The proapoptotic dp5 gene is a direct target of the MLK-JNK-c-Jun pathway in sympathetic neurons.

The death of sympathetic neurons after nerve growth factor (NGF) withdrawal requires de novo gene expression. Dp5 was one of the first NGF withdrawal-induced genes to be identified and it encodes a proapoptotic BH3-only member of the Bcl-2 family. To study how dp5 transcription is regulated by NGF withdrawal we cloned the regulatory regions of the rat dp5 gene and constructed a series of dp5-luciferase reporter plasmids. In microinjection experiments with sympathetic neurons we found that three regions of dp5 contribute to its induction after NGF withdrawal: the promoter, a conserved region in the single intron, and sequences in the 3' untranslated region of the dp5 mRNA. A construct containing all three regions is efficiently activated by NGF withdrawal and, like the endogenous dp5, its induction requires mixed-lineage kinase (MLK) and c-Jun N-terminal kinase (JNK) activity. JNKs phosphorylate the AP-1 transcription factor c-Jun, and thereby increase its activity. We identified a conserved ATF site in the dp5 promoter that binds c-Jun and ATF2, which is critical for dp5 promoter induction after NGF withdrawal. These results suggest that part of the mechanism by which the MLK-JNK-c-Jun pathway promotes neuronal apoptosis is by activating the transcription of the dp5 gene.

INTRODUCTION

Apoptosis occurs extensively during the normal development of the mammalian nervous system, and is important for establishing neuronal populations of the correct size and for eliminating neurons that have made inappropriate connections (1,2). Developing sympathetic neurons depend on nerve growth factor, synthesized by their target tissues, for survival. In the absence of nerve growth factor (NGF), these cells die by apoptosis and their death requires de novo gene expression (3). Sympathetic neurons have been widely used for in vitro studies of the molecular mechanisms of neuronal apoptosis and a considerable amount has been learned about the signalling pathways that regulate the cell death programme (4,5). Following NGF withdrawal, the stress-responsive mixed-lineage kinase (MLK) and c-Jun N-terminal kinase (JNK) protein kinase cascade is activated and JNKs phosphorylate the AP-1 transcription factor c-Jun, which increases c-Jun activity and c-Jun expression (6–10). The MLK-JNK-c-Jun pathway is required for normal NGF withdrawal-induced death and promotes the release of mitochondrial cytochrome c and caspase activation (11–15).

The release of cytochrome c and other proapoptotic proteins from mitochondria is regulated by the Bcl-2 protein family (16). In sympathetic neurons, the multidomain proapoptotic Bcl-2 family member Bax is essential for cytochrome c release and cell death after NGF deprivation (17). In contrast, the antiapoptotic proteins Bcl-2 and Bcl-xL, which can form heterodimers with Bax, inhibit cytochrome c release and protect against NGF withdrawal-induced death (14,18–20). Finally, several proapoptotic BH3-only Bcl-2 family members are expressed in sympathetic neurons and three of these are regulated by NGF withdrawal: the dp5, bim and puma mRNAs and proteins increase in level after NGF deprivation, in all cases before the cell death commitment point (14,21–24). These BH3-only proteins may promote sympathetic neuron apoptosis by binding to the antiapoptotic members of the Bcl-2 family, which would then be unable to interact with Bax, or possibly by directly binding to and activating Bax (25).

The BH3-only proteins that increase in level after NGF withdrawal are downstream targets of the MLK-JNK-c-Jun pathway. Expression of a c-Jun dominant negative mutant (JunΔ169) or the JunAA knock-in mutation in mice, which eliminates the two major JNK phosphorylation sites in c-Jun, reduce the increase in bim RNA and protein levels after NGF withdrawal (14,24). In addition, the MLK inhibitor CEP-1347, which prevents JNK activation, also reduces the increase in bim and dp5 mRNA levels after NGF deprivation (22,23). To understand in general how JNKs and AP-1 transcription factors promote neuronal apoptosis it is important to determine the molecular mechanisms by which these proteins regulate Bim and Dp5 expression and NGF-dependent sympathetic neurons have been a useful model for these studies (26,27). Here, we use a dp5 reporter gene assay, expression vectors for JNK and AP-1 inhibitor proteins, specific chemical inhibitors, and site-directed mutagenesis to investigate how NGF withdrawal activates dp5 transcription in sympathetic neurons. We show that an ATF-binding site in the dp5 promoter, closely related in sequence to the jun1 and jun2 TRE elements in the c-jun promoter, can bind c-Jun and ATF2 in vitro and in chromatin, and is critical for dp5 promoter activity in sympathetic neurons and for promoter induction following NGF withdrawal.

MATERIALS AND METHODS

5′ RACE and cloning of the dp5 promoter

5′ rapid amplification of cDNA ends (RACE) was performed on rat brain mRNA using the Marathon™ cDNA amplification kit (Clontech Laboratories Inc.) with the dp5-specific primer 5′-CTGCAGCCGCAGCGCGGTCACC-3′. To isolate the dp5 promoter, the 5′ RACE product was used as a probe to screen the rat P1 artifical chromosome (PAC) library RPCI31 (generated by P.Y. Woon and P. de Jong, UK Human Genome Mapping Project Resource Centre, Cambridge, UK). A 4-kb fragment from upstream of exon 1 was cloned and the remaining promoter sequence between the start codon and upstream 4 kb fragment was generated by PCR. These sequences were ligated at an EcoRI restriction site to generate a 4.5 kb fragment in pGEM®-T Easy (Promega UK Ltd, Southampton, UK).

Plasmid constructs

The dp5 reporter construct dp5-LUC was generated by subcloning a 1017 bp fragment containing the dp5 promoter sequence from −980 to +37, relative to the transcriptional start site, into pGL3-basic (Promega UK Ltd.), upstream of the luciferase gene. To obtain a reporter construct also containing part of the dp5 intron (dp5-LUC + I 400), a 395 bp region of the intron that is highly conserved was subcloned into dp5-LUC downstream of the SV40 polyadenylation termination sequence. The construct dp5-LUC + 3′ UTR was made by subcloning three fragments (a 645 bp fragment containing the dp5 stop codon, a 2.36 kb fragment and a 2.02 kb fragment) that spanned the entire dp5 3′ UTR into dp5-LUC downstream of the luciferase gene. dp5-LUC + ALL was generated by subcloning the 395 bp intron fragment into dp5-LUC + 3′ UTR. For all constructs, fragment orientation and positioning was confirmed by restriction enzyme analysis and DNA sequencing.

Point mutations within the ATF-binding site in the dp5 promoter were introduced into dp5-LUC + ALL using the QuikChangeTM II XL site-directed mutagenesis kit (Stratagene Europe, Amsterdam, The Netherlands) and the oligonucleotides 5′-CCCGGGCCGGATAACGTCTCCCCCTCCCCGCG-3′ and 5′-CGCGGGGAGGGGGAGACGTTATCCGGCCCGGG-3′ containing four point mutations. Incorporation of the correct mutations was confirmed by DNA sequencing.

The expression vectors for the JIP-1 JNK-binding domain (JBD) and c-Jun (ala) are described in (12) and (28).

Cell culture

Sympathetic neurons were prepared from the superior cervical ganglia (SCG) of 1-day-old Sprague Dawley rats (supplied by the Biological Services Unit, University College London). Neurons were isolated and cultured as described previously (29), in DMEM (Sigma-Aldrich, Poole, UK) containing 10% foetal calf serum, 2 mM glutamine (Invitrogen Ltd, Paisley, UK) and penicillin-streptomycin (SCG medium). Unless stated otherwise, SCG medium was supplemented immediately prior to use with 2.5 S NGF (Cedarlane Laboratories Ltd., Hornby, Ontario) at 50 ng/ml and the antimitotic agents fluorodeoxyuridine and uridine (both from Sigma-Aldrich) each at 20 µM. Sympathetic neurons were plated on 13 mm diameter glass coverslips coated with poly-l-lysine and laminin placed in 3.5 cm dishes. Cells were maintained in 2 ml of medium at 37°C in 10% CO2 for 5–7 days before being used for experiments. In NGF withdrawal experiments, the growth medium was removed and the cells were gently rinsed twice with SCG medium lacking NGF and antimitotic agents. The neurons were then refed with fresh medium containing NGF or a neutralising anti-NGF antibody (Chemicon Europe Ltd, Chandlers Ford, UK) at 100 ng/ml. The MLK3 inhibitor CEP-11004 (provided by Cephalon, Inc., West Chester, PA) was dissolved in DMSO and used at a final concentration of 400 nM.

