![S1 fibers are unrelated to cell proliferation and cellular metabolic activity. (A) RNA and DNA synthesis activities. ARL cells cultured to a 50-60% confluency (low density cells) or for 1 and 3 weeks at full confluency were incubated for 45 minutes in a fresh medium containing [ 3 H]uridine or [ 3 H]thymidine. Incorporated radioactivities were determined with respect to DNA, and compared with the low density cell values (100%). (B) ARL cells at about 20% confluency were further incubated for 3 days in the presence or absence of 20 µg/ml aphidicolin (aphid) or 6 mM hydroxyurea (HU) and stained with McAb 351. (C) ARL cells grown in complete medium were rinsed three times with serum-free medium, and further incubated for 75 hours in the presence (+) or absence (-) of 10% fetal calf serum. Cells were stained with McAb 351. Bar, 50 µm.](https://www.researchgate.net/profile/Akira-Inoue-8/publication/7868510/figure/fig5/AS:601712540782593@1520470857415/S1-fibers-are-unrelated-to-cell-proliferation-and-cellular-metabolic-activity-A-RNA_Q320.jpg)
Content uploaded by Akira Inoue
Author content
All content in this area was uploaded by Akira Inoue
Content may be subject to copyright.
Introduction
Vimentin intermediate filaments (VFs) are expressed in
tissues of mesenchymal origin. Although mice with
knockout mutations in vimentin genes develop and
reproduce without an obvious phenotype (Colucci-Guyon et
al., 1994), it has been shown that this mutation influences
cells and organs in various ways (Henrion et al., 1997; Terzi
et al., 1997; Colucci-Guyon et al., 1999). VFs afford
mechanical stability of the cells. For example, during cell
migration, subcellular structures need to be protected
from mechanical lesions. The importance of vimentin
cytoskeleton in cell motility has been demonstrated with
mouse fibroblasts (Eckes et al., 1998) and human mammary
epithelial cells (Gilles et al., 1999). Similarly, it has been
also shown to be important during mouse development with
neural crest (Cochard and Paulin, 1984), parietal endoderm
(Lane et al., 1983; Lehtonen et al., 1983) and mesenchymal
cells (Franke et al., 1982). In fact, VFs become prerequisite
not only in contraction and reorganization but also in cell
motility of connective tissues, and all of these are mandatory
events in wound healing (Eckes et al., 1998) (reviewed by
Gailit and Clark, 1994).
S1 proteins A-D are ubiquitously found in animal cells
(Inoue et al., 1983; Emura et al., 1992; and unpublished
results), and extracted as a group of proteins at pH 4.9 from
the nuclei treated with either RNase A or DNase I (Inoue et
al., 1983; Higashi et al., 1984; Inoue et al., 1986). They are
resolved by SDS-PAGE, each as doublets: A1 (an apparent
molecular mass of 74.5 kDa), A2 (69.5 kDa); B1 (47.4 kDa),
B2 (46.5 kDa); C1 (43.9 kDa), C2 (42.8 kDa); D1 (40.8 kDa)
and D2 (39.4 kDa). S1 proteins B-D are multifunctional
hnRNP proteins (Inoue et al., 2001; Inoue et al., 2003).
Among them, S1 proteins D2 and C2 are identified as CArG-
box binding factor-A (CBF-A) (Kamada and Miwa, 1992)
and its splicing isoform (GenBank accession number
AJ238854) respectively, acting as positive (Bemark et al.,
1998; Mikheev et al., 2000) or negative (Kamada and Miwa,
1992) transcriptional regulators. A hybridoma producing
monoclonal antibody McAb 351, which is highly specific for
S1 proteins C2 and D2, has been isolated. Cell staining with
this antibody showed that they occur not only in the nucleus
but also in the cytoplasm often in association with VFs in
cultured cells (Tsugawa et al., 1997). In this study, we
verified the association of S1 proteins with VFs, and
investigated its biological significance. We demonstrate that
the VF association of S1 proteins is closely related to cell
motility.
2303
S1 proteins C2 and D2 are multifunctional hnRNP
proteins acting as transcriptional regulators in the
nucleus. Immunofluorescence staining of various cells
in culture revealed that S1 proteins also occur in
the cytoplasm, often in association with vimentin
intermediate filaments (VFs). Here, we verified the
association of S1 proteins with vimentin using vimentin-
deficient cells, crosslinking and immunoprecipitation,
and further investigated the biological significance of this
association. S1 proteins on VFs, referred to here as S1
fibers, were lost in highly confluent cells, where cell
proliferation and cellular metabolic activity greatly
decreased owing to cell density-dependent arrest.
However, the disappearance of S1 fibers was not related
to these reduced activities, but to inhibited cell migration.
Although undetected in cells of non-migratory tissues as
well as in confluent cultured cells, S1 fibers were found
in all migratory cells examined, such as cultured cells in
scratch/wound experiments, blood neutrophils and
monocytes, and fibroblasts engaging in tissue healing. In
addition, S1 fibers reappeared even in confluent cells
when VFs were induced to reorganize with okadaic acid.
We propose that S1 proteins occur in association with
VFs in migratory cells. Possible participation of S1
proteins in the formation/reorganization of VFs is
discussed.
