Interaction of BAG-1 with retinoic acid receptor and its inhibition of retinoic acid-induced apoptosis in cancer cells.

BAG-1 (also known as RAP46) is an anti-apoptotic protein, which has been shown previously to interact with a number of nuclear hormone receptors, including receptors for glucocorticoid, estrogen, and thyroid hormone. We show here that BAG-1 also interacts with retinoic acid receptor (RAR). Gel retardation assays demonstrated that in vitro translated BAG-1 protein could effectively inhibit the binding of RAR but not retinoid X receptor (RXR) to a number of retinoic acid (RA) response elements (RAREs). A glutathione S-transferase-BAG-1 fusion protein also specifically bound RAR but not RXR. Interaction of BAG-1 and RAR could also be demonstrated by yeast two-hybrid assays. In transient transfection assays, co-transfection of BAG-1 expression plasmid inhibited the transactivation activity of RAR/RXR heterodimers but not RXR/RXR homodimers. When stably expressed in breast cancer cell lines, BAG-1 inhibited binding of RAR/RXR heterodimer to a number of RAREs and suppressed RA-induced growth inhibition and apoptosis. In addition, RA-induced suppression of Bcl-2 expression was abrogated by overexpression of BAG-1. These results demonstrate that BAG-1 can regulate retinoid activities through its interaction with RAR and suggest that elevated levels of BAG-1 protein could potentially contribute to retinoid resistance in cancer cells.

Development of a multicellular organism requires tightly regulated cellular processes, such as proliferation, differentiation, and cell death. Failure to maintain the balance among these fundamental and mechanistically related processes may result in abnormal cell growth, as seen in cancer cells where cell death is often inhibited (1,2). Retinoids, a group of natural and synthetic vitamin A derivatives, are currently used to treat epithelial cancer and promyelocytic leukemia and are being evaluated for prevention and therapy of other human cancers (3,4). The anti-cancer effects of retinoids are mainly due to their inhibition of cell proliferation, induction of cell differentiation, and promotion of apoptosis. Retinoids alone or in combination with other stimuli induce apoptosis during normal development and in different types of cancer cells in vitro (5)(6)(7)(8)(9)(10)(11). However, it remains largely unknown how retinoidinduced apoptosis is regulated.
The effects of retinoids are mainly mediated by two classes of nuclear receptors, the retinoic acid receptors (RARs) 1 and retinoid X receptors (RXRs). RARs and RXRs are encoded by three distinct genes (␣, ␤, and ␥) and are members of the steroid/ thyroid/retinoid hormone receptor superfamily that function as ligand-activated transcription factors (12)(13)(14). 9-cis RA is a high affinity ligand for both RARs and RXRs, whereas alltrans-RA (trans-RA) is a ligand for only RARs. RARs and RXRs primarily function as RXR/RAR heterodimers that bind to a variety of RA response elements (RAREs) and regulate their transactivation activities.
The involvement of retinoid receptors in retinoid-induced apoptosis has been demonstrated in several studies. Expression of RAR␤ may be involved in the apoptosis of mesenchyme of the interdigital regions during mouse limb development (28). RAR␤ is required for RA-induced apoptosis of breast cancer (5) and lung cancer (7) cells, whereas activation of RXR is essential for RA-induced HL-60 cell apoptosis (29). In 4-HPR-induced apoptosis, activation of RAR␥ may be involved (30,31), whereas regulation of activation-induced apoptosis of T-cells by 9-cis RA requires activation of both RARs and RXRs (32).
