Transcription factors: positive and negative regulators of cell growth and disease.

Recent advances in our understanding of transcription factors that regulate RNA polymerase II transcribed genes are described. This review considers the role of alternate protein-protein interactions to yield positive or negative regulation, the consequences of deregulated factor activity on cell function and approaches to establish factor-binding specificities.

Bag1-Hsp70 mediates a physiological stress signalling pathway that regulates Raf-1/ERK and cell growth.

Survival after stress requires the precise orchestration of cell-signalling events to ensure that biosynthetic processes are alerted and cell survival pathways are initiated. Here we show that Bag1, a co-chaperone for heat-shock protein 70 (Hsp70), coordinates signals for cell growth in response to cell stress, by downregulating the activity of Raf-1 kinase. Raf-1 and Hsp70 compete for binding to Bag1, such that Bag1 binds to and activates Raf-1, subsequently activating the downstream extracellular signal-related kinases (ERKs). When levels of Hsp70 are elevated after heat shock, or in cells conditionally overexpressing Hsp70, Bag1-Raf-1 is displaced by Bag1-Hsp70, and DNA synthesis is arrested. Mutants Bag1C204A and Bag1E208A, which cannot bind Hsp70, constitutively activate Raf-1/ERK kinases but are unaffected by Hsp70; consequently neither Bag1-Raf-1 nor DNA synthesis is negatively affected during heat shock. Likewise, mutants Hsp70F245S, Hsp70R262W and Hsp70L282R, which retain chaperone activity but do not bind to Bag1, fail to repress Bag1 activation of Raf-1/ERK kinase. We propose that Bag1 functions in the heat-shock response to coordinate cell growth signals and mitogenesis, and that Hsp70 functions as a sensor in stress signalling.

Heat-shock protein 70 inhibits apoptosis by preventing recruitment of procaspase-9 to the Apaf-1 apoptosome.

The cellular-stress response can mediate cellular protection through expression of heat-shock protein (Hsp) 70, which can interfere with the process of apoptotic cell death. Stress-induced apoptosis proceeds through a defined biochemical process that involves cytochrome c, Apaf-1 and caspase proteases. Here we show, using a cell-free system, that Hsp70 prevents cytochrome c/dATP-mediated caspase activation, but allows the formation of Apaf-1 oligomers. Hsp70 binds to Apaf-1 but not to procaspase-9, and prevents recruitment of caspases to the apoptosome complex. Hsp70 therefore suppresses apoptosis by directly associating with Apaf-1 and blocking the assembly of a functional apoptosome.

Mutational analysis of the human HSP70 protein: distinct domains for nucleolar localization and adenosine triphosphate binding.

The human HSP70 gene was modified in vitro using oligonucleotide-directed mutagenesis to add sequences encoding a peptide from the testis-specific form of human lactate dehydrogenase (LDH) to the carboxy terminus of HSP70. The peptide-tagged HSP70 can be distinguished from the endogenous HSP70 protein using an LDH peptide-specific antiserum in indirect immunofluorescence assays of cells transiently transfected with an expression vector containing the tagged HSP70 gene regulated by the human HSP70 promoter. A series of deletion mutants within the HSP70 protein coding region were generated. Using double-label indirect immunofluorescence with the LDH peptide-specific antiserum and HSP70-specific mAbs, we compared the intracellular distribution of the deletion mutants to that of endogenous HSP70. We have determined that sequences in the carboxy terminus of HSP70 are necessary for proper nucleolar localization after heat shock. In contrast, sequences in the amino terminus of HSP70 are responsible for the ATP-binding ability of the protein. Mutants that were unable to bind ATP, however, still displayed nucleolar association, indicating that ATP binding is apparently not required for interaction with substrate. Additional support that HSP70 appears to be composed of at least two domains follows from the results of trypsin digestions of wild type and mutant HSP70. Protease digestion of the mutant HSP70 proteins identified a region of HSP70 that, when deleted, affected HSP70 conformation.

Cell cycle-dependent association of HSP70 with specific cellular proteins.