Key words: CBF-A, Neutrophil, Fibroblast, Cell migration, Cell
motility, Ulcer
Summary
Association of hnRNP S1 proteins with vimentin
intermediate filaments in migrating cells
Akira Inoue1,*, Takanori Watanabe2, Kazunari Tominaga3, Katsuji Tsugawa4, Koji Nishio5,
Kenichi P. Takahashi6and Kenji Kaneda2
1Molecular Mechanisms of Biological Regulation, 2Department of Anatomy and 3Department of Internal Medicine, Osaka City University Graduate
School of Medicine, 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan
4Department of Environmental Sciences, Faculty of Science, Osaka Women’s University, 2-1 Daisen-cho, Sakai 590-0035, Japan
5Department of Anatomy and Cell Biology, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan
6Department of Developmental Anatomy, Osaka Prefecture College of Health Sciences, 3-7-30 Habikino, Habikino, Osaka 583-8555, Japan
*Author for correspondence (e-mail: ainoue@med.osaka-cu.ac.jp)
Accepted 21 February 2005
Journal of Cell Science 118, 2303-2311 Published by The Company of Biologists 2005
doi:10.1242/jcs.02345
Research Article
Journal of Cell Science
2304
Materials and Methods
Cell culture
ARL rat liver epithelial cells (ARL J301-3, Japanese Health Science
Foundation) were grown in William’s E medium supplemented with
10% fetal calf serum and HaK Syrian hamster kidney cells (CCL15,
ATCC) in Eagle’s minimum essential medium supplemented with
10% fetal calf serum. Vimentin-deficient tTA-1 mouse fibroblasts and
vimentin (+/+) MFT-6 mouse fibroblasts (Holwell et al., 1997) were
cultured in Dulbecco’s modified Eagle medium containing 5% calf
serum. Hygromycin (200 µg) was added to the culture of tTA-1 cells.
Inhibitors used in cell cultures were aphidicolin (20 µg/ml),
hydroxyurea (6 mM), actinomycin D (5.0 µg/ml), okadaic acid (50
nM), genistein (10 µM) or staurosporine (0.2 µM).
Preparation of white blood cells and chemotaxis
White blood cells were isolated by centrifugation of heparinized
peripheral blood from an adult Wistar rat (250 g) at 400 gfor 10
minutes at 20°C. The white blood cell layer was collected with a
pipette, and the cell suspension diluted 5-15 times with RPMI and
incubated at 37°C for 20 minutes between a glass slide and a sheet of
Saran wrap. In chemotaxis experiments, the diluted white blood
cell suspension (10 µl) and 10–7 M N-formylmethionyl-leucyl-
phenylalanine (fMLP) (10 µl) in RPMI were placed at a 5 mm
distance, and covered with a sheet of Saran wrap. After incubation,
the Saran wrap and excess fluid were removed, and the specimens
were air-dried and stained as described below.
Indirect immunofluorescence staining and antibodies
Cells on slides or coverslips were fixed with 4% paraformaldehyde
for 10 minutes, permeabilized with 0.2% Triton X-100 for 5 minutes,
and incubated first with 10% fetal calf serum for 20 minutes, then with
primary and secondary antibodies for 1 hour each. Each solution was
made in phosphate-buffered saline (PBS) and between each step, cells
were washed three or four times with PBS. Primary antibodies were
mouse monoclonal antibody against vimentin (V9, 1:150 dilution, IT),
rabbit (Cappel or Medak) or goat (RDI or Chemicon) polyclonal
antibody against vimentin (1:50-200 dilution), and mouse monoclonal
antibody McAb 351 against rat S1 proteins C2 and D2 (1:100 dilution)
(Tsugawa et al., 1997). Secondary antibodies (1:30-200 dilution) were
FITC-conjugated goat or rabbit anti-mouse IgG antibodies (Biosource
International or Zymed), rhodamine-conjugated goat anti-rabbit IgG
(AP156R, 1:50 dilution, Chemicon International) or TR-conjugated
donkey anti-goat IgG (Santa Cruz Biotechnology). Images were
observed under a fluorescence microscope with a mercury light source
(Olympus, model Bx50), or a confocal laser-scanning fluorescence
microscope (Carl Zeiss, model LSM510).
Preparation of proteins and immunoblotting
Reference S1 proteins were prepared from rat liver nuclei as described
(Inoue et al., 1983; Tsugawa et al., 1997). ARL cells were solubilized
in a standard SDS sample buffer (SDS and 2-mercaptoethanol were
at 2% and 1% respectively) and the lysates sonicated and heated at
95°C for 5 minutes. Immunoblots of 9.5% SDS-PAGE gels were
probed with McAb 351 (1:1000 dilution) and a horseradish
peroxidase-conjugated goat anti-mouse IgG (1:1000 dilution, Cappel)
(Inoue et al., 2001). Bands were visualized on X-ray film (Hyperfilm
ECL, Amersham) with an ECL detection kit (RPN 2106, Amersham)
and their intensities were measured using an image analyzer (BioRad,
Multi-Analyst). For normalization, DNA in the samples was isolated
by a standard method using proteinase K and RNase A dissolved in
H2O and determined by measuring absorbance at 260 nm.
Crosslinking
ARL cells grown in 10 cm dishes were incubated at room temperature
for 45 minutes with 3 mM dimethyl 3,3′-dithiobispropionimidate
(DTBP, Pierce) in PBS (pH 7.4), containing 0.4 mM PMSF and 20
µg/ml leupeptin. The cells were dissolved in an SDS sample buffer
without mercaptoethanol, sonicated and heat-treated. After SDS-
PAGE on a 6.5% gel, the sample lane was excised with the aid of pre-
stained marker proteins (Broad range, BioLabs) run on adjacent lanes.
The gel strip was equilibrated in 125 mM Tris-HCl, pH 6.8, incubated
with 40 mM DTT in 125 mM Tris-HCl, pH 6.8, 0.1% SDS and 10%
glycerol for 5 hours at room temperature, and subjected to the second
SDS-PAGE (9% gel with a flat 3% stacking gel). Proteins were
immunoblotted with McAb 351, and reprobed with a polyclonal rabbit
anti-vimentin antibody (DBS, Fremont, CA).
Immunoprecipitation
ARL cells at about 40% confluence were collected with PBS from six
10-cm culture plates, lysed for 1 hour with occasional mixing with
750 µl lysis buffer (0.15 M NaCl, 1% NP-40, and 50 mM Tris-HCl,
pH 8.0) containing leupeptin (10 µg/ml), aprotinin (1000 KIU/ml) and
0.2 mM PMSF, and centrifuged at 10,000 g. All steps were done at
4°C and the washing with lysis buffer. The supernatant extract was
pre-cleared by incubation with a 75 µl packed volume of protein A-
Sepharose (4 Fast-Flow, Pharmacia), which was previously saturated
with non-specific antibodies (100 µl fetal calf serum) and washed.