Although much interest has been directed to the role of retinoid-induced apoptosis in both physiological and pathological processes, very little is known regarding regulation of the process. It is believed that apoptosis, once triggered, proceeds through a central death pathway in which specific cellular proteases and endonucleases are activated (1,2,33). Members of the Bcl-2 family play an important role in the regulation of the central death pathway. Bcl-2 can suppress induction of apoptosis in many systems, whereas Bax promotes apoptosis. In addition, several other proteins that modulate Bcl-2 activity by interacting with Bcl-2 have been described (1,2,33). One of these genes, BAG-1 (for Bcl-2-associated anti-death គ gene 1), was cloned from a murine embryo cDNA library using a pro-tein-protein interaction technique (34). Two differently localized BAG-1 isoforms, the long BAG-1 isoform and the short BAG-1 isoform, generated by alternative translation initiation are expressed in mammalian cells (35). Whether two isoforms act differently remains to be determined. Recent studies demonstrated that co-expression of BAG-1 and Bcl-2 in Jurkat lymphoid cells, NIH 3T3 fibroblasts, and melanoma cells promoted the survival of these cells in response to a variety of apoptotic stimuli (34,36,37). In addition to Bcl-2, BAG-1 also interacts with Raf-1 (38), resulting in activation of its kinase activity. Furthermore, BAG-1 can interact with hepatocyte growth factor receptor and platelet-derived growth factor receptor and enhance the ability of these receptors to transduce signals for cell survival (39). These observations suggest that BAG-1 may function as an adaptor to mediate the interaction between survival factors and apoptotic machinery and may also play a role in regulating cellular proliferation. The recent observation that BAG-1 binds tightly to Hsp70/Hsc70-family proteins and modulates their chaperone activity (40 -42) suggests that the ability of BAG-1 to alter the activities of diverse groups of proteins involved in cell growth control may be attributed to its effects on Hsp70/Hsc70 proteins.
Interestingly, the human BAG-1 homolog (also known as RAP46) was cloned from a human liver cDNA library by virtue of its interaction with the glucocorticoid receptor (43). In vitro, RAP46 interacts with a number of nuclear hormone receptors, including estrogen receptor and thyroid hormone receptor (TR) (43). Since molecular chaperones are known to play an important role in controlling the activity of many members of the steroid/thyroid/retinoid receptor family (44), it is possible that BAG-1 could alter the function of these transcriptional regulators. Before this report, however, it was unknown whether BAG-1 regulates the activities of the nuclear hormone receptors and whether BAG-1 interacts with retinoid receptors.
Here we demonstrate that short BAG-1 isoform interacts with the RAR but not the RXR both in vitro and in vivo. GST pull-down and the yeast two-hybrid assays show that BAG-1 directly interacts with RAR but not RXR. Moreover, BAG-1 inhibits RAR/RXR heterodimer DNA binding and suppresses RA-induced transactivation activity of RARs on various RAREs. Overexpression of BAG-1 in MCF-7 and ZR-75-1 breast cancer cells reduces the ability of trans-RA to inhibit the growth and induce apoptosis, as well as its modulation of Bcl-2 expression. Taken together, our results demonstrate that BAG-1 can physically interact with RARs and is an important component in the retinoid response pathway. Our findings suggest that this protein-protein interaction may play a role in the regulation of retinoid-induced growth inhibition and apoptotic processes, potentially contributing to retinoid resistance in cancer.

MATERIALS AND METHODS
Cell Culture-Monkey kidney CV-1 cells and breast cancer MCF-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, and ZR-75-1 breast cancer cells were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum.
Growth Inhibition Assay-Cells were seeded at 1,000 -2,000 cells/ well in 96-well plates and treated 24 h later with various concentrations of trans-RA for 7 days. Media and trans-RA were changed every 48 h. Relative viable cell number was determined using the MTT assay (52).
Apoptosis Analysis-For the terminal deoxynucleotidyl transferase assay (5), cells were treated with or without 10 Ϫ6 M trans-RA. After 48 h, cells were trypsinized, washed with phosphate-buffered saline, fixed in 1% formaldehyde in phosphate-buffered saline, washed with phosphate-buffered saline, resuspended in 70% ice-cold ethanol, and stored at Ϫ20°C overnight. Cells were then labeled with biotin-16-dUTP by terminal transferase and stained with avidin-fluorescein isothiocyanate (Boehringer Mannheim). Fluorescently labeled cells were analyzed using a FACScater-Plus. Representative histograms are shown.
