Activation of BAG3 by Egr-1 in Response to FGF-2 in Neuroblastoma Cells

Oncogene (2008) 27, 5011–5018 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Activation of BAG3 by Egr-1 in response to FGF-2 in neuroblastoma cells

A Gentilella1, G Passiatore1, S Deshmane1, MC Turco2 and K Khalili1

1Department of Neuroscience, Center for Neurovirology, Temple University School of Medicine, Philadelphia, PA, USA and 2Department of Pharmaceutical Sciences, University of Salerno, Salerno, Italy

The co-chaperone protein, BAG3, which belongs to the involved in neuronal protection and repair after BAG protein family, has an established antiapoptotic ischemic or traumatic brain injury; it prevents apoptosis function in different tumor cell lines. Here we demon- by activating antiapoptotic pathways and promotes strated that treatment of the human neuroblastoma cell neurogenesis in the adult hippocampus after injury line, SK-N-MC, with fibroblast growth factor-2 (FGF-2) (Alzheimer and Werner, 2002). results in induction of BAG3 expression. Induction of Egr-1 is a zinc-finger transcription factor that BAG3 protein by FGF-2 occurs at the transcriptional responds rapidly to a variety of stimuli and functions level; it requires the extracellular regulated kinase1/2 as a convergence point for many signaling cascades and pathway and is dependent on the activity of Egr-1 upon couples short-term extracellular signals to longer-term the BAG3 promoter. Targeted suppression of BAG3 by changes in the expression of target genes (reviewed by small-interfering RNA results in dysregulation of cell- Thiel and Cibelli, 2002). It has been reported that Egr-1 cycle progression most notably at S and G2 phases, which is closely linked to FGF-2 signaling and has been corroborates the decreased level of cyclin B1 expression. implicated in orchestrating changes in gene expression These observations suggest a new role for BAG3 in that underlie neuronal plasticity and brain development regulation of the cell cycle. (O’Donovan et al., 1999; Midgley and Khachigian, Oncogene (2008) 27, 5011–5018; doi:10.1038/onc.2008.142; 2004). Here we report the involvement of BAG3 in these published online 12 May 2008 events. The human BAG family of proteins comprises six Keywords: BAG3; Egr-1; neuron/neuronal family members (BAG1–6) that are thought to function as molecular chaperone regulators (Doong et al., 2002). Recent evidence has demonstrated that a member of the BAG protein family, BAG1, has an essential role in differentiation and survival of neuronal cells in vivo (Gotz et al., 2005) and mediates neuroprotectivity Introduction (Liman et al., 2005). The BAG protein family, whose members share a highly conserved BAG domain, has Fibroblast growth factor-2 (FGF-2) has been reported been shown to participate in a wide variety of cellular to be important in the development and function of processes including cell survival (stress response), numerous organ systems, promoting proliferation and proliferation, migration and apoptosis (Doong et al., differentiation of a wide range of cell types. Angiogen- 2002). By interacting with the ATPase domain of Hsc70/ esis, hematopoiesis, chondrogenesis, myogenesis, wound Hsp70, BAG proteins modulate their activity in healing and tissue repair are among the biological events physiological and stress conditions. Some BAG proteins that are regulated by FGF-2 (Basilico and Moscatelli, have been demonstrated to inhibit Hsc70/Hsp70 cha- 1992; Schwartz and Liaw, 1993; Bikfalvi et al., 1997). perone activity in vitro (Takayama et al., 1999). BAG1, FGF-2 also has numerous effects on the nervous system, 3 and 4 have also been isolated as Bcl-2-binding proteins which span from regulation of stem cell proliferation (Bcl-2-associated AthanoGene) (Lee et al., 1999). and differentiation to neuronal survival (Gremo and Despite their common C-terminal BAG domain, BAG Presta, 2000). Moreover, FGF-2 is widely expressed in proteins differ widely in their N termini, whose features the embryonic central nervous system (CNS) (Baird, should confer different properties and functions to each 1994), and has been shown to have a mitogenic action member. on a variety of forebrain CNS progenitor cells (Temple BAG3, also known as Bis or CAIR-1, is the only and Qian, 1995; Gritti et al., 1996). FGF-2 is intimately member having a WW domain, which has not yet been linked to a particular function of this protein. The PXXP proline-rich motif, also present in BAG6, Correspondence: Dr K Khalili, Department of Neuroscience, Center mediates the interaction of BAG3 with the SH3 domain for Neurovirology, Temple University School of Medicine, 1900 North of PLC-g (Doong et al., 2000). As do other BAG 12th Street, 015-96, Room 203, Philadelphia, PA 19122, USA. E-mail: [email protected] proteins, BAG3 binds to the ATPase domain of Hsc70 Received 24 January 2008; revised 11 March 2008; accepted 28 March with high affinity and inhibits its chaperone activity in a 2008; published online 12 May 2008 Hip-repressible manner (Takayama et al., 1999; Rosati BAG3 regulation by FGF-2 A Gentilella et al 5012 et al., 2007a). Furthermore, BAG3 has a synergistic effect on the antiapoptotic activity of Bcl-2 (Lee et al., 1999). With respect to cancer, BAG3 has been found to be overexpressed in pancreatic and thyroid tumors (Liaoa et al., 2001; Chiappetta et al., 2007) and it is thought that it exerts a prosurvival effect in chronic lymphocytic leukemia and acute lymphoblastic leukemia primary cultures (Romano et al., 2003a, b). Knockout of the BAG3 gene in the mouse leads to a fulminant myopathy and an early lethality and has unveiled new roles for this protein (Homma et al., 2006). Notably, it has been also observed that BAG3 expression can be induced by some stressful agents such as high tempera- tures and heavy metals (Pagliuca et al., 2003), and HIV-1 infection (Rosati et al., 2007b). However, little is known about the role of BAG3 in nervous system. In light of observations demonstrating the effect of FGF-2 on Egr-1 activity and the presence of potential Egr-1-binding site on the BAG3 promoter, we initiated studies to assess the impact of FGF-2 on BAG3 expression in SK-N-MC cells, a neuroblastoma cell line that is responsive to FGF-2. In the present study, we show that BAG3 expression is induced by FGF-2 in human neuroblastoma cells, SK- N-MC. The signaling cascade required for this event is the mitogen-activated protein kinase (MEK)/extracellular regulated kinase (ERK) pathway, which activates the transcription factor Egr-1. Egr-1 activity is then exerted on BAG3 promoter through two Egr-1-binding sites. Furthermore, BAG3 induction due to FGF-2 affects cell progression through S/G2 phases of the cell cycle.

