1593
Journal of Cell Science 107, 1593-1607 (1994)
Printed in Great Britain © The Company of Biologists Limited 1994
The presence or absence of a vimentin-type intermediate filament network
affects the shape of the nucleus in human SW-13 cells
Alfonso J. Sarria, Jonathan G. Lieber, Steven K. Nordeen and Robert M. Evans*
Department of Pathology, University of Colorado Health Sciences Center, Denver, Colorado 80262, USA
*Author for correspondence
SUMMARY
Human SW-13 cells express the intermediate filament
protein vimentin in a mosaic pattern (Hedberg, K. K. and
Chen, L. B. (1986). Exp. Cell Res. 163, 509-517). We have
isolated SW-13 clones that do (vim+) or do not (vim−) synthesize vimentin as analyzed using anti-intermediate
filament immunofluorescence, electron microscopy and
two-dimensional gel analysis of detergent-extracted preparations. Vimentin is the only cytoplasmic intermediate
filament protein present in the vim+ cells, and the vim− cells
do not contain any detectable cytoplasmic intermediate
filament system. The presence or absence of intermediate
filaments did not observably affect the distribution of mitochondria, endoplasmic reticulum, microtubules or actin
stress fibers when these structures were visualized by fluorescence microscopy. However, electron microscopy and
anti-lamin A/C immunofluorescence studies showed that
nuclear morphology in vim− cells was frequently characterized by large folds or invaginations, while vim+ cells had
a more regular or smooth nuclear shape. When vim− cells
were transfected with a mouse vimentin expression
plasmid, the synthesis of a mouse vimentin filament
network restored the smooth nuclear morphology characteristic of vim+ cells. Conversely, when vim+ cells were
transfected with a carboxy-terminally truncated mutant
vimentin, expression of the mutant protein disrupted the
organization of the endogenous vimentin filaments and
resulted in nuclei with a prominently invaginated morphology. These results indicated that in SW-13 cells the
vimentin filament system affects the shape of the nucleus.
INTRODUCTION
cellular role of cIFs has been an inability to relate these
filaments directly to specific cellular events. While there are a
wide variety of chemical agents that are known to affect cIF
organization, none appears to be useful as a filament-specific
drug (Klymkowsky, 1988). Microinjection of cells with antifilament antibodies has been shown to produce cIF collapse or
fragmentation specifically (Klymkowsky, 1981; Lin and
Feramisco, 1981; Gawlitta et al., 1981; Klymkowsky et al.,
1983; Summerhayes et al., 1983; Tolle et al., 1986). Examination of cells with antibody disrupted cIF systems has
uniformly shown that the injected cells are still motile, capable
of division, and without any obvious alteration in other cytoplasmic structures or cell-cell interactions (Klymkowsky et al.,
1989). This experimental approach has been limited by the
small number of cells that can be microinjected in a reasonable period of time, the transient nature of the filament
collapse, and the reality that although cIF organization may be
perturbed the filaments are still present in the injected cell.
The expression of cIF proteins with dominant negative
mutations has been shown to be another effective means of
specifically disrupting endogenous filament organization
(Albers and Fuchs, 1987, 1989; van den Heuvel et al., 1987;
Coulombe et al., 1990; Gill et al., 1990; Lu and Lane, 1990;
Wong and Cleveland, 1990; Chin et al., 1991). Studies of this
type have shown that expression of dominant negative mutant
Intermediate filaments (IFs) are a major component of the
cytoplasm. Although the protein subunit composition of these
filaments differs in a cell type-specific manner, they are a
prominent feature of the cytoskeleton of nearly all vertebrate
cells (Traub, 1985). The structural and biochemical properties
of IFs have been extensively examined and described (for
reviews see: Skalli and Goldman, 1991; Klymkowsky et al.,
1989). While recent studies have suggested that IFs are vital
to maintain the mechanical integrity of cells in tissues (see
Coulombe et al., 1991), relatively little is known about the
potential intracellular function of the IF network. One of the
most conspicuous features of the cytoplasmic intermediate
filament (cIF) network is a characteristically close association
with the nucleus in many cells (Franke, 1971; Goldman et al.,
1985, 1986). This has led to suggestions that cIFs are involved
in a nuclear anchoring (Lehto et al., 1978; Virtanen et al., 1979,
Ngai et al., 1987; Sangiorgi et al., 1990) or signaling function
(Goldman et al., 1986). However, there has been little direct
evidence to support an involvement of cIFs in nuclear structure
and the functional significance of an interaction between the
cytoplasmic cIF system and some component of the nucleus
remains unclear.
One of the factors that has hindered attempts to study the
Key words: vimentin, intermediate filament, nuclear lamin
1594 A. J. Sarria and others
keratins in transgenic mice results in the formation of
epidermis that is sensitive to mechanical trauma (Coulombe et
al., 1991; Vassar et al., 1991). In embryonal carcinoma cells,
dominant negative mutant keratin expression prevents the
formation of normal extraembryonic endoderm in an in vitro
embryoid body formation assay (Trevor, 1990). However, ES
cells without a keratin filament network, following specific
keratin gene inactivation, can still form extraembryonic
endoderm in culture (Baribault and Oshima, 1991), but exhibit
significant lethality later during embryonic development in
vivo (Baribault et al., 1993). The basis of the difference in the
results obtained from these gene inactivation and negative
dominant mutant experiments is not understood.
The present studies represent a characterization of highly
related cell lines that either contain or lack a cIF network. A
number of cultured cell lines have been reported to lack
detectable cIFs (Traub et al., 1983; Venetianer et al., 1983;
Dellagi et al., 1985; Giese and Traub, 1986; Lilienbaum et al.,
1986). In particular, Hedberg and Chen (1986) found that a
human adrenal tumor cell line, designated SW-13, expressed
vimentin filaments in a mosaic pattern, and clones derived from
these cells were characterized as lacking any detectable cIFs. We
have isolated clones of SW-13 cells that either contain or lack
vimentin-type filaments (Sarria et al., 1990) and have transfected
vimentin negative cells with a mouse vimentin expression
plasmid to produce stable lines of cells that contain mouse
vimentin filaments (Sarria et al., 1992). In addition, SW-13 cells
that contain human vimentin were transfected with a carboxyterminally truncated mutant vimentin expression plasmid to
produce stable lines of cells with disrupted filament networks.
These studies indicate that while the presence or absence of an
organized cIF network in SW-13 cells does not appear to be a
major determinant of cytoplasmic organization, the presence or
absence of an organized vimentin filament network has a demonstratable effect on nuclear morphology in these cells.
length mouse vimentin cDNA, has been described previously (Sarria
et al., 1990). To create a carboxy-terminal deletion, pSP64-MMTVVimS was cleaved at a unique BglII site, 1149 bases from the 5′ end
of the vimentin coding sequence (Hennekes et al., 1990). The DNA
ends were filled in with Klenow fragment and the blunt ends religated
to create pMMTV-Vim-83∆C. This procedure resulted a single base
frame-shift in the vimentin sequence after amino acid 384, producing
a five amino acid mutant sequence (Arg-Ser-Ala-Gln-Cys) followed
by a termination sequence. The predicted translation product of this
sequence thus lacks the 83 carboxy-terminal amino acids of wild-type
vimentin, including the entire carboxy-terminal tail and 26 amino
acids of the carboxy-terminal end of the alpha-helical rod domain.
Cell culture
SW-13 cells (ATCC CCL 105) were obtained from the American
Type Culture Collection (Rockville MD). Cells were grown in
monolayer culture in a 1:1 (v/v) mixture of Ham’s F12:Dulbecco’s
MEM containing 5% fetal bovine serum. SW-13 cells were cloned as
previously described (Sarria et al., 1990). Briefly, 50 clones were
obtained and examined for human vimentin content by immunofluorescence microscopy. Twenty-five of the clones were found to be
essentially vimentin negative, 22 were mosaics with a significant
number of positive and negative cells, and three clones were obtained
that essentially appeared to contain only cells with prominent
vimentin filament networks. The three clones of the vimentin-positive
cells, designated SW-13/cl.1, cl.5 and cl.11 vim+, and three selected
vimentin-negative clones, designated SW-13/cl.2, cl.3 and cl.7 vim−,
were selected and carried as cell lines for further study. Preliminary
studies indicate that these cell lines do not substantially differ in
growth characteristics in a vimentin-dependent manner. In addition,
we have been unable to observe any significant difference in total
protein and RNA synthesis (other than vimentin expression), or the
relative proportion of Triton-insoluble RNA, between the vim+ and
vim− cell lines (Sarria, 1993).