In asynchronous populations of HeLa cells maintained at control or heat shock temperatures, HSP70 levels and its subcellular distribution exhibit substantial heterogeneity as demonstrated by indirect immunofluorescence with HSP70-specific monoclonal antibodies. Of particular interest is a subpopulation of cells in which the characteristic nuclear accumulation and nucleolar association of HSP70 is not detected after heat shock treatment. This apparent variation in the heat shock response is not observed when synchronized cells are examined. In this study, we demonstrate that three monoclonal antibodies to HSP70, in particular, do not detect nucleolar-localized HSP70 in heat-shocked G2 cells. This is not due to an inability of G2 cells to respond to heat shock as measured by increased HSP70 mRNA and protein synthesis, or due to a lack of accumulation of HSP70 after heat shock in G2. Rather the epitopes recognized by the various antibodies appear to be inaccessible, perhaps due to the association of HSP70 with other proteins. Non-denaturing immunoprecipitations with these HSP70-specific antibodies suggest that HSP70 may interact with other cellular proteins in a cell cycle-dependent manner.

Intron-independent association of splicing factors with active genes.

The cell nucleus is organized as discrete domains, often associated with specific events involved in chromosome organization, replication, and gene expression. We have examined the spatial and functional relationship between the sites of heat shock gene transcription and the speckles enriched in splicing factors in primary human fibroblasts by combining immunofluorescence and fluorescence in situ hybridization (FISH). The hsp90alpha and hsp70 genes are inducibly regulated by exposure to stress from a low basal level to a high rate of transcription; additionally the hsp90alpha gene contains 10 introns whereas the hsp70 gene is intronless. At 37 degrees C, only 30% of hsp90alpha transcription sites are associated with speckles whereas little association is detected with the hsp70 gene, whose constitutive expression is undetectable relative to the hsp90alpha gene. Upon exposure of cells to heat shock, the heavy metal cadmium, or the amino acid analogue azetidine, transcription at the hsp90alpha and hsp70 gene loci is strongly induced, and both hsp transcription sites become associated with speckles in >90% of the cells. These results reveal a clear disconnection between the presence of intervening sequences at specific gene loci and the association with splicing factor-rich regions and suggest that subnuclear structures containing splicing factors are associated with sites of transcription.

Molecular chaperone machines: chaperone activities of the cyclophilin Cyp-40 and the steroid aporeceptor-associated protein p23.

Molecular chaperones are essential proteins that participate in the regulation of steroid receptors in eukaryotes. The steroid aporeceptor complex contains the molecular chaperones Hsp90 and Hsp70, p48, the cyclophilin Cyp-40, and the associated proteins p23 and p60. In vitro folding assays showed that Cyp-40 and p23 functioned as molecular chaperones in a manner similar to that of Hsp90 or Hsp70. Although neither Cyp-40 nor p23 could completely refold an unfolded substrate, both proteins interacted with the substrate to maintain a nonnative folding-competent intermediate. Thus, the steroid aporeceptor complexes have multiple chaperone components that maintain substrates in an intermediate folded state.

Effect of sodium salicylate on the human heat shock response.

Sodium salicylate, an anti-inflammatory agent, was examined for its effects on the heat shock response in cultured human cells. Salicylate activation of DNA binding by the heat shock transcription factor (HSF) was comparable to activation attained during heat shock. However, sodium salicylate did not induce heat shock gene transcription even though the HSF was bound in vivo to the heat shock elements upstream of the heat shock protein 70 (Hsp 70) gene. These results reveal that activation of the heat shock transcriptional response is a multistep process. Modulation of extracellular pH augments sensitivity to salicylate-induced activation of HSF.

Activation of heat shock transcription factor 3 by c-Myb in the absence of cellular stress.

In vertebrates, the presence of multiple heat shock transcription factors (HSFs) indicates that these factors may be regulated by distinct stress signals. HSF3 was specifically activated in unstressed proliferating cells by direct binding to the c-myb proto-oncogene product (c-Myb). These factors formed a complex through their DNA binding domains that stimulated the nuclear entry and formation of the transcriptionally active trimer of HSF3. Because c-Myb participates in cellular proliferation, this regulatory pathway may provide a link between cellular proliferation and the stress response.

Complex regulation of heat shock- and glucose-responsive genes in human cells.

We have isolated a human genomic clone that encodes the glucose-responsive protein GRP78 and have used this cloned gene probe, together with a cloned HSP70 gene, to study the expression of both stress-induced genes in response to inhibitors of cellular metabolism. On the basis of the effects of this group of chemicals on GRP78 and HSP70 expression, we have identified three classes of stress gene inducers. The first class induces GRP78 expression and includes inhibitors of glycoprotein processing. The second class results in coordinate activation of both GRP78 and HSP70 synthesis and includes amino acid analogs and heavy metals. Chemicals in the third class coordinately induce GRP78 and repress HSP70 expression; this class includes the calcium ionophore A23187 and the glucose analog 2-deoxyglucose. Whereas induction of GRP78 or HSP70 expression is primarily due to transcriptional activation, chemicals that repress HSP70 expression act through posttranscriptional regulation. These results reveal that the regulation of GRP78 and HSP70 expression is complex and may be dependent on the specificity and magnitude of physiological damage.

