Heat shock factor
The heat shock transcription factor HSF activates expression of its target genes in response to elevated temperatures and chemical or physiological stress. A key step in the activation process involves the formation of HSF homotrimers, leading to high-affinity DNA binding. The mechanism by which HSF trimerization and DNA binding are regulated by stress signals has remained elusive. Trimerization and DNA binding of purified Drosophila HSF can be directly and reversibly induced in vitro by heat shock temperatures in the physiological range and by an oxidant, hydrogen peroxide. Other inducers of the heat shock response, including salicylate, dinitrophenol, ethanol, and arsenite, have no effect on HSF trimerization in vitro, indicating that these inducers act by indirect
mechanisms. How might the direct action of heat of hydrogen peroxide enhance the formation of HSF trimers? Trimerization is mediated by the N-terminal heptad repeats whose propensity to associate as a three-stranded alpha-helical coiled-coil is repressed under non-shock conditions by the C-terminal region of HSF. The occurrence of a conserved hydrophobic heptad repeat in this region has led to the suggestion that repression could involve an intramolecular coiled-coil,formed between the N-terminal and C-terminal heptad repeats. Heat or oxidation is likely to induce a conformational change that unmasks the hydrophobic heptad repeats of the trimerization domain, thereby allowing assembly of the trimeric coiled-coil. The active sites of sensitivity to heat or oxidation on HSF are as yet undefined. These findings also imply that other heat shock inducers (such as heat shock proteins known to associate with HSF) exert secondary effects indirectly, perhaps by generating an intermediate redox signal, or by the action of cellular factors (Zhong, 1998).
The minimal DNA sequence required for the formation of a stable complex with Drosophila heat shock factor (HSF) in vitro is an inverted repeat of a 5 bp recognition unit, -GAA-. Surprisingly, both permutations of this 5 bp unit, head-to-head and tail-to-tail, bind to HSF with similar affinity and with
striking 2-fold symmetry. HSF also binds to longer arrays of inverted 5 bp units, and the size of the HSF footprint increases with the addition of each 5 bp unit to these arrays. However, the electrophoretic mobility of the HSF-DNA complexes decreases most distinctly with the addition of every three 5 bp units. Cross-linking of purified HSF in the absence of DNA generates complexes with the sizes expected of HSF trimers. It is proposed that trimers of HSF bind to DNA and that the number of HSF subunits in direct contact with DNA is determined by the number of correctly positioned 5 bp recognition units (Perisic, 1989).
Activation of HSF occurs at two independent levels, DNA binding and the acquisition of transcriptional competence. The binding of HSF to DNA is accomplished by a stress-induced oligomeric switch of HSF protein. Calculation of the native molecular mass indicates that the two forms of Drosophila HSF are best described as a monomer and trimer, respectively, of the 77-kDa HSF polypeptide. The monomeric and trimeric states of HSF were verified by chemical cross-linking experiments. The finding of a monomeric composition for the latent form of HSF is incompatible with speculative models that suggest that molecular chaperones such as hsp70 feedback to inhibit trimerization of HSF by forming a stable heteromeric complex. Both HSF monomers and HSF trimers exhibit unusually high frictional ratios, indicating that they have asymmetric shapes. The degree of asymmetry is significantly greater for the HSF trimer, suggesting that the monomer undergoes a conformational change to a more extended structure upon trimerization. These findings are consistent with a model for the inert HSF protein that is based on a monomer constrained by intramolecular coiled-coil interactions between amino- and carboxy-terminal domains (Westwood, 1993).
The intracellular level of free heat shock proteins, in particular the 70-kDa stress protein family, has been suggested to be the basis of an autoregulatory mechanism by which the cell measures the level of thermal stress and regulates the synthesis of heat shock proteins. It has been proposed that the DNA-binding and oligomeric state of the heat shock transcription factor (HSF) is a principal step in the induction pathway that responds to the level of 70-kDa stress protein. The association between HSF and 70-kDa stress protein has been investigated by means of a coimmunoprecipitation assay. 70-kDa stress proteins associate to similar extents with both latent and active forms of HSF, although unlike other 70-kDa stress protein substrates, the association with HSF is not significantly disrupted in the presence of ATP. Gel mobility shift assays indicate that active HSF trimers purified from a bacterial expression system can not be substantially deactivated in vitro with purified 70-kDa stress protein and ATP. In addition, elevated concentrations of hsp70 alone can not significantly inhibit induction of the DNA-binding activity of endogenous HSF in cultured rat cells, and the induction is also not inhibited in cultured rat cells or Drosophila cells containing elevated levels of all members of the heat shock protein family. However, the deactivation of HSF to the non-DNA-binding state after prolonged heat stress or during recovery could be accelerated by increased levels of heat shock proteins. Hence, the level of heat shock proteins may affect the rate of disassembly of HSF trimers, but another mechanism, as yet undefined, appears to control the onset of the oligomeric transitions (Rabindran, 1994).
Heat shock factor (HSF), the transcriptional activator of eukaryotic heat shock genes, is induced to bind DNA by a monomer to trimer transition involving leucine zipper interactions. Although this mode of regulation is shared among many eukaryotic species, there is variation in the temperature at which HSF binding activity is induced. The response of a human HSF expressed in Drosophila cells and Drosophila HSF expressed in human cells was investigated. The temperature that induces DNA binding and trimerization of human HSF in Drosophila is decreased by approximately 10 degrees C to the induction temperature for the host cell, whereas Drosophila HSF expressed in human cells is constitutively active. The results indicate that the activity of HSF in vivo is not a simple function of the absolute environmental temperature (Clos, 1993).