The PC6-3 subline of the PC12 cell line (30) was cultured in RPMI 1640 medium (Invitrogen Ltd.) supplemented with 10% horse serum, 5% FCS, 2 mM glutamine and penicillin/streptomycin. Cells were grown on collagen-coated tissue culture dishes at 37°C in 5% CO2 and passaged once a week. For differentiation, cells were plated at a density of 1 × 106 cells per 9 cm dish and maintained for 7 days in RPMI 1640 medium containing 2% horse serum, 1% FCS, penicillin/streptomycin and NGF (Promega UK Ltd.) at 100 ng/ml. In NGF withdrawal experiments, the neuronally differentiated PC6-3 cells were rinsed twice with differentiation medium lacking NGF and then refed with medium containing NGF or anti-NGF antibody, as required.

Microinjection and dual luciferase assay

Sympathetic neurons were microinjected as described previously (14). The different reporter constructs and expression vectors were injected at the concentrations indicated (see Results). The injection mix contained the DNA to be tested in 0.5× PBS (–Ca2+, –Mg2+) (Sigma-Aldrich) plus 5–10 ng/µl of the Renilla luciferase construct pRL-TK (Promega) to control for variations in the volume of DNA or number of cells injected per coverslip. The DNA was injected directly into the nucleus and at least 120 neurons were injected for each condition tested. Typically 50–80% of the neurons survived injection.

At 16–24 h after injection, the neurons were harvested and luciferase activity was determined using the Dual-Luciferase reporter assay system (Promega). Neurons were rinsed off the coverslips in ice-cold PBS and were collected by centrifugation before being lysed in 25 µl of passive lysis buffer. The luciferase assay was then performed using a Lumat LB 9507 luminometer following the protocol provided with the assay system. The output for firefly luciferase was normalized to the Renilla luciferase output (firefly luciferase output divided by Renilla luciferase output). The relative induction of each construct was then calculated (for example, normalized luciferase activity in the absence of NGF divided by normalized luciferase activity in the presence of NGF). Each experiment was performed at least three times using different neuron preparations and the standard error of the mean was calculated.

For antibody co-injection experiments, the c-Jun (H-79) X and ATF2 (C-19) X rabbit polyclonal antibodies (Santa Cruz Biotechnology Inc.) were diluted in PBS (−Ca2+, −Mg2+) and centrifuged in Microcon YM-3 centrifugal filters (Millipore Corporation, Bedford, MA) to remove sodium azide. The final antibody concentration was adjusted to 2 µg/µl. Purified rabbit immunoglobulin in PBS (Jackson ImmunoResearch Laboratories, Inc.) was used as a control. Neurons were microinjected with dp5-LUC + ALL or pLuc-MCS or pCRE-Luc (Stratagene) all at 20 ng/µl, pRL-TK (10 ng/µl) and antibody (1 µg/µl). After injection, in the case of dp5-LUC + ALL, the cells were rinsed twice with SCG medium and refed with medium containing NGF or anti-NGF antibody. For pLuc-MCS and pCRE-Luc the injected cells were refed with + NGF medium containing 500 µM CPTcAMP (Sigma-Aldrich). Twenty hours later, the cells were harvested and a dual luciferase assay performed.

RT–PCR

RNA was isolated from sympathetic neurons using an RNeasy kit (QIAGEN Ltd.). Total RNA was eluted in 30 µl and, after treatment with amplification grade DNase I (Invitrogen Ltd) to remove any contaminating DNA, 10 µl of the purified RNA was reverse transcribed using Superscript II reverse transcriptase (Invitrogen Ltd). Fifty microliter PCR reactions were prepared using 1–4 µl of cDNA, 0.2 mM dNTPs, 0.1 µg of oligonucleotide primers and 0.05 U/µl of REDTaq DNA polymerase in 1× RedTaq PCR reaction buffer (Sigma-Aldrich). Cycling parameters were 94°C for 30 s, 58°C for 20 s and 72°C for 60 s and PCR cycles were performed a sufficient number of times for a product to be detected for each sample in a set, typically 30–38 cycles. The amplified products were resolved on 2.5% agarose gels containing ethidium bromide at 0.5 µg/ml. Images were captured using a UVIdoc gel documentation system (UVItec Ltd, Cambridge, UK) and quantified using ImageMaster TotalLab imaging software (GE Healthcare UK Ltd, Chalfont St. Giles, UK). Experiments were performed four times and the average of three PCR reactions and SEM were calculated for each cDNA sample. The following primers were used:

dp5, 5′-AGACCCAGCCCGGACCGAGCAA-3′ and

5′-ATAGCACTGAGGTGGCTATC-3′;

neurofilament (nf-m), 5′-ACGCTGGACTCGCTGGGCAA-3′ and 5′-GCGAGCGCGCTGCGCTTGTA-3′.

In vitro transcription and translation

ATF2, c-Fos, c-Jun, c-JunΔ169 and Δ169m0 were produced in vitro using the TNT T7 or T3 Coupled Reticulocyte Lysate system (Promega). All plasmids used contained the β-globin RNA leader sequence, which increases translation efficiency. The plasmid pBAT ATF2 (31) was used with the T3 coupled system as a template for the synthesis of ATF2. All other proteins were translated using the T7 coupled system. T7 c-Fos was provided by Curt Pfarr and Moshe Yaniv (Institut Pasteur, Paris). The plasmid pCDc-Jun and pCDFLAGΔ169, the expression vector for a c-Jun dominant negative mutant, are described in (7). Δ169m0 protein was made using the plasmid pCDFLAGΔ169m0, in which the m0 leucine zipper mutation (31) had been transferred into the c-Jun sequences in pCDFLAGΔ169 by subcloning. After coupled transcription/translation at 30°C for 60 min in a volume of 50 µl, an equal volume of 2x dialysis buffer (40 mM Hepes pH 7.9, 50 mM KCl, 2 mM MgCl2, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 20% glycerol) was added to each translation. The translated proteins were divided into aliquots, frozen on dry ice, and stored at −80°C.

In some experiments, the in vitro translated proteins were labelled with 35S-methionine and run on a 12% SDS polyacrylamide gel. The gel was fixed, treated with AmplifyTM solution (GE Healthcare UK Ltd.) and dried using a BioRad gel dryer (model 583) before being exposed to HyperfilmTM X-ray film (GE Healthcare UK Ltd).

Electrophoretic mobility shift assays

Double-stranded oligonucleotides were labelled with [α-32P]dCTP (3000 Ci/mmol; PerkinElmer) using Klenow polymerase (Roche Diagnostics Ltd, Lewes, UK) to fill in 5′ overhangs. The following pairs of oligonucleotides were used (binding sites are underlined): collagenase TRE, 5′-CTAGAGCATGAGTCAGACAC-3′ and 5′-CTAGGTGTCTGACTCATGCT-3′; jun2 TRE, 5′-CTAGAGCATTACCTCATCCC-3′ and 5′-GTACGGGATGAGGTAATGCT-3′; dp5 ATF, 5′-CTAGGCCGGATGATGTAACCCCCT-3′ and 5′-CTAGAGGGGGTTACATCATCCGGC-3′; dp5 ATF mutant, 5′-CTAGGCCGGATAACGTCTCCCCCT-3′ and 5′-CTAGAGGGGGAGACGTTATCCGGC-3′.

Binding reactions were prepared with 40 mM KCl, 20 mM Hepes pH 7.9, 5 mM MgCl2, 1 mM EGTA, 0.5 mM DTT, 10% glycerol, 0.5 µg/µl BSA, 1 µg poly(dI-dC) and 4 µl of in vitro translated protein/unprogrammed rabbit reticulocyte lysate to a final volume of 19 µl. The reaction was prepared without the radiolabelled oligonucleotide, and incubated at room temperature for 15 min. One microlitre containing 0.4 ng of 32P-labelled probe was then added before a further 15 min incubation at room temperature. The binding reactions were loaded onto a 5% polyacrylamide/0.25× TBE native protein gel. Following electrophoresis at 180 V for ∼2 h at room temperature, the gel was fixed for 15 min in 10% acetic acid, 10% methanol and dried at 80°C under vacuum. The bands were visualised by exposing the dried gel to Kodak MXB X-ray film (Kodak Ltd, Hemel Hempstead, UK) or a phosphorimager screen overnight. The exposed phosphorimager screen was scanned using a TyphoonTM 8600 phosphorimager and the image analysed using ImageQuant software and saved as a TIFF file. Exposed X-ray films were scanned using an Epson photo scanner (model 4990) and the resulting images were saved as TIFF files.

