Báo cáo khoa học: Differential recognition of heat shock elements by members of the heat shock transcription factor family ppt

Differential recognition of heat shock elements by
members of the heat shock transcription factor family
Noritaka Yamamoto
1
, Yukiko Takemori
1
, Mayumi Sakurai
2
, Kazuhisa Sugiyama
2
and Hiroshi Sakurai
1
1 Department of Clinical Laboratory Science, Kanazawa University Graduate School of Medical Science, Japan
2 Department of Ophthalmology and Visual Science, Kanazawa University Graduate School of Medical Science, Japan
Heat shock transcription factor (HSF), a protein that
is evolutionarily conserved from yeast to humans, is a
major regulator of heat shock protein (HSP) expres-
sion. Many HSPs function as molecular chaperones
that aid the folding of damaged proteins, and
increased accumulation of HSPs is essential for sur-
vival of cells exposed to protein-damaging stresses,
including heat shock. The structure of HSF comprises
a conserved DNA-binding domain (DBD), which
binds to the 5 bp sequence nGAAn, and two hydro-
phobic repeat (HR) regions (HR-A and HR-B), which
are necessary for homotrimer formation. Trimeric
HSF recognizes a heat shock element (HSE) compris-
ing at least three inverted repeats of the 5 bp unit
[1,2].
Biochemical and genetic evidence indicates that HSF
regulates the expression of genes encoding proteins

involved not only in stress resistance but also in
cell maintenance and developmental processes [3–5].
Saccharomyces cerevisiae HSF (yHSF) is encoded by a
single gene and is essential for cell viability even under
normal physiological conditions. yHSF target genes
encode proteins that function in protein folding,
protein degradation, detoxification, energy generation,
Keywords
crystallin; heat shock element; heat shock
protein; heat shock response; heat shock
transcription factor
Correspondence
H. Sakurai, Department of Clinical
Laboratory Science, Kanazawa University
Graduate School of Medical Science,
5-11-80 Kodatsuno, Kanazawa, Ishikawa
920-0942, Japan
Fax: +81 76 234 4369
Tel: +81 76 265 2588
E-mail:
(Received 5 November 2008, revised 14
January 2009, accepted 21 January 2009)
doi:10.1111/j.1742-4658.2009.06923.x
Heat shock transcription factor (HSF), an evolutionarily conserved stress
response regulator, forms trimers and binds to heat shock element (HSE),
comprising at least three continuous inverted repeats of the sequence
5¢-nGAAn-3¢. The single HSF of yeast is also able to bind discontinuously
arranged nGAAn units. We investigated interactions between three human
HSFs and various HSE types in vitro, in yeast cells, and in HeLa cells.
Human HSF1, a stress-activated regulator, preferentially bound to contin-

uous HSEs rather than discontinuous HSEs, and heat shock of HeLa cells
caused expression of reporter genes containing continuous HSEs. HSF2,
whose function is implicated in neuronal specification and spermatogenesis,
exhibited a slightly higher binding affinity to discontinuous HSEs than did
HSF1. HSF4, a protein required for ocular lens development, efficiently
recognized discontinuous HSEs in a trimerization-dependent manner.
Among four human c-crystallin genes encoding structural proteins of the
lens, heat-induced HSF1 preferred HSEs on the cA-crystallin and cB-crys-
tallin promoters, whereas HSF4 preferred HSE on the cC-crystallin
promoter. These results suggest that the HSE architecture is an important
determinant of which HSF members regulate genes in diverse cellular
processes.
Abbreviations
DBD, DNA-binding domain; EGS, ethylene glycol bis-(succinimidylsuccinate); hHSF, human heat shock transcription factor; HR, hydrophobic
repeat; HSE, heat shock element; HSF, heat shock transcription factor; HSP, heat shock protein; mHSF, mouse heat shock transcription
factor; SD, standard deviation; yHSF, Saccharomyces cerevisiae heat shock transcription factor.
1962 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
carbohydrate metabolism, and maintenance of cell
integrity [6–8]. yHSF binds to and regulates gene
expression via HSEs comprising variously arranged
nGAAn units: a continuous perfect-type HSE, con-
sisting of consecutive inverted repeats of the nGAAn
unit (nTTCnnGAAnnTTCn); and a discontinuous
gap-type or step-type HSE, which contains one
insertion [nTTCnnGAAn(5 bp)nGAAn] or two inser-
tions [nTTCn(5 bp)nTTCn(5 bp)nTTCn], respectively,
between the nGAAn units [9,10]. Schizosaccharo-
myces pombe HSF is also able to recognize these
various HSE types [10].
In mammalian cells, three related HSF proteins,

HSF1, HSF2, and HSF4, are involved in different,
but in some cases similar, biological functions. HSF1
is ubiquitously expressed and functions as a key regu-
lator for stress-induced transcription of HSP genes
and for acquisition of thermotolerance [1,2,11]. Analy-
sis of HSF1 knockout mice indicates the involvement
of HSF1 in extraembryonic development, carcinogene-
sis, and circadian control [12–14]. HSF2 is widely
expressed, and binds constitutively to the promoters
of HSP genes to modulate their expression [15]. Dur-
ing development, HSF2 is important for neuronal
specification and spermatogenesis [16–18]. Expression
of HSF4 is restricted to the brain and lung, and is
required for ocular lens development and fiber cell dif-
ferentiation [19–22]. There are two HSF4 isoforms, a
and b. HSF4b possesses a relatively weak activation
domain and activates transcription, whereas this
region is absent in HSF4a, which functions as a
repressor [23–25]. HSF1, HSF2 and HSF4 share sig-
nificantly conserved DBDs, but they exhibit slightly
different specificities for HSE binding in vitro
[19,26,27].
When human HSF (hHSF)1 is expressed in yHSF-
deficient S. cerevisiae cells, it fails to substitute for the
cell viability function of yHSF, because its trimer for-
mation is inhibited at normal growth temperatures
[28]. Mutant forms of hHSF1 that can trimerize in the
absence of stress are able to substitute for yHSF cell
viability function [10,28]. In these cells, however,
hHSF1 derivatives are defective in binding and activat-

ing transcription via discontinuous gap-type and step-
type HSEs, indicating that hHSF1 recognizes HSEs in
a different way from yHSF [10]. In this study, we ana-
lyzed in vitro binding of hHSF1, hHSF2 and hHSF4
to various HSE types and characterized S . cerevisiae
and HeLa cells expressing hHSFs. Our results show
that the members of the hHSF family differentially
recognize HSEs, and suggest that the regulated expres-
sion of different hHSF target genes is dependent upon
the architecture of the HSE.
Results
Human HSF1, HSF2 and HSF4 exhibit differential
binding specificities for various HSE types
in vitro
Interactions between hHSFs and various HSE types
were investigated using electrophoretic mobility shift
assays with in vitro-synthesized polypeptides and oligo-
nucleotide probes (Fig. 1A). Protein–DNA complexes
were formed when binding reactions were carried out
using hHSF1-programmed transcription ⁄ translation
mixtures, but not in control reaction mixtures that
A
B
Fig. 1. Binding of hHSFs to various HSE types in vitro. (A) Nucleo-
tide sequences of four model HSEs. The GAA and inverted TTC
sequences are indicated in bold upper-case letters with arrows.
These HSE oligonucleotides were used as DNA probes for electro-
phoretic mobility shift assays and as cis-acting sequences for HSE–
SV40p–LUC reporters. (B) Electrophoretic mobility shift assay of
hHSFs. Typical results obtained using in vitro-synthesized hHSF1

