2925
Journal of Cell Science 110, 2925-2934 (1997)
Printed in Great Britain © The Company of Biologists Limited 1997
JCS9684
HSF1 granules: a novel stress-induced nuclear compartment of human cells
José J. Cotto, Susan G. Fox and Richard I. Morimoto*
Department of Biochemistry, Molecular Biology, and Cell Biology, Rice Institute for Biomedical Research, Northwestern University,
Evanston, IL 60208, USA
*Author for correspondence (e-mail: r-morimoto@nwu.edu)
SUMMARY
Heat shock factor 1 (HSF1) is the ubiquitous stress-responsive transcriptional activator which is essential for the
inducible transcription of genes encoding heat shock
proteins and molecular chaperones. HSF1 localizes within
the nucleus of cells exposed to heat shock, heavy metals, and
amino acid analogues, to form large, irregularly shaped,
brightly staining granules which are not detected during
attenuation of the heat shock response or when cells are
returned to their normal growth conditions. The kinetics of
detection of HSF1 granules parallels the transient induction
of heat shock gene transcription. HSF1 granules are also
detected using an HSF1-Flag epitope tagged protein or a
chimeric HSF1-green fluorescent protein which reveals that
these nuclear structures are stress-induced and can be
detected in living cells. The spatial organization of HSF1
granules in nuclei of stressed cells reveals that they are novel
nuclear structures which are stress-dependent and provides
evidence that the nucleus undergoes dynamic reorganization in response to stress.
INTRODUCTION
of the inert monomer remains unresolved, however, HSF1
trimers appear within minutes of activation and can be detected
bound to DNA shortly thereafter. The rapid induction of the
heat shock response suggests that the relocalization of HSF1
must involve a dynamic process. Yet, unlike most proteins
which either translocate constitutively to the nucleus or exhibit
a constant pattern of localization, the mechanisms involving
HSF1 are likely to be distinct. Previous studies on the
immunolocalization of the HSFs have detected the appearance
of granules which seem to correlate with activation (Sarge et
al., 1993; Nakai et al., 1993, 1995; Nakai et al., 1997). In this
study we examine the cell biological properties of HSF1 using
a collection of monoclonal antibodies which specifically detect
the appearance of HSF1 granules within the nucleus of stressed
cells. The kinetics of induction of HSF1 granules by heat shock
and other stresses which lead to heat shock gene transcription
strongly indicates a role in the stress response.
The cellular response to adverse environmental and physiological conditions such as heat shock, exposure to amino acid
analogs, heavy metals, oxidative stress, anti-inflammatory
drugs, and arachidonic acid leads to the rapid and transient activation of genes encoding heat shock proteins (hsps) and
molecular chaperones (Lindquist, 1986; Morimoto et al., 1990,
1994; Jurivich et al., 1992). Stress-induced transcription is
regulated by a family of heat shock transcription factors (HSF).
In vertebrates, four members of the HSF gene family (HSFs 14) have been characterized (Rabindran et al., 1991; Sarge et al.,
1991; Schuetz et al., 1991; Nakai and Morimoto, 1993; Nakai
et al., 1997). The co-expression of multiple HSFs and characterization of regulatory conditions has revealed that different
members of the HSF family mediate the response to distinct
forms of cellular stress. Consistent with this, HSF1 responds
to the classical inducer of the heat shock response, HSF2 is
activated during embryogenesis, spermatogenesis and
erythroid differentiation, HSF3 functions as a high temperature
activator, and HSF4 has properties of a negative regulator of
heat shock gene expression (Sistonen et al., 1992, 1994; Sarge
et al., 1993, 1994; Nakai et al., 1995, 1997).
Under normal conditions of cell growth, HSF1 is maintained
in an inert non-DNA binding state which undergoes reversible
oligomerization to a DNA binding competent trimer in stressed
cells (reviewed by Morimoto et al., 1994 and Wu, 1995). Two
distinct mechanisms involving negative regulatory domains
and constitutive phosphorylation at serine residues participate
to maintain HSF1 in its inert state (Green et al., 1995; Shi et
al., 1995; Kline and Morimoto, 1997; Knauf et al., 1996). How
the stress signal is transduced and results in the de-repression
Key words: Subnuclear structure, Heat shock, Transcription, Stress
response
MATERIALS AND METHODS
Cell culture
Human HeLa, A431 and HOS cells were grown in Dulbecco’s
modified Eagle’s medium (DMEM) with 5% fetal calf serum (FCS).
Primary epithelial and fibroblast cells were grown in DMEM with
10% FCS supplemented with essential and non-essential amino acids,
vitamins, and buffered with 1 M Hepes, pH 7.4. HeLa S3 cells were
grown in Joklik’s medium with 5% calf serum. Cell growth, heat
shock conditions and exposure to heavy metals, amino acid analogs
and anti-inflammatory drugs were as described before (Mosser et al.,
1988; Jurivich et al., 1992; Sarge et al., 1993).
Antibodies and indirect immunofluorescence
The subcellular localization of HSF1 was determined using a panel of
2926 J. J. Cotto, S. G. Fox and R. I. Morimoto
anti-HSF1 rat monoclonal antibodies (10H8, 10H4 and 4B4) generated
from rat hybridoma cell lines using purified recombinant mouse HSF1
as the antigen. The monoclonal antibodies were characterized for
specificity to HSF1 by ELISA and western blot analysis using purified
recombinant mouse HSF1 and total cell extracts from mouse and other
vertebrates (see Table 1). The epitope recognized by each monoclonal
antibody was determined by western blot analysis of a collection of
mouse HSF1 deletion mutants (Shi et al., 1995; Kline and Morimoto,
1997, see Fig. 1). The subcellular localization of the antigens recognized by each antibody was determined by indirect immunofluorescence. The specificity of each antibody for the control monomeric
non-DNA binding form of HSF1 and the active trimer was determined
by immunoprecipitation from extracts of control and heat shocked cells
and by the use of the antibodies for antibody upshift assays. The results
of the characterization of each antibody are summarized in Table 1.
