Báo cáo khoa học: Phosphorylation of NF-jB proteins by cyclic GMP-dependent kinase A noncanonical pathway to NF-jB activation potx

Phosphorylation of NF-jB proteins by cyclic GMP-dependent kinase A noncanonical pathway to NF-jB activation Bin He 1 and Georg F. Weber 1,2 1 Department of Radiation Oncology, New England Medical Center, Boston, MA, USA; 2 Immunology Program, Sackler School of Graduate Biomedical Research, Tufts University Medical School, Boston, MA, USA The transcription factor NF-jB is activated in cellular stress responses. This requires rapid regulation of its function, which is accomplished, in part, by various modes of phos- phorylation. Even though diverse DNA binding subunits of NF-jB proteins may transactivate from distinct recognition sequences, the differential regulation of transcription from the large number of NF-jB responsive sites in various gene promoters and enhancers has been incompletely understood. The cyclic GMP-dependent kinase (PKG) is an important mediator of signal transduction that may induce gene expression through cAMP response element binding protein (CREB) and through other, yet undefined, mechanisms. We have previously characterized a signal transduction pathway that leads to activation-induced cell death in T-lymphocytes and involves the activation of PKG. Here we demonstrate that the NF-jB proteins p65, p49 (also called p52), and p50 are specific substrates for this kinase. PKG dose-dependently increases the transactivating activity of p65 from the NF-jB consensus sequence. It also mediates dose-dependently an increase in transcriptional activity by p49 or p50 from a unique CCAAT/enhance binding protein (C/EBP)-associ- ated NF-jB site, but not from the consensus site. Phos- phorylation of p65, p50, or p49 does not alter their subcellular distribution. Because the release of cytosolic p65/p50 heterodimers into the nucleus is by itself insufficient to differentiate all the numerous NF-jBpromoter sequences, phosphorylation of the DNA-binding subunits reveals a form of differential regulation of NF-jB activity and it implies a novel pathway for PKG-induced gene transcription. These observations may bear on mechanisms of programmed cell death in T-lymphocytes. They may also be relevant to ongoing efforts to induce cancer cell apoptosis through activation of PKG. Keywords: protein kinases; signal transduction; superanti- gens; transcription factors. The transcription factor NF-jB [1,2] mediates a wide range of cellular stress responses. It induces the gene expression of cytokines, acute phase proteins, and adhesion molecules. In T-lymphocytes, NF-jB regulates activation and activa- tion-induced cell death [3] and it contributes to T-cell selection at the double positive stage [4,5]. A large number of gene promoters and enhancers contain NF-jB binding sites. This poses the question how selectivity of NF-jB- dependent transactivation is accomplished during specific modes of lymphocyte activation. While the release of p65/p50 into the nucleus is insufficient to differentiate among diverse binding sites it is clear that there are mechanisms, which control the interactions of NF-jBwith its various recognition sequences. The biologic roles of NF-jB necessitate a rapid activation of the preformed cytosolic complex, which is typically accomplished by phosphorylation. NF-jB phosphorylation is functionally relevant on three levels. Firstly, it targets the inhibitor I-jB for degradation. Even though phosphoryla- tion of I-jB is not sufficient to dissociate the complexes of I-jBandNF-jB in vivo, phosphorylation on serines 32 or 36 is a prerequisite for I-jB degradation in the ubiquitin- proteasome pathway [6]. Secondly, phosphate residues contribute to the processing of the p100 and p105 NF-jB precursor proteins for p50 and p49/p52. A prerequisite for the proteolytic cleavage of p105 is the phosphorylation of serines 894 and 908, which are potential recognition sites for proline-directed serine/threonine kinases, including cyclin- dependent kinases and Erk2 kinase [7]. The phosphoryla- tion of the C-terminal region of p105 may be mediated by cyclic AMP-dependent protein kinase or protein kinase C [8]. Finally, direct phosphorylation of the DNA-binding NF-jB subunits can regulate their functions. Phosphoryla- tion of the DNA-binding NF-jB subunits may modulate DNA binding affinity, transactivation, or the interaction with other regulatory proteins. A PKA recognition sequence within the Rel homology domain, which contains DNA binding sites and nuclear localization signals, has been linked to the transformation of avian spleen cells by v-rel and to the cytoplasmic retention of c-Rel in chicken embryo fibroblasts [9]. PKA activates NF-jB in a manner that is independent of I-jB phosphorylation, does not impair NF-jB interaction with unmodified p105, and induces transactivation [8,10]. This is accomplished through the recruitment of CREB-binding protein (CBP)/p300 by Correspondence to G. F. Weber, Department of Radiation and Cancer Biology, New England Medical Center, NEMC #824, 750 Washington Street, Boston, MA 02111, USA. Fax: 617 636 1766, Tel.: 617 636 9013, E-mail: gweber@tufts-nemc.org Abbreviations: PKA, cyclic AMP-dependent kinase; PKG, cyclic GMP-dependent kinase. (Received 20 November 2002, revised 19 February 2003, accepted 14 March 2003) Eur. J. Biochem. 