Histochem Cell Biol (1998) 109:19–24 © Springer-Verlag 1998

O R I G I N A L PA P E R

Kentaro Kawasaki · Yoshitake Hayashi

Yoshihumi Nishida · Akinori Miki · Hiroshi Itoh

Developmental expression of the tight junction protein,

occludin, in the gastrointestinal tract of the chick embryo

&misc:Accepted: 30 May 1997

&p.1:Abstract The developmental expression of occludin junction), zonula adherens, and desmosome (Farquhar
was studied biochemically in whole chick embryos to and Palade 1963). The tight junction is situated on the
determine when intercellular tight junctions develop. apical surface of the epithelium and seals the paracellu-
Occludin mRNA was first detected after 3 days of incu- lar space to prevent direct penetration of the luminal
bation by the reverse transcriptase-polymerase chain re- contents into the intercellular spaces. The tight junction
action. On northern blot analysis, although occludin may also form a continuous permeability barrier be-
mRNA was not discernible in 3-day-old embryos, weak tween adjacent cells to regulate the flux of molecules
but clear expression was noted on day 4 of incubation through the paracellular space, enabling cells to carry
and increased dramatically in 5-day-old embryos. Occlu- out polarized transport (Anderson and van Itallie 1995).
din was not detectable on days 3 or 4 of incubation by In recent years, several protein components of the tight
western blot analysis, and was first detected in 5-day-old junction have been identified, including ZO-1 (Steven-
embryos. In addition, the expression of occludin was ex- son et al. 1986), ZO-2 (Jesaitis and Goodenough 1994),
amined immunohistochemically in the gastrointestinal cingulin (Citi et al. 1988), 130 kDa (Balda et al. 1993),
tract of 3- to 21-day-old embryos. Immunoreactivity for 7H6 (Zhong et al. 1993), and occludin (Furuse et al.
occludin was not expressed on day 3 of incubation. On 1993). On the basis of the biochemical characteristics of
day 4 of incubation, weak immunoreactivity was demon- these molecules, a testable hypothesis as to how the
strated in the gastrointestinal tract, and gradually became tight junction might be organized and regulated has
stronger with development. By day 11 of incubation, a been proposed (Barry 1993).
positive immunoreaction was obtained only on the apical Of these molecules, occludin has been isolated as a
surfaces of the epithelial cells, i.e., at the junctional com- unique, approximately 65-kDa protein, by raising
plexes, while weak immunoreactivity was diffusely dis- monoclonal antibodies against junction-enriched mem-
tributed throughout the epithelial cells. The possible branes from chicken liver. An electron microscopy
roles of occludin in the developing gastrointestinal tract study using the immunogold labeling method has re-
are discussed.&bdy: vealed that occludin is localized preferentially at junc-
tional complex contact sites in both epithelial and endo-
thelial cells. Occludin has also been reported to bind to
Introduction the cytoplasmic surfaces of the cell membranes and to
have a COOH-terminus which connects directly with
In mammals, intercellular junctions (junctional com- ZO-1 and forms heterodimers with ZO-2 and 130 kDa.
plexes) are composed of the zonula occludens (tight These findings suggest that occludin might be one of
the key proteins of the tight junction (Furuse et al.
1993).
K. Kawasaki (✉) · Y. Hayashi · Y. Nishida · H. Itoh To date, however, it remains unknown when and how
First Division of Pathology, Kobe University School of Medicine,
7-5-1 Kusunoki-cho, chuo-ku, Kobe 650, Japan this protein is expressed in developing epithelial tissue.
Tel. +81-78-341-7451 ext. 3282; fax +81-78-361-7269 In this study, therefore, to determine the possible roles of
occludin in the tight junction, the developmental expres-
K. Kawasaki
First Division of Surgery, Kobe University School of Medicine, sion of this protein and its mRNA was examined bio-
Kobe, Japan chemically in whole chick embryos, and the localization
A. Miki
of occludin was investigated immunohistochemically in
Faculty of Health Science Medical Science, the gastrointestinal tract.
Kobe University School of Medicine, Kobe, Japan&/fn-block:
20
Western blot analysis
Materials and methods
Western blot analysis was performed using the method described
Materials by Kameda et al. (1990). Fifty micrograms of protein samples
were subjected to 7.5% SDS–PAGE followed by electrotransfer
Fertilized white Leghorn eggs were incubated in a humidified at- onto nitrocellulose filters. The monoclonal antibody Oc-2 (Oc-2
mosphere at 37° C. Embryos used in this study were obtained mAb) was used as the primary antibody (Furuse et al. 1993) at a
from the eggs at days 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 14, 17, and 21 of dilution of 1:50. For detection of the immunocomplex, an en-
incubation. The samples were quickly frozen using liquid nitro- hanced chemiluminescence western blotting detection system
gen. On each of days 2–11 of incubation, whole chick embryos (Amersham) was used. Finally, the membrane was exposed to X-
were used as samples, while on days 14, 17, and 21, only the gas- ray film (Hyperfilm-ECL, Amersham).
trointestinal tract was used for samples.