SW-13 cell lines transfected with mouse wild-type or
mutant vimentin cDNAs
Since it is likely that vimentin is not the only gene affected in the vim−
and vim+ cells, it cannot be assumed that all phenotypic differences
between vim+ and vim− SW-13 cell lines are associated with the
presence or absence of cIFs. Therefore, one of the vim− cell lines (cl.2)
was transfected with the wild-type vimentin expression plasmid pSP64MMTV-VimS, and one of the vimentin-containing cell lines (cl.1) was
transfected with the carboxy-terminally deleted pMMTV-Vim-83∆C.
Stable cell lines were obtained by cotransfecting 0.6×106 to 0.8×106
SW-13/cl.2 vim− cells in 10 cm dishes with 40 µg of the vimentin
expression plasmid and 1.0 µg pSV2neo by calcium phosphate precipitation (Graham and van der Eb, 1973) and 15% glycerol shock (Parker
and Strak, 1979). Stable transfectants were initially selected in conditioned medium containing 400 µg/ml G-418. Medium was conditioned
by adding 10 ml of medium to a semi-confluent 75 cm2 flask of SW13 cl.2 cells. After 8 hours, the conditioned medium was removed and
filtered before use. After 4-6 weeks, stable transfectants were then
cultured in medium containing 200 µg/ml G-418. Individual colonies
were recovered and examined for mouse vimentin filament content by
indirect immunofluorescence using an antibody specific for rodent
vimentin (Sarria et al., 1990). From cl.2 vim− cells transfected with the
wild-type mouse vimentin, two stable transfectant lines, designated
T3M and T4M vim+, were found to have a substantial mouse vimentin
filament content and been described previously (Sarria et al., 1992).
Cells with disrupted vimentin IF networks were obtained by transfecting cl.1 vim+ cells with the expression vector, pMMTV-Vim83∆C, which expresses a carboxy-terminally truncated vimentin. Two
stable cell lines were selected that contained significant filamentous
material recognized by the mouse vimentin specific antibody. These
cell lines were designated 83∆C/1A4 and 83∆C/1A2. These cell lines
were then recloned to ensure that a clonal population was obtained.
Previous studies with other human cell lines transfected with the wildtype vimentin plasmid had shown that the expression of the mouse
vimentin was dependent on the addition of dexamethasone (Sarria et
al., 1990). We had also observed a dexamethasone-dependent
expression of vimentin in transiently transfected SW-13 cells (Sarria
et al., 1990). However, additional studies with SW-13 cells indicated
that, in contrast to the other cell lines that we have studied, some
mouse vimentin expression can also be observed following transient
transfection in the absence of hormone. In stably transfected SW-13
cells the expression of this plasmid does not appear to be affected by
added dexamethasone (data not shown). This basal expression is
apparently a hormone-independent phenomenon because the
expression of vimentin in these cells is not affected by the glucocorticoid antagonist RU-486 or inhibitors of corticosteroid synthesis
(data not shown). In addition, conditioned medium from these cells
did not induce the transcription of an MMTV-luciferase plasmid in a
heterologous cell system that was responsive to dexamethasone concentrations as low as 10−11 M.
Plasmid construction
The plasmid pSP64-MMTV-VimS, which contains the promoter and
hormone response element of mouse mammary tumor virus and a full-
Fluorescence microscopy
Cells were plated on sterile glass coverslips. In experiments involving
immunofluorescence of IFs, the cells were rinsed briefly in phosphate
MATERIALS AND METHODS
Intermediate filaments and nuclear morphology 1595
buffered saline (PBS) and then fixed in acetone:methanol, 70:30 (v/v),
at −20°C for 10 minutes. The coverslips were rinsed in PBS, and
processed for indirect immunofluorescence as described by Franke et
al. (1978). A rabbit antiserum specific for rodent vimentin (Moscinski
and Evans, 1987), and a monoclonal antibody that reacts with human
but not mouse vimentin (V-9, Boehringer Mannheim, Indianapolis,
IN), were used as primary anti-vimentin antibodies as described previously (Sarria et al., 1990). Lamins A and C were visualized using
a human autoantiserum, LS-1 (McKeon et al., 1983, 1986). Microtubules and stress fibers were localized in cells that had been fixed for
30 minutes in PBS containing 3.7% formaldehyde, and then extracted
with a solution containing 0.15% Triton X-100, 1 M hexylene glycol.
Tubulin and actin were visualized using a monoclonal anti-β-tubulin
(Boehringer Mannheim, Indianapolis, IN) and rhodamine-phallacidin
(Molecular Probes, Eugene, OR), respectively. Lissamine-rhodamineconjugated anti-rabbit, lissamine-rhodamine or fluorescein-conjugated anti-mouse (Boehringer Mannheim) and fluorescein-conjugated
anti-human antisera (Maloy Laboratories or Boehringer Mannheim)
were used as second antibodies, respectively. All antibodies were
diluted in PBS containing 1% ovalbumin and 1% normal goat serum.
The cells on coverslips were viewed on an Olympus microscope
equipped with epifluorescence optics. Photographic exposures were
made for 15-30 seconds using Kodak T-Max 400 film and the film
was processed with an exposure index of 1200 using Kodak HC-110
developer.
Immunoblotting
Cells were incubated for 15 minutes in methionine-free medium and
then labeled with 25-50 µCi/ml [35S]methionine for 2 hours in the
same medium containing 1% FBS. Triton-insoluble cytoskeletons
were prepared by the method of Zackroff and Goldman (1979).
Triton-insoluble proteins were separated by one or two-dimensional
PAGE (Laemmli, 1970; O’Farrell, 1975). In some experiments, the
ampholyte composition of the isoelectric focusing dimension was
modified as described by Dagenais et al. (1984) to improve the resolution of basic lamin proteins. Following electrophoresis, the proteins
were transferred to nitrocellulose (Burnette, 1981). The primary antibodies were: purified IFA (gift from M. Klymkowsky), used at a concentration of 10 µg/ml, and rabbit anti-vimentin (Moscinski and
Evans, 1987), used at a dilution of 1:100. Bound antibody was
detected with a horseradish peroxidase-conjugated goat anti-mouse or
anti-rabbit Ig (Boehringer Mannheim, Indianapolis, IN) diluted 1:500,
using diaminobenzidine (DAB) as a substrate. The nitrocellulose was
blocked and the antibodies were diluted in PBS containing 5% BSA
and 1% Tween-20. Following development of the DAB reaction
product, the nitrocellulose was dried and autoradiographed on Kodak
XAR X-ray film at −70°C.
Electron microscopy
For conventional thin sections, cells in monolayer culture were fixed
in 3% phosphate buffered glutaraldehyde and then post-fixed in 1%
phosphate buffered osmium tetroxide. The monolayers were then
dehydrated in a graded ethanol series and embedded in a mixture of
Epon and Araldite (Mollenhauer, 1964). To prepare whole-mount
cytoskeletons, cells were grown on collodion-coated gold grids.
Detergent-extracted whole-mounts were prepared by a slight modification of the method described by Wang et al. (1989). The cells were
extracted with a solution containing 0.5% Triton X-100, 100 mM
NaCl, 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, 10 mM PIPES,
pH 6.8, and 1 mM PMSF (CSK I buffer), for 3 minutes at 4°C,
followed by a second extraction in 0.5% Triton X-100, 250 mM
(NH4)2SO4, 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, 10 mM
PIPES, pH 6.8, and 1 mM PMSF (CSK II buffer), for 5 minutes. The
grids were rinsed briefly in 0.5% Triton X-100, 50 mM NaCl, 300
mM sucrose, 3 mM MgCl2, 1 mM EGTA, 10 mM PIPES, pH 6.8, and
then incubated in the same buffer containing 100 µg/ml of both DNase
I and RNase A for 20 minutes at room temperature. The grids were
then incubated in CSK II buffer for 5 minutes, rinsed in 50 mM NaCl,
300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, 10 mM PIPES, pH
6.8, fixed in 2% glutaraldehyde, 0.1 M sodium cacodylate, pH 7.2,
for 30 minutes at room temperature, and then post-fixed in 1% OsO4
in the same buffer for 10 minutes at 4°C. The grids were rinsed twice
in 0.1 M sodium cacodylate, pH 7.2, dehydrated in graded ethanol,
and air dried. Whole-mount cytoskeletons and thin sections were
examined using a Philips 201 electron microscope.