270, 2174–2185 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03574.x phosphorylated p65 [11–13]. Some investigators have not found p65, p50, or p52 to be phosphorylated by PKA or PKC [8,14]. While p65 and p50 have candidate recognition sequences for PKA or PKG, p49/p52 does not have such a consensus site [9,15], however, related sequences are found in all three molecules. The enzyme cyclic GMP-dependent kinase (cGK, PKG) is an important mediator of intracellular signal transduc- tion, involved in such diverse processes as the regulation of blood vessel tone, platelet aggregation, and long-term potentiation in memory formation [16–18]. Furthermore, PKG may regulate apoptosis positively [3,19–22] or negat- ively [23], possibly depending on other modulating bio- chemical events [24]. Efforts have been made to induce cancer cell apoptosis with sulindac sulfone (exisulind, Aptosyn), which activates PKG [25–27]. The down-regula- tion of PKG levels in neoplastic ovarian epithelial cells [28] may contribute to enhanced tumor cell survival. Although PKG, like cyclic AMP-dependent kinase (PKA), can phosphorylate and activate the transcription factor CREB, the connection of PKG to gene expression is incompletely understood. Recently, the localization of the PKG form I in the cytosol and the nucleus was reported [29,30], suggesting a broader role for PKG in the regulation of gene transcrip- tion. Our previous studies [3] suggested that NF-jBmaybe activated by PKG. Here we demonstrate that p49, p50, and p65 are substrates for the kinase and we analyze the mechanisms by which PKG induces NF-jB activation. Materials and methods Reagents Recombinant human NF-jB p49 and p50 were obtained from Promega. The a-isozyme of cGMP-dependent protein kinase, purified from bovine lung or recombinant bovine, was purchased from Promega or Calbiochem. The following oligonucleotides were utilized in gel shift reactions after radiolabeling with T4 polynucleotide kinase: NF-jB consensus (Promega, sense 5¢-AGTTGAGGGGA CTTTCCCAGGC-3¢), OCT1 (Promega, sense 5¢-TGTCG AATGCAAATCACTAGAA-3¢), H2K (sense 5¢-GGATC CCGGTCGGGGGATTCCCCATCTCGG-3¢), j enhan- cer (sense 5¢-AGCAGAGGGGACTTTCCGAGGC-3¢). The custom made oligonucleotides were obtained as single stranded and were annealed to double stranded probes after phosphorylation with T4 polynucleotide kinase and [c- 32 P]ATP. Double-stranded poly(dI-dC).poly(dI-dC) was purchased from Pharmacia. The cGMP-dependent kinase inhibitor Rp-8-pCPT-cGMPS was purchased from Biolog. The reporter constructs used in this study included a commercial NF-jB luciferase reporter (pNF-jB-luc, Clontech) that contains four consensus NF-jBsites.A luciferase reporter containing the C-reactive protein promoter (pC/EBP-luc) and a relevant control with a mutated NF-jB p50 binding site (pC/EBP-mP50-luc) were generously provided by D. Samols (Dept Biochem- istry, Case Western Reserve University, Cleveland, OH, USA). The plasmids containing human p49, p50, and p65 were obtained from the NIH AIDS Reagent Repository. J. Stavnezer generously provided the murine p50 DNA. Kinase reaction The enzymatic activity of cGMP-dependent kinase was analyzed in kinase reaction buffer (250 m M Mes, pH 6.9, 2m M EGTA, 5 m M magnesium acetate, 50 m M NaCl, 10 mgÆmL )1 BSA, 100 m M dithiothreitol, 2 m M protein kinase A inhibitor peptide) with 1 mgÆmL )1 Kemptide substrate (LRRASLG) and 1 m M [c- 32 P]ATP (30–40 c.p.m.Æpmol )1 ). The reaction was performed in the presence or absence of 200 l M cyclic GMP at room temperature for 3 min. The reaction was terminated by spotting 50 lL onto Whatman P-81 filter paper and immediate immers- ionin10mL75m M H 3 PO 4 for 2 min. This was followed by five washes in 10 mL 75 m M H 3 PO 4 , air drying, and scintillation counting. For the analysis of PKG activity in cell lyzates, the PKA inhibitor peptide (Sigma) was present. Phosphorylation of PKG substrates (232 ngÆlL )1 for recombinant NF-jB p49 and p50; 500 ngÆlL )1 for p65) was performed in 100 m M Tris/HCl,pH7.5,20m M sodium chloride, 10 m M dithiothreitol, 2 m M magnesium acetate, 200 l M ATP, at room temperature for 15 min. The kinase concentration was 3.5 UÆlL )1 , while cyclic GMP was present at 200 l M . Mops buffer was not used because it is incompatible with the electrophoretic mobi- lity shift assay after transfer. There was no loss of PKG enzymatic activity in Tris buffer at the indicated concentration. 293T cells (1 · 10 6 per 100 mm diameter Petri dish) were transiently transfected with 0.3 lg pFLAG-p49 or pRSV- p65 with CaCl 2 . Twenty-four hours after transfection, the cells were lyzed in 0.5 mL RIPA buffer (50 m M Tris/HCl pH 7.5, 150 m M NaCl, 1% NP-40, 0.5% Na-deoxycholate, 0.1% sodium dodecyl sulfate) and precleared with 30 lL 20% protein A agarose beads overnight. One microgram of anti-FLAG (mouse IgG) or anti-p65 Ig (rabbit polyclonal IgG) were added for 2 h at 4 °C followed by pulling-down with 30 lL 20% protein A agarose beads for an additional 1 h. The agarose beads were pelleted at 14 000 g for 1 min and washed four times in cold RIPA buffer, then twice in detergent-free buffer (50 m M Tris/HCl pH 7.5, 150 m M NaCl). Kinase reaction buffer was directly added to the pelleted beads with or without PKG and cGMP as indicated plus 1 lL[c- 32 P]ATP for 15 min at room temperature. The reaction solutions were then resolved on 8% reducing denaturing SDS-polyacrylamide gel and transferred to poly(vinylidene difluoride) (PVDF) membranes for auto- radiographic exposure and Western blotting. Electrophoretic mobility shift assay DNA binding was assessed by electrophoretic mobility shift according to standard protocols. The reaction mixture contained 10 m M Tris/HCl, pH 7.5, 1 m M EDTA, pH 8.0, 35 m M NaCl, 50 lgÆmL )1 poly(dI-dC)Æpoly(dI-dC), and 5% glycerol plus labeled probe. The DNA binding proteins were transferred from the kinase reaction mixture. The reaction was incubated at room temperature for 20 min before separation on a native 4% polyacrylamide gel. For supershift, the appropriate antibodies (0.05– 0.1 lg) were incubated with the nuclear extracts for 10 min at room temperature before adding DNA binding buffer. Ó FEBS 2003 Phosphorylation of NF-jB by PKG (Eur. J. Biochem. 