Báo cáo khoa học: Invasion of enteropathogenic Escherichia coli into host cells through epithelial tight junctions ppt

Invasion of enteropathogenic Escherichia coli into host
cells through epithelial tight junctions
Qiurong Li
1,2
, Qiang Zhang
1,2
, Chenyang Wang
1
, Ning Li
1
and Jieshou Li
1
1 Institute of General Surgery, Jinling Hospital, Nanjing, China
2 School of Medicine, Nanjing University, Nanjing, China
As pathogens invade the gastrointestinal tracts, they
must overcome epithelial barriers to initiate infection.
An intact epithelial barrier is essential for physiological
homeostasis and defense against extrinsic antigens. The
intestinal barrier comprises an intact layer of epithelial
cells, which are tightly connected in the apical region
of the lateral plasma membrane by specialized narrow
belt-like structures called tight junctions (TJs). TJs are
dynamic structures composed of various proteins,
including the transmembrane proteins occludin, a fam-
ily of claudins, and ZO-1, a peripheral membrane pro-
tein of TJs [1]. Disruption of tight-junction integrity is
a mechanism by which diarrhea is induced, and TJs
are a common target of various enteric pathogens [2].
The human intestinal pathogen enteropathogenic
Escherichia coli (EPEC) causes diarrheal disease by dis-
rupting the integrity of TJs. EPEC disrupts TJ archi-

tecture by loss of TJ protein–protein interactions,
redistribution of TJ proteins, and the appearance of
aberrant TJ strands in the lateral membrane [3].
Recently, TJs have been considered as specialized
plasma membrane microdomains, lipid raft-like mem-
brane compartments that are enriched in cholesterol
and sphingolipid [4]. It was suggested that these raft-
like compartments play an important role in the
spatial organization of TJs and probably in regulation
of paracellular permeability in epithelial cells [4]. The
loss of TJ barrier function has been correlated with
displacement of lipid raft-associated TJ proteins,
suggesting that this membrane microdomain is an
integral part of the TJ structure [4].
The critical role of the host cell plasma membrane
in response to pathogens has been demonstrated by
studies indicating that lipid rafts are the preferred
entry sites for several invasive pathogens including
Salmonella, Shigella, Listeria and Chlamydia [5–8]. It is
likely that Campylobacter jejuni also utilizes lipid rafts
as the entry point [9]. Coxsackievirus, adenoviruses
Keywords
enteropathogenic Escherichia coli; lipid raft;
TER; tight junction; tight junction protein
Correspondence
J. Li, Institute of General Surgery, Jinling
Hospital, 305 East Zhongshan Road, Nanjing
210002, China
Fax: +86 25 84803956
Tel: +86 25 80860064

E-mail:
(Received 22 July 2008, revised 28
September 2008, accepted 6 October
2008)
doi:10.1111/j.1742-4658.2008.06731.x
Enteropathogenic Escherichia coli (EPEC) has been shown to disrupt the
barrier function of host intestinal epithelial tissues through entering tight
junctions. However, the mechanism by which this occurs remains poorly
understood. In this study, we determined that EPEC invades host cells
through tight junctions as it initiates infection. Immunofluorescence micro-
scopy revealed redistribution of the tight-junction proteins occludin and
ZO-1 from an intercellular to a cytoplasmic location after EPEC invasion.
Flotillin-1 was recruited to sites of EPEC entry. EPEC entered host cells
through tight-junction membrane microdomains. Tight-junction ultrastruc-
ture was disrupted following EPEC infection, accompanied by loss of
barrier function. EPEC infection caused a time-dependent decrease in
trans-epithelial electrical resistance. Subcellular fractionation using discon-
tinuous sucrose density gradients demonstrated a decline in raft-associated
occludin following exposure to EPEC. These results indicate the important
role of host membrane tight-junction microdomains in EPEC invasion.
Abbreviations
EPEC, enteropathogenic Escherichia coli;M-b-CD, methyl-b-cyclodextrin; TER, transepithelial electrical resistance; TJ, tight junction.
6022 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS
and reoviruses are known to initiate infection via TJs
[10]. Microorganisms utilize lipid rafts in order to exert
their effects on host cells [11–13]. However, the role of
membrane rafts in the internalization of bacteria into
host cells during invasion has not yet been identified
clearly. The route followed by EPEC through the TJ
remains poorly understood. Membrane microdomains