Antibodies and Immunoblotting-Cells were lysed in 150 mM NaCl, 10 mM Tris, pH7.4, 5 mM EDTA, 1% Triton X-100 and protease inhibitors phenylmethylsulfonyl fluoride, aprotinin, leupeptins, and pepstatin. Equal amounts of lysates (50 g) were boiled in SDS sample buffer, resolved by SDS-polyacrylamide gel electrophoresis, and transferred onto nitrocellulose membrane. After transfer, the membranes were blocked in TBST (50 mM Tris, pH7.5, 150 mM NaCl, 0.1% Tween 20) containing rabbit anti-Bcl-2 serum. The membranes were then washed FIG. 1. Inhibition of RAR DNA binding by BAG-1. a, inhibition of RAR/RXR heterodimer binding by BAG-1. In vitro synthesized RAR␣ and RXR␣ were preincubated with the indicated molar excess of in vitro synthesized BAG-1 or estrogen receptor (ER). Unprogrammed reticulocyte lysate was used to maintain an equal protein concentration in each reaction. After this preincubation, the reaction mixtures were incubated with 32 P-labeled TREpal and analyzed by the gel retardation assay. The open arrow indicates nonspecific binding. b, inhibition of TR␣/RXR␣ binding by BAG-1. In vitro synthesized TR␣ and RXR␣ were preincubated with the indicated molar excess amount of in vitro synthesized BAG-1 and analyzed by the gel retardation assay using the TREpal as a probe. c, inhibition of RAR␣/RXR␣ binding on the ␤RARE by BAG-1. In vitro synthesized RAR␣ and RXR␣ were preincubated with the indicated molar excess amount of in vitro synthesized BAG-1 and analyzed by the gel retardation assay using the ␤ RARE as a probe. d, inhibition of DNA binding of bacterially expressed RAR but not RXR by BAG-1. Bacterially expressed RAR␥ or RXR␣ protein was preincubated with 6 l of in vitro synthesized BAG-1 and analyzed by the gel retardation assay using the TREpal as a probe. three times with TBST and then incubated for 1 h at room temperature in TBST containing horseradish peroxidase-linked anti-rabbit immunoglobulin. After three washes in TBST, immunoreactive products were detected by chemiluminescence with an enhanced chemiluminescence system (ECL, Amersham Pharmacia Biotech).
Preparation of Receptor, BAG-1, and Nuclear Protein-cDNAs for RAR␣, RXR␣, estrogen receptor, and BAG-1, which expresses short BAG-1 isoform cloned into pBluescript (Stratagene), were transcribed by using T 7 or T 3 RNA polymerase, and the transcripts were translated in the rabbit reticulocyte lysate system (Promega) as described previously (45). The relative amounts of the translated proteins were determined by separating the [ 35 S]methionine-labeled proteins on SDS-polyacrylamide gels, quantitating the amount of incorporated radioactivity and normalizing it relative to the content of methionine residues in each protein. To synthesize receptor fusion protein, RAR␥ or RXR␣ cDNAs were cloned in-frame into the bacterial expression vector pGex.2T (Amersham) as described (45). Preparation and purification of GST-BAG-1 and GST-BAG-1(⌬172-218) fusion proteins has been described (41). Preparation of nuclear extract was described previously (6).
Gel Retardation Assay-Analysis of in vitro synthesized or bacterially expressed receptor proteins or nuclear proteins by gel retardation was described previously (45). To analyze the effect of BAG-1 protein, in vitro synthesized BAG-1 protein was preincubated with receptor protein at room temperature for 10 min before the gel retardation assay. The oligonucleotides used for gel retardation assays were ␤RARE (TG-TAGGGTTCACCGAAAGTTCAGTC) (46); TREpal (TGAGGTCATGAC-CTGA) (45); DR-5-RARE(TGTAGGGTTCACACTGAGTTCACTCA); and DR-2-RARE(AGGTCAAAAGGTCAG).