Results Figure 1 Effect of ﬁbroblast growth factor-2 (FGF-2) on BAG3 in Little is known about the role of BAG3 in the CNS and SK-N-MC cells. (a) Western blot analysis of 20 mg of total proteins less has been found on its effect on the growth of cells of derived from cells untreated (lane 1), treated with FGF-2 alone neuroectodermal origin. Here, we utilized a human (lane 2) or in presence of JNK inhibitor, SP600125 (lane 3) or mitogen-activated protein kinase kinase (MEKK) inhibitor, U0126 neuroepithelioma cell culture model, SK-N-MC, to (lane 4) for 24 h. The immunoblot was reacted with polyclonal investigate the expression of BAG3 and its possible role antibody against human BAG3 and a-tubulin antibody. (b) in cell growth and differentiation. Earlier studies showed Western blot analysis of extracts from cells pretreated with that treatment of SK-N-MC cells with FGF-2 results in U0126 before FGF-2 treatment (24 h) for the detection of BAG3, morphological changes associated with the formation of phosphorylated extracellular regulated kinase (ERK)1/2, total ERK1/2 and tubulin. (c) Northern blot analysis of RNA obtained dendrites and processes. The extent of these changes is from cells as discussed in (b) (24 h after FGF-2 treatment) using noticeably diminished upon pretreatment of cells with BAG3 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) the MEK inhibitor U0126 (Kim et al., 2004). Examina- probes. tion of BAG3 in these cells revealed a substantial increase in the level of this protein in cells that were toward understanding the mechanism underlying BAG3 maintained in 2% serum overnight followed by FGF-2 activation by FGF-2, we performed RNA analysis from treatment for 24 h. Interestingly, ERK1/2 inhibition the cells after FGF-2 treatment in the absence and blocked FGF-2-mediated enhancement of BAG3, presence of the U0126 inhibitor. Results depicted in whereas inhibition of JNK, a pathway that is involved Figure 1c show enhancement of BAG3 mRNA in cells in apoptosis of neuroblastomas at later stages of cell upon treatment with FGF-2, which can be suppressed treatment with FGF-2, showed no impact on the by MEK inhibitor. These observations suggest that induced level of BAG3 in these cells (Figure 1a). FGF-2 stimulates transcription of BAG3 in SK-N-MC Further, activation of BAG3 upon FGF-2 treatment is cells via the MEK pathway. concurrent with the increased level of phosphorylated Next, we examined the activity of the BAG3 promoter ERK1/2 (Figure 1b). Similar results were obtained in in SK-N-MC cells under various conditions as indicated cells maintained in serum-free media followed by in Figure 2a. Treatment of the cells with FGF-2 FGF-2 treatment (data not shown). As a ﬁrst step enhanced the activity of the BAG3 promoter, and in