Specific antibody (1 µl, ascites fluid) was incubated for 10 minutes
with washed fresh protein A-Sepharose (20 µl packed volume). Then,
the pre-cleared cell extract (100 µl) was added and the incubation
continued for 30 minutes with occasional mixing. The beads were
washed, and bound proteins eluted with SDS sample buffer (60 µl) at
95°C for 5 minutes and collected by centrifugation. Control mouse
monoclonal antibodies (ascites fluids) were I-65 against a toxoplasma
membrane protein and 3F7 against RBP-MS (RNA binding protein
gene with multiple splicing) protein (Shimamoto et al., 1996).
Determination of cellular RNA and DNA synthesis activities
ARL cells cultured in 5 cm dishes were incubated in fresh medium
with [3H]uridine or [3H]thymidine (NEN Life Science Products) for
45 minutes. After rinsing with PBS, cells were solubilized in cold 0.4
N NaOH (3 ml) and sonicated. Half portions were mixed with 3 ml
of 10% trichloroacetic acid (TCA), placed on ice for 5 minutes and
filtered on glass filters (GF/F, Whatman) under suction. The filters
were washed three times with 10% TCA and twice with ethanol and
radioactivity was counted. The remaining half portions were
neutralized with 0.3 N HCl and 0.1 M Tris-HCl, pH 7.5 and digested
with 60 µg/ml RNase A for 4 hours at 37°C. DNA was collected by
ethanol precipitation, dissolved in 0.1 N NaOH and determined by
reading absorbance at 260 nm.
Ulcer formation
Gastric ulcers were produced in 8-week-old Wistar male rats (Japan
SLC, Hamamatsu, Japan) as described (Tominaga et al., 1997). In
brief, rats were fasted for 12 hours and subjected to laparotomy under
ether anesthesia. A plastic mold (6 mm in diameter) was tightly placed
on the anterior serosal surface of the antral-oxyntic border of the
stomach. Acetic acid (60 µl) was poured into the mold and allowed
to remain on the gastric wall for 60 seconds. After the solution was
removed, the surface of the treated area was wiped with absorbent
paper and the abdomen was closed. Control rats received the same
laparotomy, without treatment with acetic acid (sham-operated rats).
The Animal Care Committee of Osaka City University approved the
experimental procedure.
Journal of Cell Science 118 (10)
Journal of Cell Science
2305
hnRNP S1 proteins on vimentin filaments
Immunohistochemistry
Under ether anesthesia, ulcerated gastric tissues were excised 5 days
after operation and fixed with 10 mM metaperiodate, 75 mM lysine
and 2% paraformaldehyde in 40 mM PBS (pH 7.4) for 5 hours.
Cryosections were incubated first with non-immunized rabbit serum
for 30 minutes, then with a diluted specific antibody for 4 hours at
4°C and 0.3% H2O2/methanol for 15 minutes. They were washed in
PBS and stained using a LSAB streptoavidin-biotin-peroxidase kit
(DAKO, Kyoto, Japan). Incubation with biotinylated rabbit anti-
mouse IgG was for 15 minutes and with peroxidase-labeled
streptoavidin for 15 minutes. The sections were finally developed in
0.03% 3,3′-diaminobenzidine (DAB) with 0.005% H2O2and sodium
azide (650 µg/ml). Nuclei were counterstained with Methyl Green.
Images were obtained using a bright-field microscope (Nikon,
Microphoto-FXA).
Northern blotting
cDNAs of vimentin and glyceraldehyde 3-phosphate dehydrogenase
(GAPDH) (Fort et al., 1985) were labeled with [α-32P]dCTP (NEN
Life Sciences) by random primer extension using a nick-translation
kit (NEP-103L, NEN). Isolation of total RNA by the guanidium
isothiocyanate/phenol/chloroform method, electrophoresis on a 1%
agarose gel, blotting to nylon membrane (Gene Screen Plus, NEN)
and hybridization in a solution containing 50% formamide were
performed essentially as described (Sambrook et al., 1986). The
washed membrane was exposed to Kodak XAR-5 film between two
intensifying screens at –70°C. For reprobing, the membrane was
stripped of the probe in 0.1SSC solution containing 1% SDS at
100°C and rehybridized. Autoradiographic band intensity was
determined on an optical scanner (Seiko, Epson GT-8000).
GAPDH mRNA was used as an internal reference.
Results
Localization of S1 proteins on VFs
Indirect immunofluorescence staining with antibody
McAb 351, specific for S1 proteins C2 and D2, shows
localization of these proteins on vimentin filaments
(VFs) as well as in the nuclei of VF-containing cells
such as HaK cells and HeLa cells (Tsugawa et al.,
1997). S1 proteins on VFs (S1 fibers) were similarly
observed in ARL rat liver epithelial cells (Fig. 1, A1).
Localization of S1 proteins on VFs was determined by
double staining with McAb 351 (Fig. 1, A2) and a
polyclonal anti-vimentin antibody (Fig. 1, A3).
Association of S1 proteins with VFs
To confirm the association of S1 proteins C2 and D2
with VFs, we first examined S1 fibers in vimentin-
deficient mouse fibroblasts. As expected, S1 fibers were
absent in these cells, in contrast their presence in wild-
type cells (Fig. 2).