GST Pull-down Assay-To analyze the interaction between BAG-1 and RAR, GST-BAG-1 fusion protein was immobilized on glutathione-Sepharose beads as described (52). As a control, GST prepared under the same conditions was also immobilized. The beads were preincubated with bovine serum albumin (1 mg/ml) at room temperature for 5 min. 35 S-Labeled in vitro synthesized receptor proteins (2 to 5 l, depending on translation efficiency) Bcl-2 or Hsc70 were then added to the beads. The beads were then continuously rocked for 1 h at 4°C in a final volume of 200 l in EBC buffer (140 mM NaCl, 0.5% Nonidet P-40, 100 mM NaF, 200 M sodium orthovanadate, and 50 mM Tris, pH 8.0). After washing five times with NETN buffer (100 mM NaCl, 1 mM EDTA, 20 mM Tris, pH 8.0, 0.5% Nonidet P-40), the bound proteins were analyzed by SDS-polyacrylamide gel electrophoresis.
Two-hybrid Assay-For the yeast two-hybrid assay, the yeast twohybrid system from CLONTECH Inc. (Palo Alto, CA) was used (52). BAG-1 cDNA was cloned into the yeast expression vector pGAD424 to generate an in-frame fusion with the Gal4 activation domain. RAR␥ or RXR␣ cDNAs were cloned into pGBT11 or pGBT9, respectively, to produce an in-frame fusion with Gal4 DNA binding domain. RAR␥ was also cloned into pGAD426 that contains Gal4 activation domain to study the interaction between RAR␥ and RXR␣. The yeast reporter strain Y190 containing a LacZ reporter plasmid with Gal4 binding sites was used for transformation. ␤-Galactosidase activity was determined following the conditions provided by the manufacturer.

Inhibition of Retinoid Receptor DNA Binding by BAG-1-We
investigated whether BAG-1 could interact with retinoid receptors by studying the effect of BAG-1 protein on binding of retinoid receptors to their target DNA sequences. In vitro synthesized RAR and RXR formed a strong RAR/RXR heterodimer complex with the TREpal as described previously (45). When increasing amounts of in vitro synthesized short BAG-1 isoform protein were incubated with RAR and RXR, the binding of RAR/RXR heterodimers was inhibited in a BAG-1 concentrationdependent manner (Fig. 1a). At a 5 M excess of BAG-1 protein relative to RAR/RXR, the binding was almost completely inhibited. The effect of BAG-1 on RAR/RXR binding was specific because similar amounts of estrogen receptor did not show any effect. BAG-1 also effectively inhibited the binding of TR/RXR to the TREpal probe (Fig. 1b), consistent with a prior report that BAG-1 can interact with TR (43). To study whether the inhibitory effect of BAG-1 on RAR/RXR heterodimer binding is specific to the TREpal, we used another RA responsive element (␤RARE), which is derived from the RAR␤ promoter (46). As   FIG. 2. Analysis of RAR-BAG-1 interaction by the GST pulldown assay. BAG-1 or a BAG-1 C-terminal deletion mutant (BAG-1/ ⌬c) was expressed in bacteria using the pGex.4T expression vector. The GST-BAG-1 proteins were immobilized on glutathione-Sepharose beads. As a control, the same amount of GST was also immobilized. 35 S-Labeled RAR␣, RXR␣, or Bcl-2 was then mixed with the beads. After extensive washing, the bound proteins were analyzed by SDSpolyacrylamide gel electrophoresis. The input proteins are shown for comparison (left panel). For comparison, binding of GST-BAG-1 and GST-BAG-1/⌬c to Hsc70 was shown in right panel.
We next determined whether inhibition of RAR/RXR heterodimer binding by BAG-1 is due to interaction of BAG-1 with RAR or RXR. Since in vitro synthesized RAR or RXR alone does not bind efficiently to RARE, we used bacterially expressed RAR or RXR protein. When a 5-fold molar excess of BAG-1 protein was added, the binding of RAR␥ was significantly inhibited, whereas binding of RXR␣ was not affected (Fig. 1d). These results suggest that inhibition of RAR/RXR or TR/RXR heterodimer DNA binding by BAG-1 is likely due to its interaction with RAR or TR but not with RXR.
