Transcriptional activation by the nuclear protein Hap50

Download PDF
5 downloads 0 Views 220KB Size Report
Comment
Key words: BAG-1, c-Jun, c-Fos, Estrogen receptor, Glucocorticoid receptor ... The fusion construct of Hap50 with the green fluorescent protein. (GFP) was ...
RESEARCH ARTICLE

1839

Transcriptional activation by the human Hsp70associating protein Hap50 Yilmaz Niyaz, Matthias Zeiner and Ulrich Gehring* Ruprecht-Karls-Universität Heidelberg, Biochemie-Zentrum Heidelberg, Biologische Chemie, Im Neuenheimer Feld 501, D-69120 Heidelberg, Germany *Author for correspondence (e-mail: [email protected])

Accepted 22 February 2001 Journal of Cell Science 114, 1839-1845 © The Company of Biologists Ltd

SUMMARY We investigated human Hap50, the large isoform of the previously characterized Hsp70/Hsc70-associating protein Hap46, also called BAG-1, for effects on transcriptional activities. Overproduction by transient transfection led to enhanced expression of reporter gene constructs in various cell types using different promoters, suggesting independence of promoter type. Similarly, overexpression of Hap50 resulted in increased levels of poly(A)+ mRNAs in HeLa, COS-7, 3T3 and HTC cells. Concomitantly, the expression of some selected endogenous genes, such as those coding for c-Jun and the glucocorticoid receptor, was enhanced significantly relative to actin. Nuclear runoff transcription assays using HeLa cells showed that the effect is caused by increased transcription rates rather than mRNA stabilization. Activation of transcription by Hap50

occurred at 37°C and did not require prior thermal stress, as is the case for Hap46. In accordance with these biological effects, Hap50 is localized exclusively in the nuclear compartment of different cell types, whereas Hap46 is mostly cytoplasmic in unstressed cells, as revealed by use of fusion constructs with green fluorescent protein. High cellular levels of Hap50 were found to make cells less susceptible to adverse environmental effects such as heat stress. Our data suggest that Hap50 is a nuclear protein that acts in cells to increase the transcription of various genes.

INTRODUCTION

Although binding to Hsp70s involves the C-terminal region of Hap46/BAG-1 (Takayama et al., 1998; Takayama et al., 1999), a function for the N-terminal portion of the polypeptide has only recently been recognized. Direct interaction with DNA makes use of this part of the molecule (Zeiner et al., 1999), which contains clusters of basic amino acid residues and glutamic acid-rich repeats (Zeiner and Gehring, 1995). Furthermore, Hap46 can stimulate overall transcription in vitro but, in intact cells, heat stress was found to be a prerequisite for transcriptional activation (Zeiner et al., 1999). Mammalian cells can express several isoforms of Hap46/BAG-1 proteins (Packham et al., 1997; Takayama et al., 1998; Yang et al., 1998), but the relative levels of individual forms vary largely between cell lines (Brimmell et al., 1999). The largest of these isoforms was first detected by immunoblotting and was found to originate from a noncanonical CUG translational start codon (Packham et al., 1997). Because it has an apparent molecular mass of about 50 kDa we call it Hap50, implying close relation to Hap46. This large isoform has previously been detected in several cell lines and tissues (Packham et al., 1997; Takayama et al., 1998; Yang et al., 1998; Yang et al., 1999a; Yang et al., 1999b; Crocoll et al., 2000). Compared with Hap46, the isoform Hap50 is longer at the N-terminus by 71 amino acid residues (Packham et al., 1997). We wondered whether Hap50 may similarly affect transcriptional activities in cells upon overexpression. This is indeed what we observed in the present study; however, in striking contrast to Hap46, thermal stress is not required for

The ubiquitously expressed mammalian protein Hap46, also called BAG-1, BAG-1M or p46 BAG-1, has been described to interact with a great variety of proteins, most notably with various transcription factors such as nuclear receptors c-Jun, cFos, CREB and c-Myc (Zeiner and Gehring, 1995; Zeiner et al., 1997; Froesch et al., 1998; Kullmann et al., 1998; Liu et al., 1998; Kanelakis et al., 1999, Schneikert et al., 2000; Guzey et al., 2000). Significantly, Hap46/BAG-1 was found to associate directly with members of the 70 kDa heat shock protein family (i.e. stress-inducible Hsp70 and constitutively expressed Hsc70) (reviewed by Höhfeld, 1998), hence the designation Hap46 for Hsp70/Hsc70-associating protein of apparent molecular mass 46 kDa (Gebauer et al., 1997; Zeiner et al., 1997). Most likely, the majority of the above interactions with factors that are structurally very different are mediated by Hsp70 or Hsc70 (Zeiner et al., 1997; Bimston et al., 1998; Gebauer et al., 1998), as these molecular chaperones complex with a great variety of proteins by making use of their highly homologous carboxy terminal substrate binding domains. By contrast, Hap46/BAG-1 interacts with the amino terminal ATPbinding domain of Hsp70 or Hsc70 (Höhfeld, 1998; Takayama et al., 1998; Petersen et al., 2001). Thus Hap46/BAG-1 together with Hsp70s and various other proteins can readily form ternary complexes, as was demonstrated in vitro for several model proteins (Zeiner et al., 1997; Bimston et al., 1998), including c-Jun.