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5013 agreement with the results shown in Figure 1, this induction is suppressed by MEK inhibition, but not JNK inhibition. We created a series of reporter constructs, containing various regions of the human BAG3 promoter and examined their responsiveness to FGF-2 treatment in SK-N-MC cells. As illustrated in Figure 2b, removal of the sequence spanning nucleotides À969 to À205, with respect to the transcription start site at þ 1, had only minimal effect on the level of FGF-2 activation of the BAG3 promoter. Removal of the DNA sequence between À205 to À26 completely abrogated FGF-2 responsiveness of the BAG3 promoter and also decreased the basal level of BAG3 transcription (Figure 2b). Examination of the FGF-2 promoter sequence spanning À205 to À27 by Match algorithm with Transfac matrices revealed the presence of two separate Egr-1- binding sites at distal (À150 to À142) and proximal (À50 to À42) positions with respect to the transcription start site, and one Elk-1-binding site spanning nucleotides À187 to À183. This is an interesting notion in light of the earlier observations showing that both Egr-1 and Elk-1 can be activated by FGF-2 in various cell lines (Carreras et al., 2001; Kontaridis et al., 2002; Vogel et al., 2006). To assess the contribution of these two domains in the activation of BAG3 promoter by FGF-2, a mutation that alters nucleotide composition of either or both domains was introduced and the transcriptional activity of these mutant promoters in response to FGF-2 was evaluated. As shown in Figure 2c, the level of BAG3 activation by FGF-2 was slightly decreased from the promoters containing a mutation in either the proximal or distal Egr-1-binding sites. Combined activation in both proximal and distal domains showed more impact on the responsiveness of the BAG3 promoter to FGF-2 induction. These observations indicate that the Egr-1-binding sites are important for full induction of BAG3 by FGF-2. To investigate the in-vivo interaction of Egr-1 with the BAG3 DNA sequence spanning À205 to À27, we performed chromatin immunoprecipitation (ChIP) assay. As shown in Figure 3a, we observed a noticeable Figure 2 Activation of BAG3 promoter by fibroblast growth increase in the intensity of the DNA band pointing to factor-2 (FGF-2). (a) Luciferase assay using reporter plasmid enhanced in-vivo interaction of Egr-1 with the BAG3- containing the BAG3 promoter (À969/ þ 306) in SK-N-MC cells. promoter DNA sequence upon treatment of the cells At 24 h after transfection, the cells were treated with FGF-2 and with FGF-2 (compare lanes 1 and 2). Exposure of the the inhibitors for 20 h after which cells were harvested and luciferase and renilla assays were performed. The data are cells to MEK inhibitor suppressed the elevated levels of expressed as fold effect from three independent assays where Egr-1-binding to BAG DNA. The absence of amplified luciferase activity was normalized with renilla expression. DNA fragments from immunocomplexes obtained from (b) Transcription activation of deletion mutants of the normal rabbit serum and the equal level of DNA BAG3 promoter in FGF-2 treatment in SK-N-MC cells amplification from the input DNA verified the specifi- positions. Positions and nucleotide composition of two GC-rich domains located between À150 and À142 (Dist.) and À50 to city of this observation and equal loading and efficiency À42 (Prox.) are shown. Fold activation was determined from of DNA amplification. To further investigate the three independent experiments after normalization with renilla association of Egr-1 with the proximal and distal Egr-1- activity. (c) Transcriptional activation of the BAG3 promoter binding sites of the BAG3-promoter upon FGF-2 (À201/ þ 306) and its mutant variants with alterations at the proximal Egr-1-binding site (nucleotides À50 to À42) (Mut. Prox.), treatment, we performed band shift assays using nuclear distal-binding site (nucleotides À150 to À142) (Mut. Dist) or both extracts prepared from SK-N-MC cells under various sites, Mut. Prox. þ Dist. Fold activation was determined from conditions. As shown in Figure 3b, a new DNA protein three independent experiments. The asterisks represent Po0.01. complex (Cx) was detected when protein extracts from Dist., distal; Prox., proximal; Mut., mutation. FGF-2-treated cells were incubated with DNA probes corresponding to the distal or proximal Egr-1-binding sites. Treatment of the cells with U0126 negated

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5014

Figure 3 Egr-1 interacts with BAG3 promoter in response to fibroblast growth factor-2 (FGF-2). (a) Chromatin immunoprecipitation (ChIP) assay. SK-N-MC incubated with FGF-2 (pretreated with U0126 where indicated) for 16 h were formaldehyde-cross-linked, and chromatin was immunoprecipitated with anti-Egr-1 or normal rabbit serum, followed by PCR amplification with BAG3-promoter- specific primers and analysed by Southern hybridization with Egr-1 probe. Input genomic DNA was used as positive control. (b) Nuclear extracts from SK-N-MC cells, which were either unstimulated (lanes 1–4 and 7–10), treated with FGF-2 alone (lanes 2–5 and 8–11) and in presence of mitogen-activated protein kinase kinase (MEKK) inhibitor (lanes 3–6 and 9–12) were subjected to electrophoretic mobility shift assay (EMSA), using as probe the two putative Egr-1-binding sites (distal site and proximal site) from the BAG3-promoter region responsive to FGF-2. (c) Gel supershift assays were performed with specific anti-Egr-1 antibody or normal rabbit serum as described in ‘Materials and methods’ with nuclear extract from untreated or FGF-2-treated cells. Arrows point to the position of the Egr-1 complex whereas the asterisk indicates a nonspecific complex. The asterisks represent a non-specific band.

formation of this complex. Further, mutant DNA cells were treated with FGF-2 and at various times after probes derived from proximal and distal Egr-1 with treatment, cell extracts were prepared and tested for altered Egr-1-binding sites failed to form a Cx complex Egr-1 expression by western blot analysis. As seen in (Figure 3b). Treatment of the binding reaction with anti- Figure 4a, a significant increase in the level of Egr-1 was Egr-1 antibody, but not with normal rabbit serum, observed 1 h after treatment of cells with FGF-2 and resulted in supershift DNA protein complex suggesting remained constant up to 4 h. Our subsequent study the presence of Egr-1 in the Cx complex (Figure 3c). showed that Egr-1 levels remained high for up to 16 h These observations verify the ability of Egr-1 to following treatment (data not shown). To directly specifically associate with both distal and proximal correlate FGF-2 induction of Egr-1 with BAG3 expres- Egr-1-binding sites within the BAG3 promoter. Of note, sion, cells were treated with Egr-1-targeted small- from the intensity of the band it appears that under interfering RNA (siRNA) and after 24 h they were similar experimental conditions the proximal-binding treated with FGF-2 before harvesting for protein assay. site has a stronger binding affinity than the distal- As shown in Figure 4b, treatment of cells with Egr-1- binding site to Egr-1 (Figure 3b, compare lanes 2 and 8). specific siRNA, but not control nonspecific siRNA, Results from similar studies with a DNA probe diminished the level of BAG3 induction by FGF-2. As containing À187 to À183 sequence of the BAG3 anticipated, the level of Egr-1, but not the housekeeping promoter showed no evidence for the association of a-tubulin, protein was decreased by Egr-1 siRNA. Elk-1 with BAG3 DNA (data not shown). Activation of BAG3 by FGF-2 was also observed in In the next series of experiments, we asked whether another neuroblastoma cell line, SH-SY-5Y and in PC12 treatment of SK-N-MC cells with FGF-2 impacts on the cells, a cell line derived from rat with the ability level of Egr-1 and hence BAG3 expression. To this end, to differentiate toward neuronal lineage. In contrast,

