The c-Abl Tyrosine Kinase Phosphorylates the Fe65 Adaptor Protein to Stimulate Fe65/Amyloid Precursor Protein Nuclear Signaling*

The amyloid precursor protein (APP) is proteolytically processed to release a C-terminal domain that signals to the nucleus to regulate transcription of responsive genes. The APP C terminus binds to a number of phosphotyrosine binding (PTB) domain proteins and one of these, Fe65, stimulates APP nuclear signaling. Fe65 is an adaptor protein that contains a number of protein-protein interaction domains. These include two PTB domains, the second of which binds APP, and a WW domain that binds proline-rich ligands. One ligand for the Fe65WW domain is the tyrosine kinase c-Abl. Here, we show that active c-Abl stimulates APP/Fe65-mediated gene transcription and that this effect is mediated by phosphorylation of Fe65 on tyrosine 547 within its second PTB domain. The homologous tyrosine within the motif Tyr-(Leu/Met)-Gly is conserved in a variety of PTB domains, and this suggests that PTB tyrosine phosphorylation occurs in other proteins. As such, PTB domain phosphorylation may represent a novel mechanism for regulating the function of this class of protein.

recently its proteolytically processed C-terminal domain has been shown to translocate to the nucleus to regulate transcription (17)(18)(19)(20)(21)(22)(23)(24). To do so, it complexes with a number of nuclear proteins, one of which is the adaptor protein Fe65. Fe65 contains a variety of protein-protein interaction domains including two C-terminal PTB domains (the second of which binds APP) and a WW domain that interacts with proline-rich ligands. The first Fe65 PTB domain binds to two transcription factors, the histone acetyltransferase Tip60 and CP2/LSF/LBP1, and Tip60 stimulates APP/Fe65 transcriptional activity (19,25). The full complement of binding partners for the Fe65WW domain are not known, but include the mammalian homologue of Drosophila enabled (Mena) and the c-Abl tyrosine kinase (26,27).
The molecular mechanisms that control APP/Fe65 nuclear signaling are not properly understood and indeed, APP can function to inhibit Fe65 nuclear translocation (28). However, the Fe65WW domain is required for potent APP/Fe65-mediated transcription and also for nuclear translocation of Fe65 (19,28). This suggests that its binding partners may contribute to the transcriptional competency of the complex. One partner, c-Abl, is known to be present within the nucleus (29,30) and has been shown to phosphorylate APP on tyrosine 682 (27). Here, we demonstrate that active c-Abl also phosphorylates Fe65 and that this phosphorylation functions to stimulate APP/ Fe65 transcriptional activity.

EXPERIMENTAL PROCEDURES
Experiments involving immunoprecipitation, glutathione S-transferase (GST) pull-downs, and GSK3␤ labeling were all performed at least three times with similar results. Reporter gene transcription assays were repeated as indicated in the text.
Cell Culture, Transfection, and Indirect Immunofluorescence-CHO and COS-7 cells were grown in F-12(HAM) or Dulbecco's modified Eagle's medium, respectively, containing 10% (v/v) fetal bovine serum supplemented with 2 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen). Cells were transfected using Lipo-fectAMINE (Invitrogen) according to the manufacturer's instructions. For immunofluorescence staining, cells were grown and transfected on glass coverslips, and then fixed and processed for immunocytochemistry as described (31). Captured images were analyzed for fluorescent signal intensity using Metamorph image analysis software.
GAL4UAS-dependent firefly luciferase reporter pFR-Luc and transfection efficiency Renilla luciferase phRL-TK plasmids were obtained from Stratagene and Promega, respectively. Plasmid expressing the GAL4 DNA binding domain fused to the C-terminal 50 amino acids of APP (GAL4-APPc) and which corresponds to an in vivo fragment generated by ␥/⑀ cleavage (34 -36) was created by subcloning APPc-encoding sequences into pM1 (37). Full-length APP into which the GAL4 DNA binding domain had been engineered (GAL4-APP) was as described (19). Full-length APP containing the GAL4 DNA binding domain at its C terminus was created by mutating residue 148 of the GAL4 sequence to a stop codon in plasmid APP-GAL4 (kind gift of Tomasso Russo). Plasmid APP-GAL4 encodes a chimeric protein comprising full-length GAL4 fused to the C terminus of APP (38). Mutation of GAL4 residue 148 truncates the GAL4 sequence so that it contains only the DNA binding domain. Mutagenic oligonucleotides were 5Ј-C-TGTATCGATTGACTAGGCAGCTCATCATG-3Ј and 5Ј-CATGATGAG-CTGCCTAGTCAATCGATACAG-3Ј. To express the C-terminal domain of APP, sequences encoding the last 50 residues of APP were cloned into the vector pCMV-Tag2 (Stratagene).