Maximal stress-induced transcription from the human HSP70 promoter requires interactions with the basal promoter elements independent of rotational alignment.

Transcription of the human HSP70 gene is regulated by a complex array of cis-acting promoter elements that respond to conditions that include normal conditions of cell growth and induction following physiological stress. We have examined the requirements of the basal and inducible promoter elements by using promoter mutations and a transient transfection assay. Multiple forms of stress-induced transcription, including heat shock and incubation with heavy metals or amino acid analogs, are mediated by a single heat shock element (HSE) between -105 and -91 consisting of three contiguous 5-base-pair units, NGAAN, that are inverted relative to adjacent units. Maximal inducible expression requires a fully functional basal promoter. Spacing mutations which alter the relative helical orientation of adjacent genetic elements have only minimal effects on basal and stress-inducible expression and show no effects of periodicity. In addition, placement of the HSE adjacent to the basal promoter removes the requirements for a fully functional basal promoter for maximal stress-inducible expression. These results suggest that factors bound at the HSE and the basal promoter can function through multiple interactions.

Selection of new HSF1 and HSF2 DNA-binding sites reveals difference in trimer cooperativity.

Multiple heat shock transcription factors (HSFs) have been discovered in several higher eukaryotes, raising questions about their respective functions in the cellular stress response. Previously, we had demonstrated that the two mouse HSFs (mHSF1 and mHSF2) interacted differently with the HSP70 heat shock element (HSE). To further address the issues of cooperativity and the interaction of multiple HSFs with the HSE, we selected new mHSF1 and mHSF2 DNA-binding sites through protein binding and PCR amplification. The selected sequences, isolated from a random population, were composed primarily of alternating inverted arrays of the pentameric consensus 5'-nGAAn-3', and the nucleotides flanking the core GAA motif were nonrandom. The average number of pentamers selected in each binding site was four to five for mHSF1 and two to three for mHSF2, suggesting differences in the potential for cooperative interactions between adjacent trimers. Our comparison of mHSF1 and mHSF2 binding to selected sequences further substantiated these differences in cooperativity as mHSF1, unlike mHSF2, was able to bind to extended HSE sequences, confirming previous observations on the HSP70 HSE. Certain selected sequences that exhibited preferential binding of mHSF1 or mHSF2 were mutagenized, and these studies demonstrated that the affinity of an HSE for a particular HSF and the extent of HSF interaction could be altered by single base substitutions. The domain of mHSF1 utilized for cooperative interactions was transferable, as chimeric mHSF1/mHSF2 proteins demonstrated that sequences within or adjacent to the mHSF1 DNA-binding domain were responsible. We have demonstrated that HSEs can have a greater affinity for a specific HSF and that in mice, mHSF1 utilizes a higher degree of cooperativity in DNA binding. This suggests two ways in which cells have developed to regulate the activity of closely related transcription factors: developing the ability to fully occupy the target binding site and alteration of the target site to favor interaction with a specific factor.

Posttranscriptional regulation of hsp70 expression in human cells: effects of heat shock, inhibition of protein synthesis, and adenovirus infection on translation and mRNA stability.

We have examined the posttranscriptional regulation of hsp70 gene expression in two human cell lines, HeLa and 293 cells, which constitutively express high levels of HSP70. HSP70 mRNA translates with high efficiency in both control and heat-shocked cells. Therefore, heat shock is not required for the efficient translation of HSP70 mRNA. Rather, the main effect of heat shock on translation is to suppress the translatability of non-heat shock mRNAs. Heat shock, however, has a marked effect on the stability of HSP70 mRNA; in non-heat-shocked cells the half-life of HSP70 mRNA is approximately 50 min, and its stability increases at least 10-fold upon heat shock. Moreover, HSP70 mRNA is more stable in cells treated with protein synthesis inhibitors, suggesting that a heat shock-sensitive labile protein regulates its turnover. An additional effect on posttranscriptional regulation of hsp70 expression can be found in adenovirus-infected cells, in which HSP70 mRNA levels decline precipititously late during infection although hsp70 transcription continues unabated.

Heat shock-induced interactions of heat shock transcription factor and the human hsp70 promoter examined by in vivo footprinting.

