JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

Available online at www.sciencedirect.com

www.elsevier.com/locate/jprot

Low abundance of respiratory nitrate reductase is

essential for Escherichia coli in resistance to
aminoglycoside and cephalosporin
Yan Ma1 , Chang Guo1 , Hui Li, Xuan-xian Peng
Center for Proteomics, State Key Laboratory of Bio-Control, MOE Key Lab Aquat Food Safety, School of Life Sciences, Sun Yat-sen University,
University City, Guangzhou 510006, People's Republic of China

AR TIC LE I N FO

ABS TR ACT

Article history:

In the present study, we have characterized low abundance of NarG and NarH, two

Received 27 September 2012

components of respiratory nitrate reductase (Nar), in streptomycin (SM)-, gentamicine

Accepted 16 May 2013

(GEN)-, ceftazidime (CAZ)-, tetracycline (TET)- and nalidixic acid (NA)-resistant Escherichia coli

Available online 24 May 2013

strains using native/SDS-PAGE based proteomics. We validate the finding using Western

Keywords:

functional evidence indicates that loss of narG and narH results in two types of growth

Respiratory nitrate reductase

behaviors, higher and lower than control, in these antibiotic-resistant E. coli strains.

E. coli

Specifically, SM-, GEN- and CAZ-resistant bacteria grow faster, whereas NA- and TET-

Proteomics

resistant E. coli strains grow slower. Our data indicate that low abundance of respiratory Nar is

Antibiotic resistance

essential for E. coli in resistance to aminoglycoside and cephalosporin antibiotics. Meanwhile,

Native-SDS/PAGE

the results show that differential mechanisms exist in different antibiotic-resistant bacteria.

Proteinprotein interactions

The reason why the reversal growths are detected in NA- and TET-resistant E. coli strains waits

blotting and native/SDS-PAGE upon narG and narH deletion mutants. However, further

investigation. Our findings serve to propose novel strategies for controlling of aminoglycosideand cephalosporin-resistant E. coli strains through elevation of respiratory Nar activity.
Biological significance
Our data indicate that low abundance of respiratory Nar is essential for E. coli in resistance to
aminoglycoside and cephalosporin antibiotics. Meanwhile, the results show that differential
mechanisms exist in different antibiotic-resistant bacteria. Our findings serve to propose
novel strategies for controlling of aminoglycoside- and cephalosporin-resistant E. coli strains
through elevation of respiratory Nar activity.
2013 Elsevier B.V. All rights reserved.

1.

Introduction

The increasing incidence of antibiotic-resistant bacteria has

posed a serious threat to human health and aquaculture today
and thus has become a scientific issue worldwide [1,2]. A line of
evidence has indicated that four mechanisms contribute to the

bacterial resistance, including modification or hydrolysis of

enzymes, modification of drug targets, activation of efflux
pump systems, and reduce of outer membrane permeability
[3,4]. Despite these progresses over past sixty years, our
understanding of the resistant mechanisms is still incomplete,
which should be one of the major causes resulting in

Corresponding author at: School of Life Sciences, Sun Yat-sen University, University City, Guangzhou 510006, People's Republic of China.
Tel.: + 86 13580548832.
E-mail addresses: pxuanx@sysu.edu.cn, wangpeng@xmu.edu.cn (X.-X. Peng).
1
The first two authors have equally contributed.
1874-3919/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jprot.2013.05.019