(3.6 ng), hHSF2 (0.9 ng) and hHSF4 (3.6 ng) polypeptides are
shown. The reaction mixture programmed with vacant vector DNA
was used as a control (C). The binding reaction was carried out at
37 or 43 °C for 20 min with
32
P-labeled oligonucleotides HSE4Ptt
(4P), HSE3P (3P), HSEgap (G), HSEstep (S), or STRE (N). STRE oli-
gonucleotide (TCGACACCCCTTATCTAGAGACCCCTTACCTCGA)
was used as a nonspecific binding control. Samples were subjected
to gel electrophoresis and phosphorimaging. Open and closed
arrowheads indicate the positions of DNA fragments bound by one
and two hHSF trimers, respectively. The binding affinities relative
to HSE3P are shown below each lane. The experiments were
performed at least three times, and yielded similar results.
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1963
were programmed with an empty vector (Fig. 1B).
Human HSF1 was able to bind to HSE4Ptt and HSE3P
oligonucleotides containing four and three continuous,
inverted nGAAn repeats, respectively. hHSF1–HSE4Ptt
migrated more slowly than hHSF1–HSE3P, and the
amount of hHSF1–HSE4Ptt was 1.5-fold higher than
that of hHSF1–HSE3P. Previously, it was shown that
4Ptt-type HSE is bound by two trimers of Drosophila or
Saccharomyces HSFs in a cooperative manner [29,30].
This suggests that the slower migration of hHSF1–
HSE4Ptt is a result of cooperative interaction of two
hHSF1 trimers (Fig. S1). Incubation of hHSF1 with
gap-type (HSEgap) or step-type (HSEstep) HSE oligo-
nucleotides resulted in complex formation, but the

amounts were threefold and 20-fold lower, respectively,
than that of hHSF1–HSE3P. This demonstrates that
hHSF1 preferentially binds to continuous HSEs. Note
that the interaction of hHSF1 with HSEs was
stimulated without changing HSE specificity when the
binding reaction was carried out at 43 °C rather than
37 °C (Fig. 1B), as reported for binding of mouse HSF
(mHSF)1 to an HSE containing four continuous,
inverted nGAAn repeats [31].
When hHSF2 polypeptide was incubated with perfect-
type HSEs, the electrophoretic mobility and the amount
of hHSF2–HSE4Ptt were almost the same as those of
hHSF2–HSE3P (Fig. 1B), indicating that a single
hHSF2 trimer binds to HSE4Ptt. Gap-type and step-
type HSEs were recognized by hHSF2, although the
binding affinity for HSEstep was threefold lower than
that for HSE3P. The amount of hHSF2 polypeptide
used in the assay was fourfold lower than that of hHSF1
polypeptide, and the addition of fourfold more hHSF2
polypeptide to the reaction mixture caused an increase in
the amount of hHSF2–HSE complexes without chang-
ing HSE specificity (Fig. S1). Although it is unknown
whether all polypeptides synthesized are active for
binding, hHSF2 appears to have a higher binding affin-
ity for at least discontinuous HSEs than does hHSF1.
Human HSF4 was observed to bind as a single tri-
mer to 4Ptt-type HSE, as judged from the amount and
mobility of the complex (Fig. 1B). Notably, the
amount of complex formed with hHSF4 and HSEgap
or HSEstep was more than 70% that of hHSF4–

HSE3P. Therefore, like yHSF, hHSF4 possesses the
ability to bind to various HSE types comprising
different configurations of nGAAn units.
Phenotypes of S. cerevisiae cells expressing
hHSF2 and hHSF4
We constructed S. cerevisiae cells expressing hHSF2
and hHSF4, and analyzed their phenotypes. In agree-
ment with previous reports [23,28], yeast cells harbor-
ing hHSF2 and hHSF4 on low-copy-number (YCp) or
high-copy-number (YEp) plasmid grew at temperatures
below 35 and 33 °C, respectively (Fig. 2A). The
amounts of hHSF4 in cells harboring YCp-hHSF4 or
YEp-hHSF4 (0.01–0.1 ng hHSF4Ælg
)1
protein) were
markedly lower than those of hHSF2 in cells harbor-
ing YCp-hHSF2 or YEp-hHSF2 (1–2 ng hHSF2Ælg
)1
protein), for unknown reasons (Fig. 2B).
Using RT-PCR, we analyzed the mRNA levels of
yHSF target genes containing 4P-type HSE (HSP42,
HSP78, and KAR2), 3P-type HSE (APA1, HSP10,
and SSA2), gap-type HSE (CPR6, CUP1, and
HSP82), step-type HSE (FSH1, SGT2, and SSA3),
and atypical HSE consisting of directly repeating
nGAAn units and several irregular nGAAn units
(DR-type; AHP1 and TIP1) [10]. When yHSF cells
grown at 28 °C were heat-shocked at 39 °C, the
mRNA levels of target genes were significantly
increased (Fig. 2C). In yeast cells expressing hHSF2,

the heat-induced transcription of all target genes
analyzed was appropriately regulated, with the excep-
tion of transcriptional activation via step-type HSEs,
which was slightly lower in hHSF2-expressing cells than
in yHSF-expressing cells. hHSF4 was also able to
compensate for yHSF in the regulation of target gene
expression; however, mRNA levels were slightly
reduced in hHSF4 cells as compared to yHSF cells. The
low mRNA levels may be due to the relatively weak
transcriptional activity [19,23] and ⁄ or the low-level
expression of hHSF4 (Fig. 2B). Unlike trimerization-
prone hHSF1 derivatives, which fail to activate
transcription of genes containing gap-type, step-type or
DR-type HSEs in yeast cells [10], hHSF2 and hHSF4
activate transcription of various target genes, and thus
support cell viability at elevated temperatures.
Heat-induced expression of reporter genes
containing various HSE types in HeLa cells
The transcriptional activity of various HSE types in
mammalian cells was analyzed using reporter genes con-
taining HSE oligonucleotides positioned upstream of
the SV40 promoter–luciferase gene fusion (HSE–
SV40p–LUC). In HeLa cells, insertion of HSEs in the
reporter gene did not significantly affect the basal
expression level under normal culture conditions
(Fig. 3A). This suggests that endogenous hHSFs are
not involved in the expression, although it is possible
that they bind to HSEs of reporter genes without affect-
ing the expression. When cells were heat-shocked at
43 °C for 1 h and allowed to recover at 37 °C for 5 h,

expression of the reporter gene containing 4Ptt-type
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1964 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
HSE (HSE4Ptt–SV40p–LUC) was induced by more
than 15-fold (Fig. 3A). After heat shock, expression
directed by 3P-type HSE was modestly induced
( 5-fold), but discontinuous gap-type and step-type
HSEs failed to mediate the induction. This suggests that
heat-induced hHSF1 preferentially activates transcrip-
tion of genes containing continuous HSEs.
Expression of HSE–SV40p–LUC reporter genes
by hHSF–VP16 fusion proteins
The activation potential of hHSF2 is significantly
lower than that of hHSF1 [32], and hHSF4 is not
appreciably expressed in HeLa cells [19]. To explore
HSE architecture-specific functions, the herpes simplex
virus VP16 activation domain was fused to the C-ter-
mini of hHSF1, hHSF2, and hHSF4, and the resulting
hHSF–VP16 constructs were introduced into HeLa
cells. Fusion of the VP16 activation domain did not
significantly affect the HSE specificity of hHSFs, as
judged by electrophoretic mobility shift assay of
in vitro-synthesized polypeptides (Fig. S2). As shown
by immunoblot analysis with an antibody against
VP16, these fusion proteins were expressed in HeLa
cells; however, the amount of hHSF2–VP16 was much
lower than that of hHSF1–VP16 or of hHSF4–VP16,
even when cells were transfected with 10-fold more
hHSF2–VP16 expression construct (Fig. 3B). It has
been shown in HeLa cells that transfected hHSF1