Key words: BAG-1, c-Jun, c-Fos, Estrogen receptor, Glucocorticoid receptor, Hap46, Hap50, Hsp70, Hsc70

1840

JOURNAL OF CELL SCIENCE 114 (10)

Hap50 to produce this effect. This observation coincides with the localization of Hap50 in cell nuclei under normal physiological conditions.

MATERIALS AND METHODS Plasmids The cDNA encoding Hap50 was amplified by PCR from the original template (Zeiner and Gehring, 1995) using primers 5′ ATAGAAGCTTCAAGTGCGGGCATGGCTCAGCG 3′ and 5′ TGGAATTCTGCTACACCTCACTCGGC 3′, thereby introducing HindIII and EcoRI restriction sites. The Hap50 cDNA was inserted into HindIII and EcoRI sites of plasmid pcDNA3.1/HisA (Invitrogen) from which the polyhistidine coding sequence had been removed with HindIII. The fusion construct of Hap50 with the green fluorescent protein (GFP) was generated by PCR from the original cDNA template with primers 5′ ATAGAAGCTTCAAGTGCGGGCATGGCTCAGCG 3′ and 5′ AAAGGATCCACACCTAACTCGGCCAG 3′, followed by inserting the Hap50 cDNA into HindIII and BamHI sites of pEGFPN1 (CLONTECH). The CTG initiation codon was hereby changed to ATG in both constructs to improve expression of the Hap50 isoform. The Hap46-GFP construct was the same as before (Zeiner et al., 1999). The vector pcDNA3/CAT (Invitrogen) contains the chloramphenicol acetyltransferase (CAT) gene under control of the cytomegalovirus (CMV) promoter. CAT expression from reporter plasmid pBL8 CAT is driven by the Herpes simplex type 1 thymidine kinase (TK) promoter coupled to glucocorticoid response elements (GREs). The pColl reporter plasmid codes for CAT under control of the human collagenase promoter (Angel et al., 1987). Cell culture and transfections Human HeLa, monkey COS-7, mouse Swiss 3T3, and rat HTC hepatoma cells were grown in RPMI 1640 medium supplemented with 10% fetal calf serum. Cell viability was ascertained by trypan blue exclusion. For detailed microscopic analysis, a Leica TCS SP MP instrument (Leica) was used. For metabolic labelling, cells were treated as above but, 6 hours before harvesting, received phosphatefree medium supplemented with 200 µCi [33P]phosphoric acid (4000 Ci/mmol, ICN) per 50 mm plate Transfections were with 15 µl FuGENE 6 Transfection Reagent (Roche Molecular Biochemicals) and 5 µg Hap50 expression or control vector (either the empty vector or vector containing Hap50 DNA in antisense orientation) per 50 mm plate. Typically, 50-70% of cells turned out to be transfected, as determined by use of the GFP cDNA. In experiments with reporter plasmids, 4 µg Hap50 expression vector and 1 µg reporter plasmid were used. For testing reporter gene expression under control of the collagenase promoter or the TK promoter coupled to GREs, cells were cultured in the presence of 50 nM 12-O-tetradecanoylphorbol-13-acetate (TPA) (Sigma) or 1 µM dexamethasone (Sigma), respectively. Omitting induction by TPA or dexamethasone resulted in negligible levels of CAT activity (0.03 to 0.05% of controls). Cells were harvested 48 hours after transfection either for RNA isolation or for CAT assays (Ausubel et al., 1995). In some experiments (Fig. 3; Fig. 5), 24 hours after transfection a 42°C heat shock was applied for 2 hours and cells were then further incubated at 37°C for another 24 hours before analysis. In the experiment in Fig. 3, COS-7 cells were co-transfected with expression vector pSV2Wrec (1 µg) encoding the murine glucocorticoid receptor. Immunoblotting Cell extracts were prepared in buffer containing 2% SDS, analyzed by electrophoresis in 10% polyacrylamide gels and transferred to Immobilon-P membranes (Millipore) as before (Zeiner and Gehring, 1995). Hap50 was detected by mouse monoclonal antibody CC9E8