SDS-PAGE and Immunoblotting-Samples were processed for SDS-PAGE by addition of SDS-PAGE sample buffer and heating in a boiling water bath. Samples were separated on either 8.5 or 12.5% (w/v) acrylamide gels and transferred to Protran nitrocellulose membranes (Schleicher & Schuell) using a Bio-Rad TransBlot system. Following blocking and probing with primary antibodies, the blots were washed and incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse Ig (Amersham Biosciences) and developed using an enhanced chemiluminescence system (Amersham Biosciences) according to the manufacturer's instructions. The Fe65 and APP (APPab recognizing the C terminus of APP) antibodies have been described previously (32,39,40). Anti-Myc antibody 9B11 (that recognizes Myc-tagged Fe65) was obtained from Cell Signaling Technology, c-Abl antibody (24 -11) was from Santa Cruz Biotechnology, APP antibody 22C11 (recognizing the N terminus of APP) was from Roche Applied Science, phosphotyrosine antibody 4G10 was from Upstate Cell Signaling, anti-GSK3␤ was from BD Transduction Laboratories, and antibody DM1A to tubulin was from Sigma.
Immunoprecipitation and GST Pull-down Assays-Immunoprecipitation and GST pull-down assays were performed as previously described (39,41). Briefly, for immunoprecipitation assays, cells were lysed in ice-cold lysis buffer comprising 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 0.5 mM sodium orthovanadate, 30 mM sodium fluoride, and protease inhibitors (Complete, Roche Applied Science) and then cleared by centrifugation at 14,000 ϫ g for 30 min at 4°C. Cell lysates were then precleared with either protein A-or protein G-Sepharose beads (Sigma) and the target protein immunoprecipitated. Following washing of the beads in lysis buffer, immunoprecipitated proteins were resolved by SDS-PAGE and detected by immunoblotting.
GST and GST-Fe65WW domain proteins expressed in Escherichia coli BL21(DE3) were prepared essentially according to the manufacturer's instructions (Amersham Biosciences). Equimolar amounts of GST or GST-Fe65WW domain baits were used in pull-down assays from transfected cell lysates. Cell lysates were prepared by harvesting cells into lysis buffer as described above. Captured proteins were then isolated by boiling in SDS-PAGE sample buffer and analyzed by SDS-PAGE and immunoblotting.
Mass Spectrometric Sequencing of Fe65-Fe65 was sequenced by on-line liquid chromatography tandem mass spectrometry (LC/MS/MS) as recently described by us (42). Briefly, Fe65 was isolated by immunoprecipitation from Fe65 and Fe65ϩc-Abl⌬XB-co-transfected cells using antibody 9B11 to the Myc tag on Fe65 and resolved by SDS-PAGE. The bands corresponding to Fe65 were excised, reduced, alkylated, and digested with either trypsin, Asp-N, Lys-C, or chymotrypsin (Roche Applied Science) and extracted from the gel pieces with two wash cycles of 50 mM NH 4 HCO 3 and acetonitrile, lyophilized, and resuspended in 20 l of 50 mM NH 4 HCO 3 .
Chromatographic separations were performed using an Ultimate LC system (Dionex). Peptides were ionized by electrospray ionization using a Z-spray source fitted to a QTof-micro (Micromass). The instrument was set to run in automated switching mode, selecting precursor ions based on their intensity and charge state, for sequencing by collisioninduced fragmentation. The MS/MS analyses were conducted using collision energy profiles that were chosen based on the mass/charge (m/z) and the charge state of the peptide and optimized for phosphorylated peptides.