forms oligomers and binds to HSEs at physiological
temperatures [33]. Consistent with the results of heat
shock response, hHSF1–VP16 activated constitutive
expression of SV40p–LUC reporters containing contin-
uous HSEs, but the levels of activation for reporters
containing discontinuous HSEs were less than twofold
(Fig. 3C). The reporter gene expression in the presence
of hHSF2–VP16 was similar in pattern to that
observed in the presence of hHSF1–VP16, except that
HSEgap–SV40p–LUC expression was activated three-
fold. In contrast, hHSF4–VP16 was a potent activator
of reporter genes containing gap-type and step-type
HSEs. The HSE type-specific differences in transcrip-
tion of these reporters were consistent with the in vitro
binding affinity of each hHSF and HSE type, suggest-
ing that hHSF1, hHSF2 and hHSF4 differentially
recognize various HSEs in mammalian cells.
yHSF
hHSF4
45 15 0 45 15 0 45 15 0 45 15 0
hHSF2
45 15 0
YCp YEp YCp YEp
C
HSP42
HSP78
Gap
3P
4P
control

Step
CPR6
HSP10
CUP1
HSP82
SGT2
FSH1
ACT1
APA1
KAR2
SSA2
TIP1
SSA3
AHP1
DR
39
o
C (min)
A
33
o
C
37
o
C 35
o
C
hHSF2
YCp YEp
B

39
o
C (min)
15 0 15 0
50
70
yHSF
hHSF2
28
o
C
YCp
YEp
hHSF4
YEp
YCp
hHSF4
YCp YEp
15 0 15 0
NS
Fig. 2. Characterization of yeast cells expressing hHSF2 and
hHSF4. (A) Growth of hHSF cells. Cells of strains HS170T (YCp-
yHSF), YYT49 (YCp-hHSF2), YYT42 (YEp-hHSF2), YYT50 (YCp-
hHSF4) and YYT17 (YEp-hHSF4) were streaked onto YPD medium
and incubated at the indicated temperatures for 2 days. (B) Immu-
noblot analysis of hHSF proteins. Cells were grown in YPD medium
at 28 °C and heat-shocked at 39 °C for the indicated times.
Extracts of cells expressing hHSF2 (2 lg of protein) or hHSF4
(20 lg of protein) and recombinant hHSF proteins (not shown)
were subjected to immunoblot analysis with antibodies against

hHSF2 and hHSF4. The positions of molecular mass markers are
shown on the left in kilodaltons. NS denotes nonspecific band. The
experiments were performed at least twice, and yielded similar
results. Cell extracts (1 lg of protein) of YYY49 and YYT42 con-
tained approximately 1 and 2 ng of hHSF2, and those of YYT50 and
YYT17 contained approximately 0.01 and 0.1 ng of hHSF4, as
judged by the intensity of each band. (C) mRNA levels in heat-
shocked hHSF cells. Cells were grown in YPD medium at 28 °C
and heat-shocked at 39 °C for the indicated times. Total RNA pre-
pared from the cells was subjected to RT-PCR analysis. The genes
targeted by yHSF are classified according to the structure of their
HSEs: 4P, 3P, Gap, Step and DR (directly repeating nGAAn units
and several irregular nGAAn units) types. The experiments were
performed at least three times and yielded similar results.
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1965
Efficient trimerization of hHSF4–VP16 is
necessary for activation via discontinuous HSEs
To locate the HSF4 region responsible for transcrip-
tional regulation via discontinuous HSEs, we analyzed
the transcriptional activity of various hHSF4–VP16
derivatives (Fig. 4A,B). Human HSF4 contains a DBD
at the N-terminus, HR-A and HR-B in the central
region, and a relatively weak activation domain at the
C-terminus [19]. Deletion of the C-terminal half of
hHSF4 (hHSF4-n355–VP16 and hHSF4-n217–VP16)
did not significantly affect transcriptional activity or
HSE specificity, with the exception of a slight decrease
of HSE3P–SV40p–LUC expression by hHSF4-n355–
VP16. hHSF4–VP16 lacking HR-B (hHSF4-n178–

VP16) exhibited transcriptional activation via 3P-type
HSE, but failed to do so via gap-type or step-type
HSEs. An extended deletion construct leading to partial
removal of HR-A (hHSF4-n159–VP16) was abundantly
expressed but failed to activate transcription.
The roles of HR-A and HR-B were examined by
introducing amino acid substitutions. In hHSF4-
L140P–VP16 and hHSF4-I186P–VP16, a helix-destabi-
lizing residue (proline) was substituted for a hydro-
phobic residue (leucine or isoleucine) in HR-A and
HR-B, respectively. To analyze oligomer formation of
these hHSF4–VP16 derivatives, polypeptides synthe-
sized in vitro were subjected to chemical crosslinking
with ethylene glycol bis-(succinimidylsuccinate) (EGS)
(Fig. 4C). The band of approximately 220 kDa, which
corresponds to the size of a trimer, was detected by
treatment of wild-type hHSF4–VP16 with EGS. The
L140P and I186P substitutions appeared to inhibit tri-
mer formation, and most of the polypeptides migrated
at the position of a monomer (75 kDa). When an elec-
trophoretic mobility shift assay was conducted
(Fig. 4D), the substitution derivatives and 3P-type
HSE formed complexes exhibiting mobilities similar to
that of wild-type hHSF4–VP16 trimer–HSE3P complex
[this may be somewhat surprising; however, the com-
plex formation might be supported by DBD–DBD and
DBD–HSE3P interactions (see Discussion)]. However,
they exhibited reduced binding affinities for gap-type
and step-type HSEs. In HeLa cells, the L140P and
I186P substitutions in hHSF4–VP16 inhibited tran-

scriptional activation via gap-type and step-type HSEs,
AB
C
Control
Heat shock
hHSF1-VP16
hHSF4-VP16hHSF2-VP16
Gap
3P
4Ptt
Step
None
2.5 50
Fold activation
10 200
10 200
10 200
Gap
3P
4Ptt
Step
None
10 200
hHSF4-VP16
62
83
hHSF2-VP16
hHSF1-VP16
Fold activation
*

