Evaluation and target validation of indole derivatives as inhibitors of the AcrAB-TolC efflux pump

Article abstract of DOI:10.1271/bbb.100433

Indole derivatives 3-amino-6-carboxyl-indole and 3-nitro-6-amino-indole were designed and synthesized based on the TolC structure. They proved to have potent synergistic antibacterial effects on chloramphenicol, tetracycline, erythromycin, and ciprofloxacin against Escherichia coli YD2 and FJ307 with decreased minimal inhibitory concentrations (MICs) at 2-64 folds. To research its functional site, Escherichia coli BL21(DE3)-3 expressing a target-site mutated TolC was constructed by red homologous recombination and the site-directed mutagenesis technique. They did not noticeably affect antimicrobial activity against BL21(DE3)-3. All the results indicate that these compounds match our design and can be developed as efflux pump inhibitors for the AcrAB-TolC efflux pump.

Full text of DOI:10.1271/bbb.100433

Biosci. Biotechnol. Biochem., 74 (11), 2237–2241, 2010
Evaluation and Target Validation of Indole Derivatives as Inhibitors
of the AcrAB-TolC Efflux Pump
Bo ZENG, Hongning WANG,y Likou ZOU, Anyun ZHANG, Xin YANG, and Zhongbin GUAN
Animal Disease Prevention and Food Safety Key Laboratory of Sichuan Province, School of Life Sciences,
Sichuan University, Chengdu, 610064, China
Received June 9, 2010; Accepted August 24, 2010; Online Publication, November 7, 2010
[doi:10.1271/bbb.100433]
Indole derivatives 3-amino-6-carboxyl-indole and
3-nitro-6-amino-indole were designed and synthesized
based on the TolC structure. They proved to have potent
synergistic antibacterial effects on chloramphenicol,
tetracycline, erythromycin, and ciprofloxacin against
Escherichia coli YD2 and FJ307 with decreased minimal
inhibitory concentrations (MICs) at 2–64 folds. To re-
search its functional site, Escherichia coli BL21(DE3)À3
expressing a target-site mutated TolC was constructed
by red homologous recombination and the site-directed
mutagenesis technique. They did not noticeably affect
antimicrobial activity against BL21(DE3)À3. All the
results indicate that these compounds match our design
and can be developed as efflux pump inhibitors for the
AcrAB-TolC efflux pump.
Functional genomics techniques are valuable tools to
evaluate the pool of targets, and can exploited for target
validation and the determination of molecule action.
Knockout analyses and mutation studies to create mutant
strains with precise genetic alteration aid in the selection
and validation of novel potential targets.7) The applica-
tion of targeted gene knockout in identifying and
validating new drug targets and vaccine candidates in
M. tuberculosis is practiced.8) Hence, to further validate
the target of indole derivatives in our study, the strain
expressing mutant TolC had to be constructed by
knockout and site-directed mutation.
The aim of the present study was to design and
synthesize indole derivatives in order to restore the
antibiotic susceptibility of resistant E. coli strains and to
identify the functional target of indole derivatives to
improve the validity of molecular designing.
Key words: antibacterial activity; efflux pump inhibitor;
indole derivatives; TolC
Materials and Methods
In China, colibacillosis in pigs and chicken is still
prevalent, and multi-drug resistance in Escherichia coli
is also severe. The AcrAB-TloC efflux pump as the
major drug efflux system of E. coli plays an important
role in this. Hence efflux pump inhibitors (EPIs),
especially those active against multi-drug resistance
pumps, are of great interest. Small molecules with
conjugated aromatic rings, such as phenyls, naphtha-
lenes, and indoles are of high potential in EPI re-
search.1,2) Indolenine and its derivatives have been
generally developed as protein inhibitors of key targets
antivirus, antitumor, and antibacterial efforts.3,4) Espe-
cially, 2-aryl-5-nitro-1H-indoles have been reported as
NorA EPI in Staphylococcus aureus.5) The crystal
structures of TolC have now been solved. The allosteric
mechanism of TolC from a closed state to an open state
has been analyzed. The tunnel helices of TolC are
constrained at the entrance by a circular network of
intra- and inter-monomer hydrogen bonds and salt
bridges. Aromatic residues, particularly tyrosine and
phenylalanine, cluster in a ring around the base of the
TolC ꢀ-barrel, playing an important role in the closed
states of TolC due to hydrogen bonds.6) If the aperture
dilatation of TolC is inhibited by strengthening the
hydrogen bonding between the essential sites, the TolC
channel is forced to close, and antibiotic susceptibility
is increased.
