-1λpir (Biomedal,
Spain) by transformation. The transformants and TX1
were grown in LB medium to an OD600 of 1 and mixed
in 1：1 ratio. The mixed cells were washed and resus-
pended in 10 mM MgSO4 and dropped onto a LB plate.
After incubation at 28 C for 12 h, the growth on the
plate was scraped off and resuspended in 1 ml LB,
from which 100 μl was taken and spread onto a LB
plate supplemented with ampicillin and tetracycline.
The plate was incubated at 28 C for 48 h, and the colo-
nies that appeared were verified to be authentic
transconjugants by PCR and sequence analysis of the
PCR products.
 Database search and nucleotide sequence acces-
sion numbers. A database search was conducted
using the BLAST programs at the NCBI (National Cen-
194
ter for Biotechnology Information). A signal peptide
search was performed with the SignalP 3.0 server. The
nucleotide sequence of the acrR-acrAB cluster has
been deposited in GenBank database under the ac-
cession number EU082215.
Results
Characterization of the acrR-acrAB cluster of TX1
 The acrR-acrAB cluster of TX1 was cloned via the
pBU system, a signal sequence trap developed in our
laboratory for the identification of proteins with func-
tional secretion domains (Zhang and Sun, 2007). pBU
is a pBR322-based expression plasmid that carries the
coding sequence of a secretion-defective reporter
AgaV (an extracellular agarase). Since AgaV lacks its
own signal peptide, it remains in the cytoplasm. How-
ever, if a heterologous DNA encoding a secretion pro-
tein is inserted upstream of and forms an in-frame fu-
sion with agaV, then AgaV may be transported out of
the cell as a passenger protein by the fused heterolo-
gous protein. Since AgaV, once being secreted out of
the cell, can degrade agar, AgaV-secreting colonies
have a visible pit around them on LB agar plates and
therefore can be easily detected. In the selection of the
acrR-acrAB cluster, the TX1 genomic DNA selected by
the pBU system is a 1,005 bp DNA covering the 189
bp upstream and the 816 bp downstream of the trans-
lation start of acrA (Fig. 1). The up- and down-stream
regions of this 1,005 bp DNA were then obtained by
genome walking using primers based on the sequence
of this 1,005 bp DNA. In this fashion, the entire acrR-
acrAB cluster was cloned. Sequence analysis showed
that acrA and acrB are arranged in tandem and pre-
ceded by acrR, which is transcribed divergently from
Fig. 1. Schematic representation of the acrR-acrAB cluster.
 The open arrows of acrR, acrA, and acrB indicate the direc-
tion of transcription. The small solid arrows under acrA repre-
sent internal primers used in genome walking. The regions of
the cluster obtained by using the pBU system and genome
walking (GW) are indicated. The region of acrR encoding N215
is represented by a solid arrow.
HOU et al. Vol. 55
the former. AcrR, AcrA, and AcrB share the highest se-
quence identities with their respective counterparts in
Yersinia enterocolitica (56%), Salmonella enterica
(74%), and Y. enterocolitica (82%) (GenBank acces-
sion nos. AM286415 and AL627267, respectively). Pu-
tative signal peptide sequences, consisting of the first
24 and 33 amino acid residues respectively, were iden-
tified in AcrA and AcrB. Like in many RND family pro-
teins (Tseng et al., 1999; Yu et al., 2003), 12 transmem-
brane helices were found in AcrB by using the TMHMM
Server (v.2.0). Two large periplasmic loops (between
the signal peptide and helix 1 and between helices 6
and 7, respectively), which are another characteristic
of the RND exporters, were also found in AcrB. These
data indicated that the TX1 acrR-acrAB cluster pre-
sumably encodes an AcrAB efflux system similar to
those identified in other bacterial species.