Interaction of BAG-1 with Retinoid Receptors-To further study the interaction between BAG-1 and RAR, we used an in vitro GST pull-down assay. A GST-BAG-1 fusion protein was expressed in bacteria and immobilized on glutathione-Sepharose beads. The beads were then incubated with in vitro synthesized 35 S-labeled RAR or RXR protein. After extensive washing, the mixtures were analyzed on a SDS-polyacrylamide gel. In comparison to the input lane, significant amounts of labeled RAR but not RXR were retained by GST-BAG-1-Sepharose beads but not by GST control beads (Fig. 2). For control, Bcl-2, a known BAG-1 interacting protein (34), bound strongly to GST-BAG-1 beads. We also employed a BAG-1 mutant protein in which the last 47 amino acid residues are deleted from its C-terminal end (41). This mutant protein (BAG-1/⌬c) can interact with Bcl-2 but not with Hsc70. Interestingly, labeled RAR but not RXR was also retained by the mutant BAG-1, thus implying that the interaction of BAG-1 with RAR is independent of its binding to Hsc70.
Interaction between RAR and BAG-1 was also evaluated by the two-hybrid assay in yeast. Fig. 3 shows that co-transformation of BAG-1 and RAR␥ significantly activated the reporter in ␤-gal filter assay, whereas co-transformation of BAG-1 and RXR␣ did not. Interaction between BAG-1 and RAR␥ was specific because co-transformation of either BAG-1 or RAR␥ with the corresponding empty vector did not activate the reporter gene. Thus, RAR and BAG-1 also interact in intact cells.
Inhibition of Transactivation Activity of Retinoid Receptors by BAG-1-To further examine the BAG-1-RAR interaction, we studied the effects of the short BAG-1 isoform on RAR transactivation activity on a number of RAREs by transient transfection assay. When CV-1 cells were transiently transfected with RAR␣ expression vector together with either TREpal-tk-CAT (Fig. 4a) or ␤RARE-tk-CAT (Fig. 4b), trans-RA-induced reporter gene activity was markedly inhibited by co-transfection of BAG-1 expression plasmid in a concentration-dependent manner. The effect is specific to BAG-1 because co-transfection of similar amounts of empty expression vector (pcDNA3) did not inhibit trans-RA-induced gene expression. BAG-1 also showed inhibitory effect on trans-RA-induced RAR␤ activity on the TREpal. However, the trans-RA-independent RAR␤ activity was not affected. This suggests that ligand-dependent RAR activity is more sensitive to the inhibitory effect of BAG-1. A similar inhibitory effect of BAG-1 was also obtained when RAR␥ expression vector was used (data not shown). We also studied the effects of BAG-1 on thyroid hormone (T 3 )-induced TR␣ activity, and we observed a significant inhibitory effect of BAG-1 on TR␣ (Fig. 4d), consistent with the ability of BAG-1 to bind the TR protein ( Fig. 1b; Ref. 43). To determine the effect of BAG-1 on RXR homodimer activity, we co-transfected TREpaltk-CAT reporter plasmid and RXR␣ expression vector. How-ever, the 9-cis-RA-induced RXR␣ homodimer activity was not affected by co-transfection of BAG-1 (Fig. 4e), consistent with our observation that BAG-1 does not interact with RXR␣ in vitro (Figs. 1d, 2, and 3).