JO U R N A L OF P ROTE O MI CS 8 7 ( 20 1 3 ) 7 88 8

antibiotic-resistant bacteria out of control. Thus, further investigation of bacterial resistant strategies to antibiotics may be
necessary for combating antibiotic-resistant bacteria.
The recently developed technologies are expected to facilitate the study of bacterial antibiotic-resistant mechanisms. For
example, proteinprotein interactions, being important for the
majority of biological functions including antibiotic resistance,
may be potentiated using highly throughput screening methods
of 2-D native PAGE (N-PAGE) and 2-D blue native PAGE
(BN-PAGE) based proteomics [57]. The two methods have
been proven especially useful for investigation of native protein
complexes in the context of proteomics. Comparatively,
N-PAGE based on proteomics may offer more information on
labile protein complexes containing metabolic enzymes than
BN-PAGE based proteomics [6,8]. Using the two approaches,
researchers have investigated membrane and outer membrane
protein complexes in Escherichia coli and identified a series of
homogeneous and heterogeneous complexes. The identified
membrane-related protein complexes may contain cytoplasmic
and periplasmic proteins since approximately 3040% E. coli
proteins may function in the membrane of the cell envelope, in
which there are soluble proteins tethered to the inner and outer
membranes through hydrophobic patches, lipid moieties, or
charge interactions or in membrane protein complexes [68].
On the other hand, recent reports have indicated that metabolic
enzymes are related to antibiotic resistance [9,10], and the
regulation of metabolism elevates susceptibility of bacterial
persisters to antibiotics [11]. We have also shown that Na(+)
NQR complex is essential for Vibrio alginolyticus in resistance to
balofloxacin using N-PAGE based proteomics coupled with
functional validation [12]. We hypothesized that the comparative N-PAGE approach might provide metabolic-related protein
complexes that contributed to antibiotic resistance for further
functional investigation since many of metabolic enzymes are
detected in membrane protein fraction [13]. In the present
study, thus, N-PAGE based proteomics is used to identify a
metabolic protein complex in membrane fraction of E. coli K12
BW25113 and then the role of the complex in bacterial antibiotic
resistance is further investigated.
Here we present a comparative analysis of membrane protein
complexes in E. coli responsible to five antibiotics, streptomycin
(SM), gentamicine (GEN), ceftazidime (CAZ), tetracycline (TET)
and nalidixic acid (NA). The five represent four classes of
antibiotics containing aminoglycoside (SM, GEN), -lactam
(CAZ), tetracyclines (TET) and fluoroquinolone (NA) that are
used in clinic. We are interested in identifying NarG and NarH,
two components of respiratory nitrate reductase (Nar) in line.
Nar is membrane-bound heterotrimeric enzymes that catalyze
the reduction of nitrate, coupled to the generation of a
proton-motive force across the cytoplasmic membrane during
anaerobic respiration. The protein complex is validated using
Far-Western blotting and the N-PAGE analysis of narG and narH
mutants. We further reveal low abundance of NarG and NarH in
SM-, GEN-, CAZ-, TET- and NA-resistant E. coli strains (SM-R,
GEN-R, CAZ-R, TET-R and NA-R, respectively). At last, we report
the functional investigation using narG- and narH-deleted
mutants. Survival capabilities are higher in SM-R, GEN-R and
CAZ-R, but lower in TET-R and NA-R compared with control.
These findings indicate that decrease of Nar level elevates
bacterial resistance to SM, GEN, CAZ and susceptibility to NA,

79

TET. These results may have significant implications in understanding bacterial antibiotic-resistant mechanisms based on
metabolic regulation due to the importance of the protein
complex in bacterial energy metabolism.

2.

Materials and methods

2.1.

Bacterial strains and culture conditions

The bacterial strain, E. coli K12 BW25113, and its mutants, narG
and narH, used in the present study were kindly provided by
NBRP (NIG, Japan): E. coli. The antimicrobial agents SM, GEN,
CAZ, TET and NA were purchased from a commercial source
(Shanghai Sangon Biological Engineering Technology & Services
Co. Ltd. China). SM-, GEN-, CAZ-, TET- and NA-resistant strains
(SM-R, GEN-R, CAZ-R, TET-R and NA-R, respectively) were
selected by the use of ten sequential propagations in LB
medium with 1/2 MIC of these drugs of the original strain
(control) as described previously [14]. The resulting MICs were at
least four fold higher in SM-R, GEN-R, CAZ-R, TET-R and NA-R
than control, indicating that they were antibiotic-resistant
strains. These bacteria were cultured overnight in LB medium
at 37 C in a shaker bath as seed. Fresh overnight cultures were
inoculated into LB medium. The bacteria were cultured at 37 C
and grown to an OD600 nm of 0.6.

2.2.

Extraction of membrane proteins

Extraction of E. coli membrane proteins was performed as

described previously [8]. In brief, bacterial cells were harvested
by centrifugation at 4000g for 15 min at 4 C. Followed by
washing three times, the cells were resuspended in 5 mL of
50 mM, pH 7.4 TrisHCl buffer and then disrupted by intermittent ultrasonic treatment of 5.0 s once for a total of 40 min at
intervals of 9.9 s on ice. Supernatants were collected by
centrifugation at 5000g for 20 min. The supernatants were
further centrifuged at 100,000g for 1 h at 4 C in a Beckman
Coulter L-100XP centrifuge using a SW 41Ti Rotor. The collecting
pellets were resuspended in 50 mM, pH 7.4 TrisHCl buffer for
determination of concentrations of the proteins using the
Bradford method. The membrane fraction was validated to be
free of cytoplasmic proteins through measurement of superoxide dismutase (SOD) activity as described previously [13].

2.3.

Two-dimensional native/SDS-PAGE

Two-dimensional native/SDS-PAGE was performed according

to a procedure described previously [8]. In brief, the membrane
protein samples were diluted into 1:4 using buffer containing
100 mM TrisHCl, 50% glycerol, 0.2% bromophenol blue and
0.5% Triton X-100 and kept at 4 C overnight or 80 C for later
use. Triton X-100 was added since it is commonly used to
solubilize membrane proteins, especially inner membrane
proteins. The samples were directly electrophoresed in a
1 mm thick and 8% discontinuous native polyacrylamide slab
gel without SDS and beta-mercaptoethanol at 4 C with
constant voltage of 80 V for running gels until the tracking dye
reached the bottom of the gel. After the electrophoreses, the
gels were visualized by reverse staining with imidazolezinc

80

JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

sulfate and the visualized gel strips containing protein bands

were excised from the native gels. The gel strips were soaked in
equilibration buffer containing 0.06 M TrisHCl, pH 6.8, 5%
beta-mercaptoethanol (v/v), 10% glycerol and 2% SDS, on a
gentle shaker for 30 min and then switched to a seconddimensional SDS-PAGE. The second dimension SDS-PAGE was
carried out in 1.2 mm thick and 10% discontinued polyacrylamide slab gels at room temperature with constant voltage of
80 V 40 min and 120 V 2.5 h. Protein spots were visualized by
colloidal Coomassie brilliant blue (cCBB) staining method and
these gels were scanned in an AGFA white-light scanner at a
resolution of 600 by 200 m. The obtained raw images were
processed using the 2-D software Melanie 5.0. Following
background subtraction and spot detection, the gel patterns
were matched to each other by visual comparison.