Antibodies that recognized other nuclear structures included mouse
monoclonal anti-splicing factor SC-35 antibody, from Sigma (Catalog
# S-4045), human autoimmune anti-kinetochore antibodies and antiNuMA, monoclonal anti-nuclear lamin A and B antibody (provided by
Dr Robert Goldman, Northwestern University Medical School), antip80-coilin rabbit polyclonal antibody (provided by Dr Angus Lamond,
University of Dundee and Dr Edward Chen, The Scripps Research
Institute), anti-fibrillarin (provided by Dr David Spector, Cold Spring
Harbor) and anti-PML monoclonal antibody (provided by Dr Luitzen
de Jong, University of Amsterdam). The antiserum titer was established
by sequential dilution to a range of 1:100 to 1:300 before use. Horseradish peroxidase (HRP)-conjugated goat anti-rat IgG was obtained
from Pierce (catalog #31475G) and goat anti-rabbit IgG antibody was
obtained from Promega (catalog # W4011). Texas Red and fluorescein
(FITC)-conjugated goat anti-rat IgG (catalog # 712-095-153, 712-075153), goat anti-rabbit IgG (catalog # 711-075-152) and goat anti-mouse
IgG antibodies (catalog # 715-095-151) were obtained from Jackson
Immunoresearch and mouse anti-bromo-uridine triphosphate (BrUTP)
monoclonal antibody was obtained from Sigma (catalog # B2531).
For immunofluorescence analysis, adherent cells on coverslips were
washed in 1× PBS, fixed for 10 minutes with 2% paraformaldehyde in
1× PBS at room temperature, washed twice with 1× PBS, and permeabilized with 0.1% Triton X-100. The permeabilized cells were
washed twice with PBS and incubated for 1 hour with a blocking
solution consisting of 1% bovine serum albumin (BSA) in PBS at room
temperature prior to incubation with antibodies. Fixed cells were
incubated for 1 hour at 37°C with either rat monoclonal or rabbit polyclonal anti-HSF1 antibody at dilutions of 1:100 and 1:300, respectively. To study the co-localization of HSF1 granules with other known
nuclear structures, a mix of rat monoclonal anti-HSF1(1:100) with
either anti-Brdu (1:100), anti-SC-35 (1:100), anti-kinetochore (1:200),
anti-p80 coilin (1:100), anti-fibrillarin (1:1), anti-lamin A and B (1:00),
anti-NuMA (1:100) and anti-PML (1:5) was prepared and cells were
incubated for 1 hour at 37°C prior to detection. The antibodies were
detected using Texas Red or FITC-conjugated goat anti-rat and goat
anti-mouse and the staining pattern was analyzed by conventional epifluorescence microscopy on a Zeiss Axiophot microscope or by laser
confocal microscopy. To establish the spatial organization of HSF1
granules and the relationship of HSF1 granules with other nuclear
structures, Z-sections of the stained cells were collected in a confocal
laser scanning microscope (Zeiss) and the data were analyzed with the
program NIH Image, Vers. 1.60 for Macintosh.
Nuclease treatment
Cells permeabilized with 0.05% Triton X-100 were treated with either
RNAse A (100 mg/ml) or DNAse I (5 units/30 ml of RNAse free
DNAse) from Boehringer Mannheim at 37°C for pre-determined
times and washed with PBS prior to fixation and immunostaining.
Cell extraction and biochemical analysis
The procedure for in situ sequential fractionation of heat shocked cells
was performed as described (He et al., 1990; Bissoto et al., 1995) with
some variations. Briefly, heat shocked (42°C) HeLa S3 cells were
extracted in suspension with centrifugation steps (600 g, 3 minutes)
between treatments. Supernatants were collected for SDS-PAGE and
glycerol gradient analysis. Cells were first extracted in low ionic buffer
(20 mM Hepes, pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl2,
0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM DTT) for 5 minutes at 4°C.
After a brief wash with 1× PBS, cells were extracted with cytoskeleton
buffer (10 mM Pipes, pH 6.8, 100 mM NaCl, 300 mM sucrose, 3 mM
MgCl2, 1 mM EGTA, 0.5% Triton X-100, 4 mM vanayl riboside
complex, 1 mM PMSF) for 3 minutes at 4°C. After washing with 1×
PBS, DNA was digested with 25 units/ml DNAse I for 30 minutes at
25°C in digestion buffer (essentially the same as cytoskeleton buffer but
50 mM NaCl and 0.05% Triton X-100). Chromatin was then removed
by three 10 minute washes with 0.25 mM ammonium sulfate in
digestion buffer to yield the nuclear matrix intermediate filament
structure (He et al., 1990). An additional high salt treatment was applied
by washing the nuclear matrix with 2 M NaCl in digestion buffer. As a
last step the pellets were boiled in SDS sample buffer for SDS-PAGE
analysis.
Construction and expression of HSF1-Flag and HSF1-GFP
fusion proteins
To generate the carboxy-terminal epitope tagged mHSF1-Flag fusion
protein, an oligonucleotide encoding an eight amino acid peptide (nDYKDDDDK-c) recognized by the monoclonal antibody Anti•Flag
M2 (IBI Flag System, Kodak) was cloned at the 3′ end of the mouse
HSF1 cDNA (Sarge et al., 1991) in the eukaryotic expression vector
pCDNA3 (InvitroGen), and verified by sequencing across the region.
The construct corresponding to HSF1-green fluorescent protein
(HSF1-GFP) fusion protein was made using the eukaryotic expression
vector pEGFP-N1 (Clontech) to clone the green fluorescent protein
(GFP) coding sequence to the 3′ end of mouse HSF1 cDNA.
For the expression of the fusion proteins, HeLa cells at a density
of 40 to 50% confluence were transfected with 20 mg of DNA per 10
cm plate. Plasmid DNA was combined with 250 mM CaCl2 in a 500
ml final volume. After chilling on ice, the DNA-Ca2+ solution was
added dropwise to 500 ml of 2× Hepes (N-2-hydroxyethylpiperazine-
Table 1. Summary of the imunological characteristic of HSF1 specific monoclonal antibodies
Western blot
Immunoprecipitation
Supershift assay
Immunofluorescence
Antibody
Epitope*
37°C
42°C
37°C
42°C
37°C
42°C
37°C
42°C
10H8
10H4
378-395
295-378
+†(1,2,3,4)
+†(1,3,4)
+†(1,2,3,4)
+†(1,3,4)
+†(1,4)
−
+†(1,4)
−
N/A‡
N/A
+†(1,4)
+†(1,4)
+†(1,2,3,4)
+†(1,2,3,4)
+†(1,2,3,4)
+†(1,2,3,4)
4B4
425-439
+†(1,2,3,4)
+†(1,2,3,4)
+†(1,4)
+†(1,4)
N/A
+†(1,4)
+†(1,2,3,4)
+†(1,2,3,4)
*Numbers correspond to the minimal amino acid region containing the epitopes for anti-HSF1 monoclonal antibodies as detected by western blot analysis of
recombinant mouse HSF1 deletion mutants (Shi et al., 1995; Kline and Morimoto, 1997).