The association of S1 proteins with VFs was also
examined by a crosslinking reaction. Proteins in ARL
cells were first reacted with DTBS, a cleavable
bifunctional crosslinker, and subjected to the first SDS-
PAGE without prior reduction of disulfide bonds. Then,
crosslinked proteins in an excised sample lane were
separated in gel with dithiothreitol, which cleaves
DTBS in the middle of the molecule. The second
dimensional PAGE showed that proteins not
crosslinked in the crosslinking reaction run forming a curved
diagonal (Fig. 3A). On the other hand, vimentin lay on a
horizontal line at its monomer size (Fig. 3C), indicating that it
was crosslinked to larger products of progressively increasing
sizes. The results are exactly those expected for its occurrence
as polymer filaments. S1 proteins C2 and D2 were also
crosslinked in a similar manner (Fig. 3B). Their crosslinking
patterns, similar to that of vimentin, support the fact that S1
proteins are associated with polymerized structures. However,
the non-crosslinked population was significantly high with S1
proteins, probably because they occur in the nucleus as well as
in the cytoplasm.
To further examine the S1 protein association with vimentin,
ARL cell extract was immunoprecipitated with McAb 351. As
expected, the antibody against S1 proteins precipitated
vimentin (Fig. 4). No vimentin was precipitated without the
antibody or with control antibodies irrelevant to S1 proteins.
As McAb 351 does not recognize vimentin (Tsugawa et al.,
1997), the results verified that S1 proteins form complexes with
vimentin.
Relationship of S1 fibers and cell confluency
ARL cells in culture can be maintained in a confluent
monolayer for up to 3 weeks without a change of the medium.
In such confluent ARL cells, S1 fibers were found to disappear
and S1 proteins were detected only in the nuclei (Fig. 1, B1).
Fig. 1. S1 fiber formation in ARL cells at various states of confluency. ARL
cells were cultured at a low cell density for 3 days (A) or in a confluent state
for 3 weeks (B), and stained with anti-S1 protein antibody McAb 351 (A1,
B1), monoclonal anti-vimentin antibody V9 (B2), or double-stained with
McAb 351 (A2) and a polyclonal anti-vimentin antibody (A3). At low cell
densities, S1 proteins are present not only in the nucleus, but also in the
cytoplasm in association with VFs. (C) ARL cells were cultured for 3 days to
40% confluency (lane 1) and for 2 weeks at a full confluency (lane 2). Total
proteins equivalent to 0.2 µg cellular DNA were analyzed by immunoblotting
using McAb 351. Lane 3, reference S1 proteins isolated from rat liver nuclei.
Bar, 25 µm.
Journal of Cell Science
2306
The results suggested dependency of S1 fibers on cell
confluency. On the other hand, VFs remained, although their
networks had dwindled according to the decrease in cell size
(Fig. 1, B2). Their nuclei had also reduced in size, and were
stained less intensely with McAb 351 when compared with
growing cells at lower cell densities (Fig. 1, B1 compared to
A1). In fact, after 2 weeks in such a confluent state, S1 proteins
C2 and D2 decreased to 57 and 63% of the levels observed in
growing cells (Fig. 1C, lane 1 compared to lane 2).
The correlation of S1 fibers with cell confluency was
confirmed. First, a small volume of trypsinized cell suspension
was placed at a corner of a chamber slide. After the cells had
attached to the substratum, enough medium was added and the
cell culture was continued for 2.5 weeks at a
tilted position (tilted culture experiment).
Although hardly detected inside the high cell-
density area, prominent S1 fibers were
observed in the cells on the margin and in
those moving ahead of the dense area (not
shown). These cells had large nuclei with a
strong S1 protein staining intensity as well
as well-developed VFs. Likewise, ARL cells
plated at ~10% confluency and cultured for
3 days exhibited prominent S1 fibers in
contrast to the cells plated at full confluency,
which showed few S1 fibers (not shown).
These results thus confirmed that S1 fibers
become absent from confluent cells or cells
at high cell densities.
S1 fiber formation is dissociated from
cell proliferation and cellular metabolic
activity
Biosynthetic activities of ARL cells at 50-
60% confluency and those cultured for 1 and
3 weeks at full confluency were assayed by
incubation of the cells with [3H]thymidine or
[3H]uridine (Fig. 5A). The RNA synthesis
activity of the 3-week confluent cells was
reduced by 70% compared to that measured in the low-density
cells. DNA synthesis was also drastically reduced in the 3-
week confluent cells, with the activity reduced by more than
99%. These results thus indicate that under these conditions,
cellular metabolic activity greatly decreased, and cell
proliferation ceased almost completely.
In confluent states, ARL cells are restricted from migration,
and they are in the resting G0 state owing to cell density-
dependent arrest. Therefore, the disappearance of S1 fibers
in confluent cells may have arisen from the suppressed cell
proliferation, lowered metabolic activity of G0 state or
inhibited cell movement. These possibilities were examined.
First, ARL cells doubling every 18.5 hours at a low cell density
Journal of Cell Science 118 (10)
Fig. 2. Absence of S1 fibers in vimentin-deficient cells. Vimentin-deficient tTA-1 mouse
fibroblasts (–/–), and control vimentin (+/+) MFT-6 fibroblasts were double-stained with
anti-S1 protein antibody McAb 351 (left) and an anti-vimentin antibody (middle).
Images were obtained by confocal microscopy. Merged images are shown on the right.
Fig. 3. Protein crosslinking experiment. Proteins in ARL cells were crosslinked with DTBP and resolved, without prior reduction of disulfide
bonds by SDS-PAGE (1D). The sample-lane was excised, incubated with DTT to cleave DTBP and subjected to the second dimension SDS-
PAGE (2D). Blotted proteins were stained with Coomassie Brilliant Blue (A), destained and probed with McAb 351 (B) and finally reprobed
with an anti-vimentin antibody (C). Non-crosslinked proteins form a curved diagonal in 2D (A); S1 proteins C2 and D2 (B) and vimentin (C)
are seen on horizontal lines at the monomer sizes: they were crosslinked to the products of various sizes. Their non-crosslinked forms are seen
on the left as strongly stained spots.