RESULTS
Intermediate filament content of SW-13 clones
Immunofluorescence microscopy of the vim+ cell lines
indicated that greater than 99% of the cells contained
detectable vimentin filaments. Two of these lines, cl.1 vim+
and cl.11 vim+, exhibited prominent vimentin filament
networks radiating throughout the cytoplasm. A third line, cl.5
vim+, was characterized by cells that contained fewer vimentin
filaments, and these filaments appeared to be short (often but
not always) in a perinuclear region of the cytoplasm (data not
shown). A small number of cells (usually less than 1%) without
detectable vimentin filaments were always present in all three
vim+ cell lines. In contrast, examination of the three vim− SW13 cell lines (cl.2, cl.3 and cl.7) using immunofluorescence
indicated that greater than 99% of the cells had no detectable
vimentin IFs. However, in all three vim− cell lines there were
always a small number of cells (less than 0.5%) with visible
vimentin filaments. Attempts to detect keratin in SW-13 clones
with monoclonal antibodies specific for keratins 8 or 18 were
negative using immunofluorescence (data not shown). Similar
attempts to localize cIFs with IFA, a monoclonal antibody to
a shared IF epitope (Pruss et al., 1981; Osborn and Weber,
1987), detected a filament system only in the vim+ clones (data
not shown). These results are consistent with the observations
of Hedberg and Chen (1986) and Osborn and Weber (1987),
that vimentin is the only cIF protein expressed in SW-13 cells.
Examination of the cell lines transfected with a mouse
vimentin cDNA expression vector and selected for vimentin
expression showed that these clones express murine vimentin
in a heterogeneous manner. Anti-vimentin immunofluorescence of T3M vim+ cells indicated that while approximately
40% of the cells had very prominent, relatively uniformly distributed vimentin cIF networks, some cells contained smaller
amounts of vimentin, often localized in one region of the
cytoplasm, and approximately 10-15% of the cells did not
contain any detectable vimentin filaments. Similar analysis of
the second transfectant cell line, T4M vim+, showed that in
these cells only about 30% of the population contained
detectable vimentin IFs. Analysis of Triton-insoluble proteins
of wild-type and transfectant cell lines had shown that the
overall level of vimentin synthesis in T3M and T4M cells was
about 20% and 10%, respectively, of that observed in the cl.1
vim+ cells (Sarria et al., 1992). The heterogeneity of mouse
vimentin expression in these cell lines is a clonal phenomenon,
as repeated subcloning of these cells produced cell populations
with similar properties.
Northern analysis of RNA isolated from a vim+ clone and
two of the vim− clones with a vimentin cDNA probe indicated
that little if any vimentin mRNA could be detected in the vim−
cells, indicating that the failure of these cells to produce
1596 A. J. Sarria and others
vimentin involves events at the transcriptional level (data not
shown).
Two-dimensional gel and immunoblotting analysis of
Triton-insoluble extracts from [35S]methionine-labeled cells
was consistent with the immunofluorescence data. Little
vimentin synthesis could be detected in the vim− clones, while
35S-labeled vimentin was the major Triton-insoluble protein of
the vim+ cells. The T3M vim+ cells, which express a mouse
vimentin cDNA, contained less labeled vimentin than the cl.1
vim+ cells; however, vimentin was still a major component of
this preparation. In order to determine if other unknown cIF
proteins might be present as minor components in these preparations, two-dimensional gels of Triton-insoluble 35S-labeled
proteins were immunoblotted using the IFA antibody. In these
studies, vimentin was the major IFA immunoreactive protein
in the cl.1 vim+ and T3M vim+ preparations. Trace amounts of
vimentin were detected in cl.2 vim− cell preparations, presumably the contribution of the small number of vimentin-containing cells detected in immunofluorescence microscopy (data
not shown; Sarria, 1993).
Although examination of Triton-insoluble cytoskeleton
preparations from vim− cell lines failed to reveal the presence
of other known cIF proteins, it was possible that these cells
might contain filament protein as a minor component or a cIF
protein that was not recognized by the IFA antibody.
Therefore, detergent-extracted cytoskeletons were examined in
whole-mount electron microscopy. As shown in Fig. 1,
detergent- and salt-extracted cl.1 vim+ cells exhibited an
elaborate system of cytoplasmic filaments with diameters
appropriate for IFs. However, in similar preparations of cl.2
vim− cells, these filaments appeared to be completely absent
(Fig. 1B).
Morphology and distribution of cytoplasmic
organelles in vim− and vim+ SW-13 clones
Numerous studies have indicated a relationship between the
organization of cIFs and other components of the cytoskeleton. However, the precise nature of potential interactions of
cIFs with microtubules and microfilaments has not been well
characterized (see Klymkowsky et al., 1989). Anti-tubulin
indirect immunofluorescence of vim+ and vim− cells indicated
that there were no dramatic differences in the appearance of
cytoplasmic microtubules. The rhodamine-phallacidin fluorescence patterns of the vim+ cl.1 and vim− cl.2 cells indicated
that while both cell lines contained actin stress fibers, the cl.1
cells that contained a vimentin network exhibited stress fibers
that were larger and more prominent than those observed in the
cl.2 cells without vimentin filaments. However, examination of
the mouse vimentin-containing T3M vim+ line using antivimentin and phallacidin double immunofluorescence
indicated that while virtually all the cells in this population
exhibited prominent stress fibers, similar to the cl.1 vim+ cells,
this included cells that contained vimentin as well as cells that
lacked a detectable vimentin cIF network (data not shown;
Sarria, 1993). These observations indicate that the presence or
absence of a vimentin filament network does not have a
dramatic effect on the organization of other major cytoskeletal components. Visualization of mitochondria, endoplasmic
Fig. 1. Whole-mount electron micrographs of vim+ and vim− SW-13 cell lines. The figure shows cl.1 vim+ (A) and cl.2 vim− (B) cells that were
detergent- and salt-extracted, treated with nuclease, and viewed as whole-mounts. The nuclear residue (N) is visible in both cells but only the
vim+ cell (A) has filaments visible in the cytoplasm. The magnifications are identical for A and B. Bar, 300 nm.
Intermediate filaments and nuclear morphology 1597
Fig. 2. Electron micrographs of vim+ and vim− SW-13 cell lines. The figure shows thin sections of cl.1 vim+ (A) and cl.2 vim− (B) cells. The
magnifications are identical for A and B. Bar, 5 µm.
reticulum and Golgi, using rhodamine 123, 3,3′-dihexyloxacarbocyanine iodide and NBD-ceramide, respectively, did not
reveal any observable differences between cells that contain or
lack a cIF network (data not shown).
SW-13 cells that lack cytoplasmic intermediate
filaments have irregularly shaped nuclei
Examination of SW-13 clones that contain (vim+) or lack
(vim−) a cIF system using thin-section electron microscopy
revealed that many of the vim− cells exhibited a highly
irregular nuclear morphology. As shown in Fig. 2A, the nuclei
of the cl.1 vim+ cells appeared to be relatively smooth or
regular in shape. In contrast, the nuclei of the cl.2 vim− cells
often appeared to be highly folded, forming prominent lobes
and clefts (Fig. 2B). Although this irregular nuclear morphology was a prominent feature of all the vim− lines, not all the
nuclei in a given section had this appearance, and occasional
nuclei in the thin sections of the vim+ cells also exhibited some
irregularities. To determine the prevalence of the folded or
invaginated nuclear shape seen in thin sections, vim+ and vim−
cells were examined in light microscopy. This difference in
nuclear morphology between cells that contain or lack
vimentin filaments was not apparent in living or fixed cells in
phase-contrast microscopy. Therefore, cells were examined
using immunofluorescence microscopy to visualize a major
structural component of the nuclear envelope.
The nuclear lamin proteins A, B and C interact to form a
fibrillar network located along the inner surface of the nuclear
membrane, referred to as the nuclear lamina (Aebi et al., 1986).
Fig. 3. Two-dimensional immunoblots of Triton-insoluble
cytoskeletons from vim+ and vim− cells with the IFA antibody.
Preparations from similar numbers of cells were separated by twodimensional PAGE as described by Dagenais et al. (1984). The
figure shows the IFA DAB reaction product in an enlarged region
of two-dimensional immunoblots from cl.1 vim+ cell (A), cl.2 vim−
cell (B), and T3M vim+ cell (C) Triton-insoluble preparations. The
position of vimentin (V), and proteins with two-dimensional gel
mobilities similar to lamin proteins A (a), B (b) and C (c) are
indicated. Two additional minor immunoreactive proteins (*),
which may be lamin B variants (Vorburger et al., 1989), are also
shown.
1598 A. J. Sarria and others
The nuclear lamina is believed to play an essential role in the
maintenance of the structure of the nuclear envelope (Gerace
and Burke, 1988; Whytock et al., 1990). Numerous studies
have indicated that the composition of the nuclear lamina can
vary in different cell types. Although lamin B is constitutively
expressed, lamins A and C are absent in a number of cultured
cell lines (Guilly et al., 1987; Lebel et al., 1987; Stewart and
Burke, 1987; Worman et al., 1988; Paulin-Levasseur et al.,
1988, 1989a). In particular, Paulin-Levasseur et al. (1989b)
have reported that SW-13 cells can differ substantially in the
expression of lamin proteins A and C. In order to determine if
the differences in the appearance of the nuclear envelope could
be due to differences in lamin protein composition,
immunoblotting experiments were carried out using the IFA
antibody. The nuclear lamin proteins also share the IFA
epitope common to cIFs, although lamins are reported to be
less immunoreactive than vimentin (Osborn and Weber, 1987).