Immunohistochemistry
RNA extraction
Immunohistochemistry of tissue sections was performed using a
To extract mRNA, 100– to 200-mg samples were used on each Large Volume Dako LSAB Kit, peroxidase (Dakopatts, Copenha-
day. Total RNA was isolated from whole chick embryos (days gen) according to the manufacturer’s instructions. Five-micrometer-
2–7) by the guanidinium thiocyanate–phenol chloroform extrac- thick frozen sections were cut in a cryostat, mounted on glass
tion method using Isogen (Takara, Tokyo). slides, air-dried, and fixed in 95% ethanol at 4° C for 30 min and
100% acetone at room temperature for 1 min. The sections were in-
cubated with 3% hydrogen peroxide for 15 min. They were then in-
Reverse transcriptase–polymerase chain reaction (RT–PCR) cubated with blocking solution for 30 min and incubated with Oc-2
mAb diluted 1:100 in 0.01% phosphate-buffered saline overnight.
For reverse transcription, 50 pmol of oligo dT was annealed to a 5- After washing with TRIS buffer, visualization was done with ami-
µg sample of total RNA and extended by 200 U RNase reverse noethylcarbazole chromogen containing hydrogen peroxide. The
transcriptase (Takara) in a reaction volume of 50 µl. Five micro- sections were finally counterstained with hematoxylin and mounted.
grams of cDNA was used in the PCR, which included 50 pmol
of occludin primer (5′-TTGGATGAGTCCCAGTATGA-3′) as
the upstream and 50 pmol of occludin primer (5′-
GTGCTGTTATTATGGTAGTG-3′) as the downstream primer in a Results
reaction volume of 100 µl containing 1 U Taq DNA polymerase
(Takara) and a buffer supplemented with 2 mM MgCl 2 and 0.1 RT–PCR
mM deoxynucleotide triphosphates. The 622-bp occludin se-
quence was amplified by PCR carried out in a thermocycler for 30
cycles of 94° C for 1 min, 63° C for 1 min, and 72° C for 2 min. The expression of occludin mRNA was determined by
After these operations, 30 µl of the PCR reaction solution was RT–PCR on days 2–7 of incubation in whole chick em-
electrophoresed in a 1% agarose gel. bryos. On day 2 of incubation, occludin mRNA was not
amplified. PCR amplified a 622-bp band corresponding
Northern blot analysis to occludin mRNA, as shown in Fig. 1. This band was
apparent in 3- to 7-day-old embryos. There were no ob-
Northern blot analysis was performed using a modification of the vious differences in the intensity of the band between
non-radiographic method of Kohri et al. (1993). To acquire digoxi- these time points.
genin (DIG)-labeled cDNA fragments as probes for northern blot
analysis, cDNA was amplified by PCR with a PCR DIG Probe
Synthesis kit (Boehringer, Mannheim). PCR was performed using
the occludin primer described above. Northern blot analysis
Total cellular RNA was isolated from the frozen samples of
chick embryo, and 20 µg total RNA per lane was electrophoresed The amount of occludin mRNA in whole chick embryos
in 1% agarose gels containing 7.5% formaldehyde and transferred
to a nylon membrane (Hybond-N+; Amersham, Tokyo) by vacuum was determined by northern blot analysis on days 3–7 of
blotting with 20×SCC. The membranes were baked for 2 h at 80°
C and prehybridized in prehybridization buffer (0.5% SDS,
5×SSC, 10% Denhardt’s solution, 10 mM Na2HPO4, 50% form-
amide, 0.1 mg/ml sonicated genomic DNA) for 3 h at 50° C. The RT-PCR
RNA was probed by hybridization at 50° C with a DIG-labeled
174 2 3 4 5 6 7 Days
DNA probe under prehybridization conditions overnight. The
membranes were washed with 0.2×SSC containing 0.1% SDS (15
min×2 at 60° C). The membranes were rinsed in DIG buffer 1 and
incubated with 1% blocking reagent (Boehringer) in DIG buffer 1
for 60 min, and incubated with 0.5 µl anti-DIG alkaline phos- 622bp
phate-conjugated antibody diluted 1:10000 in DIG buffer 1 with
0.2% polyoxethylene sorbitan monolaurate (Wako Pure Chemical
Industries, Osaka) for 30 min. The membranes were then washed
with DIG buffer 1 (15 min ×2) and rinsed with DIG buffer 3 for 3 -actin
min. For detection of signals, the membranes were incubated at
37° C with disodium 3-(4-methoxyspiro {1,2-dioxethane-3,2′-(5′-
chloro) trichloro [3.3.1.13,7] decan}-4-phenylphosphate) (Boeh- Fig. 1 Reverse transcriptase–polymerase chain reaction
ringer) diluted 1:100 in assay solution (100 mM diethanolamine, 2 (RT–PCR) of occludin. The amplified PCR product of the 622-bp
nM MgCl2, 0.02% NaN3) for 10 min. Hybridization signals were occludin cDNA is evident (arrows). Embryos at 2 days (lane 1), 3
exposed for 3 h at room temperature to X-ray film (Hyperfilm- days (lane 2), 4 days (lane 3), 5 days (lane 4), 6 days (lane 5), and
ECL; Amersham). 7 days old (lane 6)&ig.c:/f
 21
Northern blot analysis Western blot analysis
3 4 5 6 7 Days 3 4 5 6 7 8 Days