*
*
*
*
*
*
*
*
**
NS
Fig. 3. Expression of artificial reporter genes containing various HSE types in HeLa cells. (A) Heat shock-induced expression. Cells were
transfected with DNA mixtures containing 100 ng of SV40p–LUC plasmid (none) or HSE–SV40p–LUC plasmids (4Ptt, 3P, Gap, and Step). For
heat shock experiment, cells were incubated at 43 °C for 1 h, and culture was continued at 37 °C for 5 h. Luciferase activity (fold activation)
was expressed relative to that of SV40p–LUC plasmid-transfected cells (control, left panel) or to that of cells grown at 37 °C (heat shock,
right panel). Each bar represents the mean ± standard deviation (SD) for at least five experiments. Asterisks indicate significant differences
(P < 0.01) as compared with SV40p–LUC control as determined by Student’s t-test. (B) Immunoblot analysis of hHSF–VP16 fusion proteins.
Cells were transfected with DNA mixtures containing 100 ng of reporter plasmid (lane –) and hHSF1–VP16 (10 ng), hHSF2–VP16 (100 ng) or
hHSF4–VP16 (10 ng) expression constructs. Extracts prepared from cells grown at 37 °C were subjected to immunoblotting using an anti-
body against VP16. The positions of molecular mass markers are shown on the left in kilodaltons. NS denotes nonspecific band. The experi-
ments were performed at least twice, and yielded similar results. (C) Constitutive expression in cells cotransfected with hHSF–VP16
plasmids. Transfection was carried out described for (B). Luciferase activity (fold activation) was expressed relative to that of cells
transfected with the reporter plasmid alone. Each bar represents the mean ± SD for at least five experiments. Asterisks indicate significant
differences (P < 0.01) as compared with SV40p–LUC control as determined by Student’s t-test.
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1966 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
but not via 3P-type HSE (Fig. 4A). These results show
that trimerization facilitated by HR-A ⁄ B is obligatory
for binding of hHSF4 to discontinuous HSEs.
Differential expression of c-crystallin promoter–
luciferase reporter genes by heat-induced hHSF1

and hHSF4–VP16
We next analyzed expression of the luciferase gene dri-
ven by promoters of the human c-crystallin genes,
whose mouse orthologs are transcriptionally regulated
by HSF1 and HSF4 [21,22]. As shown in Fig. 5A, the
HSE of the cA-crystallin (CRYGA) promoter contains
six GAA-like sequences, at positions 1, 2, 3, 4, 6, and
7. CRYGA–LUC expression was induced eightfold by
heat shock, and cotransfection of hHSF4–VP16 caused
a fivefold increase in expression. The cB-crystallin
(CRYGB) HSE is similar to the CRYGA HSE in
sequence and configuration of GAA-like sequences.
CRYGB–LUC expression was induced by heat shock
and by hHSF4–VP16 cotransfection. In electrophoretic
mobility shift assays, DNA fragments containing the
CRYGA and CRYGB HSEs were bound by hHSF1
and hHSF4–VP16 (Fig. 5B). The cC-crystallin (CRY-
GC) promoter contains two HSEs: a distal 3P-like
HSE, and a proximal HSE comprising six GAA-like
sequences (p1, p2, p3, p5, p6, and p7) (Fig. 5A).
Unlike the expression observed for CRYGA–LUC and
CRYGB–LUC, CRYGC–LUC expression was induced
only threefold by heat shock, whereas cotransfection
of hHSF4–VP16 caused a 14-fold increase in expres-
sion. Human HSF1 bound to the proximal but not
Gap
3P
Step
10 20
WT

n159
n355
n217
n178
A
I186P
L140P
159
178
217
355
VP16
493
HR-A/BDBD
VP16
VP16
VP16
VP16
VP16
VP16
Fold activation
0
493
493
C
B
240
100
140
70

175
62
83
47.5
n355
WT
n217
n178
L140P
n159
I186P
32.5
25
D
3P SG3PSG3PSG
WT I186PL140P
EGS
3.01.003.01.003.01.00
WT I186PL140P
(mM)
*
*
*
*
*
*
*
*
*
NS

Fig. 4. Expression of reporter genes by hHSF4–VP16 derivatives in HeLa cells. (A) Expression in cells cotransfected with hHSF4–VP16 deriv-
atives. Structures of hHSF4–VP16 derivatives are shown on the left. The DBD and two HRs (HR-A ⁄ B) are shown. Numbers indicate amino
acid positions. Vertical bars show the positions of amino acid substitutions. Cells were transfected with DNA mixtures containing 100 ng of
reporter plasmid and 10 ng of the indicated hHSF4–VP16 derivatives. Luciferase activity (fold activation) was expressed relative to that of
cells transfected with the reporter plasmid alone. Each bar represents the mean ± SD for at least four experiments. Asterisks indicate signif-
icant differences (P < 0.01) as compared with hHSF4–VP16 control as determined by Student’s t-test. (B) Immunoblot analysis of hHSF4–
VP16 derivatives. Extracts were prepared from cells transfected as described for (A), and were subjected to immunoblot analysis. Positions
of molecular mass markers are shown on the left in kilodaltons. NS denotes nonspecific band. The experiments were performed at least
twice, and yielded similar results. (C) Chemical crosslinking analysis of hHSF4–VP16 derivatives. In vitro-synthesized polypeptides (4.0 ng)
were incubated without or with 1.0 and 3.0 m
M EGS, and were subjected to immunoblot analysis. Positions of molecular mass markers are
shown on the left in kilodaltons. Open and closed circles indicate the positions of hHSF4–VP16 monomers and trimers, respectively. The
experiments were performed at least twice, and yielded similar results. (D) Electrophoretic mobility shift assay of hHSF4–VP16 derivatives.
Typical results using in vitro-synthesized polypeptides (4.0 ng) are shown as described for Fig. 1B. Open arrowheads indicate the positions
of DNA fragments bound by a single hHSF4–VP16 trimer. The experiments were performed at least three times, and yielded similar results.
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1967
A
B C
Fig. 5. Expression of c-crystallin promoter-luciferase reporter genes in HeLa cells. (A) Expression by heat-induced hHSF1 and by
cotransfected hHSF4–VP16. Structures of the c-crystallin promoter–luciferase reporter genes are shown on the left. Bars represent the
crystallin genes, and open boxes indicate the HSEs. The cC-crystallin promoter contains two HSEs, one at a distal (dHSE) position and one
at a proximal (pHSE) position. Numbers indicate nucleotide positions relative to the translation initiation site. HSE sequences are shown in
which the GAA and inverted TTC sequences are indicated by bold upper-case letters with numbers. The nucleotides alterations are shown
below the HSEs. For heat shock experiments, cells were transfected with DNA mixtures containing 200 ng of the indicated reporter
plasmid, and luciferase activity (fold activation) was determined as described for Fig. 3A. For cotransfection experiments, cells were
transfected with DNA mixtures containing 200 ng of the indicated reporter plasmid and 10 ng of hHSF4–VP16 expression construct.
Luciferase activity (fold activation) was determined as described for Fig. 3C. Each bar represents the mean ± SD for at least four
experiments. Asterisks indicate significant differences (P < 0.01) as compared with wild-type control as determined by Student’s t-test. (B)
Electrophoretic mobility shift assay of hHSF1 and hHSF4–VP16. Typical results obtained using in vitro-synthesized hHSF1 (4.8 ng) and