2.4.
In-gel protein digestion and tandem mass
spectrometric analysis
In brief, protein spots separated by N-PAGE were finely
excised and washed three times with 100 L of 50 mM
ammonium bicarbonate/50% acetonitrile (ACN) for 30 min at
ambient temperature. After vacuum drying, gel pieces were
reduced with fresh solution of 10 mM DTT in 100 mM
ammonium bicarbonate at 57 C for 45 min, and subsequently
the gel pieces were alkylated with 55 mM iodoacetamide in
100 mM ammonium bicarbonate solution at room temperature
for 45 min in the dark. After washing with 100 mM ammonium
bicarbonate and dehydrated with acetonitrile, the gels
pieces were dried again. In-gel digestion was performed using
12.5 ng/L sequencing grade-iron-deprivation porcine trypsin
(Promega, Madison, WI) in 25 mM ammonium bicarbonate at
37 C for 1215 h. The peptide mixtures were extracted from the
gel pieces with 20 mM NH4HCO3/50% (v/v) acetonitrile
(containing 5% formic acid). Finally, the extracts were
vacuum-dried and redissolved in 0.1% trifluoroacetic acid in
water. Protein samples were sequenced by a fuzzy logic
feedback control system (Reflex III MALDI-TOF system, Bruker)
equipped with delayed ion extraction. The sample solution
(30100 ppm) with equivalent matrix solution (alpha-cyano4-hydroxycinnamic acid) was applied onto the MALDI
TOF-target using alpha cyano-4-hydroxycinnamic acid (HCCA)
as a MALDI matrix for peptide mapping and was prepared for
MALDI-TOF/MS analysis. MALDI-TOF spectra were calibrated
using trypsin autodigestion peptide signals and matrix ion
signals. FdoG and FdnG were further identified using MALDI
TOF/TOF. For MS/MS spectra, the 5 most abundant precursor
ions per sample were selected for subsequent fragmentation
and 10001200 Da laser shots were accumulated per precursor
ion. The criterion for precursor selection was a minimum S/N of
50. All MALDI analyses were performed by a fuzzy logic
feedback control system (Reflex III MALDI-TOF system, Bruker)
equipped with delayed ion extraction. Both the MS and MS/MS
data were interpreted and processed by using Flexanalysis
3.0 (Bruker Daltonics), and then the obtained MS and MS/MS
spectra per spot were combined and submitted to
MASCOT search engine (V2.3, Matrix Science, London, U.K.) by
Biotools 3.1 (Bruker Daltonics) and searched with the following
parameters: the NCBI in SwissProt (http://www.matrixscience.
com), one missed cleavage site, carbamidomethyl as fixed

modification of cysteine, oxidation of methionine as a variable

modification, and tolerance of 100 ppm for MS and 0.6 Da for
MS/MS. Known contaminant ions (keratin) were excluded.

2.5.

Bioinformatics analysis

The protein subcellular locations were determined by Program

PSORTb version 2.0 (http://www.psort.org/psortb/) [15]. DAVID
(http://david.abcc.ncifcrf.gov/list.jsp) was used to analyze protein list for pathway enrichment [16]. Names of 119 proteins
were uploaded into the server. The set of identifier was official
gene symbol and the E. coli genome was selected as functional
annotation background. A total of 76 proteins were enriched in
KEGG pathways of E. coli K-12 genome (Supplementary Table 1).
Processes were selected based on P values smaller than 0.05.
Seventeen enriched pathways with P value smaller than 0.05
and their corresponding proteins were used to further mapping
(Supplementary Table 2). We used Cytoscape [17] and its plugin
Network Analyzer [18] to visual the cross-linkage between
proteins and pathways. The layout was manually adjusted to
achieve better visualization. Protein physiological interpretation was obtained by search of molecular function in EcoCyc
(http://ecocyc.org/). Complex physiological interpretation was
obtained by data integration tools in DAVID (http://david.abcc.
ncifcrf.gov/list.jsp). In detail, proteins within a single complex
were uploaded into the server and used for molecular function
enrichment. The set of identifier was official gene symbol and
the E. coli genome was selected as functional annotation
background. Enriched complexes were selected. When correlation of altered abundance complexes to antibiotic resistance
was investigated, GO terms with P value smaller than 0.05 were
selected. P values of the selected and unselected enriched GO
molecular function terms were transformed with log(1/p) 1
and into 0, respectively, and then visualized [19].