acrAB expression is required for resistance against ac-
riflavine and methyl viologen
 TX1 exhibits relatively high levels of tolerance to tet-
racycline, acriflavine, and methyl viologen (MV) (MICs
of 20, 35, and 500 μg/ml, respectively), which are
known substrates of some AcrAB efflux systems (Tsu-
kagoshi and Aono, 2000; White et al., 1997). To deter-
mine whether the TX1 AcrAB was involved in resis-
tance against these compounds, the expression of
acrAB was attenuated by antisense RNA interference
and its effect on drug resistance was examined. For
this purpose, the plasmid pAI, which expresses the an-
tisense RNA of acrAB (corresponding to the region
between the translation stop codon of acrB and posi-
tion ­70 relative to the translational start of acrA), was
introduced into TX1 via conjugation. Production of the
antisense RNA in the transconjugant, TX1/pAI, was
confirmed by qRT-PCR (data not shown). qRT-PCR
analyses showed that acrA and acrB expressions in
TX1/pAI were reduced 2.9- and 3.8-fold, respectively,
compared to those in TX1 harboring the control plas-
mid pJRA (Fig. 2). Resistance analyses indicated that
TX1/pAI exhibited reduced tolerance to acriflavine and
MV (MICs of 20 and 250 μg/ml, respectively) but re-
tained the same level of resistance to tetracycline.
Consistent with this observation, transcription of acrA
was significantly enhanced by acriflavine and MV (Fig.
3) but not affected by tetracycline (data not shown).
These results demonstrated that acrAB expression
was positively regulated by acriflavine and MV and
that full expression of acrAB was required to achieve
2009 Analysis of Edwardsiella tarda AcrR/AcrAB 195

Fig. 4. Nucleotide sequence of the intergenic region be-

Fig. 2. Quantitative real time (qRT-PCR) analysis of the ex-
tween acrR and acrA.
pression of acrA and acrB in TX1/pJRA and TX1/pAI.
 The translation starts of acrR and acrA, respectively, are in
 Total RNA was prepared from cells grown in LB medium to
capital letters at the ends of the sequence; the acrO site is in
OD600 ∼ 0.8 and used for qRT-PCR. The mRNA levels of acrA
italics; the putative ­10 and ­35 elements of PacrR and PacrA
and acrB were normalized to that of the 16S rRNA. Data are
are underlined and in capital letters.
presented as the means ± SE. **, P ＜ 0.01.

Table 2. acrA and acrR promoter activities in the absence

and presence of AcrR.

Relative β-galactosidase
Strain activity (%)
­ acriflavine + acriflavine
DH5α/pSC284 100
DH5α/pSC284M1 3.2
Fig. 3. Quantitative real time (qRT-PCR) analysis of the ex- DH5α/pSC141R 420
pression of acrA in relation to acriflavine and methyl viologen. DH5α/pSC141RM1 52.5
 Total RNA was prepared from TX1 grown in LB medium DH5α/pSC284/pBT 100 100
(hatched bar) and in LB medium supplemented with 10 μg/ml DH5α/pSC284/pAR 0.3 3
of acriflavine (AF, black bar) or 50 μg/ml of methyl viologen (MV, DH5α/pSC284/pARD5 0.4 3
open bar). The mRNA levels of acrA were normalized to that of DH5α/pSC284/pARD12 0.4 1
the 16S rRNA. Data are presented as the means ± SE. **, P ＜ DH5α/pSC284/pARD22 0.5 1
0.01. DH5α/pSC284/pARD32 96.4 100

β-Galactosidase assay was performed using cells gown in LB

the full resistance against acriflavine and MV observed
in the natural state of TX1. These data supported the
notion that AcrAB was required for acriflavine and MV
resistance in TX1.
Identification of the putative promoters of acrA and
acrR
To locate the promoter of acrA, the 284 bp DNA frag-
ment immediately upstream of the translational start of
acrA was cloned into the promoter probe plasmid
pSC11 so that in the recombinant plasmid, pSC284,
the 284 bp DNA is fused to a promoterless lacZ re-
porter gene. pSC284 was introduced into the E. coli
strain DH5α by transformation, and the transformants
formed blue colonies on X-gal plates, suggesting that
the 284 bp DNA upstream of acrA possesses an active
promoter. Inspection of this DNA fragment revealed a
putative σ70-dependent promoter (named PacrA) with
CTTACA and TACCAT as the ­35 and ­10 box, re-
medium with (+) or without (­) acriflavine (10 μg/ml) to OD600
of 1. β-Galactosidase activities are presented as percentages of
that of DH5α/pSC284.
spectively (Fig. 4). Mutational study indicated that
DH5α carrying pSC284M1, which is identical to
pSC284 except that the ­10 region of PacrA is mutated
to TACCTA, which bears less resemblance to the con-
sensus ­10 element, exhibited 31-fold less β-ga-
lactosidase activity (5 Miller units) than DH5α/pSC284
(160 Miller units) (Table 2). These results suggested
that the promoter activity observed with the 284 bp
DNA upstream of acrA was conferred by PacrA.
