MINIREVIEW Novel aspects of heat shock factors: DNA recognition, chromatin modulation and gene expression Hiroshi Sakurai and Yasuaki Enoki Department of Clinical Laboratory Science, Kanazawa University Graduate School of Medical Science, Ishikawa, Japan Background The heat shock factor (HSF) in eukaryotes is involved not only in heat shock protein (HSP) gene expression and stress resistance, but also in the expression of genes with roles in cell maintenance and differentia- tion, as well as in developmental processes. HSF forms a homotrimer that binds to gene promoters containing a heat shock element (HSE), which is composed of multiple inverted repeats of the pentanucleotide motif nGAAn. Functional conservation of HSFs among eukaryotes has been revealed by the finding that HSFs from various organisms, including insects, mammals and plants, can substitute for yeast HSF in Saccharo- myces [1–4]. HSF proteins contain two evolutionarily conserved functional modules: the DNA-binding domain (DBD) at the amino-terminus and the oligomerization domain in the central region of the protein [1,4]. The HSF DBD belongs to the ‘winged’ helix-turn-helix family of DNA-binding proteins and contains a three-helix bun- dle capped by a four-stranded antiparallel b-sheet, and a flexible loop or ‘wing’ with a less ordered structure (Fig. 1) [5,6]. The second and third a-helices comprise the helix-turn-helix motif. The oligomerization domain consists of arrays of hydrophobic heptad repeats (HRs), characteristic of helical coiled-coil structures [1,4,7]. The HRs are divided into two subdomains: HR-A and HR-B. The amino-terminal HR-A has the potential to form trimers independently of HR-B, and the carboxy-terminal HR-B can form large oligomers [7]. Keywords chromatin; heat shock element; heat shock transcription factor; histone; protein–DNA interactions 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: sakurai@kenroku.kanazawa-u.ac.jp (Received 10 May 2010, revised 9 July 2010, accepted 23 July 2010) doi:10.1111/j.1742-4658.2010.07829.x Heat shock factor (HSF) is an evolutionarily conserved stress-response reg- ulator that activates the transcription of heat shock protein genes, whose products maintain protein homeostasis under normal physiological condi- tions, as well as under conditions of stress. The promoter regions of the target genes contain a heat shock element consisting of multiple inverted repeats of the pentanucleotide sequence nGAAn. A single HSF of yeast can bind to heat shock elements that differ in the configuration of the nGAAn units and can regulate the transcription of various genes that func- tion not only in stress resistance, but also in a broad range of biological processes. Mammalian cells have four HSF family members involved in dif- ferent, but in some cases similar, biological functions, including stress resis- tance, cell differentiation and development. Mammalian HSF family members exhibit differential specificity for different types of heat shock ele- ments, which, together with cell type-specific expression of HSFs is impor- tant in determining the target genes of each HSF. This minireview focuses on the molecular mechanisms of DNA recognition, chromatin modulation and gene expression by yeast and mammalian HSFs. Abbreviations 3P, three perfect repeats; DBD, DNA-binding domain; HDAC1, ; HDAC2, ; HR, hydrophobic heptad repeat; HSE, heat shock element; HSF, heat shock factor; HSP, heat shock protein; ncRNA, noncoding RNA; PARP, poly(ADP)-ribose polymerase; Pol II, RNA polymerase II; SAGA, Spt-Ada-Gcn5 acetyltransferase; TFIIA, general transcription factor IIA. 4140 FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS Trimerization of HSF polypeptides is a prerequisite for binding to the HSE. Under physiological condi- tions, stress-inducible HSFs, such as Drosophila HSF and mammalian HSF1, are found in an inactive mono- meric form. In vitro analysis has shown that inactive monomeric HSF directly senses heat and oxidative stress, and becomes able to trimerize and bind to the HSE [8,9]. Drosophila HSF and mammalian HSF1 have a third HR, HR-C, adjacent to the carboxy-ter- minus of the protein [10]. The HR-C maintains HSFs in a monomeric form by suppressing trimerization through intramolecular coiled-coil interactions with the HR-A ⁄ B. Under stress conditions, the HR-A ⁄ B–HR-C interaction is disrupted, thereby leading to intermolec- ular coiled-coil interactions of the HR-A ⁄ B, although the detailed molecular mechanisms are unknown [10–12]. In addition, protein–protein interactions with HSPs and other proteins can regulate the monomeric– trimeric transition of HSF in cells [2]. Transcription activation domains have been located in the HSFs of several organisms; however, the transactivating ability of the domain itself, even from stress-inducible HSFs, is not stress sensitive [1,2,4]. Rather, another region of HSF possesses a stress-regulated inhibitory role that represses the transactivating ability under physiological conditions [13]. Therefore, the transcriptional activity of HSF is regulated at two steps: DNA-binding; and acquisition of the transactivating ability. HSF–HSE interactions The HSE is composed of multiple inverted repeats of the nGAAn pentanucleotide that are positioned con- tinuously without spacing. Because each individual DBD of a trimeric HSF binds to the sequence, the typ- ical HSE contains at least three nGAAn units [4]. The helical repeat of the DNA (10.5 bp), the length of the unit (5 bp) and the inverted nature of the repeats, posi- tions all of the GAA units on the same side of the double helix, with the central GAA oriented in the opposite direction. In vitro, Drosophila HSF is capable of binding to two repeat sequences, but the affinity is significantly lower