Genomic footprinting of the human hsp70 promoter reveals that heat shock induces a rapid binding of a factor, presumably heat shock transcription factor, to a region encompassing five contiguous NGAAN sequences, three perfect and two imperfect matches to the consensus sequence. Arrays of inverted NGAAN sequences have been defined as the heat shock element. No protein is bound to the heat shock element prior to or after recovery from heat shock. Heat shock does not perturb the binding of factors to other regulatory elements in the promoter which contribute to basal expression of the hsp70 gene.

Repression of the heat shock factor 1 transcriptional activation domain is modulated by constitutive phosphorylation.

Heat shock transcription factor 1 (HSF1) is constitutively expressed in mammalian cells and negatively regulated for DNA binding and transcriptional activity. Upon exposure to heat shock and other forms of chemical and physiological stress, these activities of HSF1 are rapidly induced. In this report, we demonstrate that constitutive phosphorylation of HSF1 at serine residues distal to the transcriptional activation domain functions to repress transactivation. Tryptic phosphopeptide analysis of a collection of chimeric GAL4-HSF1 deletion and point mutants identified a region of constitutive phosphorylation encompassing serine residues 303 and 307. The significance of phosphorylation at serines 303 and 307 in the regulation of HSF1 transcriptional activity was demonstrated by transient transfection and assay of a chloramphenicol acetyltransferase reporter construct. Whereas the transfected wild-type GAL4-HSF1 chimera is repressed for transcriptional activity and derepressed by heat shock, mutation of serines 303 and 307 to alanine results in derepression to a high level of constitutive activity. Similar results were obtained with mutation of these serine residues in the context of full-length HSF1. These data reveal that constitutive phosphorylation of serines 303 and 307 has an important role in the negative regulation of HSF1 transcriptional activity at control temperatures.

Characterization of a novel chicken heat shock transcription factor, heat shock factor 3, suggests a new regulatory pathway.

We have cloned three avian heat shock transcription factor (HSF) genes corresponding to a novel factor, HSF3, and the avian homologs of mammalian HSF1 and HSF2. The predicted amino acid sequence of HSF3 is approximately 40% related to the sequence of HSF1 and HSF2. The sequences for all three factors exhibit extensive identify in the DNA binding motifs and the heptad repeats of hydrophobic amino acids which are common to all eukaryotic HSFs. Despite these overall similarities, each avian HSF exhibits distinct DNA binding properties. HSF2 when expressed in vitro binds constitutively to the heat shock element promoter sequence, whereas neither HSF1 nor HSF3 expressed in vitro binds to DNA. HSF1 DNA binding is induced upon heat shock or treatment with nonionic detergents, whereas the DNA binding properties of HSF3 are not induced by these conditions in vitro. These results suggest that HSF3 activation may involve an induction pathway distinct from the traditional forms of heat shock gene induction. HSF3 DNA binding activity, however, is obtained when the carboxyl-terminal region including the distal heptad repeat is deleted, indicating the presence of negative cis-regulatory sequences. The HSF3 message, like HSF1 and HSF2 messages, is coexpressed during development and in most tissues, which suggests a general role for the regulatory pathway involving HSF3.

Coordinate changes in heat shock element-binding activity and HSP70 gene transcription rates in human cells.

Activation of human heat shock gene transcription by heat shock, heavy metal ions, and amino acid analogs required the heat shock element (HSE) in the HSP70 promoter. Both heat shock- and metal ion-induced HSP70 gene transcription occurred independently of protein synthesis, whereas induction by amino acid analogs required protein synthesis. We identified a HSE-binding activity from control cells which was easily distinguished by a gel mobility shift assay from the stress-induced HSE-binding activity which appeared following heat shock or chemically induced stress. The kinetics of HSP70 gene transcription paralleled the rapid appearance of stress-induced HSE-binding activity. During recovery from heat shock, both the rate of HSP70 gene transcription and stress-induced HSE-binding activity levels declined and the control HSE-binding activity reappeared. The DNA contacts of the control and stress-induced HSE-binding activities deduced by methylation interference were similar but not identical. While stable complexes with HSE were formed with extracts from both control and stressed cells in vitro at 25 degrees C, only the stress-induced complex was detected when binding reactions were performed at elevated temperatures.

Mouse heat shock transcription factors 1 and 2 prefer a trimeric binding site but interact differently with the HSP70 heat shock element.