 To locate the promoter of acrR, the 141 bp DNA con-
fined between acrA and acrR was cloned into pSC11
and the recombinant plasmid, pSC141R, was intro-
duced into DH5α by transformation. The transformant,
DH5α/pSC141R, produced 4.2-fold more β-galacto-
sidase activity (672 Miller units) than DH5α/pSC284,
196 HOU et al.
suggesting that the 141 bp DNA harbored a functional
promoter. Consistently, a potential σ70-dependent pro-
moter (named PacrR) was identified in this DNA frag-
ment (Fig. 4). Mutational analyses indicated that DH5α
carrying pSC141RM1, which is identical to pSC141R
except that the ­10 element (TATAAA) of PacrR is mu-
tated to TAATTA, which is divergent from the consen-
sus ­10 element, produced 8-fold less β-galactosidase
activity (84 Miller units) than DH5α/pSC141R (Table 2),
suggesting that the promoter activity of the 141 bp
DNA confined between acrA and acrR was conferred
by PacrR.
AcrR negatively regulates transcription from both PacrA
and PacrR
 To investigate whether AcrR could regulate acrAB
and its own expression, DH5α/pSC284 and DH5α/
pSC141R were transformed separately with the con-
trol plasmid pBT and the plasmid pAR, which express-
es acrR. Subsequent β-galactosidase assays showed
that the β-galactosidase activities of DH5α/pSC284/
pAR and DH5α/pSC141R/pAR (0.5 and 224 Miller
units, respectively) were, respectively, 320- and 3-fold
lower than those of DH5α/pSC284/pBT and DH5α/
pSC141R/pBT (160 and 670 Miller units, respectively),
suggesting that AcrR repressed transcription from
both PacrA and PacrR. Consistently, sequence inspec-
tion identified a potential AcrR binding site in the form
of a 24 bp palindrome-forming sequence (5 -TACATG-
CATTGATGAATGTATGTA-3 , designated acrO) that
overlaps PacrA (Fig. 4). To examine the potential signifi-
cance of acrO in the functioning of AcrR, the dyad
symmetry of acrO was disrupted by mutating the TG-
CATT sequence to AGGTAA; the plasmid pSC284M3,
which carries the 284 bp DNA upstream of acrA bear-
ing the mutated acrO, was introduced into DH5α and
DH5α/pAR by transformation. Subsequent β-galacto-
sidase assays showed that the β-galactosidase activity
of DH5α/pSC284M3 (151 Miller units) was similar to
that of DH5α/pSC284 (160 Miller units) but the β-ga-
lactosidase activity of DH5α/pAR/pSC284M3 (138
Miller units) was 276-fold higher than that of DH5α/
pAR/pSC284 (0.5 Miller units). These data demon-
strated that mutation of acrO had no impact on the
activity of PacrA but abolished the repressive effect of
AcrR on PacrA, which suggested that acrO was essen-
tial to the interaction between AcrR and PacrA.
Vol. 55
Acriflavine, MV, and ethidium bromide are among the
inducers of AcrR
 TetR-like proteins are known to respond to certain
effector molecules that can interact with and thereby
abolish the DNA-binding capacity of the regulator. To
find out the inducers of the TX1 AcrR, DH5α/pSC284/
pAR was cultured in the presence of chloramphenicol,
streptomycin, tetracycline, acriflavine, MV, and ethidi-
um bromide, respectively. Subsequent β-galactosidase
assay indicated that acriflavine (10 μg/ml), ethidium
bromide (12.5 μg/ml), and MV (80 μg/ml) caused, re-
spectively, 10-, 6-, and 12-fold increase in the β-galac-
tosidase activity of DH5α/pSC284/pAR, suggesting
that these compounds could function as AcrR induc-
ers and abolish the ability of AcrR to interact with the
target promoter. In contrast, chloramphenicol, strepto-
mycin, and tetracycline had no apparent effect on the
β-galactosidase activity of DH5α/pSC284/pAR.
