Phosphorylation of IκB-β Is Necessary for Neuronal Survival*

Cerebellar granule neurons undergo apoptosis when switched from culture medium containing depolarizing levels of potassium (high potassium or HK) to nondepolarizing medium (low potassium or LK). We showed that in healthy neurons maintained in HK medium, IκB-β is phosphorylated at a novel site, Tyr-161. LK-induced neuronal apoptosis is accompanied by a decrease in the extent of IκB-β phosphorylation at this residue. Tyr-161 shares similarity to the consensus sequence for phosphorylation by the nonreceptor tyrosine kinases Abl and Arg. Arg phosphorylates Tyr-161 differentially in vitro, and LK treatment does cause a down-regulation of Arg activity. Moreover, treatment of neurons with two structurally distinct and highly selective Abl inhibitors, PD173955 and Gleevec, blocks HK-induced phosphorylation of IκB-β at Tyr-161 and induces neuronal apoptosis. Overexpression of wild-type IκB-β blocks LK-induced apoptosis, but this effect is abolished when Arg is pharmacologically inhibited. On the other hand, forced overexpression of IκB-β in which Tyr-161 is mutated inhibits survival in HK demonstrating the importance of this residue to neuronal survival. Phosphorylation of IκB-β enhances its association with p65/RelA causing an increase in NF-κB DNA binding activity. Our results identified IκB-β phosphorylation as a key event in neuronal survival and provided a mechanism by which this is mediated.

Apoptosis plays a critical role in the normal development of the nervous system by eliminating large numbers of superfluous neurons and ensuring proper neural connections. Aberrant apoptosis often occurs during adulthood leading to an unwanted loss of neurons such as that seen in neuropathological conditions, including Alzheimer, Parkinson, or Huntington disease and following ischemic stroke (reviewed in Ref. 1). Understanding the molecular mechanisms regulating apoptosis will therefore improve our understanding of neurodevelopment and will lead to the development of useful therapeutic strategies against neurodegenerative conditions.
One family of molecules that plays a pivotal role in the maintenance of neuronal survival in a variety of in vivo and in vitro experimental paradigms is NF-B, a widely expressed transcription factor. Inhibition of NF-B activity causes neuronal death in a variety of tissue culture paradigms of neurodegeneration (1)(2)(3). Reduced NF-B activity has also been implicated in neurodegenerative diseases such as Alzheimer disease, Parkinson disease, and amyotrophic lateral sclerosis (reviewed in Ref. 4). In mammalian cells, there are five NF-B proteins, p50, p52, p65 (RelA), RelB, and c-Rel, characterized by the presence of a conserved 300-amino acid Rel homology domain that is located toward the N terminus of the protein (reviewed in Refs. [5][6][7][8]. Functional NF-B is composed of homodimers and heterodimers of these proteins, typically p65:p50, which are held in the cytoplasm by association with members of the IB protein family. Generally, but not always, activation of NF-B is mediated by the phosphorylation of IB proteins on two N-terminal serine or threonine residues by the IB kinase (IKK) 2 complex, which contains the catalytic subunits IKK␣ and IKK␤ and the regulatory IKK/ NEMO protein. Phosphorylation by the IKK complex targets IB for degradation via the ubiquitin-proteasome pathway. The released NF-B thus translocates to the nucleus, where it binds to B DNA motifs within the promoter regions of a variety of genes (6 -8). In addition to nuclear translocation, more recent evidence indicates that maximal transcription activity of NF-B requires protein-protein interaction and site-specific post-translational modifications, including phosphorylation and acetylation (6 -8). Of the seven IB family members, the best studied is IB-␣, which is phosphorylated by IKK at Ser-32 and Ser-36. Besides being a target of IKK, IB-␣ can be phosphorylated at its C-terminal PEST domain by casein kinase-2 and DNA-PK (9 -11). In contrast to IKK-mediated phosphorylation, however, IB-␣ phosphorylation by these kinases does not cause its degradation (9 -11). Another major member of the IB family is IB-␤, which is phosphorylated by IKK at Ser-19 and Ser-23 (12). In comparison to IB-␣, relatively little is known about the functional significance of IB-␤. Although frequently assumed to be functionally interchangeable with IB-␣, more recent evidence indicates that IB-␤ plays distinct roles within the cell (13,14). Although association with IB-␤ can inhibit NF-B, in some situations, IB-␤ can lead to increased NF-B activity, although the precise mechanisms involved remain to be fully addressed (15,16).
We have been studying the molecular mechanisms underlying neuronal survival using primary cultures of cerebellar granule neurons. These neurons undergo apoptosis when shifted from medium containing serum and depolarizing concentrations of potassium (high K ϩ medium, HK) to medium containing low potassium (LK) (17). NF-B is required for the survival of granule neurons by HK (2). Most interestingly, however, neither the levels nor intracellular distribution of the five NF-B proteins nor those of IB-␣ and IB-␤ are altered in neurons primed to undergo apoptosis by LK treatment (2). We have reported previously that one factor involved in the down-regulation of NF-B activity by LK treatment is a lowering interaction between p65 and cAMP-response element-binding protein-binding protein (CBP), an alteration that is associated with hyperphosphorylation (3).