The heat shock transcription factor (HSF) mediates the induction of heat shock gene expression. The
activation of HSF involves heat shock-induced trimerization, binding to its cognate DNA sites, and the
acquisition of transcriptional competence. In this study, the oligomeric properties of Drosophila HSF were
analyzed by equilibrium analytical ultracentrifugation and gel filtration chromatography. Previous findings have
shown that trimerization of purified Drosophila HSF is directly sensitive to heat and oxidation. Low pH, in the physiological range, also directly induces HSF trimerization and DNA binding
in vitro. Furthermore, the induction of HSF trimerization by low pH is synergistic with the actions of heat
and oxidation. Since heat or chemical stress leads to a moderate decrease of intracellular pH, it is suggested that
intracellular acidification may contribute to activating the heat shock response in vivo (Zhong, 1999).
A number of activators are known to increase transcription by RNA polymerase (pol) II through protein acetylation. While the physiological substrates for those acetylases are poorly defined, possible targets include general transcription factors, activator proteins and histones. Using a cell-free system to reconstitute chromatin with increased histone acetylation levels, a direct test was performed for a causal role of histone acetylation in transcription by RNA pol II. Chromatin, containing either control or acetylated histones, was reconstituted to comparable nucleosome densities and characterized by electron microscopy after psoralen cross-linking, as well as by in vitro transcription. Chromatin was reconstituted using histones from either TSA-treated CV-1 cells, which accumulates hyperacetylated histone isoforms, or from untreated cells, in which the histones are primarily non-acetylated onto a 7.75 kb plasmid containing an hsp26 minigene. This process involves the prior depletion of the endogenous histones present in the chromatin assembly extract such that chromatin is assembled quantitatively from the input, exogenous histones. The chromatin assembly reaction generates complex chromatin containing many non-histone proteins and enzymatic activities. While H1-containing control chromatin severely represses transcription of a model hsp26 gene, highly acetylated chromatin is significantly less repressive. Acetylation of histones, and particularly of histone H4, affects transcription at the level of initiation (Nightingale, 1998).
The regulatory elements of the hsp26 promoter are well-known from in vivo and in vitro studies. A proximal regulatory element includes the TATA box, proximal heat shock element (HSE) and an adjacent GAGA element, while a distal regulatory site corresponds to the distal HSE and GAGA binding sites. Monitoring the ability of the transcription machinery to associate with the promoter in chromatin, it was found that Heat shock factor, a crucial regulator of heat shock gene transcription, profits most from histone acetylation. Templates with mutated hsp26 promoters were assembled into control and acetylated chromatin and analysed for their transcription potential. A template bearing only the TATA box supports a very low level of transcription, even in the absence of chromatin; there is no discernible transcription from chromatinized templates even after prolonged exposure. Addition of the proximal HSEs however, results in significantly increased transcription from both mock assembled and chromatinized templates, confirming the important role of the activator HSF. This minimal promoter, containing only proximal HSEs and the TATA box, clearly shows increased transcription from the acetylated template. The addition of GAGA elements to the promoter enhances transcription significantly, but to a similar degree on control and acetylated chromatin templates. Interestingly, the GAGA elements do not increase transcription in the mock assembled control, confirming that GAGA factor is involved in overcoming chromatin-mediated transcriptional repression, but by a mechanism that does not profit from histone acetylation. These results suggest that the HSE and TATA box are the significant sequence elements for the increased transcription observed in acetylated chromatin. Thus histone acetylation can modulate activator access to their target sites in chromatin, and provide a causal link between histone acetylation and enhanced transcription initiation of RNA pol II in chromatin (Nightingale, 1998).
The generation of an accessible heat shock promoter in chromatin in vitro requires the concerted action of the GAGA transcription factor and NURF, an ATP-dependent nucleosome remodeling factor. NURF is composed of four subunits and is biochemically distinct from the SWI2/SNF2 multiprotein complex, a transcriptional activator that also appears to alter nucleosome structure. The 140 kDa subunit of NURF can be identified as ISWI, previously of unknown function but highly related to SWI2/SNF2, but only in the ATPase domain. The ISWI protein is localized to the cell nucleus and is expressed throughout Drosophila development at levels as high as 100,000 molecules/cell. The convergence of biochemical and genetic studies on ISWI and SWI2/SNF2 underscores these ATPases and their close relatives as key components of independent systems for chromatin remodeling (Tsukiyama, 1995a and b).
Genetic control elements are usually situated in local regions of chromatin that are hypersensitive to structural probes such as DNase I. Binding of the GAGA transcription factor on existing nucleosomes leads to nucleosome disruption, DNase I hypersensitivity at the TATA box and heat-shock elements, and rearrangement of adjacent nucleosomes. ATP hydrolysis facilitates this process, suggesting that an energy-dependent pathway is involved in chromatin remodeling (Tsukiyama, 1994).
Drosophila heat shock factor (HSF) binds to specific sequence elements of heat shock genes and can activate their transcription 200-fold. Though HSF has an acidic activation domain, the mechanistic details of heat shock gene activation remain undefined. HSF interacts directly with the general transcription factor TBP (TATA-box binding protein), and these two factors bind cooperatively to heat shock promoters. A third factor that binds heat shock promoters, GAGA factor, also interacts with HSF and further stabilizes HSF binding to heat shock elements (HSEs). The interaction of HSF and TBP is explored in some detail here and is shown to be mediated by residues in both the amino- and carboxyl-terminal portions of HSF. This HSF/TBP interaction can be disrupted specifically by competition with the potent acidic transcriptional activator VP16. The acidic domain of the largest subunit of Drosophila RNA polymerase II (Pol II) associates with TBP in vitro and is specifically displaced from TBP upon addition of HSF. The region of TBP that mediates both HSF and Pol II acidic domain binding maps to the conserved carboxyl-terminal repeats and depends on at least one of the TBP residues known to be contacted by VP16 and to be critical for transcription activation. It is thought that HSF triggers hsp70 transcription by freeing the hsp70 promoter-paused Pol II from the constraints on elongation caused by the affinity of Pol II for general transcription factors (Mason, 1997).
