Hyperphosphorylation of JNK-interacting Protein 1, a Protein Associated with Alzheimer Disease*S

The c-Jun N-terminal kinase (JNK) group of mitogen-activated protein (MAP) kinases are activated by pleiotropic signals including environmental stresses, growth factors, and hormones. JNK-interacting protein 1 (JIP1) is a scaffold protein that assembles and facilitates the activation of the mixed lineage kinase-dependent JNK module and also establishes an interaction with β-amyloid precursor protein that has been partially characterized. Here we show that, similarly to other proteins involved in various neurological diseases, JIP1 becomes hyperphosphorylated following activation of stress-activated and MAP kinases. By immobilized metal affinity chromatography and a combined microcapillary LC/MALDI-TOF/ESI-ion trap mass spectrometry approach, we identified 35 sites of mitotic phosphorylation within JIP1, among which eight were present within (Ser/Thr)-Pro sequence. This motif is modified by various kinases in aggregates of the microtubule-associated protein tau, which generates typical intraneuronal lesions occurring in Alzheimer disease. Most of the post-translational modifications found were located within the JNK, MAP kinase kinase, and RAC-α Ser/Thr protein kinase binding regions; no modifications occurred in protein Src homology 3 and phosphotyrosine interaction domains, which are essential for binding to kinesin, β-amyloid precursor protein, and MAP kinase kinase kinase. Protein phosphorylation is known to affect stability and protein-protein interactions. Thus, the findings that JIP1 is extensively phosphorylated after activation of stress-activated and MAP kinases indicate that these signaling pathways might modulate JIP1 signaling by regulating its stability and association with some, but not all, interacting proteins.

Alzheimer disease (AD), 1 the most common senile dementia, is characterized by amyloid plaques, vascular amyloid, neurofibrillary tangles (NFTs), and progressive neurodegeneration. Amyloid plaques are mainly composed of A␤ peptides (A␤40 and A␤42), which are derived from processing of the ubiquitous type I transmembrane ␤-amyloid precursor protein (APP) by various secretases (1)(2)(3)(4). A pathogenic role for APP processing in AD has been ascertained by the finding that mutations in APP (5) and presenilins (6 -9), key components of the ␥-secretase, cause AD autosomal dominant familial forms. In AD, APP is first cleaved at the plasma membrane by ␤-secretase (10) releasing the APP␤ ectodomain extracellularly or into intracellular compartments, whereas its C-terminal fragment of 99 amino acids (C99) remains membrane-bound. Then C99 is cleaved by ␥-secretase, generating amyloidogenic A␤ peptides and an intracellular product named APP intracellular domain (AID) (11)(12)(13). In the non-amyloidogenic proteolytic pathway, APP is first processed by ␣-secretase in the A␤ peptide sequence leading to the production of APP␣ and a membrane-bound C-terminal fragment of 83 amino acids (C83). Then C83 is cleaved by ␥-secretase into P3 and AID fragments. Although A␤ peptide is implicated in the pathogenesis of AD, AID mediates most of the APP signaling functions. In fact, the AID fragment is able to activate expression of various genes in combination with its binding partner Fe65 (12) or repress it (14,15). Indeed it has been shown that upon overexpression of APP, Fe65 and Tip60, a repressor complex assembled on the KAI-1 promoter is replaced by a complex containing AID, Fe65, and Tip60. All these reports reinforce the role for AID in gene regulation. Moreover, eight potential phosphorylation sites were predicted on APP cytoplasmic domain. A detailed MS analysis of AD and control brains demonstrated that this region became hyperphosphorylated following the disease (16). This modification seems to affect A␤ peptide generation.