As shown in Fig. 3, immunoblotting of Triton-insoluble
Fig. 4. Anti-lamin A/C indirect immunofluorescence of SW-13 cell lines. The figure shows the lamin A/C staining pattern for cl.1 vim+ (A),
cl.11 vim+ (B), cl.5 vim+ (C), cl.2 vim− (D), cl.3 vim− (E), and cl.7 vim− (F) cells.
Intermediate filaments and nuclear morphology 1599
extracts of the cl.1 vim+, cl.2 vim− and T3M vim+ cells
indicated expression of similar amounts of IFA reactive
proteins with two-dimensional gel mobilities similar to those
reported for lamins A, B and C (Dagenais et al., 1984). Preliminary experiments with two different A/C-specific antisera
indicated that, as shown by immunofluorescence microscopy,
these cells had similar detectable A/C lamin nuclear fluorescence intensity (see below). In addition, it has been reported
that MPC-11 cells that do not express lamins A and C lack a
salt-stable nuclear matrix (Wang and Traub, 1991). Similar
immunoblotting experiments with anti-IFA antibodies of saltand nuclease-treated SW-13 vim+ and vim− cell extracts failed
to reveal any detectable difference in the salt stability of lamins
A and C (data not shown).
To gain a better insight into the actual number of cells within
a given population that had anomalous nuclear morphology
and to specifically visualize the lamina in these cells, immunofluorescence experiments were carried out with an A/Cspecific anti-lamin antibody. Examination of the vim+ and
vim− SW-13 cell lines indicated that while all clones had
similar lamin A/C fluorescence intensity, there was a marked
difference in the appearance of the lamina and the shape of the
nucleus between the vim+ and vim− cells. The three vim+ cell
lines revealed a more regular lamina shape (Fig. 4A-C), similar
to that seen in other cultured cells (McKeon et al., 1983;
Sinensky et al., 1990), while the vim− cell lines exhibited a
lamina that appeared to be highly irregular or folded (Fig. 4DF). However, it was clear that the differences in nuclear morphology were not absolute. There were some irregular
appearing nuclei in the vim+ populations and, conversely, some
regular appearing nuclei in the vim− populations. Using the
morphological differences shown in Fig. 4 as criteria, all clones
were scored for regular or smooth and irregular or folded antiA/C lamin immunofluorescence patterns, and the results are
given in Table 1.
Expression of an organized mouse vimentin
filament network in SW-13 cells that lack
endogenous intermediate filaments restores a more
regular nuclear morphology
Since the difference in nuclear morphology between vim+ and
vim− SW-13 cell lines was consistent in three independently
isolated clones of each type, this difference is unlikely to be
Table 1. Nuclear morphology in vim+ and vim− SW-13 cell
lines
Smooth
Invaginated
Cell line
%
(range)
%
(range)
cl.1 vim+
cl.5 vim+
cl.11 vim+
cl.2 vim–
cl.3 vim–
cl.7 vim–
78
81
80
48
21
46
(86-72)
(87-76)
(84-78)
(65-33)
(22-16)
(55-30)
22
19
20
52
79
54
(28-14)
(24-13)
(22-16)
(67-35)
(84-74)
(70-45)
The anti-lamin A/C immunofluorescence staining was examined for each
of the cell lines shown in Fig. 4. The values given for each cell line are the
averages for over 350 cells in random fields, evaluated by three independent
observers for the percentage of cells exhibiting a uniform (smooth) vs
invaginated lamin A/C morphology. The numbers in parenthesis give the
range (high-low) between the values obtained by individual observers.
due to simple clonal variation. To test more directly the
hypothesis, that cIFs influence the shape of the nucleus, antivimentin and anti-lamin A/C double immunofluorescence
experiments were carried out with T3M vim+ cells, which were
derived from vim− cells, and which express mouse vimentin in
a heterogeneous fashion from a stably transfected vimentin
expression vector. As shown in Fig. 5, the anti-vimentin
immunofluorescence staining of T3M vim+ cells demonstrates
that some of the cells had prominent mouse vimentin filaments
organized throughout the cytoplasm (Fig. 5A), and some cells
had prominent vimentin filaments that appeared to be nonuniformly distributed, localized in particular regions of the
cytoplasm (Fig. 5C). In addition, there were cells that had little
or no detectable anti-vimentin fluorescence (Fig. 5E). When
vimentin filament expression and distribution was compared
with the appearance of the lamin A/C fluorescence in these
cells (Fig. 5B,D and F), the expression of uniformly distributed mouse vimentin IFs was associated with a decreased
frequency of cells with folded or invaginated nuclei. However,
similar to the appearance of the nuclear lamina in the other
SW-13 clones, the difference in lamin fluorescence patterns in
T3M vim+ cells with different levels of vimentin filament
expression was not absolute. A few cells with well organized
mouse vimentin IFs had prominent nuclear invaginations while
some cells that contained little or no vimentin had nuclei
without visible invaginations or folds. To determine the degree
of association between vimentin expression and nuclear shape,
the cells in the experiment shown in Fig. 5 were scored for
vimentin content and nuclear morphology. As shown in Table
2, the T3M vim+ cells with little or no detectable mouse
vimentin expression had a fraction of nuclei with prominent
folding similar to that of the untransfected cl.2 vim− cells. In
contrast, only a few of the T3M vim+ cells with well organized
mouse vimentin filaments exhibited irregular nuclei, similar to
the fraction of irregular nuclei in SW-13 clones with endogenous human vimentin filament networks. These results would
also indicate that not only expression but also the organization
of expressed vimentin is an important factor. T3M vim+ cells
that exhibited significant vimentin fluorescence but had a nonuniform distribution of filaments, as shown in Fig. 5C and D,
had more irregular nuclei than cells with more uniformly distributed cIFs (Fig. 5A and B). Similar results were seen in a
second transfectant line, T4M vim+, which expresses mouse
vimentin at a lower level than T3M vim+ cells (Table 2).
Vimentin filament disruption by synthesis of a
carboxy-terminally truncated mutant vimentin
results in altered nuclear shape
Since expression of a vimentin cDNA appeared to restore a
more regular nuclear morphology in SW-13 cells that lacked
an endogenous cIF system, it was also of interest to determine
whether disruption of the vimentin cIFs in vim+ cells would
produce the invaginated morphology characteristic of vim−
cells. Therefore, cl.1 vim+ cells were stably transfected with a
mutant mouse vimentin expression plasmid that codes for a
vimentin protein that lacks the 83 carboxy-terminal amino
acids of wild-type vimentin. This truncation is similar to the
mutations in other cIF proteins described in recent studies that
produce a dominant negative phenotype and can specifically
disrupt the endogenous cIF network, even when expressed at
a small fraction of the level of the wild-type subunit protein
1600 A. J. Sarria and others
(Albers and Fuchs, 1987, 1989; van den Heuvel et al., 1987;
Coulombe et al., 1990; Gill et al., 1990; Lu and Lane, 1990;
Wong and Cleveland, 1990; Chin et al., 1991; Raats et al.,
1991). Two cell lines were obtained that contained material
that was immunoreactive with a rodent-specific anti-vimentin
antibody. As shown in Fig. 6B, immunoblots of Tritoninsoluble cytoskeletons probed with this polyclonal antivimentin antibody detected an approximately 48 kDa protein
(in addition to vimentin) in both the 83∆C transfectant lines
that was not detectable in the parental cl.1 vim+ cells. Although
this antiserum does not recognize human vimentin in indirect
immunofluorescence, it does react with the wild-type human
protein on immunoblots. Examination of both the immunoblot
(Fig. 6B) and the 35S autoradiograph of the labeled protein
(Fig. 6C) showed that the 48 kDa vimentin immunoreactive
protein was present in smaller amounts than the normal human
protein. Identical blots probed with the IFA antibody, which
recognizes an epitope in the carboxy-terminal end of the helical
Fig. 5. Anti-lamin A/C and anti-vimentin double indirect immunofluorescence of T3M vim+ cells. T3M vim+ cells were obtained after stable
transfection of cl.2 vim− cells with a mouse vimentin cDNA expression plasmid. The figure shows anti-vimentin staining (RITC) (A,C,E), and
anti-lamin A/C staining (FITC) (B,D,F) for T3M vim+ cells on the same coverslip. (A and B) Examples of cells that contain significant
vimentin that is uniformly distributed (note the vimentin-negative cells also present). (C and D) Cells that contain significant vimentin that is
not uniformly distributed (also note the vimentin-negative cells present). (E and F) Cells that contain little or no detectable vimentin filaments.