Overexpression of BAG-1 Inhibits Trans-RA-induced Cancer Cell Apoptosis-The above data suggest that BAG-1 may function as a modulator of trans-RA-induced biological responses. We previously showed that trans-RA effectively inhibits the growth and induces apoptosis of some human breast cancer cell lines (5). We, therefore, stably expressed BAG-1 in human breast cancer cell lines MCF-7 and ZR-75-1. MCF-7/BAG-1(#3), 75-1/BAG-1(#5), and 75-1/BAG-1(#6) that stably expressed high levels of the transfected BAG-1 plasmid (data not shown) were chosen to examine the effect of BAG-1 overexpression on RA activities. As shown in Fig. 5a, trans-RA effectively inhibited the growth of parental MCF-7 and ZR-75-1 cells. However, the BAG-1-overexpressing clones displayed resistance to the growth inhibitory effects of trans-RA. The growth of the MCF-7/BAG-1(#3) was even stimulated by trans-RA (Fig. 5a). Although trans-RA did not stimulate the growth of 75-1/BAG-1(#5) and 75-1/BAG-1(#6) cells, its inhibitory effect on the growth of these cells was significantly reduced as compared with its effect on ZR-75-1 cells (Fig. 5b). The effect on trans-RA activity observed above was specific because clones stably transfected with the empty vector, MCF-7/neo and 75-1/neo, exhibited similar responses to trans-RA as that observed with the parental cell lines. Thus, BAG-1 partially abrogates the growth inhibitory effects of trans-RA on human breast cancer cells.
Overexpression of BAG-1 Abrogates Down-regulation of Bcl-2 by Trans-RA-Retinoids have been shown to down-regulate the expression of Bcl-2 in leukemia (11). We therefore studied whether BAG-1 affected RA-regulated expression of Bcl-2 in MCF-7 cells. Bcl-2 was highly expressed in MCF-7 cells, and its expression level was significantly reduced by trans-RA as de- termined by immunoblotting. In MCF-7/BAG-1(#3)-stable transfectant, however, treatment of trans-RA had little or no effect on Bcl-2 expression (Fig. 7). These data suggest that BAG-1 may inhibit trans-RA-induced apoptosis at least in part through its effects on trans-RA-regulated genes, such as Bcl-2.
To further study the effect of BAG-1 on RAR-mediated gene regulation, we prepared nuclear extract from MCF-7, MCF-7/ neo, and MCF-7/BAG-1(#3) cells and analyzed their binding to various RAREs, including ␤RARE, DR-5-type RARE, and DR-2-type RARE. As shown in Fig. 8a, extracts from MCF-7 and MCF-7/neo showed strong binding complex to these RAREs, whereas binding of the slow-migrating complex observed in MCF-7 and MCF-7/neo cells was strongly inhibited in MCF-7/ BAG-1(#3) cells. To determine the nature of the slow-migrating complex, extract from MCF-7 cells was incubated with either anti-RAR (␣-RAR) or anti-RXR (␣-RXR) antibody or nonspecific preimmune serum prior to DNA binding reaction. Fig. 8b shows that either anti-RAR or anti-RXR antibody, but not nonspecific serum, completely inhibited the formation of the complex, suggesting that the complex contains RAR and RXR. These data further demonstrate that overexpression of BAG-1 inhibits RAR/RXR binding and suggest that the alteration of RAR transcriptional activity may contribute to the effect of BAG-1 on RA responses. DISCUSSION Previously, it was reported that BAG-1 (RAP46) can interact with several steroid hormone receptors (43). However, the biological effects of the interaction are unknown. In this report, we show that the short BAG-1 isoform that is known to be predominantly expressed in cytoplasm (35) can antagonize RAR activity through its direct interaction with RAR. Using gel retardation assays, we observed that BAG-1 inhibits binding of RAR/RXR heterodimers to several RAREs (Fig. 1). In GST-pull down assay, we found that BAG-1 directly interacts with RAR in solution (Fig. 2). By using yeast two-hybrid assays (Fig. 3), we showed that the BAG-1-RAR interaction could occur in vivo. Moreover, a functional interaction was demonstrated by our observation that co-transfection of BAG-1 inhibits trans-RAinduced RAR transactivation activities on several RAREs (Fig.  4). Furthermore, RAR/RXR RARE binding was abrogated in MCF-7 cells that express transfected BAG-1 (Fig. 8). Thus, the interaction of BAG-1 with steroid hormone receptors can be extended to RAR. We also present evidence here that BAG-1 can similarly prevent TR DNA binding and transactivation activity. However, BAG-1 does not interact with all nuclear hormone receptors, as shown here by the failure of BAG-1 to bind to and modulate the activity of RXR (Figs. 1d, 2, 3, and 4e).