Journal of Cell Science
2307
hnRNP S1 proteins on vimentin filaments
were incubated for 3 days in the presence of high
concentrations of aphidicolin (20 µg/ml) or hydroxyurea (6
mM). These chemicals block DNA synthesis by inhibiting
DNA polymerase αand deoxyribonucleotide-producing
ribonucleotide reductases. Neither drug exerted appreciable
effects on the S1 fiber formation (Fig. 5B), which indicated that
S1 fiber formation is dissociated from cell proliferation.
To examine the effect of lowered metabolic activity in the
G0 state, ARL cells growing at a low cell density were brought
to G0 by serum starvation (Zetterberg and Skold, 1969;
Zetterberg and Larsson, 1991). S1 fibers, observed 3 days later,
appeared similar in the absence or presence of serum (Fig. 5C).
The results indicated that S1 fibers are not correlated with
resting G0 state or reduced cellular metabolic activity.
S1 fibers and cell motility
To examine the relationship of S1 fibers with cell migration,
scratch-wound experiments were performed on confluent cells,
where cells were allowed to migrate towards the cleared space.
A central area of a 3-week monolayer sheet of ARL cells was
removed by scratching and observed after 30 hours. The ARL
cells exposed to open substratum detached from the high cell-
density area and migrated into the cleared space exhibiting
well-developed S1 fibers (Fig. 6A,B). Essentially the same
results were obtained with HaK cells (a hamster kidney cell
line): S1 fibers appeared in the cells on the margin of dense
regions as early as 8 hours after scratching (data not shown).
Furthermore, even under the conditions where aphidicolin (20
µg/ml) was added to scratched monolayers, cell migration and
reappearance of S1 fibers were still observed (Fig. 6C), thereby
confirming that S1 fibers are unrelated to cell proliferation.
With hydroxyurea, the same conclusion was obtained (not
shown).
In addition, when 3-week confluent ARL cells were
trypsinized and replated at a lower cell-density, S1 fibers
appeared within 10 hours. Similar results were obtained with
confluent HaK cells. All of these results indicated that the S1
fibers that had disappeared in confluent monolayer appeared
again in a reversible manner when the cells were placed under
non-restricted conditions for cell movement.
Further evidence for the correlation of S1 fibers and cell
migration
Different types of cells have different migratory
potentials. Accordingly, we analyzed S1 proteins by
immunohistochemistry in various cells of the rat tissues. It was
evident that S1 proteins were commonly detected only in the
nuclei in all non-migratory cells examined (K.T. and N. Ikeda,
unpublished data). The results also confirmed that different
metabolic activities between the various cell types had no
correlation with the presence of S1 fibers. In the cerebellum
(Fig. 7A) for example, the large nuclei of Purkinje cells were
strongly stained with McAb 351 in contrast to the granule cell
nuclei, probably because of high RNA synthesis activity in the
Fig. 4. Immunoprecipitation of vimentin with McAb 351. ARL cell
extract was immunoprecipitated with control Tris-buffered saline
(lane 1), control antibody I-65 against toxoplasma membrane (lane
2), control 3F7 against an RNA-binding protein RBP-MS (lane 3) or
McAb 351 against S1 proteins C2 and D2 (lane 4) and the
precipitated proteins were immunoblotted for vimentin. Lane 5,
reference total proteins from ARL cells. Positions of molecular
weight protein markers (M) are indicated on the right-hand side of
the blot.
Fig. 5. S1 fibers are unrelated to cell proliferation and cellular
metabolic activity. (A) RNA and DNA synthesis activities. ARL cells
cultured to a 50-60% confluency (low density cells) or for 1 and 3
weeks at full confluency were incubated for 45 minutes in a fresh
medium containing [3H]uridine or [3H]thymidine. Incorporated
radioactivities were determined with respect to DNA, and compared
with the low density cell values (100%). (B) ARL cells at about 20%
confluency were further incubated for 3 days in the presence or
absence of 20 µg/ml aphidicolin (aphid) or 6 mM hydroxyurea (HU)
and stained with McAb 351. (C) ARL cells grown in complete
medium were rinsed three times with serum-free medium, and
further incubated for 75 hours in the presence (+) or absence (–) of
10% fetal calf serum. Cells were stained with McAb 351. Bar, 50
µm.
Journal of Cell Science
2308
former cells. The S1 proteins were confined in the nuclei in
both Purkinje cells and granule cells (Fig. 7Ab) and did not
merge with cytoplasmic VFs in the VF-containing Purkinje
cells (Fig. 7Aa).
On the other hand, among the various cell types examined,
only migratory cells revealed S1 proteins in the cytoplasm, as
expected. White blood cells were isolated from the blood of
the rat, and incubated for 20 minutes between a slide glass
and Saran wrap. Monocytes with kidney-shaped nuclei, which
represent typical migratory cells (Alberts et al., 2002),
contained many S1 fibers in the cytoplasm (Fig. 7B). The
neutrophil is another typical migratory cell that has
characteristic polymorphonuclear structures and these cells
similarly exhibited large amounts of S1 fibers. When
stimulated with the chemoattractant, N-formylmethionyl-
leucyl-phenylalanine (fMLP), neutrophils showed distinct S1
fibers (Fig. 7B, panels 2-4); without fMLP, S1 fibers were less
distinct, even though cytoplasmic S1 proteins were as abundant
as in stimulated cells (Fig. 7, panel 5).
To examine other migratory cells, we chose fibroblasts
engaging in tissue remodeling (Alberts et al., 2002). In the
healing stage of a gastric ulcer, fibroblasts on the edge of the
ulcer migrate actively towards the center of ulcerated area,
forming granulation tissue (Tominaga et al., 1997). The
spindle-shaped fibroblasts exhibited S1 proteins in the
cytoplasm as well as in the nuclei (Fig. 8A,C). In contrast, the
fibroblasts in normal gastric tissues showed S1 proteins only
in the nuclei (Fig. 8B,D). VFs were present in the cytoplasm
of both the ulcerated (Fig. 8E,G) and intact tissue cells (Fig.