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5015

Figure 4 Effect of fibroblast growth factor-2 (FGF-2) on Egr-1 and BAG3 overexpression. (a) Cells were treated with FGF-2 for the time points indicated, and evaluated for Egr-1 expression by western blot. (b) Cells were transfected with 150 nM small- interfering RNA (siRNA) for 24 h, then treated with and without FGF-2 for an additional 24 h. Expression of Egr-1 and BAG3 was Figure 5 SK-N-MC cell-cycle analysis and cyclin B1 expression. evaluated by western blot using a-tubulin as a loading control. (a) Cells were transfected with 100 nM nonspecific or BAG3-specific (The arrow indicates the specific band recognized by anti-Egr-1). small-interfering RNA (siRNA), treated with FGF-2 and analysed for cell-cycle progression by fluorescence-activated cell sorting FGF-2 treatment of MCF-7 cells, which are derived (FACS) analysis. The average of three experiments is shown. from mammary gland carcinoma and are responsive to (b) Total protein (20 mg) extract from cells as shown in (a) was prepared and analysed by western blot using anti-cyclin B1, anti- FGF-2, had no effect on BAG3 levels (data not shown). BAG3 and anti-tubulin antibodies. To investigate the impact of BAG3 induction by FGF-2 on cell growth and proliferation, cells were transfected with BAG3 siRNA then maintained in media Discussion containing aphidicholine (5 mM) and 2% serum for 16 h. Distribution of cells in cell-cycle stages was assessed by FGF-2 and Egr-1 are important in self-renewal of CNS fluorescence-activated cell sorting analysis. As shown in progenitor stem cells (Temple and Qian, 1995; Gritti Figure 5a, in control cells (transfected with nonspecific et al., 1996). This process is of crucial importance during siRNA), FGF-2 treatment increased cell accumulation at brain development and in response to injuries such as G2/M from 36.1–50% (black bars). At the same time, the ischemia and seizure (O’Donovan et al., 1999; Lowe appearance of cells at S phase was slightly reduced (from et al., 2003). Our results demonstrate that BAG3 is a 16.3 to 11.2%) (light gray bars). Treatment of the cells new target for FGF-2, and that activation of BAG3 with BAG3-specific siRNA, which caused a nearly 50% upon treatment with FGF-2 is mediated by Egr-1. This decrease in the level of BAG3 in the presence of FGF-2 is an interesting observation in light of other reports that (Figure 5b, middle panel), altered distribution of cells in Egr-1 is activated by several other growth factors G2/M and S phases in the presence of FGF-2. As shown in including NGF and BDNF in different cellular contexts Figure 5a, only 40.6% of the cells were accumulated in G2/M (Her et al., 2004; Ro¨ ssler and Thiel, 2004). Curiously, we phase whereas 17.7% of the cells were detected in S observed that NGF does not have any effect on BAG3 phase. Thus, as seen, suppression of BAG3 levels upon expression in SK-N-MC, yet it stimulates BAG3 FGF-2 treatment decreased (19.8%) of the cell distribu- expression in PC12 cells (data not shown), suggesting tion at G2/M and increased (58.0%) cell appearance at S that additional cellular factor(s) is involved in the phase. In the absence of FGF-2, cells transfected with activation of BAG3 by Egr-1 in various cell types. We BAG3 siRNA in the absence of FGF-2 showed no also observed that BAG3 induction by FGF-2 impacts noticeable difference in G2/M distribution compared to cell-cycle progression at S/G2. As previously reported, that from control siRNA. Examination of the level of BAG1, through its BAG domain, interacts with and cyclin B1, a major regulator that orchestrates cell-cycle activates Raf-1 in a Ras-independent manner. This progression through the G2 phase, was lowered in cells event, in turn, activates ERK1/2 and their downstream treated with BAG3 siRNA (Figure 5b). These observa- targets, leading to positive regulation of DNA synthesis tions highlight the possible role of BAG3 on cell-cycle (Song et al., 2001). Due to the degree of homology progression in response to FGF-2 by affecting S/G2 among BAG domains of BAG1 and 3, one may checkpoint that results in the accumulation of cells at S anticipate that BAG3 binds and activates Raf-1 and phase, possibly controlling their entry into G2 phase as a its downstream pathways, thus modulating DNA result of cyclin B1 expression. synthesis. This is interesting in light of our observations