Posttranslational modifications of the N-terminal tails of the core
histones within the nucleosome particle are thought to act as signals
from the chromatin to the cell for various processes. The experiments
presented here show that the acetylation of histones H3 and H4 in
polytene chromosomes does not change during heat shock. In contrast,
the global level of phosphorylated H3 decreases dramatically during a
heat shock, with an observed increase in H3 phosphorylation at the heat
shock loci. Additional experiments confirm that this change in
phosphorylated H3 distribution is dependent on functional heat shock
transcription factor activity. These experiments suggest that H3
phosphorylation has an important role in the induction of transcription
during the heat shock response (Nowak, 2000).
The acetylation of the N-terminal tails is the best-studied
modification of the core histones. Several transcription factors, such
as GCN5, and the TAFII250 subunit of TFIID, as well as subunits of the RNA polymerase complex show intrinsic histone
acetyltransferase (HAT) activity, which suggests a potential role for
histone acetylation in either the activation or maintenance of
transcription. The acetylation of the N-terminal tail domains of core
histones H3 and H4 at various lysine residues is essential for the
normal implementation of various cellular processes, such as
promoter-transcription factor association, gene transcription, and
dosage compensation (Nowak, 2000 and references therein).
Phosphorylation of serine 10 of the N-terminal arm of histone H3 has
been shown to be essential for proper mitotic chromosomal condensation and segregation. In addition, recent studies have outlined the possibility that
histone H3 phosphorylation may have a role in the regulation of
transcription. Ser 10 H3 phosphorylation is found to rapidly increase
in quiescent cells during mitogenic stimulation,
as well as during immediate-early gene induction via the epidermal
growth factor (EGF)-signaling pathway. In
addition, recent experiments performed in vitro have suggested that
EGF-stimulated H3 phosphorylation may act as a signal for histone
acetyltransferase binding and subsequent acetylation of a particular
locus during transcription initiation (Nowak, 2000 and references therein).
Acetylation of core histones H3 and H4 at lysines 14 and 8, respectively, has been linked to gene transcription. In addition, deacetylation of core histones is thought to have a role in silencing specific loci. Because of the
near-total repression of cellular gene products during a heat shock, it might be expected that the distribution of acetylated H3 and H4 would
radically change during thermal stress in a manner reflective of the
transcriptional profile of the cell. Because acetylation of H3 at Lys 14 of the
N-terminal arm has been described as essential for transcription, the
distribution of acetylated H3 was examined by staining polytene chromosomes with an
antibody specific for Lys 14 acetylated histone H3.
Lys 14 acetylated H3 staining is observed at the puffs, which are
active sites of transcription in polytene chromosomes, and distributed throughout the
chromosomes in discrete bands before heat shock. One locus,
subdivision 62A, which becomes puffed during larval development in
response to ecdysone, is intensely labeled with the
Lys 14 acetylated H3 antibody. In addition, other
chromosomal subdivisions such as 89B display Lys 14 acetylated H3
staining but are not puffed before heat shock. The
chromosomal subdivision 93D, which is known to become puffed during
heat shock, is Lys 14 acetylated but not puffed before heat shock. Examination of polytene chromosomes from larvae that were
subjected to a 20-min heat shock shows that the 87A and 87C heat shock
puffs, which contain the hsp70gene cluster, are stained by the anti-Lys 14 acetylated antibody, although
the staining at these puffs appears to be less intense and rather
diffuse. This might not represent a reduction in the
level of acetylation, but rather a decrease in signal intensity due to
the large puffing at the heat shock loci. After heat shock, the overall
number of discrete stained bands does not appear to change significantly
and regions that were stained before heat shock, such as 89B, remain
acetylated. Loci with acetylated H3 staining that were
puffed before heat shock, such as 62A, are no longer puffed after heat shock but remain acetylated. The observation that the
heat shock genes are acetylated before heat shock, at a time when they
are not transcribed, and non-heat shock genes, which are not
transcribed during heat shock, are acetylated during heat shock,
suggests that the presence of Lys 14-acetylated H3 does not necessarily
denote an actively transcribed locus (Nowak, 2000).
Examination of H3 acetylation during EGF stimulation raises the
issue that antibodies against Lys 14 acetylated H3 may show decreased
recognition of their epitope when other modifications, such as
phosphorylation, coexist on the same histone tail. This problem can be overcome by using antibodies against histone H3 acetylated at lysines 9 and 14. To ensure that these results were not caused by this potential artifact, the distribution of hyperacetylated H3 was examined using antibodies against H3 acetylated at lysines 9 and 14 on the N-terminal tail before and after heat shock. The results suggest that the distribution of diacetylated H3 is similar to
the distribution of Lys 14 acetylated H3 before and after heat shock.
Diacetylated H3 staining appears to be more widespread than
monoacetylated staining, which is probably caused by the antibody's
recognition of acetylation of H3 at lysine 9. The intensity of staining
of the Lys 9,14-acetylated H3 antibody at several of the heat shock
puffs examined appears to be similar to that observed with the Lys
14-acetylated H3 antibody. These results suggest that the diffuse staining at the heat shock puffs is not an artifact attributed to the masking of the acetylated Lys 14 epitope by Ser 10 phosphorylation (Nowak, 2000).