(Yang et al., 1998), peroxidase-conjugated secondary antibody (Sigma), and enhanced chemiluminescence (ECL) (Amersham). RNA analysis Equal numbers of cells were ruptured by use of QIAShredder spin columns (Qiagen) in RLT chaotropic lysis buffer (Qiagen). [33P]labelled poly(A)+ mRNA was isolated on Oligotex oligo-dT affinity matrix (Qiagen). Aliquots (1/50 of total) were used for direct scintillation counting of Cerenkov radiation, whereas the major portion of samples were run on formaldehyde-containing 1% agarose gels, transferred onto nylon filters (Ausubel et al., 1995) and submitted to autoradiography with a 16 hour exposure. After radioactive decay for at least two half-lives, the same filters were used for hybridizations under high stringency conditions (Ausubel et al., 1995) with human cDNAs for glucocorticoid receptor, c-Jun and β-actin labelled with [32P]CTP (3000 Ci/mmol, ICN). Electrophoresis was in formaldehydecontaining 1.2% agarose gels with 0.2 µg/ml ethidium bromide. For nuclear runoff transcription assays, nuclei from 3×107 transfected and control HeLa cells were prepared by Dounce homogenization in lysis buffer containing 0.5% NP-40 (Ausubel et al., 1995). Runoff transcription in the presence of [32P]UTP (3000 Ci/mmol, ICN) was carried out according to a standard protocol (Ausubel et al., 1995). After rupturing nuclei by QIAShredder spin columns and DNase I treatment, total labelled RNA was extracted by use of the RNeasy Mini Kit (Qiagen) and used for hybridization with cDNA probes spotted onto nylon membranes (Ausubel et al., 1995).

RESULTS Hap50 is localized in nuclei of various cell types in contrast to Hap46 We first checked the cell lines used in this study for the presence of Hap46 and the large isoform Hap50. By immunoblotting of total cell extracts, we observed that the levels of these proteins varied greatly between cell types. Although untransfected HeLa human cervical carcinoma cells and HTC rat hepatoma cells yielded distinct Hap46 and Hap50 immunosignals, murine 3T3 fibroblasts and COS-7 monkey kidney cells were found to contain only minute amounts of immunochemically detectable levels of Hap46/BAG-1 isoforms (data not shown). To investigate the intracellular distribution of Hap50 and Hap46 isoforms, we used the respective cDNAs and coupled them to cDNA encoding GFP from the jellyfish Aequorea victoria. The plasmids were transfected into HeLa, 3T3 and HTC cells, and in situ analysis was by confocal laser-scanning microscopy (CLSM) (Fig. 1). In cells expressing the Hap50GFP fusion protein, the GFP signal was exclusively localized within nuclei (Fig. 1, upper panels). As controls, the same live cells were viewed by differential interference contrast microscopy (DICM) to visualize the respective cell shapes and nuclei (Fig. 1, middle panels). Our observation of exclusive nuclear localization of Hap50 conforms with and extends previous results of cell fractionation studies (Packham et al., 1997; Takayama et al., 1998; Yang et al., 1998; Brimmell et al., 1999). By contrast, cells transfected with the Hap46-GFP construct and kept at 37°C showed predominantly cytoplasmic fluorescence (Fig. 1, lower panels). Hap50 stimulates the expression of reporter gene constructs To check for the effects on reporter gene expression, various