than binding to three repeat sequences [14]. The crystal structure of the yeast Kluy- veromyces HSF DBD complexed with the sequence ggTTCtaGAAcc (containing one set of the nGAAn inverted repeat) has been determined [15]. In the com- plex, the third helix of the DBD is docked into the major groove, approximately perpendicular to the heli- cal axis of the DNA. The evolutionarily conserved arginine residue in this recognition helix makes two hydrogen bonds with G 2 of the pentanucleotide n 1 G 2 A 3 A 4 n 5 (Fig. 1). Additional direct contacts include van der Waals bonds between the arginine and a conserved serine residue and the thymine comple- mentary to the third A 3 . The remaining contacts with the DNA are to the phosphate backbone. Consistent with the structural data, in vitro-binding assays have shown that the order of importance of the bases in the nGAAn repeat is G 2 >A 3 >A 4 [16–18]. In addition, HSF binds with highest affinity when the n 1 residue is adenine. HSF trimers bind to long arrays of the nGAAn sequence in a cooperative manner [19]. In an electro- phoretic mobility shift assay, an HSE containing four continuous nGAAn units (four perfect repeats) was found to bind two trimers of Saccharomyces HSF, Drosophila HSF, or human HSF1, with two subunits PAFLTKLWTLVSDPDTDALICWSPSGNSFHVFDQGQFAKEVLPKYFKHNNMASFVRQLNMYGFRKVVHIEQGGLVKPERDDTEFQHPCFLR S3H3 S4 Turn Wing H1 S2 H2 Linker PAFVNKLWSMLNDDSNTKLIQWAEDGKSFIVTNREEFVHQILPKYFKHSNFASFVRQLNMYGWHKVQDVKSGSIQSSSDDKWQFENENFIR Sc Hs1 PAFVNKLWSMVNDKSNEKFIHWSTSGESIVVPNRERFVQEVLPKYFKHSNFASFVRQLNMYGWHKVQDVKSGSMLSNNDSRWEFENENFKR Kl PAFLAKLWRLVDDADTNRLICWTKDGQSFVIQNQAQFAKELLPLNYKHNNMASFIRQLNMYGFHKITSIDNGGL-RFDRDEIEFSHPFFKR Dm PAFLSKLWTLVEETHTNEFITWSQNGQSFLVLDEQRFAKEILPKYFKHNNMASFVRQLNMYGFRKVVHIDSGIVKQERDGPVEFQHPYFKQ Hs2 PAFLGKLWALVGDPGTDHLIRWSPSGTSFLVSDQSRFAKEVLPQYFKHSNMASFVRQLNMYGFRKVVSIEQGGLLRPERDHVEFQHPSFVR Hs4 DNA contact Aromatic-aromatic interaction Disulfide bond formation in mammalian HSF1 S1 PHFLTKLWILVDDAVLDHVIRWGKDGHSFQIVNEETFAREVLPKYFKHNKITSFIRQLNMYGSRKVFALQTEKTSQENKISIEFQHPLFKR Mm3 Fig. 1. Amino acid sequences of HSF DBDs. The DBD contains three a-helices (H1, H2 and H3), four b-sheets (S1, S2, S3 and S4), a ‘turn’ of the helix-turn-helix motif and a disordered loop referred to as a ‘wing’. The linker region connects between the DBD and the HR-A ⁄ B motifs. The amino acid sequences of the DBDs from Kluyveromyces HSF (Kl), Saccharomyces HSF (Sc), Drosophila HSF (Dm), human HSF1 (Hs1), HSF2 (Hs2) and HSF4 (Hs4), and mouse HSF3 (Mm3) are shown. Residues making contacts with DNA are shown in red, residues that have aromatic–aromatic interactions are shown in blue, cysteine residues that form a disulfide bond are shown in green and conserved residues are shown in black. H. Sakurai and Y. Enoki HSF–HSE interaction FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 4141 (possibly one from each trimer) not making contact with the DNA (Fig. 2A, B) [14,20,21]. The dissociation constant of Drosophila HSF for an HSE containing six or more units (2 · 10 )15 m at 25 °C) is significantly lower than that for an HSE containing three perfect repeats (3P) (4 · 10 )12 m at 25 °C) [19]. Multiple inverted repeats of nGAAn found at the promoter regions of many HSP genes provide high affinity-bind- ing sites for HSF. The DBD–DBD interaction is important in the HSF–HSF and HSF–HSE interactions. In HSF DBD– HSE co-crystals, the protein–protein interface consists of the helix 2 amino-terminus, the turn and the wing [15]. Unlike other winged helix-turn-helix proteins, the wing of the HSF does not appear to contact the DNA. The wing of the Saccharomyces HSF is necessary for efficient binding of a single trimer to the HSE [22], and that of mammalian HSF1 is involved in trimer–trimer interactions [23]. The tryptophan and phenylalanine residues in the mammalian HSF1 DBD form intermo- lecular aromatic–aromatic interactions, which stimulate trimerization (Fig. 1) [24]. Furthermore, two cysteine residues near the aromatic amino acids in HSF1 form intermolecular disulfide bonds in a stress-inducible manner [9,24]. The aromatic amino acids are conserved in HSFs of various organisms and would generally enhance HSF trimerization; however, the disulfide bond formation is specific for mammalian HSF1 because the cysteines are not conserved in the HSFs of other organisms (Fig. 1). Mutations in the linker region that connects the DBD and HR-A ⁄ B have been shown to alter the monomer–trimer equilibrium of mamma- lian HSF1, indicating that this region is crucial for HSF1 trimerization [25]. HSE-specific transcriptional regulation by Saccharomyces HSF Saccharomyces HSF constitutively forms a trimer, localizes in the nucleus and binds to and regulates basal expression via HSEs of target genes [26]. Deletion of the HSF gene is lethal to cells, even at physiological temperatures. The genes targeted by HSF encode proteins that function in protein folding and degradation, detoxification, energy generation, carbohydrate metabolism and cell wall organization [27,28]. Heat shock and oxidative stress lead to enhanced binding of HSF to HSEs in yeast [27,29]. However, whether the trimerization status and ⁄ or the affinity of each DBD for nGAAn changes under stress conditions is not known. The activation domains at the amino- and carboxy-termini are required for the transient and sustained heat shock responses, respectively [30]. Phosphorylation of HSF, regulated by heat and oxidative stress, is involved in the activation and inactivation of the transactivating ability [26]. Interestingly, Saccharomyces HSF is unu- sual among transcriptional activators because it can bypass a need for critical general transcription factors and co-activators, including general transcription fac- tor IIA (TFIIA) [31], Kin28 (a C-terminal-domain kinase of TFIIH) [32], the Taf9 subunit of TFIID and Spt-Ada-Gcn5 acetyltransferase (SAGA) [33,34], and the Med17 and Med22 subunits of Mediator [32]. Although the archetypical HSE is a sequence of con- tinuous perfect inverted repeats of nGAAn (perfect- type), the Saccharomyces HSF tolerates the presence of gaps between the units [28,35]. The gap-type HSE con- sists of two inverted nGAAn units followed by another unit after a gap of 5 bp, and the step-type HSE con- sists of three direct units, each interrupted by 5 bp (Fig. 2A). In the discontinuous gap- and step-type Two trimers on 4P-type HSE One trimer on step-type HSE One trimer on gap-type HSE One trimer on 3P-type HSE 4P type A B a b c d nTTCnnGAAnnTTCnnGAAn 3P type nTTCnnGAAnnTTCn Gap type nTTCnnGAAnnnnnnnGAAn Step type nTTCnnnnnnnTTCnnnnnnnTTCn Fig. 2. HSE types and HSF–HSE interactions. (A) Nucleotide sequences of the different types of HSEs. The GAA and inverted TTC sequences are indicated by bold uppercase letters with arrows. (B) HSF binding to the different types of HSEs. Thick lines represent DNA with GAA and inverted TTC sequences (white squares). Gray ovals represent HSF monomers making contact with the sequences, and light gray ovals represent HSF monomers that do not make contact with the DNA. On the gap- and step-type HSEs, a single HSF trimer is in equilibrium between the two bind- ing states that are shown. HSF–HSE interaction H. Sakurai and Y. Enoki 4142 FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS HSEs, all three GAA units are positioned on the same side of the double helix, as in the continuous HSE. HSF binds to discontinuous HSEs with similar affinity for the continuous 3P-type HSE [20,28]. On discontin- uous HSEs, the HSF trimer dissociates from two units and quickly rebinds to another two units, thereby sta- bilizing the protein–DNA complex (Fig. 2B). Gap-type HSEs are involved in moderate stress-induced tran- scription, whereas step-type HSEs are involved in basal constitutive transcription and in low-level activation [28,36]. These discontinuous HSEs are widely used as HSF-binding sequences, at least in the Saccharomyces genome, because approximately half of the HSF target genes contain these types of HSE [37]. The number of HSF trimers bound to an HSE affects the subsequent acquisition of the transactivat- ing capacity. In HSF, a C-terminal modulator domain is necessary for stress-induced hyperphosph- orylation of the protein and for transcriptional activa- tion of a gene by a single trimer bound to HSEs with three nGAAn units (3P-, gap-, or step-type). How- ever, the C-terminal modulator, and thereby hyper- phosphorylation, are not necessary for transcriptional activation by trimers cooperatively bound to HSEs containing four or more repeat units [20,29]. The HR-A ⁄ B is indispensable for binding of a single tri- mer to three-unit HSEs and for HSF hyperphosph- orylation, but HR-A ⁄ B is not required for binding HSEs with four or more units or for transactivation at these HSEs. These results indicate that trimer–tri- mer interactions are implicated not only in the coop- erative binding of HSF to the DNA, but also in transcriptional activation [36]. Although the role of hyperphosphorylation in transcriptional activation is not understood, this modification may induce a con- formational change in a single trimer that converts it to an active form. As hyperphosphorylation is not required for activation at HSEs containing four or more repeats, a similar conformational change may be mediated by trimer–trimer interactions. The essen- tial role of the HR-A ⁄ B is to maintain the structural integrity of the HSF trimer, because an HSF deriva- tive containing the dimerization domain of transcription factor Gcn4 instead of the HR-A ⁄ Bis hyperphosphorylated and is capable of activating transcription via three-unit HSEs [36]. HSE binding by the four mammalian HSFs Mammalian HSF1 is a bona fide stress-inducible HSF: in response to stress, monomeric HSF1 located in the cytoplasm trimerizes, translocates into the nucleus and binds to the HSEs of HSP genes [10,11]. Under physi- ological conditions, the activity of the carboxy-termi- nal activation domain is repressed by the regulatory region of the protein, and repression is relieved in response to heat shock [13]. The transcriptional activ- ity of HSF1 is further regulated by covalent modifica- tions, including phosphorylation, SUMOylation and acetylation [2]. It should be noted, however, that HSF1-null mice exhibit multiple phenotypes, including placentation defects, growth retardation and exagger- ated production of the pro-inflammatory cytokine tumour necrosis factor-a, without affecting basal HSP expression [38]. Although HSF1 gains DNA-binding and transactivating abilities upon stress, a little HSF1 trimer is present constitutively in cells and binds to and regulates the expression of genes without any apparent stress intervention [1,3]. HSF2 and HSF4 may not be directly involved in the stress response, but rather in cell differentiation and development [3]. In addition to covalent modifications, including phosphorylation, SUMOylation and ubiquiti- nation, the regulatory roles of HSF2 and HSF4 are dependent on their cellular concentrations [2]. The non-DNA-binding form of HSF2 exists primarily in the cytoplasm as a dimer, which has been suggested to require an interaction between HR-A⁄ B and HR-C. Although the signals responsible for activation of HSF2 remain enigmatic, activated HSF2 trimerizes and binds to HSEs [39]. HSF2 possesses an activation domain at the carboxy-terminus, but the activity is sig- nificantly lower than the activation domain of HSF1 [2]. During development, HSF2 is important for neural specification and