†HSF1 species reactivity: 1, human; 2, monkey; 3, rat; 4, mouse.
‡Not applicable.
HSF1 granules 2927
N′-2-ethanesulfonic acid)-buffered saline (pH 7.06; 2× HeBes = 280
mM NaCl, 1.5 mM Na2HPO4, 50 mM Hepes). After 20 minutes at
room temperature the precipitate was added to the cells dropwise and
allowed to settle on the cells for 6 to 8 hours at 37°C. The plates were
then removed from the incubator and washed twice with 10 ml of 1×
phosphate-buffered saline (PBS)-1 mM ethylene glycol-bis(Baminoethyl ether)N,N,N′,N′-tetraacetic acid (EGTA) and replaced
with fresh medium for 48 hours before analysis.
Gel mobility shift assay and western blot analysis of HSF1
HSF1 DNA-binding activity was analyzed using the gel mobility shift
assay as described previously (Mosser et al., 1988) using a 32P-labeled
double-stranded synthetic HSE containing four inverted nGAAn
repeats (Sarge et al., 1991). Western blot analysis was performed
using whole cell extracts and rat monoclonal anti-HSF1 antibody
(10H8). The immune complexes were analyzed using the ECL
detection system (Amersham).
RESULTS
HR C
212
4B4
HR A,B
120 132
10H4
DNB
1 16
10H8
HSF1 localizes to discrete nuclear granules upon
heat shock
Upon exposure of HeLa cells to heat shock, HSF1 relocalizes
to form brightly staining nuclear foci or granules which were
detected using polyclonal antisera (Sarge et al., 1993). To
further characterize these granules, we prepared a collection of
rat monoclonal antibodies (10H8, 10H4 and 4B4) that specifically recognized HSF1 in cells from mouse, human and other
vertebrate species. Although the epitopes for each antibody
mapped to different regions of HSF1 (see Fig. 1 and Table 1),
both the inactive and active trimeric form of HSF1 can be
detected by various immunological assays including, immunoprecipitation and gel mobility shift-supershift assays (data not
shown, see Table 1). Indirect immunofluorescence analysis
using any of these monoclonal antibodies in heat shocked cells
+
+
+
+
+
+
+
+
-
+
+
+
-
-
+
-
-
+
+
-
-
-
-
-
503
378 407
mHSF1
227
451
I
288
425
II
295
498
III
395
503
IV
425
503
V
378
407
VI
439
503
4B4
10H8
VII
10H4
revealed the presence of HSF1 granules for which antibody
10H8 gave typical results (Fig. 2B).
To examine whether the immunofluorescence staining
pattern detected by these monoclonal antibodies was either the
consequence of how the cells were prepared for indirect
immunofluorescence or an unusual feature of these antibody
reagents, we constructed a Flag epitope-tagged mouse HSF1
gene (mHSF1-Flag) and a green fluorescent protein (GFP)HSF1 chimeric gene for transient expression studies in HeLa
cells. The endogenous and heterologous HSF1 proteins were
detected by double-label immunofluorescence using rat monoclonal antibody 10H8 and the anti-Flag antiserum or the
intrinsic fluorescence of GFP, respectively. The mHSF1-Flag
protein (Fig. 3B) co-localizes with human HSF1 (Fig. 3A) as
a general diffuse nuclear staining pattern under control conditions (Fig. 3C). Upon heat shock, mHSF1-Flag (Fig. 3E)
localizes to the same granules detected with the monoclonal
antibodies which recognize human HSF1 (Fig. 3D and F).
Similar results of co-localization were observed using HSF1GFP (Fig. 4). Whereas GFP alone is distributed in a diffuse
pattern throughout the cell under control (Fig. 4A) and heat
shocked conditions (Fig. 4B), mHSF1-GFP is primarily
nuclear in control cells (Fig. 4C) and upon heat shock relocalizes to form HSF1 granules (Fig. 4D). These results establish
that the heat shock induced formation of HSF1 granules are a
specific property of HSF1, they can be detected with an epitope
tag and in living stressed cells.
We next examined whether HSF1 granules are found in other
human tissue culture cell lines. Primary fibroblasts and epithelial cells were exposed to various heat shock temperatures and
examined by indirect immunofluorescence. Brightly staining
HSF1 granules were readily detected upon exposure of either
primary cell to heat shock with the majority of cells containing two large foci and occasional smaller speckles (Fig. 5B and
D). The optimal temperatures required to detect HSF1 granules
in primary human cells was higher (43-45°C) than required in
transformed human cells (see companion paper in this issue:
Jolly et al., 1997). Examination of the HSF1 staining pattern
in two other human transformed cell lines, HOS (hyperdiploid
osteosarcoma cell line) and A431 (hypotetraploid epidermal
carcinoma cell line) revealed that HOS cells have an average
of 4 to 5 granules per nucleus (Fig. 5F) and HeLa (hypotetraploid cervical carcinoma cell line) and A-431 cells contain
Fig. 1. Schematic representation of various mouse HSF1 deletion
mutants and a summary of the pattern of recognition detected by
anti-HSF1 monoclonal antibodies 10H8, 10H4 and 4B4. E. coli
whole cell extracts expressing the recombinant mouse HSF1 deletion
mutants were analyzed by western blot and the minimal peptide
regions recognized by the different antibodies are indicated.
Conserved structural motifs corresponding to DNA binding domain
(DNB) and heptad repeats A,B and C (HR A,B and C) are shown.
Fig. 2. HSF1 nuclear granules are formed upon heat shock in HeLa
cells. Cells cultured at 37°C or heat shock at 42°C were subjected to
immunofluorescence analysis using anti-HSF1 rat monoclonal
antibody 10H8 (A and B). Bar, 5 µm.
2928 J. J. Cotto, S. G. Fox and R. I. Morimoto
HSF1-Flag
Fig. 3. Immunofluorescence analysis of
HeLa cells transfected with mouse HSF1Flag expression vector. Transfected HeLa
cells were cultured at 37°C and stained with
anti-HSF1 (A) or anti-FLAG (B) antibodies.
(C) Co-localization of HSF1 and mHSF1FLAG. (D-F) Cells incubated at 42°C heat
shock and treated as in A-C. Bar, 5 µm.
Heat Shock
Control
HSF1
an average of 7 granules per nucleus (Fig. 5H and J). The
number of HSF1 granules per cell is relatively constant in any
particular cell line with fewer granules being detected in
primary cells than in transformed cell lines. These results and
those of Jolly et al. (1997) suggest a possible relationship
between the number of granules and chromosomal ploidy.
