A novel splice variant of occludin deleted in exon 9 and its
role in cell apoptosis and invasion
Jin-Mo Gu
1
, Seung Oe Lim
1
, Young Min Park
2
and Guhung Jung
1
1 Department of Biological Sciences and Seoul National University, Korea
2 Hepatology Center and Laboratory of Hepatocarcinogenesis, Bundang Jesaeng General Hospital, Kyungkido, Korea
Occludin is a tight junction (TJ) protein and the first
identified TJ-associated molecule [1]. Occludin is the
product of a single gene located on human chromo-
some band 5q13.1 and produces several different
mRNAs as a result of alternative splicing [2].
Recently, several variants of occludin, such as
occludin 1B and a fourth transmembrane domain
(TM4)-deleted variant, have been discovered [3,4].
Occludin 1B contains a 193 bp insertion correspond-
ing to an alternatively spliced exon in the gene
encoding a unique N-terminus. Conversely, the TM4-
deleted variant has a missing fourth TM correspond-
ing to exon 4. In addition to these variants, there is
evidence to show that two distinct promoters, P1 and
P2, confer separate transcriptional start sites [5]. Pro-
moter P2 is located downstream of promoter P1, and
both promoter regions are regulated by tumor nec-
rosis factor-a [5].
Occludin ( 65 kDa) is composed of two extracellu-

(Occ
DE9
). Furthermore, comparison analysis of wild-type occludin (Occ
WT
)
and Occ
DE9
revealed that exon 9 played important roles in the induction of
mitochondria-mediated apoptosis and the inhibition of invasion, along with
the downregulation of matrix metalloproteinase expression. In addition, by
using the calcium indicator X-rhod-1, and the inositol trisphosphate recep-
tor inhibitor 2-aminoethoxydiphenyl borate, we found that Occ
WT
but not
Occ
DE9
increased calcium release from the endoplasmic reticulum. In con-
clusion, our results showed that occludin mediates apoptosis and invasion
by elevating the cytoplasmic calcium concentration and that exon 9 of
occludin is an important region that mediates these effects.
Abbreviations
2-APB, 2-aminoethoxydiphenyl borate; 5¢-aza-dC, 5¢-aza-2¢-deoxycytidine; BrdUTP, 5-bromo-2¢-deoxyuridine-5¢-triphosphate; ER, endoplasmic
reticulum; ERK1 ⁄ 2, extracellular signal-regulated kinase 1 ⁄ 2; HA, hemagglutinin; IP
3
, inositol trisphosphate; IP
3
R, inositol trisphosphate
receptor; JNK, Jun N-terminal kinase; MMP, matrix metalloproteinase; MSP, methylation-specific PCR; shRNA, short hairpin RNA; TJ, tight
junction; TM, transmembrane domain; TSA, trichostatin A; TUNEL, terminal deoxyribonucleotide transferase-mediated nick-end labeling.
FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS 3145

3
R) located on the endoplasmic reticu-
lum (ER), thereby releasing calcium into the cytoplasm
[20,21]. When calcium concentrations in the cytoplasm
increase, calcium accumulation occurs along the inner
membrane of mitochondria. This destabilizes mem-
brane potentials and facilitates the initiation of apop-
tosis [22]. In addition, elevated calcium levels
contribute to cell invasion [23,24]. In contrast, in the
case of occludin, an ER-mediated increase in intra-
cellular calcium levels is not considered to be the
mechanism underlying its effects in cancer cell apopto-
sis and invasion [21].
In this article, we provide evidence of an additional
occludin variant deleted in exon 9 (Occ
DE9
). On the
basis of a comparative analysis of the involvement of
wild-type occludin (Occ
WT
) and variant occludin in
apoptosis and invasion, as determined by assay, we
revealed that exon 9 played a major role in the induc-
tion of mitochondria-mediated apoptosis and the
reduction of cell invasiveness, along with the down-
regulation of matrix metalloproteinase (MMP) expres-
sion. In addition, on the basis of an analysis using
X-rhod-1, a calcium indicator, and 2-aminoethoxy-
diphenyl borate (2-APB), which inhibits IP
3