2.6.
Gene cloning, protein purification and antiserum
preparing
Two pairs of primers for genes narG and narH were designed
according to E. coli K12 genomic sequences. For narG, the sense
primer was 5-ACGGGATCC ATGAGTAAATTCCTG-3 and the
antisense primer was 5-TGCCTCGAG TTTCACATCGTCATAG3. For narH, the sense primer was 5-GCGGAATTC AGCGTAA
AATGAAAA-3 and the antisense primer was 5-ATACTCGAG
CGATCATGGATGCGG-3. Standard PCR and molecular biology
protocols were utilized to amplify the two genes using E. coli K12
genomic DNA as a template. PCR products were obtained when
these primers were separately used. The two PCR fragments
were directionally cloned into plasmid pET-32a digested with
the same enzyme, and then expressed in E. coli BL21. Recombinant plasmids were detected by restriction enzyme analysis
and sequencing. Sequencing was carried out by BGI, Shenzhen,
Guangzhou. The overnight cultures of E. coli BL21 harboring
recombinant plasmid were diluted 1:100 (V/V) in fresh
Luria Broth with ampicillin (100 g/mL), then incubated at
37 C until the optical density at an OD 600 nm of 0.6. The
protein expressions were induced with 1 mM isopropyl--Dthiogalactopyranoside (IPTG) (BBI, Colorado, USA) for 3 h at
37 C after the optimization of expression conditions. Bacterial
cells were harvested by centrifugation, then washed and

JO U R N A L OF P ROTE O MI CS 8 7 ( 20 1 3 ) 7 88 8

resuspended in buffer D (8 M urea, 0.1 M NaH2PO4, 10 mM Tris

HCl, pH 8.0). The cell suspensions were disrupted by sonication
in an ice bath (300 W, 3 10 min) and centrifuged (7000g for
15 min at 4 C). The supernatants containing the recombinant
proteins were subsequently purified by affinity chromatography on Ni-NTA Super flow resin (Qiagen) according to the

81

manufacturer's instructions. Non-specific proteins were

washed by buffer E (8 M urea, 0.1 M NaH2PO4, 10 mM TrisHCl,
pH 6.3). Finally, the purified recombinant proteins were eluted
with buffer F (8 M urea, 0.1 M NaH2PO4, 10 mM TrisHCl, pH 5.9)
and buffer G (8 M urea, 0.1 M NaH2PO4, 10 mM TrisHCl, pH 4.5).
Concentration of proteins was determined by the Bradford

Fig. 1 Investigation of respiratory nitrate reductase from membrane protein fraction of E. coli K12 BW25113 cultured under LB
medium using N-PAGE. A, A representative N-PAGE map, showing approximately 100 cleared protein spots. B, NarG and NarH
of Nar complex were lined. C, Enlarged sixteen proteins out of the 100 protein spots were validated using 2-DE Western
blotting. D, Pathway enrichment analysis with DAVID (http://david.abcc.ncifcrf.gov/list.jsp). Box indicates enriched pathways,
and the circle represents the node and the color of the nodes corresponding to the number of connection. E and F, Heatmap of
transformed P value of significant enriched GO molecular function terms with the package gplots in the R platform.
P value < 0.05 was transformed with the formula of Log(1/P) 1. Unsupervised hierarchical clustering was applied to GO terms.
E, Complexes containing altered increased abundance of protein (s), F, Complexes containing altered decreased abundance of
protein(s).

82

JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

Fig. 1 (continued).

method. The purified recombinant proteins were used for

antiserum preparing (Guangzhou Chengxue Biotech. Corp.
China).

2.7.

Western blotting and far-Western blotting analyses

The two assays were performed as described previously [20].

Rabbit antisera to NarG and NarH were used as the primary
antibodies and horseradish peroxidase (HRP)-conjugated goat
anti-rabbit antibody was used as the secondary antibody
(Guangzhou Chengxue Biotech. Corp. China). For 1-DE and
2-DE Western blotting, the separated proteins from the 1-DE
and 2-DE gels were transferred to nitrocellulose (NC) membranes using constant voltage of 70 V for 1 h in transfer buffer
(48 mM Tris, 39 mM glycine, and 20% methanol) at 4 C and
were stained with Ponceau S to evaluate the transfer
efficiency. The membranes were blocked overnight with 5%
non-fat milk in Tris-buffered saline buffer containing 0.05%
Tween-20 (TTBS) at 4 C. After rinsing three times for 10 min
with TTBS buffer, the membranes were separately incubated
with rabbit antibodies to NarG and NarH for 2 h on a gentle

shaker at room temperature. The membranes were rinsed

again, and then incubated with the secondary antibody for 2 h
at the same conditions. The membranes were washed and
developed with dimethylaminoazobenzene (DAB) substrate
system until maximum color appearance. For Far-Western
blotting, the membranes used for the Western blotting were
washed with TTBS, and then incubated in blocking buffer
containing 2 mL of E. coli cell lysates overnight at 4 C. The
membranes were incubated with the same primary and
secondary antibodies in the Western blotting assay and were
developed with DAB system as described above.

2.8.
Investigation of antimicrobial susceptibility by
minimal inhibitory concentration (MIC) and survival
capability assays
MIC was performed according to National Committee for
Clinical Laboratory Standards. Survival capability assay was
performed as described previously [21]. In brief, inoculums of
narG-, narH-deleted mutants and control were separately
cultured overnight and diluted 1:1000 into 5 mL fresh LB

JO U R N A L OF P ROTE O MI CS 8 7 ( 20 1 3 ) 7 88 8

3.