58kDa
2.Okbp

Fig. 3 Western blot analysis of occludin using anti-occludin

-actin monoclonal antibody (Oc-2 mAb). Arrows show recognized bands
corresponding to occludin. Embryos at 3 days (lane 1), 4 days
(lane 2), 5 days (lane 3), 6 days (lane 4), 7 days (lane 5), and 8
days old (lane 6)&ig.c:/f

Fig. 2 Northern blot analysis of occludin. Arrows show hybrid- kers” developed and formed simple networks (Fig. 4d).
ized bands corresponding to occludin. Embryos at 3 days (lane 1),
4 days (lane 2), 5 days (lane 3), 6 days (lane 4), and 7 days old Thereafter, the occludin-staining network became more
(lane 5)&ig.c:/f intense and complex as embryonic development pro-
gressed (Fig. 4e). From days 4–11 of incubation, occlu-
din staining was localized only on the apical surfaces of
incubation. A probe prepared for detection of occludin the epithelial cells. In 14-day-old embryos, besides
mRNA revealed a 2-kbp band (Fig. 2). No band corre- strong immunoreactivity on the apical surfaces, weak
sponding to occludin mRNA was detected in 3-day-old immunoreactivity was diffusely distributed throughout
embryos. A weak but clear band was first detected in the cytoplasm of the epithelial cells, but not in the nucle-
samples from 4-day-old embryos, and this band became us (Fig. 4f). In 17- and 21-day-old embryos, the distribu-
much stronger on days 5–7 of incubation; no significant tion of immunoreactivity in the epithelium of the gastro-
differences in the intensity of these bands were apparent. intestinal tract as almost the same as that in 14-day-old
embryos (Fig. 4g, h).