shown). To determine the detailed occludin variant
sequence, we performed 3¢-RACE and obtained an
exon 9-deleted splice variant, which was compatible
with the results of PCR (Fig. 1C).
In a previous report, occludin was shown to involve
two different promoters [5]. Depending on the pro-
moter, either exon 1 or exon 1a was selected for the
alternative splicing process [5]. Using primers designed
for exon 1 and exon 1a, we compared the promoters
involved in Chang and Huh7 cells. As shown in
Fig. 1D, PCR products containing exon 1 were pro-
duced in Huh7 cells but not in Chang cells. To deter-
mine whether the different splice variants were caused
by a difference in the boundary sequence deciding the
splicing points between exon 8 and exon 9, we com-
pared the sequences of these regions in Huh7 and
Chang cells; no difference was found (Fig. 1E).
Figure 1F depicts the schematic of Occ
DE9
obtained
from Chang cells as compared to Occ
WT
, based on
PCR analysis.
Occ
WT
is epigenetically silenced by promoter
hypermethylation in Chang cells
On the basis of the different usages of exon 1 and
exon 1a in Huh7 and Chang cells, we decided to ana-

genes into pCMV ⁄ HA and performed immuno-
cytochemistry. The expression of each construct in
Chang, Hep3B and Huh7 cells was analyzed using
immunoblotting (Fig. 3A). As shown in Hep3B cells
transfected with a control plasmid, endogenous occlu-
din was intermittently localized in the membrane and
diffused through the cytoplasm. Occludin in the mem-
brane of Huh7 cells was localized in a more discon-
nected fashion than that in Hep3B cells. In Chang
cells, stained occludin was observed in the cytoplasmic
region. Additionally, in Occ
WT
-overexpressing Hep3B
and Huh7 cells, regions stained with antibody to hem-
agglutinin (HA), indicating the expression of exoge-
nous occludin, were located in the cell membranes. In
Occ
DE9
-overexpressing cells, however, no regions
stained with antibody to HA were observed in the
membrane (Fig. 3B). In contrast, neither Occ
WT
nor
A
B
DE
F
C
Fig. 1. Identification of a human occludin splice variant in liver cell lines. (A) Occludin protein expression was examined with an antibody to
the N-terminus of occludin in HepG2, Hep3B, Huh7, Chang and HLE cells. (B) cDNA synthesized from Chang cells was amplified with prim-

Occ
DE9
were localized in the membranes of Chang cells
(Fig. 3B).
Mitochondrial apoptosis is induced by Occ
WT
overexpression but not Occ
DE9
overexpression
Consistent with previous observations [12], in Occ
WT
-
overexpressing cells, cell proliferation decreased, and
the number of apoptotic cells increased as compared
to findings in control cells. These changes, however,
were not observed in Occ
DE9
-overexpressing cells
(Fig. 4A,B).
To identify the pathways linked to apoptosis
induced by Occ
WT
, we analyzed caspase 3 activity and
the expression of several apoptosis-related genes.
Owing to the overexpression of Occ
WT
, the expression
of the apoptotic genes BAX and Apaf-1 increased, and
that of the antiapoptotic gene Bcl-2 decreased. Occ
DE9

-overexpressing
cells but not in Occ
DE9
-overexpressing cells
Using a Matrigel invasion assay and analysis of MMP
expression, we next determined whether occludin
expression in cancer cells was responsible for invasion.
The Matrigel invasion assay revealed that significantly
fewer Occ
WT
-overexpressing Chang and Huh7 cells
were invasive as compared to the number of invasive
control and Occ
DE9
-overexpressing cells (Fig. 6A). Spe-
cifically, in Chang and Huh7 cells, Occ
WT
significantly
decreased the expression of MMP2, MMP7, MMP9
and MMP14, and of MMP1, MMP2, MMP3, MMP7,
MMP9 and MMP14, respectively (Fig. 6B).
To determine the function of occludin, we reduced
occludin expression in Huh7 cells by using occludin
short hairpin RNA (shRNA). After determining the
reduced expression level of occludin with immuno-
blotting (Fig. 7A), we analyzed the effects of down-
regulated occludin on apoptotic sensitization for H
2
O
2

WT
and Occ
DE9
were cloned into pCMV ⁄ HA, a mammalian overexpression vector
(pCMV ⁄ HA-Occ
WT
and pCMV ⁄ HA-Occ
DE9
), and their expression
was confirmed by immunoblot analysis. (B) Green fluorescence
indicates endogenous occludin stained with antibodies to occludin
in Chang, Hep3B and Huh7 cells transfected with pCMV ⁄ HA. Red
fluorescence indicates exogenous occludin stained with antibodies
to HA in Chang, Hep3B and Huh7 cells transfected with pCMV ⁄ HA-
Occ
WT
or pCMV ⁄ HA-Occ
DE9
. Cont, pCMV ⁄ HA-transfected
cells; WT, Occ
WT
-overexpressing cells; DE9, Occ
DE9
-overexpressing
cells.
Exon 9 of occludin in apoptosis and invasion J M. Gu et al.
3148 FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS
To confirm the effects of occludin on the cytoplas-
mic calcium concentration, we analyzed calcium
concentrations by using X-rhod-1, a calcium indicator.