In this study, we show that cerebellar granule neuron survival by membrane-depolarizing stimuli such as HK involves phosphorylation of IB-␤. HK-induced phosphorylation of IB-␤ occurs at a novel site, Tyr-161. We also report that phosphorylated IB-␤ associates with * This work was supported by Department of Defense Grant DAMD 17-99-1-9566 and NF-B and stimulates its DNA binding activity. In neurons primed to die by LK treatment, there is a reduction in IB-␤ phosphorylation and binding activity of NF-B. We present evidence that HK-induced phosphorylation of IB-␤ is likely mediated by nonreceptor tyrosine kinases, Abl/Arg, molecules shown previously to regulate neuronal morphogenesis and axon guidance in the developing nervous system (18 -22). Overexpression of wild-type IB-␤ prevents LK-induced apoptosis in neurons, whereas addition of PD173955, a highly selective Abl/Arg inhibitor, abolishes the neuroprotection effect by IB-␤. Moreover, when IB-␤ (Y161F) mutant is overexpressed in neurons, neuronal survival in HK is reduced. Our results also show that mutation on Tyr-161 of IB-␤ decreases the interaction between IB-␤ and p65, which is similar to the effect of LK treatment and HK treatment with the presence of Abl inhibitor. Thus, our results identify IB-␤ phosphorylation as a key event in neuronal survival and provide a mechanism by which this is mediated.
Cell Culture and Treatments-Granule neuron cultures were obtained from dissociated cerebella of 7-8-day-old rats as described previously (17). Cells were plated in Basal Eagle's Medium with Earle's salts (BME) supplemented with 10% fetal bovine serum, 25 mM KCl, 2 mM glutamine (Invitrogen), and 100 g/ml gentamycin on dishes coated with poly-L-lysine in 24-well dishes at a density 1.0 ϫ 10 6 cells/ well, 1.2 ϫ 10 7 cells/60-mm dish, or 3.0 ϫ 10 7 cells/100-mm dish. Cytosine arabinofuranoside (10 M) was added to the culture medium 18 -22 h after plating to prevent replication of non-neuronal cells. Unless indicated otherwise, cultures were maintained for 6 -7 days prior to experimental treatments. For treatment, the cells were rinsed twice and then maintained in LK medium (serum-free BME medium, 5 mM KCl) or HK medium (serum-free BME medium, supplemented with 20 mM KCl). Unless indicated otherwise in the figure legends, treatment of cultures with pharmacological inhibitors was initiated 15 min prior to rinsing and was maintained through the subsequent incubation in LK or HK medium. The control cultures were treated with Me 2 SO.
Plasmid Construction and Mutagenesis-pGEX-KG constructs containing full-length wild-type and the S19A/S23A mutation of IB-␤ were the generous gifts from Dr. Richard B. Gaynor (Eli Lilly Co.). The truncation mutants were generated by PCR and cloned into pGEX-KG vector. Site-specific mutations in IB-␤ were generated with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The full-length IB-␤ containing wild-type or mutations on S19A/S23A and Y161F, respectively, were subcloned into pEGFP-N1 (Clontech) vector at the N terminus of enhanced green fluorescence protein (EGFP). All plasmid constructs were confirmed by sequencing.

Expression of Bacterially Produced GST-IB-␤ Proteins-
The pGEX-KG-IB-␤ constructs were transformed into Escherichia coli BL21 (DE3) PlyS (Promega, Madison, WI). Cultures (250 ml) of E. coli were grown to an absorbance at 600 nm of 0.4 to 0.6 and induced with 0.4 mM isopropyl ␤-D-thiogalactopyranoside for 3 h to induce the expression of the GST fusion protein. Cells were pelleted, resuspended in buffer A (20 mM HEPES, pH 7.9, 400 mM NaCl, 5 mM dithiothreitol (DTT), 10% glycerol, 0.1 mM EDTA, 0.1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride (PMSF)), mildly sonicated, and centrifuged. The supernatant was incubated with 0.5 ml of glutathione-agarose beads (G-4510, Sigma) for 2 h at 4°C. The matrix was washed four times with buffer A and resuspended in buffer A for GST pull-down assay. To obtain soluble fusion proteins, the beads were further washed two times with 5 ml each of buffer B (50 mM HEPES, pH 8.3, 150 mM NaCl, 0.5% Nonidet P-40, 5 mM DTT, 1 mM PMSF). The fusion proteins were eluted off beads by 10 mM glutathione in buffer B at 37°C for 30 min and stored at Ϫ80°C.