270) 2175 Cloning of relevant gene products To clone mouse PKG Ia cDNA, frozen mouse kidney tissue was homogenized and total RNA was isolated by using RNeasy mini kit from Qiagen (Valencia, CA, USA) following the manufacturer’s protocol. One microgram of total RNA was used for cDNA synthesis with Super- script II RNase H – reverse transcriptase (Gibco BRL, USA). The coding sequence of PKG Ia was amplified with the primers 5¢-AGCATGGGCACCCTGCGGGAT TTA-3¢ and 5¢-ATTAGAAGTCTATGTCCCAGCCTGA GTTG-3¢. The amplified product was cloned into the vector pCR3.1 (Invitrogen Carlsbad, CA) followed by subcloning into the vector pEF6/His B (Invitrogen, Carlsbad, CA). Sequence fidelity and accurate reading frame were verified by DNA sequencing analysis. Targeted mutations in p65 were generated in positions 276 and 305 by PCR cloning with the Quickchange site directed mutagenesis kit (Stratagene) according to the protocol by the manufacturer. The sense mutagenic oligo- nucleotides used were 5¢-GCGGCGGCCTGCCGACCGG GAGCTCAGT-3¢ for S276A and 5¢-AAACGTAAAAG GGCATATGAGACCTTCAAGAGCATC-3¢ for T305A (mutations in bold). The accuracy of the mutations was confirmed by DNA sequencing. p49 (obtained from the NIH AIDS Reagent Repository) was Flag-tagged at the 5¢-end by PCR using the pri- mers 5¢-CTGCAGCATGGACTACAAGGACGACGA TGACAAGGAGAGTTGCTACAACCCAGGTCTG-3¢ and 5¢-GAGAGTTGCTACAACCCAGGTCTG-3¢ with pRSV-p49 as a template. The amplified fragment was cloned into the vector pCR3.1 and sequence fidelity was confirmed by DNA sequencing. Reporter gene assays 293T cells were plated at 1 · 10 6 cells per 100 mm diameter Petri dish and were grown for 24 h before transfection with CaCl 2 . The commercial pNF-jB-luc reporter (Clontech, Palo Alto, CA, USA) contains four NF-jB response elements and was used at 0.5 lgper transfection (to amplify the signal for the analysis of transactivation by endogenous p65, we used 2 lgof reporter DNA). The common internal transfection standard Renilla in pRL-SV40 (10 ng per transfection) served as a control for transfection efficiency. This reporter construct was not sensitive to cotransfection of PKG or NF-jB and provided stable reference values. Renilla was not used in transfection experiments with the noncommercial reporter constructs because the lumines- cence intensity is too high compared to the specific readout. Twenty-four hours after transfection, the cells were harvested in 1 mL reporter lysis buffer (Promega) and dual luciferase reporter assays were performed following the protocol provided by the manufacturer. Lyzates were diluted 1 : 40 and 10 lLwereusedfor measurement in a luminometer (Turner Designs TD-20/20). In reporter gene experiments without Renilla, lumine- scene was measured in 40 lg (total protein) of lyzate. The protein concentrations were determined by the BCA protein assay reagent kit (Pierce). As confirmation of protein expression, 20 lg of the same lyzates were also used for separation on 8% SDS-polyacrylamide gels followed by Western blotting on PVDF membranes. PKG kinase activity in the lyzates was confirmed by phosphorylation of the standard substrate LRRASLG (ÔKemptideÕ)where indicated. Western blotting Cells were lyzed in RIPA (50 m M Tris/HCl pH 7.5, 150 m M NaCl, 1% NP-40, 0.5% Na-deoxycholate, 0.1% sodium dodecyl sulfate) or NTEN buffer (20 m M Tris/ HCl, pH 8.0, 120 m M NaCl, 0.5% NP40) containing 1m M phenylmethanesulfonyl fluoride, 10 lgÆmL )1 pep- statin, and 1 m M dithiothreitol. The lyzates were centri- fuged at 15 000 g for 5 min and the protein concentration was determined in the supernatants. Twenty micrograms of total protein were resolved on reducing denaturing SDS/polyacrylamide gels and trans- ferred to PVDF membranes. The membranes were probed with appropriate antibodies followed by horse- radish peroxidase-conjugated secondary antibodies and development using enhanced chemiluminescence. Anti- p65 Ig (C-terminus, rabbit polyclonal), anti-p50 Ig (rabbit antiserum), and anti-p52 Ig (rabbit antiserum, used to detect p49) were obtained from Upstate Biotechnology (Lake Placid, NY, USA). Anti-PKG Ig (C-terminal, amino acids 657–671, rabbit polyclonal) was purchased from Calbiochem (La Jolla, CA, USA). Reprobing of the membranes with anti-tubulin Ig (mouse IgG isotype, Sigma) served as additional loading control. For pull-down assays, transiently transfected 293T cells were lyzed in 0.5 mL NTEN buffer. The lyzates were precleared at 4 °Cwith25lL 20% agarose beads, before addition of 0.5–1 lg of the indicated antibody for 2 h and precipitation with 25 lL 20% agarose beads for 1 h. The beads were washed, resuspended in SDS/PAGE sample buffer, and the bound proteins were resolved on 8% SDS/ polyacrylamide gels. The proteins of interest were detected by Western blotting. Cells Reporter assays were performed by transient transfections of 293T cells. O3 is a CD4 + T-helper cell clone derived from BALB/c mice after in vitro selection for proliferation to ovalbumin in association with BALB/c antigen-present- ing cells [31]. O3 cells express V b 6 and respond to conventional antigen (ovalbumin) and to the retroviral superantigen MTV-7 (Mls-1 a ) [32]. The AF3.G7 hybri- doma was generated by fusing beef insulin immune C57BL/6 lymph node cells with the BW5147 thymoma line. It bears V b 6andV a 3.2 and responds to MTV-7 according to interleukin-2 production [33]. After stimula- tion by conventional antigen or superantigen, the T-cells were obtained by passage through Cell-ect TM columns (Biotex Laboratories Inc., Edmonton, Alberta, Canada) for the preparation of nuclear extracts. As control for activation, measurement of 3 H-thymidine incorporation by O3 cells after stimulation with mitomycin C-treated LBB cells as antigen-presenting cells was performed as described previously [32]. 2176 B. He and G. F. Weber (Eur. J. Biochem. 270) Ó FEBS 2003 Results NF-jB is activated by PKG in T-lymphocytes T-lymphocytes proliferate in response to engagement of their antigen receptor by conventional antigen peptide. We have previously described an alternative signal transduc- tion pathway, associated with the T-cell antigen receptor that is induced by superantigen and leads to activation of PKG and activation-induced cell death [3]. Because T-lymphocyte stimulation is often characterized by NF-jB translocation to the nucleus and binding to cognate DNA sequences, we compared the induction of NF-jB following stimulation by conventional antigen or retroviral super- antigen in the T-cell clone O3. Consistent with previous observations [34], the induction of NF-jB by physiologic T-lymphocyte stimulation is moderate. In electrophoretic mobility shift assays measuring the binding to a NF-jB consensus probe, the larger and smaller NF-jB complexes were induced by the conventional antigen. In contrast, predominantly the smaller complex, represented by the lower band on the gel, was induced by stimulation with superantigen. This was not due to quantitative