spermatogenesis [3]. HSF4 lacks the HR-C, and trimeric HSF4 is able to constitutively bind to the HSEs of target genes [40]. Alternative splicing of the HSF4 transcript results in the production of two isoforms, HSF4a and HSF4b. HSF4b has the potential to activate transcription, whereas HSF4a does not [2]. HSF4 is required for ocular lens develop- ment and fibre cell differentiation [3]. The HSF3 gene was recently identified in mouse as an orthologue of the chicken HSF3 gene; however, the human HSF3 gene is a pseudogene [41]. Mouse HSF3 translocates into the nucleus upon heat shock and may activate nonclassical heat shock genes, but does not activate classical heat shock genes (i.e. HSP genes). HSF1 prefers long arrays of the nGAAn unit, while HSF2 prefers short arrays [18]. In addition, a four per- fect repeat-type HSE binds two trimers of HSF1 or a single trimer of HSF2 [21]. These differences are related to differences in the cooperativity of the tri- mers; the wing regions of the HSF1 and HSF2 DBDs H. Sakurai and Y. Enoki HSF–HSE interaction FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 4143 affect the cooperativity [23]. HSF4 trimers exhibit weak cooperativity [21,40]. The genomic sequences to which HSF1 binds are continuous perfect inverted repeats of nGAAn [42]. In contrast, the HSF4-binding consensus sequence is more ambiguous than that of HSF1 and HSF2 [43]. The human cA-crystallin and cC-crystallin promoters contain HSEs of different configurations, and the for- mer is recognized by both HSF1 and HSF4, while the latter is preferentially recognized by HSF4 [21]. The mouse p35 gene, a specific target of HSF2, contains a putative HSE that diverges from the canonical HSE [44]. The HSE specificity of three HSF members has been systematically characterized using model 3P-, gap- and step-type HSEs [21]. HSF1 preferentially binds to continuous 3P-type HSE, HSF2 exhibits a slightly higher binding affinity to discontinuous HSEs than does HSF1, and HSF4 efficiently recognizes dis- continuous HSEs. When HSF1, HSF2 and HSF4 are expressed in yeast cells, transcription of various genes is differentially regulated by the three HSFs and corre- lated with the type of HSE, namely perfect, gap, or step [21,45]. These observations indicate that the con- figuration of the nGAAn unit is an important determi- nant of HSF–HSE interactions. This is consistent with the notion that although DNA-binding transcription factors of the same family bind the same highest-affin- ity sites, they prefer different lower-affinity sites, and that low-affinity sites do contribute to factor binding and gene expression [46]. Differences in the HSE-binding specificity are in part explained by the oligomerization of HSFs. When an amino acid substitution is made in the HR-A⁄ Bof HSF4, the protein exhibits a reduced ability to trimerize and is unable to bind to discontinuous gap- and step-type HSEs, but binding to the continuous 3P-type HSE is not affected [21]. In contrast, an amino acid substitution in the HR-C of HSF1 enables HSF1 to constitutively form trimers that bind to discontinuous HSEs [45]. Similar results have been obtained by introducing oligomeriza- tion-prone mutations into the HSF1 DBD [45]. On discontinuous HSEs, in which two pairs of two nGAAn units provide a platform for binding two DBDs of a single HSF trimer, stable oligomerization inhibits disso- ciation and ⁄ or stimulates rebinding of HSF (Fig. 2B). Interestingly, HSF1 and HSF2 form heterotrimers via the HR-A ⁄ B regions [47–49]. When HSF1 and HSF2 bind as a complex to satellite III DNA in nuclear stress bodies, elevated expression of HSF2 causes transcriptional activation of noncoding satellite III RNA by the heterotrimer [49]. In addition, HSF2 modulates the activity of stress-induced HSF1 in a gene-specific manner: the presence of HSF2 leads to stimulation of HSP70 expression but to inhibition of HSP40 and HSP110 expression [48]. Differences between HSF1 and HSF2 in HSE specificity may enable an HSF1–HSF2 heterotrimer to have a distinct HSE specificity. Binding of HSF to chromatin and changes in chromatin structure Packaging of DNA into nucleosome arrays, which can be folded into higher-order structures, provides way with which to tightly control access to the DNA sequence. Many sequence-specific DNA-binding fac- tors, including HSF, as well as general transcription factors, cannot bind to nucleosomal DNA [50]. In cells, there are various proteins and protein complexes that change chromatin structure: histone variants, histone-modifying enzymes, histone chaperones and chromatin remodelling complexes [51]. In yeast In Saccharomyces cerevisiae cells, heat shock induces the binding of general transcription factors and RNA polymerase II (Pol II) to the HSP promoter, resulting in rapid RNA synthesis [52]. Chromatin regulators, including the SAGA histone–acetylase complex, the Rpd3 histone–deacetylase complex, ATP-dependent chromatin remodelling complexes (SWI ⁄ SNF, ISW1 and RSC) and histone chaperones (Asf1, Spt6 and Spt16), are involved in histone acetylation and eviction [52–55]. Stress-inducible or constitutive binding of HSF to a promoter is dependent on the structure of the HSE. The HSP12 promoter possesses a low-affinity HSE and a binding site for the general stress response transcription factors Msn2 and Msn4 (Msn2 ⁄ 4). However, HSF and Msn2 ⁄ 4 are not preloaded on this promoter. Heat-induced binding of HSF to the pro- moter in nucleosomes is assisted by Msn2 ⁄ 4 and the chromatin remodelling complexes SWI ⁄ SNF, ISW1 and RSC [56,57]. These remodellers play partially overlapping, but not redundant, function [57]. In con- trast, the HSP82 promoter possesses a high-affinity HSE, which constitutively binds HSF. The promoter contains remodelled nucleosomes