Transcriptional activation by the nuclear protein Hap50

1841

experiments with a tetracycline-regulated system for the expression of Hap50 (data not shown), we obtained effects on reporter gene expression comparable with those presented in Fig. 3. Hap50 stimulates the expression of selected endogenous genes The above observed stimulation of reporter gene expression prompted us to ask whether Hap50 elicits overall effects on the transcription of endogenous genes. Gene transfer-mediated overexpression of Hap50 resulted in two- to fivefold levels of total poly(A)+ [33P]phosphate-labelled mRNAs in HeLa, COS-7, 3T3 and HTC cells, as quantified by scintillation counting. Gel electrophoretic analysis of mRNA preparations disclosed somewhat different patterns amongst cell lines, as shown for HeLa and COS-7 cells (Fig. 4A), which is consistent with cell type specificity in gene expression. Interestingly, there were no gross changes detectable in the individual mRNA patterns upon Hap50 overproduction, but the levels of Fig. 1. Intracellular localization of Hap50 and Hap46. HeLa, 3T3, and HTC cells were gene expression were increased (Fig. 4A). transfected either with Hap50-GFP or Hap46-GFP constructs, as described in We also checked some of the above Materials and Methods. After 24 hours, cells were used either for CLSM (upper and lower panels) or for DICM (middle panels), using the same optical fields for both poly(A)+ mRNA preparations for individual techniques in the case of cells transfected with cDNA for Hap50-GFP. For better messages using specific hybridization probes. analysis, optical fields containing only few transfected cells were selected. DICM Clearly, the specific mRNAs for the photographs also depict all the nontransfected cells within the respective fields. glucocorticoid receptor and c-Jun were Control experiments with cDNA encoding solely GFP showed uniform distribution of present at significantly higher levels upon the fluorescence signal throughout cells. overexpression of Hap50, while the abundance of actin messages remained unaffected (Fig. cell lines were transiently transfected with Hap50 cDNA. 4B). Levels of 18S and 28S rRNAs in total RNA preparations Roughly a 10-fold overexpression of Hap50 was achieved in were unaffected by overexpression of Hap50 in these cells HeLa cells relative to endogenous levels, as detected by (data not shown). immunoblotting (Fig. 2, lane 3 vs 1). We observed that the To distinguish between increased transcription rates and levels of Hap50 were essentially the same either without added message stability, we carried out nuclear runoff transcription DNA (lane 1), or upon transfection with empty control vector assays. [32P]UTP-labelled nascent RNA transcripts obtained (not shown) or Hap50 DNA in the antisense orientation (lane from HeLa cell nuclei after transfection with Hap50 cDNA 2). (Fig. 4C, lower panels) or control vector (upper panels) were Transcriptional activation was investigated by cohybridized with several specific cDNAs. Amongst nuclear transfection of HeLa cells with CAT reporter constructs. CAT receptors, both glucocorticoid and estrogen receptor transcripts under the control of either the CMV promoter, the human were found to be significantly increased (roughly tenfold). collagenase promoter, or the TK promoter coupled to GREs, Similarly, the message levels for the AP-1 proteins c-Jun and produced a significant increase in CAT activity upon c-Fos were much higher in cells upon Hap50 overexpression. overexpression of Hap50 (Fig. 3, bars 2, 5, 8 vs 1, 4, 7, However, the expression of other genes may not be influenced respectively). significantly by Hap50, as was observed for actin. Interestingly, this transcriptional stimulation by high cellular Upon realizing that gene transfer-mediated increases of levels of Hap50 occurred upon maintaining the cells at 37°C and was not further enhanced in response to a 42°C heat shock treatment (Fig. 3, bars 3, 6, 9). This is in clear contrast to observations with Hap46, which required thermal stress to elicit a stimulatory effect on transcription (Zeiner et al., 1999). To check for cell type and species specificity, we carried out Fig. 2. Expression of Hap50 in transfected and control cells. HeLa similar experiments with COS-7 cells and 3T3 fibroblasts. Fig. cells were submitted to the transfection protocol either without DNA 3 (lanes 7, 8; 10, 11; 12, 13) compares the results obtained in (lane 1), with Hap50 cDNA (lane 3) or with Hap50 DNA in the three different cell lines using CAT under the control of the TK antisense orientation (lane 2). Extracts from equal numbers of cells promoter and GREs. Overexpression of Hap50 produced (equivalent to 3.5×105 cells each) were used for specific approximately CAT activities fivefold those of controls. In immunoblotting.

1842

JOURNAL OF CELL SCIENCE 114 (10) 1

2

3

4

5

6

7

8

9

10

11

12

13

-

+

+

-

+

-

+

+

+ +

-

+

-

-

+

+

-

+

-

7

CAT activity [fold increase]

6

5

4

3

2

1

0

Hap50 heat shock

-

-

Fig. 3. Hap50 stimulation of reporter gene expression. HeLa cells (bars 1-9), COS-7 cells (bars 10, 11), and 3T3 fibroblasts (bars 12, 13) were transfected with Hap50 expression vector or control vector, as indicated, in combination with CAT constructs driven by the CMV promoter (bars 1-3), the human collagenase promoter (bars 4-6), or the TK promoter in conjunction with GREs (bars 7-13). In some experiments (bars 3, 6, 9), cells received a heat shock as described in Materials and Methods. CAT activity in cell extracts is expressed as fold stimulation of acetylated chloramphenicol formed. Data show averages of three independent experiments with error bars indicating maximum deviations.

Hap50 levels may result in augmented expression of cellular genes, we wondered whether this could lead to overall effects on the physiology of cells. Because heat shock is known to cause large and generalized reductions of transcriptional activities, we examined whether Hap50 overexpression might improve the tolerance under such stress conditions. The viability of 3T3 fibroblasts drastically decreased following a 42°C treatment (Fig. 5, bar 3), but overexpression of Hap50 significantly reduced the extent of this effect on cell viability (Fig. 5, bar 4). By contrast, the growth rate of unstressed cells was not affected by Hap50 overexpression (Fig. 5, bar 2 vs 1). We used 3T3 cells for this heat shock experiment because they were most sensitive to thermal stress amongst the cell lines studied here. For comparison, HeLa cells readily survived extended treatment at 45°C (results not shown) and have been found to express Hsp70, even at 37°C, in addition to Hsc70 (Zeiner et al., 1997). DISCUSSION Originally, Hap46/BAG-1 was discovered and its cDNA cloned by interaction screening using several completely unrelated bait proteins, all expressed in the baculovirus system (Takayama et al., 1995; Zeiner and Gehring, 1995; Bardelli et al., 1996). This curious convergence is explained by the fact that crude extracts of virus-infected Sf9 cells had been used that contained Hsp70s, and these chaperones were subsequently found to mediate a host of interactions by forming ternary complexes (Zeiner et al., 1997; Bimston et al., 1998). BAG-1 was also detected by interaction screening with an oligonucleotide sequence from a human polyomavirus