Another pathological feature of AD is the presence of intraneuronal lesions known as NFTs. These are generated by hyperphosphorylated aggregates of the microtubule-associated protein tau and coincide with cell death in several neu-rodegenerative diseases known as tauopathies (17,18). Mutations in the tau gene have been linked to NFT formation in a number of tauopathies in which no ␤-amyloid deposits develop (19,20). However, no tau gene mutations have been found in AD, and the relationship between tau pathology, amyloid deposition, and neuronal degeneration is still not clear (21,22). Rampant hyperphosphorylation of tau seems to be the key event associated with NFT formation (17,18). A complete characterization of tau phosphorylation sites in normal and AD brain required a massive analysis by different mass spectrometric methodologies (23)(24)(25). Around 30 modified Ser/Thr residues were identified in tau samples from AD brain, among which a dozen are present in (Ser/Thr)-Pro sites. These motifs, similar to those occurring in other diseaserelated proteins, are subjected to physiological isomerization and dephosphorylation events (controlled by the peptidylprolyl cis/trans isomerase Pin1 and the phosphatase PP2A, respectively), which regulate protein conformation/function and NFT formation (18,26,27). A growing body of evidence suggests that inappropriate activation of the cell cycle, especially mitotic events, may contribute to this abnormal hyperphosphorylation and thereby play an important role in the development of the disease (18, 28 -33).
Previously we have identified JNK-interacting protein 1 (JIP1) as an APP-interacting protein (34,35). JIP1 was initially identified as an inhibitor of JNK activation but was soon shown to be a scaffold protein that bound various components of the JNK cascade including MKK7, mixed lineage kinase, dual leucine zipper kinase, and other upstream kinases (36). The significance of this APP-JIP1 interaction is still unknown. It is tempting to speculate that JIP1-APP interaction could provide the molecular basis for activation of signaling cascades involving mitogen-activated and stress-activated kinases in the brain of AD patients that may ultimately be responsible for tau hyperphosphorylation and NFT formation (37)(38)(39)(40)(41). Preliminary site-directed mutagenesis experiments demonstrated that JIP1 is also phosphorylated, and its modification regulates JNK module dynamics and activation (42). In an attempt to define a possible role of JIP1 in hyperphosphorylation events present in AD cases, which are accompanied by activation of various kinases, we studied the phosphorylation of JIP1 in response to activation of MAP and stress-activated kinases. This investigation, based on the massive use of some recent MS approaches developed to study protein phosphorylation, demonstrates that JIP1 is also hyperphosphorylated.

EXPERIMENTAL PROCEDURES
Reagents-Monoclonal anti-APP N-terminal antibody 22C11 was purchased from Chemicon, and anti-␣-tubulin DM 1A was obtained from Sigma. Polyclonal anti-APP C-terminal antibody was purchased from Zymed Laboratories. Polyclonal anti-JNK or -phosphorylated JNK, anti-ERK or -phosphorylated ERK, and anti-p38 or -phosphorylated p38 were obtained from Cell Signaling Technology. Anti-FLAG was from Sigma, and anti-phospho-Tyr and anti-phospho-Thr anti-bodies were from Cell Signaling Technology. Sequencing grade proteolytic enzymes and phosphatases were purchased from Roche Applied Science. Ultrapure HPLC grade solvents were from Baker. All chemical reagents were purchased from Sigma.
Immunoprecipitation and Immunoblotting Analysis-Twenty-four to 48 h following transfection, cells were washed with ice-cold PBS and lysed in lysis buffer containing a protease inhibitor tablet (Roche Applied Science). Lysis was allowed to continue for 10 min on ice, and cells were centrifuged at full speed at 4°C for 10 min. Some lysate representing the total lysate was removed and boiled with SDS loading buffer containing DTT, whereas the rest was immunoprecipitated for 2 h at 25°C with a monoclonal antibody directed against the FLAG epitope already bound to agarose beads (Sigma). The beads were washed five times with lysis buffer and boiled with SDS loading buffer with DTT. Immunoprecipitates and total lysates were separated by SDS-PAGE and blotted onto nitrocellulose (Bio-Rad). Membranes were probed with the same antibodies as were used for immunoprecipitation, and horseradish peroxidase-conjugated secondary antibodies were used. Proteins were detected using the Supersignal West Pico chemiluminescence system (Pierce).
In Vivo Orthophosphate Labeling of Cells-HEK293 cells were transfected with the indicated construct. After 24 h, cells were washed three times with phosphate-free Dulbecco's modified Eagle's medium and incubated at 37°C in the same medium for 30 min. Cells were then incubated in phosphate-free Dulbecco's modified Eagle's medium containing 0.5% dialyzed fetal bovine serum and 1 mCi/ml [␥-32 P]ATP orthophosphate (Amersham Biosciences) for 4 h at 37°C.