H4 acetylation was also examined using antibodies specific for Lys
8-acetylated histone H4 to stain polytene chromosomes isolated from third instar larvae. The distribution of Lys 8 acetylated histone H4 is similar to that of
acetylated H3, with H4 acetylation observed in discrete bands in
nonpuffed regions, such as subdivision 89B, and at ecdysone-induced
puffed regions, such as 62A, before heat shock. Chromosomal
subdivisions 87A and 87C, which contain the hsp70 heat shock
genes, are acetylated before and after
heat shock. Similar to acetylated H3, heat shock does not
significantly affect the observed distribution of Lys 8 acetylated H4
in polytene chromosomes. Taken together, the above results
suggest that the acetylation state of H3 and H4 does not change
substantially during heat shock and that a gene locus can be acetylated
when it is not actively transcribed (Nowak, 2000).
The absence of a drastic change in H3 acetylation during heat shock
is rather surprising, given current models that indicate that H3
acetylation is a crucial step in transcription initiation. This would lead to the expectation that the heat shock loci would not be acetylated before heat shock and should become intensely acetylated during thermal stress. To determine if
other histone modifications occur during the heat shock response, whether changes in histone H3 phosphorylation occur after
temperature elevation was examined. Stimulation of quiescent cells with EGF leads to
rapid and transient phosphorylation of histone H3 at Ser 10 of the
N-terminal arm in vivo. This EGF-mediated
phosphorylation of H3 is targeted to a small subpopulation of total
histone H3 that is acetylated at the Lys 14 position. In addition, in vitro studies have shown that phosphorylated H3
may serve as an affinity-increasing substrate for HAT activity in H3
acetylation, which raises the possibility that phosphorylation may be
tied to transcription. If histone phosphorylation were implicated in transcription, then the distribution of
phosphorylated H3 might change in response to heat shock and would most
likely be localized primarily to the heat shock puffs while
disappearing from other loci after heat shock. Because histone H3
phosphorylation is a robust marker for mitotic cells, analysis of the distribution of phosphorylated H3 in polytene
chromosomes, rather than isolation of phosphorylated H3 from whole cell
extracts, allows for the examining of phosphorylation of H3 in
a nonmitotic environment. To examine whether the heat shock-induced puffs contain N-terminal phosphorylated H3 molecules, polytene chromosomes were stained with antibodies specific for Ser 10 phosphorylated histone H3. Before heat shock,
phosphorylated H3 staining is found in discrete bands throughout the
chromosomes, with the most intense staining observed in the naturally
occurring ecdysone-induced developmental puffs. After a 20-min heat shock at 37°C, phosphorylated H3 staining is not distributed throughout the
chromosomes but is instead concentrated at a few specific sites. The most prominent of these regions corresponds to chromosomal
divisions 63BC, 67B, and 87AC. These regions contain the hsp83
gene, the hsp22, hsp23, hsp26, and
hsp27 gene cluster, and hsp70 gene clusters,
respectively. These regions become reproducibly puffed during the heat shock response. Although in some chromosomes examined there are
several non-heat shock loci that remain slightly phosphorylated
during heat shock, the intensity of staining at these regions is much
lower than the staining observed at the heat shock loci (Nowak, 2000).
The regions of the chromosome where the heat shock genes are located do
not contain histone H3 phosphorylated at Ser 10 before heat shock. After temperature elevation, the only puffs that possess
phosphorylated histone H3 are the heat shock puffs. The appearance of
phosphorylated histone H3 in the heat shock puffs, accompanied by the
nearly complete reduction of staining at all other loci during heat
shock, leads to the conclusion that the presence of the Ser 10 phosphorylated isoform of histone H3 might be required for the
transcriptional activation of the heat shock genes (Nowak, 2000).
Induction of the heat shock genes and cessation of normal gene
expression is rapid and reproducible in response to heat shock. Transcription run-on assays reveal that after only 1 min at 37°C, the levels of many normal cellular gene transcripts have greatly diminished, with the heat shock gene
transcripts dominating the population of total mRNA in the cell. Following a heat
shock, the normal pattern of gene expression within the cell is
restored gradually over time. Therefore an examination was made of
the change in phosphorylated histone H3 staining over time during and
after heat shock, to determine whether or not the appearance of
phosphorylated H3 closely parallels the induction of transcription
of the heat shock genes and whether or not the non-heat shocked H3
distribution might be restored following recovery from heat shock. After only 1 min at 37°C, there is a noticeable change in the distribution of Ser 10 phosphorylated H3. The
level of global H3 phosphorylation decreases, with several regions
remaining intensely phosphorylated. Within 5 min of incubation at 37°C, many of the less intense regions of staining have disappeared. After 10 min at 37°C, the only remaining intense regions of
staining are those at the heat shock puffs. When larvae were
allowed to recover at room temperature from a 20-min heat shock at
37°C, H3 phosphorylation reappears in several non-heat shock loci
after 10 min of recovery. After 30 min of recovery from heat
shock, the number and distribution of loci that contained
phosphorylated H3 appears to be virtually indistinguishable from
normal (i.e., non-heat shocked) chromosomes. This restoration
of the normal (non-heat shocked) H3 phosphorylation pattern closely
mimics previously described restoration of normal gene expression in
cells experiencing thermal stress (Nowak, 2000).