The mass spectral data were processed into peak lists containing the m/z value of each precursor ion and the corresponding fragment ion m/z values and intensities. Data were searched against a custom built data base containing the full-length sequence of Fe65 using the Mascot searching algorithm (Matrix Science). Peptides and phosphopeptides of Fe65 were identified by matching the MS/MS data against mass values generated from the theoretical fragmentation of peptides based on the search criteria set (i.e. the cleavage enzyme used with up to 3 missed cleavages, carbamidomethyl modification of cysteine residues, oxidized methionine, deamidation of asparagine and glutamine residues, and N-acetylation of the protein). Phosphorylated peptides were identified by selecting for serine/threonine and tyrosine phosphorylation as a variable modification. The exact location of phosphorylation within each peptide was determined by the pattern of fragment ions produced.
Luciferase Assays-Luciferase assays were performed using a Dual-Glo luciferase assay system according to the manufacturer's instructions (Promega). Briefly, cells were harvested into Glo lysis buffer (Promega) 24-h post-transfection and the lysates then transferred to a 96-well luminometer plate (Wallac). An equal volume of Dual-Glo luciferase substrate was added and firefly luciferase activities produced by the GAL4UAS-dependent firefly luciferase reporter plasmid pFR-Luc measured using a Wallac Trilux luminometer. Renilla luciferase activities produced by the phRL-TK transfection efficiency control plasmid were then assayed by adding an equal volume of Dual-Glo Stop&Glo substrate (comprising the stop solution for firefly luciferase and substrate for Renilla luciferase) and remeasuring in the luminometer. All luciferase transfections received the same number and amount of plasmids, which was achieved by transfection of vector pCIneo-CAT where appropriate; pCIneo is the vector used for expression of Fe65 in these assays. Firefly luciferase activities were standardized to the cor-responding Renilla luciferase activities and statistical analyses performed using one-way analysis of variance tests. Results shown were obtained using CHO cells, but similar data were obtained using COS-7 cells.
In Vitro Phosphorylation of Fe65 by c-Abl⌬XB-c-Abl⌬XB was isolated from transfected CHO cells by immunoprecipitation using c-Abl antibody 24-11. For in vitro phosphorylation of recombinant GST fusion proteins, 1 g of each substrate was incubated with immunoprecipitated kinase (prepared from 200 g of precleared lysate) in 25 mM HEPES pH 7.5 containing 20 mM MgCl 2 , 20 mM ␤-glycerophosphate, 20 mM p-nitrophenylphosphate, 0.1 mM sodium orthovanadate, 2 mM dithiothreitol, 0.185 MBq [␥-32 P]ATP, and 20 M ATP for 20 min at 30°C in a final volume of 40 l. Reactions were stopped by addition of SDS sample buffer and heating in a boiling water bath. Samples were separated on 10% (w/v) acrylamide SDS-PAGE gels, and the gels then subjected to autoradiography.

RESULTS
We initially confirmed that Fe65 interacts with c-Abl and a dominantly active mutant of c-Abl (c-Abl⌬XB) using immunoprecipitation and GST pull-down assays. Fe65 from Fe65ϩc-Abl-or Fe65ϩc-Abl⌬XB-transfected CHO was immunoprecipitated using the Myc tag on Fe65 and bound c-Abl detected on immunoblots. Fe65 interacted with both c-Abl and c-Abl⌬XB but the binding to c-Abl⌬XB was stronger (Fig. 1A). In complementary experiments, GST or GST-Fe65WW domain baits were used in pull-down assays from c-Abl-or c-Abl⌬XB-transfected CHO cells. Again, both c-Abl and c-Abl⌬XB bound Fe65 with the c-Abl⌬XB interaction stronger (Fig. 1B). Mutation of the conserved aromatic residues Tyr-Tyr-Trp within the Fe65WW domain to Ala-Ala-Ala inhibited binding of c-Abl⌬XB (Fig. 1C). We also monitored c-Abl expression by immunostaining of Fe65ϩc-Abl-and Fe65ϩc-Abl⌬XB-co-transfected CHO and COS-7 cells. Fe65, c-Abl, and c-Abl⌬XB were all present in both the cytoplasm and nuclei of these cells, and Fe65 and c-Abl/c-Abl⌬XB showed a marked overlap in their distributions (Fig. 1D). These results are in agreement with previous observations, which demonstrate that Fe65 binds to c-Abl through its WW domain, that this interaction is stronger with active isoforms of c-Abl, and that a proportion of both Fe65 and c-Abl are present within the nucleus (27,30).