of the TJ may be used by EPEC to penetrate the
epithelial barrier.
In this study, the specific involvement of TJ mem-
brane microdomains in EPEC invasion into host cells
was determined. The results demonstrate that TJ mem-
brane microdomains are required for bacterial inva-
sion, and that the distribution of TJ proteins in TJ
membrane microdomains is altered in response to
EPEC infection. EPEC invasion has a significant effect
on localization of TJ proteins in TJ membrane microd-
omains. Infection of intestinal epithelial cells with
EPEC disrupts TJ structure and barrier function by
altering the distribution of TJ proteins in TJ mem-
brane microdomains.
Results
EPEC co-localizes with flotillin-1 at the entry site
In order to investigate the role of lipid rafts in EPEC
invasion of epithelial cells, we used confocal fluores-
cence microscopy to observe the entry of the bacteria
into epithelial cells, and confirmed that penetration of
EPEC through TJ membrane microdomains was
related to infection. We examined the role of a known
lipid raft protein, flotillin-1, in EPEC invasion using a
specific probe for flotillin-1. As shown in Fig. 1, we
found that flotillin-1 is intimately associated with
invading bacteria at the point of attachment in EPEC-
infected cells. Accumulation of flotillin-1 was observed
at the entry foci and around intracellular EPEC. The
bacteria co-localized with flotillin-1 in the plasma
membrane and also in the cytoplasm (Fig. 1), where a

noticeable shift in cellular flotillin-1 to the location of
intracellular bacteria was observed. Flotillin-1 accumu-
lated at the sites of EPEC entry, and was transported
Flotillin-1 EPEC Merge
20 min
40 min
60 min
80 min
Fig. 1. EPEC co-localizes with flotillin-1. Cell
monolayers exposed to EPEC for 20-80 min
were stained with antibodies to flotillin-1
(red). EPEC (green) was visualized using a
CFDA SE cell tracer kit. Flotillin-1 accumu-
lated at the attachment site and was
recruited around the internalized bacteria.
Arrows indicate co-localization (yellow) of
EPEC (green) with flotillin-1 (red). These
experiments were repeated three times
with three replicates in each experiment.
Q. Li et al. EPEC enters host cells through tight junctions
FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS 6023
from the apical cell surface to the TJ, and then to the
cytoplasm. The lipid raft marker protein was recruited
to the sites of entry of the bacteria, indicating that
lipid rafts are involved in bacterial invasion. The asso-
ciation of lipid rafts with bacteria revealed that EPEC
entered host cells by a lipid raft-dependent route.
These results indicated that, once bacteria have
reached the TJ, the lipid raft is responsible for inter-
nalization and invasion of EPEC. Thus, lipid rafts are

required for EPEC penetration into the TJ and subse-
quent infection.
EPEC infection alters the distribution of TJ
proteins
We explored the impact of EPEC infection on the dis-
tribution of occludin and ZO-1. Caco-2 cells were
infected with EPEC for 1, 3 or 5 h, and then immuno-
stained for occludin and ZO-1. A chicken-wire pattern
of distribution for occludin and ZO-1 was seen in
uninfected control monolayers (Fig. 2A,B). In unin-
fected cells, occludin and ZO-1 were found to be
primarily localized to the cell membrane of the TJ.
Determination of the localization of these two proteins
showed a continuous and uniform distribution pattern
at the apical membrane of the cells and revealed
marked changes in the distribution of TJ proteins fol-
lowing EPEC infection that progressed in severity.
Breaks in occludin staining were seen (Fig. 2A) as pre-
viously reported [14]. EPEC infection of monolayers
resulted in redistribution of occludin from the lateral
membrane to the cytoplasm, with more pronounced
changes at 5 h post-infection. Consistent with these
results, the distribution of ZO-1 was also altered
following EPEC infection (Fig. 2B). These findings
suggest that occludin and ZO-1 dissociate from the
membrane and are redistributed to the cytoplasm in
response to EPEC infection. Immunofluorescence
microscopy revealed accumulation of ZO-1 around the
internalized EPEC (Fig. 3) during the entry process.
EPEC invasion into host cells was accompanied by

specific recruitment of TJ protein ZO-1. ZO-1 moved
from the apical membrane to the TJ, and then to the
cytoplasm concurrently with EPEC. These data indi-
XY
XZ
ZO-1
Uninfected
1
h
3 h 5 h
XY
XZ
Occludin
A
B
Fig. 2. EPEC infection alters the distribution of TJ proteins. Uninfected cell monolayers (left column) and monolayers infected with EPEC for
1, 3 and 5 h were immunostained with antibodies to occludin or ZO-1. In uninfected monolayers, occludin and ZO-1 are primarily limited to
the cell–cell interfaces, and the staining is continuous and uniform. After infection with EPEC, the distribution of occludin and ZO-1 was
significantly altered and breaks in occludin and ZO-1 staining were seen. These experiments were repeated three times with three replicates
in each experiment.
EPEC enters host cells through tight junctions Q. Li et al.
6024 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS
cate that EPEC infection has significant effects on the
distribution of TJ proteins in the cell monolayer, and
that EPEC invasion into host cells occurs through TJ
membrane microdomains.
Cholesterol depletion is responsible for
non-pathogenic E. coli entry into the cells
In our study, methyl-b-cyclodextrin (M-b-CD), a
cholesterol-depletion reagent that is commonly used to