During heat shock, the heat shock transcription factor (HSF) rapidly
trimerizes in solution, localizes to the heat shock loci, binds to heat
shock response promoter elements (HSEs), and induces the expression of
the heat shock gene products. The appearance of phosphorylated H3 at the heat shock loci could therefore be due to HSF recruitment of a specific histone kinase on
binding to the HSEs of the heat shock genes. To test this hypothesis,
the staining pattern of phosphorylated H3 was examined in polytene
chromosomes isolated from hsf4-mutant larvae, which
lack functional HSF at restrictive temperatures and do not respond to
thermal stress. Before heat shock, the
distribution of phosphorylated H3 in hsf4-mutant
chromosomes is similar to wild-type chromosomes, with staining
observed in discrete bands and at the developmental puffs. In contrast to wild-type chromosomes, histone H3 at the heat shock loci does not become phosphorylated in
hsf4-mutant chromosomes during heat shock, which
suggests that phosphorylation of histone H3 at the heat shock loci
depends on functional HSF activity. In addition, no H3 phosphorylation was detected in the rest of the genome during heat
shock in hsf4 mutants, suggesting that repression of
normal transcription and loss of H3 phosphorylation at non-heat shock
loci does not require the presence of an active HSF protein (Nowak, 2000).
To determine if the loss of the HSF transcription factor could also
alter the distribution of acetylated H3 and H4 during heat shock, acetylation of each of these histones was examined in
hsf4-mutant polytene chromosomes. The distribution
of Lys 14 acetylated histone H3 before and after heat shock in
hsf4 mutants was indistinguishable from the
wild-type distribution, with staining observed at both the
developmental puffs and nonpuffed regions. H3 acetylation
was observed at the 87A and 87C chromosomal subdivisions, which
normally are puffed during heat shock but these regions do not become
puffed in hsf4-mutant chromosomes.
Examination of acetylated H3 using antibodies for Lys 9- and Lys
14-acetylated H3 shows a pattern similar to that observed for the Lys
14 acetylated H3 antibody. In addition, H4 acetylation does
not change after heat shock in hsf4 mutants. Because the heat shock genes are not induced in
hsf4 mutants during thermal stress and because
hsf4-mutant chromosomes are acetylated, but not
phosphorylated after heat shock, it is concluded that H3 phosphorylation,
and not acetylation, depends on the presence of a functional heat shock
transcription factor (Nowak, 2000).
How might acetylation and phosphorylation of histones H3 and H4 work
together to promote transcription of a particular gene? The data
suggest that acetylated histones might define a particular locus that
is primed for possible phosphorylation and subsequent transcription. This acetylated locus would attract transcription factors that interact with the
acetylated residues on histones H3 and H4, known to be essential for
proper association of several transcription factors with their
promoters. Once bound to this locus, the transcription factor would then
recruit a particular histone, which phosphorylates Ser
10 of the N-terminal arm of histone H3. The most logical site
of phosphorylation would be an H3 molecule with a Lys 14 acetylated
N-terminal arm, a species that has been shown to exist in vivo. The presence of this dimodified H3 would define that locus as 'active' for transcription (Nowak, 2000).
There are several kinases known to localize to specific loci on
polytene chromosomes that phosphorylate H3 in vitro, such as JIL-1 on
the X chromosome and P-TEFb kinase at the heat shock
loci (Lis, 2000). This raises the possibility that the
specificity of a kinase for activation of a particular gene through H3
phosphorylation might be regulated by the specific transcription
factors that control expression of this gene. It has yet to be determined
whether phosphorylation of H3 is required for assembly of the RNA
polymerase II complex or if phosphorylation is a by-product of complex
formation and polymerase procession during transcription. If
phosphorylation of H3 were indeed the critical step in activating gene
transcription, then a reasonable hypothesis is that deactivation of a
particular gene would be dependent on either regulated or unregulated
phosphatase activity to remove the activating phosphate group from the
N-terminal tails of H3. The disappearance of phosphorylated H3 at
nontranscribing loci and appearance of phosphorylated H3 at actively
transcribing loci during heat shock suggests that a functional
transcription complex might actively maintain the phosphorylated state
of histone H3, which would be subject to ready dephosphorylation by
either passive or regulated phosphatase activity in a nontranscribing state (Nowak, 2000).
P-TEFb kinase recruitment to heat shock loci during the heat shock response and functions to stimulate promoter-paused RNA polymerase II (Pol II) to enter into productive elongation. P-TEFb is located at >200 distinct sites on Drosophila polytene chromosomes.
Upon heat shock, P-TEFb, like the regulatory factor heat shock factor (HSF), is rapidly recruited to heat shock loci, and this recruitment is blocked in an
HSF mutant. Yet, HSF binding to DNA is not sufficient to recruit P-TEFb in vivo, and HSF and P-TEFb immunostainings within a heat
shock locus are not coincident. Insight to the function of P-TEFb is offered by experiments showing that the direct recruitment of a
Gal4-binding domain P-TEFb hybrid to an hsp70 promoter in Drosophila cells is sufficient to activate transcription in the absence of heat shock. Analyses of point
mutants show this P-TEFb stimulation is dependent on Cdk9 kinase activity and on Cdk9's interaction with cyclin T. These results, coupled with the frequent
colocalization of P-TEFb and the hypophosphorylated form of Pol II found at promoter-pause sites, support a model in which P-TEFb acts to
stimulate promoter-paused Pol II to enter into productive elongation (Lis, 2000).
P-TEFb is required to produce full-length transcripts from a variety of cellular DNA templates in an in vitro transcription system that
accurately recapitulates the normal DRB-sensitive transcription seen in cells (Marshall, 1995). These results suggest that
P-TEFb may have a role in transcription of many cellular genes. If so, this kinase may localize to chromosomal loci that possess genes that are the target of its
activity. The chromosomal distribution of P-TEFb was examined by staining salivary gland polytene chromosomes with a highly specific antibody to the cyclin T
regulatory subunit. This cyclin T subunit binds tightly to Cdk9 and is a critical component of the P-TEFb activity (Peng, 1998a). Moreover,
immunodepletion experiments show that the vast majority of Cdk9 is associated with a cyclin T subunit (Peng, 1998a), and probing of phosphocellulose
fractions from Drosophila Kc cell nuclear extracts indicates that cyclin T is present only where P-TEFb activity is found. Therefore,
the cyclin T antibody provides a good means of tracking the P-TEFb complex (Lis, 2000).