We next examined the effect of c-Abl⌬XB on APP/Fe65-mediated transcription. To do so, we utilized a previously described GAL4-dependent reporter system that involves monitoring the transcriptional activity of APP-GAL4 DNA binding domain fusion genes using a GAL4UAS-luciferease reporter (19). GAL4-APP fusions involved either the GAL4 DNA binding domain fused to the C-terminal domain of APP (GAL4-APPc) or full-length APP into which the GAL4 DNA binding domain had been engineered (GAL4-APP) (19). Both GAL4-APPc and GAL4-APP transcription was stimulated by transfection of Fe65, and this stimulation was enhanced further by co-transfection with c-Abl⌬XB (Fig. 2). c-Abl⌬XB also stimulated transcription of the reporter in the absence of co-transfected Fe65, although to a lesser extent (Fig. 2). This effect of c-Abl⌬XB in the absence of co-transfected Fe65 may well involve endogenous Fe65, because Fe65 is expressed in CHO and COS-7 cells (42). Mutation of the conserved aromatic residues Tyr-Tyr-Trp within the Fe65WW domain to Ala-Ala-Ala markedly inhibited its stimulatory effect (Fig. 2). This finding is consistent with the GST pull-down assays, which showed that this mutation inhibited binding of the WW domain to c-Abl⌬XB (Fig. 1C).
Tyrosine 682 within the intracellular C-terminal domain of APP has been shown to be phosphorylated by active c-Abl (27), and so one possibility is that the effect of c-Abl⌬XB on APP/ Fe65-mediated transcription is due to phosphorylation of this residue. We therefore confirmed that c-Abl⌬XB phosphorylated APP Tyr-682 by monitoring the reactivity of APP and a mutant of APP in which tyrosine 682 was altered to phenylalanine with antibody 4G10 that detects phosphotyrosines. APP immunoprecipitated from APPϩc-Abl⌬XB-but not APP-transfected cells was reactive with antibody 4G10, and this reactivity was abolished by mutation of APP Tyr-682 (Fig. 3A). Thus, we confirmed that APP Tyr-682 is phosphorylated in APPϩc-Abl⌬XBtransfected cells as previously described (27).
An alternative possibility is that c-Abl also phosphorylates Fe65, and the stimulatory effect of c-Abl⌬XB on transcription is caused by Fe65 phosphorylation. We therefore isolated Fe65 by immunoprecipitation from Fe65-or Fe65ϩc-Abl⌬XB-co-transfected cells, and it was sequenced by mass spectrometry. We obtained over 80% sequence coverage and detected one peptide with a phosphorylated tyrosine residue (Fig. 3B). This peptide contains two adjacent tyrosines (tyrosines 546/547), although despite repeated sequence runs, we were unable to unambiguously distinguish which tyrosine was phosphorylated. We therefore prepared mutants of Fe65 in which either of these tyrosines were mutated to phenylalanine to preclude phosphorylation and examined their reactivities with antibody 4G10 in Fe65ϩc-Abl⌬XB-co-transfected cells. Immunoprecipitated Fe65 Y546F but not Fe65 Y547F was reactive with 4G10 in these assays (Fig. 3C). To confirm that no other tyrosines in Fe65 were phosphorylated in the Fe65ϩc-Abl⌬XB-co-transfected cells, we mutated individually, the remaining seven tyrosines to phenylalanine, and in a similar fashion, monitored the reactivities of these mutants with antibody 4G10. Mutation of these other tyrosines had no effect on 4G10 labeling (Fig.  3C). Thus, Fe65 is phosphorylated on a single residue, tyrosine 547, in c-Abl⌬XB-transfected cells.