exhibiting DNase I hypersensitivity, and the gene is transcribed at low levels under physiological conditions [58,59]. Binding of HSF to the HSP82 promoter has been suggested to occur soon after DNA replication and before nucleo- somes have assembled, in the S ⁄ G 2 phase of the cell cycle [59], depending on ISW1 and RSC activities [57]. In response to heat shock, HSF occupancy at HSE- containing promoters increases, the SAGA and HSF–HSE interaction H. Sakurai and Y. Enoki 4144 FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS SWI ⁄ SNF complexes are recruited to both the pro- moter and coding regions, and histones are acetylated and displaced [56–58]. However, the presence of SAGA and SWI ⁄ SNF, and a high density of Pol II within the coding region, are not sufficient to elicit histone dis- placement, suggesting that histone eviction is modu- lated by factors that are not linked to elongating polymerase [55]. Recently, a candidate factor was iden- tified in Drosophila [60, see below]. The Rpd3-contain- ing histone–deacetylase complex, which deacetylates histones to repress transcription, is also recruited to induced genes. It is possible that the opposing activities of histone acetylase and deacetylase modulate chroma- tin structure and fine-tune transcription [58]. The his- tone chaperones Asf1, Spt6 and Spt16 are involved in histone eviction and redeposition, and inactivation of Spt16 leads to sustained transcription [61]. In Drosophila Studies in Drosophila provide many insights into how HSF and Pol II overcome the nucleosome barrier. In brief, on polytene chromosomes of Drosophila, heat shock leads to visible changes in the heat shock loci, termed ‘‘puffing’’. Puffing reflects the changes in chro- matin structure that lead to the disruption of nucleo- somes along the coding region of HSP genes. Unlike yeast HSP genes, the DNase I hypersensitive region of the hsp70 promoter of nonheat-shocked cells contains paused Pol II, which is transcriptionally engaged and paused 20–40 bp downstream of its initiation site [62– 64]. The promoter-bound GAGA factor is required for efficient pause-site entry of Pol II. Upon heat shock, escape of the paused polymerase from the site is trig- gered by HSF-mediated recruitment of positive tran- scription elongation factor-b (P-TEFb), which phosphorylates the polymerase [62–64]. Concomitant nucleosome loss at hsp70 requires poly(ADP)-ribose polymerase (PARP), which is capable of binding nucleo- somes and is important for puff formation [60]. The possible roles of PARP in nucleosome loss are as fol- lows, (a) PARP prebound to nucleosomes is released from chromatin to reverse any repressive effects of the chromatin structure, (b) PARP ADP-ribosylates histones and thus destabilizes nucleosomes or (c) ADP-ribose polymers generated by PARP, which look much like a nucleic acid, bind to histones, facilitating nucleosome dissociation [60]. In mammals Similarly to Drosophila, the promoter region of mam- malian HSP70 is hypersensitive to DNase I and contains paused polymerase [62–64]. Although GAGA factor-like DNA-binding activity has been reported, the involvement of a mammalian GAGA factor-like protein in polymerase pausing has not been docu- mented [65]. Instead, the HSP70 promoter is bound by HSF2 in mitotic cells, which prevents compaction at the site by condensin and maintains the gene in a tran- scription-competent state, thereby enabling robust acti- vation by HSF1 if cells in the early G 1 phase are exposed to a stressor [66]. The bookmarking function of HSF2 is achieved by recruitment of protein phos- phatase 2A to inhibit nearby condesin complexes by dephosphorylation. The bookmarking of the HSP70 promoter and other HSE-containing promoters by HSF2 is different from that of many active promoters, which are marked by binding of TATA-binding pro- tein and recruitment of protein phosphatase 2A [67,68]. In mouse lens, HSF4 binds to nonclassical heat shock genes and induces demethylation of histone H3 lysine 9; methylation of this residue is correlated with transcriptionally repressed chromatin [43]. This HSF4 function is required for heat-induced transcription of these genes, in part by facilitating HSF1 binding via chromatin modification. The roles of HSF1 in chroma- tin modulation have been shown: under unstressed conditions, HSF1 binds to the promoter of the non- heat-shock interleukin-6 gene and functions in the opening of its chromatin structure through recruitment of the CBP histone acetyltransferase and the BRG1- containing SWI ⁄ SNF chromatin remodeller [69]. It is unknown whether binding of HSF4 and HSF1 to DNA is maintained in mitotic chromatin. Upon transcriptional activation, HSF1 bound to the HSP70 promoter recruits the BRG1-containing SWI ⁄ SNF chromatin remodeller and the CBP ⁄ p300 histone acetyltransferases [70,71]. In the HSF1 activa- tion domain, phenylalanine residues are involved in promoting elongation mediated by Pol II and recruit- ment of the SWI⁄ SNF complex, while acidic amino acids are involved in transcriptional initiation [70]. Interestingly, mouse HSF3, which fails to activate clas- sical heat shock genes (including HSP70), does not bind SWI ⁄ SNF, suggesting the importance of the chro- matin remodeller in HSF member-specific transcription [41]. Histone acetylation at the HSP70 promoter occurs under treatment with heat and arsenite; how- ever, phosphorylation of histone H3 occurs as a result of treatment with arsenite [72]. The arsenite-induced phosphorylation of H3 and HSP70 transcription is dependent on p38 mitogen-activated protein kinase activation. However, it remains to be explored how phosphorylation of H3 is involved in arsenite-induced, but not in heat-induced, transcription by HSF1. H. Sakurai and Y. Enoki HSF–HSE