Activation or repression of gene transcription by nuclear hormone receptors requires their interaction with multiple cellular co-regulatory factors. These include receptor co-activators, which exert their effect on receptor transactivation activ-ity by mediating transcription-initiation complex formation and affecting chromatin structure (15), and receptor co-repressors, which bind to receptors in the absence of ligand and actively repress target gene transcription by impairing the activity of the basal transcription machinery (25). BAG-1 appears to function differently from receptor co-activators or corepressors. It does not induce ligand-dependent gene activation nor does it cause repression of target gene transcription in the absence of ligand (Fig. 4a). Instead, interaction of RAR and BAG-1 resulted in inhibition of RAR DNA binding (Fig. 1) and trans-RA-induced RAR transactivation activities (Fig. 4). Thus, it may function as a modulator of RAR activities through the mechanisms that resemble the effect of AP-1, which was previously shown to inhibit trans-RA-induced RAR activity by preventing RAR binding to target DNA sequences in promoters (27). However, whether BAG-1 affects recruitment of co-activator or co-repressor by RAR remains to be determined.
The mechanism by which BAG-1 inhibits RAR binding to DNA and transactivation activity remains to be elucidated. Recently, BAG-1 was reported to bind tightly with Hsp70/ Hsc70-family proteins and modulate their activity (40 -42). Although the role of molecular chaperones in transcriptional activation by retinoid receptors remains controversial (49,50), Hsp70/Hsc70 and other heat shock proteins are known to participate in the regulation of several other steroid hormone receptors (44). It is therefore tempting to speculate that BAG-1 may also influence RAR activity through Hsp70/Hsc70-mediated conformational changes that prevent it from binding DNA and transactivating retinoid-responsive target genes. Interestingly, a deletion mutant of BAG-1 lacking its C-terminal 47 amino acids, which does not bind to Hsc70, was capable of a, an equal amount (3 g) of nuclear extracts prepared from MCF-7, MCF-7/neo, and MCF-7/BAG-1(#3) was analyzed by gel retardation assay using ␤RARE, DR-5-type RARE (DR-5) or DR-2-type RARE (DR-2) as a probe. b, effects of anti-RXR (␣RXR) and anti-RAR (␣RAR) antibodies on the binding of the slow-migrating binding complex. Nuclear extract (3 g) from MCF-7 cells was incubated with antibody (Ab) for 30 min at room temperature before performance of the gel retardation assay using the ␤RARE as a probe. The arrow indicates the slow-migrating complex, which had a binding that was inhibited by either ␣RXR or ␣RAR but not by nonspecific preimmune serum (NI). binding to RAR in vitro (Fig. 2). Thus, the domains in BAG-1 required for interactions with RAR and Hsp70 appear to be separable. Unfortunately, when expressed in mammalian cells, the BAG-1 mutant protein was unstable, precluding functional evaluation of its effects on RAR-mediated gene expression. Thus, the relevance of BAG-1 interactions with Hsp70-family proteins to its function as inhibitor of RAR remains to be determined.
One of the interesting features of BAG-1 is its ability to promote cell survival (34,36,37). The effect was previously attributed to its interaction with Bcl-2 (34). In addition, BAG-1 interacts with platelet-derived growth factor and hepatocyte growth factor receptors, enhancing their ability to transduce signals that promote cell survival (39). BAG-1 can also bind to Raf-1 and activate its kinase activity (38). Our observations that overexpression of BAG-1 inhibits trans-RA-induced apoptosis (Fig. 6) and prevents trans-RA-induced down-regulation of Bcl-2 expression (Fig. 7) suggest that interaction with RAR may represent another mechanism by which BAG-1 promotes cell survival. These observations also suggest that overexpression of BAG-1 may contribute to retinoid resistance in certain malignancies. Taken together with recent observations that BAG-1 protein levels are elevated in breast and prostate cancers (51), BAG-1 may represent an important regulator of cell survival and growth, which may contribute in multiple ways to tumorigenesis and resistance to therapy.