disrupt lipid rafts, was used to deplete membrane
cholesterol. Caco-2 cells were pretreated with M-b-CD
for 30 or 45 min, and incubated with non-pathogenic
E. coli strain DH5a for 1 h. Bacteria were detected
using a CFDA SE cell tracer kit. After treatment with
M-b-CD, the typical chicken-wire pattern distribution
of ZO-1 was disrupted (Fig. 4A,B, left column). More
interestingly, the non-pathogenic E. coli was found to
enter M-b-CD-treated cells. ZO-1 was found to bind
to the internalized bacteria (Fig. 4A,B, right column).
When DH5a was applied to cells in the absence of
M-b-CD, the ultrastructures of TJs and the desmosome
were similar to those of controls (Fig. 4D). It was also
shown that ZO-1 is expressed along the lateral
membranes of the cells with a continuous and uniform
distribution (Fig. 4C). These results indicate that TJ
morphology is not affected by DH5a. Non-pathogenic
E. coli entered into the cells as a result of cholesterol
depletion. Thus, the TJ membrane microdomains are
responsible for bacterial entry into epithelial cells.
EPEC infection disrupts TJ ultrastructure
As the distribution pattern of TJ proteins was changed
in EPEC-infected cells, we examined whether EPEC
infection of epithelial cells disrupted TJ morphology.
Transmission electron microscopy was performed on
cell monolayers of uninfected and EPEC-infected cells
to investigate the morphological features of EPEC
invasion. Uninfected Caco-2 cells showed intact TJ
and desmosome structures (Fig. 5A). Transmission
electron micrographs of Caco-2 cells at various stages

ZO-1 EPEC Merge
20 min
40 min
60 min
80 min
Fig. 3. TJ proteins are responsible for EPEC
entry into host cells. Monolayers were
infected with EPEC for the indicated times
and stained for ZO-1 (red) and bacteria
(green) to assess bacterial entry. Yellow
areas (arrows) represent spatial overlap
(co-localization) of the red-stained ZO-1 with
the green bacterial signal. Nuclei (blue) were
visualized using 4¢,6-diamidino-2-pheny-
lindole. These experiments were repeated
three times with three replicates in each
experiment.
Q. Li et al. EPEC enters host cells through tight junctions
FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS 6025
MergeZO-1 DH5a
30 min
45 min
A
B
C
200n
D
Fig. 4. TJ membrane microdomains are
required for non-pathogenic E. coli delivery
into Caco-2 cells. Caco-2 monolayers pre-

treated with 10 m
M methyl-b-cyclodextrin
(M-b-CD) were exposed to the non-invasive
E. coli strain DH5a. Disruption of continuity
in ZO-1 staining was observed (left column).
Arrows indicate areas of co-localization (yel-
low) of bacteria (green) with ZO-1 (red).
DAPI was used to detect nuclear DNA
(blue). These experiments were repeated
three times with three replicates in each
experiment.
CBA
200 nm 200 nm 200 nm
F
E
200 nm
D
G
200 nm
200 nm
200 nm
Fig. 5. Transmission micrographs of the tight junction in cells infected with EPEC. (A) TJs and desmosomes are intact in uninfected cells.
(B) EPEC has not attached intimately and effaces microvilli. (C) The microvilli are lost at the site of attachment. (D,E) EPEC are attached inti-
mately to the host cell membrane and TJ membrane fusions (‘kisses’) are partly lost. (F) The bacterium enters the host cell through the tight
junction. Arrows indicate the location of tight junctions. Arrowheads show desmosomes. Scale bar = 200 nm.
EPEC enters host cells through tight junctions Q. Li et al.
6026 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS
of EPEC infection show alterations of TJ ultrastruc-
ture (Fig. 5B–G) characterized by the loss of microvilli
at sites of attachment, intimate attachment of bacteria