The above studies are consistent with a direct phosphorylation of Fe65 by c-Abl⌬XB but do not eliminate the possibility that c-Abl⌬XB activates some other tyrosine kinase that then phosphorylates Fe65. To determine whether c-Abl⌬XB can directly phosphorylate Fe65, we therefore performed in vitro phosphorylation assays using the second Fe65-PTB domain (which contains tyrosine 547) prepared as a GST fusion protein substrate. These assays revealed that c-Abl⌬XB could directly phosphorylate Fe65 (Fig. 3D). Thus c-Abl can directly phosphorylate both APP (on tyrosine 682) and Fe65 (on tyrosine 547).
To examine whether the stimulatory effect of c-Abl⌬XB on APP/Fe65-mediated transcription involves phosphorylation of either APP Tyr-682 or Fe65 Tyr-547 , we performed further GAL4-APPc-dependent transcription assays using mutants in which these residues were altered to phenylalanine to preclude phosphorylation. Fe65 and c-Abl⌬XB both stimulated transcription of mutant GAL4-APPc Y682F , and this stimulation was greater than that of wild-type GAL4-APPc (Fig. 4A). Others have also shown that mutation of APP 682 to phenylalanine enhances transcriptional activity of GAL4-APP fusions (43). The mechanisms that underlie this effect are not known but mutation of APP 682 to phenylalanine does not influence either APP/Fe65 or APP/JIP-1 interactions (Ref. 43 and see below). Thus, the stim-ulatory effect of c-Abl⌬XB on APP/Fe65 transcription cannot be through phosphorylation of APP 682 .
We next tested whether the stimulatory effect of c-Abl⌬XB on APP/Fe65 transcription was through phosphorylation of Fe65 Tyr-547 . Mutation of this residue to phenylalanine to preclude phosphorylation (Fe65 Y547F ) completely eliminated the effect of c-Abl⌬XB (Fig. 4A). However, Fe65 Y547F was still capable of stimulating GAL4-APPc-dependent transcription (Fig.  4A). This latter observation suggests that a component of the Fe65 stimulatory effect on GAL4-APPc-mediated transcription is not dependent upon phosphorylation of Fe65 Tyr-547 . It also suggests that mutation of Fe65 Tyr-547 to phenylalanine does not induce some conformational change to this domain of the protein that precludes the ability of Fe65 to stimulate transcription from GAL4-APPc. This was further confirmed by monitoring the effect of mutating the adjacent tyrosine (Fe65 Tyr-546 ) to phenylalanine (Fe65 Y546F ). This mutant behaved in a similar fashion to wild-type Fe65 (Fig. 4A). We also tested whether mutation of Fe65 Tyr-547 inhibited transcription from full-length GAL4-APP and obtained similar results to GAL4-APPc (Fig.  4B). Thus, the stimulatory effect of c-Abl⌬XB on APP/Fe65 transcription is through phosphorylation of Fe65 Tyr-547 .
Tyrosine 547 resides within the second PTB domain of Fe65 and because this domain binds APP, it is possible that its phosphorylation influences Fe65-APP interactions in some manner. We therefore performed immunoprecipitation assays to determine whether c-Abl⌬XB altered binding of APP and GAL4-APPc to Fe65. Fe65 isolated from Fe65ϩAPPor Fe65ϩAPPϩc-Abl⌬XB-transfected cells bound identical amounts of APP (Fig. 5A). Likewise, the presence of c-Abl⌬XB did not influence the amounts of GAL4-APPc bound to Fe65 (Fig. 5B). To determine whether c-Abl⌬XB influenced binding of APP to endogenous Fe65, we immunoprecipitated APP from APP-or APPϩc-Abl⌬XB-transfected cells and probed for bound Fe65. However, we again could detect no differences in the amounts of co-immunoprecipitating Fe65 (Fig. 5C). We also probed these samples with antibody 4G10, and this demonstrated that at least a proportion of Fe65 that was bound to APP was tyrosine-phosphorylated.