Western blot analysis

Discussion
The amount of occludin protein in whole chick embryos
was measured by western blot analysis on days 3–8 of The full-length cDNA sequence of chicken occludin pre-
incubation. The (Oc-2 mAb, detected a 58-kDa band dicts a novel 504-amino acid protein with four potential
(Fig. 3). No positive signal was noted in 3- or 4-day-old hydrophobic transmembrane helices and two extracellu-
embryos. In 5- to 8-day-old embryos, as clear 58-kDa lar loops. This amino acid composition suggests that the
band corresponding to occludin was detected. These extracellular loops could create a tight paracellular seal
bands had almost the same intensities. through hydrophobic contact with loops from adjacent
cells or, more probably, by dissolving the lipid bilayer of
the adjacent cell membrane. There are several similar
Immunohistochemical study proteins, such as connexin, a protein of the gap junction,
synaptophysin in synaptic vesicles, and proteolipopro-
To examine the localization of occludin in the epitheli- tein, which fuse myelin membranes. These molecules
um, an immunohistochemical study was performed using have also been reported to possess four transmembrane
tissue from the gastrointestinal track of chick embryos helices, and are considered to be responsible for the fu-
3–11, 14, 17, and 21 days old (Fig. 4). In this study, two sion or sealing of membranes. Although the primary
embryos were examined on each day of incubation. The amino sequence of occludin is not homologous with
intensity and distribution of the occludin signal in both those of these molecules, occludin might also play an
samples were almost the same. important role in the sealing mechanism at the apical
In 3-day-old embryos, no distinct immunoreactivity surface of epithelial cells.
for occludin could be found anywhere in the gastrointes- It has also been reported that when chicken occludin
tinal tract. In 4-day-old embryos, faint immunoreactivity cDNA was used to transfect a cultured mammalian epi-
was recognized along the apical surface of the epitheli- thelial cell line (Madin–Darby bovine kidney cells),
um (Fig. 4a). In 5-day-old embryos, occludin staining occludin was expressed at the endogenous tight junction
was markedly stronger than on the previous days. In ad- and participated in its functions (Furuse et al. 1994).
dition, short spike-like zones of reactivity were noted These findings suggest that occludin might provide a
between the epithelial cells, mainly near the apical sur- common molecular basis for the sealing function at tight
face (Fig. 4b, c). In 7-day-old embryos, small “whis- junctions in both mammalian and chicken organs.
22
 23

In this study, occludin mRNA was first detected by ment of the chick intestine (Kimura et al. 1996). In intes-
RT–PCR in the chick embryo on day 3 of incubation, tinal samples from embryos aged 13 days or more, the
and by northern blot analysis on day 4 of incubation, structure of the tight junction was already apparent. In the
suggesting that synthesis of occludin begins in the chick present study, on days 14, 17, and 21 of incubation, be-
embryo at a very early stage of development. On western sides a distinct immunoreaction for occludin in the apical
blot analysis, occludin protein was first detected after portions, weak immunoreactivity was diffusely distribut-
day 5, and its band reached an intensity plateau in 6- to ed throughout the cytoplasm of the epithelial cells. These
8-day-old embryos. This implies that the formation of findings suggest that occludin is actively synthesized in
the tight junction also starts at an early stage of develop- the epithelium of the gastrointestinal tract during the later
ment. It has been demonstrated that primitive cell–cell embryonic stages. This abundant expression of occludin
contacts are already present in the neural tubes and so- might be responsible for the rapid functional maturation
mites of 2-day-old chick embryos (Miki and Mizoguti of the tight junction in the gastrointestinal tract.
1982). In this study, we were unable to detect any ex- Occludin binds to the cytoplasmic surfaces of cell
pression of occludin mRNA or protein in 3-day-old em- membranes and its COOH-terminus connects directly
bryos. However, we found that occludin immunoreactivi- with ZO-1,which forms heterodimers with ZO-2 and 130
ty was present not only in the gastrointestinal tract, but kDa (Furuse et al. 1994). This strongly suggests that
also in the liver, lung, and neural tube (data not shown). occludin might be one of the key proteins of the tight
These findings suggest that occludin is widespread in junction. Thus it is suggested that the developmental ex-
various organs. pression of occludin might be a useful tool for helping to
Many reports have described the development of the clarify the morphological and functional developmental
junctional complex in the embryonic epithelium of ani- of the tight junction. At present, however, the interaction
mals, including the chick. Using a freeze-etching proce- of occludin with other proteins, such as cingulin and 7H6,
dure, Okamoto and Ishimura (1978) studied the develop- has not been fully clarified. Further biochemical and his-
ment of intercellular junctions in the duodenal epithelium tochemical studies of the relationships between these pro-
of chick embryos, and reported that incomplete tight teins will be necessary in order to elucidate the possible
junctions were already present after 6–7 days of incuba- role of occludin at the tight junction during development.
tion in the apical portion of the lateral plasma membrane,
suggesting that tight junctions begin to form in the gas- &p.2:Acknowledgements We thank Dr. Mikio Furuse and Shoichiro
trointestinal tract at an early developmental stage. There- Tsukita for providing us with Oc-2 monoclonal antibody and Miss
after, tight junctions develop rapidly and form complicat- Chiyomi Ikeuchi for her excellent technical assistance.
ed networks (Okamoto and Ishimura 1978). This devel-
opmental process has also been reported in the rat small
intestine and thyroid follicles (Yamamoto et al. 1992). References
Our immunohistochemical study revealed that occludin
began to be expressed in thegastrointestinal tract of the 4- Anderson JM, Itallie CM van (1995) Tight junctions and the mole-
cular basis for regulation of paracellular permeability. Am J
day-old chick embryo. At this early developmental stage, Physiol 269:467–475
occludin immunoreactivity was very weak, and was de- Balda MS et al. (1993) Assembly of the tight junction: the role of
tected only along the apical surface of the epithelium. As diacylglycerol. J Cell Biol 123:293–302
the embryo developed, the immunoreactivity gradually Barry MG (1993) Breaking through the tight junction barrier. J
Cell Biol 1631–1633
became stronger and formed more complex networks Citi S et al. (1988) Cingulin, a new peripheral component of tight
near the apical surface. This indicates that the develop- junctions. Nature 333:272–276
mental expression of occludin in the gastrointestinal tract Fallon MB et al. (1995) Altered hepatic localization and expres-
is closely correlated with the morphological as well as sion of occludin after common bile duct ligation. Am J Physiol
269:1057–1062
functional development of the tight junction. Farquhar MG, Palade GE (1963) Junctional complex in various
Okamoto and Ishimura (1978) reported that formation epithelia. J Cell Biol 17:375–409
of the tight junction in chick duodenum might be com- Fujimoto K (1995) Freeze-fracture replica electron microscopy
plete by day 18 of incubation. Recently, the distribution combined with SDS digestion for cytochemical labeling of in-
of 7H6 and occludin has been studied during develop- tegral membrane proteins. Application to the immunogold la-
beling of intercellular junctional complexes. J Cell Sci 108:
3443–3449
Furuse M et al. (1993) Occludin: a novel integral membrane pro-
tein localizing a tight junctions. J Cell Biol 123:1777–1788
▲