83

Results

3.1.
Separation of membrane proteins and detection of Nar
complex in E. coli K12 BW25113 membrane protein fraction
using N-PAGE based proteomics

Fig. 2 Validation of Nar complex in E. coli K12 BW25113.

A, NarG and NarH interaction was identified by far-Western
blotting. 1, NarG and NarH location on the gel; 2, NarG was
detected using anti-NarG; 3, the NC membrane used in 2 was
incubated with E. coli K12 cell extraction and incubated with
anti-NarG again. B, N-PAGE maps were obtained using narG
and narH mutants, showing disappearance of NarG and
NarH.

medium with and without an antibiotic as tested group and

control, respectively, two tubes for each. Bacteria were cultured
at 37 C on a gentle shaker (150 rpm) and OD (600 nm) of
bacterial cultures was detected at 10 h. Survival capability was
characterized by division of the OD of the cultures with the
antibiotic over that of the cultures without the drug, and was
termed survival rates of bacterial strains cultured in medium
with the drug. The experiment was repeated at least three
times. Difference in survival rates was investigated with SPSS
program.

N-PAGE was used to separate the membrane proteins of E. coli

K12 BW25113. Reproducible N-PAGE gel maps were obtained
and good spot resolution was achieved when triplicate
experiments were carried out under identical conditions. A
representative map was shown in Fig. 1A. Approximately a
hundred of clear protein spots were observed on each of the
2-DE gels stained with cCBB R-250, most of which were located
in the range of 25120 kDa with minimal streaks. The streaks
could be resulted from the diffusion of the same proteins
during the first dimensional native-PAGE as reported previously [6,8,22,23]. These well visualized protein spots were
excised from 2-DE gels and analyzed by MALDI/TOFTOF
analysis. A hundred of proteins were successfully identified
(Supplementary Table 3). They represented ninety-four
unique proteins since five proteins had more than one spot.
The five proteins included TnaA (spots 14, 37, 38), BamA
(formerly YaeT, spots 11, 119), GlpD (spots 31, 32), TsF (spots
61, 62), NuoG (spots 77, 91). In addition, two proteins, FdoG
and FdnG, were identified in both spots 1 and 7. The two
proteins belonged to formate dehydrogenase system with
high sequence similarity so that they were identified in each
of the two spots, but they could be identified by their
molecular masses. The molecular masses of FdoG and FdnG
are 113 kDa and 90 kDa, and thus spots 1 and 7 were identified
as FdoG and FdnG, respectively, which were validated using
MALDI/TOFMS/MS analysis (Supplementary Fig. 1). We interestingly found that the two components, NarG and NarH, of
Nar complex were lined (Fig. 1B), which did not show in our
previous N-PAGE to detection of E. coli DH5 membrane
protein fraction [8]. Nar complex includes four polypeptides
NarG, NarH, NarJ and NarI of molecular weight 138.7, 57.7, 26.5
and 25.5 kDa, respectively [24]. NarJ and NarI did not appear in
the gel probably due to the low molecular weight. NarG and
NarH were further validated in 2-DE Western blotting (Fig. 1C).
The complex is capable of reducing nitrate using normal
physiological substrates and is clearly the major respiratory
Nar in E. coli since it accounts for 98% of the Nar activity when
fully induced [25]. The result provides a chance to investigate
a role of the complex contributed into bacterial energy
metabolism in bacterial antibiotic resistance. In addition,
sixteen out of the 94 proteins were validated using Western
blotting (Fig. 1C) and these possible protein complexes and
their identification criteria were grouped in Supplementary
Tables 4, 5 and Supplementary Fig. 2. We identified the
subunits A and B of the BAM complex in two distinct protein
complexes. The BAM complex consists of one outer membrane protein, BamA (formerly YaeT), and four lipoproteins,
BamB, BamC, BamD and BamE (formerly YfgL, NlpB, YfiO and
SmpA, respectively). The other three were not seen, which
might be related to protein abundance, molecular weight and
binding force. Further pathway analysis was carried out and
obtained 90 annotation out of the 119 proteins identified.
Among the 90 annotated proteins, 76 were enriched in 16

84

JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

Fig. 3 Low abundance of Nar complex was detected in antibiotic-resistant bacteria. SM-R: streptomycin-resistant strain,
GEN-R: gentamicine-resistant strain, CAZ-R: ceftazidime-resistant strain, TET-R: tetracycline-resistant strain, NA-R: nalidixic
acid-resistant strain.

pathways as shown in Fig. 1D. They are mainly involved in

carbon metabolism, nitrogen metabolism, oxidative phosphorylation. Furthermore, we investigated the relationships
between protein complexes containing altered abundance of
protein(s) and antibiotic resistance. Two and fourteen GO
molecular function terms were enriched in the complexes
containing higher and lower abundance of protein(s) than
control, respectively (Fig. 1E, F; Supplementary Table 6). The
two, 3-oxoacyl-[acyl-carrier-protein] synthase activity and
fatty-acid synthase activity, were enriched only in CAZ-R,
GEN-R and SM-R (Fig. 1E), whereas 4 iron, 4 sulfur cluster
binding, the most significant enriched molecular function
term out of the fourteen, was determined in the five
antibiotic-resistant bacteria (Fig. 1F). Nar complex belongs to
the molecular function term of 4 iron, 4 sulfur cluster binding.