Fig. 4. Effect of Hap50 on the expression of endogenous genes. HeLa (lanes 1, 2) and COS-7 (lanes 3, 4) cells were either transfected with Hap50 cDNA (lanes 2, 4) or control vector (lanes 1, 3) and metabolically labelled with [33P]phosphate. Poly(A)+ mRNA was submitted to electrophoresis and autoradiography. Positions of 18S and 28S rRNAs are indicated. (A) The same filters were subsequently used for hybridization with [32P]-labelled cDNAs for human glucocorticoid receptor (GR), c-Jun and β-actin. Detection was by autoradiography. (B) For nuclear runoff transcription assays, HeLa cells either transfected with Hap50 cDNA (lower panels) or control vector (upper panels) were used. Labelled RNAs were probed with dot-blotted cDNAs encoding the glucocorticoid receptor (GR), the estrogen receptor (ER), c-Jun, c-Fos and actin, as indicated (C).

(Devireddy et al., 2000), suggesting the potential to interact with nucleic acids. In fact, human Hap46 was previously recognized to directly interact with DNA through a positively charged region at the N-terminus (Zeiner et al., 1999). Such DNA binding is a prerequisite for the ability of Hap46 to elicit general transcriptional activation in in vitro assays. However, in intact cells, thermal stress is required for Hap46 to produce similar enhancement of transcription (Zeiner et al., 1999). This agrees with the observation that, under normal cell culture

Transcriptional activation by the nuclear protein Hap50 1

2

3

4

Hap50

-

+

-

+

heat shock

-

-

+

+

2.5

2.0

Viable cells [×106]

conditions, Hap46 is localized mostly in the cytoplasm (Fig. 1, lower panels) and preferentially accumulates in nuclei only upon heat shock (Zeiner et al., 1999). Thermal stress thus causes Hap46 to get transferred to the cell nucleus, possibly in concert with Hsp70, which itself has long been known for such nuclear translocation (Velazquez and Lindquist, 1984). The large isoform Hap50 exhibits a very different cellular distribution pattern. Biochemical fractionations disclosed preferential nuclear localization of Hap50 in various cell types (Packham et al., 1997; Takayama et al., 1998; Yang et al., 1998; Brimmell et al., 1999). Using the fusion protein with GFP, we observed in situ expression of Hap50 exclusively in the nuclear compartment of different cell types, even under normal temperature conditions (Fig. 1, upper panels). Interestingly, the distribution of Hap50-GFP is not homogeneous within cell nuclei. The speckled patterns that we observe are strongly reminiscent of recent descriptions of nuclear clusters of RNA polymerase II activity (Cook, 1999; Szentirmay and Sawadogo, 2000). It will thus be interesting to find out whether Hap50 colocalizes with such sites of active transcription or a subset thereof. In addition to a potential bipartite nuclear localization sequence roughly in the middle of the Hap46 sequence (Zeiner and Gehring, 1995), the positively charged region in Hap50 may function as another nuclear localization signal (Packham et al., 1997; Takayama et al., 1998; Brimmell et al., 1999). This is located exactly at the beginning of the N-terminal extension in Hap50. These differences between Hap46 and Hap50 perfectly account for the observations presented here; in particular, Hap50 causing transcriptional activation independent of heat stress, as it already resides in cell nuclei under normal physiological conditions. In nuclear runoff transcription assays, we show for some selected genes that overexpression of Hap50 indeed exerts a positive effect on transcriptional activity rather than stabilizing the respective messages (Fig. 4C). We suppose that most genes activated by Hap50 are subject to multiple and rather subtle regulations. This contention is strengthened by the observation that message levels but not their patterns are changed upon overexpression of Hap50 (Fig. 4A). Thus, the expression of housekeeping genes such as actin (Fig. 4B,C) may not be affected. It is possible that Hap50 exerts its effects by interacting with other transcription factors and/or the basal transcriptional apparatus itself. Furthermore, the effects of Hap50 may be specific for genes transcribed by RNA polymerase II, as we did not observe any changes in the levels of 18S and 28S rRNAs. Protective effects of Hap46/BAG-1 proteins in terms of cell survival under apoptosis-inducing conditions or heat stress were repeatedly observed (Höhfeld, 1998; Kullmann et al., 1998; Yang et al., 1999a; Zeiner et al., 1999; Hayashi et al., 2000). Even though some cooperation with Bcl2 has been described (Takayama et al., 1995; Schulz et al., 1997; EversoleCire et al., 2000), the anti-apoptotic action of Hap46 and Hap50 may well occur independent of Bcl2, as these proteins are not coordinately expressed in various cell lines (Brimmell et al., 1999; Yang et al., 1999a; Yang et al., 1999b). We rather suppose that the beneficial cellular effects of Hap46 under stress or other adverse conditions result from the potential to stimulate transcription (Zeiner et al., 1999). This applies in an even more stringent way to the large isoform Hap50, which is