Neuronal Survival-Neuronal survival was assessed by staining neurons with 4Ј,6-diamidine-2Ј-phenylindole dihydrochloride (DAPI) to show the apoptotic characteristics of nuclei after treatment. Briefly, cells growing on coverslips were washed in 0.1 M PB, pH 7.4, and fixed in 4% paraformaldehyde in PB, pH 7.4, for 15 min at room temperature. Cells were then washed once with PB for 5 min at room temperature followed by staining with DAPI solution (0.5 g/ml) for 10 min at room temperature. After staining, cells were washed again with PB for 5 min. Coverslips were then mounted on slides. Cells were visualized under a fluorescence microscope, and the number was counted. Results from at least three separate experiments were statistically analyzed.
Preparation of Nuclear and Cytosolic Extracts-To obtain cytoplasmic proteins, cells were washed with cold phosphate-buffered saline, pH 7.2, resuspended in buffer containing 10 mM HEPES, pH 7.9, 0.1 mM EDTA, 10 mM KCl, 1 mM DTT, 50 mM NaF, 1 mM sodium orthovanadate, 50 mM ␤-glycerophosphate, 5% glycerol, and protease inhibitor mixture (Roche Applied Science), and incubated on ice for 15 min. At the end of incubation, 1/20 volume of 10% Nonidet P-40 was added. Cells were vortexed for 30 s and then subjected to centrifugation for 30 s at 14,000 rpm. Supernatants were collected as cytosolic proteins.
Nuclei from neurons were resuspended in buffer containing 20 mM HEPES, pH 7.9, 50 mM KCl, 420 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 10% glycerol, protease inhibitor mixture and extracted on ice for 30 min, followed by centrifugation at 14,000 rpm for 5 min at 4°C. The supernatants were collected as nuclear extracts. Protein concentrations of the cellular proteins were determined by the Bradford assay using Bio-Rad reagent.
Western Blots-50 g of cellular extracts or protein complex pulled down by GST fusion protein or antibodies were resolved on SDS-polyacrylamide gels. The proteins were then transferred to nitrocellulose membranes. Membranes were blocked and incubated with the various antibodies mentioned previously. After washes, membranes were incubated with horseradish peroxidase-conjugated secondary antibody. Following the incubation, membranes were washed extensively and developed with ECL luminal reagent (Amersham Biosciences). The image was captured on x-ray film. Data were quantified using ImageQuant software (Amersham Biosciences).
Immunoprecipitation-100 -200 g of cellular lysates from treated neurons was incubated with control and specific antibodies (1 to 2 g) on ice for 30 min. Protein A/G-agarose beads were added to the mixture, and further incubation was carried out at 4°C overnight. After incuba-tion, beads were pelleted by centrifugation followed by three washes with buffer containing 10 mM HEPES, pH 7.9, 50 mM KCl, 1 mM EDTA, 1 mM DTT, 5% glycerol, and protease inhibitor mixture. The beads were then used in variety of experiments. Supernatants from the immunoprecipitation were collected and resolved on SDS-polyacrylamide gels for Western blotting of ␣-tubulin antibody.
Analysis of Phosphorylation on Endogenous IB-␤-100-mm dishes of 7-8-day-old neurons were washed twice with warm, phosphate-free DMEM (Invitrogen) and incubated in phosphate-free DMEM containing 20 mM KCl overnight. The cultures were then incubated for 6 h in medium containing [ 32 P]orthophosphate (MP Biomedicals, Irvine, CA) with the indicated treatment. After being lysed in ice-cold RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, 0.1% SDS, 1 mM Na 3 VO 4 , 50 mM NaF, 30 mM ␤-glycerophosphate, 1 mM EDTA, protease inhibitors mixture), the lysates were subjected to immunoprecipitation as described, and the proteins were separated on SDS-polyacrylamide gel. After electrophoretic transfer to nitrocellulose membrane, labeled proteins were visualized by autoradiography. Results were obtained by scanning on Storm860 (Amersham Biosciences). Data were quantified using ImageQuant software (Amersham Biosciences).
GST Pull-down and in Vitro Kinase Assay-GST-IB-␤ proteins were bound to glutathione-agarose beads. The beads were incubated with whole cell lysates prepared from HK-and LK-treated neuronal FIGURE 2. Differential phosphorylation of GST-IB-␤ proteins by cellular extracts from HK-or LK-treated neurons. Bacterially expressed GST-IB-␤ (wt), (S19A/S23A, shown as dm), C-terminally truncated (1-204dm), and N-terminally truncated (195-359) proteins were bound to glutathioneagarose beads. These beads were incubated with cellular extracts prepared from cerebella granular neurons treated for 6 h under HK or LK conditions. In vitro kinase assay was conducted with [␥- 32  cultures. The whole cell lysates were generated from cultures plated in 100-mm dishes (30 ϫ 10 6 cells/dish) and lysed in a volume of 250 l. The kinase assay was performed in the kinase reaction buffer containing 20 mM HEPES, pH 7.9, 100 mM KCl, 5% glycerol, 0.2 mM EDTA, 4 M ATP, 10 Ci of [␥-32 P]ATP, 5 mM NaF, 1 mM Na 3 VO 4 , 40 M MgCl 2 , protease inhibitor mixture at 30°C for 30 min. The beads were pelleted by brief centrifugation. After addition of 4ϫ SDS sample buffer to the pellet, the samples were heated at 95°C for 5 min. The proteins were resolved on 10% SDS-polyacrylamide gels followed by autoradiography. Results were obtained by scanning with Storm860. Data were quantified using ImageQuant.