8F,H). However, the vimentin staining intensity was several
fold stronger in the ulcerated gastric tissues. In accordance with
this, the expression of vimentin mRNA measured by Northern
blotting on day 5 increased 5.6-fold in the ulcerated tissues
compared with expression levels in sham-operated tissue (Fig.
8I, mean value of six determinations; P<0.05).
Journal of Cell Science 118 (10)
Fig. 6. Scratch-wound
experiments. The center areas
of 3-week confluent
monolayers of ARL cells grown
on coverslips were removed by
scratching with a pipette tip,
and the incubation was
continued in fresh medium for
30 hours. Cells were stained
with McAb 351 (A), or double-
stained with McAb 351 (B1)
and an anti-vimentin antibody
(B2). (C) After scratching, ARL
cells were further incubated in
the presence or absence of
aphidicolin (20 µg/ml) for 30
hours. Cells were stained with
McAb 351. In all panels, right-
hand areas correspond to the
marginal regions of cleared
areas made by scratching. Bar,
50 µm.
Fig. 7. S1 fibers in animal tissues. (A) A
section of the cerebellum from an adult rat
was double-stained with an anti-vimentin
antibody (a, red) and McAb 351 (b, green).
Purkinje cells are seen on the diagonal, and
the granule cells on the right-hand side. Bar,
50 µm. (B) White blood cells from an adult
rat were incubated at 37°C for 20 minutes
between a slide glass and a sheet of Saran
wrap in the presence (2-4) or absence (1, 5)
of fMLP. The cells were stained with McAb
351 (a), DAPI (1b-3b), and an anti-vimentin
antibody (4b, 5b). Panels 1, monocytes;
panels 2-5, neutrophils. Bar, 20 µm.
Journal of Cell Science
2309
hnRNP S1 proteins on vimentin filaments
Possible involvement of S1 fibers in the
formation/reorganization of VFs
When RNA synthesis of ARL cells was inhibited with
actinomycin D (5 µg/ml), well-developed S1 fibers diminished
relatively rapidly in intensity and fiber length. The decrease
continued over 9 hours, with a reciprocal increase in the
nuclear S1 protein staining intensity (Fig. 9A). The results
suggest the dependence of S1 fibers on RNA synthesis and
relocalization of S1 proteins from VFs to the nucleus upon
inhibition of RNA synthesis.
Effects of S1 protein phosphorylation on S1 fibers
were examined. When ARL cells after 2 weeks of
confluent culture were treated with staurosporine and
genistein, inhibitors of PKC and tyrosine kinases
respectively, no effects on the S1 fibers were observed.
In contrast, the phosphatase inhibitor okadaic acid,
which induces reorganization of VFs through hyper-
phosphorylation of vimentin molecules (Inagaki et al.,
1987; Lee et al., 1992), exerted a strong effect: in
accord with the alteration of VF networks, S1 fibers
appeared as thick bundles in the confluent cells (Fig.
9B).
Discussion
We have shown previously by cell staining using McAb
351 that in addition to their presence in the nucleus,
hnRNP S1 proteins C2 and D2 are found in the cytoplasm
in association with VFs and not with microtubules,
microfilaments, cytokeratin filaments or desmin filaments
(Tsugawa et al., 1997). In the present study, we verified
the VF association of S1 proteins by three different
approaches. First, vimentin-deficient fibroblasts were
shown not to form S1 fibers, as expected. Second,
crosslinking experiments showed that S1 proteins were
indeed associated with polymer filament structures. Third,
immunoprecipitation with McAb 351 coprecipitated
vimentin.
In this study, the biological significance of the association
of the S1 proteins with VFs was also studied. Prominent S1
fibers formed in ARL cells growing at low cell densities, and
when the cells became confluent and entered into cell density-
dependent arrest, S1 fibers disappeared. The disappearance of
S1 fibers was not related to inhibited cell proliferation nor to
a reduced cellular metabolic activity, but to inhibited cell
movement. This conclusion is well supported by the following
findings. Inhibition of cell proliferation with specific
inhibitors, aphidicolin and hydroxyurea, exerted no
Fig. 8. Appearance of S1 proteins in the cytoplasm of
migratory fibroblasts. (A-H) Fibroblasts in the healing tissue
of gastric ulcer (left panels) and those in the sham-operated
intact tissue of the stomach (right panels) were stained with
McAb 351 (A-D) and with an anti-vimentin antibody (E-H).
Nuclei (arrows) were counterstained with Methyl Green.
Spindle-shaped fibroblasts were cut along the short (A,B,E,F)
and long axes (C,D,G,H). In addition to the nucleus, the
cytoplasm was stained with McAb 351 in ulcerated healing
tissue (A,C compare with B,D). Vimentin was present in the
cytoplasm of both ulcerated (E,G) and intact (F,H) tissues,
with more vimentin in the former. (I) Vimentin mRNA levels
during gastric ulcer healing. Total RNA from gastric tissues on
day 5 after ulcer formation was analyzed by Northern blotting
for vimentin and GAPDH. One representative autoradiogram
of six determinations is shown. Bar, 12.5 µm.
Fig. 9. Effects of actinomycin D and okadaic acid on S1 fiber
formation. (A) ARL cells grown to 65% confluency were
incubated in the presence of actinomycin D (5 µg/ml) for up
to 9 hours and immunostained with McAb 351. (B) ARL cells
cultured for 2 weeks at full confluency were treated with 50
nM okadaic acid (+OA) for 3 hours, and immunostained with
McAb 351. Arrowheads indicate S1 fibers. Bar, 25 µm.