Intermediate filaments and nuclear morphology 1601
Table 2. Effect of expression and distribution of mouse
vimentin IFs on nuclear morphology in cells that lack
endogenous human vimentin IFs
Smooth
Vimentin dist. (% total)
T3M cells
Uniform vim+ (7)
Non-uniform vim+ (41)
vim+/– (37)
vim– (15)
T4M cells
Uniform vim+ (3)
Non-uniform vim+ (17)
vim+/– (9)
vim– (71)
kDa
Invaginated
%
(range)
%
(range)
76
46
33
41
(82-72)
(49-41)
(34-33)
(44-35)
24
54
67
59
(28-18)
(59-51)
(67-66)
(65-56)
68
37
32
27
(74-63)
(40-34)
(36-27)
(31-23)
32
63
68
73
(37-26)
(66-60)
(73-64)
(77-69)
T3M and T4M cells were derived from the cl.2 vim– cell line by stable
transfection with a mouse vimentin cDNA expression vector (Sarria et al.,
1992). These cell lines express the mouse vimentin transgene
heterogeneously and on the basis of the anti-vimentin immunofluorescence
staining pattern were subdivided into cells that contained prominent vimentin
networks distributed uniformly throughout the cytoplasm (Uniform vim+),
cells that exhibited prominent vimentin networks that were restricted to only a
portion of the cytoplasm (Non-uniform vim+), cells that contained only a
small number of faintly visible vimentin IFs (vim+/–), and cells that had no
observable vimentin IFs (vim–). Examples of the anti-vimentin and anti-lamin
A/C immunofluorescence staining patterns for the T3M cell line are shown in
Fig. 5. The values given for each cell line are the averages for over 2000 cells
in random fields, evaluated by two (T4M cells) or three (T3M cells)
independent observers for the percentage of cells exhibiting a uniform
(smooth) vs invaginated lamin A/C morphology.
rod domain of vimentin (Magin et al., 1987) that should be
absent from the mutant protein, did not detect the 48 kDa
protein (Fig. 6A). These results indicate that a carboxy-terminally truncated anti-vimentin immunoreactive protein of
appropriate size to be the mutant mouse vimentin can be
detected in the 83∆C transfectant cell lines and is not present
in the parental cl.1 vim+ cells.
To determine the effect of expression of this mutant
vimentin on the endogenous human vimentin IFs, double
immunofluorescence of these cells was performed with mouse
and human anti-vimentin-specific antibodies (Sarria et al.,
1990). As shown in Fig. 7C and D, the transfected cells
contained material that was recognized by the mouse vimentinspecific antibody that colocalized with the endogenous human
vimentin network. In these cells the vimentin network
appeared to have collapsed into perinuclear aggregates. Similar
examination of the parental cl.1 vim+ cells (Fig. 7A and B)
revealed no staining with the mouse vimentin-specific
antibody, and a normal, extended human vimentin network. In
both 83∆C cell lines there was a clear relationship between the
amount of immunofluorescent material detected by the mouse
vimentin-specific antibody and the degree of disruption of the
human vimentin IFs. To determine the effect of vimentin
filament disruption on nuclear shape, double immunofluorescence with anti-vimentin and anti-lamin antibodies was carried
out. As shown in Fig. 8, 83∆C cells with prominently collapsed
endogenous vimentin IFs often exhibited nuclei with irregular
lamin staining patterns (C and D), which were similar in
appearance to those found in vim− cells (see Fig. 5), while the
parental cl.1 vim+ cells had extended filament networks and
few irregularly shaped nuclei (A and B). To determine the
Fig. 6. Anti-vimentin immunoblot analysis of SW-13 cells
transfected with a carboxy-terminally deleted mutant mouse
vimentin cDNA. Triton-insoluble cytoskeletons were prepared from
[35S]methionine-labeled cells, the proteins were separated on 10%
SDS-PAGE, and transferred to nitrocellulose. The figure shows the
separation of Triton-insoluble proteins from untransfected cl.1 vim+
cells (lane 1), 83∆C/1A4 cells (lane 2), and 83∆C/1A4 cells (lane 3).
(A and B) The DAB staining product of duplicate samples probed
with IFA antibody (A) and rabbit anti-vimentin (B). (C) The
autoradiograph of the nitrocellulose strip in B. The position of
vimentin (V) and molecular mass standards (Sigma SDS-6H) are
indicated. In (B and C) an arrowhead indicates the position of an
approximately 48 kDa anti-vimentin immunoreactive protein present
in the transfectant cell lines, not detected in untransfected cells or by
the IFA antibody.
degree of association between filament disruption and nuclear
morphology, cells were scored for collapsed endogenous
vimentin filaments, mutant vimentin expression, and nuclear
shape. As shown in Table 3, there was a strong correspondence
between the percentage of the transfectant cells that contained
significant amounts of mutant protein, the fraction of the cells
with collapsed endogenous filament networks, and the fraction
of the cells that exhibited an invaginated or folded nuclear
shape.
Disruption of microtubules does not restore a
regular nuclear morphology
The difference in the morphology of the nucleus observed
between cell lines that contain or lack cIFs could be due to a
direct effect of cIFs on the nucleus or an indirect reflection of
changes in other cytoplasmic structures that are normally associated with cIFs, which in turn affect the shape of the nucleus.
Since cIFs are often described as having an interaction with
microtubules (see Klymkowsky et al., 1989), and invagination
of the nucleus around areas of centrosomal microtubules is
seen in early prophase cells, studies were conducted to
determine if the nuclear morphology of cl.2 vim− cells might
be due to the action of microtubules freed of the constraint of
a cIF network. Cells were treated with 5 µg/ml nocodazole for
6 hours and nuclear morphology was compared with untreated
cells using anti-lamin A/C immunofluorescence. As shown in
Table 4, the relative number of cl.2 vim− cells with a folded or
irregular nuclear morphology was not affected by nocodazole.
However, nocodazole treatment increased the fraction of cl.1
vim+ cells with folded or irregular nuclei, although this
1602 A. J. Sarria and others
Fig. 7. The effect of expression of a carboxy-terminally deleted vimentin on the endogenous filament network in SW-13 vim+ cells. The figure
shows double immunofluorescence of untransfected cl.1 cells (A and B) and 83∆C/1A4 cells transfected with a carboxy-terminally truncated
mouse vimentin cDNA (C and D) stained with rabbit anti-mouse vimentin (FITC; B and D) and monoclonal anti-human vimentin (RITC; A
and C). A second, independently derived transfectant cell line, 83∆C/1A2, exhibited a similar vimentin staining pattern to that seen in
83∆C/1A4 (data not shown).
treatment did not result in vim+ cells with the same fraction of
irregular nuclei as vim− cells (Table 4). The distribution of
tubulin and lamin A/C was examined in these cells using
double immunofluorescence. As shown in Fig. 9 (E and G),
there were no detectable microtubules in the nocodazoletreated cells. In addition, since anti-microtubule agents,
including nocodazole, indirectly affect vimentin filaments
(Geuens et al., 1983), the distribution of vimentin and lamin
A/C was also examined in the vim+ cells. As expected, nocodazole-treated vim+ cells exhibited a prominent perinuclear
collapse of the cIF network (Fig. 9A-D). These experiments
show that disruption of the microtubule network in cells that
lack cIFs does not restore a regular nuclear shape, and that the
cytoplasmic microtubule network is not responsible for the
irregular or folded nuclear morphology in these cells.
Moreover, it seemed that the cIF collapse associated with MT
disruption increased the frequency of invaginated nuclear morphology in vim+ cells, but not in vim− cells.
DISCUSSION
The development of SW-13 cell clones that contain or lack
vimentin cIFs provides a useful model to evaluate systematically the functional significance of the cIF component of the
cytoskeleton at a cellular level. In agreement with previous
studies on cultured cells with antibody-disrupted cIF systems
(Klymkowsky, 1981; Lin and Feramisco, 1981; Gawlitta et al.,
1981; Klymkowsky et al., 1983; Summerhayes et al., 1983;
Tolle et al., 1986), the presence or absence of cIFs does not
appear to have a profound effect on the organization of the
cytoplasm in SW-13 cells. We have been unable to discern any
prominent difference in the distribution of major organelles
visible in light and thin-section electron microscopy in cells
that contain or lack cIFs. While this does not mean that cIFs
do not have important interactions with other cytoplasmic components, particularly other elements of the cytoskeleton, it does
indicate that cIFs do not have a central organizing function.