DE9
. The concentration of
2-APB was determined by a proliferation assay, and
20 lm 2-APB did not induce apoptosis in Chang cells
(data not shown). In 2-APB-treated Chang cells, no
difference in apoptosis was seen among the control and
A
BC
D
Fig. 4. Occ
WT
but not Occ
DE9
induces mitochondria-mediated apoptosis by the modulation of expression of apoptosis-related genes and
caspase 3 activity. Chang, Hep3B and Huh7 cells transfected with pCMV ⁄ HA, pCMV ⁄ HA-Occ
WT
or pCMV ⁄ HA-Occ
DE9
. (A) The viability of
cells was determined by CCK-8 assay. (B) Apoptotic cell numbers were calculated by counting BrdUTP-incorporating cells. (C) The levels of
BAX, Bcl-2 and Apaf-1 were determined using real-time RT-PCR. (D) Caspase 3 activity was determined using the CaspACE colorimetric
assay. All results in (A), (B), (C) and (D) are expressed as the fold ratio relative to control (Cont). All numerical data represent mean and stan-
dard deviation of three independent experiments. Cont, pCMV ⁄ HA-transfected cells; WT, Occ
WT
-overexpressing cells; DE9, Occ
DE9
-overex-
pressing cells; *P < 0.05.
Fig. 5. Phosphorylation of ERK1 ⁄ 2 in Occ
WT

whereas their upregulation induces apoptosis and
inhibits invasion [8]. Occludin, one of the TJ proteins,
has also been shown to have several types of splice
variant and antitumorigenic activity [3,4,27]; however,
its effects on and implications for tumorigenesis are
poorly understood. Here, we present a novel occludin
splice variant deleted in exon 9; in Chang cells, this
variant acts via the P2 promoter. As the P2 promoter
was involved in occludin action, and as no wild-type
form was observed in Chang cells, the methylation
status of the occludin promoter P1 could be
examined. Therefore, we tested the occludin promoter
P1-containing CpG island region in liver cell lines, and
observed that the occludin promoter was strongly
methylated only in Chang cells.
Osanai et al. reported that an occludin mutant
deleted in region 478–522, which contains exon 9,
could not localize in the membrane or induce apopto-
sis in mammary cell lines [12]. In addition, Wang et al.
described another occludin mutant deleted in the
whole cytoplasmic tail. This mutant did not exert
a potent inhibitory effect on Raf1-mediated tumori-
genesis [15]. In liver cells, as in mammary cell lines, we
observed that Occ
WT
was localized in the membrane,
whereas Occ
DE9
was localized in the cytoplasm. In pre-
vious reports, several factors, such as the proteins

WT
or pCMV ⁄ HA-Occ
DE9
. All numerical data represent the mean and standard deviation of three independent
experiments. Cont, pCMV ⁄ HA-transfected cells; WT, Occ
WT
-overexpressing cells; DE9, Occ
DE9
-overexpressing cells; *P < 0.05.
Exon 9 of occludin in apoptosis and invasion J M. Gu et al.
3150 FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS
Apaf-1 and the induction of caspase 3 activity in
Occ
WT
-overexpressing cells indicated that occludin
affected mitochondria-mediated apoptosis [25]. The
mitogen-activated protein kinase pathway, which is
known to exhibit some correlation with occludin
expression, is inactivated by increased expression of
this protein in HeLa cells [12]. Our data showed that
Occ
WT
activated only ERK1 ⁄ 2 and not p38 and JNK.
On the basis of our data, exon 9 of occludin may play
a role in altering cell behavior via the internal cellular
signaling pathway. Furthermore, we newly confirmed
that Occ
WT
located in the cytoplasmic region induced
the same effects (Fig. 3).