Heat shock causes a rapid and dramatic activation of transcription of heat shock
genes and a concomitant reduction in transcription of many normally expressed
genes. Immunofluorescence analysis of polytene chromosomes reveals that Pol II relocates to heat shock loci after a brief heat shock. P-TEFb distribution also changes dramatically following heat shock. In uninduced larvae, P-TEFb is undetectable
at major heat shock loci 87A and 87C, which contain the five hsp70 genes, or at 59B, which, in this strain, contains an
hsp70-lacZ transgene. After a 20-min heat shock, these and all the other major heat shock loci at 63B, 67B, 93D, and 95D are the
prominent sites of labeling. Loci that had high levels of P-TEFb before heat shock now have a reduced level. Therefore, P-TEFb redistributes to heat
shock loci following heat shock (Lis, 2000).
The Pol II level on the 5' end of the hsp70 gene begins to be elevated in as little as 70 sec following a very rapid heat shock induction (mixing cells with warm
medium), and Pol II is detected beyond the pause region and in the middle of the gene in as little as 2 min. This rapid transcriptional
activation leads to a very high density of hyperphosphorylated Pol II on these genes. Could P-TEFb be
playing a role in the transition of Pol II to its hyperphosphorylated, elongationally competent mode? If so, then one might expect P-TEFb to be recruited as rapidly as
Pol II to these newly activated heat shock sites (Lis, 2000).
The kinetics of localization to heat shock loci at 87A and 87C and to 59B, which in this strain contains an Hsp70-lacZ transgene, were examined. No P-TEFb is detected at the native or the transgenic sites before heat shock. However, within 2 min, staining is apparent at 87A and 87C, each of
which contain multiple copies of hsp70. Some staining is also detectable at the transgenic copy of Hsp70-lacZ. By 5 min of heat shock, staining at all heat shock loci
is strong and this high level persists and may even increase in the 10- and 15-min time points. The level remains high during heat shock measured out to 60 min. A
shift back to normal fly culture temperature (e.g., a 60-min recovery) reduces heat shock gene transcription and the normal pattern transcription is largely
re-established (Lis, 2000).
The recruitment of P-TEFb to heat shock loci is completely dependent on HSF. A Drosophila temperature-sensitive mutant HSF strain, hsf4, shows a much
reduced induction of heat shock gene transcription and chromosome puffing. In this strain, heat shock fails to concentrate P-TEFb at
heat shock loci. Additionally, heat shock does not lead to a dramatic loss of P-TEFb at the normally active chromosomal sites in the HSF mutant strain as
exemplified at 88D (Lis, 2000).
Heat shock rapidly stimulates the trimerization and binding of HSF to the heat shock elements (HSEs) located upstream of every heat shock gene. HSF acquires
strong DNA-binding activity and localizes to heat shock loci on polytene chromosomes within 2 min following heat shock. Therefore, the rapid induction of HSF binding is similar to the rapid recruitment of P-TEFb seen here. Could HSF itself be sufficient to recruit
P-TEFb through a stable interaction? This hypothesis was tested in vivo using a transgenic line containing a polymer of native HSF-binding sites that are unlinked to the
rest of the hsp70 promoter. Following heat
shock, HSF is known to localize to sites on polytene chromosomes containing this polymer. This anti-HSF staining is more than an order of magnitude
stronger than that seen at the regulatory region of a single hsp70 gene, and can be compared with the 87A and 87C loci that
contain two and three copies of native hsp70, respectively. The 87C signal is considerably stronger than 87A (Lis, 2000).
If HSF is sufficient to recruit P-TEFb to heat shock loci in vivo, then one would expect to see high levels of P-TEFb at the polymer site. There
is detectable P-TEFb at the polymer site, but the level is less than at the native heat shock loci 87A and 87C. Moreover, the ratio of P-TEFb to HSF staining is much
higher at heat shock genes than at the polymer site. These results indicate that HSF does not on its own recruit P-TEFb, and other features of the heat shock
promoters are required to provide P-TEFb's strong recruitment to heat shock genes (Lis, 2000).
P-TEFb appears to resolve from HSF at the 87A locus. In most extended chromosomes examined, it is observed that the P-TEFb label
separates into a doublet with HSF overlapping and falling between the peaks of the P-TEFb doublet. This can be interpreted in terms of the known arrangement of
hsp70 genes at 87A. The hsp70 genes are divergently transcribed and the regulatory DNA containing the binding sites for
HSEs resides in this region between the genes. HSF binds these regulatory regions as was seen from a band of fluorescence in the middle of the puff. In contrast,
the centers of P-TEFb staining appear to reside downstream of the HSEs on both copies of the hsp70 gene. The partial separation of P-TEFb and HSF is also
consistent with the idea that P-TEFb does not derive its stable association with heat shock genes solely through interaction with HSF (Lis, 2000).
A biochemical assay of the interaction of HSF and P-TEFb adds further support to the conclusion that these proteins do not interact strongly. Plasmids that express HSF, Cdk9-Flag, and cyclin T-6His were cotransfected into Drosophila cells. Following a standard heat shock treatment, cleared lysates were prepared from
these cells, and the lysates were then chromatographed over nickel-NTA beads, which bind the 6His-tag. Portions of the lysates and nickel-bound fractions were
then examined by Western blotting using HSF or Flag antibodies. Whereas Cdk9 is efficiently recovered in the Ni-bound fraction, HSF is not
recovered at levels exceeding the background from cells lacking cyclin T-6His. These results and the in vivo results indicate that the high levels of P-TEFb
association with heat shock loci cannot be explained by an interaction of HSF with P-TEFb (Lis, 2000).