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5016 on the accumulation of cells at S phase of the cell cycle. treatment, cells were transferred to DMEM plus 2% fetal BAG3 has also been reported to abrogate Hsp-70- bovine serum (FBS) 16 h prior to the addition of human mediated proteasomal degradation of client proteins as recombinant FGF-2 (Invitrogen) at a concentration of Akt, EGF-R and Raf-1 (Doong et al., 2003). By 40 ng/ml. sustaining the levels of these proteins, which are involved in growth and proliferation, BAG3 may affect Materials U0126 and SP600125 were purchased from EMD Biosciences FGF-2-mediated S/G2 progression. 32 32 Induction of BAG3 by FGF-2 may have other (Darmstadt, Germany). [a- P]ATP and [g- P]ATP were obtained from PerkinElmer Life Sciences (Waltham, MA, impacts on cell behavior. For example, FGF-2 promotes USA). LipofectAMINE2000 reagent was from Invitrogen, differentiation of SK-N-MC cells toward a neuronal- Fugene 6 from Roche (Basel, Switzerland). Human recombi- like phenotype. This event is associate with increased nant FGF-2 was purchased from Invitrogen. Bradford expression of neurofilament L and class III b-tubulin Reagent was from Bio-Rad (Hercules, CA, USA). Polyclonal (Kim et al., 2004; Russo et al., 2004). However, our antibody to BAG3 (TOS-2) and monoclonal antibody to preliminary results showed that suppression of BAG3 BAG3 (AC-1) were obtained from Alexis Biochemicals (San expression by siRNA has no impact on the expression of Diego, CA USA). A specific siRNA-targeting BAG3 mRNA 0 0 these neuronal markers upon FGF-2 treatment of the (5 -AAGGUUCAGACCAUCUUGGAA-3 ) and a nonspecific siRNA (50-CAGUCGCGUUUGCGACUGG-30) were cells (data not shown). Despite the increase in G2/M population, FGF-2 treatment decreases the growth rate purchased from Dharmacon, (LaFayette, CO, USA). BAG3 siRNA has been selected among three different BAG3- of SK-N-MC cells (Smits et al., 2000). The induction of targeting siRNA sequences that have been evaluated for high BAG3 expression by FGF-2 is not accountable for any specificity and lack of off target effect at the concentration change in proliferation rate in SK-N-MC (data not used. Polyclonal anti-Egr-1 (588) was from Santa Cruz shown). Preliminary data in neural progenitor cells, on Biotechnology Inc. (Santa Cruz, CA, USA). Polyclonal anti- the other hand, suggest that in this cellular context p42/p44 MAP kinases (137F5), polyclonal anti-phospho-p42/ BAG3 expression is sustained by FGF-2 has a p44 (Thr202/Tyr204) were from Cell Signaling Technology proliferative effect. (Danvers, MA, USA). Anti-a-tubulin (clone B-5-1-2) was In light of our new findings, we conclude that BAG3 purchased by Sigma-Aldrich (St Louis, MO, USA). Goat anti- is a new downstream effector of growth factors, in (mouse immunoglobulin G, IgG)-peroxidase conjugate and particular FGF-2, and its activity has a considerable goat anti-(rabbit IgG)-peroxidase conjugate were from Pierce (Rockford, IL, USA). effect on the progression of S/G2 cell-cycle phases. This ascribes a new role for BAG3 in control of the cell cycle Mutagenesis and proliferation, corroborating earlier observations Site-directed mutagenesis was performed in plasmid-contain- demonstrating overexpression of BAG3 in a variety of ing À205/ þ 306 of the BAG3 promoter by PCR using the cancer cells and its possible role in the response of tumor QuikChange II site-directed mutagenesis kit (Strategene, La cells to chemotherapy-induced apoptosis (Romano Jolla, CA, USA) according to the supplier’s protocol. The et al., 2003a, b; Chiappetta et al., 2007). However, the following nucleotides were used as primers for creating the mechanism involved in this event remains to be nucleotide substitutions. Egr-1-binding sites are in bold and investigated. On the other hand, the involvement of the altered nucleotides are underlined: the MEK/ERK pathway in BAG3 upregulation pro- Egr-1 DS, 50-GGCTTCGGCGCGGGCGCGGGCGGGCG CCG-30 vides a rationale for studying the impact of BAG3 in 0 tumors that exhibit dysregulated MEK/ERK pathway. Egr-1 DS, mut, 5 -GGCTTCGGCGCTAGCGCGGGCG GGCGCCG-30 The level of BAG3 was elevated in the brain after Egr-1 PS, 50-GCTCCGGGGCGGGGGCGAGGAGACG transient forebrain/ischemia and seizure induced in GGG-30 experimental animals. These events are associated with Egr-1 PS, mut, 50-GCTCCGGGGCTAGGGCGAGGAGA a high level of FGF-2 expression in affected regions of CGGGG-30. the brain (Issa et al., 2005). This implies a possible role for BAG3 in neural cell survival or repair (Lee et al., Protein analysis 2002a, b). Creation of an experimental animal model Cell protein extracts were prepared as described previously with conditional expression of BAG3 in brain will (Zapata et al., 1998) and quantitated by Bradford Assay (Bio- provide important clues regarding the involvement of Rad). Cell protein extracts (15–20 mg) were resuspended in BAG3 in the development of brain and responses to SDS-sample buffer and after treatment at 95 1C for 5 min, stress and injury. proteins were size-fractionated by 10% SDS–polyacrylamide gel electrophoresis (PAGE), and transferred to nitrocellulose membranes. Blots were stained with Ponceau S to confirm equal loading and transfer of proteins and then reacted with Materials and methods the indicated antibodies. Immunoblots were developed using appropriate secondary horseradish peroxidase-coupled anti- Cell culture bodies and an enhanced chemiluminescence plus kit (Pierce). SK-N-MC are human neuroepithelioma cells that we obtained from the American Type Culture Collection (Rockville, MD, RNA analysis USA) and maintained in Dulbecco’s modified Eagle’s medium Total cellular RNA (1.8 Â 105 cells per sample) was isolated (DMEM) supplemented with 10% heat-inactivated fetal using TRIzol reagent (Invitrogen) according to the manufac- bovine serum (Invitrogen, Carlsbad, CA, USA). For FGF-2 turer’s instructions. Total RNA (12 mg) was resolved on