Journal of Cell Science
2310
appreciable effects on S1 fibers in ARL cells. Also, the
neutrophils (the terminally differentiated non-dividing cells),
and the ARL cells that were brought to G0 state by serum
starvation still exhibited prominent S1 fibers. In ARL cells, S1
fibers persisted under reduced cellular metabolic activities
brought about by the serum starvation. Also, S1 fibers were
commonly absent from most of the tissue cells, irrespective of
their different levels of metabolic activity. These results
suggested that S1 fibers are also unrelated to the metabolic
activity of the cell. In confluent monolayers where cell
migration is largely reduced, S1 fibers disappeared. However,
S1 fibers appeared again when the cells were allowed to
migrate into vacant space by scratching of confluent cell
monolayers. Similarly, in the tilted culture experiment, the
cells moving ahead the edge of the high cell density
monolayer region had prominent S1 fibers, but the cells inside
the dense region did not. S1 fibers were not detected in any of
the non-migratory cells of various tissues examined. In
contrast, neutrophils and monocytes, the migratory cells in the
blood, exhibited large amounts of S1 fibers. Similarly, the
fibroblasts in wound/healing, another representative migratory
cells, contained significant amounts of S1 proteins in the
cytoplasm, and contrasted to those in sham-operated tissues,
which contained no detectable cytoplasmic S1 proteins. From
these results, we concluded that S1 proteins become
associated with VFs in migrating cells. Although no direct
measurement of cell migration was made in the present study,
the epithelial ARL cells that detached from high cell density
regions in scratch-wound experiments, the cells moving ahead
of dense areas in the tilted culture experiment, the neutrophils
treated with chemoattractant fMLP and fibroblasts in tissue-
remodeling (Alberts et al., 2002) as well as ARL cells at low
cell-densities are all thought to be in migratory motion.
How well cultured cells develop S1 fibers depends on cell
density. It is known that many cultured epithelial cells do move
around, relative to their neighbors, even in confluent cells. This
may be the reason why it took ARL cells more than a week to
reach an undetectable level of S1 fibers. Even after reaching a
confluent state, the cells increase in number and become
smaller and denser. Thus time is required to form more packed
monolayer, producing greater restraints on cell movement and
leading to the disappearance of S1 fibers.
It is known that VFs are prerequisite for cell motility (Gailit
and Clark, 1994; Eckes et al., 1998; Gilles et al., 1999). The
function or the role of S1 proteins on VFs may be understood
in terms of formation/reorganization of VFs. First, inhibition
of RNA synthesis brought about relatively rapid reduction in
S1 fibers. Second, appearance of S1 proteins in the cytoplasm
seemed coupled with increased vimentin synthesis as shown
with augmented vimentin mRNA and protein levels in tissue-
healing fibroblasts in the gastric ulcer. Similarly, whenever S1
fibers were formed in ARL cells, VFs and their networks
became more prominent. From these findings, we propose that
S1 fibers may be involved in the synthesis of vimentin proteins.
Furthermore, in ARL cells that were allowed to migrate by
scratching of the confluent cell-monolayer, prominent S1 fibers
appeared and VF networks were apparently reorganized in
densely orienting, seemingly migrating directions (Fig. 6).
Also, S1 fibers reappeared even in confluent cells when VFs
underwent reorganization by treatment with okadaic acid
(Inagaki et al., 1987). Hence S1 proteins may be involved in
the formation/reorganization of VFs, which are needed for cell
migration.
S1 proteins C2 and D2 act as hnRNA binding proteins in the
nucleus (Inoue et al., 2001), and as presented in this study,
appear to be associated with VFs in the cytoplasm. It is
interesting to speculate whether the S1 proteins participate in
the localization of a particular set of mRNAs on VFs, thereby
facilitating the site-oriented production of proteins required
for formation or reorganization of VFs. However, these
possibilities await clarification by further study. In conclusion,
we found that S1 proteins localize on VFs and propose that
they may be involved in cell migration through participation in
the formation/reorganization of VFs.
We are grateful to Robert M. Evans, University of Colorado Health
Sciences Center, Denver, for providing us with the cell lines of
vimentin-deficient and control vimentin (+/+) mouse fibroblasts. We
thank also Isao Kimata, Osaka City University, Graduate School of
Medicine, Osaka for antibody I-65 and Akira Shimamoto and Saori
Kitao, Agene Research Institute, Kanagawa for antibody 3F7.
References
Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K. and Walter, P.
(2002). Molecular Biology of the Cell, 4th edn. New York: Garland
Science.
Bemark, M., Olsson, H., Heinegard, D. and Leanderson, T. (1998).
Purification and characterization of a protein binding to the SP6 kappa
promoter. A potential role for CArG-box binding factor-A in kappa
transcription. J. Biol. Chem. 273, 18881-18889.
Cochard, P. and Paulin, D. (1984). Initial expression of neurofilaments and
vimentin in the central and peripheral nervous system of the mouse embryo
in vivo. J. Neurosci. 4, 2080-2094.
Colucci-Guyon, E., Portier, M.-M., Dunia, I., Paulin, D., Pournin, S. and
Babinet, C. (1994). Mice lacking vimentin develop and reproduce without
an obvious phenotype. Cell 79, 679-694.
Colucci-Guyon, E., Gimenez, Y. R.-M., Maurice, T., Babinet, C. and
Privat, A. (1999). Cerebellar defect and impaired motor coordination in
mice lacking vimentin. Glia 25, 33-43.
Eckes, B., Dogic, D., Colucci-Guyon, E., Wang, N., Maniotis, A., Ingber,
D., Merckling, A., Langa, F., Aumailley, M., Delouvee, A. et al. (1998).
Impaired mechanical stability, migration and contractile capacity in
vimentin-deficient fibroblasts. J. Cell Sci. 111, 1897-1907.
Emura, T., Fujimoto, S., Takahashi, K. P., Morisawa, S. and Inoue, A.
(1992). Nuclear S1 proteins from the starfish Asterina pectinifera. Biochem.
Int. 26, 531-535.
Fort, P. H., Piechaczyk, M., Sabrouty, S. E. L., Dani, C. H., Jeanteur, P.
H. and Blanchard, J. M. (1985). Various rat adult tissues express only one
major mRNA species from the glyceraldehyde-3-phosphate-dehydrogenase
multigenic family. Nucleic Acids Res. 13, 1431-1442.