Bacterial strains and antibiotics. E. coli strain Fj307 (chosen from
180 clinical multi-drug resistant isolates and found to be a highly drug
efflux strain) served as positive control. E. coli strain TW1b with the
tolC gene partly knocked out (donated by the Coli Genetic Stock
Center at Yale University) was used as negative control. ATCC25922
(conserved by our laboratory), was used as quality control. The
antibiotics chloramphenicol (CHL), erythromycin (ERY), ciproflox-
acin (CIP), gentamicin (GM), ampicillin (AMP), and tetracycline
(TET) were purchased from Sigma-Aldrich (Shanghai, China).
Design and synthesis of indole derivatives. There are four major
links that establish a circular network constraining six helical pairs of
TolC monomers to form a closed state. Hydrogen bonds Asp153-Tyr362
and Gln136-Glu359 connect the inner coiled coils to the outer coils of
the same monomer, and a salt bridge between Asp153-Arg367 and
hydrogen bonds Arg367-Thr152 connects each inner coil to the outer coil
of the monomer. When the AcrAB-TolC efflux system is active, the
inner coils move outward to enlarge the entrance diameter of TolC, and
these links are broken.9) Indole derivatives were designed to contain
two chemical groups that can bind to TolC at the O@Asp153 and
H@Tyr362 sites with strong hydrogen bonds and are hard to break, so,
with its strong connection, TolC is hard to open naturally and the
function of the AcrAB-TolC system is inhibited.
The hydrogen bond between tyrosine and phenylalanine around the
base of the TolC ꢀ-barrel was screened using software the Swiss-Pdb
Viewer (v3.7). The indole derivatives were designed according to the
TolC structure with the GAUSSIAN98 package of programs.10) All
the derivatives were synthesized by the College of Chemistry and
Environmental Protection Engineering, Southwest University for
Nationalities.11)
y
To whom correspondence should be addressed. Fax: +86-28-8547-1599; E-mail: whongning@163.com
Abbreviations: CFUs, colony-forming units; EPIs, efflux pump inhibitors; MICs, minimal inhibitory concentrations; CLSI, clinical and laboratory
standards institute; CHL, chloramphenicol; ERY, erythromycin; CIP, ciprofloxacin; GM, gentamicin; AMP, ampicillin; TET, tetracycline; SEM,
standard error of mean

2238
B. ZENG et al.
Fig. 1. Chemical Structures and Synthetic Protocol of Indole Derivꢁatives.
ꢁ
.
Reagents and conditions: (a) HNO3, (CH3CO)2O, ꢀ10 C to 0 C; (b) 40% NaOH, heat reflex, pH 1.0–2.0; (c) SnCl2 2H2O, ethyl acetate, N2,
60 ꢁC reflex; (d) (Boc)2O, NaOH, 1,4-dioxane, 0 ꢁC; (e) AgNO3, benzoyl chloride, acetonitrile, ꢀ10 ꢁC to 25 ꢁC; (f) 68% HNO3, CH2Cl2, 0 ꢁC
to 25 ꢁC.
3-Amino-6-carboxyl-indole (1c) and 3-nitro-6-amino-indole (2c)
were synthesized from 6-methyl formate-indole (1m) and 6-amino-
indole (2m). The molecular formula and synthetic route are shown in
Fig. 1. The structures were confirmed by 1H NMR, MS, and IR spectra
and elemental analysis.
colony-forming units (CFU) bacteria was inoculated into 50 ml MH
broth containing CHL every time. The induced concentration of CHL
started from 1/4 fold to 1/2 fold, 1 fold, 2 fold . . . of the MIC. We
stabilized the inducing bacteria by inoculating it in the same drug
concentration repeatedly. The MIC of the resistant E. coli against
CHL, CIP, GM, TET, and AMP was tested after inducing. The final
induced strain was designated E. coli YD02, and it served as a positive
control.