Sympathetic neuron whole-cell extracts for EMSA experiments were prepared as follows. The SCG neurons isolated from 30 1-day-old rats were plated in four 3.5 cm dishes coated with poly-l-lysine and laminin. After 7 days in vitro, the medium was removed, the cells were rinsed twice with SCG medium, and two dishes were refed with medium containing NGF and two with medium containing anti-NGF antibody. Sixteen hours later, the dishes were placed on ice and the medium was removed. The neurons were then rinsed off using 1 ml of ice-cold PBS per dish and transferred to 1.5 ml microfuge tubes on ice. The cells were pelleted by centrifugation for 5 min at 4°C and the cells from each pair of dishes were pooled and resuspended in 1 ml of ice-cold PBS. The cells were spun again for 5 min at 4°C and then suspended in 30 µl of whole cell extract buffer (0.1% NP-40, 250 mM KCl, 50 mM Hepes pH 7.9, 10% glycerol, 0.2 mM EDTA, 0.2 mM EGTA, containing the following inhibitors added just before use, 4 mM NaF, 4 mM Na3VO4, 1 mM DTT, 0.5 mM PMSF, 2% v/v Sigma mammalian protease inhibitor cocktail). The lysates were incubated on ice for 30 min and were gently pipetted up and down from time to time. Cell debris was removed by centrifugation for 30 min at 16 000 × g in a microfuge at 4°C, and the supernatants were frozen in aliquots on dry ice and stored at −80°C. Protein concentration was determined using the Bio-Rad protein assay.

EMSA experiments with sympathetic neuron extracts were carried out as described above, except that the volume of the binding reaction was 25 µl and 4 µl of whole cell extract buffer was included. From 4 to 8 µg of whole cell extract was used per binding reaction and for supershift assays 2–4 µl of antibody was added. The antibodies were c-Jun (H-79) X, ATF2 (C-19) X, phospho-c-Jun (KM-1) X (all from Santa Cruz Biotechnology, Inc.), phospho-ATF2 (Thr71) (Cell Signaling Technology) and, as a control, Bim AB17003 (Chemicon). Binding reactions were for 2 h at 4°C, after which 0.4 ng of the 32P-labelled double-stranded oligo was added. The samples were incubated for 15 min at room temperature, and then electrophoresed on a 5% polyacrylamide, 0.25× TBE gel.

Chromatin immunoprecipitation (ChIP)

ChIP was performed as described (32) with a number of modifications. Most of the buffers used were from an Upstate ChIP assay kit (Upstate Ltd, Dundee, UK). PC6-3 cells were plated at a density of 1 × 106 cells per 9 cm dish and differentiated for 7 days. After 7 days, the cell density had increased to ∼4 × 106 cells per 9 cm dish. The differentiated cells were rinsed twice and cultured in medium containing NGF or anti-NGF antibody for 16 h. Proteins and DNA were cross linked by adding 37% formaldehyde to the culture medium to a final concentration of 1% and incubating at room temperature for 3 h. Two hundred microlitre of aliquots of fragmented chromatin (∼8 × 106 nuclei) were used per ChiP, and the c-Jun (H-79) X and ATF2 (C-19) X antibodies were diluted 1:1000 in ChIP dilution buffer in a volume of 2 ml. The phospho-c-Jun (serine 63) KM-1 X antibody was diluted 1:100. For preclearing and for recovery of the immune complexes 50% protein A/G-agarose (without DNA, Santa Cruz Biotechnology, Inc.) in ChIP dilution buffer containing BSA (10 µg/µl) was used.

After immunoprecipitation, washing and DNA purification, ChIP samples were analysed by PCR using Taq DNA polymerase and CoralLoad PCR buffer (QIAGEN Ltd.). To detect binding of c-Jun and ATF2 to the jun1 and jun2 TREs in the rat c-jun promoter the following primer pair was used: 5′-TGGAGAAAGAAGGGCCCAACTGTAG-3′ and 5′-GTGCAACTCTGAGTCCTTATCCAGC-3′.PCR conditions were: 5 min at 94°C followed by 30–35 cycles of 30 s at 94°C, 45 s at 52°C, 1 min at 72°C, followed by 10 min at 72°C. For the dp5 ATF site the following primers were used: 5′-AAGTTACCTCTCGGCTTTTTCTCC-3′ and 5′-CTGGACCCCAAGTTTCGCTC-3′. PCR conditions were the same as those used for c-jun except that annealing was for 45 s at 62°C and Q-Solution (Qiagen Ltd, Crawley, UK) was added to the PCR reactions to facilitate amplification of the GC-rich dp5 promoter fragment. The PCR products were run on non-denaturing 8% polyacrylamide/1× TBE gels and then stained with SYBR Green 1 (Sigma-Aldrich). Images were captured using an UVIdoc gel documentation system (UVItec Ltd) and saved as TIFF files.

Adenovirus infection

Recombinant adenoviruses that express the FLAG-tagged c-Jun dominant negative mutant FLAGΔ169 (AdvJunΔ169) or Escherichia coli β-galactosidase (AdvlacZ) under the control of the CMV promoter were previously described (14). Purified adenovirus preparations were titred by TCID50 assay using HEK 293A cells, and by infecting sympathetic neurons and performing immunocytochemistry with the FLAG-specific M2 monoclonal antibody (Sigma-Aldrich) or a monoclonal antibody against β-galactosidase (Promega) as described (14).

For infection experiments, sympathetic neurons growing on glass coverslips were transferred into the wells of 24-well plates with 1 ml of SCG medium per well. At 5 days in vitro, the medium was replaced with 0.5 ml of medium containing recombinant adenovirus particles at the lowest multiplicity of infection (MOI) that would lead to the expression of the FLAGΔ169 or β-galactosidase protein in ∼50% of infected neurons (14). After overnight infection at 37°C, the virus-containing medium was removed and replaced with 1 ml of fresh SCG medium containing NGF. After 24–36 h, the cells were gently rinsed twice with SCG medium lacking NGF and antimitotic agents and then refed with 1 ml of SCG medium containing NGF or anti-NGF antibody, as appropriate. After a further 16–24 h, the cells were harvested for protein analysis or RNA analysis, as described in the immunoblotting and RT–PCR sections, respectively.

Immunoblotting

Neurons were washed off the coverslips into microfuge tubes using ice cold PBS. After centrifugation, the neurons were lysed in sample buffer (2% SDS, 2 mM β-mercaptoethanol, 60 mM Tris, pH 6.8, 0.01% bromophenol blue) by incubating at 100°C for 15 min. Proteins were separated on 12% SDS polyacrylamide gels and transferred to Immobilon-P (Millipore) using the Bio-Rad Mini-PROTEAN III transfer system. Protein detection was performed as described previously (14) or following protocols supplied with the primary antibodies. The following primary antibodies were used: the M2 anti-FLAG epitope mouse monoclonal antibody (Sigma-Aldrich), a mouse monoclonal anti-β-galactosidase antibody (Promega), a rat monoclonal anti-α-tubulin antibody (Serotec Ltd, Kidlington, UK), a mouse monoclonal anti-c-Jun antibody (BD Transduction Laboratories), a mouse monoclonal phospho-c-Jun (serine 63) antibody (KM-1; Santa Cruz Biotechnology, Inc.), a rabbit polyclonal phospho-c-Jun (serine 73) antibody (9164; Cell Signalling Technology), a rabbit polyclonal anti-ATF2 antibody (C-19; Santa Cruz Biotechnology, Inc.), a rabbit polyclonal phospho-ATF2 (threonine 71) antibody (9221; Cell Signalling Technology), a rabbit polyclonal ERK1/2 antibody (Cell Signalling Technology).

RESULTS

The dp5 promoter, intron and 3′ UTR contain sequences that respond to NGF withdrawal in sympathetic neurons

The rat dp5 gene contains two exons divided by an 18.9 kb intron, of which only a short region towards the 5′ end is conserved between the rat, mouse and human DNA sequences. The Dp5 open reading frame (ORF) is located in exon 1, which also contains part of the 3′ UTR, the remainder of which is contained within exon 2 (Figure 1a).

Figure 1.