WT
and Occ
DE9
, the exon 9 region played
a significant role in promoting apoptosis and inhibiting
invasion by regulating signaling pathways. Calcium
release from the ER has been described as one of the
mechanisms of this regulation. On the basis of the
above findings, our research provides new insights into
the role of exon 9 in the regulation of apoptosis and
invasion in liver cell lines.
A
B
Fig. 7. Effects of occludin shRNA on apoptosis and invasiveness. Huh7 cells were transfected with five constructs producing predesigned
occludin shRNA. (A) Immunoblot analysis confirmed the silencing effects of occludin shRNA. The degree of occludin downregulation is repre-
sented as a bar graph in the right panel. (B) Apoptosis of Huh7 cells transfected with occludin shRNA or control shRNA was analyzed using
a TUNEL assay in the presence of 400 and 800 l
M H
2
O
2
. The number of apoptotic cells is expressed as the ratio of these cells to the num-
ber of apoptotic control shRNA-transfected cells (Cont shRNA). (C) The invasiveness of Huh7 cells transfected with occludin shRNA or con-
trol shRNA was analyzed using a Matrigel assay. The number of invasive cells was determined by counting stained cells and then
normalizing to the number in the case of control shRNA-transfected cells. All numerical data represent the means and standard deviations of
three independent experiments. –, control shRNA-transfected cells; +, occludin shRNA-transfected cells; *P < 0.05.
J M. Gu et al. Exon 9 of occludin in apoptosis and invasion
FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS 3151
Experimental procedures
Cell cultures and treatment

microscope. Chang cells were cotransfected with pCMV ⁄ HA, pCMV ⁄ HA-Occ
WT
or pCMV ⁄ HA-Occ
DE9
and with pEGFP-N1 vector. The cells
were stained with X-rhod-1, a calcium indicator, and then analyzed. Each image shows one cell in the center. (B) The relative fluorescence
intensity is represented as a bar graph. Chang cells were transfected with pCMV ⁄ HA, pCMV ⁄ HA-Occ
WT
or pCMV ⁄ HA-Occ
DE9
, and then trea-
ted with 20 l
M 2-APB. (C) Apoptotic cell numbers were calculated by counting BrdUTP-incorporating cells. (D) The levels of BAX, Bcl-2 and
Apaf-1 were determined using real-time RT-PCR analysis. The relative expression levels are represented as a bar graph. (E) Immunoblot anal-
yses using specific antibodies against ERK1 ⁄ 2, pERK1 ⁄ 2 and b-actin were performed. (F) Invasiveness of cells was analyzed using a Matrigel
assay. Cells penetrating the Matrigel-coated inserts were stained with crystal violet. The number of invasive cells was determined by count-
ing the number of stained cells. All results in (B), (C), (D), (E) and (F) are expressed as the fold ratio relative to control (Cont). All numerical
data represent the means and standard deviations of three independent experiments. Cont, pCMV ⁄ HA-transfected cells; WT, OccWT-over-
expressing cells; DE9, OccDE9-overexpressing cells; +, 2-APB-treated cells; *P < 0.05.
Exon 9 of occludin in apoptosis and invasion J M. Gu et al.
3152 FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS
constructs expressing occludin shRNA were purchased from
Sigma and transfected with FuGENE 6 (Roche). Transfect-
ed cells were selected with puromycin (0.5 lgÆmL
)1
)
(Sigma). The pLKO.1 vector that does not contain an
shRNA insert was used as a control.
3¢-RACE
For 3¢-RACE, modified protocols from Scotto-Lavio et al.