differences in stimulation because both modes of T-cell activation induced comparable levels of tritiated thymidine incorpor- ation in the same experiment (Fig. 1A). To confirm the dependence of the superantigen mediated induction of NF-jB, we treated the O3 clones with the cell permeable PKG inhibitor Rp-8-pCPT-cGMPS before stimulation. Because superantigen induces predominantly the faster migrating band the gel shift analysis was performed with the H2K probe, which has a higher affinity to p50/p52 than the consensus probe. Expectedly, the presence of the PKG inhibitor suppressed the superantigen-dependent induction of NF-jB but had no effect on T-cell stimulation by conventional antigen. To test whether PKG could induce the lower NF-jB band, we treated nuclear extracts from resting O3 T-cell clones or AF3.G7 T-cell hybridoma cells with PKG and cGMP and found it to induce DNA binding of the lower band to a NF-jB consensus Fig. 1. Correlation between PKG activity and NF-jB induction in T-lymphocyte activation. (A,leftpanel)TimecourseofNF-jB induction after activation of O3 T-cell clones. O3 cells were stimulated by the conventional antigen ovalbumin (OVA) or retroviral super- antigen (MTV-7) for 0, 2, or 4 h. The T-cells were selected for pre- paration of nuclear extracts, which were then subjected to incubation with a 32 P-labelled NF-jB consensus probe followed by electropho- retic mobility shift assay. The relative intensity of the slower and faster migrating bands after stimulation with MTV or OVA was quantitated by densitometric measurement. The relative density units for the lower band are O3 7.0, OVA 2 h 9.8, OVA 4 h 12.8, MTV 2 h 7.9, MTV 4 h 9.7.TherelativedensityunitsfortheupperbandareO36.9,OVA2h 10.5, OVA 4 h 13.3, MTV 2 h 5.5, MTV 4 h 6.7. Comparable levels of T-cell stimulation were confirmed by tritiated thymidine incorporation (c.p.m.). (A, right panel) Before stimulation for 4 h, the O3 cells were pretreated with the inhibitor Rp-8-pCPT-cGMPS. After T-cell selec- tion, nuclear extracts were prepared and analyzed by gel shift assay for binding to the H2K probe. (B) Nuclear extracts from the AF3.G7 hybridoma or the O3 clone were phosphorylated in vitro by PKG plus cGMP followed by analysis of DNA binding to an oligonucleotide containing the H2K sequence in gel shift assays. The nuclear extract from O3 cells that had been treated with plate-bound anti-CD3e antibody served as a positive control for induction of the faster migrating NF-jB band. Note that the resting levels of NF-jBbinding are relatively high in O3 cells, because the clone depends on the pre- sence of interleukin-2 in the growth medium, whereas the hybridoma AF3.G7 does not. (C) The induction by PKG plus cGMP of DNA binding by AF3.G7 cytosol is inhibitable by addition of high con- centrations of a competing PKG substrate peptide, GRTGRRNSI (ÔPKI substrateÕ,amountsarel M ). In addition, cGMP and PKG did not affect Octamer-1 binding in the same experiment (not shown). Similar results were obtained in at least three additional experiments. Ó FEBS 2003 Phosphorylation of NF-jB by PKG (Eur. J. Biochem. 270) 2177 oligonucleotide (Fig. 1B). A similar induction was seen in AF3.G7 cytosol and this was inhibitable by increasing amounts of the competing PKG substrate peptide GRTGRRNSI (ÔPKI substrateÕ) (Fig. 1C). We therefore set out to investigate the role of PKG in the induction of NF-jB. PKG increases the transactivating activity of NF-jB proteins from distinct recognition sites We asked whether PKG can alter the transactivating activity of p65 as judged by luciferase assays with a commercial NF- jB luciferase reporter that contains four consensus NF-jB Fig. 2. PKG increases the transactivating activity of p65 from consensus sites but not from a nonconsensus NF-jBreporter.(A) 293T cells were transiently transfected with 300 ng pRSV-p65 and increasing amounts of pEF6/HisB-PKG as indicated. Transactivation from the cotransfected pNF-jB-luc reporter (500 ng) was measured by luciferase activity. The values are normalized to luminescence induced by 10 ng cotransfected Renilla construct pRL-SV40 and the results (fold induction) are presented as mean ± standard deviation of three samples (top panel). The protein expression of the transfected molecules correlated with the amounts of DNA introduced into the cells, while tubulin (used as a loading control) remained constant (bottom panel). The PKG activity in 20 lg of cell lyzates, as judged by phosphorylation of LRRASLG peptide (ÔkemptideÕ) in vitro, reflected the amounts of PKG transfected. The kinase activity is indicated as mean ± standard deviation (middle panel). (B) 293T cells were transiently cotransfected with pC/EBP-wt-luc reporter (1 lg) and pRSV-p65 (0.3 lg) with or without 3 lg pEF6/HisB-PKG. Transfection of 0.5 lg pRSV-p50 with pC/EBP-wt-luc reporter (1 lg) served as a positive control. Twenty-four hours after transfection, the cells were harvested in reporter lysis buffer. Forty micrograms of lyzate samples were used for luciferase assays by luminometer and the values obtained for the vector control group were normalized to 1. The results represent mean ± standard deviation of triplicate samples (top panel). Similar results were obtained in three independent experiments. Twenty micrograms of lyzates were used for Western blotting to confirm the expression levels of the transfected proteins (bottom panel). (C) 293T cells were transiently transfected with increasing amounts of PKG in conjunction with either 2 lg NF-jB consensus reporter or 2 lg pC/EBP-wt-luc reporter. Twenty-four hours after transfection, the cells were harvested in reporter lysis buffer. Luciferase activity was measured in 40 lg of lyzate samples and the values obtained for the vector control group were normalized to 1. The results represent mean ± standard deviation of triplicate samples. 2178 B. He and G. F. Weber (Eur. J. Biochem. 