Effect of uncontrolled overexpression of acrR
 Since acrR expression is autoregulated, we won-
dered what would ensue from disruption of this regula-
tion. To investigate this question, the plasmid pJT48,
which constitutively expresses acrR, was conjugated
into TX1. qRT-PCR analyses showed that acrR expres-
sion in the transconjugant, TX1/pJT48, was increased
∼7-fold whereas that of acrA was decreased ∼4-fold
compared to those in TX1/pJRA, suggesting that in
TX1/pJT48, acrR was indeed overexpressed, which in
turn repressed the expression of acrA. Growth studies
showed that the growth profile of TX1/pJT48 was simi-
lar to that of TX1/pJRA when grown in LB medium but
differed from the latter when grown in LB medium sup-
plemented with acriflavine, under which condition TX1/
pJT48 exhibited slower doubling time and lower maxi-
mum cell density (Fig. 5). Since our recent study
(Zhang et al., 2008) showed that in E. tarda, the LuxS/
AI-2 quorum sensing system is involved in bacterial
growth and pathogenicity, we also examined AI-2 pro-
duction in TX1/pJT48. The result showed that TX1/
pJT48 displayed 8.1-fold less AI-2 activity than TX1/
pJRA.
 To investigate whether overexpression of acrR had
any effect on bacterial virulence, Japanese flounder
were infected with the same dose of TX1/pJT48 or
TX1/pJRA. The livers of the fish were taken at 24 h
post-infection and homogenized in PBS. The homoge-
nates were plated on selective LB plates and exam-
ined for bacterial recovery. The results showed that the
2009 Analysis of Edwardsiella tarda AcrR/AcrAB
Fig. 5. Growth profiles of TX1/pJT48 and TX1/pJRA under
different conditions.
 Cells grown overnight in LB medium were diluted to 1×106
CFU/ml in fresh LB medium supplemented with or without acri-
flavine (AF, 10 μg/ml); growth was continued into the stationary
phase and aliquots of the culture were taken at various time
points for the measurement of absorbance at OD600.
average amount of the recovered TX1/pJT48 was 22-
fold lower than that of the recovered TX1/pJRA. To ex-
amine whether the difference in bacterial recovery be-
tween TX1/pJT48 and TX1/pJRA was due to selective
loss of pJT48 during infection, the loss rate of pJT48
was determined. For this purpose, the liver homoge-
nates (as described above) of the TX1/pJT48-infected
fish were plated on both LB plates supplemented with
tetracycline alone (for the selection of TX1 and TX1/
pJT48) and LB plates supplemented with tetracycline
plus ampicillin (for the selection of TX1/pJT48). The
result showed that the number of TX1/pJT48 that ap-
peared on the tetracycline plus ampicillin plates was
12.1% lower than the number of colonies that ap-
peared on the tetracycline plates; PCR analysis using
TX1- and pJT48-specific primers indicated that 88.7%
(71/80) of the colonies on the tetracycline plates were
TX1/pJT48. Hence, the plasmid loss rate of pJT48 was
at maximum 12.1%, which was comparable to that of
pJRA (∼10%, Zhang et al., 2008). Taken together,
these results demonstrated that uncontrolled overex-
pression of acrR impaired the tissue dissemination
and survival ability of TX1.
Identification and analysis of a mutant AcrR that is only
partially inducible
 AcrR possesses an N-terminal DNA-binding motif
(16 62) and a C-terminal effector domain (84 204)
that are conserved among the TetR family of proteins.
To examine the essentialness of the C-terminal region
to the overall activity of AcrR, the plasmids pARD5,
pARD12, pARD22, and pARD32, which express acrR
bearing deletions of the C-terminal 5, 12, 22, and
197
32 residues respectively, were introduced into DH5α/
pSC284 by transformation. Subsequent β-galactosi-
dase assays showed that the β-galactosidase activity
of DH5α/pSC284/pARD32 was 321-fold more than that
of DH5α/pSC284/pAR, but the β-galactosidase activi-
ties of all other transformants were comparable to that
of DH5α/pSC284/pAR (Table 2). These results indicat-
ed that the C-terminal 5, 12, and 22 amino acid resi-
dues were not essential to the activity of AcrR. When
acriflavine was present, the β-galactosidase activity of
DH5α/pSC284/pARD5 was of the same level as that of
DH5α/pSC284/pAR, whereas the β-galactosidase ac-
tivities of DH5α/pSC284/pARD12 and DH5α/pSC284/
pARD22 were only one third of that of DH5α/pSC284/
pAR (Table 2); therefore the last 5 amino acid residues
(Val233 to Glu237) of AcrR were not essential to the in-
duction of PacrA but the 17 amino acid residues (Pro216
to Cys232) preceding Val233 were required for full induc-
tion of PacrA by acriflavine.