Fig. 4a–h Immunohistochemistry of occludin in the gastrointesti- Furuse M et al. (1994) Direct association of occludin with ZO-1
nal tract of the chick embryo using Oc-2 mAb. Embryos at 4 days and its possible involvement in the localization of occludin at
(a), 5 days (b), 6 days (c), 7 days (d), 10 days (e), 14 days (f), 17 tight junctions. 127:1617–1626
days (g), and 21 days old (h). In a–c the immunoreaction is evi- Jesaitis LA, Goodenough DA (1994) Molecular characterization
dent only at the apical surface (arrowheads) of the epithelium of of tissue distribution of ZO-2, a tight junction protein homolo-
the gastrointestinal tract. In contrast, in f–h, besides the apical sur- gous to ZO-1 and the Drosophila discs-large tumor suppressor
face of the epithelium, the cytoplasm of the epithelial cells is protein. J Cell Biol 124:949–961
stained brown, whereas the cytoplasm of the mesenchymal cells is Kameda T et al. (1990) Expression of eRBB2 in human gastric
not. (E Epithelium of the gastrointestinal tract, M mesenchymal carcinomas: relationship between p185eRBB2 expression and
cells of the gastrointestinal tract). Magnification ×725 the gene amplification. Cancer Res 50:8002–8009
24
Kimura M et al. (1996) Comparison between the distribution of Stevenson BR et al. (1986) Identification of ZO-1: a high molecu-
7H6 tight junction-associated antigen and occludin during the lar weight polypeptide associated with the tight junction (zon-
development of chick intestine. Cell Struct Funct 21:91–96 ula occludens) in a variety of epithelia. J Cell Biol 103:
Kohri K et al. (1993) Structure and expression of the mRNA en- 755–766
coding urinary stone protein (osteopontin). J Biol Chem 268: Yamamoto M, Toyota T, Kataoka K (1992) Electron microscope
15180–15184 observations on the formation of primitive villi in rats small
Miki A, Mizoguti H (1982) Acetylcholinesterase activity in the intestine with special reference to intercellular junctions. Arch
myotome of the early chick embryo. Cell Tissue Res 227:23–40 Histol Cytol 55:551–560
Okamoto H, Ishimura K (1978) Freeze-etching observation on the Zhong Y et al. (1993) Monoclonal antibody 7H6 react with a nov-
development of intercellular junctions of the duodenum epi- el tight junction-associated protein distinct from ZO-1, ci-
thelial cells in the chick embryo. Arch Histol Cytol 41: ngulin and ZO-2. J Cell Biol 120:477–483
325–327