transcribed with oligo-dT by using avian myeloblastosis
virus (AMV) reverse transcriptase (Promega, Madison, WI,
USA). For PCR of occludin variants, the primers used were:
5¢-ACTCGACAATGAACAATCCGTCAGAA-3¢ (sense)
and 5¢-AGAGTATGCCATGGGACTGTCA-3¢ (antisense)
for exon 5; 5¢-TGCAGG-TGCTCTTTTTGAAGGT-3¢
(antisense) for exon 6; 5¢-GC-TCTTGTATTCCTGTAGGC
CAG-3¢ (antisense) for exon 7; 5¢-GTATTCATCAGCAG
CAGCC-3¢ (antisense) for exon 8; and 5¢-CTGTCTATCA
TAGTCTCCAACCAT-CTTC-3¢ (antisense) for exon 9.
PCR was carried out at 53 °C for 32 cycles.
Genomic DNA was extracted from the cells using a
standard phenol protocol. For PCR, the primers
used were: 5¢-CAGCAATTGTCACACATCAAGAA-3¢
(sense) and 5¢-T-ACATGTAGGTATGAAGACATCGTC
T-3¢ (antisense) for exon 9; 5¢-TCCCTGCTTCCTCTGGC
GGA-3¢ (sense) and 5¢-AGCCATAGCCATAGCCACTTC
C-3¢ (antisense) for exon 1; 5¢-CCGGAGGGTCGGGCC
CAGTT-3¢ (sense) and 5¢-AGCCATAGCCATAGCCACT
TCC-3¢ (antisense) for exon 1a; 5¢-TAATAGGCTGCTGC
TGATGAATA-3¢ (sense) and 5¢-GGTATGTGGTCACAT
TGTGAAAATT-3¢ (antisense) for the exon 8–intron
boundary; and 5¢-ACTGCCAGGCACCTTGCGTATTT-3¢
(sense) and 5¢-TATCATAGTCTCCAACCATCTTCTTGA
-3¢ (antisense) for the intron–exon 9 boundary. PCR was
carried out at 58 °C for 30 cycles. All PCR products
were analyzed on agarose gels and stained with ethidium
bromide.
Real-time RT-PCR analysis
Real-time RT-PCR analysis was performed with specifically

criptase (Promega) and oligo-dT primers. Transcript levels
were assessed by quantitative real-time PCR (ABI 7300;
Applied Biosystems); all experiments were normalized to
b-actin.
J M. Gu et al. Exon 9 of occludin in apoptosis and invasion
FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS 3153
Immunoblot analysis
Cells were washed three times in NaCl ⁄ P
i
and scraped with
lysis buffer (ReadyPrep Sequential extraction kit, Reagent 3;
Bio-Rad, Hercules, CA, USA). Next, the protein concentra-
tion was determined by measuring with Bradford reagent
(Bio-Rad). Cell lysates (20 lg) were resolved by SDS ⁄ PAGE
and transferred onto poly(vinylidene difluoride) membranes.
The blots were blocked with 5% nonfat milk in NaCl ⁄ P
i
containing 0.1% Tween-20, and probed with antibodies to
occludin (Zymed), ERK1 ⁄ 2 (pERK1 ⁄ 2; Cell Signaling Tech-
nology, Danvers, MA), pERK1 ⁄ 2 (Cell Signaling Techno-
logy), and b-actin (Sigma Aldrich). After being washed with
NaCl ⁄ Tris containing 0.1% Tween-20, the membranes were
incubated for 1 h with horseradish peroxidase-conjugated
secondary antibodies. Detection of peroxidase-coupled anti-
bodies was performed using the Western Lightning chemilu-
minescence kit (Perkin-Elmer, Boston, MA, USA).
Cell proliferation assay
For the cell proliferation assay, cells were seeded on the
96-well plates, transfected using FuGENE 6 according to
the manufacturer’s manual, and incubated in 5% CO

5
) were resuspended in serum-free medium
and then plated in the upper part of the chamber. The
lower chamber was filled with NIH ⁄ 3T3 conditioned med-
ium [29]. Boyden chambers were incubated for 48 h. Fol-
lowing removal of noninvading cells from the upper surface
with a cotton swab, invading cells were fixed, counted, and
normalized to control sample.
MSP
One microgram of genomic DNA was treated with sodium
bisulfate using the One Day MSP Kit (IN2GEN, Seoul,
Korea). To analyze the occludin promoter, we designed
MSP primers using the Methyl Primer Express sys-
tem (Applied Biosystems). Primers for methylated DNA
were: 5¢-AAGTAGGCGGAGTATCGAAC-3¢ (sense) and
5¢-GAAAAAACGCGATCCTACTT-3¢ (antisense). Primers
for unmethylated DNA were: 5¢-GAAGTAGGTGGAGT
ATTGAAT-3¢ (sense) and 5¢-CAAAAAAACACAATCCT
ACTT-3¢ (antisense).
Caspase 3 activity
Cells subjected to the caspase assay were seeded on a
24-well plate and transfected with FuGENE 6. The caspase
assay was performed using the CaspACE colorimetric assay
kit as described by the manufacturer (Promega). Twenty-
four hours after transfection, cells were harvested and lysed
with supplied lysis buffer by freeze–thawing. Ac-DEVD-
pNA, caspase 3 substrate, was added to the cell extract and
incubated for 4 h. Measurement of the caspase activity was
done at 405 nm.
Green fluorescent protein expression construct