Phosphoamino Acid Analysis-Cellular extracts were resolved by SDS-PAGE and transferred to a PVDF membrane; 32 P-labeled JIP1 was identified by autoradiography and excised from the membrane. Blotted protein was subjected to vapor-phase hydrolysis with 6 N HCl at 110°C for 3 h (43). Two-dimensional phosphoamino acid analysis was carried out on thin layer cellulose plates using the Hunter thin layer electrophoresis system (44). The pattern of phosphoamino acids was visualized by exposing the dried cellulose plates to a model Storm-860 PhosphorImager plate (Amersham Biosciences).
LC-ESI-Q-MS Analysis-Aliquots of protein extracts containing human JIP1 were analyzed by using an API-100 single quadrupole mass spectrometer (PerkinElmer Sciex) equipped with an electrospray source connected to a Phoenix 40 pump (ThermoFinnigan). Protein extracts were separated on a capillary Hypersil-Keystone Aquasil C 4 Kappa column (100 ϫ 0.32 mm, 5 m) using a linear gradient from 2 to 60% acetonitrile in 0.1% (v/v) formic acid over 80 min at a flow rate of 5 l/min. Spectra were acquired in the range 200 -2000 m/z as reported previously (45). Data were analyzed using the BioMultiView 3.1 software provided by the manufacturer. Mass calibration was performed by means of the multiply charged ions from a separate injection of horse heart myoglobin (molecular mass, 16,951.5 Da). Masses are reported as average values.
In Situ Digestion with Proteolytic Enzymes-Samples containing 10 -50 g of total proteins were subjected to electrophoresis under denaturing conditions on 12% polyacrylamide gels (46). Separated bands were visualized by colloidal Coomassie G250 staining. JIP protein was excised from the gel, triturated, in-gel reduced, S-alkylated, and digested either with trypsin or endoprotease AspN, as previously reported (47). Gel particles were extracted with 25 mM NH 4 HCO 3 /acetonitrile (1:1 v/v) by sonication and peptide mixtures were concentrated.
LC Analysis-Aliquots of whole JIP1 digests or phosphopeptideenriched samples were resolved on a Phoenix 40 LC system (Ther-moFinnigan, USA). Peptides were separated on a microbore Jupiter Proteo C12 column (150 ϫ 0.5 mm, 4 m) using a linear gradient from 2% to 60% of acetonitrile in 0.1% (v/v) trifluoroacetic acid, over 80 min, at flow rate of 20 l/min. Individual components were collected manually and freeze-dried.

FIG. 1. Anisomycin induces stress-and mitogen-activated kinase-mediated phosphorylation of human JIP1 on Ser and Thr residues.
A, Western blotting analysis of HEK293 cells transfected with a construct coding for FLAG-tagged JIP1 using anti-␣-tubulin, anti-JNK or -phosphorylated JNK, anti-ERK or -phosphorylated ERK, and anti-p38 or -phosphorylated p38 antibodies. Extracts from cells treated (ϩ) or not (Ϫ) with anisomycin are shown in comparison. B, HEK293 transfected cells before (TL) and following (IP␣FLAG) JIP1 purification with anti-FLAG (␣FLAG) antibody-agarose beads were assayed by Western blotting with anti-FLAG, anti-phospho-Tyr (␣Y p ), and anti-phospho-Thr (␣T p ) antibodies. C, SDS-PAGE analysis of human JIP1 from HEK293 transfected cells following purification with anti-FLAG antibody bound to agarose beads; a Coomassie-stained gel is shown. D, qualitative phosphoamino acid analysis of human JIP1. A sample of SDS-PAGEresolved 32 P-labeled JIP1 was electrotransferred onto PVDF membranes, subjected to acid hydrolysis, and analyzed by two-dimensional cellulose thin layer electrophoretic analysis (43,44). Anis., anisomycin; W.B., Western blot; IP, immunoprecipitation; TL, total; PSer, phospho-Ser, PThr, phospho-Thr; PTyr, phospho-Tyr. (48) using the dried droplet technique. Samples were analyzed with a Voyager-DE PRO spectrometer (Applera). Spectra were acquired in positive and negative polarity using the instrument in linear and reflectron mode. Internal or external mass calibration was performed with peptides derived from enzyme autoproteolysis or with molecular mass markers (Applera). Data were elaborated using the DataExplorer 5.1 software (Applera). Masses are reported as monoisotopic values.