hHSF4–VP16 (3.0 ng) are shown as described for Fig. 1B. Probe fragments were prepared by PCR with primers flanking the putative HSEs
of CRYGA ()217 to )157), CRYGB ()228 to )168), CRYGC (distal, )367 to )307; proximal, )234 to )174), and CRYGD ()247 to )187). The
binding reaction was carried out at 37 °C (hHSF4–VP16) or 43 °C (hHSF1) for 20 min. Brackets indicate protein–DNA complexes. (C)
Electrophoretic mobility shift assay of hHSF1. The gel was electrophoresed longer than the gels of (B) to resolve DNA fragments bound by
one (open arrowhead) and two (closed arrowhead) hHSF1 trimers.
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1968 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
distal HSE, and hHSF4–VP16 bound to both HSEs
(Fig. 5B). The cD-crystallin (CRYGD) promoter con-
tains an HSE-like sequence comprising four GAA-like
sequences, and CRYGD–LUC expression was induced
twofold by heat shock (Fig. 5A). However, hHSF1
may not be involved in the heat shock response,
because it failed to bind the putative HSE under our
assay conditions (Fig. 5B). The putative HSE was only
weakly bound by hHSF4–VP16, but hHSF4–VP16
cotransfection did not affect the expression of
CRYGD–LUC (Fig. 5A,B). These results show that
heat shock-induced hHSF1 prefers cA-crystallin and
cB-crystallin promoters, whereas hHSF4–VP16 prefers
the cC-crystallin promoter.
HSF–HSE interactions were analyzed by introducing
base alterations in the HSEs of reporter genes
(Fig. 5A). In CRYGA–LUC derivatives, the m1 repor-
ter gene contained a 3P-like HSE (units 2, 3, and 4), a
gap-like HSE (units 3, 4, and 6), and a step-like HSE
(units 2, 4, and 6); the m2 reporter gene contained a
gap-like HSE (units 3, 4, and 6); the m3 reporter con-
tained a step-like HSE (units 2, 4, and 6); the m4
reporter gene lacked any apparent HSE; and the m6

reporter contained a 3P-like HSE (units 2, 3, and 4).
The m1 and m6 reporter gene alterations significantly,
but not completely, inhibited heat shock-induced
expression, although these reporters contained a
3P-like HSE. This result could be explained by a
model in which wild-type CRYGA promoter is bound
by two hHSF1 trimers; one trimer binding to units 1,
2, and 3, and the other binding to units 4 and 6.
Consistently, hHSF1–wild-type complex migrated more
slowly than hHSF1–m1 complex or hHSF1–m6 com-
plex in electrophoretic mobility shift assays (Fig. 5C).
Expression of m2 and m3 reporters was reduced to the
level of the m4 reporter, suggesting that gap-like and
step-like HSEs of these reporters are nonfunctional for
binding by hHSF1 (Fig. 5A). In hHSF4–VP16 cotrans-
fection experiments, unit 1 of the CRYGA HSE was
dispensable for activation (m1), although alterations of
other units, including unit 6, caused significant inhibi-
tion of activation (m2, m3, m4, and m6). Similar
results were obtained in electrophoretic mobility shift
assays (Fig. 5B). The observation that hHSF4–VP16
did not bind stably to sequences containing either
3P-like, gap-like or step-like HSEs might be explained
by the divergence of GAA-like sequences from the
canonical GAA sequence.
An alteration of either the distal or proximal HSE
of CRYGC–LUC caused inhibition of hHSF4–VP16-
induced expression, suggesting that both HSEs are
involved in hHSF4–VP16 binding (mdHSE and mp3,6)
(Fig. 5A). Notably, changing GAC to GAA in unit 3

resulted in robust activation by hHSF4–VP16, without
changing the magnitude of the heat shock response
(cp3). The mutational analysis of the CRYGA and
CRYGC HSEs suggests that the nucleotide sequences
and configuration of nGAAn-like units are important
for interaction with hHSF1 and hHSF4.
Discussion
In this study, we demonstrate that hHSF1, hHSF2 and
hHSF4 differentially recognize HSEs comprising
diversely arranged nGAAn units in vitro, in yeast cells,
and in HeLa cells. All three hHSFs bind to HSEs with
continuous, inverted repeats of nGAAn. In addition,
hHSF4 exhibits a relatively higher affinity for discon-
tinuous HSEs containing gaps between nGAAn units.
Trimerization facilitated by the HR-A ⁄ B is obligatory
for hHSF4 recognition of discontinuous HSEs. In
addition to these results obtained with synthetic, model
HSEs, hHSF4 exhibited a different specificity from
heat shock-induced hHSF1 in interactions with the
human c-crystallin promoters. These results show that
the configuration of nGAAn units in the promoter is
important in determining which hHSF members are
involved in the regulation of the gene.
Footprint analysis of the hHSF4–HSE interaction
has shown that hHSF4 binding on the HSP70 pro-
moter corresponds to a region that is identical to that
observed with mHSF1 but is distinct from
that observed with mHSF2 [19]. It has been reported
that, similar to hHSF1, hHSF4 expressed in yeast cells
strongly activates transcription of SSA3, but only

slightly activates transcription of CUP1 [23]. Recently,
Fujimoto et al. [34] have shown that mHSF4 is
required for induction of a set of genes in response to
heat shock, in part by facilitating mHSF1 binding.
Although these results implied a similarity between
hHSF1 and hHSF4 in HSE binding specificity, our
data show that hHSF4 exhibits a binding specificity
clearly distinguishable from that of hHSF1 and
hHSF2, and is able to recognize discontinuously posi-
tioned nGAAn units. We suggest that genes containing
discontinuous HSEs are preferred targets for hHSF4
but not for hHSF1 or hHSF2. Consistently, Fujimoto
et al. [34] identified genomic regions that are occupied
by only mHSF4, and showed that the HSF4 binding
consensus sequence is more ambiguous than that of
HSF1 and HSF2. Two hHSF4 isoforms, hHSF4a and
hHSF4b, share the same DBD and HR-A ⁄ B, but func-
tion as a repressor and activator, respectively [23–25].
Phosphorylation of HSF4b by extracellular signal-
related kinase leads to increased ability of hHSF4b to
bind DNA [35]. Therefore, genes containing discontin-
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1969
uous HSEs are subject to positive and negative regula-
tion by phosphorylated and unphosphorylated hHSF4
isoforms.
The L140P and I186P substitutions in hHSF4–VP16
inhibited binding to gap-type and step-type HSEs but
not to 3P-type HSE. In mammalian HSF1, HR-A ⁄ B
interacts with the third HR region (HR-C) and main-

tains HSF1 in an inactive monomeric form under
physiological conditions. HR-A ⁄ B of transcriptionally
active HSF1 mediates interactions among three mono-
mers to form a trimer, thereby facilitating binding to
the HSE [33,36,37]. The linker region located between
the DBD and HR-A ⁄ B also plays a role in oligomer
formation [38]. However, it remains unknown whether
HR-A ⁄ B of mammalian HSFs is involved in the speci-
ficity of the HSF–HSE interaction. In yHSF, HR-A ⁄ B
has been shown to be necessary for interaction with
promoters containing not only gap-type or step-type
HSEs but also 3P-type HSE, but not with those con-
taining four or more nGAAn units [9]. Notably, the
yHSF HR-A ⁄ B region can be substituted with a
dimerization domain from an unrelated protein with
no effect on the HSE-binding properties of the protein,
which suggests that yHSF HR-A ⁄ B does not play an
important role in the binding of HSEs, other than olig-
omerization [9,39]. It has been shown that DBD–DBD
interaction affects HSF trimerization and HSE binding
[9,40–42]. The HR-A ⁄ B-facilitated trimerization of
hHSF4 was not obligatory for binding to continuous
HSEs, suggesting that other, as yet unknown regions,
including, potentially, the DBD, have roles in hHSF4–
HSE interaction. However, efficient trimerization was
required for the interaction of hHSF4 with discontinu-
ous HSEs, which indicates that this as yet unidentified
region is not sufficient for binding to discontinuous
HSEs.
By using synthetic model HSEs, we have shown that