The kinetics of HSF1 granule formation parallels the
activation of HSF1 in stressed cells
The activation of HSF1 is associated with a series of rapidly
occurring events including oligomerization of the non-DNA
binding monomer to the DNA binding trimer, inducible serine
phosphorylation, and transcriptional induction of heat shock
genes (Sorger and Pelham, 1988; Baler et al., 1993; Sarge et
al., 1993; Cotto et al., 1996). During continuous exposure to
heat shock, HSF1 activity attenuates as reflected by the loss of
transcriptional activity, dephosphorylation, and conversion of
trimers to monomers (Abravaya et al., 1991a,b; Sarge et al.,
1993). Therefore, we examined whether the appearance of
HSF1 granules is temporally associated with its activity as a
transcriptional activator.
Within 30 minutes of heat shock, HSF1 granules ranging in
size from 0.5 to 1.5 µm were detected in 80-90% of HeLa cells
(Fig. 6A), and by 60 minutes of heat shock, 95% of the cells
exhibited brightly staining HSF1 granules. Up through two
hours of heat shock, HSF1 granules were ubiquitous in all
cells. Analysis of the size distribution of the granules in HeLa
cells heat shocked at 42°C for 2 hours reveals that they can be
described as two populations of which 60% of the foci correspond to smaller (0.5 to 1.5 µm) brightly staining granules or
speckles and the remaining are larger (1.5 to 2.5 µm) clustered
or ring-like granular structures (Fig. 7A). The majority (55%)
of the cells contained an average of 7 HSF1 granules, although
there was substantial cell-to-cell variation (Fig. 7B). During
continuous exposure to heat shock, both the fluorescence
intensity and the numbers of HSF1 foci increased. Comparison to the level of HSF1 DNA binding activity (Fig. 6B)
reveals that the appearance of HSF1 granules correlates
closely with both the acquisition of HSF1-DNA binding
activity and the inducibly phosphorylated state of HSF1 (Fig.
6B and C). After 2 hours of continuous heat shock, the fraction
of cells which exhibit HSF1 granules rapidly declines to 5%
of the population; likewise HSF1 DNA binding attenuates and
is dephosphorylated to the control state (Fig. 6B and C).
The activation of HSF1 is a multi-step process which involves
the stable appearance of intermediate states (Jurivich et al.,
1992; Lee et al., 1995; Cotto et al., 1996). To examine whether
the appearance of HSF1 granules reflects the formation of HSF1
DNA binding trimers alone or requires complete activation to
Fig. 4. mHSF1-GFP fusion protein localizes to nuclear granules in
heat shocked HeLa cells. HeLa cells at 37°C or heat shocked at 42°C
were transfected with the GFP vector (A and B) or the mHSF1-GFP
vector (C,D) and visualized by confocal microscopy. Bar, 5 µm.
HSF1 granules 2929
Fig. 5. Subcellular localization of HSF1 in various primary and transformed human cells. Cells either at control or 42°C heat shock conditions
were analyzed by immunofluorescence using rat monoclonal antibody 10H8. Human primary fibroblasts (A,B), epithelial cells (C,D), human
HOS cells (E,F), HeLa (G,H), and A431 (I,J) cells were examined. Bar, 5 µm.
Fig. 6. Kinetics of HSF1 granule formation and comparison with DNA binding activity and inducible phosphorylation during heat shock.
(A) Immunofluorescence analysis in HeLa cells prior to heat shock (0 minutes) or after incubation at 42°C for 30, 60, 120, 180 and 240 minutes
using HSF1 rat monoclonal antibody 10H8. (B) Kinetics of HSF1 DNA binding activity measured using gel mobility shift analysis, and (C)
western blot analysis of whole cell extracts from samples indicated in B. The slower mobility of HSF1 in heat shocked cells is due to inducible
serine phosphorylation. Bar, 5 µm.
2930 J. J. Cotto, S. G. Fox and R. I. Morimoto
ditionally, these results reveal that HSF1 granules are not the
consequence of heat shock-induced aggregation of HSF1.
Fig. 7. Quantitative analysis of the number and size of HSF1
granules in HeLa cells. (A) The average diameter was calculated for
200 granules using the program NIH image. In HeLa cells, HSF1
granules can be divided in two sub-populations based on their
average size; small granules (~0.5 to 1.6 µm) and large granules
(~1.6 to 3 µm). (B) The number of HSF1 granules was established
by analysis of 100 nuclei. The mean number of HSF1 granules/cell is
6.8±2.4.
the transcriptionally competent trimer state, HeLa cells were
exposed to three inducers of HSF1 activity. Sodium salicylate
induces HSF1 trimers which are nuclear-localized, transcriptionally inert, and not-inducibly phosphorylated (Jurivich et al.,
1992; Cotto et al., 1996), the amino acid analogue azetidine
induces a transcriptionally active form of HSF1 which is not
inducibly phosphorylated, and the heavy metal cadmium
induces a transcriptionally active and inducibly phosphorylated
form of HSF1 (Sarge et al., 1993) (Fig. 8). Although each of
these conditions activate equivalent levels of HSF1 DNA
binding activity (Fig. 8A), HSF1 granules were only detected
in azetidine or cadmium treated cells and not in salicylate
treated cells (Fig. 8B). These results reveal, that the appearance
of HSF1 granules is a reliable visual indicator of the transcriptional activity of HSF1 and that HSF1 granules are induced by
other stresses which activate the heat shock response. Ad-
HSF1 granules are novel sub-nuclear structures
To determine whether HSF1 granules represent a novel nuclear
compartment or correspond to previously characterized subnuclear structures, we used double immunofluorescence with a
number of antibodies to nuclear antigens and analysis by laser
confocal microscopy. The fluorescence labeling patterns of each
antibody and 8 to 10 horizontal optical sections of each field were
scanned from top to bottom of the cell and the results of stacked
images are presented. The data in Fig. 9 represent the superimposed images of fluorescein (FITC) and Texas Red-coupled
secondary antibody labeling of different primary antibodies. In
Fig. 9A, the sites of DNA replication were marked by incorporation of BrdU prior to heat shock and detected with anti-BrdU
antibody. No co-localization of HSF1 granules with sites of DNA
replication were detected. Likewise, the immunofluorescence
pattern of mitotic cells stained with anti-HSF1 and anti-kinetochore antibody (Fig. 9B), the anti-splicing factor SC-35 antibody
(Fig. 9C), and the coiled body marker anti-p80-coilin antibody
(Fig. 9D) did not reveal co-localization with HSF1 granules. In
this study, no heterogeneity was noticed in the appearance or morphology of the HSF1 granules during the cell cycle.