LC-ESI-IT-MS-MS Analysis-Aliquots of whole JIP1 digests or phosphopeptide-enriched samples were analyzed by using a LCQ Deca Xp Plus mass spectrometer (ThermoFinnigan) equipped with an electrospray source connected to a Phoenix 40 pump (ThermoFinnigan). Phosphopeptides were separated on a capillary Hypersil-Keystone Aquasil C 18 Kappa column (100 ϫ 0.32 mm, 5 m) using a linear gradient from 2 to 60% acetonitrile in 0.1% (v/v) formic acid over 80 min at a flow rate of 5 l/min. Spectra were acquired in the range 200 -2000 m/z. The mass spectrometer was set up to automatically acquire a full MS scan followed by three MS 2 spectra of the three most abundant ions in the MS spectrum. Spectra were acquired using a dynamic exclusion of 3 min with a mass width of Ϯ2.0 m/z. Data were elaborated using the BioWorks 3.1 software provided by the manufacturer. Masses are reported as monoisotopic values.
Alkaline Phosphatase Treatment-Phosphopeptides were dephosphorylated by mixing 3 l of each sample solved in 50 mM NH 4 HCO 3 , pH 8.5, with 2 units of shrimp alkaline phosphatase (Roche Applied Science) solved in 2 l of 50 mM NH 4 HCO 3 . Reaction mixtures were incubated for 60 min at 37°C followed by addition of 1 l of 10% (v/v) formic acid. In the case of whole digests, peptide mixtures were desalted on ZipTip C 18 pipette tips (Millipore) before MALDI-TOF-MS analysis.
Phosphopeptides Subdigestion-Phosphopeptides solved in 5 l of 50 mM NH 4 HCO 3 , pH 8, were subdigested with 30 ng of endoprotease AspN, chymotrypsin, trypsin, or carboxypeptidase A. Reaction mixtures were incubated for 360 min at 37°C followed by addition of 1 l of 10% (v/v) formic acid. Digests were directly analyzed by FIG. 3. MALDI spectrum of human JIP1 phosphopeptides captured by the Ga 3؉ IMAC column before and after treatment with alkaline phosphatase. The spectrum of the trypsin digest before and after treatment with alkaline phosphatase is shown in A and B, respectively. The spectrum of the endoprotease AspN digest before and after treatment with alkaline phosphatase is shown in C and D, respectively. Metastable peaks formed by the loss of HPO 3 2Ϫ from phosphopeptides (62) are marked with an asterisk. Assignment of signals present in A and C to specific phosphopeptides was performed on the basis of the measured mass, protease specificity, presence of the amino acids susceptible to eventual post-translational modification in the peptide sequence, and signals observed in B and D.

MALDI-TOF-MS.
Data Analysis-Observed MALDI-TOF mass values were assigned to JIP1 phosphopeptides/peptides using the GPMAW 4.23 software (Lighthouse Data). This software generated a mass/fragment database output based on JIP sequence, protease selectivity, nature of the amino acid susceptible to eventual post-translational modification, and molecular mass of the modifying groups. Mass values were matched to protein regions using a 0.02% mass tolerance value. Sequest software allowed automated identification of JIP1 phosphopeptides/peptides using LC-ESI-IT-MS-MS data (49,50). This software was run against an indexed database containing human JIP1 and protease sequences by using Met oxidation and Ser/Thr phosphorylation as differential modifications. Values of 2.5 and 1.0 m/z were used as mass tolerance for parent and fragment ions, respectively. Candidates with a Sequest Xcorr value Ͼ2 were further evaluated by visual inspection of MS 2 spectra. GPMAW 4.23 software was also used to confirm peptide identification by tandem mass spectrometry data.
Sequence Analysis-A computer-assisted prediction of phosphorylation sites occurring in human JIP1 sequence was obtained by NetPhosK 1.0, NetPhos 2.0, KinasePhos, and GPS software (51)(52)(53)(54). Analysis was performed on the FLAG-tagged JIP1 construct coding sequence using program default parameters. Protein domain prediction was performed by the Prosite program (55) using default parameters.