 Since the mutant AcrR (named N215, Fig. 1) bearing
deletion of the last 22 residues was only partially in-
ducible and retained more repressor activity than the
wild-type AcrR in the presence of the inducer (Table 2),
we examined its effect on AI-2 production and bacte-
rial virulence. The results showed that the AI-2 activity
produced by TX1 harboring pJN215, which expresses
N215, was 15-fold less than that produced by TX1/
pJRA. Bacterial dissemination analyses (performed as
described above) showed that the bacterial recovery
from the livers of TX1/pJN215-infected fish was 29-fold
less than the bacterial recovery from the livers of TX1/
pJRA-infected fish. Taken together, these results dem-
onstrated that, like overexpression of the wild-type
acrR, overexpression of N215 in TX1 attenuated AI-2
activity and bacterial virulence, but the effect of N215
appeared to be moderately stronger than that of the
wild-type acrR.
Discussion
 As multi-drug efflux pumps, the AcrAB exporters of
many bacterial species are responsible for resistance
against a wide range of chemical and biochemical
compounds that include antimicrobial agents, deter-
gents, and dyes (Cobos et al., 2007; Su et al., 2007).
The genetic mechanism of multiple resistances in-
volves the interaction of chemical compounds with
AcrR and the subsequent induction of the AcrAB efflux
systems (Ahmed et al., 1994; Zgurskaya and Nikaido,
198
2000). In TX1 we found that acrAB expression was re-
quired for resistance against acriflavine and MV but
not for resistance against tetracycline. In line with this
observation, repression of acrAB expression by AcrR
was relieved by acriflavine and MV but not by tetracy-
cline, chloramphenicol, or streptomycin, which are
substrates of AcrAB pumps identified in other bacteria.
Although it is possible that in our system, in which the
AcrR target promoter (i.e., PacrA) exists in multiple cop-
ies, some subtle inducing effects may have escaped
detection, these results nevertheless suggested that
the TX1 AcrR may have an effector range, which partly
defines the substrate range of the AcrAB pump, that is
different from or more restricted compared to that of
known AcrR.
 In prokaryotic cells, transcriptional repression can
happen via several mechanisms. Studies of the E. coli
TetR and the Staphylococcus aureus QacR, a TetR-like
repressor that controls the expression of the QacA
multidrug transporter (Brown and Skurray, 2001), have
indicated that transcriptional inhibition of tetA by TetR
is through blocking the access of RNA polymerase to
the promoter whereas repression of qacA by QacR is
via impeding the isomerization process (Grkovic et al.,
1998). In TX1, since the acrO site is located down-
stream of PacrR and overlaps PacrA, it is likely that inhibi-
tion of PacrA by AcrR is through spatial blockage which
prevents the interaction between PacrA and RNA poly-
merase, whereas inhibition of PacrR is by hindering the
formation of the DNA-RNA polymerase open complex
that is required for transcriptional initiation. The obser-
vation that the repression of AcrR upon PacrA is strong-
er than that upon PacrR is supportive of this hypothesis,
though it can also be accounted for by the fact that
PacrR is a stronger promoter than PacrA and therefore is
less repressible.
 For pathogens that have to survive within a host it is
vital to possess the facilities that enable them to cope
with the unfavorable conditions encountered within
the host environment, such as the presence of anti-
bacterium compounds produced by the host. For this
reason multidrug efflux transporters constitute part of
the virulence mechanisms that are important for bacte-
rial infection and survival. Recently Posadas et al.
(2007) have demonstrated that BepC, a TolC homo-
logue identified in Brucella suis, is involved in patho-
genesis. In our study we found that disruption of the
regulated expression of acrR by overexpressing acrR
or N215 retarded the growth of TX1 under stress con-
HOU et al. Vol. 55
ditions caused by acriflavine. These results raised the
possibility that stress conditions induced by other yet
unidentified substrates of the AcrR/AcrAB system may
also impair the growth and infectivity of TX1. The ob-
servation that acrR and N215 overexpression affected
AI-2 activity suggested that AcrR may regulate not only
the acrAB operon but also other cellular systems.
Since the temporal production of AI-2 is required for
full bacterial virulence, the reduction of AI-2 activity in
TX1/pJT48 and TX1/pJN215 caused by acrR and N215
overexpression probably contributed to the decreased
tissue dissemination and survival ability of these
strains. Based on our result, it is reasonable to specu-
late that the TX1 AcrAB transporter is involved in the
efflux of biological compounds produced by the host
as a defense mechanism against bacterial invasion;
constitutive production of AcrR leads to constitutive
repression of the acrAB operon and thus the accumu-
lation of toxic compounds in the bacteria which in turn
vitiates the survival ability of the pathogen.