(a): 6-methyl formate-indole (1m, 1.75 g, 10 mM) was dissolved
with 50 ml of (CH3CO)2O (acetic anhydride) in an erlenmeyer flask;
peroxyacetyl nitrate was dropped into the flask constantly with stirring
at ꢀ10 ꢁC, then a 0 ꢁC reaction for 6 h; product was extracted with
ethyl acetate, and washed with saturated NaCl solution, and 1a was
gained after chromatography (red solid, 90% yield). (b): 1a (2.19 g,
10 mM) and NaOH solution (40%) was added to the flask (containing
reflex condenser), heat reaction for 6 h; adjustment of solution into PH
1.0–2.0 by HCl, and subsidence 1b was collected (yellow solid, 92%
Evaluation of reversal activities by MIC test. Minimal inhibitory
concentrations (MICs) of CHL, ERY, CIP, and TET co-adminstrated
with indole derivatives were conducted by the microdilution broth
method following the CLSI (Clinical and Laboratory Standards
Institute). Four different E. coli strains (Fj307, YD02, TW1b, and
ATCC25922) were cultured in MH broth, and inoculated in tubes at a
concentration of 105 CFU/ml. The concentration of indole derivatives
was diluted to 0.5 mM in every tube. All tubes were cultured for 16 h
at 37 ꢁC.
.
yield). (c): Ethyl acetate (20 ml), 1b (2.05 g, 10 mM), SnCl2 2H2O
(1.128 g, 5 mM) was added to the flask (containing reflex condenser),
60 ꢁC reaction for 4 h in the presence of protective agent N2; then the
reaction solution was mixed with 100 ml of saturated NaHCO3 solution
with constant stirring, yellow substance was extracted with ethyl
acetate, and dried with Na2SO4, 1c was gained after purification by
column chromatography. (orange solid, 95% yield, m.p. 118–121). (d):
6-amino-indole (2m, 1.36 g, 0.01 M), 1,4-dioxane (65 ml) were mixed
in the flask at 0 ꢁC; then NaOH (1.0 ml, 2 mol/l) and (Boc)2O (2.4 g,
11 mM) was dropped into the flask separately, reaction for 2–3 h;
Product 2a was extracted from solution with ethyl acetate, and dried
with Na2SO4. (e): Solution I (acetonitrile, 5 ml; benzoyl chloride,
1.6 ml), and solution II (acetonitrile, 10 ml; AgNO3, 2.55 g, 15 mM)
were prepared; then solution I was poured into solution II very slowly
at 0 ꢁC to form solution III (milk-like solution). 2a (1.2 g, 4 mM) and
acetonitrile (20 ml) were added to the flask, reaction at ꢀ10 ꢁC for
15 min, and then solution III (6.18 ml) was added and this was
mechanically stirred for 1 h at ꢀ10 ꢁC, then continued stirred at 25 ꢁC
for another 30 min, and 50 ml 0 ꢁC H2O was added in at last. Product
2b was extracted from the solution with ethyl acetate, dried with
Na2SO4. (f): CH2Cl2 (25 ml) and 2b (3 mM, 0.8 g) were added to the
flask, and HNO3 (0.65 ml, 68%) was added slowly at 0 ꢁC, and the
solution was set at 0 ꢁC for 1 h and again at 25 ꢁC for 3 h to finish the
reaction. Sediment 2c was collected and purified by column chroma-
tography (red solid, 75% yield, m.p. 119–122).
Construction of mutant variant BL21(ED3)ꢀ2. Primers shown in
Table 1 were synthesized by Takara Biotechnology (Dalian, China). In
order to construct an E. coli strain that expresses mutant TolC, we need
to have its nature TolC expression to be blocked and set it be prepared
to recombine the new mutant tolC gene. So, the tolC gene in the E. coli
BL21(DE3) was partly knocked out by the red homogenous recombi-
nant technique.12) First, pKD46 (presented by the Coli Genetic Stock
Center at Yale University) was transferred into BL21(DE3), and
cultured in 50 ml of LB (Luria-Bertani) medium at 30 ꢁC for 12 h. Four
ml of bacteria culture was transferred into 200 ml fresh LB medium,
and cultured at 30 ꢁC until its ½OD600ꢂ ꢃ 0:25, and then 5 mmol/l
L-arabinose was added and this was incubated for another 1 h to induce
pKD46 expression (ensuring that ½OD600ꢂ < 0:6). The bacteria culture
was cooled on ice for 10 min, then washed 3 times with 10% glycerin
with centrifugation at 4 ꢁC, 4,000 rpm for 10 min, and the competent
BL21(DE3)/pKD46 was concentrated and prepared for recombination.