to the host cell membrane, and formation of surface
structures embracing the bacteria. In particular, TJ
membrane fusion ‘kisses’ were partly lost following
EPEC entry (Fig. 5D–F).
EPEC infection reduces transepithelial electrical
resistance
These morphological changes were correlated with dys-
function of the TJ barrier as determined by measuring
transepithelial electrical resistance (TER) (Fig. 6A).
Cell monolayers grown on Transwell filters were
exposed to EPEC, and TER was measured at the
indicated time points. EPEC infection significantly
decreased TER in a time-dependent manner, with
mean reductions in TER of 42.9 ± 12.2 and
61.1 ± 9.6% at 2 and 6 h post-infection, respectively.
The influence of cholesterol depletion on the TER of
cell monolayers was also examined (Fig. 6B). The low-
est concentration of M-b-CD (2 mm) caused only a
moderate decrease in TER of 21.1 ± 6% at 5 h. How-
ever, the TER was reduced by 54.6 ± 7.9
and 60.6 ± 3.1% after treatment with 5 and 10 mm
M-b-CD, respectively, for 5 h.
EPEC infection causes redistribution of TJ
proteins in TJ membrane microdomains
We next examined whether EPEC affected the molecu-
lar organization of the TJ proteins, and further investi-
gated possible molecular aspects of TJ membrane
microdomains with respect to the entry of EPEC into
host cells. TJ membrane microdomains in infected cell
monolayers were fractionated by discontinuous sucrose

density gradients. Following sucrose density gradient
centrifugation, the majority of the flotillin-1 (approxi-
mately 75% by densitometry) was found in
Triton X-100-insoluble fractions (fraction 4) of unin-
fected cells (Fig. 7A), which confirmed localization of
TJ membrane microdomains in the sucrose gradients.
Redistribution of flotillin-1 was detected in EPEC-
infected cells (Fig. 7A). Flotillin-1 shifted from
Triton X-100-insoluble fractions (fraction 4) to Tri-
ton X-100-soluble fractions (fractions 6–9) following
EPEC infection. Densitometric analysis of flotillin-1
bands indicated that fractions 6–9 contain 61.7% and
100% of total flotillin-1 after 1 and 5 h infection,
respectively (Fig. 7B).
In uninfected cells, occludin was detected both in
Triton X-100-insoluble fractions (fractions 3–5) and
Triton X-100-soluble fractions (fractions 6–9)
(Fig. 7C). There was a change in the distribution of
occludin in TJ membrane microdomains after infec-
tion with EPEC compared to uninfected cells. The
amount of occludin in fractions 3-5 from cells
infected with EPEC decreased by 22.1% after 1 h of
infection compared with uninfected cells (Fig. 7D).
After 5 h infection, all the occludin was found in
Triton X-100-soluble fractions (Fig. 7C). EPEC inva-
sion induced displacement of occludin from deter-
gent-insoluble fractions into detergent-soluble
fractions, and thus may be associated directly with
changes in the structure, function and morphology of
TJs induced by EPEC. However, claudin-1 and -4

were not detected in Triton X-100-insoluble fractions
but were present in Triton X-100-soluble fractions.
The distribution of claudin-1 and -4 was almost
0
20
40
60
80
100
120
140
01234567
Control
EPEC
Time (h)
A
***
***
***
***
0
20
40
60
80
100
120
140
0123456
Control

M-β-CD 10 m
M
M-β-CD 5 m
M
M-β-CD 2 m
M
TER (%) TER (%)
Time (h)
B
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
Fig. 6. EPEC invasion decreases TER in epithelial cell monolayers.
(A) Caco-2 monolayers were infected with EPEC, and TER was
measured at hourly time intervals. (B) Influence of increasing con-
centrations of M-b-CD on monolayer TER. Monolayers were treated
with increasing concentrations of M-b-CD (M-b-CD was added to
both the apical and basolateral chambers of the Transwells). Aster-

isks indicate that the TER value for monolayers infected with EPEC
(A) or treated with M-b-CD (B) was significantly less than that
for the control (***P < 0.001) at the same time point. Data are
means ± SEM from three experiments.
Q. Li et al. EPEC enters host cells through tight junctions
FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS 6027
unchanged after infection with EPEC for 5 h
(Fig. 7E,F). These results may indicate that claudin-1
and -4 are not involved in the process of EPEC
invasion, or that the presence of bacteria did not
influence the localization of these proteins in TJ
membrane microdomains. Thus, redistribution of TJ
Control
EPEC 30 min
EPEC 1 h
EPEC 5 h
Fraction 23456789

TX-100 insoluble TX-100 soluble
A
0
20
40
60
80
100
120
3–5 6–9
Control
EPEC 30 min

EPEC 1 h
EPEC 5 h
% of density in fraction 2–9
Fraction
B
Control
EPEC 30 min
EPEC 1 h
EPEC 5 h
Fraction 23456789
TX-100 insoluble TX-100 soluble
C
0
20
40
60
80
100
120
3–5 6–9
Control
EPEC 30 min
EPEC 1 h
EPEC 5 h