National R&D Program for Cancer Control, the Min-
istry of Health & Welfare, Korea (0520020). Seung Oe
Lim and Jin-Mo Gu were supported by a BK21
Research Fellowship from the Korea Ministry of
Education and Human Resources Development.
References
1 Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura
S, Tsukita S & Tsukita S (1993) Occludin: a novel inte-
gral membrane protein localizing at tight junctions.
J Cell Biol 123, 1777–1788.
2 Saitou M, Ando-Akatsuka Y, Itoh M, Furuse M, Inaz-
awa J, Fujimoto K & Tsukita S (1997) Mammalian
occludin in epithelial cells: its expression and subcellular
distribution. Eur J Cell Biol 73, 222–231.
3 Mankertz J, Waller JS, Hillenbrand B, Tavalali S,
Florian P, Schoneberg T, Fromm M & Schulzke JD
(2002) Gene expression of the tight junction protein
occludin includes differential splicing and alternative
promoter usage. Biochem Biophys Res Commun 298,
657–666.
4 Ghassemifar MR, Sheth B, Papenbrock T, Leese HJ,
Houghton FD & Fleming TP (2002) Occludin TM4(–):
an isoform of the tight junction protein present in pri-
mates lacking the fourth transmembrane domain. J Cell
Sci 115, 3171–3180.
5 Mankertz J, Tavalali S, Schmitz H, Mankertz A, Riec-
ken EO, Fromm M & Schulzke JD (2000) Expression
from the human occludin promoter is affected by tumor
necrosis factor alpha and interferon gamma. J Cell Sci
113, 2085–2090.

lular permeability and transepithelial electrical resistance
and disruption of the apical–basolateral intramembrane
diffusion barrier by expression of a mutant tight junction
membrane protein. J Cell Biol 134, 1031–1049.
14 Chen Y, Merzdorf C, Paul DL & Goodenough DA
(1997) COOH terminus of occludin is required for tight
junction barrier function in early Xenopus embryos.
J Cell Biol 138, 891–899.
15 Wang Z, Mandell KJ, Parkos CA, Mrsny RJ & Nusrat
A (2005) The second loop of occludin is required for
suppression of Raf1-induced tumor growth. Oncogene
24, 4412–4420.
16 Balda MS, Gonzalez-Mariscal L, Matter K, Cereijido
M & Anderson JM (1993) Assembly of the tight junc-
tion: the role of diacylglycerol. J Cell Biol 123 , 293–302.
17 Balda MS, Gonzalez-Mariscal L, Contreras RG,
Macias-Silva M, Torres-Marquez ME, Garcia-Sainz JA
& Cereijido M (1991) Assembly and sealing of tight
junctions: possible participation of G-proteins, phos-
pholipase C, protein kinase C and calmodulin. J Membr
Biol 122, 193–202.
18 Chen YH, Lu Q, Goodenough DA & Jeansonne B
(2002) Nonreceptor tyrosine kinase c-Yes interacts with
occludin during tight junction formation in canine kid-
ney epithelial cells. Mol Biol Cell 13, 1227–1237.
19 Basuroy S, Seth A, Elias B, Naren AP & Rao R (2006)
MAPK interacts with occludin and mediates EGF-
induced prevention of tight junction disruption by
hydrogen peroxide. Biochem J 393, 69–77.
20 Matter K & Balda MS (2003) Signalling to and from

form of cathepsin B contributes to invasiveness in
cancer. Cancer Res 61, 3493–3500.
30 Giannelli G, Bergamini C, Fransvea E, Sgarra C &
Antonaci S (2005) Laminin-5 with transforming growth
factor-beta1 induces epithelial to mesenchymal transi-
tion in hepatocellular carcinoma. Gastroenterology 129,
1375–1383.
31 Mori K, Shibanuma M & Nose K (2004) Invasive
potential induced under long-term oxidative stress in
mammary epithelial cells. Cancer Res 64, 7464–7472.
Exon 9 of occludin in apoptosis and invasion J M. Gu et al.
3156 FEBS Journal 275 (2008) 3145–3156 ª 2008 The Authors Journal compilation ª 2008 FEBS