Anisomycin Induces JIP1 Phosphorylation on Ser and Thr
Residues-HEK293 cells were transfected with a construct coding for FLAG-tagged JIP1 (35). Twenty-four hours after transfection, cells were treated with anisomycin and then lysed. To determine whether anisomycin treatment activated stress-and mitogen-activated kinases, lysates were analyzed for phosphorylation of p38, JNK, and ERK. In fact, phosphorylation of these proteins activates their kinase activity. As shown in Fig. 1A, anisomycin induced phosphorylation of all the kinases tested, while the total amounts of these proteins were not changed as compared with untreated controls.
Having determined that this experimental condition activated MAP and stress-activated kinases, we purified transfected JIP1 with an anti-FLAG antibody bound to agarose beads. Western blot with an anti-FLAG antibody showed that treated and untreated cells expressed equivalent levels of JIP1; this protein was efficiently immunoprecipitated from both samples (Fig. 1B). Moreover immunoblotting with an anti-phospho-Tyr and an anti-phospho-Thr antibody revealed that anisomycin treatment induced phosphorylation of JIP1 on Thr but not Tyr residues. As expected, a Coomassiestained SDS-PAGE gel of purified JIP1 revealed an abundant protein component with a molecular mass of almost 90 kDa, although polypeptide species presenting lower molecular masses were also detected (Fig. 1C). The qualitative nature of JIP1 phosphorylated residues was directly ascertained by phosphoamino acid analysis (43,44). A sample of 32 P-labeled JIP1 was resolved by SDS-PAGE, electrotransferred onto PVDF membranes, and subjected to acid hydrolysis. Twodimensional cellulose thin layer electrophoretic analysis definitively demonstrated the occurrence of phospho-Ser and phospho-Thr in JIP1 hydrolysates together with the absence of phospho-Tyr (Fig. 1D). To determine the extent to which cellular kinases phosphorylated JIP1 in vivo, the molecular mass of the intact protein was determined by LC-ESI-Q-MS measurements. The spectrum showed a heterogeneous distribution of protein molecular masses, each differing by 80 Da (Fig. 2). The mass difference with respect to the non-phosphorylated protein (theoretical average mass, 78,432.4 Da) and the range of the mass distribution suggested that each mole of JIP1 was modified with 8 -19 mol of phosphate. However, the presence of additional phosphate groups at low stoichiometry could not be ruled out as the spectrum was rather noisy. In fact, mass spectra of multiply phosphorylated proteins are inherently noisy relative to non-phosphorylated ones as the signal is being distributed across multiple forms, and metal ion adducts can occur. All these data underlined a reduced estimation of JIP1 phosphorylation (42).
Assignment of Phosphorylation Sites in JIP1-To locate phosphorylation sites in the JIP1 sequence, protein was excised from the gel, alkylated, and digested either with trypsin or endoprotease AspN; small digest aliquots were also treated with alkaline phosphatase (56,57). MALDI-TOF-MS spectra were obtained and compared, revealing the disappearance of peaks in untreated samples and concomitant appearance of new peaks at masses differing by multiples of Ϫ80 Da in treated samples. Thus, four and five phosphorylated peptides could have been identified from whole JIP1  trypsin and endoprotease AspN digests, respectively (data not shown). These numbers were far fewer than the minimum of 19 suggested by the molecular mass distribution of phosphorylated JIP1.
Because the ionization of phosphopeptides might be suppressed by the presence of large amounts of non-phosphorylated peptides in non-fractionated mixtures (58), both digests were enriched for phosphopeptides by IMAC resin (59 -61). The MALDI spectrum of the peptides from trypsin and endoprotease AspN digestion eluted from an IMAC column is shown in Fig. 3, A and C, respectively. In contrast to previous reports (62), a reduced number of poor satellite signals associated with metastable decomposition of phosphorylated peptides was observed. To identify phosphopeptides, we treated an aliquot of each IMAC fraction with alkaline phosphatase and compared the MALDI spectra taken before and after treatment (57). The spectra of the treated samples were greatly simplified as a result of the collapse of multiply phosphorylated peptides into a single non-phosphorylated peak (Fig. 3, B and D). Thus, assignment of signals to specific phosphorylated peptides was achieved on the basis of the measured mass, JIP1 sequence, protease selectivity, and nature of the amino acid susceptible to eventual post-translational modification. This analysis demonstrated that most of the signals observed in the spectrum of untreated digests were associated to mono-, di-, tri-, and tetraphosphorylated peptides, thus confirming the high efficiency of the adopted IMAC procedure in selectively enriching phosphopeptides.