% of density in fraction 2–9
Fraction
D
Control
EPEC 30 min

EPEC 1 h
EPEC 5 h
Fraction 23456789
Fraction
23456789
E
TX-100 insoluble TX-100 soluble
F
Control
EPEC 1 h
EPEC 5 h
TX-100 insoluble TX-100 soluble
EPEC 30 min
Fig. 7. Redistribution of flotillin-1 and the TJ proteins of occludin and claudins in TJ membrane microdomains by EPEC infection. Mono-
layers were exposed to EPEC for the indicated times and lysed. TJ membrane microdomains were isolated by sucrose gradient centrifu-
gation. Equal amounts of protein from each fraction were subjected to SDS–PAGE and analyzed by immunoblotting with mAb specific
for flotillin-1, occludin and claudin-1 and -4. (A) EPEC infection changes the distribution of flotillin-1 in TJ membrane microdomains. The
immunoblots were analyzed quantitatively (B), and the amount of flotillin-1 in raft fractions 3–5 is shown as a percentage of the total
density detected in fractions 2–9. Results are means ± SEM. (C) EPEC infection displaced occludin from Triton X-100-insoluble fractions
to Triton X-100-soluble fractions. (D) Densitometric analysis of occludin distribution in fractions 2–9. (E,F) Distribution of claudin-1 and
claudin-4 in EPEC-infected cells.
EPEC enters host cells through tight junctions Q. Li et al.
6028 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS
proteins from TJ membrane microdomains is
required for invasion of EPEC into host cells and
subsequent infection.
Discussion
An increasing number of pathogens, including some
bacteria, viruses and parasites, have been found to
enter host cells through lipid rafts [15–17]. Recently,

it has been suggested that lipid rafts represent a spe-
cial membrane microdomain that may facilitate virus
entry [18]. Some bacteria have been found to interact
with lipid rafts of the host plasma membrane. The
mechanisms that underlie this interaction are starting
to be unraveled. As EPEC invade the gastrointestinal
tract, they must cross the TJ barrier in epithelial
cells. The entry of EPEC into cells is a complex
process, and the underlying molecular and cellular
mechanisms of EPEC-induced disruption of the
epithelial barrier are not completely understood. In
this study, we demonstrated the importance of TJ
membrane microdomains in invasion of EPEC into
host cells through epithelial TJs. We have shown that
EPEC infection induces a dramatic redistribution of
TJ proteins along the lateral membrane, with a
progressive loss in barrier function. In addition,
EPEC-induced disruption in TJ structures is associ-
ated with the redistribution of TJ proteins in TJ
membrane microdomains.
A crucial role of TJs is to prevent commensal and
pathogenic microbes from entering the hosts. The
effect of EPEC infection on TJ barrier function is
well documented [14,19]; however, the mechanisms
and molecular changes that induce the disruption are
unclear. In this study, we utilized a combination of
biochemical, immunofluorescence and ultrastructural
analyses to characterize the effect of EPEC on TJ
structure. We found that infection of intestinal epithe-
lial cells with EPEC caused dramatic disruption of TJ

structure with a decrease in TER. This was accom-
panied by a redistribution of TJ proteins along the
lateral membrane. TJ protein complexes play a key
role in maintenance of intestinal barrier integrity
[1,20]. EPEC invasion led to a redistribution of TJ
structural proteins in TJ membrane microdomians
(Fig. 2). Our data indicate that the ultrastructural
alterations of TJs most likely account for redistribu-
tion of occludin and flotillin-1 in TJ membrane micro-
domians in EPEC-infected cells. This is the first
report demonstrating that loss of TJ barrier function
following EPEC infection is correlated with displace-
ment of TJ proteins. The distribution of occludin in
TJ membrane microdomians were altered in EPEC-
infected cells, but claudins were minimally affected
(Fig. 7E,F). EPEC infection induced marked decreases
in the TER with a concomitant selective loss of
occludin, but not claudins, from TJ membrane micr-
odomians to the cytosolic compartment. It has been
suggested that loss of occludin constitutes the most
significant alteration in TJ transmembrane proteins,
and our results are similar to those reported previ-
ously [21]. EPEC infection of intestinal epithelial cells
triggers two major functional effects: inflammation
and disruption of barrier function. We have previ-
ously shown using an in vitro model that proinflam-
matory cytokines disrupt epithelial barrier function by
altering lipid composition in membrane microdomains
of TJ and displacing occludin from TJ membrane
microdomains to detergent-soluble fractions. The dis-