The dp5 promoter, intron and 3′ UTR contain sequences that respond to NGF withdrawal in sympathetic neurons. (a) Structure of the rat dp5 gene. The rat dp5 gene consists of two exons and an 18.9 kb intron. The transcriptional start site mapped by 5′ RACE is indicated as +1 (see Figure 3 for DNA sequence). Exon 1 contains the Dp5 ORF (black box) as well as a small region of the 3′ UTR, the remainder of which is in exon 2. A short region (<400 bp long) towards the 5′ end of the intron is conserved between the rat, mouse and human dp5 genes (grey box). (b) dp5 reporter gene structure and induction levels in sympathetic neurons following NGF deprivation. The construct dp5-LUC contains the dp5 promoter sequence from –980 to +37 cloned upstream of the firefly luciferase gene (Luc) and the SV40 poly-A termination signal (striped box). In addition to these sequences, other constructs contained either ∼400 bp from the conserved region of the dp5 intron inserted downstream of the poly-A termination signal (dp5-LUC + I 400), the dp5 3′ UTR inserted downstream of luciferase (dp5-LUC + 3′ UTR), or both the intron fragment and 3′ UTR inserted as above (dp5-LUC + ALL). The induction factor for each reporter construct in sympathetic neurons following NGF withdrawal is shown. (c) Relative luciferase activity of dp5 reporter constructs in sympathetic neurons maintained in the presence and absence of NGF. Sympathetic neurons were microinjected with dp5-LUC (5 ng/µl), dp5-LUC + I 400 (5 ng/µl), dp5-LUC + 3′ UTR (15 ng/µl) or dp5-LUC + ALL (10 ng/µl) together with pRL-TK (5 ng/µl). Following microinjection, cells were rinsed twice with medium lacking NGF and were refed either with medium supplemented with NGF or medium containing neutralising anti-NGF antibody. After 20–24 h luciferase activity was measured. Firefly luciferase activity was normalized as outlined in the Materials and Methods and luciferase induction –NGF (black bars) was calculated relative to +NGF (white bars), which was set as 1. The mean of at least four independent experiments ±SEM is shown.

Figure 3.

Alignment of the promoter sequences for the rat, mouse, human and cow dp5 genes. Shaded regions indicate a conserved ATF site, a GC box, an E box and a TATA box. Asterisks represent bases conserved in all four species. Overall, 80% of the nucleotides are conserved. The transcriptional start site of the rat dp5 gene determined by 5′ RACE is indicated as +1, together with the direction of transcription.

As a first step in studying how dp5 expression is regulated, we used the 5′ RACE technique to map the major dp5 RNA start site in rat brain mRNA. This is 37 bp upstream of the start of the Dp5 ORF and we designated it as +1 (Figures 1a and 3). We then cloned different regions of the rat dp5 gene and constructed a series of dp5-luciferase reporter plasmids (Figure 1b). These contained 1 kb of promoter sequence including the dp5 transcriptional start site upstream of the firefly luciferase gene, either alone or with the entire dp5 3′ UTR and/or ∼400 bp of the conserved intron sequence (Figure 1b). We microinjected the reporter constructs, together with the control Renilla luciferase construct pRL-TK, into the nuclei of sympathetic neurons cultured in vitro. After injection, the cells were maintained for 16–24 h in the presence of NGF or a neutralising anti-NGF antibody, after which time relative luciferase activity was determined by dual luciferase assay (33,34). In each case, the different dp5 reporter constructs were compared to the construct containing only 1 kb of the promoter sequence (dp5-LUC). In microinjected neurons, this construct has a much higher level of luciferase activity, i.e. promoter activity, than the promoterless vector pGL3-Basic (data not shown) and, after NGF withdrawal, it is induced 2.03-fold (±0.14) on average (Figure 1b). This indicates that the 1 kb dp5 promoter fragment contains sequences that respond to NGF withdrawal. Following NGF deprivation, both constructs containing the 1 kb dp5 promoter and either the conserved intron region or the 3′ UTR sequence showed higher levels of luciferase activity than the construct with the promoter alone, with induction factors of 3.61- and 3.23-fold, respectively (Figure 1c). This suggests that these two regions also contain elements that contribute to the increase in dp5 expression after NGF withdrawal. The construct dp5-LUC + ALL, which contained 1 kb of dp5 promoter sequence, the conserved region of the intron, and the 3′ UTR was induced the most after NGF deprivation, 4.81-fold (Figure 1c). Since dp5-LUC + ALL was efficiently activated by NGF withdrawal it was used as the standard dp5 reporter construct in all subsequent experiments.

Inhibition of the MLK-JNK pathway reduces the increase in endogenous dp5 mRNA level and dp5 reporter activity after NGF deprivation

The MLK-JNK protein kinase cascade is activated by NGF withdrawal in sympathetic neurons (9,10,11). Mixed lineage kinases (MLKs) phosphorylate and activate the JNK kinases MKK4 and MKK7, which in turn phosphorylate and activate JNKs (35). MLK activity can be directly inhibited by the compound CEP-1347 (36). Treatment of sympathetic neurons with CEP-1347 blocks NGF withdrawal-induced death (37) and reduces dp5 mRNA induction following NGF deprivation by ∼75% (22). In addition, it has been reported that CEP-1347 reduces the increase in dp5 RNA level in cortical neurons undergoing amyloid-beta-induced apoptosis (38).

To further study the role of the MLK-JNK pathway in regulating dp5 expression we used CEP-11004, a MLK inhibitor closely related to CEP-1347, which also inhibits NGF withdrawal-induced death (39). We investigated the effect of CEP-11004 (at 400 nM) on the level of c-Jun and ATF2 and c-Jun and ATF2 phosphorylation in sympathetic neurons cultured in the presence or absence of NGF for 16 hours (Figure 2a). As previously described (7), c-Jun protein levels and c-Jun N-terminal phosphorylation increased after NGF withdrawal. The increased phosphorylation was associated with a decrease in the mobility of c-Jun in SDS PAGE. These changes were reversed by CEP-11004. Similarly, phosphorylation of c-Jun at serine 63 and serine 73 was significantly increased after NGF withdrawal and this increase was inhibited by CEP-11004. In contrast, ATF2 protein levels did not change after NGF withdrawal and phosphorylation of ATF2 at threonine 71 was only marginally increased at 16 h after the removal of NGF, as previously reported (12). To determine whether CEP-11004 had the same effect on dp5 expression as CEP-1347, we treated sympathetic neurons with CEP-11004 at 400 nM in the presence or absence of NGF for 16 h and then isolated RNA and measured dp5 mRNA levels by semi-quantitative RT–PCR (Figure 2b). Following NGF deprivation, the dp5 mRNA increased in level by 4.27-fold in untreated control cells but this induction was reduced to 2.28-fold in cells treated with 400 nM CEP-11004 (Figure 2b and c). We then investigated the effect of CEP-11004 on the activity of the dp5-LUC + ALL reporter construct in microinjection experiments. The dp5 reporter construct was induced 4.53-fold after NGF withdrawal and this was reduced to a relative induction of 1.76-fold following treatment with 400 nM CEP-11004 (Figure 2d). This result suggests that, like the endogenous dp5 gene, induction of the dp5 reporter construct depends on MLK activity in sympathetic neurons.

Figure 2.

Inhibition of the MLK-JNK pathway reduces the increase in endogenous dp5 mRNA level and the activation of a dp5 reporter construct following NGF withdrawal. (a) CEP-11004 inhibits the increase in c-Jun protein level and c-Jun N-terminal phosphorylation that occurs after NGF withdrawal. Sympathetic neurons were maintained in the presence or absence of NGF and were treated with either DMSO or CEP-11004 at 400 nM for 16 h. The neurons were then harvested, lysed in SDS gel sample buffer and immunoblotting was performed with the antibodies indicated. The level of ERK1/2 was used as a loading control. Three independent experiments were performed and representative immunoblots are shown. (b) CEP-11004 reduces endogenous dp5 mRNA expression after NGF deprivation. Sympathetic neurons were maintained in the presence or absence of NGF and were treated with either DMSO or CEP-11004 at 400 nM for 16 h. RNA was extracted and RT–PCR analysis performed using dp5 and neurofilament (nf-m) specific primers. Representative images of agarose gels are shown. (c) Relative levels of dp5 mRNA in sympathetic neurons treated with CEP-11004 after NGF withdrawal. Following RT–PCR, band intensity was measured and the normalized dp5 level in each condition was calculated relative to the level in cells treated with NGF and DMSO, which was set as 1. The mean of four independent experiments ±SEM is shown. At 400 nM, CEP-11004 significantly reduced the increase in dp5 mRNA level after NGF withdrawal (−NGF+CEP-11004 compared to −NGF). *P < 0.05, Student's t-test. (d) CEP-11004 reduces induction of a dp5 reporter construct after NGF withdrawal in sympathetic neurons. Sympathetic neurons were microinjected with dp5-LUC+ALL (10 ng/µl) and pRL-TK (5 ng/µl). Cells were maintained in the presence or absence of NGF and were treated with either DMSO or 400 nM CEP-11004 for 16 h. Normalized luciferase activity was calculated relative to the level +NGF treated with DMSO, which was set as 1. The mean of three independent experiments ±SEM is shown. At 400 nM, CEP-11004 significantly reduced the increase in the level of luciferase activity for dp5-LUC + ALL after NGF withdrawal (−NGF + CEP-11004 compared to −NGF). #P < 0.02, Student's t-test. (e) Expression of the JIP-1 JBD reduces induction of a dp5 reporter construct after NGF withdrawal. Sympathetic neurons were microinjected with dp5-LUC+ALL (20 ng/µl), pRL-TK (10 ng/µl), and pcDNA3.1 or pCD JIP-1 JBD (100 ng/µl). Cells were maintained +NGF or –NGF for 20 h and then luciferase activity was measured. Luciferase activity was calculated relative to the level +NGF injected with pcDNA3.1, which was set as 1. The mean of six independent experiments ±SEM is shown.