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5017 formaldehyde-containing 1.2% agarose gel and transferred to 1%) to crosslink chromatin. Cell lysates were then prepared Hybond-N nylon membranes (GE Healthcare) and crosslinked and sonicated on ice to break chromatin DNA to an average by ultraviolet irradiation. A 823 bp Sal1/Xho1 DNA fragment length of 500 bp. Cell lysates (from 106 cells per sample) were corresponding to the BAG3-coding region was radiolabeled treated with Egr-1 antibody (2 mg). DNA amplification of with [a-32P]dCTP and used as a probe. The filter was pre- the captured DNA immunocomplex was performed using hybridized 1 h at 48 1C in UltraHyb solution (Ambion, Austin, primers complementary to the BAG3-promoter region harboring TX, USA) and hybridized for 20 h at 48 1C with cDNA probe both putative Egr-1-binding sites. The primer sequences were for BAG3 (106 c.p.m./ml). The blot was then rinsed twice 5 min 50-GACGGCCCCAGTCCAGCTCG-30 as forward and with 2 Â SSC 0.1% SDS at 50 1C and washed twice with 0.2 Â 50-CTGAGTCATCGGCTATAATCG-30 as reverse to generate SSC, 0.1% SDS at 50 1C for 20 min. The filter was exposed at a 204 bp amplification product. DNA samples were analysed À80 1C to Fuji XAR-5 film with Kodak intensifying screen. To by electrophoresis on 1.2% agarose gels, stained with ethidium verify equivalent RNA loading on northern blot membrane, bromide then transferred to Hybond N nylon membranes the filter was rehybridized with glyceraldehyde 3-phosphate (Amersham Biosciences, Piscataway, NJ, USA). The filter was dehydrogenase cDNA probe. prehybridized 1 h at 42 1C with Ultrahyb (Ambion), then probed using 106 c.p.m./ml of Egr-1 PS oligo terminally labeled Transfection and luciferase assays with T4 polynucleotide kinase (Roche) and [g-32P]dATP. After The SK-N-MC cells were transfected using Fugene 6 (Roche). 20 h hybridization, the blot was rinsed twice (5 min) with 2 Â Cells were plated at a density of 1.8 Â 105 cells per well in a six- SSC, 0. % SDS at 42 1C and washed twice with 0.2 Â SSC, well plate 48 h prior to transfection. Transfection was 0.1% SDS at 42 1C for 20 min. The filter was exposed at performed according to the manufacturer’s protocol with À80 1C to Fuji XAR-5 film with Kodak intensifying screen. 2 mg of reporter plasmid and 250 ng of Renilla luciferase control plasmid (Promega, Madison, WI, USA) for 24 h. Following this period, the transfection mixture was removed Cell-cycle analysis and replaced with media with or without 40 ng/ml FGF-2 for Cell-cycle analysis was performed using the Guava cell-cycle 20 h. Cell extracts were subsequently prepared and assayed reagent protocol (Guava Technologies, Hayward, CA, USA). using the Dual Luciferase kit (Promega) as per the manufac- Briefly, SK-N-MC cells were grown overnight in 2% serum turer’s instructions. Luciferase activities were normalized to plus 5 mM aphidicholine DMEM then transfected with 100 nM the Renilla control plasmid, and values shown are the mean of BAG3 siRNA or nonspecific siRNA using Lipofectamine three independent experiments. 2000. After 12 h, cells were maintained in DMEM containing 2% serum for additional 12 h prior to the addition of FGF-2. After 24 h, cells were harvested, fixed with 70% ethanol and Electrophoretic mobility shift assay treated for 20 min with Guava cell-cycle reagent. Samples were Nuclear extracts were prepared as described previously evaluated on Guava EasyCyte Mini System. The data are (Andrews and Faller, 1991). Electrophoretic mobility shift representative of three different experiments. assays (EMSAs) were performed by incubating 10 mgof nuclear extracts with [32P]-end-labeled double-stranded oligo- nucleotide probes (5.4 Â 104 c.p.m.) in binding buffer contain- Statistical analysis ing 10 mM Tris (pH 7.5), 75 mM KCl, 10% glycerol, 0.1 mM All statistical analyses have been done by STDEV and EDTA, 2.5 mM MgCl2, 0.25 mM DTT and 1.0 mg poly(dI-dC) AVERAGE programs. at 4 1C for 20 min. Supershifts were performed by adding 2.0 mg antibody to the extract 30 min prior to incubation with probe. Protein/DNA complexes were resolved on 6% poly- Acknowledgements acrylamide gel run in 0.5 Â TBE, at 170 V at room temperature. The primers used for mutagenesis were also used for We thank past and present members of the Department of EMSA. Neuroscience and Center for Neurovirology for sharing of ideas and reagents, and their encouragement throughout this Chromatin immunoprecipitation assays study. We also thank Dr Francesca Peruzzi, Dr Davide Eletto Chromatin immunoprecipitation assays were performed using and Dr Thersa M Sweet for their advices, William Yen and the ChIP assay kit (Upstate Biotechnology Inc., Lake Placid, Jessica Otte for their technical support, Dr Martyn White for NY, USA), following the manufacturer’s instructions. Briefly, his time and assistance in the preparation of the manuscript SK-N-MC were incubated in presence or absence of FGF-2 and C Schriver for editorial assistance. This study is dedicated (pretreated with 15 mM U0126 as specified) for 16 h. Cells were to our friend and colleague, Dr Arturo Leone. This work was harvested and treated with formaldehyde (final concentration made possible by grants awarded by NIH to KK.