tribution of claudin isoforms was unaffected by cyto-
kine treatment [22]. Molecular evidence regarding the
inflammatory response caused by EPEC is consistent
with these results [22]. In the previous study, it was
reported that EPEC applied to fibroblast cells could
also cause barrier function injury characterized by
formation of an actin ‘pedestal’ resulting from rear-
rangement of the host cytoskeleton beneath adherent
bacteria [23].
Lipid rafts are the preferred entry sites for several
invasive pathogens including Salmonella, Shigella,
Listeria and Chlamydia [24]. It is likely that EPEC
also utilizes lipid rafts as the entry point. Our study
shows clearly using confocal microscopy that intracel-
lular EPEC is surrounded by vesicles enriched in the
lipid raft component flotillin-1 (Fig. 1). The results
indicate that lipid rafts are required for EPEC
invasion. Fluorescence microscopy of epithelial cell
monolayers using ZO-1-specific antibody revealed
accumulation of ZO-1 around internalized EPEC
after exposure to EPEC (Fig. 3), implying that TJ
proteins are specifically mobilized to sites of bacterial
entry. Methyl-b-cyclodextrin is commonly used to
specifically disrupt the structure of lipid rafts by
depleting cholesterol from cells. The change induced
by M-b-CD allowed non-pathogenic bacteria to
recruit of ZO-1 for penetration through epithelial
monolayers into cells. This observation also indicates
that TJ membrane microdomains are required for
bacterial invasion.

Taken together, these results indicate that EPEC
utilizes TJ membrane microdomains of plasma mem-
branes for penetration into host cells. Internalization
of EPEC occurred in TJ membrane microdomains,
and the barrier function of TJ was disrupted after
exposure to EPEC. Our results suggest for the first
time that TJ membrane microdomains serve as
Q. Li et al. EPEC enters host cells through tight junctions
FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS 6029
the portal of EPEC entry into host cells. The TJ
membrane microdomains have been shown to mediate
internalization of EPEC into the host cell and hence
infection events.
Experimental procedures
Antibodies and reagents
Flotillin-1 mAb was purchased from BD Transduction Lab-
oratories (Lexington, KY, USA). Occludin and ZO-1 were
purchased from Zymed Laboratories Inc. (San Francisco,
CA, USA). Alexa Fluor 635 secondary antibody, 4¢,6-di-
amidino-2-phenylindole (DAPI) nucleic acid stain and the
Vybrant CFDA SE cell tracer kit were purchased from
Molecular Probes (Eugene, OR, USA). Complete protease
inhibitor tablets were purchased from Boehringer Mann-
heim (Indianapolis, IN, USA). The ECL western blotting
analysis system was purchased from Amersham (Piscata-
way, NJ, USA).
Cell culture
Caco-2 cells (American Type Culture Collection, Rockville,
MD, USA) were cultured in Dulbecco’s modified Eagle’s
medium supplemented with 10% heat-inactivated newborn

calf serum and antibiotics (100 UÆmL
)1
penicillin,
100 lgÆmL
)1
streptomycin) at 37 °C in a humidified atmo-
sphere of 5% CO
2
. Before infection, the cells were placed
in antibiotic-free medium with 0.5% newborn calf serum
overnight.
Bacterial growth and infection of host cells
The wild-type strain EPEC 2348 ⁄ 69 was a generous gift from
J. Kaper and J. Michalski (Center for Vaccine Development,
University of Maryland, Baltimore, MD, USA). The non-
pathogenic strain of E. coli DH5a was also used in this
study.
Wild-type EPEC 2348 ⁄ 69 and E. coli DH5a were
grown overnight in Luria–Bertani broth. On the day of
experimentation, the bacterial cultures were diluted into
antibiotic-free cell culture medium containing 0.5% new-
born calf serum and 0.5% mannose. The bacteria were
grown at 37 °C in a shaking incubator overnight. The
bacterial suspension was then centrifuged at 1500 g for
10 min at room temperature and resuspended in culture
medium. Bacteria were added to the apical surface of the
cells on Transwell filters (CoStar, Cambridge, MA, USA)
to a final concentration of approximately 5 · 10
7
colony-