3.2.
Validation of Nar complex in E. coli K12 BW25113
membrane proteins
Then, the Nar complex was validated using two assays. First,
far-Western blotting was used to investigate the interaction of
NarH with NarG. Our results indicated that the bait protein
was NarH when NarG was used as the prey protein (Fig. 2A).
Then, the complex was further validated by N-PAGE analysis
of narH and narG deletion mutants. This was based on the
hypothesis that when one protein within a protein complex is
absent, the others may disappear from N-PAGE gel. Our
results showed that loss of narG caused disappearance of
NarH, while deletion of narH resulted in loss of NarG in the
same line, meaning that NarGNarH complex existed in the
line (Fig. 2B). These results indicate that Nar complex detected
by the N-PAGE based proteomics is reliable, and thus may be
used for investigation of antibiotic resistance.

3.3.
Characterization of Nar complex in five antibioticresistant E. coli strains
In order to investigate whether abundance of Nar complex
components was altered to resist antibiotics, the N-PAGE
based proteomic approach was used to detect NarG and NarH
levels in SM-R, GEN-R, CAZ-R, TET-R and NA-R. Significant
changes were detected, showing that NarG and NarH
disappeared in the SM-R, CAZ-R, TET-R and decreased in
GEN-R and NA-R (Fig. 3). These results not only indicate that
Nar complex may play an important role in antibiotic
resistance, but also show that the role is related to the classes
of antibiotics. In addition, we summarized the altered
abundance of proteins in Supplementary Table 7. They were
mainly involved in oxidative phosphorylation and protein
synthesis.

3.4.
Functional characterization of narG and narH deletion
mutants in response to five antibiotics
We next performed functional characterization of the Nar
complex proteins using narG or narH deletion mutants. The
performance was carried out using two antimicrobial susceptibility assays: MIC assay and survival capability assay. Fig. 4AE
was a summary of MICs of narG and narH. Deletion of narG or
narH resulted in distinct elevation of MICs in medium with SM
or GEN, and two-fold decrease of MICs in medium with TET, and
no significant changes in medium with CAZ or NA (Fig. 4AE).
The survival capability assay was also carried out to further
investigate functional characterization of the mutants due that
the assay is more sensitive than MIC assay [20,26]. These
bacteria including narI and narJ were cultured in tubes with
series of two-fold dilution of SM, GEN, CAZ, TET and NA

JO U R N A L OF P ROTE O MI CS 8 7 ( 20 1 3 ) 7 88 8

concentrations from 0.625 to 5 g/mL, 0.19531.1719 g/mL,

0.07810.3125 g/mL, 0.781253.25 g/mL, and 1.257.5 g/mL,
respectively. The resulting cultures were assayed by measurement of optical density at 600 nm and the survival capability
was determined by division of OD values of the cultures with an
antibiotic over those without the drug. The survival rates of the
mutants cultured in medium with SM, GEN or CAZ were
significantly increased (P < 0.01) in a dose-dependent manner.
However, lower survival capabilities were detected in the
mutants cultured in medium with TET and NA (Fig. 4FJ).
Generally, the results obtained from survival capability assay
were in consistence with those obtained from MIC assay. In
summary, the results obtained from functional characterization of narG and narH were consistent with the changes of

85

NarG and NarH at the protein abundance which was detected by

2-DE analysis of these antibiotic-resistant bacterial strains.

4.

Discussion

N-PAGE based proteomics permits the study of membraneliable protein complexes in conditions formerly inaccessible
to previous techniques [6,8,12]. Different from our previous
report, which constructed a map of E. coli DH5 membrane
protein fraction [8], the present study, which described E. coli
K12 BW25113 membrane protein complexes, displays NarG and
NarH, two components of Nar complex that were not detected
in the previous report, on the gels. We were interested in the

Fig. 4 Histogram of MIC assay by broth microdilution method and testing for survival capability. AE, Histogram of MIC assay
using narG and narH deletion mutants and control E. coli K12 BW25113 for SM (A), GEN (B), CAZ(C), TET (D), NA (E). FJ, Testing
for survival capabilities using narG, narH deletion mutants and control E. coli K12 BW25113 for SM (F), GEN (G), CAZ(H), TET
(I), NA (J). * P < 0.05, ** P < 0.01.

86

JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

Fig. 4 (continued).

complex due to its role in bacterial energy metabolism [24].