The tetr gene as a resistance selection gene was amplified from
pBR322 with primers FPQS and RPQS, which contained homologous
recombination arms. Then approximately 500 ng PCR product was
electro transformed into the competent BL21(DE3)/pKD46 with
Gene Pulser (Xcell, Bio-Rad, USA) at 200 ꢀ, 25 mF, 1.8 kV. The
recombined strain was then cultured on an LB plate with 34 mg/ml of
tetracycline at 37 ꢁC. BL21(ED3)/pKD46/tolC(M)tetr was selected,
and identified by PCR with primers FPJD and RPJD, and was
designated BL21(DE3)ꢀ2.
3-Amino-6-carboxyl-indolenine (1c), IR (KBr) ꢁ: 3445 (NH2),
3422 (NH), 2923 (Ar–H), 3651 (–COOH), 1710 (C=O), 1620 (C=C),
1468 (C–N). 1H NMR (DMSO) ꢂ: 12.98 (br, 1H, –COOH), 11.61
(s, 1H, NH), 8.85 (s, 1H), 8.16–8.18 (m, 2H, ArH), 7.93–7.96 (m, 1H,
ArH), 4.65 (br, 2H, NH2). ESI-MS (70 eV) m=z (%): 177.19 (Mþ þ 1,
65). 3-Nitro-6-amino-indole (2c), IR (KBr) ꢁ: 3403 (NH), 3132 (Ar–
H), 1620 (C=C), 1468 (C–N), 1383 (NO2). 1H NMR (DMSO) ꢂ: 12.82
(m, 1H, NH), 8.56 (s, 1H), 8.05–8.27 (m, 3H, ArH), 4.60 (s, 2H,
–NH2). ESI-MS (70 eV) m=z (%): 178.01 (Mþ þ 1, 100).
Preparation of the outer membrane fraction of E. coli and detection
of the TolC band on the SDS–PAGE gel by immunoblotting were
performed as previously described.13) Western blotting was conducted
following the operation manual of Bio-Rad.
In the analysis, the interaction energy between the indole
derivatives and TolC reduced ꢀ16.51 and ꢀ40.51 kcal/mol respec-
tively, as compared with the natural binding energy. In the binding
˚
state, the TolC diameter was approximately 17 A, smaller than the
˚
diameter of the opening state, 54 A.
Creation of TolC mutant variant BL21(DE3)ꢀ3. To create site
mutations in tolC gene, the wild tolC gene was amplified from
BL21(DE3) with primers FPEC and RPEC, and then inserted into
pUC18. Mutations were introduced into pUC18-tolC at Asp153 and
Tyr362 of tolC using the site-directed mutagenesis kit from SBS
Genetech (Beijing, China) with the mutant primers (Table 1), and
pUC18-tolCM was confirmed by nucleotide sequencing analysis and
blasted with that of wild E. coli at GenBank. tolCM was inserted into
Bacterial inducing. Construction of a highly efflux strain by
artificial induction: E. coli ATCC25922 was cultured in MH broth. The
bacterial concentration was manually adjusted by testing the optical
density (½OD600ꢂ ꢃ 0:5) and by centrifugation, then approximately 109

Indole Derivatives as AcrAB-TolC Efflux Pump Inhibitors
2239
Table 1. Primers Used in This Study
Primers
Function
Sequence
FPQS
RPQS
Amplify tet from
pBR322 for
recombination
5-ACAGTTTGATCGCGCTAAATACTGCTTCACCACAAGGAATGCAAGTTTGACAGCTTATCAT-
CGA-3
5-CATTCAACACGTTGAAATAAGCGGTCGCGGTGTTGAGGATCAAGGTTTTGGTGAATCCGTT-
AGCGAGGT-3
5-AACGGGCAGGTTGTCTGGCTTAA-3
FPJD
RPJD
FPEC
RPEC
Identifying mutant
isolate of tolC
Amplifying wild
tolC and inserting
into pET30b
5-CCGTTAAATCCAGAGTCGGTAAGTGA-3
5-TCTTATTCATATGAAGAAATTGCTCCCCATTCTTATCG-3 (Nde I)
5-TCAATAAGAATTCAGCCCCGTCGTCGTCATCAG-3 (EcoR I)
FPD153
A
Introducing site-
directed mutagenesis
in tolC
5-GCCTGGTAGCGATCACGGCCGTGCAGAACGCCCG-3
5-CGGGCGTTCTGCACGGCCGTGATCGCTACCAGGC-3
RPD153A
FPY362
F
5-GCGATGGAAGCGGGCTTCTCGGTCGGTACCCGTACCATTGTT-3
5-AACAATGGTACGGGTACCGACCGAGAAGCCCGCTTCCATCGC-3
RPY362
F
The sequences of the homologous recombination arms and the mutation site are underlined.