To understand the function of multiple heat shock transcription factors in higher eukaryotes, we have characterized the interaction of recombinant mouse heat shock transcription factors 1 and 2 (mHSF1 and mHSF2) with their binding site, the heat shock element (HSE). For our analysis, we utilized the human HSP70 HSE, which consists of three perfect 5'-nGAAn-3' sites (1, 3, and 4) and two imperfect sites (2 and 5) arranged as tandem inverted repeats. Recombinant mHSF1 and mHSF2, which exist as trimers in solution, both bound specifically to this HSE and stimulated transcription of a human HSP70-CAT construct in vitro. Footprinting analyses revealed differential binding of mHSF1 and mHSF2 to the HSP70 HSE. Specifically, mHSF1 bound all five pentameric sites, whereas mHSF2 failed to interact with the first site of the HSE but bound to sites 2 to 5. Missing-nucleoside analysis demonstrated that the third and fourth nGAAn sites were essential for mHSF1 and mHSF2 binding. The binding of the initial mHSF1 trimer to the HSE exhibited preference for sites 3, 4, and 5, and then binding of a second trimer occurred at sites 1 and 2. These results suggest that HSF may recognize its binding site through the dyad symmetry of sites 3 and 4 but requires an adjacent site for stable interaction. Our data demonstrate that mHSF1 and mHSF2 bind specifically to the HSE through major groove interactions. Methidiumpropyl-EDTA footprinting revealed structural differences in the first and third repeats of the HSE, suggesting that the DNA is distorted in this region. The possibility that the HSE region is naturally distorted may assist in understanding how a trimer of HSF can bind to what is essentially an inverted repeat binding site.

E1a transactivation of the human HSP70 promoter is mediated through the basal transcriptional complex.

We have examined the promoter sequence requirements for E1a transactivation of the human HSP70 gene by using a transient cotransfection assay. A 5' deletion study has defined a basal transcription unit extending to -74 relative to the transcription initiation site which was fully E1a responsive. Further deletion, abolishing a CCAAT element at -67, drastically reduced basal and E1a-induced expression. A linker-scanner analysis has identified four functional elements within the basal transcription unit which may interact with CTF, SP1, TFIID, and an ATF/AP1-like factor. Sequences between -100 and -188 can partially compensate for mutations in these elements. No mutation specifically abolished E1a inducibility. Any reduction in absolute E1a-induced levels was accompanied by a corresponding reduction in absolute basal levels, thereby maintaining a constant relative fold induction. We conclude that E1a transactivation of the human HSP70 promoter does not require any single basal transcription element. We also examined an HSP70 promoter fragment, containing the CCAAT element at -67 and the purine-rich element at -54, out of its normal context by fusing it upstream of a transcriptionally inactive herpes simplex virus thymidine kinase deletion construct containing only the TATA box. The resulting chimeric promoter was fully E1a responsive. Mutagenesis of this promoter fusion demonstrated that the CCAAT element was essential for detectable basal and E1a-induced expression. Mutations in the purine-rich element resulted in an approximately 10-fold elevation in basal levels and rendered the promoter nonresponsive to E1a.

Human heat shock factors 1 and 2 are differentially activated and can synergistically induce hsp70 gene transcription.

Two members of the heat shock transcription factor (HSF) family, HSF1 and HSF2, both function as transcriptional activators of heat shock gene expression. However, the inducible DNA-binding activities of these two factors are regulated by distinct pathways. HSF1 is activated by heat shock and other forms of stress, whereas HSF2 is activated during hemin-induced differentiation of human K562 erythroleukemia cells, suggesting a role for HSF2 in regulating heat shock gene expression under nonstress conditions such as differentiation and development. To understand the distinct regulatory pathways controlling HSF2 and HSF1 activities, we have examined the biochemical and physical properties of the control and activated states of HSF2 and compared these with the properties of HSF1. Our results reveal that the inactive, non-DNA-binding forms of HSF2 and HSF1 exist primarily in the cytoplasm of untreated K562 cells as a dimer and monomer, respectively. This difference in the control oligomeric states suggests that the mechanisms used to control the DNA-binding activities of HSF2 and HSF1 are distinct. Upon activation, both factors acquire DNA-binding activity, oligomerize to a trimeric state, and translocate into the nucleus. Interestingly, we find that simultaneous activation of both HSF2 and HSF1 in K562 cells subjected to hemin treatment followed by heat shock results in the synergistic induction of hsp70 gene transcription, suggesting a novel level of complex regulation of heat shock gene expression.