We next tested whether the Fe65 Y547F and APP Y682F mutations altered binding of APP and Fe65, respectively, in immunoprecipitation experiments. However, neither of these mutants had altered binding properties (Fig. 5D). Mutant APP Y682F has previously been shown to bind Fe65 (43). Thus, c-Abl⌬XB phosphorylates Fe65 within its second PTB domain to stimulate APP/Fe65 transcriptional activity, but this stimulation does not appear to be through an overt effect on Fe65/ APP interactions.
Recently, APPc has been shown to induce expression of the GSK3␤ gene (44). We therefore asked whether c-Abl⌬XB stimulated the APPc-dependent expression of GSK3␤ by analyzing GSK3␤ protein levels in cells transfected with APPc either alone or with Fe65 and c-Abl⌬XB using immunoblotting techniques. Because we obtain 30 -40% transfection efficiencies, any changes observed in these pooled samples of transfected and non-transfected cells are likely to be less than that seen in individual transfected cells. Nevertheless, although we observed little change to GSK3␤ protein levels following co-transfection of APPc with Fe65 or c-Abl⌬XB alone, we detected a marked increase in GSK3␤ signal in cells transfected with all three plasmids (Fig. 6A). We also studied the effect of c-Abl⌬XB on GSK3␤ protein levels by immunostaining, and this revealed that transfection of APPcϩFe65ϩc-Abl⌬XB increased the GSK3␤ signal compared with non-transfected cells (Fig. 6B). DISCUSSION The functions of APP are not properly understood. However, several recent studies have demonstrated that the C-terminal domain of APP, produced by ␥-secretase activity, can translocate to the nucleus to regulate transcriptional events (19 -24, 43, 45, 46). One APP binding partner that is involved in this process is the adaptor protein Fe65 (19,(21)(22)(23)(24)43). Fe65 is present within the nucleus and, aside from APP, binds to at least two transcription factors, CP2/LSF/LBP1 and Tip60 (19,25). The Fe65WW domain is required for its stimulatory effect on APP-mediated transcription (19) and also for nuclear translocation of Fe65 (28). This suggests that WW domain ligands are required for the nuclear functions of Fe65. One Fe65WW domain ligand is the tyrosine kinase c-Abl (25). c-Abl phosphorylates APP on tyrosine 682 (25) and here, we demonstrate that it additionally phosphorylates Fe65 on tyrosine 547. We also show that active c-Abl stimulates APP/Fe65-mediated transcription and that this is through phosphorylation of Fe65 Tyr-547 but not APP Tyr-682 .
Tyrosine phosphorylation of the C-terminal domain of APP does not influence APP binding to Fe65, although it can modulate interactions with ShcA and disabled, two other PTB domain proteins (5,10,14,47). The residue within Fe65 that is phosphorylated by active c-Abl (tyrosine 547) resides within the second PTB domain, and this domain binds to APP. However, Fe65 and Fe65 Y547F (which cannot be phosphorylated) both bound APP equally well in immunoprecipitation experiments, and we could detect no changes in Fe65/APP or Fe65/ APPc-GAL4 interactions in cells co-transfected with c-Abl⌬XB. As such, phosphorylation of APP Tyr-682 or Fe65 Tyr-547 by active c-Abl appears not to influence APP/Fe65 interactions in any overt manner. Thus, while the stimulatory effect of c-Abl⌬XB on APP/Fe65 transcription is mediated by Fe65 Tyr-547 phosphorylation, this does not involve any marked changes in binding of APP to Fe65.