1843

1.5

1.0

0.5

0.0

Fig. 5. Enhanced viability of heat stressed cells overexpressing Hap50. 3T3 cells (7×105 per 50 mm plate) were transfected with Hap50 cDNA (bars 2, 4) or control vector (bars 1, 3). Some cultures (bars 3, 4) received a heat shock, as detailed in Materials and Methods. Cells excluding Trypan blue were counted as viable and numbers per plate are given. Data show averages of two independent experiments with error bars indicating maximum deviations.

a nuclear protein in unstressed cells and has the ability to activate transcription in such cells (Fig. 3; Fig. 4). By contrast, shorter isoforms, particularly in the range of 33 kDa apparent molecular mass, mainly affect Hsp70/Hsc70-mediated protein folding reactions (Lüders et al., 2000; Nollen et al., 2000), whereas the intermediate form Hap46 exerts pleiotropic effects and shows an overlap in biological activities. In the cytoplasm, Hap46 functions as a regulator of Hsp70 chaperoning activity (Gebauer et al., 1997; Zeiner et al., 1997; Höhfeld, 1998), and upon transfer to the nucleus under stress conditions it can stimulate transcription (Zeiner et al., 1999). We suppose that Hap50, which is a nuclear protein and a transcriptional activator, causes enhanced expression of some critical cellular genes whose products are important for cell viability, for example heat shock proteins (Jolly and Morimoto, 2000). In this context, it is of interest that the messages for the AP-1 components c-Jun and c-Fos are produced at increased levels in Hap50-overexpressing cells (Fig. 4B,C). AP-1 is known as a transcription factor of wide importance for cellular responses to environmental impacts (Curran and Franza, 1988) and, as such, it may be critical for cell viability, differentiation and apoptosis. Thus AP-1 has been implicated in the protection of various mammalian cells against apoptosis, but its activation may also play a role in growth suppression and induction of cell death (Karin et al., 1997; Liebermann et al., 1998). Similarly, upon overexpression, Hap50 may promote the apoptotic response, depending on the specific stimulus and cell type (Yang et al., 2000). Clearly, regulation of apoptosis is rather complex and may depend greatly on the cell system involved and on specific experimental conditions. Stressful conditions under which cells grow in culture or to which solid tumours are exposed in vivo may readily lead to selection of subpopulations that express increased levels of

1844

JOURNAL OF CELL SCIENCE 114 (10)

Hap46/BAG-1 proteins, in particular, the large isoform Hap50. Consequently, cells harbouring relatively large amounts of these proteins are much better equipped to cope with conditions of stress. Overexpression of Hap46/BAG-1 proteins has indeed been observed in various cancer cells (Takayama et al., 1998; Yang et al., 1998; Yang et al., 1999a; Yang et al., 1999b; Brimmell et al., 1999; Shindoh et al., 2000). Such survival-promoting actions may well occur in concert with Hsp70s, which by themselves exert cytoprotective effects on cancer cells, especially under chemotherapeutic treatment (Jäättelä, 1999; Jolly and Morimoto, 2000). The observation that Hap50 is expressed at increased levels in cells upon developing multidrug resistance (Ding et al., 2000) fits closely to the view that Hap50 exerts a rather general cell-protective effect. Even though Hap46/BAG-1 proteins are about to become useful in the clinics as molecular tumour markers, it will be more important in the future to find out how their expression can be downregulated in cancer cells. We thank A. Pater (Memorial University of Newfoundland, St John’s, Newfoundland, Canada) for monoclonal antibody CC9E8, P. Chambon (CNRS, INSERM, Illkirch, France) for human estrogen and glucocorticoid receptor cDNA constructs, and P. Angel and H. Sültmann (German Cancer Research Center, Heidelberg, Germany) for providing cDNAs coding for c-Fos, c-Jun, and actin. We are grateful to M. Grabenbauer (Ruprecht-Karls-Universität Heidelberg, Germany) for help with the confocal laser-scanning microscope. This work was supported by the Deutsche Forschungsgemeinschaft.