270) Ó FEBS 2003 sites. Cotransfection of increasing amounts of PKG dose- dependently enhanced the transactivation by transfected p65. The results were consistent with the protein expres- sion levels and kinase activities in the cell lyzates (Fig. 2A). Therefore, PKG increases the transactivating ability of p65. We also performed cotransfection experiments with a reporter that contains a nonconsensus NF-jB binding site. Consistent with earlier reports [35–37], p65 does not transactivate the nonconsensus motif associated with C/EBP, which is found in the C-reactive protein promoter. In this case, the lack of transactivation is not overcome by cotransfection of PKG (Fig. 2B). We noted that the transfection of PKG alone was sufficient to increase the consensus reporter activity by approximately fourfold (compare Fig. 2A). We therefore tested whether PKG could stimulate transactivation by the endogenous NF-jB. 293T cells express substantial amounts of endogenous p65, but very little p49 and p50. Consistently, transfected PKG increased the activity of the consensus reporter, but not of the C/EBP-associated promoter, in a dose-dependent manner, reaching Fig. 3. Transactivation by p50/p49 from a nonconsensus sequence is enhanced by PKG. (A) Transactivation by transfected p50 of a luci- ferase reporter gene containing a nonconsensus NF-jBsitethat overlaps with a C/EBP site (sequence in top panel). 293T cells were transiently cotransfected with 1 lg pC/EBP-wt-luc reporter (or the control construct pC/EBP-mp50-luc, in which the p50 binding site is mutated, represented as pC/EBP (mP50)) with 0.5 lgpRSV-p50and the indicated amounts of pEF6/HisB-PKG. Twenty-four hours after transfection, the cells were harvested in reporter lysis buffer and 40 lg lyzate samples were used for luciferase assays. The results, measured as fold induction, are presented as mean ± standard deviation of three samples and the values obtained for the vector control group have been normalized to 1. The transcriptional activity is induced dose-depend- ently by cotransfection of increasing doses of PKG (second panel from top). Twenty micrograms of the lyzates were analyzed for PKG kinase activity (third panel from top), while another 20 lg of lyzates were used for Western blotting to confirm the protein expression levels (fourth panel). PKG activity in cell lyzate and Western blotting for the transfected molecules served as transfection controls. (B) PKG enhances p49-mediated transactivation from the C/EBP-associated nonconsensus site. 293T cells were transiently cotransfected with 1 lg pC/EBP-wt-luc reporter and 0.5 lgpRSV-p49plusincreasing amounts of pEF6/HisB-PKG for analysis of luciferase reporter gene activity. The results are represented as fold induction and the values obtained for the vector control group are normalized to 1. Western blotting confirmed the expression levels of the transfected molecules (bottom panel). (C) Neither p49 (top panel) nor p50 (bottom panel) transactivate from the commercial NF-jB reporter gene containing four consensus sites. 293T cells were transiently cotransfected with 0.5 lgpNF-jB-luc plus 0.3 lgp49or0.3lgpRSV-p50plusincreas- ing amounts of pEF6/HisB-PKG. 10 ng of the Renilla construct pRL- SV40 was also cotransfected to normalize the data for transfection efficiency. Transactivation by p65 (0.3 lgpRSV-p65)inthesame experiment is shown as a positive control (at the chosen concentrations of reporter DNA, PKG induces a less than twofold induction of transactivation by endogenous p65). Expression of the transfected proteins was confirmed by Western blotting. All panels show the results of one representative experiment from at least three replicates. Ó FEBS 2003 Phosphorylation of NF-jB by PKG (Eur. J. Biochem. 270) 2179 approximately fivefold increase in luminescence readout (Fig. 2C). The NF-jB protein p50 contains a DNA binding domain, but no transactivation domain. Nevertheless, transactivation may be observed after transfection of p50 into cells, presumably due to its binding to endogenous interaction partners. Those include most prominently p65, but also Bcl-3 [38]. In addition, p50-dependent transactiva- tion can occur from a nonconsensus site in conjunction with C/EBP [35,36]. We used the nonconsensus reporter con- struct in transient cotransfection assays. No reporter activity was induced by PKG alone, whereas p50 dose-dependently increased the luciferase activity (data not shown). Co-transfection of PKG with low amounts of p50 (0.3 lg DNA) dose-dependently enhanced its transactivating acti- vity (Fig. 3A), consistent with an increased affinity of p50 to this DNA sequence after phosphorylation by PKG. Com- parable results were obtained with murine p50 (data not shown). P49 and p50 are related NF-jB subunits [39]. We found p49 to also transactivate from the nonconsensus NF-jB site in a manner that could be increased dose- dependently by cotransfected PKG (Fig. 3B). We then tested whether p50 or p49 transactivate the luciferase reporter that contains four NF-jB consensus sites and whether transactivation under these conditions might be modulated by PKG. Luciferase activity was not induced by transfection of p49 or p50 alone (under the conditions used here, PKG enhances the transactivation by endo- genous p65 less than twofold). Furthermore, cotransfection of p50 or p49 with increasing amounts of PKG did not lead to measurable transactivation from the NF-jB consensus luciferase reporter (Fig. 3C). Fig. 4. The NF-jB proteins p49, p50, and p65 are substrates for cGMP-dependent kinase. (A) Substrate phosphorylation of p49 or p50 depends on the presence of PKG and is enhanced by the addition of cGMP, whereas cGMP in the absence of the kinase does not mediate measurable incorporation of phos- phate. Autophosphorylation of PKG is rep- resented as the upper band on all gels and reflects a specificity control for effects on the enzyme. (B) Phosphorylation of p49 or p50 by PKG (as well as PKG autophosphorylation) is reversible by titration of a competing substrate peptide for the kinase (GRTGRRNSI), but not a control peptide with a mutated serine (GRTGRRNAI). (C) An analog of cGMP, Rp-8-pCPT-cGMPS, which can act as an inhibitor of PKG, reverses the enzymatic phosphorylation of p49 or p50. Consistent with the competition for binding to the kinase between cGMP and Rp-8-pCPT-cGMPS, the inhibition is more complete in the absence of cGMP than in its presence. Autophosphory- lation of PKG serves as a positive control for kinase activity. (D) Recombinant His-tagged p65 was phosphorylated by PKG in vitro in the absence or in the presence of cGMP (top panel). 