Table 2. MIC Test Results for ATCC25922 and E. coli YD02 before
and after Induction by CHL
Table 3. Evaluation of the Ability of Indole Derivatives to Increase
Antimicrobial Sensitivity
Drug
MIC
(mg/ml) E. coli YD02
CHL TET
CIP
0.5
128
GM AMP
MIC(mg/ml)
Compounds
Strains
0.5 mM
ATCC25922
4
128
2
64
2
128
4
128
CHL
TET
ERY
CIP
0.5
None
None
None
None
1m
1m
1m
1m
2m
2m
2m
2m
1c
ATCC25922
YD02
FJ307
4
128
256
—
2
2
64
64
—
2
0.25
64
64
128
128
—
pET30b with Nde I and EcoR I (Takara), and then transformed into
BL21(DE3)ꢀ2. BL21(DE3)ꢀ2/pET30b-tolCM was selected with Kanr
and designated BL21(DE3)ꢀ3. Expression of mutant TolC in
BL21(DE3)ꢀ3 was induced with IPTG and confirmed by electro-
phoresis and western blotting.
TW1ba
—
ATCC25922
YD02
FJ307
0.25
64
64
0.25
128
256
—
2
64
64
—
1
128
128
—
TW1ba
—
The tolC gene in BL21(DE3), BL21(DE3)ꢀ2, and BL21(DE3)ꢀ3
was examined by PCR with primers FPJD and RPJD.
ATCC25922
YD02
FJ307
0.25
64
64
0.5
128
256
—
1
64
64
—
1
128
128
—
Target validation. In order to determine whether these indole
moleculars are functional with the target site according to our design,
MICs of BL21(DE3), BL21(DE3)ꢀ2, and BL21(DE3)ꢀ3 against four
different antibiotics (CHL, ERY, CIP, and TET) were tested with and
without indole derivatives. The method was almost the same for CLSI,
but the drug concentration gradient was adjusted into 10 lower degrees
(0, 0.125, 0.25, 0.5, 1, 1.5, 2, 3, 4, 6 mg/ml). The statistical data were
analyzed by SPSS12.
TW1ba
—
ATCC25922
YD02
FJ307
0.25
32
32
0.25
1c
1c
4
32
16
—
1
4
16
—
8
1c
2c
TW1ba
—
1
—
0.25
16
ATCC25922
YD02
FJ307
0.25
2c
2c
4
16
16
—
8
16
—
4
—
4
—
2c
TW1ba
Results and Discussion
1m, 6-methyl formate-indole; 2m, 6-amino-indole; 1c, 3-amino-6-carboxyl-
indole; 2c, 3-nitro-6-amino-indole; YD02, Efflux E. coli by artificial
induction; FJ307, Efflux E. coli from clinical isolates; TW1ba, Mutant
E. coli, tolC gene was partly knocked out; —, Bacteria were not grown at
every concentration of antibiotics.
Bacterial inducing
Multi-drug resistant strain E. coli YD02 was pro-
duced after induction from ATCC25922. The results are
shown in Table 2.
1,049 bp amplicon was detected in normal BL21(DE3).
In BL21(DE3)ꢀ2, the plasmid pKD46 contained three
genes, ꢃ, ꢀ, and exo, whose products were Gam, Bet,
and Exo. These three proteins were recombinant protein
of the ꢄ bacteriophage and all of them are vital to the red
system. The recombination arms on tetr were designed
to be homologous to tolC. Hence, pKD46 can express
the red system under the control of its regulated
promoter, and can recombine tetr into the tolC. The tetr
gene (1,305 bp) was recombined into the tolC between
302 bp and 723 bp, so tolC gene was partly knocked out,
and a 1,981 bp amplicon was detected. All the amplicons
were verified by PCR from bacterial chromosomes,
which was consistent with the expected size (Fig. 2).
No expression of TolC in BL21(ED3)ꢀ2 was detected
by SDS–PAGE or western blot (Fig. 2). The results
showed that the function of tolC in BL21(DE3)ꢀ2
was inactive, but in BL21(DE3)ꢀ3, the site mutated
Evaluation of reversal activities
Indole derivatives restored the activity of CHL, TET,
ERY, and CIP to E. coli (Table 3).