both hHSF1 and hHSF2 preferentially bind to contin-
uous HSEs in vitro and in HeLa cells. hHSF1 and
hHSF2 consistently share the same target genes as
judged by chromatin immunoprecipitation analysis
[43]. In binding to continuous HSEs, mHSF1 prefers
long arrays of the nGAAn unit, whereas mHSF2 pre-
fers short arrays [27]. These differences are related to
differences in the capability for cooperative interac-
tions of trimers [26,27], which was confirmed by our
electrophoretic mobility shift assay (Fig. 1B). The
wing region of the DBD facilitates interactions among
mHSF1 trimers [41]. We have shown that in vitro and
in HeLa cells, hHSF2 exhibits a slightly higher bind-
ing affinity for discontinuous HSEs than observed for
hHSF1, and that unlike hHSF1 [10], hHSF2 expressed
in yeast cells properly regulates gene expression via
atypical HSEs as well as discontinuous HSEs. In this
regard, it should be noted that the mouse p35 gene, a
specific target of mHSF2, contains a putative HSE
that diverges from the canonical HSE [17]. It was
recently reported that mammalian HSF1 and HSF2
form heterotrimers and that HSF2 modulates the
activity of stress-induced HSF1 in a gene-specific
manner [44,45]. Differences in cooperativity and HSE
specificity are likely to be important determinants of
the interaction between HSF1–HSF2 heterotrimers
and HSEs.
In mouse, expression of the c-crystallin gene family
in the lens is regulated by various transcription factors,
including HSF1 and HSF4 [21,22,46]. Our analysis of

the four human c-crystallin promoters has shown that
heat-induced hHSF1 preferentially activates the
CRYGA and CRYGB promoters, and hHSF4–VP16
activates the CRYGC promoter, whereas neither acti-
vates the CRYGD promoter. Five GAA-like sequences
in the CRYGA promoter may provide a site for coop-
erative hHSF1 binding. In the CRYGC promoter, the
proximal and distal HSEs were necessary for activation
by hHSF4–VP16. Mice lacking HSF4 develop cata-
racts during the early postnatal period, probably due
to decreased expression of c-crystallin and ⁄ or HSP25
[21,22]. In humans, the CRYGC and CRYGD genes
encode abundant lens c-crystallins [47], and CRYGC
transcription is regulated by HSF4. This may be one
of the reasons why missense mutations in the HSF4
gene are associated with congenital cataracts [20].
Transcriptional regulation of genes by three mam-
malian HSFs is implicated in a variety of cellar
processes, including cell maintenance and differentia-
tion, as well as stress resistance [3–5]. Which HSF
members are expressed in cells is important in deter-
mining which genes are activated or repressed.
Although hHSF1, hHSF2 and hHSF4 contain similar
DBDs and HR-A ⁄ B regions, they possess differential
binding specificities for various HSE types. This
differential specificity may give HSFs the ability to
distinguish their target genes.
Experimental procedures
Plasmids
The ORFs of hHSF1, hHSF2 and hHSF4b were cloned into

plasmid pcDNA3.1(+) (Invitrogen, Carlsbad, CA, USA).
For expression of hHSFs in Escherichia coli, hHSF1,
hHSF2 and hHSF4b (amino acids 220–493) were cloned
into plasmid pGEX6P-1 (GE Healthcare, Piscataway, NJ,
USA). For expression in yeast cells, hHSF2 and hHSF4b
were inserted between the ADH1 promoter and terminator
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1970 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
of low-copy-number plasmid pSK484 (YCp-TRP1–P
ADH1
–
T
ADH1
) [48] and high-copy-number plasmid pK346 (YEp-
LEU2–P
ADH1
–T
ADH1
) [10]. For expression of hHSF–VP16
fusion proteins in HeLa cells, hHSF1, hHSF2 and hHSF4b
were cloned into pK543, a derivative of pcDNA3.1(+) con-
taining an activation domain of herpes simplex virus VP16
(amino acids 413–490). Derivatives of hHSF4–VP16 were
created by using standard methods. The reporter gene
HSE–SV40p–LUC contained an HSE oligonucleotide (see
Fig. 1A) upstream of the SV40 promoter–firefly luciferase
gene fusion (SV40p–LUC) of pGL3-Promoter vector (Pro-
mega, Madison, WI, USA). The promoter regions of the
human cA-crystallin, cB-crystallin, cC-crystallin and cD-
crystallin genes were cloned upstream of the luciferase gene

of pGL3-Basic vector (Promega) to create CRYGA–LUC,
CRYGB–LUC, CRYGC–LUC, and CRYGD–LUC,
respectively.
In vitro polypeptide synthesis, electrophoretic
mobility shift assay, and chemical crosslinking
analysis
hHSF polypeptides were synthesized by in vitro transcrip-
tion ⁄ translation reaction (TNT Coupled Reticulocyte
Lysate System with T7 RNA polymerase; Promega) using
pcDNA3.1(+) derivatives that carried the hHSF ORFs as
templates. The synthesized polypeptides were detected as a
single band of the expected molecular mass, as judged by
immunoblot analysis with antibodies against hHSF1,
hHSF2, and hHSF4 (kindly provided by A. Nakai, Yama-
guchi University School of Medicine, Japan). The amounts
of polypeptides were determined by immunoblot analysis
using purified recombinant hHSFs as references (data not
shown). The recombinant proteins were expressed in
E. coli as fusion proteins with glutathione S-transferase.
Fusion proteins were purified on glutathione Sepharose 4B
beads and were treated with PreScission Protease accord-
ing to the manufacturer’s protocol (GE Healthcare).
For electrophoretic mobility shift assays, the binding
reaction was carried out in 16 lL of mixture containing
0.2–2.0 lLofin vitro transcription ⁄ translation reaction
mixture (0.9–5.8 ng of hHSF polypeptides), 25 mm
Hepes ⁄ KOH (pH 7.6), 25 mm NaCl, 2 mm EDTA, 5%
glycerol, 200 ng of poly(dI-dC) and 0.02 ng of
32
P-labeled

HSE oligonucleotide for 20 min at 37 or 43 °C. The
samples were electrophoresed on a 3.5% polyacrylamide gel
at room temperature, and subjected to phosphorimaging as
described previously [9,10].
Oligomer formation of polypeptides was analyzed by
chemical crosslinking with EGS [9]. In vitro-synthesized
polypeptides (1.0–1.5 lL, 4.0 ng of protein) in 5 lLof
13 mm Tris ⁄ Cl (pH 7.6) and 100 mm NaCl were incubated
without or with 1.0 and 3.0 mm EGS for 20 min at room
temperature. The reaction was quenched by the addition of
glycine to 75 mm. Samples were subjected to SDS ⁄ PAGE
and immunoblot analysis using an antibody against VP16
(Abcam, Cambridge, UK).
Yeast strains, immunoblot analysis, and RT-PCR
Yeast strain HS126 (MATa ade2 his3 leu2 trp1 ura3 can1
hsf1::HIS3 YCp-URA3–yHSF) contains a null mutation of
the chromosomal yHSF gene and bears wild-type yHSF on
a URA3-marked centromeric plasmid [10]. For construction
of strains HS170T, YYT49, YYT42, YYT50, and YYT17,
HS126 was transformed respectively with YCp-TRP1–
yHSF, YCp-TRP1–P
ADH1
–hHSF2–T
ADH1
, YEp-LEU2–
P
ADH1
–hHSF2–T
ADH1
, YCp-TRP1–P