Does the redistribution of P-TEFb to heat shock loci influence transcription of the heat shock genes? The effects of directly recruiting P-TEFb subunits,
Cdk9 or cyclin T, to the hsp70 promoter were tested. A pair of Gal4-binding sites (UASgal) was introduced upstream of a Drosophila hsp70-M reporter gene. The
expression of this hybrid reporter gene can be distinguished from native hsp70 genes since it is marked by fusion to a bacterial DNA sequence. This reporter construct
and copper-inducible expression vectors, which express the Gal4 DNA-binding alone (G4) or G4 fused to Cdk9, cyclin T and a variety of controls, were
cotransfected into Drosophila cells. The inserted UASgal sequences are upstream of the regions critical for heat shock expression, so, as anticipated, transcription of
this reporter gene is heat inducible, albeit at about a twofold lower level than the control containing no UASgal insert. The reporter gene
containing UASgal sites is strongly activated without heat shock when cells are cotransfected with G4 fused to the activation domain of HSF (G4-HSF). The reporter gene carrying the UASgal sites is also strongly activated without heat shock when cells are cotransfected with plasmids expressing G4-Cdk9 or
G4-cyclin T. A point mutation that disrupts the activity of the kinase subunit, Cdk9/D199N (Peng, 1998a), also disrupts the ability of
the G4-Cdk9 hybrid protein to activate transcription from the hsp70 reporter. The levels of expression of wild-type and mutant G4-Cdk9 are
similar. Also, a pair of mutations in cyclin T that disrupt its ability to interact with Cdk9, the double point mutant CycT/2XMut (Bieniasz, 1999),
greatly impairs the ability of G4-cyclin T to activate transcription. These results demonstrate that artificially recruiting P-TEFb to the promoter by
directly recruiting either of its two subunits is sufficient to strongly activate an hsp70 gene. A similar activation by G4-HSF, G4-Cdk9, and G4-cyclin T was observed
with UASgal sequences inserted further upstream at -256, although the level of activation was reduced two to threefold (Lis, 2000).
A model is proposed in which P-TEFb acts on promoter-paused Pol II complexes to stimulate their escape into productive elongation. If P-TEFb is a major kinase
that acts on the promoter-paused Pol II complex, its distribution should overlap at least some of the chromosomal sites that accumulate unphosphorylated RNA polymerase II (Pol IIa). However, the
correlation need not be perfect, since the rate of formation of a promoter-paused Pol IIa is likely to be governed by mechanisms distinct from those that are responsible
for recruiting P-TEFb. These mechanisms appear to be quite independent in an extreme case of heat shock genes, in which Pol IIa is present at full occupancy on the
uninduced hsp70 promoter, and heat shock is needed to trigger both high levels of transcription and recruitment of P-TEFb. However, when a gene is active, Pol IIa is being continuously recruited to the promoter and maturing into a productive hyperphosphorylated polymerase II (Pol IIo) elongation complex. In the
case of heat shock genes, the entry is fast enough to keep the pause region fully occupied with Pol II even when the gene is fully induced. Therefore, both the kinase responsible for phosphorylation and the Pol IIa would be expected to be present on active promoters, and their respective
levels would be dictated by the relative rates of Pol entry and its maturation into a productive elongation complex (Lis, 2000 and references therein).
Chromosome were stained with antibodies to P-TEFb and Pol IIa. Most chromosomal sites in unstressed larvae that are labeled strongly by
the P-TEFb (cyclin T) antibody are also labeled to various extents by the Pol IIa antibody; however, the ratio of labeling by these two antibodies varies at different sites. Therefore, the level of Pol IIa must be governed by factors that act at least somewhat independently from factors that govern the level of P-TEFb at specific
sites. Nonetheless, the strong tendency of these proteins to colocalize is consistent with a model in which Pol IIa is a substrate for P-TEFb, and this phosphorylation
serves to convert Pol II into a productive elongation complex (Lis, 2000).
The hyperphosphorylated form, PoI IIo, labels many more sites than does P-TEFb. Numerous sites are strongly labeled with antibody to Pol IIo, but not
detectably labeled with antibody to P-TEFb. This pattern does not easily fit a model in which P-TEFb has a universal role in all Pol II transcription elongation.
Presumably, there are distinct mechanisms (and other kinases) for producing Pol IIo that do not require the stable and continuous association of P-TEFb with a locus (Lis, 2000).
In contrast, there are few chromosomal sites that have P-TEFb, but no Pol IIo. A simple interpretation of this result, which is consistent with the known properties of
P-TEFb, is that the recruitment of P-TEFb to a locus generally leads to efficient formation of transcription elongation complexes. These results also indicate that there
is little recruitment of P-TEFb to sites that are not transcriptionally active (Lis, 2000).