However, because the ionization of highly phosphorylated species might be suppressed by the presence of large amounts of poorly phosphorylated peptides in non-fractionated mixtures, IMAC eluates were also resolved by LC, and the resulting fractions were analyzed by MALDI-TOF-MS in linear and reflectron modality as well as in negative and positive polarity (63). As an example, Fig. 4A shows the spectrum of fraction 25 from the JIP1 endoprotease AspN digest acquiredinpositivelinearmode.Inadditiontodi-totetraphosphorylated forms of peptide 190 -214 (already observed in Fig. 3C), mass spectrometric analysis revealed the concomitant occurrence of the penta-and hexaphosphorylated peptide species as also confirmed following digestion with alkaline phosphatase (Fig. 4B). Table I summarizes all phosphorylated peptides identified in the trypsin digest; their identity was confirmed following MALDI-TOF-MS analysis of their alkaline phosphatase digests. Peptides deriving from aspecific hydrolysis at Trp-188 and Phe-480 were also detected. Similarly Table II reports all phosphorylated peptides identified in the endoprotease AspN digest. All these results highlighted the importance of LC fractionation for detection of highly phosphorylated peptides. The eventual loss of small hydrophilic phosphopeptides from JIP1 digests during IMAC enrichment was also evaluated (62,64). However, MALDI-TOF analysis of both digests after purification on graphite powder microcolumns according to the improved procedure described by

Hyperphosphorylation of JNK-interacting Protein 1
Larsen et al. (64) did not allow detection of phosphopeptides other than those already reported in Tables I and II. Phosphopeptides identified by MALDI-TOF analysis were also analyzed by tandem mass spectrometry to confirm their identity and determine the site(s) of phosphorylation (56,58). Thus, IMAC eluates of trypsin and endoprotease AspN digests were subjected to LC-ESI-IT-MS-MS analysis (49,50,65). Phosphopeptides were easily recognized by examination of their positive mode MS 2 spectra for an abundant ion resulting from the neutral loss of 98 m/z (H 3 PO 4 ) for a singly charged, 49 m/z (H 3 PO 4 /2) for a doubly charged, 32.7 m/z (H 3 PO 4 /3) for a triply charged, or 24.5 m/z (H 3 PO 4 /4) for a quadruply charged ion for each phospho-Ser or phospho-Thr in the peptide (33,50,51,53). Most of the peaks observed in chromatograms were associated to phosphorylated peptides, confirming the high efficiency of the adopted IMAC procedure in selectively enriching phosphorylated species (Fig. 5A). In most cases, fragmentation data allowed assigning the exact sites of modification as clearly shown in the case of the tryptic peptide (201-229)P2 (Fig. 5, B and C). In total, sequencing data were obtained for 40 of 68 tryptic phosphopeptides already detected by MALDI-TOF-MS analysis (Table I and Supplemental Material 1); in addition, eight new phosphopeptides were identified. Similarly sequencing data were obtained for 54 of 103 phosphopeptides from the endoprotease AspN digest already detected by MALDI-TOF-MS analysis (Table II and Supplemental Material 2) with 12 new phosphopeptides identified. Tables I and II show all identified  To complete assignment of phosphorylation sites and confirm data derived by MALDI-TOF-MS and LC-ESI-IT-MS-MS analysis, selected peptide fractions from LC purification of trypsin and endoprotease AspN digests (see Tables I and II) were subjected to further digestion with various proteases and then subjected to MALDI-TOF-MS analysis. Table III summarizes all phosphorylated peptides identified in subdigestion mixtures; their identities were also ascertained following alkaline phosphatase treatment and further MALDI-TOF-MS analysis (data not shown). This approach allowed demonstration that phosphorylation occurred at Ser-152, Ser-366, Ser-469, Ser-471, Ser-472, and Ser-473. Thus, all data reported in Fig.  3 and Tables I, II, and III were consistent with each other and allowed the identifing of all phosphorylation sites present in vivo in human JIP1.