We also investigated the effect of directly inhibiting JNK activity by using the JBD of the scaffold protein JNK interacting protein 1 (JIP-1), which binds to and specifically inhibits JNKs but not other MAP kinases (35). It has previously been shown that expression of the JBD in sympathetic neurons inhibits JNK activity and promotes cell survival following NGF withdrawal (12,13). We microinjected an expression vector for the JBD or the empty vector pcDNA3 into sympathetic neurons together with dp5-LUC + ALL and luciferase activity was measured 20–24 h after NGF withdrawal (Figure 2e). Expression of the JBD strongly reduced the induction of the dp5-LUC + ALL reporter construct after NGF deprivation from 4.31-fold with pcDNA3 to 1.51-fold. This result indicates that JNK activity is required for normal induction of the dp5 reporter construct following NGF withdrawal.

A conserved ATF site in the dp5 promoter binds c-Jun and ATF2 and contributes to dp5 basal promoter activity and induction after NGF withdrawal

The AP-1 family of basic/leucine zipper transcription factors includes the Jun and ATF subfamilies. Both c-Jun and activating transcription factor 2 (ATF2) are phosphorylated by JNKs after NGF deprivation in sympathetic neurons (9,10,12), although, in the case of ATF2, the increase in phosphorylation (at threonine 71) is more transient (12). Importantly, microinjection of a neutralising c-Jun antibody, expression of a c-Jun dominant negative mutant or conditional knockout of the c-jun gene in sympathetic neurons in culture protects the cells from NGF withdrawal-induced death (6,7,15). AP-1 family members form homo- and heterodimers with each other, which depending on the particular dimer, specifically bind to the AP-1 site/TPA responsive element (TRE) (5′-TGA G/C TCA-3′) or the ATF site/cAMP response element (CRE) (5′-TGA CG TCA-3′) (35). The c-jun promoter itself contains two ATF sites, the jun1 and jun2 TREs, that bind heterodimers of c-Jun and ATF2 or ATF2 homodimers, and which are required for c-jun promoter activation following NGF withdrawal in sympathetic neurons (10).

To identify conserved, potential transcription factor binding sites in the dp5 promoter, we aligned the DNA sequence of the rat dp5 promoter from −115 to +24, in relation to the transcriptional start site, with the corresponding regions of the mouse, human and cow dp5 genes (Figure 3). Interestingly, we identified a potential ATF-binding site (5′-TGA TG TAA-3′) at positions −96 to −89 in the rat dp5 promoter, which was conserved between the rat, mouse, human and cow sequences (Figure 3). Closer analysis of this site indicated that it only differs from the ATF/CRE consensus site by 2 bases and it is only 1 base different from the jun2 TRE site in the c-jun promoter (Figure 4a).

Figure 4.

The dp5 promoter contains an ATF site that binds AP-1 proteins in vitro. (a) Sequences of ATF sites. The ATF consensus sequence is shown together with the jun2 TRE sequence, the dp5 ATF site and a mutant dp5 ATF site. (b) In vitro translated proteins used for EMSA experiments. ATF2, c-Fos, c-Jun, c-JunΔ169 and Δ169m0 were transcribed and translated in a rabbit reticulocyte lysate. Unprogrammed lysate (RRL) was used as a negative control. The in vitro translated proteins were labelled with 35S-methionine and separated on a 12% SDS polyacrylamide gel. The size of molecular weight markers run on the same gel is shown. (c) ATF2 homodimers, c-Jun/ATF2 and c-Jun/c-Fos can bind to the dp5 ATF site. 32P-labelled, double-stranded oligonucleotides containing either the collagenase TRE, the jun2 TRE, or the dp5 ATF site were incubated with in vitro translated c-Jun, c-Fos, ATF2, or unprogrammed lysate (RRL), as indicated, and separated on a 5% polyacrylamide, 0.25× TBE gel. Several experiments were performed and a representative result is shown. Free probe, a non-specific complex (NS) and the AP-1 complexes are indicated. (d) Point mutations in the dp5 ATF site abolish the binding of AP-1 proteins in vitro. Oligonucleotides containing either the wild type dp5 ATF site or a mutant site, in which four nucleotides have been changed (a) were used for EMSA experiments with in vitro translated c-Jun, c-Fos, ATF2 or unprogrammed lysate (RRL), as indicated. Several experiments were performed and a representative result is shown.

We tested the ability of representative AP-1 family members to bind to the dp5 ATF site in vitro in an electrophoretic mobility shift assay (EMSA). We translated c-Fos, c-Jun and ATF2 in vitro (Figure 4b). As a control, we tested the binding of these AP-1 proteins to two well characterized binding sites, the collagenase TRE (a consensus AP-1 site) and the jun2 TRE (an ATF site) (Figure 4c). It has been reported that c-Jun homodimers have a very low affinity for both of these sites whereas c-Jun/c-Fos heterodimers bind with high affinity to the collagenase TRE, and c-Jun/ATF2 heterodimers bind with high affinity to the jun2 TRE (31). In agreement with these results, we observed a c-Jun/c-Fos/DNA complex, but no binding of c-Jun or c-Fos alone, with the collagenase TRE (lanes 1–4, Figure 4c), and a c-Jun/ATF2/DNA complex, but no binding of c-Jun alone, with the jun 2 TRE (lanes 5–8, Figure 4c). We then tested the binding of the AP-1 proteins to the dp5 ATF site (lanes 9–14, Figure 4c). Neither c-Jun nor c-Fos alone bound to the dp5 sequence but c-Jun/c-Fos heterodimers bound with a low affinity (compare lane 11 with lane 3). In contrast, c-Jun/ATF2 heterodimers and ATF2 homodimers bound the dp5 site with an affinity similar to the affinity of c-Jun/ATF2 for the jun2 TRE (compare lanes 12 and 13 with lane 7, Figure 4c). Mutation of the dp5 ATF site by the introduction of four base changes (Figure 4a) abolished the binding of all of the AP1 proteins (Figure 4d, lanes 8–14).

To investigate whether c-Jun and ATF2 in sympathetic neuron extracts can bind to the dp5 ATF site we prepared extracts from neurons that had been cultured in the presence or absence of NGF for 16 h and performed an EMSA experiment (Figure 5a). Extracts from neurons maintained in the presence of NGF contained proteins that bound to the dp5 ATF site (lane 2) and this binding (marked AP-1) was abolished by point mutations in the ATF site (lanes 5 and 6). After NGF withdrawal, a similar amount of specific AP-1-binding activity was observed but the pattern of bands was slightly different (compare lanes 2 and 3). To determine whether the specific protein complexes contained c-Jun and ATF2 we added antibodies specific for c-Jun or ATF2 or a Bim antibody, as a negative control, to the binding reactions. The Bim antibody did not alter the binding pattern observed with the –NGF extract (compare lanes 7 and 8) whereas the c-Jun antibody supershifted some of the AP-1 complexes (lane 9). Phosphorimaging and quantitation revealed that the c-Jun antibody supershifted 45% of the AP-1/dp5 complexes. Addition of the ATF2 antibody did not cause a clear supershift but did displace 42% of the AP-1/dp5 complexes (compare lanes 10 and 7) and addition of both the c-Jun and ATF2 antibodies together supershifted 71% of the specific protein complexes bound to the dp5 ATF site (compare lanes 11 and 7). We also performed an EMSA experiment in which we tested the effect of phospho-c-Jun (serine 63) and phospho-ATF2 (threonine 71) antibodies (Figure 5b). With the –NGF extract the phospho-c-Jun antibody supershifted the majority of the AP-1 proteins bound to the dp5 ATF site (compare lane 15 to lanes 14 and 13). We confirmed that this was a supershift by measuring the distribution of peaks in a vertical line through the –NGF lanes in a phosphorimage of the gel shown in Figure 5b. The AP-1 proteins were clearly displaced towards the top of the gel in lane 15 compared to lanes 14 and 13. In contrast, the phospho-ATF2 antibody only supershifted a small fraction of the dp5/AP-1 complex (compare lane 16 to lanes 14 and 13). These results demonstrate that after NGF withdrawal the c-Jun in sympathetic neuron extracts that binds to the dp5 ATF site in vitro is phosphorylated at serine 63. In the case of ATF2, a smaller fraction of the AP-1 complex bound to the dp5 ATF site is phosphorylated at threonine 71. We then compared the amount of phospho-c-Jun supershifted in DNA-binding assays performed with +NGF and –NGF extracts (Figure 5c). Phosphorimaging and quantitation of the supershifted complex in lanes 19 and 20 revealed that the amount of c-Jun phosphorylated at serine 63 was 2-fold greater at 16 h after NGF withdrawal.