P-TEFb is a kinase/cyclin heterodimer that was critical for overcoming an early block to transcriptional elongation
(Marshall, 1995). Interestingly, the short transcripts of 20-40 nucleotides that are produced in the absence of P-TEFb
are remarkably similar in size to those measured in vivo at genes that show promoter-associated pausing. Such pausing has been observed at a variety of genes; however, the heat shock promoters of Drosophila are perhaps the most
thoroughly studied in eukaryotes. P-TEFb stimulates production of full-length transcripts in vitro (Marshall, 1995), and also
from HIV templates in vivo (Mancebo, 1997). P-TEFb is normally located at >200 loci, but upon heat shock, it redistributes to native and transgenic heat shock loci with a robustness and rapidity that make it a
good candidate for playing a critical role in the activation of heat shock gene transcription. The normally broad distribution of P-TEFb is simplified during heat shock,
where the bulk of P-TEFb concentrates at all the major heat shock genes. The resolution of HSF and P-TEFb staining within the 87A locus is consistent with the long-held view that the DNA within this activated heat shock locus is in a very
extended configuration. The two divergently transcribed hsp70 genes at this locus are separated by only 2 kb, and yet the
P-TEFb staining resolves as two distinct bands. The major HSF staining resides between the two P-TEFb bands. HSF binding sites are known to reside in
the region between the start sites of these genes. If the DNA in a highly decondensed puff approximates B-form DNA, that is, it has a
chromatin packing ratio similar to that of highly transcribed ribosomal DNA, then the distance between the start sites of the two hsp70 genes would be ~0.7 µm. The
centers of the two bands of staining are approximately twice that distance, implying that P-TEFb may be distributed over a region that extends downstream of the
hsp70 start sites. Higher resolution biochemical methods will be required to precisely define the limits of the P-TEFb distribution. Nonetheless, the partial resolution
of HSF and P-TEFb staining supports a view that these two components act at distinct points in the process of activating heat shock genes (Lis, 2000).
P-TEFb is not simply recruited by the hypophosphorylated Pol IIa. Pol IIa is the form of Pol II that is at the promoter pause region of hsp70 and other heat shock
genes. Yet very little or no P-TEFb is detected at these sites prior the heat shock. It is speculated that a
separate event must occur at these promoters to cause the association of P-TEFb, for example, another protein or proteins could recruit P-TEFb to these promoters.
In the case of HIV, the Tat protein interacts with cyclin T to recruit the P-TEFb complex (Garber, 1998). For normal cellular genes, other host
transcription factors may also play such a role; P-TEFb has been shown recently to be functionally recruited to MHC class II gene promoters by the CIITA activator
(Kanazawa, 2000). Alternatively, transcription activation may normally allow a paused Pol IIa, a likely in vivo substrate of P-TEFb, to undergo a change
or unmasking that now allows its association with P-TEFb (Lis, 2000).
P-TEFb is normally located at many chromosomal sites that are transcriptionally active. The chromosomal loci scored as positive with the cyclin T antibody may
represent only a fraction of the genes that could be regulated by P-TEFb, owing to the dynamic developmental regulation of the Drosophila genome. Also, the
existence of additional cyclin subunits that can couple with Cdk9 may produce a P-TEFb activity lacking cyclin T. Although whether P-TEFb activates transcription at all of the loci containing cyclin T cannot be evaluated, in the case of heat shock genes, the direct recruitment (via a Gal4 DNA-binding
domain) of P-TEFb to an hsp70 promoter leads to an activation of this gene in the absence of heat shock. Although this activation is less than the very high level of
activation caused by directly recruiting the activation domain of HSF, it is nonetheless clearly dependent on P-TEFb kinase activity. Interestingly, related kinases,
Cdk2 and Cdk7, fail to activate this promoter and critical point mutations in the P-TEFb kinase or cyclin T disrupt the activation. The fact that Cdk7, the kinase of
the TFIIH complex, fails to activate is worth noting, because it, like P-TEFb, can phosphorylate efficiently the CTD of Pol II. Perhaps these kinases have specificity
for discrete steps in transcription. For example, P-TEFb may be capable of stimulating the elongation of the paused Pol II, whereas TFIIH kinase fulfills another role
such as providing activity for an earlier step in transcription that does not necessarily lead to the full phosphorylation and maturation of elongationally competent Pol
II. This specificity issue requires further, more focused investigation (Lis, 2000 and references therein).
Direct recruitment of P-TEFb to an HIV promoter has been shown to activate HIV transcription fully and bypasses the need for Tat (Bieniasz, 1999).
Although the activation of hsp70 by directly recruiting P-TEFb that is observed in uninduced cells is strong, it is still less than that seen when the HSF activation
domain is directly recruited. This fact suggests that HSF may be providing a function beyond triggering the events that lead to P-TEFb recruitment. The HSF
activation domain is large enough to accommodate multiple interactions and functions (Lis, 2000 and references therein).
The colocalization of the hypophosphorylated Pol IIa with P-TEFb is intriguing, because the promoter-paused Pol II associated with all genes examined in Drosophila is
hypophosphorylated. If the Pol IIa distribution is a general indicator of sites in which promoter-pausing is a part of the transcription
mechanism, then P-TEFb may be stimulating the maturation of Pol II and its entry into productive elongation at a significant subset of active genes. Three
of the four constitutively active genes that have been reported to have promoter-paused Pol II are at chromosomal sites that
show significant P-TEFb. The fourth, Gapdh-2, is at 13F, a region that shows light P-TEFb staining. A
higher resolution analysis will be required for a rigorous assignment of the P-TEFb signals to these specific genes (Lis, 2000 and references therein).
The failure to see a quantitative correlation of the intensity of staining of anti-Pol IIa and anti-P-TEFb at specific sites on polytene chromosomes is consistent
with models in which the mechanism of generating paused Pol IIa is distinct from the mechanism that recruits P-TEFb. The extreme case of this is hsp70, in which,
before heat shock, the promoter is fully occupied with Pol IIa, but has very little P-TEFb. Heat shock triggers the dramatic recruitment of P-TEFb, and the
accumulation of Pol IIo on heat shock puffs. During heat shock, the paused Pol IIa still forms, but it escapes
into productive elongation faster, once every 4 sec as compared with the uninduced level of once every 10 min. It is hypothesized that P-TEFb participates in this
escape at heat shock genes and the subset of other genes that have promoter-paused Pol II (Lis, 2000).
back to Heat shock factor: Regulation part 1/2
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Heat shock factor:
Biological Overview
| Evolutionary Homologs
| Effect of mutation
| References
Society for Developmental Biology's Web server.