REFERENCES Angel, P., Baumann, I., Stein, B., Delius, H., Rahmsdorf, H. J. and Herrlich, P. (1987). 12-O-tetradecanoyl-phorbol-13-acetate induction of the human collagenase gene is mediated by an inducible enhancer element located in the flanking region. Mol. Cell. Biol. 7, 2256-2266. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A. and Struhl, K. (1995). Current Protocols in Molecular Biology. New York: Wiley. Bardelli, A., Longati, P., Albero, D., Goruppi, S., Schneider, C., Ponzetto, C. and Comoglio, P. M. (1996). HGF receptor associates with the anti-apoptotic protein Bag-1 and prevents cell death. EMBO J. 15, 6205-6212. Bimston, D., Song, J., Winchester, D., Takayama, S., Reed, J. C. and Morimoto, R. I. (1998). BAG-1, a negative regulator of Hsp70 chaperone activity, uncouples nucleotide hydrolysis from substrate release. EMBO J. 17, 6871-6878. Brimmell, M., Burns, J. S., Munson, P., McDonald, L., O’Hare, M. J., Lakhani, S. R. and Packham, G. (1999). High level expression of differentially localized BAG-1 isoforms in some estrogen receptor-positive human breast cancers. Br. J. Cancer 81, 1042-1051. Cook, P. R. (1999). The organization of replication and transcription. Science 284, 1790-1795. Crocoll, A., Blum, M. and Cato, A. C. (2000). Isoform-specific expression of BAG-1 in mouse development. Mech. Dev. 91, 355-359. Curran, T. and Franza, B. R. (1988). Fos and Jun: The AP-1 connection. Cell 55, 395-397. Devireddy, L. R., Kumar, K. U., Pater, M. M. and Pater, A. (2000). BAG1, a novel Bcl-2-interacting protein, activates expression of human JC virus. J. Gen. Virol. 81, 351-357. Ding, Z., Yang, X., Pater, A. and Tang, S. C. (2000). Resistance to apoptosis is correlated with the reduced caspase-3 activation and enhanced expression of antiapoptotic proteins in human cervical multidrug-resistant cells. Biochem. Biophys Res. Commun. 270, 415-420. Eversole-Cire, P., Concepcion, F. A., Simon, M. I., Takayama, S., Reed, J. C. and Chen, J. (2000). Synergistic effect of Bcl-2 and BAG-1 on the prevention of photoreceptor cell death. Invest. Ophthalmol. Vis. Sci. 41, 1953-1961.

Froesch, B. A., Takayama, S. and Reed, J. C. (1998). BAG-1L protein enhances androgen receptor function. J. Biol. Chem. 273, 11660-11666. Gebauer, M., Zeiner, M. and Gehring, U. (1997). Proteins interacting with the molecular chaperone Hsp70/Hsc70: physical associations and effects on refolding activity. FEBS Lett. 417, 109-113. Gebauer, M., Zeiner, M. and Gehring, U. (1998). Interference between proteins Hap46 and Hop/p60 which bind to different domains of the molecular chaperone Hsp70/Hsc70. Mol. Cell. Biol. 18, 6238-6244. Guzey, M., Takayama, S. and Reed, J. C. (2000). BAG1L enhances transactivation function of the vitamin D receptor. J. Biol. Chem. 275, 4074940756. Hayashi, T., Sakai, K.-i., Sasaki, C., Yasuto, I. and Abe, K. (2000). Loss of bag-1 immunoreactivity in rat brain after transient middle cerebral artery occlusion. Brain Res. 852, 496-500. Höhfeld, J. (1998). Regulation of the heat shock cognate Hsc70 in the mammalian cell: The characterization of the anti-apoptotic protein BAG-1 provides novel insights. Biol. Chem. 379, 269-274. Jäättelä, M. (1999). Escaping cell death: Survival proteins in cancer. Exp. Cell Res. 248, 30-43. Jolly, C. and Morimoto, R. I. (2000) Role of the heat shock response and molecular chaperones in oncogenesis and cell death. J. Natl. Cancer Inst. 92, 1564-1572. Kanelakis, K. C., Morishima, Y, Dittmar, K. D., Galigniana, M. D., Takayama, S, Reed, J. C. and Pratt, W. B. (1999). Differential effects of the Hsp70-binding protein BAG-1 on glucocorticoid receptor folding by the hsp90-based chaperone machinery. J. Biol. Chem. 274, 34134-34140. Karin, M., Liu, Z. and Zandi, E. (1997) AP-1 function and regulation. Curr. Opin. Cell Biol. 9, 240-246. Kullmann, M., Schneikert, J., Moll, J., Heck, S., Zeiner, M., Gehring, U. and Cato, A. C. B. (1998). RAP46 is a negative regulator of glucocorticoid receptor action and hormone induced apoptosis. J. Biol. Chem. 273, 1462014625. Liebermann, D. A., Gregory, B. and Hoffman, B. (1998). AP-1 (Fos/Jun) transcription factors in hematopoietic differentiation and apoptosis. Int. J. Oncol. 12, 685-700. Liu, R., Takayama, S., Zheng, Y., Froesch, B., Chen, G., Zhang, X., Reed, J. C. and Zhang, X.-k. (1998). Interaction of BAG-1 with retinoic acid receptor and its inhibition of retinoic acid-induced apoptosis in cancer cells. J. Biol. Chem. 273, 16985-16992. Lüders, J., Demand, J., Papp, O. and Höhfeld, J. (2000). Distinct isoforms of the cofactor BAG-1 differentially affect Hsc70 chaperone function. J. Biol. Chem. 275, 14817-14823. Nollen, E. A. A., Brunsting, J. F., Song, J., Kampinga, H. H. and Morimoto, R. I. (2000). Bag-1 functions in vivo as a negative regulator of Hsp70 chaperone activity. Mol. Cell. Biol. 20, 1083-1088. Packham, G., Brimmel, M. and Cleveland, J. L. (1997). Mammalian cells express two differently localized Bag-1 isoforms generated by alternative translation initiation. Biochem. J. 328, 807-813. Petersen, G., Hahn, C. and Gehring, U. (2001). Dissection of the ATP binding domain of the molecular chaperone Hsc70 for interaction with the cofactor Hap46. J. Biol. Chem. 276, 10178-10184. Schneikert, J., Hübner, S., Langer, G., Petri, T., Jäättelä, M., Reed, J and Cato, A. C. B. (2000). Hsp70-RAP46 interaction in downregulation of DNA binding by glucocorticoid receptor. EMBO J. 19, 6508-6516. Schulz, J. B., Bremen, D., Reed, J. C., Lommatzsch, J., Takayama, S., Wullner, U., Loschmann, P. A., Klockgether, T. and Weller, M. (1997). Cooperative interception of neuronal apoptosis by BCL-2 and BAG-1 expression: prevention of caspase activation and reduction of reactive oxygen species. J. Neurochem. 69, 2075-2086. Shindoh, M., Adachi, M., Higashino, F., Yasuda, M., Hida, K., Nishioka, T., Ono, M., Takayama, S., Reed, J. C., Imai, K. et al. (2000). BAG-1 expression correlates highly with the malignant potential in early lesions (T1 and T2) of oral squamous cell carcinoma. Oral Oncol. 36, 444-449. Szentirmay, M. N. and Sawadogo, M. (2000). Spacial organization of RNA polymerase II transcription in the nucleus. Nucleic Acids Res. 28, 20192025. Takayama, S., Sato, T., Krajewski, S., Kochel, K., Irie, S., Millan, J. A. and Reed, J. C. (1995). Cloning and functional analysis of BAG-1: A novel Bcl-2-binding protein with anti-cell death activity. Cell 80, 279-284. Takayama, S., Krajewski, S., Krajewska, M., Kitada, S., Zapata, J. M., Kochel, K., Knee, D., Scudiero, D., Tudor, G., Miller, G. J. et al. (1998). Expression and location of Hsp70/Hsc-binding anti-apoptotic protein BAG1 and its variants in normal tissues and tumor cell lines. Cancer Res. 58, 3116-3131.