Figure 5.

c-Jun and ATF2 in sympathetic neuron extracts bind to the dp5 ATF site in vitro. Sympathetic neurons were cultured for 7 days and then refed with medium containing NGF (+N) or anti-NGF antibody (−N). Sixteen hours later, whole cell extracts were prepared. (a) EMSA experiments were performed using oligonucleotides containing either the wild type dp5 ATF site or mutant site, in which four nucleotides have been changed (Figure 4a). Eight micrograms of +N or –N whole cell extract was used per binding reaction, as indicated. A control antibody (against Bim) or c-Jun or ATF2 antibodies were added as shown. The binding reactions were separated on a 5% polyacrylamide, 0.25× TBE gel. Several experiments were performed and a representative result is shown. Free probe, a non-specific complex (NS) and the AP-1 complexes are indicated. The four point mutations inhibited the binding of AP-1 proteins to the dp5 ATF site. The c-Jun antibody caused a supershift (s s) whereas the ATF2 antibody partially reduced binding of the AP-1 complexes. (b) Effect of phospho-c-Jun (Serine 63) and phospho-ATF2 (Threonine 71) antibodies on the binding of AP-1 proteins to the dp5 ATF site. Only the top of the gel is shown. The positions of the AP-1 complexes and supershifted complexes (s s) are indicated. (c) The amount of c-Jun phosphorylated at serine 63 bound to the dp5 ATF site increases after NGF withdrawal. Only the top of the gel is shown. The positions of the AP-1 complexes and supershifted complexes (s s) are indicated.

To confirm that c-Jun and ATF2 can bind to the region of the dp5 promoter that contains the ATF site in living cells, we performed ChIP assays using the Bim, c-Jun and ATF2 antibodies. Since large numbers of cells are required for conventional ChIP assays we used neuronally differentiated PC12 cells that had been cultured in the presence or absence of NGF for 16 h, rather than sympathetic neurons. The binding of c-Jun or ATF2 to the c-jun and dp5 promoters was studied by PCR using primers that flank the jun1 and jun2 TREs (a positive control) or the dp5 ATF site, respectively (Figure 6a). The control antibody (Bim) did not immunoprecipitate either the c-jun or dp5 promoter (lanes 3 and 4), whereas both the c-Jun and ATF2 antibodies precipitated the region of the c-jun promoter that contains the jun1 and jun2 TREs and the dp5 promoter region that contains the ATF site (lanes 5–8). These results indicate that c-Jun and ATF2 can bind to the dp5 ATF site in living cells. In the case of both c-jun and dp5 the amount of c-Jun or ATF2 bound to the promoter had not increased significantly at 16 h after NGF withdrawal. These results are in agreement with previous studies that showed that the jun1 and jun2 TREs in the c-jun promoter are already bound by c-Jun and ATF2-containing complexes in a variety of unstimulated cells in culture and that the protein-DNA contacts are unchanged during gene activation by TPA and UV (40). Using an antibody specific for c-Jun phosphorylated at serine 63 we also investigated whether the amount of phospho-c-Jun associated with the c-jun and dp5 promoters increased after NGF withdrawal (Figure 6b). We found that the c-Jun bound to the c-jun promoter was phosphorylated at serine 63 and that this increased in level after NGF withdrawal (lanes 14 and 15). Similarly, we also observed that there was increase in the amount of phospho-c-Jun associated with the dp5 promoter at 16 h after NGF deprivation. In the case of both promoters, we found that phosphorylation of ATF2 at threonine 71 had not increased at 16 h after NGF withdrawal (data not shown), consistent with the results of the immunoblotting experiments with sympathetic neuron extracts (Figure 2a).

Figure 6.

c-Jun and ATF2 bind to the dp5 ATF site in the chromatin of neuronally differentiated PC12 cells. The PC6-3 subline of PC12 cells was treated with NGF at 100 ng/ml for 7 days. The neuronally differentiated cells were then refed with fresh differentiation medium containing NGF (+N) or anti-NGF antibody (−N). After 16 h, the cells were cross-linked with 1% formaldehyde and chromatin immunoprecipitations were performed using (a) the c-Jun(H-79) or ATF2(C-19) antibodies or a Bim antibody as a negative control or (b) the phospho-c-Jun (serine 63) (KM-1) or Bim antibodies. As a positive control, PCR was performed using primers that flank the jun1 and jun2 TREs in the c-jun promoter, which have previously been shown to bind c-Jun and ATF2 in chromatin. Binding of c-Jun and ATF2 to the dp5 promoter was detected by performing PCR using primers that flank the dp5 ATF site. The ChIP experiments were performed several times and representative gel images are shown. The positions of the c-jun and dp5 PCR products are indicated. The equivalent of 1% of the +N and –N chromatin used for each ChIP assay was also run on each gel (input lanes). As a negative control a PCR reaction without chromatin was performed (H2O).

We then investigated the effect of introducing the four point mutations shown to abolish AP-1 binding into the construct dp5-LUC + ALL. The wild-type and mutant constructs were microinjected into sympathetic neurons and luciferase activity was measured after 20–24 h in the presence or absence of NGF (Figure 7). The mutations in the ATF site in the dp5 promoter completely obliterated its induction after NGF withdrawal by (i) decreasing the basal activity of the construct by 86% (Figure 7a), and (ii) by reducing the induction factor after NGF withdrawal from 7.08- to 3.05-fold (Figure 7b). This demonstrates that abolishing the binding of c-Jun and ATF2 to the dp5 ATF site not only reduces the induction of dp5 expression following NGF withdrawal but also decreases basal promoter activity. Finally, we investigated the effect of the MLK inhibitor CEP-11004 (at 400 nM) on the activity of the mutant construct (Figure 7c). CEP-11004 did not significantly reduce the 3-fold increase in the activity of the reporter with mutated ATF site that occurs after NGF withdrawal (Figure 7c). This is different to the effect of CEP-11004 on the wild type dp5-LUC + ALL construct (Figure 2d) and suggests that the 3-fold induction observed in the absence of the ATF site is independent of the MLK-JNK pathway.

Figure 7.

Mutation of an ATF-binding site in the dp5 promoter reduces basal promoter activity and induction after NGF withdrawal. Sympathetic neurons were microinjected with dp5-LUC + ALL or a reporter construct in which the ATF-binding site in the dp5 promoter had been mutated (dp5-LUC + ALLmut) (both at 10 ng/µl) as well as pRL-TK (5 ng/µl). Cells were maintained in medium containing or lacking NGF for 20 h, after which time luciferase activity was measured. (a) Normalized firefly luciferase activity was calculated relative to the level +NGF microinjected with the wild-type construct, which was set as 1. The mean of six independent experiments ±SEM is shown. (b) The induction factor for the same series of experiments was calculated for each reporter construct = (normalized firefly luciferase activity –NGF)/(normalized firefly luciferase activity +NGF). (c) CEP-11‱004 does not reduce the activity of the reporter with mutated ATF site after NGF withdrawal. The reporter construct in which the ATF-binding site in the dp5 promoter had been mutated (dp5-LUC + ALLmut) (10 ng/µl) was microinjected into sympathetic neurons together with pRL-TK (5 ng/µl). Cells were maintained in medium containing or lacking NGF with either DMSO or 400 nM CEP-11‱004 as indicated for 20 h, after which time luciferase activity was measured. Normalized firefly luciferase activity was calculated relative to the level +NGF, which was set as 1. The mean of six independent experiments ± SEM is shown.