Trypsin and endoprotease AspN digests not subjected to IMAC enrichment were also analyzed by LC-ESI-IT-MS-MS procedures. As reported in the case of MALDI-TOF analysis, the majority of the most intense ions in chromatogram were assigned to non-phosphorylated peptides (data not shown). A combination of data deriving from MALDI-TOF-MS and LC-ESI-IT-MS-MS analyses allowed the obtaining of complete overall sequence coverage for JIP1 protein, also demonstrating that a partial modification occurred on all assigned phosphorylation sites. In fact, all peptides detected as phosphorylated species were also present as non-phosphorylated counterparts. This result was in perfect agreement with the extent of phosphorylation on intact protein revealed by ESI-MS measurement (Fig. 2).
A theoretical prediction of phosphorylation sites occurring in human JIP1 sequence was achieved by four independent computer-assisted methodologies (51)(52)(53)(54). This approach was chosen to consider the maximum number of kinases (eventually not included by a single program). Analysis revealed that human JIP1 sequence contains 118 putative modification sites recognized by a broad number of kinases; 20, 28, 38, and 32 sites were simultaneously ascertained by four, three, two, and one programs, respectively. Predicted phosphorylated Ser, Thr, and Tyr residues are reported in Fig. 6. In contrast, our mass spectrometry-based investigation showed the absence of Tyr modification, demonstrating the occurrence of 35 phosphorylation sites, among which eight were present in (Ser/Thr)-Pro motifs. All residues characterized as phosphorylated by previous site-directed mutagenesis studies (42) were confirmed by our investigation; in addition, 24 novel phosphorylation sites were identified, among which one (Ser-214) was not predicted by computer-assisted analysis.
Most post-translational modifications were located within the JNK, MAP kinase kinase, and RAC-␣ Ser/Thr protein kinase binding regions of JIP1 sequence not affecting protein Src homology 3, APP, and phospho-Tyr interaction domains (34 -36, 42, 55). DISCUSSION Progression through different phases of the cell cycle is controlled by highly regulated sequential activation and inactivation of cyclin-dependent kinases (28). Cyclin-dependent kinases belong to the family of Pro-directed kinases and also include ERKs, JNKs, and glycogen synthase kinase-3␤, which phosphorylate almost exclusively on (Ser/Thr)-Pro motifs. Therefore, phosphorylation of (Ser/Thr)-Pro motifs is the central mechanism governing progression through different phases of the cell cycle. At the G 2 -M transition, activation of cyclin B/Cdc2 leads to the phosphorylation of a number of proteins. Many of these phosphoproteins (such as Cdc25, Wee1, Myt1, and RNA polymerase II) are important mitotic regulators and also contain multiple (Ser/Thr)-Pro signatures (18,66). An intriguing feature common to these proteins is that they are phosphorylated on multiple Ser/Thr residues clustered at the regulatory domain of molecules during mitosis. Phosphorylation on multiple sites is necessary for activation/ inactivation of their function in M phase (67,68). A major mechanism that reverses these phosphorylation events is dephosphorylation by protein phosphatases. In this regard, PP2A is an important enzyme that dephosphorylates the conserved (Ser/Thr)-Pro motifs present in a number of mitotic proteins and plays an essential role in mitotic regulation (27).