As previous studies have shown that the nucleolar morphology is affected by heat shock (Welch and Suhan, 1985), we
investigated whether the HSF1 granules would correspond to
an accumulation of the factor into nucleoli. This assumption
was strongly supported by the granular morphology of HSF1
foci which is similar to that of nucleoli detected with a marker
of the dense fibrillar center (Roussel et al., 1993). HSF1 was
detected by immunofluorescence together with a marker of the
dense fibrillar center, the UBF cofactor (upstream binding
factor) using an anti-UBF antibody (Roussel et al., 1993),
however, no co-distribution between HSF1 foci and nucleoli
was observed (Fig. 9E). Likewise, HSF1 granules do not colocalize with PML, a nuclear localized protein involved in
promyelocytic leukemia (Koken et al., 1994; Weis et al., 1994)
or with nuclear matrix proteins, such as nuclear lamins (A or
B) and NuMA (nuclear matrix mitotic apparatus protein) which
have been shown to accumulate into nuclear domains at
specific points in the cell cycle (Moir et al., 1994) (data not
shown). Overall, these results clearly demonstrated that HSF1
accumulates into discrete nuclear substructures which are
distinct from other previously described nuclear granules.
Co-localization experiments carried out in our laboratory
have clearly shown that HSF1 granules do not contain other heat
shock transcription factors (data not shown). Indeed, in earlier
studies, HSF2, HSF3, and HSF4 have been visualized as
nuclear speckles, not granules (Sarge et al., 1993; Sheldon and
Kingston, 1993; Nakai et al., 1995, 1997). Furthermore, these
speckles are detected constitutively and do not correlate with
gene activation and heat shock response. In contrast, the appearance of HSF1 granules correlates with transformation of the
factor from the inert to the active state and with gene activation.
Recently, an increasing number of transcription factors,
including the mineralocorticoid and glucocorticoid receptors,
the haemopoietic factors GATA-1 and -3, and p53, have been
shown to accumulate into nuclear domains (Jackson et al.,
1994; van Steensel et al., 1995, 1996; Elefanty et al., 1996).
Although we have not performed a comprehensive comparison
HSF1 granules 2931
Fig. 8. Effects of different
stress conditions on the
activation of HSF1 and the
formation of HSF1 granules.
(A) Gel mobility shift
analysis of HSF1 DNA
binding activity in whole
cell extracts from HeLa
cells at control (37°C)
conditions or treated with
20 mM salicylate, 5 mM
azetidine, 30 µM CdSO4,
or 42°C heat shock.
(B) Intracellular localization
of HSF1 in HeLa cells
exposed to conditions
indicated in A and stained
with monoclonal anti-HSF1
antibody. Bar, 5 µm.
Control
20 mM salicylate
between HSF1 and each one of these transcription factors,
HSF1 does not co-localize with members of the GATA family
(data not shown) thus ruling out a common mechanism for
compartmentalization of transcription factors.
To assess whether nucleic acids are a component of HSF1
granules, heat shocked cells were permeabilized and incubated
with either DNAse I or RNAse A (Fig. 10). Interestingly, the
number, size, or distribution of HSF1 granules were not
affected despite the substantial reduction in nuclear DNA
observed in DNase I treated cells as detected by staining with
Hoechst dye (Fig. 10A). Likewise, RNAse A treatment did not
have an effect on HSF1 granules (Fig. 10B, a-c). These results
reveal that the general features of the granules are not altered
by depolymerization of nucleic acids. However, given the relatively large size of the HSF1 granules, it is also possible that
their location within the nucleus may be unaffected even if
nucleic acids are an important structural component.
The possible association of HSF1 granules with the nuclear
matrix was examined using biochemical fractionation conditions known to extract the nuclear matrix (He et al., 1990;
Bissoto et al., 1995). The majority of the HSF1 which was
removed by extraction with a low ionic strength buffer (Fig.
11, compare lanes 1 and 2) corresponds to trimeric HSF1; the
remaining HSF1 can be extracted with the non-ionic detergent
Triton X-100 (lane 4). HSF1 was not detected in the insoluble
pellet corresponding to the nuclear matrix (lane 7). The ease
of extraction of HSF1 by low ionic buffer and relatively mild
non-ionic detergents reveal that HSF1 is neither in an insoluble
fraction nor associated specifically with the nuclear matrix.
5 mM azetidine
30 µM CdSO4
Heat shock (42°C)
These results also suggest that HSF1 granules are labile and
readily disrupted even upon gentle lysis of the nucleus.
DISCUSSION
HSF1 granules represent a unique class of subcellular structures which appear transiently in the nucleus of human cells
when heat shock genes are transcriptionally induced and
disappear rapidly during attenuation of the heat shock response
as the transcription of heat shock genes diminishes to control
levels. These results establish HSF1 granules as a novel
dynamic feature of the heat shock response and underscores
the potential for new information on the effects of stress on
nuclear structure.
The detection of HSF1 granules is likely to reflect the
general response to stress as exposure of human cells in culture
to heat shock, cadmium and azetidine gave indistinguishable
results. Furthermore, these results rule out the possibility that
HSF1 granules result from the aggregation of HSF1 at elevated
temperatures, as other stresses which induce these granules are
effective at 37°C. Additionally, the effects of elevated temperatures and other stresses on the biochemical properties of
HSF1 result in the oligomerization to a trimeric DNA binding
state rather than association of HSF1 with the nuclear matrix
or other high molecular sized nuclear structures. Another
observation presented here which links the appearance of
HSF1 granules with the transcriptional activity of heat shock
genes is the absence of HSF1 granules in salicylate-treated
2932 J. J. Cotto, S. G. Fox and R. I. Morimoto
Fig. 9. HSF1 does not co-localize with other known subnuclear structures. Immunofluorescence analysis of double-stained HeLa cells. Green
channel represents HSF1 staining and red channel represents (A) DNA replication sites (anti-Brdu), (B) kinetochores, (C) splicing factor SC35, (D) coiled body and (E) nucleolar dense fibrillar center staining. Bar, 5 µm.
cells. Although salicylate treatment induces HSF1 trimers
which exhibit complete DNA binding activity, this form of
HSF1 is transcriptionally inert. In contrast, stress-inducers
such as heavy metals, amino acid analogues, and heat shock
results in the fully active form of HSF1 which corresponds with
stress-induced granules. Finally, the temporal link in the well
established kinetics of the heat shock response and the rapid
recovery during attenuation parallels precisely the transient
appearance and disappearance of HSF1 granules. Taken
Fig. 10. Analysis of the effects of
nuclease treatment on HSF1
granules. (A) Heat shocked HeLa
cells grown on coverslips were
permeabilized in 0.05% Triton X100 and incubated in DNAse I for 0,
15 or 30 minutes prior to
immunofluorescence analysis with
HSF1 monoclonal antibody 10H8
(a-c). Reduction of Hoechst staining
(d-f) indicates digestion of DNA in
these cells. (B) Same as A, but cells
were treated with RNAse A prior to
immunofluorescence analysis (a-c).