The reverse ability of 1c, 2c as measured was at least
2-fold MIC decrease on TET and ERY, but at least
8-fold on CHL and CIP to resistant E. coli. 1m and 2m
also showed a little reverse ability on ATCC25922, but
had no effect on positive strains FJ307 and YD02. TW1b
was sensitive to all the antibiotics, since tolC was partly
knocked out.14) These results are well accordant to the
prediction by molecular design.
Construction of TolC mutant variants BL21(ED3)ꢀ2
and BL21(DE3)ꢀ3
In BL21(DE3), Natural tolC are 2,096 bp. Primers
FPJD and RPJD for identification were designed to
detect the tolC gene sequence between 72–1,121 bp. A

2240
B. ZENG et al.
A
B
C
Fig. 2. Identification of TolC in Various E. coli Strains by PCR, SDS–PAGE, and Western Blotting.
A, PCR test of the tolC gene. M, DNA markers III; 1, BL21(DE3)ꢀ2; 2, BL21(DE3). B, SDS–PAGE of membrane protein expression. M,
Low molecular weight protein marker; 1, BL21(DE3)ꢀ3; 2, BL21(DE3)ꢀ2; 3, BL21(DE3). Bacteria were grown to exponential phase in LB
medium. Twenty ml bacteria were washed by centrifugation, and treated with lysozyme (0.2 mg/ml) in sucrose solution (12%) for 30 min. Then
spheroplasts were broken down gently by sonication on ice and centrifugation at 10;000 g for 20 min at 4 ꢁC to remove cell debris. The
supernatants obtained were subjected to ultracentrifugation at 100;000 g for 1 h at 4 ꢁC, and then the membrane fraction was suspended in Tris-
Mg buffer (10 mM, pH 7.8) containing sarcosyl (2%), and incubated at 20 ꢁC for 30 min. Then the tubes were centrifuged at 70;000 g for 1 h at
4 ꢁC. Fraction sediment was collected with 200 ml ddH2O and used for SDS–PAGE. C, Western blotting to check TolC expression. 1,
BL21(DE3)ꢀ3; 2, BL21(DE3)ꢀ2; 3, BL21(DE3). Dilution of anti-TolC antiserum (presented by P. C. Tai, Georgia State University) was used to
detect TolC. Bound antibody alkaline phosphatase-conjugated goat anti-rabbit secondary antibody (Boster Biological Technology, WuHan,
China) was detected using nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate.
TolCM can be expressed. The sequencing result dem-
onstrated that Asp153 and Tyr362 were mutated into Ala
and Phe. BL21(DE3)ꢀ2 and BL(DE3)-3 were con-
structed successfully.
Target validation
Statistical results: Means with SEM (standard error
of mean) of MICs in the testing groups are shown in
Fig. 3. In BL21(DE3) control groups, indole deriva-
tives 1c and 2c could significantly potentiate the
antibiotic activity of CHL, CIP, TET, and ERY,
especially in CHL: BL21(DE3) 4 mg/ml, BL21(DE3)þ1c
1 mg/ml, BL21(DE3)þ2c 0:83 ꢄ 0:17 mg/ml, and CIP:
BL21(DE3) 0:83 ꢄ 0:17 mg/ml, BL21(DE3)þ1c 0.125
Fig. 3. MIC Test of Mutant E. coli against Different Antibiotics with
and without Indole Derivatives.