Dominant negative c-Jun reduces the increase in dp5 reporter activity and the increase in the level of endogenous dp5 RNA after NGF withdrawal

Expression of a c-Jun dominant negative mutant (JunΔ169) in sympathetic neurons protects the cells against NGF withdrawal-induced death (7,14) and reduces the increase in the level of the c-Jun and BimEL proteins that normally occurs after the removal of NGF (10,14). We therefore investigated the effect of dominant negative c-Jun on dp5 reporter activity and the endogenous dp5 gene in sympathetic neurons.

First, using in vitro translated proteins we studied how JunΔ169 binds to the dp5 ATF site in an EMSA experiment (Figure 8a). Like c-Jun, JunΔ169 was unable to bind to the dp5 ATF site on its own, whereas ATF2 homodimers were able to do so (lanes 1–5, Figure 8a). When c-Jun was mixed with ATF2, c-Jun/ATF2 heterodimers formed a protein/DNA complex that ran just below the ATF2/DNA complex (compare lanes 6 and 5). JunΔ169 also formed heterodimers with ATF2 that could bind to the dp5 ATF site (lane 7). However, when the m0 leucine zipper mutation (31), which blocks binding to ATF2, was present in JunΔ169 (Δ169m0), the dominant negative c-Jun/ATF2/DNA complex was no longer observed (lane 8). Thus, JunΔ169 can only bind to the dp5 ATF site as a heterodimer with ATF2. We then investigated the effect of dominant negative c-Jun on the dp5-LUC + ALL reporter construct in a co-microinjection experiment (Figure 8b). We injected an expression vector for JunΔ169 or the empty vector pcDNA1 into sympathetic neurons together with dp5-LUC + ALL and luciferase activity was measured 20–24 h after NGF withdrawal. Expression of JunΔ169 reduced the induction of the dp5-LUC + ALL reporter construct after NGF withdrawal from 2.275 ± 0.322-fold with pcDNA1 (P < 0.002 when –NGF is compared to +NGF) to 1.492 ± 0.29-fold, which was not a significant induction (P > 0.1) when –NGF was compared to +NGF. This result suggests that AP-1 activity is required for normal induction of the dp5 reporter construct following NGF deprivation.

Figure 8.

Expression of dominant negative c-Jun reduces the induction of a dp5 reporter construct and the endogenous dp5 RNA after NGF withdrawal. (a) c-JunΔ169 binds to the dp5 ATF site as a heterodimer with ATF2. An oligonucleotide containing the dp5 ATF site was incubated with in vitro translated c-Jun, c-JunΔ169, Δ169m0, ATF2 or unprogrammed rabbit reticulocyte lysate (RRL) as indicated, and an EMSA experiment was performed. Only the regions of the gel that contain the protein/DNA complexes and the unbound probe are shown, to make the figure more compact. (b) Co-injection of an expression vector for dominant negative c-Jun (JunΔ169) reduces induction of a dp5 reporter construct after NGF withdrawal. Sympathetic neurons were microinjected with dp5-LUC + ALL (20 ng/µl), pRL-TK (10 ng/µl) and pcDNA1 or pCDJunΔ169 (100 ng/µl). Cells were maintained +NGF or –NGF for ∼20 h and then luciferase activity was measured. Firefly luciferase levels were normalized and luciferase activity was calculated relative to the level +NGF, which was set as 1, for either pcDNA1 or pCDFLAGΔ169. The mean of nine independent experiments ±SEM is shown. In the case of empty vector the increase in reporter gene activity after NGF withdrawal was significant: *P < 0.002, Student's t-test. For pCDJunΔ169 the increase in reporter activity after NGF deprivation was not significant: #P > 0.1, Student's t-test. (c) Sympathetic neurons infected with AdvJunΔ169 or AdvlacZ express recombinant proteins of the predicted size. Sympathetic neurons were cultured for 5 days in vitro and then infected with recombinant adenoviruses as indicated. After overnight infection, the cells were refed with fresh SCG medium containing NGF. Forty eight hours later, protein extracts were prepared and immunoblots performed with antibodies to β-galactosidase, α-tubulin (a loading control) and the FLAG epitope. The positions of protein molecular weight markers run on the same 12% gel are shown. (d) RT–PCR analysis of dp5 and nfm mRNA levels in adenovirus-infected neurons expressing β-galactosidase or the JunΔ169 protein. Sympathetic neurons were cultured for 5 days in vitro and then infected with the recombinant adenoviruses indicated. After overnight infection, the cells were refed with fresh SCG medium containing NGF (+NGF) or anti-NGF antibody (-NGF). Sixteen hours later, RNA was isolated and semi-quantitative RT–PCR performed with primers specific for dp5 or nfm. Thirty-five cycles of PCR were performed and the products were run on a 2.5% agarose gel. (e) AdvJunΔ169 reduces the increase in dp5 RNA levels after NGF withdrawal. RT–PCR analysis of dp5 and nfm mRNA levels in adenovirus-infected neurons was carried out as described in (d). dp5 mRNA levels were normalized to nfm RNA levels to control for any differences in the amount of cDNA used for each RT–PCR reaction. The induction factor (the normalized dp5 RNA level –NGF/+NGF) was then calculated for each virus. The average of five experiments ± SEM is shown.

To determine whether AP-1 activity contributes to the induction of the endogenous dp5 mRNA after NGF withdrawal we used a recombinant adenovirus that expresses dominant negative c-Jun (AdvJunΔ169) (14) and, as a control, an adenovirus that expresses E. coli β-galactosidase (AdvlacZ). In an immunoblotting experiment with protein extracts prepared from uninfected sympathetic neurons and adenovirus-infected cells we confirmed that AdvJunΔ169 and AdvlacZ express recombinant proteins with the predicted molecular weights (Figure 8c). We then tested the effect of the two adenoviruses on the level of dp5 mRNA (Figure 8d and e). Sympathetic neurons were cultured for 5 days in vitro and then infected overnight with the JunΔ169 or β-galactosidase adenovirus at MOIs that led to expression of the recombinant proteins in ∼50% of the infected neurons (14). After infection, the cells were refed with +NGF medium and ∼24 h later were rinsed and maintained in medium containing NGF or a neutralising anti-NGF antibody for 16 h. RNA was then isolated and semi-quantitative RT–PCR was performed to measure dp5 and nfm mRNA levels (Figure 8d). Expression of JunΔ169 did not alter the level of the nfm mRNA (Figure 8d). In the case of sympathetic neurons infected with AdvlacZ, the dp5 mRNA increased 6.02-fold after NGF withdrawal, but in cells infected with AdvJunΔ169 the induction was reduced to 2.34-fold (Figure 8e). This result suggests that in sympathetic neurons AP-1 activity contributes to the induction of the endogenous dp5 mRNA following NGF withdrawal.

Microinjection of an antibody against c-Jun, but not Jun B or Jun D antibodies or control IgG, has been shown to inhibit the NGF withdrawal-induced death of sympathetic neurons (6). To study the individual contributions of c-Jun and ATF2 to the activation of dp5-LUC + ALL after NGF withdrawal we therefore carried out an antibody co-injection experiment (Figure 9a). The dp5 reporter construct was injected into the nuclei of sympathetic neurons together with control rabbit immunoglobulin or the c-Jun or ATF2 antibodies that had been used for chromatin immunoprecipitation. With the control antibody, dp5-LUC + ALL was activated 5.3-fold on average after NGF withdrawal whereas co-injection of the c-Jun or ATF2 antibodies reduced the induction factor to 2.83- or 2.47-fold, respectively. At the concentration tested, neither of these antibodies affected basal promoter activity in the presence of NGF. Importantly, the c-Jun and ATF2 antibodies did not affect the activation of pCRE-Luc, a CREB reporter construct, when CPTcAMP, a membrane permeable cAMP analogue, was added to the cells to activate protein kinase A and CREB (Figure 9b).

Figure 9.