ADH1
–hHSF4b–T
ADH1
,
and YEp-LEU2–P
ADH1
–hHSF4b–T
ADH1
, and the resident
YCp-URA3–yHSF was evicted by streaking transformed
cells on medium containing 5-fluoroorotic acid [10]. Cells
were grown in YPD medium consisting of 1% yeast extract,
2% polypeptone, and 2% glucose.
Cells expressing hHSF2 and hHSF4 were disrupted by
vortexing with glass beads as described previously [9]. After
centrifugation at 20 000 g for 5 min, protein concentration
was measured by the Bio-Rad assay. The cleared extracts
and recombinant hHSF proteins were subjected to immuno-
blot analysis with antibodies against hHSF2 and hHSF4.
Total RNA was prepared from yeast cells, and mRNA
levels of genes were analyzed by RT-PCR as described pre-
viously [10]. The amounts of PCR products were compared
after normalizing RNA samples to the levels of control
ACT1 mRNA (encoding actin).
Cell culture, transfection, luciferase assay, and
immunoblot analysis
HeLa cells (cell number RCB0007; RIKEN Bio Resource
Center, Ibaraki, Japan) were cultured in minimal essential
medium supplemented with 10% newborn bovine serum at
37 °C in a 5% CO

2
atmosphere. Cells grown in 12-well plates
were transfected using HilyMax (Dojindo Laboratories,
Kumamoto, Japan), with DNA mixtures including 100 or
200 ng of firefly luciferase reporter plasmid, 10 ng of pRL-
TK control plasmid containing the Renilla luciferase gene
driven by the HSV-TK promoter (Promega), 10 or 100 ng of
hHSF–VP16 expression plasmid, and sufficient carrier
pcDNA3.1(+) to bring the total amount of DNA to 1.6 lg.
Cells were cultured for 20–24 h following transfection, and
firefly and Renilla luciferase activities were measured using
the Dual-Luciferase Reporter Assay System (Promega) and a
luminometer (AB-2200-R; ATTO Co., Tokyo, Japan). The
Renilla luciferase activity of each sample was used to nor-
malize firefly luciferase for transfection efficiency.
The expression of hHSF–VP16 fusion proteins in trans-
fected cells was analyzed as follows. Cells were lysed in
buffer containing 50 mm Tris ⁄ Cl (pH 8.0), 150 mm NaCl,
1% Triton X-100, 0.5 mm phenylmethanesulfonyl fluoride,
and protease inhibitor cocktail (Nakarai Tesque, Kyoto,
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1971
Japan). After centrifugation at 20 000 g for 10 min, the
supernatant (100 lg of protein) was subjected to
SDS ⁄ PAGE and immunoblot analysis using an antibody
against VP16 (Abcam).
Acknowledgements
We thank A. Nakai for providing the cDNA clones
and antibodies. This work was supported in part by
Grants-in-Aid for Scientific Research from the Minis-

try of Education, Culture, Sports, Science, and Tech-
nology of Japan to H. Sakurai.
References
1 Pirkkala L, Nykanen P & Sistonen L (2001) Roles of the
heat shock transcription factors in regulation of the heat
shock response and beyond. FASEB J 15, 1118–1131.
2 Voellmy R (2004) On mechanisms that control heat
shock transcription factor activity in metazoan cells.
Cell Stress Chaperones 9, 122–133.
3 Westerheide SD & Morimoto RI (2005) Heat shock
response modulators as therapeutic tools for diseases
of protein conformation. J Biol Chem 280, 33097–
33100.
4 Akerfelt M, Trouillet D, Mezger V & Sistonen L (2007)
Heat shock factors at a crossroad between stress and
development. Ann NY Acad Sci 1113, 15–27.
5 Morimoto RI (2008) Proteotoxic stress and inducible
chaperone networks in neurodegenerative disease and
aging. Genes Dev 22, 1427–1438.
6 Hahn JS, Hu Z, Thiele DJ & Iyer VR (2004) Genome-
wide analysis of the biology of stress responses through
heat shock transcription factor. Mol Cell Biol 24, 5249–
5256.
7 Yamamoto A, Mizukami Y & Sakurai H (2005) Identi-
fication of a novel class of target genes and a novel type
of binding sequence of heat shock transcription factor
in Saccharomyces cerevisiae. J Biol Chem 280, 11911–
11919.
8 Eastmond DL & Nelson HC (2006) Genome-wide anal-
ysis reveals new roles for the activation domains of the

Saccharomyces cerevisiae heat shock transcription factor
(Hsf1) during the transient heat shock response. J Biol
Chem 281, 32909–32921.
9 Hashikawa N, Yamamoto N & Sakurai H (2007) Dif-
ferent mechanisms are involved in the transcriptional
activation by yeast heat shock transcription factor
through two different types of heat shock elements.
J Biol Chem 282, 10333–10340.
10 Sakurai H & Takemori Y (2007) Interaction between
heat shock transcription factors (HSFs) and divergent
binding sequences: binding specificities of yeast HSFs
and human HSF1. J Biol Chem 282, 13334–13341.
11 McMillan DR, Xiao X, Shao L, Graves K & Benjamin
IJ (1998) Targeted disruption of heat shock transcrip-
tion factor 1 abolishes thermotolerance and protection
against heat-inducible apoptosis. J Biol Chem 273,
7523–7528.
12 Xiao X, Zuo X, Davis AA, McMillan DR, Curry BB,
Richardson JA & Benjamin IJ (1999) HSF1 is required
for extra-embryonic development, postnatal growth and
protection during inflammatory responses in mice.
EMBO J 18, 5943–5952.
13 Dai C, Whitesell L, Rogers AB & Lindquist S (2007)
Heat shock factor 1 is a powerful multifaceted modifier
of carcinogenesis. Cell 130, 1005–1018.
14 Reinke H, Saini C, Fleury-Olela F, Dibner C, Benjamin
IJ & Schibler U (2008) Differential display of DNA-
binding proteins reveals heat-shock factor 1 as a circa-
dian transcription factor. Genes Dev 22, 331–345.
15 Xing H, Wilkerson DC, Mayhew CN, Lubert EJ,

Skaggs HS, Goodson ML, Hong Y, Park-Sarge OK &
Sarge KD (2005) Mechanism of hsp70i gene bookmark-
ing. Science 307, 421–423.
16 Kallio M, Chang Y, Manuel M, Alastalo TP, Rallu M,
Gitton Y, Pirkkala L, Loones MT, Paslaru L, Larney S
et al. (2002) Brain abnormalities, defective meiotic chro-
mosome synapsis and female subfertility in HSF2 null
mice. EMBO J 21, 2591–2601.
17 Chang Y, Ostling P, Akerfelt M, Trouillet D, Rallu M,
Gitton Y, El Fatimy R, Fardeau V, Le Crom S, Mor-
ange M et al. (2006) Role of heat-shock factor 2 in
cerebral cortex formation and as a regulator of p35
expression. Genes Dev
20, 836–847.
18 Akerfelt M, Henriksson E, Laiho A, Vihervaara A,
Rautoma K, Kotaja N & Sistonen L (2008) Promoter
ChIP-chip analysis in mouse testis reveals Y chromo-
some occupancy by HSF2. Proc Natl Acad Sci USA
105, 11224–11229.
19 Nakai A, Tanabe M, Kawazoe Y, Inazawa J, Morimoto
RI & 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.
20 Bu L, Jin Y, Shi Y, Chu R, Ban A, Eiberg H, Andres
L, Jiang H, Zheng G, Qian M et al. (2002) Mutant
DNA-binding domain of HSF4 is associated with auto-
somal dominant lamellar and Marner cataract. Nat
Genet 31, 276–278.
21 Fujimoto M, Izu H, Seki K, Fukuda K, Nishida T,
Yamada S, Kato K, Yonemura S, Inouye S & Nakai A