Acknowledgments
 This work was supported by the National Natural Science
Foundation of China (NSFC) grant 40676090 and the 863 High
Technology Project grant 2008AA092501.
References
Ahmed, M., Borsch, C. M., Taylor, S. S., Vazques-Laslop, N.,
and Neyfakh, A. A. (1994) A protein that activates expres-
sion of a multidrug efflux transporter upon binding the
transporter substrates. J. Biol. Chem., 269, 28506 28513.
Akinbowale, O. L., Peng, H., and Barton, M. D. (2006) Antimi-
crobial resistance in bacteria isolated from aquaculture
sources in Australia. J. Appl. Microbiol., 10, 1103 1113.
Alcaide, E., Herraiz, S., and Esteve, C. (2006) Occurrence of
Edwardsiella tarda in wild European eels Anguilla anguilla
from Mediterranean Spain. Dis. Aquat. Organ., 73, 77 81.
Barbosa, T. M. and Levy, S. B. (2000) Differential expression of
over 60 chromosomal genes in Escherichia coli by consti-
tutive expression of MarA. J. Bacteriol., 182, 3467 3474.
Brown, M. H. and Skurray, R. A. (2001) Staphylococcal multi-
drug efflux protein QacA. J. Mol. Biotechnol., 3, 163 170.
Cobos, S. G., Campos, J., Lazaro, E., Roman, F., Cercenado,
E., Garcia-Rey, C., Perez-Vazquez, M., Oteo, J., and de
Abajo, F. (2007) Ampicillin-resistant non-β-lactamase-
producing Haemophilus influenzae in Spain: Recent emer-
gence of clonal isolates with increased resistance to cefo-
taxime and cefixime. Antimicrob. Agents Chemother., 51,
2564 2573.
Eswaran, J., Koronakis, E., Higgins, M. K., Hughes, C., and Ko-
2009 Analysis of Edwardsiella tarda AcrR/AcrAB
ronakis, V. (2004) Three s company: Component structures
bring a closer view of tripartite drug efflux pumps. Curr.
Opin. Struct. Biol., 14, 741 747.
Fralick, J. A. (1996) Evidence that TolC is required for function-
ing of the Mar/AcrAB efflux pump of Escherichia coli. J.
Bacteriol., 178, 5803 5805.
Grkovic, S., Brown, M. H., Roberts, N. J., Paulsen, I. T., and
Skurray, R. A. (1998) QacR is a repressor protein that regu-
lates expression of the Staphylococcus aureus multidrug
efflux pump QacA. J. Biol. Chem., 273, 18665 18673.
Grkovic, S., Brown, M. H., and Skurray, R. A. (2002) Regulation
of bacterial drug export systems. Microbiol. Mol. Biol. Rev.,
66, 671 701.
Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L.
R. (1989) Site-directed mutagenesis by overlap extension
using the polymerase chain reaction. Gene, 77, 51 59.
Koronakis, V., Sharff, A., Koronakis, E., Luisi, B., and Hughes, C.
(2000) Crystal structure of the bacterial membrane protein
TolC central to multidrug efflux and protein export. Nature,
405, 914 919.
Ma, D., Alberti, M., Lynch, C., Nikaido, H., and Hearst, J. E.
(1996) The local repressor AcrR plays a modulating role in
the regulation of acrAB genes of Escherichia coli by global
stress signals. Mol. Microbiol., 19, 101 112.
Nikaido, H. (1996) Multidrug efflux pumps of gram-negative
bacteria. J. Bacteriol., 178, 5853 5859.
Orth, P., Schnappinger, D., Hillen, W., Saenger, W., and Hinrichs,
W. (2000) Structural basis of gene regulation by the tetracy-
cline inducible Tet repressor-operator system. Nat. Struct.
Biol., 7, 215 219.
Poole, K. and Srikumar, R. (2001) Multidrug efflux in Pseudomo-
nas aeruginosa: Components, mechanisms and clinical
significance. Curr. Top. Med. Chem., 1, 59 71.