In contrast to the relative similarity in cytoplasmic organization, the appearance of the nucleus was a prominent structural
feature of SW-13 cells that did appear to differ between cells
that contain or lack cIFs. Cells that lacked cIFs had a nuclear
morphology that often appeared irregular, with prominent
folding or invaginations, while cells with vimentin cIFs characteristically exhibited a uniform or smooth nucleus. This was
specifically related to the cIF network because the highly
invaginated appearance of the nucleus in many cells without a
cIF system was restored to a more uniform appearance by
expression of a vimentin cDNA. Also, specific disruption of
the endogenous cIF network by expression of a truncated
Intermediate filaments and nuclear morphology 1603
Fig. 8. Nuclear morphology in cells with mutant vimentin disrupted endogenous filament networks. The figure shows the double
immunofluorescence of untransfected SW-13 vim+ cells (A and B) and 83∆C/1A4 cells transfected with a carboxy-terminally truncated mouse
vimentin cDNA (C and D) stained with anti-human vimentin (RITC; A and C) and anti-lamin A/C (FITC; B and D). A second, independently
derived transfectant cell line, 83∆C/1A2, exhibited a similar vimentin staining pattern to that seen in 83∆C/1A4 (data not shown).
mutant vimentin cDNA appeared to produce the same invaginated nuclear morphology found in cells that lack cIFs. These
observations indicate that the presence or absence of a
vimentin type intermediate filament network can influence the
morphology of the nucleus and may provide more than a
passive anchoring function. However, these studies also show
Table 3. Endogenous vimentin cIF network organization
and nuclear morphology in cell lines that express a
carboxy-terminally deleted mutant vimentin
Cell line
% Cells with
collapsed cIFs*
SW–13/cl.1 vim+ 15.7 (19-12)
83∆C–1A4
45.1 (49-43)
83∆C–1A2
48.2 (49-47)
% Cells expressing
% Cells with
mutant vimentin† invaginated nuclei‡
NO4
57.0 (59-58)
54.2 (63-47)
18.0 (21-14)
43.5 (52-38)
45.2 (51-34)
The values given were obtained from double immunofluorescence
experiments as shown in Fig. 8. For each of the 83∆C transfectant cell lines
and the parental cl.1 vim+ cell line, over 500 cells in random fields in two
separate experiments were independently examined by two observers. The
numbers in parenthesis give the range (high-low) between the values obtained
by individual observers.
*% Cells that had prominent cIF collapse determined by indirect
immunofluorescence using a human vimentin-specific antibody.
†% Cells that contained filaments that were positive in indirect
immunofluorescence with a mouse vimentin-specific antibody.
‡% Cells that exhibited folded or invaginated nuclei determined by antilamin A/C indirect immunofluorescence.
§NO, not observable.
that the effect of vimentin filaments on the invaginations or
folding in the nucleus is clearly not an absolute. The presence
or absence of vimentin filaments appears to affect the
frequency of cells in a given population with visibly invaginated or folded nuclei. One possible explanation for this is that
the frequency of invaginations in the nucleus is a reflection of
the flexibility of the nuclear envelope and underlying nuclear
structures. Interaction with the cIFs could make the envelope
less flexible, or more stable, which would result in a structure
that is less likely to be invaginated and more uniform in appear-
Table 4. Effect of nocodazole on the nuclear morphology
of vim+ and vim– SW-13 cell lines
% Total (range)
−Nocodazole
+Nocodazole
Cell line
Smooth
Invaginated
Smooth
Invaginated
cl.1 vim+
cl.2 vim−
75 (71-80)
24 (25-22)
25 (29-20)
76 (78-75)
54 (58-51)
17 (23-14)
46 (49-42)
83 (86-77)
Anti-lamin A/C immunofluorescence was examined in untreated cl.1 vim+
and cl.2 vim− cells and in cells exposed to 5 µg/ml nocodazole for 6 hours.
For each treatment group, over 500 cells in random fields were evaluated by
three independent observers for the percentage of the cells exhibiting a
uniform vs invaginated lamin A/C morphology. The numbers in parenthesis
give the range (high-low) between the values obtained by individual
observers.
1604 A. J. Sarria and others
ance. Alternatively, the folded or invaginated nucleus could be
a result of cytoplasmic forces acting on the nucleus and the differences that we have observed could be a reflection of a more
highly organized cytoplasm in the presence of a cIF network.
This latter explanation would seem unlikely in view of observations made in cells with antibody-disrupted cIF networks.
Fig. 9. The effect of nocodazole on the distribution of lamins A/C, vimentin and tubulin. The cells in A and B, E and F are untreated controls.
The cells in C and D, G and H were exposed to 1 µg/ml nocodazole for 6 hours. (A-D) Double immunofluorescence of cl.1 vim+ cells stained
with anti-lamin A/C (B,D), and anti-vimentin (A,C) antibodies. (E-H) Double-immunofluorescence of cl.2 vim− cells stained with anti-lamin
A/C (F,H) and anti-tubulin (E,G) antibodies.
Intermediate filaments and nuclear morphology 1605
Numerous studies have shown that the organization and
motion of other major cytoplasmic organelles do not appear to
be affected by disruption of cIFs (see Klymkowsky et al.,
1989). Yet another possibility is that the folded configuration
of the nucleus could be an indirect effect of a metabolic difference between cells that contain or lack organized cIFs,
which does not affect the other cellular characteristics that we
have studied thus far.
Although the present studies did not directly address the
issue of the nature of the association of cIFs with the nucleus,
previous work has indicated that cIFs may specifically
associate with the nuclear lamina. In certain electron
microscopy preparations, the nuclear pore-lamina complex
appears to interact with cIFs, suggesting some structural continuity (Fey et al., 1984; Katsuma et al., 1987; Carmo-Fonseca
et al., 1988; Wang et al., 1989; Wang and Traub, 1991). In
vitro binding studies have shown that cIF proteins can interact
with nuclear envelope preparations from nucleated avian erythrocytes (Georgatos and Blobel, 1987; Georgatos et al., 1987)
and bind to lamin B in solution (Georgatos and Blobel, 1988).
In addition, animals immunized with a synthetic peptide representing the region of cIFs that is required for lamin B binding
in vitro, produce anti-idiotypic antibodies that recognize lamin
B (Djabali et al., 1991). While these observations indicate that
cIFs may directly interact with the nuclear envelope, antifilament antibody microinjection studies have produced
somewhat mixed results. Some studies have shown that disruption of cIFs had no observable effect on the re-formation of
the nuclear lamina following mitosis (Gawlitta et al., 1981;
Klymkowsky, 1982; Klymkowsky et al., 1983), but recently
Kouklis et al. (1993) reported that microinjection of antivimentin antibodies affected the rate of re-formation of the
nuclear lamina around chromatin. In addition, a direct interaction between cIFs and the lamina has generally not been
apparent in conventional thin-section electron microscopy.
Therefore, while it would be tempting to speculate that in SW13 cells, the observed effect of a cIF network on nuclear shape
might involve a direct effect on the nuclear envelope, further
experiments are clearly needed to assess this possibility.
Other cell types that lack cIFs also appear to exhibit
abnormal nuclear morphology. Wang et al. (1989) have shown
that MPC-11 cells that lack a cIF system have anomalous
nuclei that resemble in morphology those we have observed
in vim − SW-13 clones. However, in the case of MPC-11 cells,
it was reported that induction of vimentin synthesis by a 20
hour treatment of the cells with 12-O-tetradecanoylphorbol13-acetate (TPA) did not result in normalization of the nuclear
architecture. In these studies the vimentin organization was
described as concentrated near the nucleus and it was speculated that these newly formed filaments might not be capable
of correcting the nuclear anomalies in the time period during
which vimentin was expressed (Wang et al., 1989). In
addition, MPC-11 cells not only lack cIFs, but also do not
express lamins A and C (Paulin-Levasseur et al., 1988), and
are reported to lack a salt-stable nuclear matrix (Wang and
Traub, 1991).
The present studies do not provide an immediate insight into
the functional significance of an interaction between the cIF
network and the nucleus. The difference that we have observed
in the shape of the nucleus of SW-13 vim+ and vim− cells does
not appear to be reflected in any obvious effect on the growth
of these cells in monolayer culture, their cloning efficiency in
soft agar, or the overall synthesis of macromolecules.
However, these observations are consistent with a role for cIFs
in maintaining the mechanical integrity of cells as recently
proposed by Coulombe et al. (1991).
This work was supported by National Institutes of Health grants
GM-42770 and HL-51850. We are grateful to Alan Jones for the
preparation of the thin sections for electron microscopy. We are also
grateful to Frank McKeon (Harvard Medical School) for his kind gift
of the LS-1 antiserum. We thank Michael Klymkowsky (University
of Colorado) for the anti-IFA antibody and his helpful comments on
this manuscript.
REFERENCES
Aebi, U., Cohn, J., Buhle, L. and Gerace, L. (1986). The nuclear lamina is a
meshwork of intermediate-type filaments. Nature 323, 560-564.
Albers, K. and Fuchs, E. (1987). The expression of mutant epidermal keratin
cDNAs transfected in simple epithelial and squamous cell carcinoma lines. J.