Treatment with RNAse did not
affect Hoechst staining (d-f).
together, these results reveal that HSF1 granules represent a
new component of the heat shock response in human cells. Our
results and those presented in the accompanying paper reveal
that HSF1 granules are in all human cells. The relationship
between the number of granules per nucleus and chromosomal
ploidy noted here and by the accompanying paper (Jolly et al.,
1997) is intriguing and suggests the existence of specific chromosomal targets for HSF1 foci.
The many unique features of HSF1 granules encouraged us
HSF1 granules 2933
Fraction no.
1
2
3
4
5
6
7
69 kDa
Fig. 11. Biochemical fractionation of heat shocked cells. (A) Heat
shocked (42°C) HeLa S3 cell pellets (approximately 2×107
cells/pellet) were either directly solubilized with SDS-sample buffer
to detect total amount of HSF1 in the cell (lane 1) or extracted
sequentially with Buffer C (lane 2), a 1× PBS wash (lane 3), 0.5%
Triton X-100 in cytoskeleton buffer (lane 4), DNAse (25 units/ml)
and 0.25 M ammonium sulfate to remove chromatin (lane 5), a high
salt wash in 2 M NaCl (lane 6) and the insoluble material was
solubilized directly in SDS-sample buffer (lane 7). Extracted material
was dialyzed and lyophilized before solubilization in sample buffer
(lanes 2-6). Extracts were analyzed directly by western blot analysis
using monoclonal antibody 10H8.
to redouble our efforts to ensure that the granules could be
detected by other reagents or methodologies which did not
depend on indirect immunofluorescence or traditional forms of
cell fixation. Two complementary approaches were described in
which an epitope tagged HSF1 allowed the detection of HSF1
granules via an antibody to a heterologous epitope and by direct
visualization of HSF1 granules in living cells using an HSF1GFP fusion protein. Thus, the detection of HSF1 granules is not
an artifact of either the polyclonal or the newly characterized
monoclonal antibodies. The ability to detect HSF1 granules in
living heat shocked cells by direct visualization of HSF1-GFP
chimeric protein also rules out possible artifacts inherent in the
procedures employed for visualization and detection of the antiHSF1 antibodies. Additionally, the latter result reveals that other
proteins (e.g. GFP) can be translocated into HSF1 granules.
Our knowledge on nuclear structure, while still limited, has
grown rapidly in recent years. Most nuclear functions including
replication, transcription, RNA splicing and RNA transport are
localized within the nucleus rather than distributed diffusely
(reviewed by de Jong et al., 1990; van Driel et al., 1991, 1995;
Moen et al., 1995). Among the most prominent sub-nuclear
structures is the nucleolus which organizes events associated
with the transcription and processing of ribosomal RNA (Scheer
and Benavente, 1990; Hernandez-Verdun, 1991). The structure
and function of other nuclear domains are less well characterized, in part because functional compartmentalization is less
apparent. Some nuclear domains for which a function has been
suggested include the small nuclear RNP clusters involved in
pre-mRNA processing (Piñol-Roma et al., 1989; Fu and
Maniatis, 1990; Spector, 1990; Lamond and Carmo-Fonseca,
1993b) and coiled bodies, which may play a role in spliceosome
assembly or recycling and in the metabolism of small nuclear
RNAs (Carmo-Fonseca et al., 1991; Lamond and CarmoFonseca, 1993a; Bohmann et al., 1995). Analysis of the subcellular distribution of certain transcription factors also reveals
a non-homogeneous distribution within the nucleus (Jackson et
al., 1994; van Steensel et al., 1995, 1996; Elefanty et al., 1996).
Yet, for each of these factors, granules or speckles are constitutive and do not appear to correspond with other functionally
characterized sub-nuclear structures. Indeed, other heat shock
factors (HSF2, 3, 4) have been visualized as nuclear speckles
(Sarge et al., 1993; Sheldon and Kingston, 1993; Nakai et al.,
1995, 1997). However, these speckles are present constitutively
and do not correlate with gene activation and heat shock
response, and they are distinct from the HSF1 granules. Only in
the case of HSF1 does the appearance of these granules correlate
with the activation of the factor from the inert to active states.
Despite the cell biological analysis and biochemical characterization presented in this study and the accompanying paper
(Jolly et al., 1997), the role of HSF1 granules in the heat shock
response remains enigmatic. Perhaps this is because the
analysis only encompasses our current expectations of a traditional role for HSF1 in the heat shock response. While there
has been substantial progress to understand the organization
and regulation of HSFs, only a few molecular targets for HSF
binding, corresponding to the Hsp90 and Hsp70 genes, have
been characterized. Others have shown that HSFs are associated with other chromosomal loci and suggested that HSFs
may have roles in the repression of non heat-shock genes
during heat shock (Westwood et al., 1991; Cahill et al., 1996).
Indeed, heat shock causes a complete arrest in the transcription of non-heat shock genes by mechanisms which have yet
to be addressed.
We thank Dr Robert Goldman for the anti-lamin A and B monoclonal antibody and the NuMA autoantibody, Dr Angus Lamond for
the anti-coilin polyclonal antibody, Dr Edward Chen for the anti-p80coilin polyclonal antibody R288, Dr David Spector for anti-fibrillarin
autoantibodies, and Dr Luitzen de Jong for the anti-PML monoclonal
antibody 5E10. We thank Dr Claire Vourc’h and Caroline Jolly for
their valuable discussion during the course of this work and their
comments on the manuscript and Dr Robert Holmgren for advice in
microscopy. These studies were supported by a grant to R.I.M. from
the National Institutes of General Medical Sciences.
REFERENCES
Abravaya, K., Phillips, B. and Morimoto, R. I. (1991a). Heat shock-induced
interactions of heat shock transcription factor and the human hsp70 promoter
examined by in vivo footprinting. Mol. Cell. Biol. 11, 586-592.