(2004) HSF4 is required for normal cell growth and
differentiation during mouse lens development. EMBO
J 23, 4297–4306.
22 Min JN, Zhang Y, Moskophidis D & Mivechi NF
(2004) Unique contribution of heat shock transcription
factor 4 in ocular lens development and fiber cell differ-
entiation. Genesis 40, 205–217.
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1972 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS
23 Tanabe M, Sasai N, Nagata K, Liu XD, Liu PC, Thiele
DJ & Nakai A (1999) The mammalian HSF4 gene gen-
erates both an activator and a repressor of heat shock
genes by alternative splicing. J Biol Chem 274, 27845–
27856.
24 Frejtag W, Zhang Y, Dai R, Anderson MG & Mivechi
NF (2001) Heat shock factor-4 (HSF-4a) represses basal
transcription through interaction with TFIIF. J Biol
Chem 276, 14685–14694.
25 Tu N, Hu Y & Mivechi NF (2006) Heat shock tran-
scription factor (Hsf)-4b recruits Brg1 during the G1
phase of the cell cycle and regulates the expression
of heat shock proteins. J Cell Biochem 98, 1528–
1542.
26 Kroeger PE, Sarge KD & Morimoto RI (1993) Mouse
heat shock transcription factors 1 and 2 prefer a tri-
meric binding site but interact differently with the
HSP70 heat shock element. Mol Cell Biol 13, 3370–
3383.
27 Kroeger PE & Morimoto RI (1994) Selection of new
HSF1 and HSF2 DNA-binding sites reveals difference

in trimer cooperativity. Mol Cell Biol 14, 7592–
7603.
28 Liu XD, Liu PC, Santoro N & Thiele DJ (1997)
Conservation of a stress response: human heat shock
transcription factors functionally substitute for yeast
HSF. EMBO J 16, 6466–6477.
29 Perisic O, Xiao H & Lis JT (1989) Stable binding of
Drosophila heat shock factor to head-to-head and tail-
to-tail repeats of a conserved 5 bp recognition unit. Cell
59, 797–806.
30 Hashikawa N, Mizukami Y, Imazu H & Sakurai H
(2006) Mutated yeast heat shock transcription
factor activates transcription independently of
hyperphosphorylation. J Biol Chem 281, 3936–
3942.
31 Sarge KD, Zimarino V, Holm K, Wu C & Morimoto
RI (1991) Cloning and characterization of two mouse
heat shock factors with distinct inducible and
constitutive DNA-binding ability. Genes Dev 5, 1902–
1911.
32 Yoshima T, Yura T & Yanagi H (1998) Function of
the C-terminal transactivation domain of human
heat shock factor 2 is modulated by the adjacent
negative regulatory segment. Nucleic Acids Res 26,
2580–2585.
33 Zuo J, Rungger D & Voellmy R (1995) Multiple layers
of regulation of human heat shock transcription
factor 1. Mol Cell Biol 15, 4319–4330.
34 Fujimoto M, Oshima K, Shinkawa T, Wang BB,
Inouye S, Hayashida N, Takii R & Nakai A (2008)

Analysis of HSF4 binding regions reveals its necessity
for gene regulation during development and heat shock
response in mouse lenses. J Biol Chem 283, 29961–
29970.
35 Hu Y & Mivechi NF (2006) Association and regulation
of heat shock transcription factor 4b with both extracel-
lular signal-regulated kinase mitogen-activated protein
kinase and dual-specificity tyrosine phosphatase
DUSP26. Mol Cell Biol 26, 3282–3294.
36 Rabindran SK, Haroun RI, Clos J, Wisniewski J & Wu
C (1993) Regulation of heat shock factor trimer forma-
tion: role of the conserved leucine zipper. Science 259,
230–234.
37 Zuo J, Baler R, Dahl G & Voellmy R (1994) Activation
of the DNA-binding ability of human heat shock tran-
scription factor 1 may involve the transition from an
intramolecular to an intermolecular triple-stranded
coiled-coil structure. Mol Cell Biol 14, 7557–7568.
38 Liu PC & Thiele DJ (1999) Modulation of human heat
shock factor trimerization by the linker domain. J Biol
Chem 274, 17219–17225.
39 Drees BL, Grotkop EK & Nelson HC (1997) The
GCN4 leucine zipper can functionally substitute for the
heat shock transcription factor’s trimerization domain.
J Mol Biol 273, 61–67.
40 Littlefield O & Nelson HC (1999) A new use for the
‘wing’ of the ‘winged’ helix–turn–helix motif in the
HSF–DNA cocrystal. Nat Struct Biol 6, 464–470.
41 Ahn SG, Liu PC, Klyachko K, Morimoto RI & Thiele
DJ (2001) The loop domain of heat shock transcription

factor 1 dictates DNA-binding specificity and responses
to heat stress. Genes Dev 15, 2134–2145.
42 Ahn SG & Thiele DJ (2003) Redox regulation of mam-
malian heat shock factor 1 is essential for Hsp gene
activation and protection from stress. Genes Dev 17,
516–528.
43 Trinklein ND, Chen WC, Kingston RE & Myers RM
(2004) Transcriptional regulation and binding of heat
shock factor 1 and heat shock factor 2 to 32 human
heat shock genes during thermal stress and differentia-
tion. Cell Stress Chaperones 9, 21–28.
44 Ostling P, Bjo
¨
rk JK, Roos-Mattjus P, Mezger V &
Sistonen L (2007) Heat shock factor 2 (HSF2) contrib-
utes to inducible expression of hsp genes through inter-
play with HSF1. J Biol Chem 282, 7077–7086.
45 Loison F, Debure L, Nizard P, Le Goff P, Michel D &
Le Drean Y (2006) Up-regulation of the clusterin gene
after proteotoxic stress: implication of HSF1–HSF2
heterocomplexes. Biochem J 395, 223–231.
46 Cvekl A & Duncan MK (2007) Genetic and epigenetic
mechanisms of gene regulation during lens development.
Prog Retin Eye Res 26, 555–597.
47 Brakenhoff RH, Aarts HJ, Reek FH, Lubsen NH &
Schoenmakers JG (1990) Human c-crystallin genes: a
gene family on its way to extinction. J Mol Biol 216,
519–532.
48 Sakurai H & Fukasawa T (2003) Artificial recruitment
of certain mediator components affects requirement of

basal transcription factor IIE. Genes Cells 8, 41–50.
N. Yamamoto et al. HSE-type specific recognition by human HSFs
FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS 1973
Supporting information
The following supplementary material is available:
Fig. S1. Electrophoretic mobility shift assay of hHSF–
HSE complexes.
Fig. S2. Electrophoretic mobility shift assay of hHSF–
VP16 polypeptides.
This supplementary material can be found in the
online version of this article.
Please note: Wiley-Blackwell is not responsible for
the content or functionality of any supplementary
materials supplied by the authors. Any queries (other
than missing material) should be directed to the corre-
sponding author for the article.
HSE-type specific recognition by human HSFs N. Yamamoto et al.
1974 FEBS Journal 276 (2009) 1962–1974 ª 2009 The Authors Journal compilation ª 2009 FEBS