Co-injection of antibodies specific for c-Jun or ATF2 reduces the induction of a dp5 reporter gene after NGF withdrawal. (a) Sympathetic neurons were cultured for 6 days in vitro and then microinjected with dp5-LUC + ALL (20 ng/µl), pRL-TK (10 ng/µl) and the c-Jun (H-79) or ATF2 (C-19) antibodies or rabbit immunoglobulin as a control (each at 1 µg/µl) as indicated. After injection, the cells were maintained in medium containing NGF (+) or anti-NGF antibody (−) for 20 h before luciferase activity was measured. For each experiment the level of normalized luciferase activity (firefly output/Renilla output) obtained for neurons injected with dp5-LUC + ALL + control IgG and maintained in medium containing NGF was set as 1 and other values were calculated relative to this (relative luciferase activity). The mean ± SEM for five independent experiments is shown, and induction factors (−NGF/+NGF) are given above the graph. Student's t-test was used to determine whether inductions were significant: *P < 0.05; #P < 0.02; +P > 0.1. (b) Sympathetic neurons were microinjected with pLuc-MCS or pCRE-Luc at 20 ng/µl, pRL-TK (10 ng/µl) and the indicated antibodies (1 µg/µl). After injection the cells were treated with 500 µM CPTcAMP to activate PKA and CREB. Normalized luciferase activity was calculated relative to pLuc-MCS, which lacks CRE sites and which was set as 0.1. The mean ± SEM for four independent experiments is shown. (c) Expression of c-Jun (ala) significantly reduces the induction of the dp5 promoter after NGF withdrawal. Dp5-LUC+ALL (20 ng/µl), pRL-TK (10 ng/µl) and pcDNA3 or CMV c-Jun (ala) (100 ng/µl) were microinjected into sympathetic neurons as indicated. The cells were then maintained in medium containing NGF (+) or anti-NGF antibody (−) for 20 h before luciferase activity was measured. Firefly luciferase levels were normalized and luciferase activity was calculated relative to the level +NGF, which was set as 1, for either pcDNA3 or CMV c-Jun (ala). The mean of seven independent experiments ± SEM is shown. *P < 0.05 when Dp5-LUC+ALL + CMV c-Jun (ala) –NGF is compared to Dp5-LUC + ALL + pcDNA3 –NGF in Student's t-test.

Finally, we investigated the effect of co-injecting an expression vector for c-Jun (ala) with dp5-LUC + ALL (Figure 9c). c-Jun (ala) is a mutant of c-Jun in which all of the potential JNK phosphorylation sites (serines 63 and 73 and threonines 91 and 93) have been mutated to alanine (28). Expression of c-Jun (ala) significantly reduced the induction factor after NGF withdrawal from 5.47 ± 1.35-fold with pcDNA3 to 2.87 ± 0.72-fold. These results suggest that following NGF withdrawal both c-Jun and ATF2 contribute to the activation of the dp5 promoter in dp5-LUC+ALL and that the JNK phosphorylation sites in c-Jun are necessary for full induction of the dp5 promoter.

DISCUSSION

In sympathetic neurons the level of the dp5 RNA increases substantially after NGF withdrawal, and this depends on the MLK-JNK pathway (21,22). Here, we have investigated the mechanism by which this protein kinase cascade regulates dp5 expression, and propose a model by which this occurs (Figure 10). To identify regulatory sequences in the dp5 gene we made reporter constructs containing different regions of dp5 linked to luciferase. In microinjection experiments with sympathetic neurons, we found that constructs containing 1 kb of promoter sequence alone or with a short sequence from the intron and/or the whole 3′ UTR were all activated following NGF deprivation. The highest level of induction was observed when all three regions were present together in the dp5-LUC+ALL construct. The increase in dp5 RNA level after NGF deprivation was reduced by treating cells with the MLK inhibitor CEP-11004, confirming the results of Harris and Johnson (22), who used the related compound CEP-1347. Not only did CEP-11004 decrease dp5 expression after NGF withdrawal, but it also reduced the induction of a dp5 luciferase reporter plasmid, suggesting that this construct behaves in a similar manner to the endogenous dp5 gene. We also co-injected sympathetic neurons with the dp5 reporter construct and an expression vector for the JIP-1 JBD. This JNK inhibitor protein strongly reduced the induction of dp5-LUC + ALL after NGF withdrawal, suggesting that the activation of dp5 transcription by MLKs is mediated by JNKs.

Figure 10.

Proposed model showing the mechanism by which the JNK pathway contributes to increased expression of Dp5 after NGF withdrawal. Following NGF withdrawal, the MLK-MKK4/7-JNK pathway is activated leading to an increase in JNK activity and increased phosphorylation and increased expression of the AP-1 transcription factor c-Jun, which can form heterodimers with ATF2 (6–13). The dp5 promoter contains a conserved ATF-binding site (5′-TGATGTAA-3′), which is bound by c-Jun and ATF2 in vitro and in chromatin. These AP-1 proteins contribute to increased transcription of the dp5 promoter after NGF withdrawal, which depends on JNK and AP-1 activity. Mutation of the ATF site, expression of JunΔ169 or the JIP1-JBD, or treatment of sympathetic neurons with the MLK inhibitor CEP-11‱004 all reduce induction of the dp5 promoter after NGF withdrawal. Together with the other BH3-only proteins induced by NGF withdrawal (BimEL and Puma) the Dp5 protein may promote cell death by binding to antiapoptotic members of the Bcl-2 family, such as Bcl-2 and Bcl-xL (25,26).

We identified a conserved ATF-binding site in the dp5 promoter that is similar to the jun2 TRE in the c-jun promoter. We showed that this site is able to bind c-Jun and ATF2 in vitro and in chromatin, and to study its role in the dp5 promoter we mutated it in a dp5 reporter construct so as to abolish AP-1 binding. This reduced the induction of the reporter plasmid after NGF withdrawal and significantly reduced basal promoter activity suggesting that this site is an important dp5 promoter element. In sympathetic neurons cultured in the presence of NGF, the dp5 ATF site may contribute to basal promoter activity by binding c-Jun/ATF2 heterodimers. After NGF withdrawal, the level of active, phosphorylated c-Jun present in the nucleus increases (9,10) and this would lead to an increase in the activity of c-Jun/ATF2 heterodimers, which could contribute to the induction of dp5 transcription.

To further study the role of AP-1 in the induction of dp5 transcription we tested the effect of overexpressing a c-Jun dominant negative mutant, JunΔ169, that can form heterodimers with ATF2 and bind to ATF sites, but which lacks the first 168 amino acids of c-Jun including the N-terminal transactivation domain, the JNK docking site and the JNK phosphorylation sites. We found that JunΔ169 reduced the induction of a dp5 reporter construct and also decreased endogenous dp5 RNA and protein levels after NGF withdrawal suggesting that dp5 is an AP-1 target gene and that the MLK-JNK pathway activates dp5 transcription via the ATF site (Figure 10). In addition, co-microinjection of either a c-Jun or ATF2 antibody together with dp5-LUC + ALL reduced the activation of the reporter construct after NGF withdrawal suggesting that both c-Jun and ATF2 contribute to the activation of the dp5 promoter (Figure 9).

The expression of Dp5 increases in a range of neural cell types following injury, stress or survival factor withdrawal, and several studies suggest that JNK may regulate Dp5 levels in response to a variety of apoptotic stimuli. dp5 mRNA levels increase in cultured rat cortical neurons following treatment with β-amyloid peptide (41) and this increase can be blocked by CEP-1347 (38). Work carried out with oligodendrocytes also suggests that the β-amyloid-induced upregulation of dp5 and cell death is mediated by the JNK pathway since treatment with β-amyloid peptide increased JNK phosphorylation and AP-1 DNA-binding activity (42). Recently, Ma et al. (43) reported that expression of shRNAs against dp5 can partially protect rat cerebellar granule neurons (CGNs) against apoptosis induced by KCl deprivation. In addition, these authors also demonstrated that the ATF site in the dp5 promoter is bound by c-Jun in CGNs and important for dp5 promoter activation following survival signal withdrawal, in agreement with the results reported here.

Since the discovery that NGF withdrawal-induced death requires de novo gene expression (3) our understanding of the events that occur in sympathetic neurons following NGF deprivation has greatly increased and a number of the genes involved in this process have been identified. The MLK-JNK-c-Jun pathway is one of the important proapoptotic signalling pathways activated after NGF withdrawal (Figure 10). Activation of the MLK-JNK pathway leads to the phosphorylation of c-Jun bound to the c-jun promoter, which increases the rate of transcription of the c-jun gene and increases the level of c-Jun, an important JNK substrate which is required for the normal NGF withdrawal-induced death of sympathetic neurons (15). c-Jun contributes to the increased expression of the BH3-only proteins BimEL (14,24) and Dp5 (43, and this study), which together with Puma, bind to antiapoptotic Bcl-2 family members and promote mitochondrial outer membrane permeabilisation. In addition, JNKs can regulate BimEL at the post-translational level by phosphorylating serine 65 and potentiating its proapoptotic activity (44). Therefore in the future it will be important to identify other transcriptional targets of the JNK pathway as well as further investigate the post-translational effects of JNK phosphorylation.

FUNDING

Wellcome Trust (Senior Research Fellowship in Basic Biomedical Science 057700 to J.H.). Funding for open access charge: The Wellcome Trust.

Conflict of interest statement. None declared.