Abravaya, K., Phillips, B. and Morimoto, R. I. (1991b). Attenuation of the
heat shock response in HeLa cells is mediated by the release of bound heat
shock transcription factor and is modulated by changes in growth and heat
shock temperatures. Genes Dev. 5, 2117-2127.
Baler, R., Dahl, G. and Voellmy, R. (1993). Activation of human heat shock
genes is accompanied by oligomerization, modification, and rapid
translocation of heat shock transcription factor HSF1. Mol. Cell. Biol. 13,
2486-2496.
Bisotto, S., Lauriault, P., Duval, M. and Vincent, M. (1995). Colocalization
of a high molecular mass phosphoprotein of the nuclear matrix (p255) with
spliceosome. J. Cell. Sci. 108, 1873-1882.
Bohmann, K., Ferreira, J., Santama, N., Weis, K. and Lamond, A. I. (1995).
Molecular analysis of the coiled body. J. Cell Sci. 19, 107-113.
Cahill, C. M., Waterman, W. R., Xie, Y., Auron, P. E. and Calderwood, S. K.
(1996). Transcriptional repression of the prointerleukin 1beta gene by heat
shock factor 1. J. Biol. Chem. 271, 24874-24879
Carmo-Fonseca, M., R. Pepperkok, B. S. Sproat W. Ansorge, M. S.
Swanson and Lamond, A. I. (1991). In vivo detection of snRNP-rich
organelles in the nuclei of mammalian cells. EMBO J. 10, 1863-1873.
Cotto, J. J., Kline, M. and Morimoto, R. I. (1996). Activation of heat shock
factor 1 DNA binding precedes stress induced serine phophorylation: Evidence
for a multistep pathway of regulation. J. Biol. Chem. 271, 3355-3358.
de Jong, L., vanDriel, R., Stuurman, N., Meejne, A. M. L. and van
Renswoude, J. (1990). Principles of nuclear organization. Cell Biol. Intl.
Rep. 14, 1051-1074.
Elefanty, A. G., Antoniou, M., Custodio, N., Carmo-Fonseca, M. and
2934 J. J. Cotto, S. G. Fox and R. I. Morimoto
Grosveld, F. G. (1996). GATA transcription factors assoicate with a novel class
of nuclear bodies in erytoblasts and megakaryocytes. EMBO J. 15, 319-333.
Fu, X. D. and Maniatis, T. (1990). Factor required for mammalian
spliceosome assembly is localized to discrete regions in the nucleus. Nature
343, 437-441.
Green, M., Schuetz, T. J., Sullivan, E. K. and Kingston, R. E. (1995). A heat
shock-responsive domain of human HSF1 that regulates transcription
activation domain function. Mol. Cell. Biol. 15, 3354-3362.
He, D., Nickerson, J. A. and Penman, S. (1990). Core filaments of the nuclear
matrix. J. Cell. Biol. 110, 569-580.
Hernandez-Verdun, D. (1991). The nucleolus today. J. Cell Sci. 99, 465-471.
Jackson, D. A., Hassan, A. B., Errington, R. J. and Cook, P. R. (1994). Sites in
human nuclei where damage induced by ultraviolet light is repaired: localization
relative to transcription sites and concentrations of proliferating cell nuclear
antigen and the tumour suppressor protein, p53. J. Cell Sci. 107, 1753-1760.
Jolly, C., Morimoto, R. I., Robert-Nicoud, M. and Vourc’h, C. (1997). HSF1
transcription factor concentrates in nuclear foci during heat shock: II.
Relationship with transcription sites. J. Cell Sci. 110, 2935-2941.
Jurivich, D. A., Sistonen L., Kroes R. A. and Morimoto, R. I. (1992). Effect of
sodium salicylate on the human heat shock response. Science 255, 1243-1245.
Kline, M. P. and Morimoto, R. I. (1997). Repression of the heat shock factor 1
transcriptional activation domain is modulated by constitutive
phophorylation. Mol. Cell. Biol. 17, 2107-2115.
Knauf, U., Newton, E. M., Kyriakis, J. and Kingston, R. E. (1996).
Repression of human heat shock factor 1 activity at control temperature by
phosphorylation. Genes Dev. 10, 2782-2793.
Koken, M. H., Puvion-Dutilleul, F., Guillemin, M. C., Viron, A., LinaresCruz, G., Stuurman, N., de Jong, L., Szostecki, C., Calvo, F. and
Chomienne, C. (1994). The t(15;17) translocation alters a nuclear body in a
retinoic acid-reversible fashion. EMBO J. 13, 1073-1083.
Lamond, A. I. and Carmo-Fonseca, M. (1993a). The coiled body. Trends Cell
Biol. 3, 198-204.
Lamond, A. I. and Carmo-Fonseca, M. (1993b). Localization of splicing
snRNPs in mammalian cells. Mol. Biol. Rep. 18, 127-133.
Lee, B. S., Chen J., Angelidis C., Jurivich D. A. and Morimoto, R. I. (1995).
Pharmacological modulation of Heat Shock Factor 1 by anti-inflammatory
drugs results in protection against stress-induced cellular damage. Proc. Nat.
Acad. Sci. USA 92, 7207-7211.
Lindquist, S. (1986). The heat shock response. Ann. Rev. Biochem. 55, 11511191.
Moen, P. T., Smith, K. P. and Lawrence, J. B. (1995). Compartmentalization
of specific pre-mRNA metabolism: an emerging view. Hum. Mol. Genet. 4,
1779-1789.
Moir R. D., Montag-Lowy M. and Goldman R. D. (1994). Dynamic
properties of nuclear lamin: Lamin B is associated with sites of DNA
replication. J. Cell Biol. 125, 1201-1212.
Morimoto, R. I., Tissieres, A. and Georgopoulos, C. (1990). Stress Proteins
in Biology and Medicine. (Ed. R. I. Morimoto, A. Tissieres and C.
Georgopoulos). Cold Spring Harbor, N. Y. Cold Spring Harbor Laboratory
Press, 1-36.
Morimoto, R. I., Jurivich, D. A., Kroeger, P. E., Mathur, S. K., Murphy, S.
P., Nakai, A., Sarge, K. D., Abravaya, K. and Sistonen, L. T. (1994). The
Biology of Heat Shock Proteins and Molecular Chaperones. (Ed. R. I.
Morimoto, A. Tissieres and C. Georgopoulos). Cold Spring Harbor, N. Y.