mg/ml, BL21(DE3)þ2c 0.125 mg/ml (p < 0:01) but in
TolC mutant BL21(DE3)ꢀ3 groups, 1c and 2c could not
significantly restore the activity of all the antibiotics,
CHL: BL21(DE3) 4 mg/ml, BL21(DE3)þ1c 3:83 ꢄ
0:17 mg/ml, BL21(DE3)þ2c 3:33 ꢄ 0:17 mg/ml, CIP:
BL21(DE3) 0:83 ꢄ 0:17 mg/ml, BL21(DE3)þ1c 0:83 ꢄ
0:17 mg/ml, BL21(DE3)þ2c 0:67 ꢄ 0:17 mg/ml (p >
0:05). Similarly to TW1b, BL21(DE3)ꢀ2 could not be
cultured in MH broth containing any antibiotics, while
tolC was not expressed.9)
In BL21(DE3)ꢀ3, the site-mutated tolCM gene was
inserted into pET30b with the natural signal peptide of
TolC, the signal peptide was used as a conductor of
TolC located at the natural membrane position. Com-
pared to the MIC results for BL21(DE3)ꢀ3 to
BL21(DE3) without indole derivatives (Fig. 3), the data
were almost same, indicating that the natural function of
mutant strain BL21(DE3)ꢀ3 was not affected by
experimental manipulation and that the signal peptide
itself can lead TolC to the right location in
BL21(DE3)ꢀ3. It also reflects that the site mutation
on TolC itself can not reverse the antibiotic sensitivity of
BL21(DE3)ꢀ3. This conforms to a study by Koronakis
(2004), in which the efflux function of TolC was not
C, Controls of each antibiotic group. þ, MICs showed significant
difference from control group statistically. ꢀ, MICs showed no
difference from control group statistically. Error bars, SEM (stand-
ard error of mean). BL21(DE3), BL21(DE3)ꢀ2, and BL21(DE3)ꢀ3
were cultured in MH broth and inoculated in tubes at a concentration
of 105 CFU/ml. Concentrations of 1c and 2c were diluted to 0.5 mM
in every tube. Antibiotics and indole derivatives were added to tubes
separately, and three repeating groups were tested at every
concentration degree. All the tubes were cultured for 16 h at 37 ꢁC.
been affected when Asp153 and Tyr362 were mutated
into Ala and Phe.6) Compared to the MIC results for
BL21(DE3)ꢀ3 and BL21(DE3) plus 1c and 2c, the
major difference was the change in the target site in
TolC according to our design. Therefore, due to that
indole derivatives interact with the Asp153 and Tyr362 on
the TolC and inhibit its aperture expanding to restore the
antimicrobial sensitivity of antibiotics.
Conclusions
3-Amino-6-carboxyl-indolenin and 3-nitro-6-amino-
indole can be used as efflux pump inhibitors of TolC and
can be co-administrated with antibiotics to potentiate

Indole Derivatives as AcrAB-TolC Efflux Pump Inhibitors
2241
4) Weng JR, Tsai CH, Kulp SK, and Chen CS, Cancer Lett., 262,
153–163 (2008).
antibiotic susceptibility. TolC can be developed as a
suitable target of efflux pump inhibitors. The conception
of designing EPIs targeting TolC is worth further
research.
5) Samosorn S, Bremner JB, Ball A, and Lewis K, Bioorg. Med.
Chem., 14, 857–865 (2006).
6) Koronakis V, Eswaran J, and Hughes C, Annu. Rev. Biochem.,
73, 467–489 (2004).
Acknowledgment
7) Freiberg C and Brotz-Oesterhelt H, Drug Discov. Today, 10,
927–935 (2005).
¨
8) Kana BD and Mizrahi V, Tuberculosis, 84, 63–75 (2004).
9) Andersen C, Koronakis E, Bokma E, Eswaran J, Humphreys D,
and Koronakis V, Proc. Natl. Acad. Sci., 99, 11103–11108
(2002).
This work was supported by the National Eleventh
Five-Year Plan, Science and Technology Supporting
Programs of China (2006BAK02A19). We also thank
Dr. Phang C. Tai (Georgia State University) for
providing anti-TolC antiserum.
10) Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
and Cheeseman JR, ‘‘V. G. Gaussian 98’’ Revision A.11,
Gaussian Inc., Pittsburgh (2001).
11) Li QH and Zhao ZG, Chem. Res. Appl., 20, 348–351 (2008).
12) Datsenko KA and Wanner BL, Proc. Natl. Acad. Sci., 97, 6640–
6645 (2000).
References
`
1) Pages JM, Masi M, and Barbe J, Trends Mol. Med., 11, 382–388
(2005).
13) Yamanaka H, Nomura T, Morisada N, Shinoda S, and Okamoto
K, Microb. Pathog., 33, 81–89 (2002).
2) Forunier dit Chabert J, Marquez B, Neville L, Jouchla L,
Broussous S, and Bouhours P, Bioorg. Med. Chem., 15, 4482–
4497 (2007).
14) Yamanaka H, Tadokoro S, Miyana M, Takahashi E,
Kobayashi H, and Okamoto K, Microb. Pathog., 42, 184–192
(2007).
3) Zhao ZJ, Wolkenberg SE, Lu MQ, Munshi V, Moyer G, and
Feng MZ, Bioorg. Med. Chem. Lett., 18, 554–559 (2008).