293T cells were transiently transfected with p65 or FLAG-tagged p49. The cells were lyzedinRIPAbufferandthetransfected molecules were pulled down by antibodies to the p65 or Flag. Kinase reaction buffer plus [c- 32 P]-ATP was directly added to the pelleted beads with or without PKG and cGMP for 15 min at room temperature. The reaction mixtures were analyzed by autoradiography and Western blotting (bottom panel). 2180 B. He and G. F. Weber (Eur. J. Biochem. 270) Ó FEBS 2003 In summary, the transactivation experiments using reporter assays indicated that PKG can enhance the transcriptional activity of the DNA binding NF-jBproteins p65, p50, and p49 from their cognate recognition sites. In contrast, PKG does not confer transactivating potential from the, respectively, noncognate NF-jB sequences. Because the proteins, p49, p50, and p65 direct transcription from distinct DNA sequences, their activation by PKG enhances their differential effects. P49, p50, and p65 are substrates for cyclic GMP- dependent kinase A possible mechanism to account for PKG-enhanced transactivation by NF-jB is the phosphorylation of the DNA binding proteins by the enzyme. We therefore tested whether p49, p50, and p65 are substrates of the kinase. NF- jB p49 and p50 were phosphorylated by cyclic GMP- dependent kinase and the phosphorylation levels were enhanced by addition of cyclic GMP (Fig. 4A). The substrate peptide GRTGRRNSI (ÔPKI substrateÕ), but not the control peptide GRTGRRNAI, inhibited phosphory- lation of p49 or p50 as well as autophosphorylation of cGMP-dependent kinase in a dose-dependent manner (Fig. 4B). The kinase reactions on p49 and p50 were also inhibited by the cGMP-dependent kinase inhibitor Rp-8- pCPT-cGMPS at a high concentration. Consistent with the competitive function of the compound, the inhibition was complete in the absence of cGMP but partial in the presence of cGMP (Fig. 4C). There is substantial substrate overlap between PKA and PKG. Cyclic AMP-dependent kinase also phosphorylated p50 with comparable efficiency, but p49 was phosphorylated strongly by cGMP-dependent kinase and only very weakly by PKA (data not shown). Fig. 5. PKG binds NF-kB proteins. (A) 293T cells were transfected with vector or p65, or were cotransfected with NF-jBp65plusPKG. Alternatively, p49 was transfected with or without PKG. The cells were lyzed in NTEN buffer. After preclearing, immunoprecipitation was performed with anti-p65 antibody or anti-Flag antibody (for pull- down of p49). The immunoprecipitates and 10% of the input were resolved on SDS/PAGE and the resulting Western blot was probed with antibodies to p65, to p52 (recognizes p49), and to PKG. Detectable levels of endogenous p65 are expressed in 293T cells and are recognized by the specific antibody, accounting for the band on the p65 blot from untransfected cells. Similar results were obtained in a repeat experiment. No bands were detected with control immunoglobulin or in a sample without cell lyzate. (B) 293T cells were transiently trans- fected by calcium phosphate precipitation with vector control or 3 lg pEF6-PKG.After24 h,thecellswerelyzedinNTENbuffer.PKGwas pulled down, the bound proteins were resolved on SDS/PAGE, and the Western blot was probed with antibodies to PKG and to p65. The input (2%) is shown in the two left lanes. The two right lanes show the no cell control and the pull-down with an irrelevant antibody, respectively. (C) 293T cells were transiently transfected with 3 lg vector, pRSV-p50, pRSV-p49, or pRSV-p65. After 24 h, the cells were lyzed for immunoprecipitation. The left panel (Input) shows the Western blots of whole cell lyzates as a control for transfection effi- ciency. The right panel (IP) coimmunoprecipitated PKG (top row) and precipitated or coprecipitated p65 (second row). The immunoglobulin heavy chain interferes with the detection of pulled-down p50 or p49. Therefore, the successful immunoprecipitation in these cases was confirmed by Western blotting for the endogenous precursor proteins p100, which is recognized by anti-p52 antiserum, and p105, which is recognized by antibody to p50 (bottom rows). Immunoprecipitation with an irrelevant antibody and immunoprecipitation without cells served as negative controls (two right lanes). Ó FEBS 2003 Phosphorylation of NF-jB by PKG (Eur. J. Biochem. 270) 2181 Similar to p49 and p50, bacterial recombinant His-tagged p65 was phosphorylated by PKG in vitro and phosphate incorporation was enhanced by the presence of cGMP (Fig. 4D, top panel). We expressed p49 and p65 by transient transfection in 293T cells. P65 was immunoprecipitated with an anti-p65 Ig and p49 was immunoprecipitated with an antibody to Flag-tag. The pulled-down proteins were phosphorylated by PKG in vitro.BothNF-jB subunits incorporated radioactive phosphate, although the phos- phorylation of p65 was substantially weaker than the phosphorylation of p49 (Fig. 4D, bottom panel). We further confirmed the interaction between p65 or p49 and PKG in vivo by coimmunoprecipitation of the kinase with an antibody to either p65 or Flag-tag (for p49) after cotransfection of the kinase with either of the NF-jB proteins (Fig. 5A). Because 293T cells express substantial levels of NF-jB p65, we tested whether transfected PKG could pull down endogenous p65. The immunoprecipitation with anti-PKG antibody efficiently yielded endogenous as well as transfected PKG. In both cases, endogenous NF-jB p65 was bound, and the band intensity on Western blot correlated to the amount of PKG present (Fig. 5B). To extend this analysis, we transfected 293T cells with p65, p50, or p49, immunoprecipitated with antibodies specific to the transfected gene products, and probed for pulled-down endogenous PKG. In all cases, coimmunoprecipitation of substantial amounts of PKG was detected. Due to the high expression levels of p65 in 293T cells, the antibody to p65 also pulled-down PKG from untransfected cells. Not unexpectedly, the immunoprecipitations with anti-p50 and anti-p52 Ig pulled down endogenous