Cell Biol. 105, 791-806.
Albers, K. and E. Fuchs, E. (1989). Expression of mutant keratin cDNAs in
epithelial cells reveals possible mechanisms for initiation and assembly of
intermediate filaments. J. Cell Biol. 108, 1477-1439.
Baribault, H. and Oshima, R. G. (1991). Polarized and functional epithelia
can form after the targeted inactivation of both mouse keratin 8 alleles. J. Cell
Biol. 115, 1675-1684.
Baribault, H., Price, J., Miyai, K. and Oshima, R. G. (1993). Mid-gestational
lethality in mice lacking keratin 8. Genes Dev. 7, 1191-1202.
Burnette, W. H. (1981). Western blotting: electrophoretic transfer of proteins
from SDS-polyacrylamide gels to unmodified nitrocellulose and
radiographic detection with antibody and radioiodinated Protein A. Anal.
Biochem. 112, 195-203.
Carmo-Fonseca, M., Cidadao, A. J. and David-Ferreira, J. F. (1988).
Filamentous crossbridges link intermediate filaments to the nuclear pore
complexes. Eur. J. Cell Biol. 45, 282-290.
Chin, S. S. M., Macioce, P. and Liem, R. K. H. (1991). Effects of truncated
neurofilament proteins on the endogenous intermediate filaments in
transfected fibroblasts. J. Cell Sci. 99, 335-350.
Coulombe, P. A., Chan, Y.-M. Albers, K. and Fuchs, E. (1990). Deletions in
epidermal keratins leading to alterations in filament organization in-vivo and
in intermediate filament assembly in-vitro. J. Cell Biol. 111, 3049-3064.
Coulombe, P.A., Hutton, M. E., Vassar, R. and Fuchs, E. (1991). A function
for keratins and a common thread among different types of epidermolysis
bullosa simplex diseases. J. Cell Biol. 115, 1661-1674.
Dagenais, A., Bibor-Hardy, V. and Simard, R. (1984). Characterization of
lamin proteins in BHK cells. Exp. Cell Res. 155, 435-447.
Dellagi, K., Tabilio, A., Portier, M.-M., Vainchenker, W., Castaigne, S.,
Guichard, J., Breton-Gorius, J. and Brouet, J.-C. (1985). Expression of
vimentin intermediate filament cytoskeleton in acute nonlymphoblastic
leukemias. Blood 65, 1444-1452.
Djabali, K., Portier, M.-M., Gros, F., Blobel, B. and Georgatos, S. D.
(1991). Network antibodies identify nuclear lamin B as a physiological
attachment site for peripherin intermediate filaments. Cell 64, 109-121.
Fey, E. G., Wan, K. M. and Penman, S. (1984). Epithelial cytoskeletal
framework and nuclear matrix-intermediate filament scaffold: three
dimensional organization and protein composition. J. Cell Biol. 98, 19731984.
Franke, W. W. (1971). Relationships of nuclear membranes with filaments and
microtubules. Protoplasma 73, 263-292.
Franke, W. W., Schmid, E., Osborn, M. and Weber, K. (1978). Different
intermediate-sized filaments distinguished by immunofluorescence
microscopy. Proc. Nat. Acad. Sci. USA 75, 5034-5038.
Gawlitta, W., Osborn, M. and Weber, K. (1981). Coiling of intermediate
filaments induced by the microinjection of a vimentin-specific antibody does
not interfere with locomotion and mitosis. Eur. J. Cell Biol. 26, 83-90.
Georgatos, S. D. and Blobel, G. (1987). Two distinct attachment sites for
vimentin along the plasma membrane and the nuclear envelope in avian
erythrocytes: a basis for vectorial assembly of intermediate filaments. J. Cell
Biol. 105, 105-115.
1606 A. J. Sarria and others
Georgatos, S. D., Weber, K., Geisler, N. and Blobel, G. (1987). Binding of
two desmin derivatives to the plasma membrane and the nuclear envelope of
avian erythrocytes, evidence for a conserved site-specificity in intermediate
filament-membrane interactions. Proc. Nat. Acad. Sci. USA 84, 6780-6784.
Georgatos, S. D. and G. Blobel, G. (1988). Lamin B constitutes an
intermediate filament attachment site at the nuclear envelope. J. Cell Biol.
105, 117-125.
Gerace, L. and Burke, B. (1988). Functional organization of the nuclear
envelope. Annu. Rev. Cell Biol. 4, 335-374.
Geuens, G., De Brabander, M., Nuydens, R. and Demay, J. (1983). The
interaction between microtubules and intermediate filaments in cultured cells
treated with taxol and nocodazole. Cell Biol. Int. Rep. 7, 35-47.
Giese, G. and Traub, P. (1986). Induction of vimentin synthesis in mouse
myeloma cells MPC-11 by 12-O-tetradecanoylphorbol-13-acetate. Eur. J.
Cell Biol. 40, 266-274.
Gill, S. R., Wong, P. C. and Cleveland, D. W. (1990). Assembly properties of
dominant and recessive mutations in the small mouse neurofilament (NF-L)
subunit. J. Cell Biol. 111, 2005-2020.
Goldman, R. D., Goldman, A., Green, K., Jones, J., Lieska, N. and Yang,
H.-K. (1985). Intermediate filaments: possible functions as links between the
nucleus an cell surface. Ann. NY Acad. Sci. 455, 1-17.
Goldman, R. D., Goldman, A. E., Green, K. J., Jones, J. C. R., Jones, S. M.
and Yang H.-Y. (1986). Intermediate filament networks: organization and
possible functions of a diverse group of cytoskeletal elements. J. Cell Sci.
Suppl. 5, 69-97.
Graham, F. L. and van der Eb, A. J. (1973). A new technique for the assay of
infectivity of human adenovirus 5 DNA. Virology 52, 456-467.
Guilly, M. N., Bensussan, A., Bouge, J. F., Bornens, M. and Courvalin, J. C.
(1987). A human T lymphoblastic cell line lacks lamins A and C. EMBO J. 6,
3795-3799.
Hedberg, K. K. and Chen, L. B. (1986). Absence of intermediate filaments in
a human adrenal cortex carcinoma derived cell line. Exp. Cell Res. 163, 509517.
Hennekes, H., Kuhn, S. and Traub, P. (1990). Coding sequence and flanking
regions of the mouse vimentin gene. Mol. Gen. Genet. 221, 33-36.
Katsuma, Y., Swierenga, S. H. H., Marceau, N. and French, S. W. (1987).
Connections of intermediate filaments with the nuclear lamina and the cell
periphery. Biol. Cell 59, 193-204.
Klymkowsky, M. W. (1981). Intermediate filaments in 3T3 cells collapse after
the intracellular injection of a monoclonal anti-intermediate filament
antibody. Nature 291, 249-251.
Klymkowsky, M. W. (1982). Vimentin and keratin intermediate filament
systems in cultured PtK2 epithelial cells are interrelated. EMBO J. 1, 161165.
Klymkowsky, M. W., Miller, R. H. and Lane, E. B. (1983). Morphology,
behavior, and interaction of cultured epithelial cells after the antibodyinduced disruption of keratin filament organization. J. Cell Biol. 96, 494-509.
Klymkowsky, M. W. (1988). Metabolic inhibitors and intermediate filament
organization in human fibroblasts. Exp. Cell Res. 174, 282-290.
Klymkowsky, M. W., Bachant, J. B. and Domingo, A. (1989). Functions of
intermediate filaments. Cell Motil. Cytoskel. 14, 309-331.
Kouklis, P. D., Merdes, A., Papamarcaki, T., and Georgatos, S. D. (1993).
Transient arrest of 3T3 cells in mitosis and inhibition of nuclear lamin
reassembly around chromatin induced by anti-vimentin antibodies. Eur. J.
Cell Biol. 62, 224-236.
Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature 227, 680-685.
Lebel, S., Lampron, C., Royal, A. and Raymond, Y. (1987). Lamins A and C
appear during retinoic acid-induced differentiation of mouse embryonal
carcinoma cells. J. Cell Biol. 105, 1099-1104.
Lehto, V.-P., Virtanen, I. and Kurki, P. (1978). Intermediate filaments
anchor the nuclei in nuclear monolayers of cultured human fibroblasts.
Nature 272, 175-177.
Lilienbaum, A., Legagneux, U., Portier, M.-M., Dellagi, K. and Paulin, D.
(1986). Vimentin gene expression in human lymphocytes and in Burkitt’s
lymphoma cells. EMBO J. 5, 2809-2814.
Lin, J. J. C. and Feramisco, J. R. (1981). Disruption of the in-vivo distribution
of intermediate filaments in fibroblasts through the microinjection of a
specific monoclonal antibody. Cell 24, 185-193.