Platelet-derived growth factor has recently been shown to induce ␤,␥-secretase cleavage of APP and this was characterized by monitoring the transcriptional activity of an APP-GAL4 FIG. 5. Neither c-Abl⌬XB nor mutation of either Fe65 Y547F or APP Y682F alters APP-Fe65 interactions. A, equal amounts of APP in Fe65 immunoprecipitates from Fe65ϩAPP-and Fe65ϩAPPϩ c-Abl⌬XB-transfected cells. B, similar experiment from Fe65ϩGAL4-APPc-and Fe65ϩGAL4-APPcϩc-Abl⌬XB-transfected cells. In both A and B, Fe65 was immunoprecipitated using the Myc tag and coimmunoprecipitated APP/GAL4-APPc and c-Abl⌬XB were detected with APPab and c-Abl antibodies, respectively. Fe65 was detected using a polyclonal antibody. C, equal amounts of endogenous Fe65 bound to APP in APP-and APPϩc-Abl⌬XB-transfected CHO cells. APP was immunoprecipitated and detected with antibody 22C11 and coimmunoprecipitated Fe65 and c-Abl⌬XB detected as above; tyrosine-phosphorylated Fe65 in the APP immunoprecipitates was detected with antibody 4G10. D, equal amounts of APP in Fe65 immunoprecipitates from Fe65ϩAPP-, Fe65 Y546F ϩAPP (Fe65Y546FϩAPP)-, Fe65 Y547F ϩAPP (Fe65Y547FϩAPP)-, and Fe65ϩAPP Y682F (Fe65ϩAPPY682F)-transfected cells. Fe65 and APP were immunoprecipitated and detected as in A and B. (ϩ) and (Ϫ) refer to the presence or absence of immunoprecipitating antibodies in the assays.
fusion gene similar to the ones we used here (38). In the course of this study, active c-Abl was shown to have no effect on APP-GAL4-mediated transcription (38). However, the APP-GAL4 fusion used in this earlier work comprised full-length GAL4 (containing both its DNA binding and transactivation domains) such that while transcription of the reporter gene required APP cleavage, it did not require the transactivation capability of the APP C-terminal domain. In contrast, our GAL4-APP and GAL4-APPc fusions involve only the GAL4 DNA binding domain, and the APP C-terminal domain is thus essential for transcriptional activity (see also Ref. 19). Thus, these different findings are most probably the result of the different GAL4-APP fusions used in these two studies. Indeed, the earlier work of Gianni et al. (38) was an elegant study aimed at understanding the mechanisms regulating APP cleavage and not APP/Fe65-mediated transcription as in our work. Nevertheless, we tested whether c-Abl⌬XB also stimulated reporter gene activity that was driven by an APP-GAL4 DNA binding domain fusion in which GAL4 DNA binding domain sequences were fused to the C terminus of APP (APP-GAL4). c-Abl⌬XB also stimulated transcription from this fusion protein (APP-GAL4 versus APP-GAL4ϩc-Abl⌬XB transcription, 1:1.65; p Ͻ 0.001).
The precise mechanisms by which phosphorylation of Fe65 by c-Abl stimulates its transcriptional activity are not clear. Because the phosphorylated residue (tyrosine 547) is within the second Fe65 PTB domain, we initially anticipated that it might influence Fe65/APP interactions in some manner. However, our immunoprecipitation assays have not provided exper-imental evidence to support this notion. Nevertheless, it remains likely that phosphorylation of Fe65 (and perhaps APP) by c-Abl somehow alters the various protein-protein interactions of the Fe65/APP transcriptional complex. One way to gain insight into this would be to solve the structure of non-phosphorylated and phosphorylated Fe65 bound to the C terminus of APP.
Tyrosine 547 is located toward the N terminus of the Fe65 PTB domain and falls within the motif Tyr-Leu-Gly. This motif is conserved in a number of other PTB-bearing proteins including Fe65like-1 and like-2 (Fe65like-2 contains the motif Tyr-Met-Gly), X11␣, X11␤, Shc, and Numb. We are unaware of any report describing phosphorylation of the homologous tyrosine in these other PTB domain proteins. However, our finding that the Fe65 PTB domain is phosphorylated raises the possibility that these other proteins may also be tyrosine-phosphorylated on their PTB domains. Tyrosine phosphorylation of PTB domains might therefore be a novel mechanism for regulating the function of this class of protein.
Mis-metabolism of APP is believed to be central to the pathogenesis of Alzheimer's disease. Altered APP processing leading to increased production of A␤ is one favored pathogenic event but such changes are also likely to influence APP/Fe65 nuclear signaling, and this too may contribute to the neurodegenerative process (48). Indeed, familial Alzheimer's disease mutant presenilin-1 has recently been shown to have altered nuclear signaling function and this has been causally related to familial forms of Alzheimer' s disease (49). Thus, defective phosphorylation of Fe65 by active c-Abl may alter APP/Fe65 nuclear signaling, and this might also contribute to Alzheimer's disease.