Posadas, D. M., Martin, F. A., Sabio y Garcia, J. V., Spera, J. M.,
Delpino, M. V., Baldi, P., Campos, E., Cravero, S. L., and
Zorreguieta, A. (2007) The TolC homologue of Brucella suis
is involved in resistance to antimicrobial compounds and
virulence. Infect. Immun., 75, 379 389.
Rahmati, S., Yang, S., Davidson, A. L., and Zechiedrich, E. L.
(2002) Control of the AcrAB multidrug efflux pump by quo-
rum-sensing regulator SdiA. Mol. Microbiol., 43, 677 685.
Randall, L. P. and Woodward, M. J. (2002) The multiple antibi-
otic resistance (mar) locus and its significance. Res. Vet.
Sci., 72, 87 93.
Rosenberg, E. Y., Bertenthal, D., Nilles, M. L., Bertrand, K. P.,
and Nikaido, H. (2003) Bile salts and fatty acids induce the
expression of Escherichia coli AcrAB multidrug efflux pump
through their interaction with Rob regulatory protein. Mol.
Microbiol., 48, 1609 1619.
Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular
199
Cloning: A Laboratory Manual, 2nd ed., Cold Spring Har-
bor Laboratory Press, Cold Spring Harbor, N.Y.
Stock, I. and Wiedemann, B. (2001) Natural antibiotic suscepti-
bilities of Edwardsiella tarda, E. ictaluri, and E. hoshinae.
Antimicrob. Agents Chemother., 45, 2245 2255.
Su, C. C., Rutherford, D. J., and Yu, E. W. (2007) Characteriza-
tion of the multidrug efflux regulator AcrR from Escherichia
coli. Biochem. Biophys. Res. Commun., 361, 85 90.
Sun, L., vanderSpek, J., and Murphy, J. (1998) Isolation and
characterization of iron-independent positive dominant
mutants of diphtheria toxin repressor, DtxR. Proc. Natl.
Acad. Sci. USA, 95, 14985 14990.
Tikhonova, E. B. and Zgurskaya, H. I. (2004) AcrA, AcrB, and
TolC of Escherichia coli form a stable intermembrane mul-
tidrug efflux complex. J. Biol. Chem., 279, 32116 32124.
Tseng, T. T., Gratwick, K. S., Kollman, J., Park, D., Nies, D. H.,
Goffeau, A., and Saier, M. H. (1999) The RND permease
superfamily: An ancient, ubiquitous and diverse family that
includes human disease and developmental proteins. J.
Mol. Microbiol. Biotechnol., 1, 107 125.
Tsukagoshi, N. and Aono, R. (2000) Entry into and release of
solvents by Escherichia coli in an organic-aqueous tow-
liquid-phase system and substrate specificity of the AcrAB-
TolC solvent-extruding pump. J. Bacteriol., 182, 4803
4810.
Wei, Y., Lee, J. M., Smulski, D. R., and LaRossa, R. A. (2001)
Global impact of sdiA amplification revealed by compre-
hensive gene expression profiling of Escherichia coli. J.
Bacteriol., 183, 2265 2272.
White, D. G., Goldman, J. D., Demple, B., and Levy, S. B. (1997)
Role of the acrAB locus in organic solvent tolerance medi-
ated by expression of marA, soxS, or robA in Escherichia
coli. J. Bacteriol., 179, 6122 6126.
Yu, E. W., Aires, J. R., and Nikaido, H. (2003) AcrB multidrug
efflux pump of Escherichia coli: Composite substrate-bind-
ing cavity of exceptional flexibility generates its extremely
wide substrate specificity. J. Bacteriol., 185, 5657 5664.
Yu, E. W., McDermott, G., Zgurskaya, H. I., Nikaido, H., and
Koshland, D. E. (2003) Structural basis of multiple drug-
binding capacity of the AcrB multidrug efflux pump. Sci-
ence, 300, 976 980.
Zgurskaya, H. I. and Nikaido, H. (2000) Multidrug resistance
mechanisms: Drug efflux across two membranes. Mol. Mi-
crobiol., 37, 219 225.
Zhang, M., Sun, K., and Sun, L. (2008) Regulation of autoin-
ducer 2 production and luxS expression in a pathogenic
Edwardsiella tarda strain. Microbiology, 154, 2060 2069.
Zhang, W. and Sun, L. (2007) Cloning, characterization and mo-
lecular application of a β-agarase gene from Vibrio sp.
strain V134. Appl. Environ. Microbiol., 73, 2825 2831.