Analysis of DNA Fragmentation-The neurons (60-mm plate) were treated with various reagents. After 24 h, the cells were harvested and washed with cold phosphate-buffered saline. The cell pellets were resuspended in 0.1 ml of lysis buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, and 0.2% Triton X-100) and incubated on ice for 20 min. Lysed cells were centrifuged at 12,000 rpm for 10 min at 4°C. The DNA was precipitated from the supernatants by adding a 0.1 volume of 5.0 M NaCl and a 0.5 volume of isopropyl alcohol. After incubation at Ϫ20°C, the samples were centrifuged at 12,000 rpm for 10 min. The resulting pellets were resuspended in 50 l of Tris/EDTA buffer containing proteinase K (300 g/ml) and RNase A (100 g/ml) and incubated for 30 min at 50°C. The samples were subsequently analyzed on a 1.5% agarose gel in 1ϫ Tris borate/EDTA buffer.
Transient Transfection-Transient transfection of granule neurons was carried out as outlined by Koulich et al. (2). Briefly, on day 5 after plating, 10 g/well of DNA was precipitated by calcium phosphate method at room temperature for 30 min and added dropwise to cultures that has been washed once with DMEM without L-glutamine, serum, and antibiotics (transfection DMEM). Cells were incubated in DMEM at 37°C for 75-90 min followed by two washes with the medium. The transfection DMEM was then replaced with the original medium. The next day, cells were treated in HK or LK or with the addition of PD173955 (2 M) for 24 h. Cells were then fixed, and DAPI staining was performed. Expression of EGFP and the DAPI-stained nucleus was visualized under a fluorescence microscope. Transfected cells were determined by merging images from GFP and UV filters.

Phosphorylation of IB-␤ but Not IB-␣ Is Reduced during Neuronal
Apoptosis-Cerebellar granule neurons undergo apoptosis when switched from HK medium to LK medium. Although cell death begins at about 16 h, previous studies have shown that commitment to death occurs within 6 h after the switch to LK medium (23)(24)(25)(26). As shown in Fig. 1A, IB-␤ is phosphorylated in HK medium, and LK treatment leads to a reduction in the level of IB-␤ phosphorylation. The reduction in IB-␤ phosphorylation is detectable as early as 4 h after LK treatment, suggesting that it is causally involved in the induction of neuronal apoptosis (Fig. 1A). In contrast to IB-␤, the phosphorylation level of IB-␣ is similar in HK or LK medium (Fig. 1B). The overall pattern of protein phosphorylation (Fig. 1C) and protein levels of both IB-␣ and IB-␤ are not altered by the treatment with HK and LK conditions (Fig. 1D), indicating that the reduced phosphorylation of IB-␤ is specific.
HK-induced IB-␤ Phosphorylation Occurs at Site(s) Independent of IKK-It is known that IKK phosphorylates IB-␤ at Ser-19 and Ser-23. We examined whether HK-induced IB-␤ phosphorylation was medi- ated by IKK by using a mutant GST-IB-␤ construct in which both Ser-19 and Ser-23 were mutated (GST-IB-␤ S19A/S23A , see Fig. 2C). This construct can therefore not be phosphorylated by IKK. As shown in Fig.  2A, although the level of phosphorylation on GST-IB-␤ in both HK and LK was somewhat reduced, the difference in phosphorylation between HK and LK extracts seen with wild-type GST-IB-␤ was also observed in the double mutant. The same result was obtained using a C-terminally truncated form of GST-IB-␤, which lacks the two target sites of IKK (Fig. 2B). The similarity in the difference of phosphorylation elicited in HK and LK medium between wild-type IB-␤ and IB-␤ S19A/S23A or the C-terminally truncated mutant suggests that the differential phosphorylation on IB-␤ during neuronal apoptosis happens on residue(s) that are not sensitive to IKK.
Mapping Apoptosis-regulated Phosphorylation Site-To map the site within IB-␤ that is differentially regulated during LK-induced apoptosis, several deletion constructs were generated and used in in vitro kinase assays with extracts from HK-and LK-treated neurons (Fig. 2C). As shown in Fig. 2B, GST-IB-␤ fragments spanning residues 1-204 S19A/S23A displayed differential phosphorylation similar to that seen with the full-length GST-IB-␤ protein. In contrast, fragment spanning residues 195-359 were not differentially phosphorylated.