References

Alzheimer C, Werner S. (2002). Fibroblast growth factors and Carreras I, Rich CB, Jaworski JA, Dicamillo SJ, Panchenko MP, neuroprotection. Adv Exp Med Biol 513: 335–351. Goldstein R et al. (2001). Functional components of basic Andrews NC, Faller DV. (1991). A rapid micropreparation technique fibroblast growth factor signaling that inhibit lung elastin for extraction of DNA binding proteins from limiting numbers of gene expression. Am J Physiol Lung Cell Mol Physiol 281: mammalian cells. Nucleic Acids Res 19: 2499. L766–L775. Baird A. (1994). Fibroblast growth factors: activities and significance Chiappetta G, Ammirante M, Basile A, Rosati A, Festa M, of non-neurotrophic growth factors. Curr Opin Neurobiol 4: 78–86. Monaco M et al. (2007). The antiapoptotic protein Basilico C, Moscatelli D. (1992). The FGF family of growth factors BAG3 is expressed in thyroid carcinomas and modulates and oncogenes. Adv Cancer Res 59: 115–165. apoptosis mediated by tumor necrosis factor-related Bikfalvi A, Klein S, Pintucci G, Rifkin DB. (1997). Biological roles of apoptosis-inducing ligand. J Clin Endocrinol Metab 92: fibroblast growth factor-2. Endocr Rev 18: 26–45. 1159–1163.