Franke, W. W., Grund, C., Kuhn, C., Jackson, B. W. and Illmensee, K.
(1982). Formation of cytoskeletal elements during mouse embryogenesis.
III. Primary mesenchymal cells and the first appearance of vimentin
filaments. Differentiation 23, 43-59.
Gailit, J. and Clark, R. F. A. (1994). Wound repair in the context of
extracellular matrix. Curr. Opin. Cell Biol. 6, 717-725.
Gilles, C., Polette, M., Zahm, J.-M., Tournier, J.-M., Volders, L., Foidart,
J.-M. and Birembaut, P. (1999). Vimentin contributes to human mammary
epithelial cell migration. J. Cell Sci. 112, 4615-4625.
Henrion, D., Terzi, F., Matrougui, K., Duriez, M., Boulanger, C. M.,
Colucci-Guyon, E., Babinet, C., Briand, P., Friedlander, G., Poitevin, P.
et al. (1997). Impaired flow-induced dilation in mesenteric resistance
arteries from mice lacking vimentin. J. Clin. Invest. 100, 2909-2914.
Higashi, Y., Uda, M., Yukioka, M., Morisawa, S. and Inoue, A. (1984).
Selective extraction of a set of non-histone proteins from rat liver nuclei.
Biochem. Int. 9, 697-704.
Holwell, T. A., Schweitzer, S. C. and Evans, R. M. (1997). Tetracycline
regulated expression of vimentin in fibroblasts derived from vimentin null
mice. J. Cell Sci. 110, 1947-1956.
Inagaki, M., Nishi, Y., Nishizawa, K., Matsuyama, M. and Sato, C. (1987).
Journal of Cell Science 118 (10)
Journal of Cell Science
2311
hnRNP S1 proteins on vimentin filaments
Site-specific phosphorylation induces disassembly of vimentin filaments in
vitro. Nature 328, 649-652.
Inoue, A., Higashi, Y., Hasuma, T., Morisawa, S. and Yukioka, M. (1983).
A set of non-histone proteins isolated from the nuclei of various rat tissues.
Eur. J. Biochem. 135, 61-68.
Inoue, A., Higashi, Y., Hasuma, T., Yukioka, M. and Morisawa, S. (1986).
Presence of a unique set of nuclear proteins at the sites sensitive to RNase
A as well as DNase I. J. Biochem. 100, 747-751.
Inoue, A., Omori, A., Ichinose, S., Takahashi, K. P., Kinoshita, Y. and
Mita, S. (2001). S1 proteins C2 and D2 are novel hnRNPs similar to the
transcriptional repressor, CArG box motif-binding factor A. Eur. J. Biochem.
268, 3654-3663.
Inoue, A., Arao, Y., Omori, A., Ichinose, S., Nishio, K., Yamamoto, N.,
Kinoshita, Y. and Mita, S. (2003). Identification of S1 proteins B2, C1,
and D1 as AUF1 isoforms and their major role as hnRNP proteins. Biochem.
J. 372, 775-785.
Kamada, S. and Miwa, T. (1992). A protein binding to CArG box motifs and
to single-stranded DNA functions as a transcriptional repressor. Gene 119,
229-236.
Lane, E. B., Hogan, L. M., Kurkinen, M. and Garrels, J. I. (1983).
Coexpression of vimentin and cytokeratin in parietal endoderm cells of early
mouse embryo. Nature 303, 701-704.
Lee, W.-C., Yu, J.-S., Yang, S.-D. and Lai, Y.-K. (1992). Reversible
hyperphosphorylation and reorganization of vimentin intermediate filaments
by okadaic acid in 9L rat brain tumor cells. J. Cell. Biochem. 49, 378-393.
Lehtonen, E., Lehto, V. P., Paasivuo, R. and Virtanen, I. (1983). Parietal
and visceral endoderm differ in their expression of intermediate filaments.
EMBO J. 2, 1023-1028.
Mikheev, A. M., Mikheev, S. A., Zhang, Y., Aebersold, R. and Zarbl,
H. (2000). CArG binding factor A (CBF-A) is involved in
transcriptional regulation of the rat Ha-ras promoter. Nucleic Acids Res.
28, 3762-3770.
Sambrook, J., Fritsch, E. F. and Maniatis, T. (1986). Molecular Cloning,
2nd edn. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory
Press.
Shimamoto, A., Kitao, S., Ichikawa, K., Suzuki, N., Yamabe, Y., Imamura,
O., Tokutake, Y., Satoh, M., Matsumoto, T., Kuromitsu, J. et al. (1996).
A unique human gene that spans over 230 kb in the human chromosome
8p11-12 and codes multiple family proteins sharing RNA-binding motifs.
Proc. Natl. Acad. Sci. USA 93, 10913-10917.
Terzi, F., Henrion, D., Colucci-Guyon, E., Federici, P., Babinet, C., Levy,
B. I., Briand, P. and Friedlander, G. (1997). Reduction of renal mass is
lethal in mice lacking vimentin. Role of endothelin-nitric oxide imbalance.
J. Clin. Invest. 100, 1520-1528.
Tominaga, K., Arakawa, T., Kim, S., Iwao, H. and Kobayashi, K. (1997).
Increased expression of transforming growth factor-β1during gastric ulcer
healing in rats. Dig. Dis. Sci. 42, 616-627.
Tsugawa, K., Takahashi, K. P., Watanabe, T., Mai, A., Fujimoto, D. and
Inoue, A. (1997). Association of hnRNP S1 proteins C2 and D2 with
vimentin intermediate filaments. Cell Struct. Funct. 22, 239-246.
Zetterberg, A. and Skold, O. (1969). The effect of serum starvation on DNA,
RNA and protein synthesis during interphase in L-cells. Exp. Cell Res. 57,
114-118.
Zetterberg, A. and Larsson, O. (1991). Coordination between cell growth
and cell cycle transit in animal cells. Cold Spring Harbor Symp. Quant. Biol.
56, 137-147.
Journal of Cell Science