Another fragment spanning residues 1-168 S19A/S23A also displayed elevated phosphorylation level in HK compared with LK (data not shown). These results localize the HK-induced phosphorylation site to the region spanning residues 1-168 of IB-␤.
To determine whether the apoptosis-regulated phosphorylation site was a Tyr residue or a Ser/Thr residue, we used genistein, a broad spectrum pharmacological inhibitor of tyrosine kinases. Addition of genistein to the in vitro kinase assay abolished the differential phosphorylation of GST-IB-␤ (Fig. 3A), suggesting that the HK-induced phosphorylation occurred at a Tyr residue. In comparison, other inhibitors, such as the phosphatidylinositol 3-kinase inhibitor wortmannin, the Akt inhibitor ML-9, and the casein kinase-II inhibitor 5,6-dichlorobenzimidazole riboside, did not affect the differential phosphorylation pattern (Fig. 3A). Western blot analysis of GST-IB-␤ in the in vitro kinase assay with a phosphotyrosine antibody also showed increased immunoreactivity in HK compared with LK, providing additional evidence that the HK-induced phosphorylation of IB-␤ occurred at Tyr residue(s) (Fig. 3B). The higher level of phosphotyrosine immunoreactivity was abolished with the addition of genistein (Fig. 3B). To confirm that tyrosine phosphorylation on I B-␤ protein also occurred in vivo, immunoprecipitation was performed with IB-␤ antibody following by Western . Identification of a tyrosine residue that is phosphorylated differentially in HK or LK conditions. A, tentative tyrosine residues in IB-␤ (1-168 S19A/S23A ) predicted by NetPhos 2.0 Sever (www.cbs.dtu.dk/services/NetPhos/) that can be phosphorylated. Site-specific mutagenesis was conducted on pGEX-KG IB-␤ (1-168 S19A/S23A ) on Tyr-86, Tyr-114, and Tyr-161, respectively. These mutated IB-␤ constructs were expressed into proteins. GST-IB-␤ proteins bound with glutathione-agarose beads were incubated with cytosolic extracts prepared from HK-and LK-treated neurons (6 h), followed by in vitro kinase assay with [␥- 32  blotting with Tyr(P) antibody. Results in Fig. 3C reveal that a higher level of Tyr(P) immunoreactivity was detected with the HK-treated neuronal lysate, whereas Western blotting results with the IB-␤ antibody show similarity to total IB-␤ immunoreactivity. These in vivo and in vitro results suggest that HK-induced phosphorylation on IB-␤ happens on tyrosine residue(s).
Analysis of the amino acid sequence of IB-␤ in the region between residues 1 and 168 revealed four tyrosine residues, Tyr-53, Tyr-86, Tyr-114, and Tyr-161, that are conserved among rat, mouse, and human IB-␤ proteins. Among them, Tyr-86, Tyr-114, and Tyr-161 are identified by NetPhos2.0(www.cbs.dtu.dk/services/NetPhos/) as potential phosphorylation sites (Fig. 4A). We investigated whether any of these residues represented the apoptosis-regulated phosphorylation site by using GST-IB-␤ in which each of these three sites were mutated in in vitro kinase assays (Fig. 4D). As shown in Fig. 4, B and C, only mutation of Tyr-161 reduced the extent of phosphorylation seen with HK extracts.
Apoptosis-regulated Phosphorylation of IB-␤ Is Mediated by Arg in Vitro-Tyr-161 resides within the consensus sequence for phosphorylation by the nonreceptor tyrosine kinase Abl (analyzed by ScanSite server at scansite.mit.edu). Two Abl kinases are expressed in mammalian cells, c-Abl and Arg. As a step toward examining whether these kinases were responsible for HK-induced IB-␤ phosphorylation, we immunoprecipitated c-Abl or Arg from neuronal cultures treated with HK or LK medium. The ability of the immunoprecipitated kinase to phosphorylate GST-IB-␤ was analyzed using a construct containing residues 1-168 S19A/S23A of IB-␤ and another construct in which Tyr-161 was mutated to Phe. As shown in Fig. 5, A and B, Arg phosphorylated IB-␤ to a greater extent in HK. The elevated phosphorylation in HK was abolished when Tyr-161 was mutated. PD173955 is a highly selective inhibitor of c-Abl and Arg (27,28). As shown in Fig. 5A, treatment of neuronal cultures with PD173955 blocks the ability of Arg to phosphorylate GST-IB-␤ in HK medium. As seen with the GST-IB-␤ construct, treatment of neuronal cultures with PD173955 also blocked HK-mediated phosphorylation of endogenous IB-␤ (Fig. 6, E and F). However, phosphorylation of the GST-IB-␤ proteins by c-Abl immunoprecipitants did not exhibit the differential pattern (Fig. 5, C and D). These results suggest that Arg is likely the kinase that mediates the differential phosphorylation of Tyr-161 in IB-␤. It has been reported that activity of c-Abl and Arg is tightly regulated by phosphorylation on tyrosine residues in these two kinases. Higher kinase activity of c-Abl/ Arg correlates to higher tyrosine phosphorylation in these two proteins. We performed Western blot experiments to examine the tyrosine phosphorylation status of immunoprecipitated c-Abl and Arg in HK, LK, and HK with PD173955-treated neurons. HK treatment induces the tyrosine phosphorylation on Arg (Fig. 5E) and only slightly higher levels of tyrosine phosphorylation on c-Abl (Fig. 5F). PD173955 treatment abolishes the tyrosine phosphorylation on both Arg and c-Abl (Fig. 5, E and  F), indicating that this drug does affect its targets in vivo. Moreover, the levels of immunoprecipitated Arg and c-Abl remain similar, regardless of the treatment, when the same blots were probed with Arg and c-Abl antibody, respectively. These data suggest that the regulation of Arg and c-Abl activity under the conditions we used occurs at the activity level, and not because of a change in the protein level.