p65, suggesting the possibility of a trimeric complex containing PKG, p65, and p50/p49 (Fig. 5C). The phosphorylation of p65 occurs on nonconsensus sites P65 is a substrate for PKA [11], an enzyme, whose substrate specificity is similar to PKG. Likely recognition sites for both enzymes are in positions serine 276 and threonine 305 on p65, and serine 276 has been demonstrated to be phosphorylated by cyclic AMP-dependent kinase [11]. We mutated both candidate phosphorylation sites. Phosphory- lation of the serine in position 276 is known to be essential for p65 dependent transactivation [11]. In accord with these previous observations, the transactivating activity was diminished moderately by the mutation T305A and sub- stantially by S276A, however, cotransfection of PKG led to comparable dose-dependent increases in reporter activity in all cases (Fig. 6A). Consistent with this observation, synthetic peptides covering the threonine 305 (EKRKRT YETF)ortheserine276(MQLRRPSDRE) did not incorporate radioactive phosphate during incubation with the kinase, whereas the standard substrate peptide LRRASLG (ÔKemptideÕ) did (Fig. 6B). We also phospho- rylated bacterial recombinant p65-His [40] with PKG in vitro and were able to compete the kinase reaction with the substrate peptide GRTGRRNSI, but not with the peptides covering threonine 305 or serine 276 (Fig. 6C). These results suggest the hypothesis that PKG phos- phorylates p65 in positions distinct from the amino acids 276 and 305. Fig. 6. The phosphorylation of p65 by PKG does not occur on the consensus recognition sites. (A) 293T cells were transiently cotrans- fected with 0.5 lgpNF-jB-luc and 0.3 lgpRSV-p65oritsmutants T305A or S276A plus increasing amount of pEF6/HisB-PKG. Co-transfected 10 ng of the Renilla construct pRL-SV40 served as a control for transfection efficiency. Twenty-four hours after transfec- tion, the cells were harvested in reporter lysis buffer and 10 lLof1 : 40 diluted lyzates were assayed for luciferase activity. The data are pre- sented as fold induction with the values obtained from the vector (pEF6/HisB) transfected cells normalized to 1. The inset shows the PKG-dose dependent increase in transactivation by p65S276 A on an adjusted scale. (B) Ten micrograms of the synthetic peptides EK- RKRTYETF (T305p65), MQLRRPSDRE (S276p65), or Kemptide were incubated with or without 1 U PKG plus 200 l M cGMP plus 1m M [c- 32 P]ATP in total volume of 100 lL for 3 min at room tem- perature. One half of the reaction volume was spotted onto filter paper and the reaction was stopped by washes in phosphoric acid. The levels of peptide phosphorylation were determined by scintillation counting and the results are presented as mean values ± standard deviation of three samples. (C) Recombinant p65 (10 lgÆmL )1 ) was phosphoryl- ated by 1 U PKG in kinase reaction buffer at room temperature for 15 min, in the presence or absence of increasing concentrations of the synthetic peptides GRTGRRNSI (ÔPKI substrateÕ), or EKRKRT YETF (T305p65), or MQLRRPSDRE (S276p65). The extent of p65 phosphorylation and PKG autophosphorylation were determined by resolution on reducing denaturing 8% SDS/polyacrylamide gel fol- lowed by autoradiography. 2182 B. He and G. F. Weber (Eur. J. Biochem. 270) Ó FEBS 2003 Phosphorylation of p50 by PKG directly impacts its DNA binding characteristics Changes in transactivating activity may reflect alterations in DNA binding affinity. We studied the effects of PKG on oligonucleotide binding by NF-jB proteins in elec- trophoretic mobility shift assays. Cyclic GMP-dependent kinase is inactive under standard gel shift assay condi- tions. Conversely, gel shift assays cannot be performed in the PKG reaction buffer. We therefore adjusted the kinase reaction buffer so that we could phosphorylate NF-jB proteins and then transfer an aliquot to the standard DNA binding buffer for analysis of phosphorylation- dependent changes in the DNA binding characteristics. Phosphorylation by cGMP-dependent kinase did not affect the binding of p49 or p50 to the H2K probe, to which these proteins already have high affinity without being phosphorylated [41]. In contrast, in vitro binding of recombinant p50 to the NF-jB consensus sequence or to the nonconsensus NF-jB-C/EBP sequence was increased by PKG (Fig. 7A,B). The binding affinity of nuclear extracts from 293T cells transiently transfected with p50 and PKG to the same probes was similarly increased (Fig. 7C,D). The specificity of the main DNA binding band from transfectants of p50 or p50 plus PKG was confirmed by supershift (Fig. 7D). The subcellular localization of NF-jB subunits is not affected by phosphorylation with PKG NF-jB is an inducible transcription factor, which is retained in the cytosol in resting cells. It was therefore possible that the phosphorylation of DNA binding subunits might affect their nuclear import as would be reflected in their subcel- lular distribution. Transfection of increasing amounts of PKG did not alter the relative fractions of cotransfected p49, p50, or p65 in the cytosols and nuclei. Although not definitive, these observations made phosphorylation- induced changes in the transport and half-lives of these NF-jB subunits unlikely (Fig. 8). Discussion The incorporation of phosphate into p49 or p50 depends on PKG, is enhanced by cGMP, can be competed by a PKG substrate peptide but not by a control peptide, and is inhibitable by a cGMP analog with inhibition being more efficient in the absence of cGMP than in its presence. The sum of these observations indicates that NF-jB p49 and p50 are specific substrates for PKG. Similarly, phosphorylation of p65 is dependent on the kinase, is increased in the presence of cGMP, and is competed out by a standard substrate peptide. Cyclic GMP-dependent kinase, like cyclic AMP- dependent kinase, has a preference for the phosphorylation of serines or threonines found close to at least two consecutive N-terminal basic residues. The standard PKG recognition site is (R,K)(R,K)X(S,T). It is important to note, however, that there are a number of exceptions to this rule. The p50 precursor p105 has PKG recognition sites in positions 335 and 940 (GenBank accession numbers M57999 or NM_003998). Although p49 (GenBank accession number A57034) does not contain any PKG consensus sites, there are five similar sites with the sequence X(R,K)X(S,T) in p49 (amino acids 76, 195, 201, 231, 430) that may conceivably serve as candidate recognition motifs. Interestingly, we have found p49 to be a poor substrate for PKA. The consensus sites for PKG in p65 are at positions 276 and 305 (GenBank accession number M62399). The phosphorylation of amino Fig. 7. PKG dependent phosphorylation of p50 increases its binding affinity to consensus and nonconsensus (jB-C/EBP) sequences. (A and B) Recombinant p50, 25 ng per sample, were phosphorylated at room temperature for 15 min by 1 U purified PKG with or without 200 l M cGMP in total volume of 10 lL. The reaction mixtures were transferred to DNA binding buffer and incubated for additional 20 min with the indicated 32 P-labeled probes. The reactants were resolved on native 4% polyacrylamide gels and exposed to autoradiography film. (C and D) 293T cells were transiently transfected with 0.3 lg pRSV-p50 with or without 3 lg pEF6/HisB-PKG. Twenty-four hours after transfection, the cells were har- vested, washed in NaCl/P i , and nuclear extracts were prepared. Ten micrograms of nuclear protein was used for electrophoretic mobility shift assay with radiolabeled NF-jB consensus or OCT1 oligonucleotides (C), or with radiolabeled jB-C/EBP oligonucleotides (D). The identity of the major DNA-binding band was confirmed by supershift with 50 ng anti-p50 antiserum, added to the nuclear extracts at room temperature for 10 min before DNA binding (D). The upper arrow indicates the supershifted band. Ó FEBS 2003 Phosphorylation of NF-jB by PKG (Eur. J. Biochem. 270) 2183 [...]... SRE, FAP, and CRE [42] The connection of PKG and NF-jB may have broad implications, as a pathway that involves phospholipase A2 , hydroxyl radical, guanylate cyclase, and PKG has been mapped to T-cell signal transduction leading to apoptosis [3] as well as to long-term potentiation in the central nervous system [18] Cellular NF-jB activation by H2O2 has been shown to involve protein phosphorylation and... therefore to be mediated indirectly via induction of a kinase [8] The cGMP-dependent kinase may thus provide a link for the NF-jB activation by reactive oxygen species because the activity of guanylate cyclase, which synthesizes the cofactor cGMP, can be regulated by redox reactions Acknowledgements Fig 8 The subcellular distribution of NF-jB proteins is not a ected by PKG 293T cells were transiently cotransfected... Calcium Homeostasis and Signaling in Aging Advances in Cell Aging and Gerontology, Vol 10 (Mattson, M.P., ed), pp 127–145 Elsevier, Amsterdam Takuma, K., Phuagphong, P., Lee, E., Mori, K., Baba, A & Matsuda, T (2001) Anti-apoptotic effect of cGMP in cultured astrocytes: inhibition by cGMP-dependent protein kinase of mitochondrial permeable transition pore J Biol Chem 276, 48093–48099 Canals, S., Casarejos,... specific antibodies to the indicated proteins In the top panel, NF-jB p105 is shown as a marker protein for cytosol to indicate proper fractionation acid 276 by PKA is of major importance for p65 transcriptional activity Remarkably, our experiments indicate that PKG phosphorylates p65 in positions other than the two consensus sites The differential phosphorylation of p49 by PKG but much less by PKA, and... and may differ in their affinities to various subunits of the transcription factor The release of cytosolic p65/p50 heterodimers into the nucleus is by itself insufficient to differentiate among the numerous NF-jB promoter sequences Our results suggest a mechanism of differential transactivation of various NF-jB dependent genes by phosphorylation of the DNA-binding subunits Phosphorylation of p65 by PKG... T., Lohmann, S.M & Pilz, R.B (1997) Regulation of gene expression by cyclic GMP-dependent protein kinase requires nuclear translocation of the kinase: identification of a nuclear localization signal Mol Cell Biol 17, 5244–5254 Gudi, T., Casteel, D.E., Vinson, C., Boss, G.R & Pilz, R.B (2000) NO activation of fos promoter elements requires nuclear translocation of G -kinase I and CREB phosphorylation but... 26 Phosphorylation of NF-jB by PKG (Eur J Biochem 270) 2185 structurally linked to transformation by p59v-rel and cytoplasmic retention of p68c-rel Mol Cell Biol 11, 5867–5877 Ghosh, S & Baltimore, D (1990) Activation in vitro of NF-jB by phosphorylation of its inhibitor IjB Nature 344, 678–682 Zhong, H., SuYang, H., Erdjument-Bromage, H., Tempst, P & Ghosh, S (1997) The transcriptional activity of NF-jB. .. is regulated by the IjB-associated PKAc subunit through a cyclic AMPindependent mechanism Cell 89, 413–424 Zhong, H., Voll, R.E & Ghosh, S (1998) Phosphorylation of NF-jB p65 by PKA stimulates transcriptional activity by promoting a novel bivalent interaction with the coactivator CBP/ p300 Mol Cell 1, 661–671 Zhong, H., May, M.J., Jimi, E & Ghosh, S (2002) The Phosphorylation status of nuclear NF-jB. .. phosphorylation of p65 by PKA and PKG on distinct residues implies multiple modes of regulation of NF-jB function It also indicates that the position, in which phosphorylation occurs, is an important determinant for transactivation Recognition sequences for NF-jB have been identified in the promoters of a large number of genes Beside the consensus sequence, additional NF-jB binding sites are known and... B., Sperl, G., Ahnen, D & Pamukcu, R (2000) Exisulind induction of apoptosis involves guanosine 3¢,5¢ -cyclic monophosphate phosphodiesterase inhibition, protein kinase G activation, and attenuated beta-catenin Cancer Res 60, 3338–3342 Wong, A. S., Kim, S.O., Leung, P.C., Auersperg, N & Pelech, S.L (2001) Profiling of protein kinases in the neoplastic transformation of human ovarian surface epithelium . Reagent Repository) was Flag-tagged at the 5¢-end by PCR using the pri- mers 5¢-CTGCAGCATGGACTACAAGGACGACGA TGACAAGGAGAGTTGCTACAACCCAGGTCTG-3¢ and 5¢-GAGAGTTGCTACAACCCAGGTCTG-3¢ with. Phosphorylation of NF-jB proteins by cyclic GMP-dependent kinase A noncanonical pathway to NF-jB activation Bin He 1 and Georg F. Weber 1,2 1 Department