Furthermore, we revealed that the low abundance of the two
proteins became a characteristic feature in five antibioticresistant E. coli strains. Thus, Nar complex may play a crucial
role in antibiotic resistance and shows antibiotic-relateddependent characteristics.
Nar complex is a membrane-bound and multi-subunit
oxidoreductase comprising a large catalytic subunit (NarG), an
electron-transfer subunit (NarH) and a membrane-anchoring
subunit (NarI) as well as a nitrate reductase chaperone
(NarJ), which is not a part of the active nitrate reductase
[24]. It catalyzes the following reaction: ferrocytochrome +
nitrate Y ferricytochrome + nitrite in E. coli. An electron transport scheme from formate to nitrate and including possible
quinone participation has been reported. NarGHI is cytoplasmically oriented [27]. The function, activity, location and
biological characteristics of the enzyme complex have been
widely investigated in E. coli [24,28,29], but information regarding
to association with antibiotic resistance is not available. In the
present study, we have demonstrated that the deletion of narG
or narH results in higher survival capabilities in exposure to SM,
GEN and CAZ compared with the control. MIC assay further
validates the critical role of the two proteins in SM and GEN
exposure. We have no proof why the loss of narG or narH leads
to lower survival capabilities when the mutants exposed to NA
and TET, but the present study demonstrates that low

abundance of Nar complex may be a critical strategy for E. coli

to resist aminoglycoside antibiotics and CAZ. Allison et al. show
that specific metabolic stimuli enable the killing of E. coli
persisters with aminoglycosides by promoting generation of a
proton-motive force which facilitates aminoglycoside uptake
[11]. Nar complex contributes to the generation of protonmotive force. Our results show the low abundance of Nar
complex in SM-R and GEN-R. The low abundance of Nar
complex may decrease the generation of proton-motive force
and thus limits the uptake of aminoglycoside, which results in
the elevation of bacterial resistance to aminoglycoside antibiotics. Thus, the present study indicates that decrease of the
proton-motive force generation is not only a resistant mechanism in the persisters [11], but also an important strategy in
aminoglycosides- and CAZ-resistant E. coli. Our results further
highlight the importance of the metabolic environment to
antibiotic treatment.
Interestingly, the mutants grew faster in the medium with
aminoglycosides and CAZ and lower in the medium with NA
and CTC, which may suggest that the loss of the genes results in
the limitation of NA and CTC uptake. The results indicate that
differential regulation mechanisms to different classes of
antibiotics are related to Nar complex. Allison et al. also indicate
that the specific metabolic stimuli enable the killing of E. coli
persisters with aminoglycosides by the generation of a
proton-motive force that is determined only in aminoglycosides

JO U R N A L OF P ROTE O MI CS 8 7 ( 20 1 3 ) 7 88 8

but not other antibiotics [11]. Our results demonstrate that low
abundance of Nar complex promotes bacterial resistance to
both aminoglycosides and CAZ, but further investigation is
required to answer why the loss of the complex elevates the
bacterial susceptibility to NA and TET.

5.

Conclusion

N-PAGE based proteomic approach was used to investigate

E. coli K12 BW25113 membrane protein complexes and resulted
in the detection of NarG and NarH, two main components of
Nar complex. The decreased abundance of the two proteins
was characterized in SM-R, GEN-R, CAZ-R, TET-R and NA-R.
Functional tests using narG- and narH-deleted mutants demonstrated distinctly elevated resistance to aminoglycosides
and CAZ. It is documented that low abundance of Nar complex
is required for E. coli resistance to aminoglycosides and CAZ,
which is the first report on E. coli Nar metabolism which alters
bacterial resistance to antibiotics. This finding can serve to
propose novel strategies for treating infections caused by
aminoglycoside- and CAZ-resistant E. coli through activation
of Nar activity.
Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.jprot.2013.05.019.

[9]
[10]

[11]

[12]

[13]

[14]

[15]

[16]

Acknowledgments
This work was sponsored by grants from the NSFC projects
(30972279, 40976080), Guangdong Provincial Science and
Technology projects (2012A031100004) and Doctoral Fund of
Ministry of Education of China (100171110029).

[17]

[18]

[19]

REFERENCES
[20]
[1] Radhouani H, Pinto L, Poeta P, Igrejas G. After genomics, what
proteomics tools could help us understand the antimicrobial
resistance of Escherichia coli? J Proteomics 2012;75:277389.
[2] Pinto L, Poeta P, Vieira S, Caleja C, Radhouani H, Carvalho C,
Vieira-Pinto M, Themudo P, Torres C, Vitorino R, Domingues
P, Igrejas G. Genomic and proteomic evaluation of antibiotic
resistance in Salmonella strains. J Proteomics 2010;73:153541.
[3] Alekshun MN, Levy SB. Molecular mechanisms of
antibacterial multidrug resistance. Cell 2007;128:103750.
[4] Lin XM, Yang JN, Peng XX, Li H. A novel negative regulation
mechanism of bacterial outer membrane proteins in
response to antibiotic resistance. J Proteome Res 2010;9:
59529.
[5] van Gestel RA, van Solinge WW, van der Toorn HW, Rijksen G,
Heck AJ, van Wijk R, Slijper M. Quantitative erythrocyte
membrane proteome analysis with blue-native/SDS PAGE.
J Proteomics 2010;73:45665.
[6] Pan JY, Wu HK, Liu X, Li PP, Li H, Wang SY, Peng XX.
Complexome of Escherichia coli cytosolic proteins under
normal natural conditions. Mol Biosyst 2011;7:265163.
[7] Dresler J, Klimentova J, Stulik J. Francisella tularensis
membrane complexome by blue native/SDS-PAGE.
J Proteomics 2011;75:25769.
[8] Pan JY, Li H, Ma Y, Chen P, Zhao P, Wang SY, Peng XX.
Complexome of Escherichia coli envelope proteins under