Abl Inhibitors Induce Apoptosis in HK Condition and Reduce IB-␤ Phosphorylation in Vivo-The above results indicated that elevated IB-␤ phosphorylation in HK happens at a site(s) not sensitive to IKK. We also showed that Tyr-161 of IB-␤ is related to the induced phosphorylation and that Arg is the candidate kinase for this phosphorylation. We then conducted experiments to address the biological relevance of the above observations with regulating apoptosis in neurons. We treated neuronal cultures with PD173955, a highly selective inhibitor of Abl and Arg (27). As shown in Fig. 6A, treatment with PD173955 blocked HK-mediated survival in a dose-dependent manner. This result was confirmed using a second and structurally distinct inhibitor of Abl/ Arg, Gleevec (29,30), as shown in Fig. 6B. Arguing against a nonspecific toxic effect of these inhibitors is the finding that treatment with PD173955 caused nuclear condensation and the characteristically nonrandom fragmentation of DNA (Fig. 6, C and D). To confirm that the reduction in neuronal survival observed after pharmacological inhibition of Arg/Abl involved a reduction in IB-␤ phosphorylation, we performed 32 P metabolic labeling experiments. IB-␤ was immunoprecipitated from lysates of neuronal cultures that were labeled with 32 P and treated with PD173955. As shown in Fig. 6E, PD173955 treatment reduces phosphorylation but has no effect on protein level of IB-␤ in vivo. The overall phosphorylation pattern in these cultures was similar, indicating that the reduction of IB-␤ phosphorylation is specific (Fig. 6F). The image of an agarose gel with DNA samples prepared from HK, LK, or HK with PD173955 and LK with PD173955-treated neurons. E, neurons were metabolically labeled with [ 32 P]orthophosphate and treated with HK or HK with PD173955 for 6 h. Lysates prepared from these neurons were immunoprecipitated with control IgG (Con. IgG) or IB-␤ antibody. Immunoprecipitants (IP) were resolved on SDS-polyacrylamide gel followed by autoradiograph to PhosphorImager screen. Image for phosphor-IB-␤ is shown (top). Same membrane was reprobed with IB-␤ antibody (bottom), whereas the total tyrosine phosphorylation profile for the HK or HK/PD-treated neurons is shown in F. IB, immunoblot.
Phosphorylated IB-␤ Associates with p65 and Regulates NF-B Binding Activity-In non-neuronal cell lines, IB-␤ has been found to associate with p65 under some conditions causing an increase in NF-B DNA binding activity (16). NF-B activity in cerebellar granule neurons is higher in HK (2). Because the phosphorylation of IB-␤ is also higher in HK, we examined whether IB-␤ associated with p65 in neurons and, if so, what effect the phosphorylation of IB-␤ had on its interaction with p65. Full-length GST-IB-␤-bound glutathione-agarose beads were incubated with cellular extracts prepared from neurons treated with HK or LK medium. Following the pull-down assay, p65 that associated with the exogenously added GST-IB-␤ was then studied by Western blot by using a p65 antibody. As shown in Fig. 7A, association between GST-IB-␤ and p65 was detectable under both HK and LK conditions. As observed with IB-␤ phosphorylation, the extent of interaction was reduced in LK consistent with the possibility that association between p65 and IB-␤ was regulated by the phosphorylation of IB-␤. To investigate this further, endogenous IB-␤ was immunoprecipitated from 32 P-labeled HK-or LK-treated neuronal cultures, and the extent of p65 interaction was evaluated. As shown in Fig. 7B, both phosphorylation of IB-␤ and its interaction with p65 was reduced in LK medium. The interaction between p65 and IB-␤ was disrupted in the presence of excess amounts on an IB-␤ peptide against which the antibody was made, but not by an unrelated peptide against the p35 protein.