Transcriptional activation by the nuclear protein Hap50 Takayama, S., Xie, Z. and Reed, J. C. (1999). An evolutionarily conserved family of Hsp70/Hsc70 molecular chaperone regulators. J. Biol. Chem. 274, 781-786. Velazquez, J. M. and Lindquist, S. (1984). Hsp70: nuclear concentration during environmental stress and cytoplasmatic storage during recovery. Cell 36, 655-662. Yang, X., Chernenko, G., Hao, Y., Ding, Z., Pater, M. M., Pater, A. and Tang, S.-C. (1998). Human BAG-1/RAP46 protein is generated as four isoforms by alternative translation initiation and overexpressed in cancer cells. Oncogene 17, 981-989. Yang, X., Hao, Y., Ferenczy, A., Tang, S.-C. and Pater, A. (1999a). Overexpression of anti-apoptotic gene BAG-1 in human cervical cancer. Exp. Cell Res. 247, 200-207. Yang, X., Hao, Y., Ding, Z., Pater, A. and Tang, S.-C. (1999b). Differential

1845

expression of antiapoptotic gene BAG-1 in human breast normal and cancer cell lines and tissues. Clin. Cancer Res. 5, 1816-1822. Yang, X., Hao, Y., Ding, Z. and Pater, A. (2000). BAG-1 promotes apoptosis induced by N-(4-hydroxyphenyl)retinamide in human cervical carcinoma cells. Exp. Cell Res. 256, 491-499. Zeiner, M. and Gehring, U. (1995). A protein that interacts with members of the nuclear hormone receptor family: Identification and cDNA cloning. Proc. Natl. Acad. Sci. USA 92, 11465-11469. Zeiner, M., Gebauer, M. and Gehring, U. (1997). Mammalian protein RAP46: an interaction partner and modulator of 70 kDa heat shock proteins. EMBO J. 16, 5483-5490. Zeiner, M., Niyaz, Y. and Gehring, U. (1999). The Hsp70-associating protein Hap46 binds to DNA and stimulates transcription. Proc. Natl. Acad. Sci. USA 96, 10194-10199.