forming units per mL, corresponding to a multiplicity of
infection of 25 (25 bacteria per cell). Monolayers and
bacteria were then co-incubated in antibiotic-free medium
for specified times as indicated.
Immunofluorescence microscopy
Caco-2 cells were infected as described above. The bacteria
were preincubated with the CFDA SE cell tracer kit for 1 h
following the manufacturer’s instructions. For M-b-CD
treatment, cell monolayers were washed with NaCl ⁄ P
i
and
incubated with 10 mm M-b-CD for 1 h. The monolayers
were then co-incubated with non-pathogenic E. coli DH5a.
Control and infected monolayers of Caco-2 cells were
washed with ice-cold NaCl ⁄ P
i
to remove non-adherent bac-
teria, fixed in 3.7% paraformaldehyde in NaCl ⁄ P
i
, pH 7.4,
for 15 min at room temperature, and permeabilized in
0.2% Triton X-100 for 10 min. Cells were washed three
times with cold NaCl ⁄ P
i
and blocked in 5% goat serum for
10 min. Monolayers were then labeled with anti-occludin
(1 : 100) or anti-ZO-1 (1 : 100) serum overnight, and then
stained with Alexa Fluor 635-labeled goat anti-mouse sec-
ondary IgG (1 : 100) for 1 h. After rinsing, the specimens
were examined using an LSM510 laser scanning confocal

microscope (Zeiss, Jena, Germany).
Transmission electron microscopy
The EPEC-infected cell monolayers were washed, fixed with
2.5% glutaraldehyde and then post-fixed with 1% OsO
4
,
embedded in Epon 812 (Fluka AG, Buchs, Switzerland)
and thin-sectioned. Sections were stained with 2% uranyl
acetate and 0.2% lead citrate and viewed with an electron
microscope (H-600, Hitachi, Tokyo, Japan).
Transepithelial electrical resistance measurement
Confluent monolayers grown on 0.33 cm
2
Transwell filters
were used for TER assessment. TER was measured using
an EVOM epithelial volt-ohm meter (World Precision
Instruments, Stevenage, UK) as previously described
[25]. Results are expressed as a percentage of initial
resistance.
Isolation of TJ membrane microdomains and
Western blot analysis
After bacterial infection, cells were washed with NaCl ⁄ P
i
and
harvested by scraping into ice-cold lysis buffer (50 mm Tris,
25 mm KCl, 5 mm MgCl
2
Æ6H
2
O, 2 mm EDTA, 40 mm NaF,

4mm Na
3
VO
4
, pH 7.4, containing 1% Triton X-100 and a
protease inhibitor mixture). TJ membrane microdomain frac-
tions were isolated as previously described [4]. Protein con-
centrations were determined by the Bradford method.
Aliquots of equal protein content were isolated by
SDS–PAGE and transferred to poly(vinylidene difluoride)
membranes as previously described [22]. Western blots were
quantified by densitometric analysis using quantity one
1d analysis software (Bio-Rad, Hercules, CA, USA).
EPEC enters host cells through tight junctions Q. Li et al.
6030 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS
Statistical analysis
Data are expressed as means ± SEM. The significance of
differences was determined using a paired Student’s t test.
P values < 0.05 were considered to be statistically
significant.
Acknowledgements
This work was supported by grants from the National
Basic Research Program (973 Program) in China
(numbers 2007CB513005 and 2009CB522405) and the
Key Project of the National Natural Science Founda-
tion in China (30830098) to J. L., and from the
National Natural Science Foundation in China
(30672061) and the Key Project of Nanjing Military
Command (06Z40) and the Military Scientific Research
Fund (0603AM117) to Q. L. We would like to thank

the Deutscher Akademischer Austauschdienst
Researcher Fellowship (Bioscience Special Program,
Germany) for support of Q. L., and James Kaper and
Jane Michalski (Center for Vaccine Development,
University of Maryland, Baltimore) for the generous
gift of EPEC strain 2348 ⁄ 69.
References
1 Schneeberger EE & Lynch RD (2004) The tight junc-
tion: a multifunctional complex. Am J Physiol 286,
C1213–C1228.
2 Balkovetz DF & Katz J (2003) Bacterial invasion by a
paracellular route: divide and conquer. Microbes Infect
5, 613–619.
3 Muza-Moons MM, Schneeberger EE & Hecht GA
(2004) Enteropathogenic Escherichia coli infection leads
to appearance of aberrant tight junctions strands in the
lateral membrane of intestinal epithelial cells. Cell
Microbiol 6, 783–793.
4 Nusrat A, Parkos CA, Verkade P, Foley CS, Liang
TW, Innis-Whitehouse W, Eastburn KK & Madara JL
(2000) Tight junctions are membrane microdomains.
J Cell Sci 113, 1771–1781.
5 Knodler L, Vallance B, Hensel M, Ja
¨
ckel D, Finlay BB
& Steele-Mortimer O (2003) Salmonella type III effec-
tors PipB and PipB2 are targeted to detergent-resistant
microdomains on internal host cell membranes. Mol
Microbiol 49, 685–704.
6 Lafont F, Tran Van Nhieu G, Hanada K, Sansonetti