To examine more directly the significance of IB-␤ phosphorylation to its association with p65, we performed GST pull-down and in vitro kinase assay with nonradioactive ATP with lysates from HK-or LKtreated neuronal cultures. As shown in Fig. 7C, GST-IB-␤ pulled down from HK-treated cultures was tyrosine-phosphorylated. The extent of tyrosine phosphorylation was greatly reduced when PD173955 was added to the in vitro kinase reaction. Not unexpectedly and consistent with what was observed in intact neurons, phosphorylation of GST-IB-␤ was also reduced in LK-treated cultures. The amount of tyrosine phosphorylation correlated well with the amount of p65 that associated with GST-IB-␤. Thus, association between GST-IB-␤ and p65 was clearly detectable when HK lysates were used but was barely detectable in the presence of PD173955, LK-treated lysates, or GST-IB-␤Y161F. This mutant IB-␤ displayed significantly lower tyrosine phosphorylation than wild-type IB-␤, and no interaction between p65 and IB-␤ Y161F was observed.
To examine whether Abl/Arg-mediated phosphorylation of IB-␤ affected the activity of NF-B during neuronal apoptosis, we performed EMSA. As shown in Fig. 8A and as reported previously (2), the DNA binding activity of NF-B activity is reduced in neurons primed to die by LK treatment. NF-B binding activity is also reduced in HK with the presence of PD173955 (Fig. 8B). In contrast, neither LK nor PD173955 had a substantial effect on the DNA binding activity of Sp1 or TFIID. The DNA binding activity of NF-B (Fig. 8C) was also reduced when the nuclear extracts were co-incubated with IB-␤ or p65 antibody (Fig.  8D), suggesting the presence of these proteins in the complex. Taken together, these results suggest that I〉-␤ and NF-B are associated in a DNA-binding complex that is stimulated by Arg-mediated phosphorylation of IB-␤.
Overexpression of Wild-type or Y161F Mutant of IB-␤ Affects Neuronal Survival-To address further the in vivo function of IB-␤ and the relationship of Tyr-161 phosphorylation to neuronal survival, we transiently expressed wild-type IB-␤ or a mutant form of the protein in which Tyr-161 was mutated (IB-␤Y161F) in cultured neurons. As shown in Fig. 9, forced expression of wild-type IB-␤ protects neurons from LK-induced cell death. Consistent with the requirement for Arg-mediated phosphorylation, inclusion of PD173955 blocked protection by wild-type IB-␤. Further underscoring the importance of Arg-mediated IB-␤ phosphorylation to neuronal survival is the finding that overexpression of IB-␤Y161F decreases neuronal survival in both HK and LK conditions.

DISCUSSION
We report that IB-␤ is phosphorylated in cerebellar granule neurons and that the level of phosphorylation is reduced in neurons primed to apoptosis by LK treatment. The higher phosphorylation of IB-␤ in HK medium is not because of IKK. A mutant form of IB-␤ with mutations in the two IKK-mediated phosphorylation sites and a C-terminally truncated form of IB-␤ lacking the two phosphorylation sites still display differential phosphorylation in HK versus LK medium. We have mapped the apoptosis-regulated phosphorylation site to Tyr-161. A comparison of sequences from different mammalian species reveals that this is an evolutionarily conserved residue within IB-␤ and hence one that may be important for the function of the protein. Indeed, over-

FIGURE 7. Differential association of IB-␤-p65 in the HK-and LK-treated neurons and effect of Tyr-161 mutation and PD173955 on GST-IB-␤ phosphorylation in vitro.
A, cellular extracts prepared from neurons with 2 and 6 h of treatment with HK and LK were applied to GST pull-down assay with GST-IB-␤-bound glutathione-agarose (full-length, wild-type). Western blot experiment was conducted with p65 antibody (top), GST antibody (middle), and ␣-tubulin (bottom) for input protein levels. B, neurons labeled metabolically with 32 P and treated with HK or LK for 2 or 6 h were lysed in RIPA buffer. Immunoprecipitation of whole cell lysates with IB-␤ antibody was performed in the presence of specific or nonspecific competitive peptides. The autoradiograph is shown at the top. The same membrane was probed with p65 antibody (middle). Supernatants from the immunoprecipitation were subjected to Western blot with ␣-tubulin antibody (bottom). C, effect of tyrosine phosphorylation of IB-␤ on its association with p65. Full-length GST-IB-␤ (WT) or GST-IB-␤ Y161F was bound to glutathione-agarose beads. GST pull-down and nonradioactive in vitro kinase assay in the presence of ATP (see "Experimental Procedures") was performed. Neurons were treated in HK or LK conditions for 6 h. Cells were lysed in RIPA buffer after treatment. Upper panel, PD-HK refers to a sample in which the pull down was performed from HK lysates, and 2 M PD173955 added to the kinase reaction following the pull-down. The reaction samples on the beads were resolved on SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose membrane. Western blotting was performed with phosphotyrosine (1st row), p65 (2nd row), or GST (3rd row) antibodies. Lower panel, Western blotting experiment was conducted for the flow-through samples with ␣-tubulin antibody. WT, wild type. expression of a mutant form of IB-␤ in which Tyr-161 is mutated to a nonphosphorylatable residue inhibits neuronal survival even in HK.