Oncogene BAG3 regulation by FGF-2 A Gentilella et al 5018 Doong H, Price J, Kim YS, Gasbarre C, Probst J, Liotta LA et al. reverse-transcription polymerase chain reaction using self-quenched, (2000). CAIR-1/BAG-3 forms an EGF-regulated ternary complex fluorogenic primers. Anal Biochem 315: 95–105. with phospholipase C-g and Hsp70/Hsc70. Oncogene 19: 4385–4395. Midgley VC, Khachigian LM. (2004). Fibroblast growth factor-2 Doong H, Rizzo K, Fang S, Kulpa V, Weissman AM, Kohn EC. induction of platelet-derived growth factor-C chain transcription in (2003). CAIR-1/BAG3 abrogates heat shock protein-70 chaperone vascular smooth muscle cells is ERK-dependent but not JNK- complex-mediated protein degradation: accumulation of poly- dependent and mediated by Egr-1. J Biol Chem 279: 40289–40295. ubiquinated Hsp90 client proteins. J Biol Chem 278: 28490–28500. O’Donovan KJ, Tourtellotte WG, Millbrandt J, Baraban JM. (1999). Doong H, Vrailas A, Kohn EC. (2002). What’s in the ‘BAG’? The EGR family of transcription-regulatory factors: progress at the A functional domain analysis of the BAG-family proteins. Cancer interface of molecular and systems neuroscience. Trends Neurosci 22: Lett 188: 25–32. 167–173. Gotz R, Wiese S, Takayama S, Camarero GC, Rossoll W, Schweizer U Pagliuca MG, Lerose R, Cigliano S, Leone A. (2003). Regulation by et al. (2005). Bag1 is essential for differentiation and survival of heavy metals and temperature of the human BAG-3 gene, a hematopoietic and neuronal cells. Nat Neurosci 8: 1169–1178. modulator of Hsp70 activity. FEBS Lett 541: 11–15. Gremo F, Presta M. (2000). Role of fibroblast growth factor-2 in Romano MF, Festa M, Pagliuca G, Lerose R, Bisogni R, Chiurazzi F human brain: a focus on development. Int J Dev Neurosci 18: et al. (2003a). BAG3 protein controls B-chronic lymphocytic 271–279. leukaemia cell apoptosis. Cell Death Diff 10: 383–385. Gritti A, Parati EA, Cova L, Frolichsthal P, Galli R, Wanke E et al. Romano MF, Festa M, Petrella A, Rosati A, Pascale M, Bisogni R (1996). Multipotential stem cells from the adult mouse brain et al. (2003b). A. BAG3 protein regulates cell survival in childhood proliferate and self-renew in response to basic fibroblast growth acute lymphoblastic leukemia cells. Cancer Biol Ther 2: 508–510. factor. J Neurosci 16: 1091–1100. Rosati A, Ammirante M, Gentilella A, Basile A, Festa M, Pascale M Her S, Arimochi H, Morita K. (2004). Nerve growth factor induces et al. (2007a). Apoptosis inhibition in cancer cells: a novel molecular elevation of steroid 5 alpha-reductase mRNA levels in rat C6 glioma pathway that involves BAG3 protein. Int J Biochem Cell Biol 39: cells through expression of transcription factor Egr-1. Brain Res Mol 1337–1342. Brain Res 126: 157–164. Rosati A, Leone A, Del Valle L, Amini S, Khalili K, Turco MC. Homma S, Iwasaki M, Shelton GD, Engvall E, Reed JC, Takayama S. (2007b). Evidence for BAG3 modulation of HIV-1 gene transcrip- (2006). AG3 deficiency results in fulminant myopathy and early tion. J Cell Physiol 210: 676–683. lethality. Am J Pathol 169: 761–773. Ro¨ ssler OG, Thiel G. (2004). Brain-derived neurotrophic factor-, Issa R, AlQteishat A, Mitsios N, Saka M, Krupinski J, Tarkowski E epidermal growth factor-, or A-Raf-induced growth of HaCaT et al. (2005). Expression of basic fibroblast growth factor mRNA keratinocytes requires extracellular signal-regulated kinase. Am J and protein in the human brain following ischaemic stroke. Physiol Cell Physiol 286: C1118–C1129. Angiogenesis 8: 53–62. Russo VC, Andaloro E, Fornaro SA, Najdovska S, Newgreen DF, Kim MS, Kim CJ, Jung HS, Seo MR, Juhnn YS, Shin HY et al. Bach LA et al. (2004). Fibroblast growth factor-2 over-rides insulin- (2004). Fibroblast growth factor 2 induces differentiation and like growth factor-1 induced proliferation and cell survival in human apoptosis of Askin tumour cells. J Pathol 202: 103–112. neuroblastoma cells. J Cell Physiol 199: 371–380. Kontaridis MI, Liu X, Zhang L, Bennett AM. (2002). Role of Schwartz SM, Liaw L. (1993). Growth control and morphogenesis in SHP-2 in fibroblast growth factor receptor-mediated suppression the development and pathology of arteries. J Cardiovasc Pharmacol of myogenesis in C2C12 myoblasts. Mol Cell Biol 22: 21(Suppl): S31–S49. 3875–3891. Smits VA, van Peer MA, Essers MA, Klopmaker R, Rijksen G, Lee JH, Takahashi T, Yasuhara N, Inazawa J, Kamada S, Tsujimoto Y Medema RH. (2000). Negative growth regulation of SK-N-MC cells et al. (1999). A Bcl-2-binding protein that synergizes with Bcl-2 in by bFGF defines a growth factor-sensitive point in G2. J Biol Chem preventing cell death. Oncogene 18: 6183–6190. 275: 19375–19381. Lee MY, Kim SY, Choi JS, Choi YS, Jeon MH, Lee JH et al. (2002b). Song J, Takeda M, Morimoto R. (2001). BAG1-Hsp70 mediates a Induction of Bis, a Bcl-2-binding protein, in reactive astrocytes of physiological stress signaling pathway that regulates Raf-1/ERK the rat hippocampus following kainic acid-induced seizure. Exp Mol and cell growth morphogenesis. Nat Cell Biol 3: 276–282. Med 34: 167–171. Takayama S, Xie Z, Reed JC. (1999). An evolutionarily conserved Lee MY, Kim SY, Shin SL, Choi YS, Lee JH, Tsujimoto Y et al. family of Hsp70/Hsc70 molecular chaperone regulators. J Biol Chem (2002a). Reactive astrocytes express bis, a bcl-2-binding protein, 274: 781–786. after transient forebrain ischemia. Exp Neurol 175: 338–346. Temple S, Qian X. (1995). bFGF, neurotrophins, and the control of Liaoa Q, Ozawaa F, Friessa H, Zimmermann A, Takayama S, Reed JC cortical neurogenesis. Neuron 15: 249–252. et al. (2001). The anti-apoptotic protein BAG-3 is overexpressed in Thiel G, Cibelli G. (2002). Regulation of life and death by the zinc pancreatic cancer and induced by heat stress in pancreatic cancer cell finger transcription factor Egr-1. J Cell Physiol 193: 287–292. lines. FEBS Letters 503: 151–157. Vogel S, Kubin T, von der Ahe D, Deindl E, Schaper W, Zimmermann Liman J, Ganesan S, Dohm CP, Krajewski S, Reed JC, Bahr M et al. R. (2006). MEK hyperphosphorylation coincides with cell cycle shut (2005). Interaction of BAG1 and Hsp70 mediates neuroprotectivity down of cultured smooth muscle cells. J Cell Physiol 206: 25–34. and increases chaperone activity. Mol Cell Biol 25: 3715–3725. Zapata JM, Takahashi R, Salvesen GS, Reed JC. (1998). Granzyme Lowe B, Avila HA, Bloom FR, Gleeson M, Kusser W. release and caspase activation in activated human T-lymphocytes. (2003). Quantitation of gene expression in neuronal precursors by J Biol Chem 273: 6916–6920.

Oncogene