interaction FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 4145 During attenuation of RNA synthesis, the chaperone HSP70 behaves as a corepressor of HSF1 in a nega- tive-feedback loop and interacts with CoREST, a com- ponent of a histone deacetylase complex containing HDAC1 and HDAC2 [73]. CoREST binds to the HSP70 promoter at low levels under physiological conditions and suppresses basal expression. Global transcriptional changes and histone modifications in heat-shocked mammalian cells When cells are subjected to heat shock, transcription of many mRNA genes is rapidly repressed. In contrast, the levels of the noncoding RNAs (ncRNAs) – B2 RNA and Alu RNA – which are transcribed from short interspersed elements (SINEs) by RNA polymer- ase III, increase in mouse and human cells, respec- tively. Both ncRNAs are transacting transcriptional repressors during the heat shock response: they block transcription by binding Pol II and entering complexes at some mRNA gene promoters [74,75]. On the acti- vated mouse and human HSP70 promoters, however, Pol II is present, but B2 RNA and Alu RNA are absent, implying that mechanisms must exist to over- come ncRNA repression on the promoter, for example, degradation of the ncRNA repressors by an RNase, removal of the ncRNA repressors by an RNA helicase, or binding of another ncRNA to Pol II to displace the ncRNA repressors [75]. In this context, it is notewor- thy that the ncRNA heat shock RNA-1 (HSR1) is an important regulator for heat-induced HSF1 trimeriza- tion and HSP expression [76]. A global deacetylation of core histones is associated with heat shock. This is achieved by HSF1 and the his- tone deacetylases HDAC1 and HDAC2 [77]. In heat- shocked cells, HSF1 binds to HDAC1 and HDAC2, and their histone-deacetylase activities increase. The heat-induced deacetylation of histone H4 is not observed in HSF1 knockout mouse cells. In addition, loss of HSF1 causes hyperacetylation of histone H4 in nonheat-shocked cells [77]. The relationship between HSF1 and histone deacetylases has also been shown in human breast cancer cells: treatment of cells with the transforming factor heregulin leads to increases in the levels of HSF1 and MTA1 (a component of the NuRD complex containing HDAC1 and HDAC2) proteins and to binding of HSF1 to MTA1 [78]. The complex is recruited to the promoters of estrogen-responsive genes and participates in the repression of transcrip- tion, an effect linked to metastasis [78]. Therefore, HSF1 is a potent regulator of global histone acetyla- tion ⁄ deacetylation in stressed and unstressed cells. Conclusion In the HSF trimer, binding of each DBD to a 5-bp unit (nGAAn) is necessary to achieve a stable protein– DNA interaction, because monomeric HSF binds to the repeat unit with significantly less affinity than does the trimer. In addition to the continuous repeat units, three discontinuous units in the gap- and step-type HSEs can be bound by HSFs. Efficient trimerization of HSFs is important to establish interactions with the gap- and step-type HSEs. The different types of HSEs result in the induction, by yeast HSFs, of different lev- els of gene expression and probably in distinguishing target genes by mammalian HSFs. In response to stress, yeast HSF bound to an HSE recruits the gen- eral transcription factors and Pol II to the promoter, while the target promoters of mammalian HSF1 con- tain paused polymerase that switches to productive elongation in response to activated HSF1. Chromatin regulators that disrupt the nucleosome structure facili- tate the binding of HSF, general transcription factors and Pol II to the promoters, and subsequent RNA synthesis by polymerase. In mammalian cells, HSF1 target promoters are also bound by HSF2 and HSF4 under physiological conditions, suggesting that these factors play a positive role in the opening of the chro- matin structure and in stress-induced transcription by HSF1. In addition, HSF1 itself associates with histone deacetylases and regulates global histone acetylation. Therefore, mammalian HSFs function not only as transcription activators but also as chromatin-modu- lating factors. Acknowledgements This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to H. Sakurai. References 1 Fujimoto M & Nakai A (2010) The heat shock factor family and adaptation to proteotoxic stress. FEBS J 277, 4112–4125. 2 Bjo ¨ rk JK & Sistonen L (2010) Regulation of the members of the mammalian heat shock factor family. FEBS J 277, 4126–4139. 3 Abane R & Mezger V (2010) Roles of heat shock factors in gametogenesis and development. FEBS J 277, 4150–4172. 4 Pirkkala L, Nykanen P & Sistonen L (2001) Roles of the heat shock transcription factors in regulation of the HSF–HSE interaction H. Sakurai and Y. Enoki 4146 FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS heat shock response and beyond. FASEB J 15 , 1118– 1131. 5 Harrison CJ, Bohm AA & Nelson HC (1994) Crystal structure of the DNA binding domain of the heat shock transcription factor. Science 263, 224–227. 6 Vuister GW, Kim SJ, Orosz A, Marquardt J, Wu C & Bax A (1994) Solution structure of the DNA-binding domain of Drosophila heat shock transcription factor. Nat Struct Biol 1, 605–614. 7 Peteranderl R, Rabenstein M, Shin YK, Liu CW, Wemmer DE, King DS & Nelson HC (1999) Biochemi- cal and biophysical characterization of the trimerization domain from the heat shock transcription factor. Biochemistry 38, 3559–3569. 8 Zhong M, Orosz A & Wu C (1998) Direct sensing of heat and oxidation by Drosophila heat shock transcription factor. Mol Cell 2, 101–108. 9 Ahn SG & Thiele DJ (2003) Redox regulation of mammalian heat shock factor 1 is essential for Hsp gene activation and protection from stress. Genes Dev 17, 516–528. 