Tyr-161 bears the consensus sequence for phosphorylation by the Abl nonreceptor tyrosine kinases. The Abl family consists of c-Abl and its paralogue Arg. Both c-Abl and Arg are expressed in most neurons during development and in the adult brain with Arg being much more abundant than c-Abl in the adult brain (19). Although the functional significance of these Abl kinases in the brain are unclear, roles in neuronal migration, axonal guidance, and synaptic communication have been suggested (18,20,22,31). In Drosophila deletion of the single Abl gene, d-abl, leads to severe central nervous system defects (21,32). Although gross abnormalities have not been detected in Abl Ϫ/Ϫ or Arg Ϫ/Ϫ mutant mice (19), this is likely to be due to their overlapping roles. Double mutant Abl Ϫ/Ϫ Arg Ϫ/Ϫ mouse embryos exhibit defects in neurulation, although the early death of these mice at embryonic day E10.5 precludes analysis of the roles of these proteins in brain development and maturation (19).
Consistent with their involvement in IB-␤ phosphorylation, the activities of c-Abl and Arg are reduced when neurons are switched from HK to LK medium. Treatment of neurons with PD173955, a specific Abl inhibitor, reduced GST-IB-␤ phosphorylation in HK-treated cultures to a level comparable with that observed in LK medium. Although both c-Abl and Arg interact with IB-␤, the differential pattern of endogenous IB-␤ phosphorylation is recapitulated in vitro only by Arg implicating it in the phosphorylation of IB-␤. Underscoring the importance of Abl-mediated phosphorylation of IB-␤ to neuronal survival is our finding that inhibition of this phosphorylation event using two distinct pharmacological inhibitors, PD173955 and Gleevec, leads to cell death.
How does the phosphorylation of IB-␤ at Tyr-161 help neuronal survival? We find that IB-␤ associates with p65, and association is enhanced by IB-␤. Mutation of Tyr-161, or the pharmacological inhibition of Abl kinases, reduces the association of IB-␤ with NF-B. The importance of NF-B to neuronal survival has been well established in a number of neuronal paradigms (1,2). In cerebellar granule neurons, the activity of NF-B is reduced by LK treatment (2,33). A similar reduction of NF-B activity is observed when IB-␤ phosphorylation is inhibited by PD173955 treatment. The crystal structure (Protein Data Bank codes IK3Z, crystal form I, and 1OY3 crystal form II) of a IB-␤-p65 complex reported recently reveals that this complex can bind to DNA (34). We have confirmed that IB-␤-bound p65 can bind DNA. 3 Studies in other laboratories have shown that IB-␤ bound to p65 can enhance NF-B DNA binding activity within the nucleus (15,16). Taken together, these observations suggest that the phosphorylation of IB-␤ increases NF-B activity by stimulating DNA binding.
The best studied member of the IB protein family is IB-␣. Although IB-␣ and IB-␤ have often been assumed to be interchange-  No (no competitive nucleotides), nonradioactive B, and AP1 oligos were included in the assay for competition experiments. D, same nuclear extract as in C was incubated with control (con) IgG, IB-␤, and p65 antibody before addition of radiolabeled B oligos for supershift experiment. able, a portion of IB-␤ that is important for its subcellular localization is missing in IB-␣ (34). Tyr-161 resides within this region. EMSA results from our laboratory using recombinant IB-␣ shows that it inhibits the DNA binding of p65. 3 In cerebellar granule neurons and other neuronal cell types, overexpression of IB-␣ inhibits NF-B activity and induces apoptosis, whereas overexpression of either wild-type or a mutant form of IB-␤ in which IKK-mediated phosphorylation sites are altered results in the survival of neuronal cells under apoptotic conditions. 3 Thus, in the context of neuronal survival, IB-␣ and IB-␤ have different effects.
In contrast to other members of the IB family of proteins, mice lacking IB-␤ have yet to be generated, and hence the physiological function of IB-␤ is poorly understood. Our results implicate a novel role for IB-␤ in the regulation of neuronal survival. This action of IB-␤ is regulated by its phosphorylation by Abl kinases and leads to the activation of NF-B. By understanding in more detail the molecular events leading to the inactivation of c-Abl and Arg in dying neurons and the mechanism by which reduced IB-␤ phosphorylation ultimately leads to neuronal death may shed insight into the mechanisms underlying neurodegenerative diseases.