Cold Spring Harbor Laboratory Press, 417-456.
Mosser, D. D., Theodorakis, N. G. and Morimoto, R. I. (1988). Coordinate
changes in heat shock element-binding activity and hsp70 gene transcription
rates in human cells. Mol. Cell. Biol. 8, 4736-4744.
Nakai, A. and Morimoto, R. I. (1993). Characterization of a novel chicken
heat shock transcription factor, HSF3, suggests a new regulatory pathway.
Mol. Cell. Biol. 13, 1983-1997.
Nakai, A., Y. Kawazoe, M. Tanabe, K. Nagata and Morimoto, R. I. (1995).
The DNA-binding properties of two heat shock factors, HSF1 and HSF3, are
induced in the avian erythroblast cell line HD6. Mol. Cell. Biol. 15 (10),
5268-5278.
Nakai, A., Tanabe, M., Kawazoe, Y., Inazawa, J., Morimoto, R. I. and
Nagata, K. (1997). HSF4 a New Member of the human heat shock factor
family which lacks properties of a transcriptional activator. Mol. Cell. Biol.
17, 469-481.
Piñol-Roma, S., Swanson M. S., Gall J. G. and Dreyfuss, G. (1989). A novel
heterogeneous nuclear RNP protein with a unique distribution on nascent
transcripts. J. Cell Biol. 109, 2575-2587.
Rabindran, S. K., Giorgi, G., Clos, J. and Wu, C. (1991). Molecular cloning
and expression of a human heat shock factor, HSF1. Proc. Nat. Acad. Sci.
USA 88, 6906-6910.
Roussel, P., André, C., Masson, C., Géraud, G. and Hernandez-Verdun, D.
(1993). Localization of the RNA polymerase I transcription factor hUBF
during the cell cycle. J. Cell Sci. 104, 327-337.
Sarge, K. D., Zimarino, V., Holm, K., Wu, C. and Morimoto, R. I. (1991).
Cloning and characterization of two mouse heat shock factors with distinct
inducible and constitutive DNA-binding ability. Genes Dev. 5, 1902-1911.
Sarge, K. D., Murphy, S. P. and Morimoto, R. I. (1993). Activation of heat
shock gene transcription by HSF1 involves oligomerization, acquisition of
DNA binding activity, and nuclear localization and can occur in the absence
of stress. Mol. Cell. Biol. 13, 1392-1407.
Sarge, K. D., Park-Sarge, O.-K., Kirby, J. D., Mayo, K. E. and Morimoto,
R. I. (1994). Expression of heat shock factor 2 in muse testis: potential
roleaas a regulator of heat shock protein gene expression during
spermatogenesis. Biol. Reprod. 50, 1334-1343.
Scheer, U. and Benavente, R. (1990). Functional and dynamic aspects of the
mammalian nucleolus. BioEssays. 12, 14-21.
Sheldon, L. A. and Kingston, R. E. (1993). Hydrophobic coiled-coil domains
regulate the subcellular localization of human heat shock factor 2. Genes
Dev. 7, 1549-1558.
Shi, Y., Kroeger, P. E. and Morimoto, R. I. (1995). The carboxyl-terminal
transactivation domain of heat shock factor 1 is negatively regulated and
stress responsive. Mol. Cell Biol. 15, 4309-4318.
Schuetz, T. J., Gallo, G. J., Sheldon, L., Tempst, P. and Kingston, R. E.
(1991). Isolation of a cDNA for HSF2: evidence for two heat shock factors
genes in humans. Proc. Nat. Acad. Sci. USA. 88, 6911-6915.
Sistonen, L., Sarge, K. D., Phillips, B., Abravaya, K. and Morimoto, R. I.
(1992). Activation of heat shock factor 2 during hemin-induced differentiation
of human erythroleukemia cells. Mol. Cell. Biol. 12, 4104-4111.
Sistonen, L., Sarge, K. D. and Morimoto, R. I. (1994). Human heat shock
factors 1 and 2 are differentially activated and can synergistically induce
hsp70 gene transcription. Mol. Cell. Biol. 14, 2087-2099.
Sorger, P. K. and Pelham, H. R. B. (1988). Yeast heat shock factor is an
essential DNA-binding protein that exhibits temperature-dependent
phosphorylation. Cell 54, 855-864.
Spector, D. L. (1990). Higher order nuclear organization: three dimensional
distribution of small nuclear ribonucleoprotein particles. Proc. Nat. Acd. Sci.
USA. 87, 147-151.
van Driel, R., Humbel B. and de Jong, L. (1991). The nucleus: a black box
being opened. J. Cell Biochem. 47, 1-6.
van Driel, R., Wansink, D. G., van Steensel, B., Grande, M. A., Schul, W.
and de Jong, L. (1995). Nuclear domains and the nuclear matrix. Int. Rev.
Cyt. 162A, 151-189.
van Steensel, B., Brink, M., van der Meulen, K., van Binnendijk, E. P.,
Wansink, D. G., de Jong, L., de Kloet, E. R. and van Driel, R. (1995).
Localization of the glucocorticoid receptor in discrete clusters in the cell
nucleus. J. Cell Sci. 108, 3003-3011.
van Steensel, B., van Binnendijk, E. P., Hornsby, C. D., van der Voort, H. T.
M., Krozowski, Z. S., de Kloet, E. R. and van Driel, R. (1996). Partial
colocalization of glucocorticoid and mineralocorticoid receptors in discrete
compartments in nuclei of rat hippocampus neurons. J. Cell Sci. 109, 787-792.
Weis, K., Rambaud, S., Lavau, C., Jansen, J., Carvalho, T., CarmoFonseca, M., Lamond, A. and Dejean, A. (1994). Retinoic acid regulates
aberrant nuclear localization of PML-RAR alpha in acute promyelocytic
leukemia cells. Cell 76, 345-356.
Welch, W. J. and Suhan, J. P. (1985). Morphological study of the mammalian
stress response: characterization of changes in cytoplasmic organelles,
cytoskeleton, and nucleoli, and appearance of intranuclear actin filaments in
rat fibroblasts after heat-shock treatment. J. Cell Biol. 101, 1198-1211.
Westwood, J. T., Clos, J. and Wu, C. (1991). Stress-induced oligomerization
and chromosomal relocalization of heat-shock factor. Nature 353, 822-827.
Wu, C. (1995). Heat shock transcription factors: structure and regulation. Annu.
Rev. Cell Dev. Biol. 11, 441-469.
(Accepted 1 October 1997)