10 Rabindran SK, Haroun RI, Clos J, Wisniewski J & Wu C (1993) Regulation of heat shock factor trimer formation: role of the conserved leucine zipper. Science 259, 230–234. 11 Zuo J, Baler R, Dahl G & Voellmy R (1994) Activation of the DNA-binding ability of human heat shock transcription factor 1 may involve the transition from an intramolecular to an intermolecular triple-stranded coiled-coil structure. Mol Cell Biol 14, 7557–7568. 12 Farkas T, Kutskova YA & Zimarino V (1998) Intramolecular repression of mouse heat shock factor 1. Mol Cell Biol 18, 906–918. 13 Newton EM, Knauf U, Green M & Kingston RE (1996) The regulatory domain of human heat shock factor 1 is sufficient to sense heat stress. Mol Cell Biol 16, 839–846. 14 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. 15 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. 16 Cunniff NF & Morgan WD (1993) Analysis of heat shock element recognition by saturation mutagenesis of the human HSP70.1 gene promoter. J Biol Chem 268, 8317–8324. 17 Fernandes M, Xiao H & Lis JT (1994) Fine structure analyses of the Drosophila and Saccharomyces heat shock factor-heat shock element interactions. Nucleic Acids Res 22, 167–173. 18 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. 19 Xiao H, Perisic O & Lis JT (1991) Cooperative binding of Drosophila heat shock factor to arrays of a conserved 5 bp unit. Cell 64, 585–593. 20 Hashikawa N, Mizukami Y, Imazu H & Sakurai H (2006) Mutated yeast heat shock transcription factor activates transcription independently of hyperphosph- orylation. J Biol Chem 281, 3936–3942. 21 Yamamoto N, Takemori Y, Sakurai M, Sugiyama K & Sakurai H (2009) Differential recognition of heat shock elements by members of the heat shock transcription factor family. FEBS J 276, 1962–1974. 22 Cicero MP, Hubl ST, Harrison CJ, Littlefield O, Hardy JA & Nelson HC (2001) The wing in yeast heat shock transcription factor (HSF) DNA-binding domain is required for full activity. Nucleic Acids Res 29, 1715– 1723. 23 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. 24 Lu M, Lee YJ, Park SM, Kang HS, Kang SW, Kim S & Park JS (2009) Aromatic-participant interactions are essential for disulfide-bond-based trimerization in human heat shock transcription factor 1. Biochemistry 48, 3795–3797. 25 Liu PC & Thiele DJ (1999) Modulation of human heat shock factor trimerization by the linker domain. J Biol Chem 274, 17219–17225. 26 Trott A & Morano KA (2003) The yeast response to heat shock. In Yeast Stress Responses (Hohmann S & Mager PWH eds), pp 71–119. Springer-Verlag, Heidel- berg, Germany. 27 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. 28 Yamamoto A, Mizukami Y & Sakurai H (2005) Identifi- cation 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. 29 Yamamoto A, Ueda J, Yamamoto N, Hashikawa N & Sakurai H (2007) Role of heat shock transcription factor in Saccharomyces cerevisiae oxidative stress response. Eukaryot Cell 6, 1373–1379. 30 Sorger PK (1990) Yeast heat shock factor contains separable transient and sustained response transcriptional activators. Cell 62, 793–805. 31 Chou S, Chatterjee S, Lee M & Struhl K (1999) Transcriptional activation in yeast cells lacking transcription factor IIA. Genetics 153 , 1573–1581. 32 Lee D & Lis JT (1998) Transcriptional activation inde- pendent of TFIIH kinase and the RNA polymerase II mediator in vivo. Nature 393, 389–392. 33 Apone LM, Virbasius CA, Holstege FC, Wang J, Young RA & Green MR (1998) Broad, but not univer- H. Sakurai and Y. Enoki HSF–HSE interaction FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 4147 sal, transcriptional requirement for yTAFII17, a histone H3-like TAFII present in TFIID and SAGA. Mol Cell 2, 653–661. 34 Moqtaderi Z, Keaveney M & Struhl K (1998) The histone H3-like TAF is broadly required for transcription in yeast. Mol Cell 2, 675–682. 35 Silar P, Butler G & Thiele DJ (1991) Heat shock transcription factor activates transcription of the yeast metallothionein gene. Mol Cell Biol 11, 1232– 1238. 36 Hashikawa N, Yamamoto N & Sakurai H (2007) Different 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. 37 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. 38 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. 39 Sistonen L, Sarge KD & Morimoto RI (1994) Human heat shock factors 1 and 2 are differentially activated and can synergistically induce hsp70 gene transcription. Mol Cell Biol 14, 2087–2099. 40 Nakai A, Tanabe M, Kawazoe Y, Inazawa J, Morimot- o 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. 41 Fujimoto M, Hayashida N, Katoh T, Oshima K, Shinkawa T, Prakasam R, Tan K, Inouye S, Takii R & Nakai A (2010) A novel mouse HSF3 has the potential to activate non-classical heat shock genes during heat shock. Mol Biol Cell 21, 106–116. 42 Trinklein ND, Murray JI, Hartman SJ, Botstein D & Myers RM (2004) The role of heat shock transcription factor 1 in the genome-wide regulation of the mammalian heat shock response. Mol Biol Cell 15, 1254–1261. 43 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. 44 Chang Y, Ostling P, Akerfelt M, Trouillet D, Rallu M, Gitton Y, El Fatimy R, Fardeau V, Le Crom S, Morange 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. 45 Takemori Y, Enoki Y, Yamamoto N, Fukai Y, Adachi K & Sakurai H (2009) Mutational analysis of human heat-shock transcription factor 1 reveals a regulatory role for oligomerization in DNA-binding specificity. Biochem J 424, 253–261. 46 Pan Y, Tsai CJ, Ma B & Nussinov R (2010) Mecha- nisms of transcription factor selectivity. Trends Genet 26, 75–83. 47 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. 48 Ostling P, Bjo ¨ rk JK, Roos-Mattjus P, Mezger V & Sistonen L (2007) Heat shock factor 2 (HSF2) contributes to inducible expression of hsp genes through interplay with HSF1. J Biol Chem 282, 7077– 7086. 