P & van der Goot F (2002) Initial steps of Shigella
infection depend on the cholesterol ⁄ sphingolipid
raft-mediated CD44–IpaB interaction. EMBO J 21,
4449–4457.
7 Seveau S, Bierne H, Giroux S, Prevost M & Cossart P
(2004) Role of lipid rafts in E-cadherin- and HGF-
R ⁄ Met-mediated entry of Listeria monocytogenes into
host cells. J Cell Biol 166, 743–753.
8 Jutras I, Abrami L & Dautry-Varsat A (2003) Entry of
the lymphogranuloma venereum strain of Chlamydia
trachomatis into host cells involves cholesterol-rich
membrane domains. Infect Immun 71, 260–266.
9 Wooldridge KG, Williams PH & Ketley JM (1996)
Host signal transduction and endocytosis of Campylo-
bacter jejuni. Microb Pathog 21, 299–305.
10 Coyne CB & Bergelson JM (2006) Virus-induced Abl
and Fyn kinase signals permit coxsackievirus entry
through epithelial tight junctions. Cell 124, 119–131.
11 Duncan MJ, Shin JS & Abraham SN (2002) Microbial
entry through caveolae: variations on a theme. Cell
Microbiol 4, 783–791.
12 Lafont F & van der Goot FG (2005) Bacterial invasion
via lipid rafts. Cell Microbiol 7, 613–620.
13 Simons K & Ehehalt R (2002) Cholesterol, lipid rafts
and disease. J Clin Invest 110, 597–603.
14 Simonovic I, Rosenberg J, Koutsouris A & Hecht G
(2000) Enteropathogenic Escherichia coli dephosphory-
lates and dissociates occludin from intestinal epithelial
tight junctions. Cell Microbiol 2, 305–315.
15 Haldar K, Mohandas N, Samuel BU, Harrison T, Hil-

ler NL, Akompong T & Cheresh P (2002) Protein and
lipid trafficking induced in erythrocytes infected by
malaria parasites. Cell Microbiol 4, 383–395.
16 Pelkmans L & Helenius A (2003) Insider information:
what viruses tell us about endocytosis. Curr Opin Cell
Biol 15, 414–422.
17 Lafont F, Abrami L & van der Goot FG (2004) Bacterial
subversion of lipid rafts. Curr Opin Microbiol 7, 4–10.
18 Takeda M, Leser GP, Russell CJ & Lamb RA (2003)
Influenza virus hemagglutinin concentrates in lipid raft
microdomains for efficient viral fusion. Proc Natl Acad
Sci USA 100, 14610–14617.
19 Spitz J, Yuhan R, Koutsouris A, Blatt C, Alverdy J &
Hecht G (1995) Enteropathogenic Escherichia coli
adherence to intestinal epithelial monolayers diminishes
barrier function. Am J Physiol 268, G374–G379.
20 Mitic LL, Van Itallie CM & Anderson JM (2000)
Molecular physiology and pathophysiology of tight
junctions I. Tight junction structure and function: les-
sons from mutant animals and proteins. Am J Physiol
279, G250–G254.
21 McNamara BP, Koutsouris A, O’Connell CB, Nou-
gayre
´
de JP, Donnenberg MS & Hecht G (2001) Trans-
located EspF protein from enteropathogenic Escherichia
coli disrupts host intestinal barrier function. J Clin
Invest 107, 621–629.
22 Li QR, Zhang Q, Wang M, Zhao SM, Ma J, Luo N, Li
N, Li YS, Xu GW & Li JS (2008) Interferon-c and

tumor necrosis factor-a disrupt epithelial barrier func-
tion by altering lipid composition in membrane micro-
domains of tight junction. Clin Immunol 126, 67–80.
Q. Li et al. EPEC enters host cells through tight junctions
FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS 6031
23 Rohde M. (2007) Bacterial pathogenesis: insights into a
new world discovered by high-resolution field emission
scanning electron microscopy (FESEM). In Research
Report 2006 ⁄ 2007: Special Features pp. 46–53.
Helmholtz Centre for Infection Research, Braun-
schweig, Germany.
24 Manes S, del Real G & Martinez AC (2003) Pathogens:
raft hijackers. Nat Rev Immunol 3, 557–568.
25 Li QR, Zhang Q, Wang M, Zhao SM, Xu GW & Li JS
(2008) n-3 polyunsaturated fatty acids prevent disrup-
tion of epithelial barrier function induced by proinflam-
matory cytokines. Mol Immunol 45, 1356–1365.
EPEC enters host cells through tight junctions Q. Li et al.
6032 FEBS Journal 275 (2008) 6022–6032 ª 2008 The Authors Journal compilation ª 2008 FEBS