Although the etiology of AD is variable, all AD cases exhibit a similar hierarchical and progressive pattern of relentless neuronal death in brain, resulting from disruption of cytoskeleton networks and formation of hallmark NFTs and A␤ peptide plaques. One biochemical hallmark of degenerating neurons is rampant tau hyperphosphorylation on multiple (Ser/Thr)-Pro motifs (17,(23)(24)(25). Similarly an abnormal hyperphosphorylation of (Ser/Thr)-Pro sites has been depicted as a distinct feature characterizing neurofilament protein NF-H and vimentin in many neurological diseases (AD, Parkinson disease, and amyotrophic lateral sclerosis) (33,69), tumor suppressor BARD1 and BRCA2 in breast cancer (32,70), epidermal growth factor receptor in mammary and squamous carcinomas (31), stathmin in fibroblast transformation (30), retinoid X receptor in apoptosis (29), and heat shock factor I in stress conditions (71). These findings emphasize that a deregulation of the processes controlling a proper protein phosphorylation should determine non-functional hyperphosphorylated species and result in other mitotic events, apoptosis, and ultimately cellular death. Strong support for the role of cell cycle mechanisms in AD comes from the findings that AD neurons contain multiple markers spanning different phases of the cell cycle (72). Furthermore Yang et al. (73) demonstrated recently that AD neurons may have synthesized DNA but fail to com-  plete M phase, so that the cells remain tetraploid. Probably the most striking feature common to the normal cell cycle and degenerating neurons of AD brain is that they share many mitotic events especially phosphorylation on certain (Ser/Thr)-Pro motifs in proteins such as tau. However, monoclonal antibodies recognizing independent phosphorylated epitopes in NFTs showed strong immunoreactivity with cells in mitosis but not in interphase in various cell types (18, 74 -76). Reciprocally the phosphospecific monoclonal antibody MPM-2 raised against mitotic cells specifically labeled affected neurons with aberrant activation of the mitotic Cdc2 kinase in AD brain but with no immunoreactivity in normal neurons (77,78). Of significance is the incorporation of mitotic phosphoepitopes into NFTs and their appearance prior to NFT formation in AD brain, suggesting their involvement in early stages of AD pathology. In addition, phosphorylation of APP and production of amyloidogenic fragments from APP are seen in normal mitotic cells (79). Finally there is a tight link between the mitotic mechanism and apoptosis; induction of mitosis in postmitotic neuronal cells leads inevitably to cell death (80,81). Indeed apoptosis is elevated in hippocampal brain tissue in AD (82,83). A closer comparison of normal mitotic cells with AD neurons has revealed not only many similarities but also some striking differences. The similarities are seen in the type of players involved, but the dissimilarities are observed in the timing of their appearance and their subcellular localization. For example, activation of cyclin B/Cdc2 occurs in the cytoplasm and then in the nucleus during normal mitosis, but such activation apparently occurs only in the cytoplasm of AD neurons (84 -86). It is suggested that these differences simply would not support normal mitosis and may be important determinants of the degenerative fate of neurons (87). Because of the central role of phosphorylation in the regulation of cellular processes, much effort has been focused on the development of methodologies for characterizing this protein modification. Nowadays mass spectrometry is the most used analytical technique to study protein phosphorylation (56,58). This modification is often substoichiometric, and phosphopeptide enrichment procedures are a necessary prerequisite for characterization of hyperphosphorylated proteins by modern MALDI and ESI mass spectrometric methods. As an example, an integrated use of these methodologies re- cently has allowed assignment of 10 phosphorylation sites in BARD1 tumor suppressor (32), 12 sites in human heat shock factor I (71), 19 sites in mitotic regulator Net1 (62), 33 sites in tau (23)(24)(25), and 38 sites in neurofilament protein NF-H (33). In this study, we used an integrated MALDI and ESI mass spectrometric approach to demonstrate that, under experimental conditions mimicking activation of mitogen-activated and stress-activated kinases, a protein able to interact with JNK, APP, and various components of the JNK cascade (34 -36), namely JIP1, is in vivo hyperphosphorylated. A total number of 35 phosphorylation sites were assigned on its polypeptide sequence, among which eight occurred in (Ser/Thr)-Pro motifs. Accordingly JIP1 can be considered among the most hyperphosphorylated of known proteins so far.
What is the biological function of JIP1 phosphorylation? This post-translational modification can affect protein function in distinct ways. In some cases, phosphorylation targets proteins for ubiquitination and degradation. Important examples are cell cycle proteins (88) and IB, the inhibitor of the transcription factor NFB (89). In others cases, phosphorylation generates docking sites for protein-protein interaction. This is the case for some phospho-Tyr residues that can create docking sites for proteins containing Src homology 2 domains (90) or for phospho-Ser/phospho-Thr binding modules, which include 14-3-3 proteins, WW domains, forkheadassociated domains, WD40 repeats, and the Polo-box domains of Polo-like kinases (91). As for JIP1, it is conceivable to postulate that phosphorylation of some residues might affect JIP1 stability and conformation or impact on the association of JIP1 with its various binding partners (34 -36, 42). A detailed analysis is required to address the significance of each JIP1 phosphorylation site, and it will be the topic of further investigation by site-directed mutagenesis.