49 Sandqvist A, Bjo ¨ rk JK, Akerfelt M, Chitikova Z, Grichine A, Vourc’h C, Jolly C, Salminen TA, Nymalm Y & Sistonen L (2009) Heterotrimerization of heat- shock factors 1 and 2 provides a transcriptional switch in response to distinct stimuli. Mol Biol Cell 20, 1340– 1347. 50 Taylor IC, Workman JL, Schuetz TJ & Kingston RE (1991) Facilitated binding of GAL4 and heat shock factor to nucleosomal templates: differential function of DNA-binding domains. Genes Dev 5, 1285–1298. 51 Li B, Carey M & Workman JL (2007) The role of chromatin during transcription. Cell 128, 707–719. 52 Zanton SJ & Pugh BF (2006) Full and partial genome-wide assembly and disassembly of the yeast transcription machinery in response to heat. Genes Dev 20, 2250–2265. 53 Schwabish MA & Struhl K (2006) Asf1 mediates histone eviction and deposition during elongation by RNA polymerase II. Mol Cell 22, 415–422. 54 Shivaswamy S & Iyer VR (2008) Stress-dependent dynamics of global chromatin remodeling in yeast: dual role for SWI ⁄ SNF in the heat shock stress response. Mol Cell Biol 28, 2221–2234. 55 Zhao J, Herrera-Diaz J & Gross DS (2005) Domain- wide displacement of histones by activated heat shock factor occurs independently of Swi ⁄ Snf and is not correlated with RNA polymerase II density. Mol Cell Biol 25, 8985–8999. 56 Erkina TY, Tschetter PA & Erkine AM (2008) Differ- ent requirements of the SWI ⁄ SNF complex for robust nucleosome displacement at promoters of heat shock factor and Msn2- and Msn4-regulated heat shock genes. Mol Cell Biol 28, 1207–1217. 57 Erkina TY, Zou Y, Freeling S, Vorobyev VI & Erkine AM (2009) Functional interplay between chromatin remodeling complexes RSC, SWI ⁄ SNF and ISWI in regulation of yeast heat shock genes. Nucleic Acids Res 38, 1441–1449. HSF–HSE interaction H. Sakurai and Y. Enoki 4148 FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 58 Kremer SB & Gross DS (2009) The SAGA and Rpd3 chromatin modification complexes dynamically regulate heat shock gene structure and expression. J Biol Chem 284, 32914–32931. 59 Venturi CB, Erkine AM & Gross DS (2000) Cell cycle-dependent binding of yeast heat shock factor to nucleosomes. Mol Cell Biol 20, 6435–6448. 60 Petesch SJ & Lis JT (2008) Rapid, transcription- independent loss of nucleosomes over a large chromatin domain at Hsp70 loci. Cell 134, 74–84. 61 Jensen MM, Christensen MS, Bonven B & Jensen TH (2008) Requirements for chromatin reassembly during transcriptional downregulation of a heat shock gene in Saccharomyces cerevisiae. FEBS J 275, 2956– 2964. 62 Core LJ & Lis JT (2008) Transcription regulation through promoter-proximal pausing of RNA polymerase II. Science 319, 1791–1792. 63 Price DH (2008) Poised polymerases: on your mark get set go! Mol Cell 30, 7–10. 64 Gilmour DS (2009) Promoter proximal pausing on genes in metazoans. Chromosoma 118, 1–10. 65 Bevilacqua A, Fiorenza MT & Mangia F (2000) A developmentally regulated GAGA box-binding factor and Sp1 are required for transcription of the hsp70.1 gene at the onset of mouse zygotic genome activation. Development 127, 1541–1551. 66 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. 67 Xing H, Vanderford NL & Sarge KD (2008) The TBP- PP2A mitotic complex bookmarks genes by preventing condensin action. Nat Cell Biol 10, 1318–1323. 68 Sarge KD & Park-Sarge OK (2009) Mitotic bookmarking of formerly active genes: keeping epigenetic memories from fading. Cell Cycle 8, 818–823. 69 Inouye S, Fujimoto M, Nakamura T, Takaki E, Hayashida N, Hai T & Nakai A (2007) Heat shock transcription factor 1 opens chromatin structure of interleukin-6 promoter to facilitate binding of an activator or a repressor. J Biol Chem 282, 33210– 33217. 70 Corey LL, Weirich CS, Benjamin IJ & Kingston RE (2003) Localized recruitment of a chromatin-remodeling activity by an activator in vivo drives transcriptional elongation. Genes Dev 17, 1392–1401. 71 Xu D, Zalmas LP & La Thangue NB (2008) A transcription cofactor required for the heat-shock response. EMBO Rep 9, 662–669. 72 Thomson S, Hollis A, Hazzalin CA & Mahadevan LC (2004) Distinct stimulus-specific histone modifications at hsp70 chromatin targeted by the transcription factor heat shock factor-1. Mol Cell 15, 585–594. 73 Go ´ mez AV, Galleguillos D, Maass JC, Battaglioli E, Kukuljan M & Andre ´ s ME (2008) CoREST represses the heat shock response mediated by HSF1. Mol Cell 31, 222–231. 74 Espinoza CA, Allen TA, Hieb AR, Kugel JF & Goodrich JA (2004) B2 RNA binds directly to RNA polymerase II to repress transcript synthesis. Nat Struct Mol Biol 11, 822–829. 75 Mariner PD, Walters RD, Espinoza CA, Drullinger LF, Wagner SD, Kugel JF & Goodrich JA (2008) Human Alu RNA is a modular transacting repressor of mRNA transcription during heat shock. Mol Cell 29, 499–509. 76 Shamovsky I, Ivannikov M, Kandel ES, Gershon D & Nudler E (2006) RNA-mediated response to heat shock in mammalian cells. Nature 440, 556–560. 77 Fritah S, Col E, Boyault C, Govin J, Sadoul K, Chiocca S, Christians E, Khochbin S, Jolly C & Vourc’h C (2009) Heat-shock factor 1 controls genome- wide acetylation in heat-shocked cells. Mol Biol Cell 20, 4976–4984. 78 Khaleque MA, Bharti A, Gong J, Gray PJ, Sachdev V, Ciocca DR, Stati A, Fanelli M & Calderwood SK (2008) Heat shock factor 1 represses estrogen-dependent transcription through association with MTA1. Oncogene 27, 1886–1893. H. Sakurai and Y. Enoki HSF–HSE interaction FEBS Journal 277 (2010) 4140–4149 ª 2010 The Authors Journal compilation ª 2010 FEBS 4149 . MINIREVIEW Novel aspects of heat shock factors: DNA recognition, chromatin modulation and gene expression Hiroshi Sakurai and Yasuaki Enoki Department of Clinical. & Sakurai H (2005) Identifi- cation of a novel class of target genes and a novel type of binding sequence of heat shock transcription factor in Saccharomyces