Lu, X., and Lane, B. (1990). Use of retrovirus vectors to express defective
keratins in non-epithelial cells: two types of binding sites involved in keratin
polymerization. Cell 62, 681-696.
Magin, T. M., Hatzfeld, M. and Franke, W. W. (1987). Analysis of
cytokeratin domains by cloning and expression of intact and deleted
polypeptides in Escherichia coli. EMBO J. 6, 2607-2615.
McKeon, F. D., Tuffanelli, D. L. Fukuyama, K. and Kirschner, M. W.
(1983). Autoimmune response directed against conserved determinants of
nuclear envelope proteins in a patient with linear scleroderma. Proc. Nat.
Acad. Sci. USA 80, 4374-4378.
McKeon, F. D., Kirschner, M. W. and Caput, D. (1986). Homologies in both
primary and secondary structure between nuclear envelope and intermediate
filament proteins. Nature 319, 463-468.
Mollenhauer, H. H. (1964). Plastic embedding mixture for use in electron
microscopy. Stain Technol. 39, 111-114.
Moscinski, L. M. and Evans, R. M. (1987). Changes in the organization of
vimentin-type intermediate filaments during retinoic acid induced
differentiation of embryonal carcinoma cells. In International Symposium on
the Cytoskeleton in Differentiation and Development (ed. J. Arechaga and R.
Maccioni), pp. 267-279. ICSU Press, London.
Ngai, J., Bond, V. C., Wold, B. J. and Lazarides, E. (1987). Expression of
transfected vimentin genes in differentiating murine erythroleukemia cells
reveals divergent cis-acting regulation of avian and mammalian vimentin
sequences. Mol. Cell. Biol. 7, 3955-3970.
O’Farrell, P. H. (1975). High resolution two-dimensional electrophoresis of
proteins. J. Biol. Chem. 250, 4007-4021.
Osborn, M. and Weber, K. (1987). Cytoplasmic intermediate filament
proteins and the nuclear lamin-A, lamin-B and lamin-C share the IFA
epitope. Exp. Cell Res. 170, 195-203.
Parker, B. A. and Strak, G. R. (1979). Regulation of simian virus 40
transcription: sensitive analysis of the RNA species present early in
infections by virus or viral DNA. J. Virol. 31, 360-369.
Paulin-Levasseur, M., Scherbarth, A., Traub, U. and Traub, P. (1988).
Lack of lamins A and C in mammalian hemopoietic cell lines devoid of
intermediate filament proteins. Eur. J. Cell Biol. 47, 121-131.
Paulin-Levasseur, M., Giese, G., Scherbarth, A. and Traub, P. (1989a).
Expression of vimentin and nuclear lamins during the in vitro differentiation
of human promyelocytic leukemia cells HL-60. Eur. J. Cell Biol. 50, 453461.
Paulin-Levasseur, M., Scherbarth, A., Giese, G., Roser, K., Bohn, W. and
Traub, P. (1989b). Expression of nuclear lamins in somatic cells lacking
cytoplasmic intermediate filament proteins. J. Cell Sci. 92, 361-370.
Pruss, R. M., Mirsky, R. and Raff. M. (1981). All classes of intermediate
filaments share a common antigenic determinant defined by a monoclonal
antibody. Cell 27, 419-428.
Raats, J. M. H., Henderik, J. B. J., Verdijk, M., van Oort, F. L. G., Gerards,
W. L. H. Ramaekers, F. C. S. and Bloemendal, H. (1991). Assembly of
carboxy-terminally deleted desmin in vimentin-free cells. Eur. J. Cell Biol.
56, 84-103.
Sangiorgi, F., Woods, C. M. and Lazarides, E. (1990). Vimentin
downregulation is an inherent feature of murine erythropoiesis and occurs
independently of lineage. Development 110, 85-96.
Sarria, A. J., Nordeen, S. K. and Evans, R. M. (1990). Regulated expression
of vimentin cDNA in cells in the presence and absence of a pre-existing
vimentin filament network. J. Cell Biol. 111, 553-566.
Sarria, A. J., Panini, S. R. and Evans, R. M. (1992). A functional role for
vimentin intermediate filaments in the metabolism of lipoprotein-derived
cholesterol in human SW-13 cells. J. Biol. Chem. 267, 19455-19463.
Sarria, A. J. (1993). Effect of vimentin cytoplasmic intermediate filaments on
nuclear morphology and intracellular cholesterol homeostasis in SW-13 cell
lines. PhD thesis, University of Colorado Health Sciences Center, Denver.
Sinensky, M., Beck, L. A., Leonard, S. and Evans, R. M. (1990). Differential
inhibitory effects of lovastatin on protein isoprenylation and sterol synthesis.
J. Biol. Chem. 265, 19937-19941.
Skalli, O., and Goldman, R. D. (1991). Recent insights into the assembly,
dynamics, and function of intermediate filament networks. Cell Motil.
Cytoskel. 19, 67-79.
Stewart, C. and Burke, C. B. (1987). Teratocarcinoma stem cells and early
mouse embryos contain only a single major lamin polypeptide closely
resembling lamin B. Cell 51, 383-392.
Summerhayes, I. C., Wong, D. and Chen, L. B. (1983). Effects of
microtubules and intermediate filaments on mitochondrial distribution. J.
Cell Sci. 61, 87-105.
Tolle, H. G., Weber, K. and Osborn, M. (1986). Microinjection of
monoclonal antibodies to vimentin, desmin, and GFA in cells which contain
more than one IF type. Exp. Cell Res. 162, 464-474.
Traub, U. E., Nelson, W. J. and Traub, P. (1983). Polyacrylamide gel
electrophoretic screening of mammalian cells cultured in vitro for the
Intermediate filaments and nuclear morphology 1607
presence of the intermediate filament protein vimentin. J. Cell Sci. 62, 129147.
Traub, P. (1985). Intermediate Filaments. A Review. Springer-Verlag, Berlin.
Trevor, K. (1990). Disruption of keratin filaments in embryonic epithelial cell
types. New Biol. 2, 1004-1014.
van den Heuvel, R. M. N., van Eys, G. J. J. M., Raemakers, F. C. S., Quax,
W. J., Vree Egberts, W. T. M., Schaart, G, Cuypers, H. T. M. and
Bloemendal, H. (1987). Intermediate filament formation after transfection
with modified hamster vimentin and desmin genes. J. Cell Sci. 88, 475-482.
Vassar, R., Coulombe, P. A., Degenstein, L., Albers, K. and Fuchs, E.
(1991). Mutant keratin expression in transgenic mice causes marked
abnormalities resembling a human genetic skin disease. Cell 64, 365-380.
Venetianer, A., Schiller, D. L., Magin, T. and Franke, W. W. (1983).
Cessation of cytokeratin expression in a rat hepatoma cell line lacking
differentiated functions. Nature 305, 730-733.
Virtanen, I., Kurkinen, M. and Lehto, V.-P. (1979). Nucleus-anchoring
cytoskeleton in chicken red blood cells. Cell Biol. Int. Rep. 3, 157-162.
Vorburger, K., Lehner, C. F., Kitten, G. T., Eppenberger, H. M. and Nigg,
E. A. (1989). A second higher vertebrate B-type lamin. cDNA sequence
determination and in vitro processing of chicken lamin B2. J. Mol. Biol. 208,
405-415.
Wang, X., Willingale-Theune, J., Shoeman, R. L., Giese, G. and Traub, P.
(1989). Ultrastructural analysis of cytoplasmic intermediate filaments and
the nuclear lamina in the mouse plasmacytoma cell line MPC-11 after the
induction of vimentin synthesis. Eur. J. Cell Biol. 50, 462-474.
Wang, X., and Traub, P. (1991). Resinless section immunogold electron
microscopy of karyo-cytoskeletal frameworks of eukaryotic cells cultured in
vitro. J. Cell Sci. 98, 107-122.
Whytock, S., Moir, R. D. and Stewart, M. (1990). Selective digestion of
nuclear envelopes from Xenopus oocyte germinal vesicles: possible
structural role for the nuclear lamina. J. Cell Sci. 97, 571-580.
Wong, P. C. and Cleveland, D. W. (1990). Characterization of dominant and
recessive assembly defective mutations in mouse neurofilament NF-M. J.
Cell Biol. 111, 1987-2004.
Worman, H. J., Lazarides, I. and Georgatos, S. D. (1988). Nuclear lamina
heterogeneity in mammalian cells. Differential expression of the major
lamins and variations in lamin B phosphorylation. J. Biol. Chem. 263, 1213512141.
Zackroff, R. V. and Goldman, R. D. (1979). In vitro assembly of intermediate
filaments from baby hamster kidney (BHK-21) cells. Proc. Nat. Acad. Sci.
USA 76, 6226-6230.
(Received 21 January 1994 - Accepted 2 March 1994)