[21]

[22]
[23]

[24]

[25]
[26]

[27]

87

normal physiological conditions. J Proteome Res 2010;9:

373040.
Martnez JL, Rojo F. Metabolic regulation of antibiotic
resistance. FEMS Microbiol Rev 2011;35(5):76889.
Yeung AT, Bains M, Hancock RE. The sensor kinase CbrA is a
global regulator that modulates metabolism, virulence, and
antibiotic resistance in Pseudomonas aeruginosa. J Bacteriol
2011;193:91831.
Allison KR, Brynildsen MP, Collins JJ. Metabolite-enabled
eradication of bacterial persisters by aminoglycosides.
Nature 2011;473:21620.
Li PP, Liu XJ, Li H, Peng XX. Downregulation of Na(+)-NQR
complex is essential for Vibrio alginolyticus in resistance to
balofloxacin. J Proteomics 2012;75:263848.
Huang CZ, Lin XM, Wu LN, Zhang DF, Liu D, Wang SY, Peng
XX. Systematic identification of the sub-proteome of
Escherichia coli cell envelope reveals the interaction network
of membrane proteins and membrane-associated peripheral
proteins. J Proteome Res 2006;5:326876.
Li H, Pan JY, Liu XJ, Gao JX, Wu HK, Wang C, Peng XX.
Alterations of protein complexes and pathways in genetic
information flow and response to stimulus contribute to
Escherichia coli resistance to balofloxacin. Mol Biosyst 2012;8:
230311.
Gardy JL, Laird MR, Chen F, Rey S, Walsh CJ, Ester M,
Brinkman FS. PSORTb v.2.0: expanded prediction of bacterial
protein subcellular localization and insights gained from
comparative proteome analysis. Bioinformatics 2005;21:
61723.
Huang da W, Sherman BT, Lempicki RA. Systematic and
integrative analysis of large gene lists using DAVID
bioinformatics resources. Nat Protoc 2009;4:4457.
Smoot ME, Ono K, Ruscheinski J, Wang PL, Ideker T.
Cytoscape 2.8: new features for data integration and network
visualization. Bioinformatics 2011;27:4312.
Assenov Y, Ramrez F, Schelhorn SE, Lengauer T, Albrecht M.
Computing topological parameters of biological networks.
Bioinformatics 2008;24:2824.
Rigbolt KT, Prokhorova TA, Akimov V, Henningsen J,
Johansen PT, Kratchmarova I, Kassem M, Mann M, Olsen JV,
Blagoev B. System-wide temporal characterization of the
proteome and phosphoproteome of human embryonic stem
cell differentiation. Sci Signal 2011;4:rs3.
Liu Y, Zhang H, Liu Y, Li H, Peng XX. Determination of the
heterogeneous interactome between Edwardsiella tarda and
fish gills. J Proteomics 2012;75:111928.
Li H, Wang BC, Xu WJ, Lin XM, Peng XX. Identification and
network of outer membrane proteins regulating
streptomysin-resistance in Escherichia coli. J Proteome Res
2008;7:40409.
Wittig L, Schagger H. Advantages and limitations of
clear-native PAGE. Proteomics 2005;5:433846.
Sun H, Pan YC. Using native gel in two-dimensional PAGE for
the detection of protein interactions in protein extract.
J Biochem Biophys Methods 1999;39:14351.
Blasco F, Dos Santos JP, Magalon A, Frixon C, Guigliarelli B,
Santini CL, Giordano G. NarJ is a specific chaperone required
for molybdenum cofactor assembly in nitrate reductase A of
Escherichia coli. Mol Microbiol 1998;28:43547.
Bonnefoy V, Demoss JA. Nitrate reductases in Escherichia coli.
Antonie Van Leeuwenhoek 1994;66:4756.
Li H, Lin XM, Wang SY, Peng XX. Identification and
antibody-therapeutic targeting of chloramphenicol-resistant
outer membrane proteins in Escherichia coli. J Proteome Res
2007;6:362836.
Arias-Cartin R, Grimaldi S, Pommier J, Lanciano P, Schaefer C,
Arnoux P, Giordano G, Guigliarelli B, Magalon A.
Cardiolipin-based respiratory complex activation in bacteria.
Proc Natl Acad Sci U S A 2011;108:77816.

88

JO U R N A L OF PR O TE O MI CS 87 ( 20 1 3 ) 7 8 88

[28] Stewart V, Bledsoe PJ, Chen LL, Cai A. Catabolite repression

control of napF (periplasmic nitrate reductase) operon
expression in Escherichia coli K-12. J Bacteriol 2009;191:
996-05.

[29] Smith DJ, Oliver CE, Shelver WL, Caesar T, Anderson RC.
Chlorate metabolism in pure cultures of Escherichia coli O157:
H7 pretreated with either nitrate or chlorate. J Agric Food
Chem 2009;57:1021624.