Phosphorylation of 69-kDa Choline Acetyltransferase at Threonine 456 in Response to Amyloid- (cid:1) Peptide 1– 42*

Choline acetyltransferase synthesizes acetylcholine in cholinergic neurons. In the brain, these neurons are especially vulnerable to effects of (cid:1) -amyloid (A (cid:1) ) peptides. Choline acetyltransferase is a substrate for several protein kinases. In the present study, we demonstrate that short term exposure of IMR32 neuroblastoma cells expressing human choline acetyltransferase to A (cid:1) -(1–42) changes phosphorylation of the enzyme, resulting in increased activity and alterations in its interaction with other cellular proteins. Using mass spectrometry, we identified threonine 456 as a new phosphorylation site in choline acetyltransferase from A (cid:1) -(1–42)-treated cells and in purified recombinant ChAT phosphorylated in vitro by calcium/calmodulin-dependent protein kinase II (CaM kinase II). Whereas phosphorylation of choline acetyltransferase by protein kinase C alone caused a 2-fold increase in enzyme activity, phosphorylation by CaM kinase II alone did not alter enzyme activity. A 3-fold increase in choline acetyltransferase activity was found with coordinate phosphorylation of threonine in of with A in two phases of of the and associated with by and the by 10 and with CaM kinase II and protein kinase that from A (cid:1) and of the that two-dimensional acid In Vitro Parent Ion Scanning— Experiments were performed on a Q-TOF2 mass spectrometer (Micromass), equipped with a nanoflow source. The instrument was calibrated with [Glu 1 ]Fibrinopeptide B (Sigma), and following desalting on a ZipTip C18TM (4 (cid:4) l), the concentrated protein digest sample was loaded into a borosilicate capillary (type F; Micromass). A voltage between 600 and 1000 V was applied to the capillary in order to produce nanomolar flow. Parent ion scanning was performed over the m / z range of 300–1500 monitoring neutral loss of 97.9769 and 49.9885 with a collision energy of 32 V. Masses that resulted in such neutral losses were subsequently further fragmented to obtain sequence information. Resultant spectra were background- subtracted and deconvoluted using MaxEnt3 software provided in the Micromass MassLynx version 3.5 software package. The sequence of the peptides was determined using PepSeq version 3.3 software also provided in the MassLynx 3.5 software. Peptide Sequencing by Mass Spectrometry— Amino acid sequences of tryptic phosphopeptides of ChAT isolated from control and A (cid:1) -treated cells or for identification of unknown proteins that co-immunoprecipi-tate with ChAT were obtained by mass spectrometry performed on a Micromass Q-TOF2 mass spectrometer equipped with a nanospray source and an online Waters CapLC (Waters). In all cases, 1 (cid:4) l of sample was injected from the autosampler. The instrument was calibrated with [Glu 1 ]Fibrinopeptide B (Sigma). A gradient consisting of 5–65% B in 15.5 min (A (cid:6) 0.1% formic acid, B (cid:6) acetonitrile with 0.1% formic acid) flowing at 1 (cid:4) l/min was used to elute peptides from a 300- (cid:4) m inner diameter reversed-phase precolumn

Choline acetyltransferase synthesizes acetylcholine in cholinergic neurons. In the brain, these neurons are especially vulnerable to effects of ␤-amyloid (A␤) peptides. Choline acetyltransferase is a substrate for several protein kinases. In the present study, we demonstrate that short term exposure of IMR32 neuroblastoma cells expressing human choline acetyltransferase to A␤-(1-42) changes phosphorylation of the enzyme, resulting in increased activity and alterations in its interaction with other cellular proteins. Using mass spectrometry, we identified threonine 456 as a new phosphorylation site in choline acetyltransferase from A␤-(1-42)-treated cells and in purified recombinant ChAT phosphorylated in vitro by calcium/calmodulin-dependent protein kinase II (CaM kinase II). Whereas phosphorylation of choline acetyltransferase by protein kinase C alone caused a 2-fold increase in enzyme activity, phosphorylation by CaM kinase II alone did not alter enzyme activity. A 3-fold increase in choline acetyltransferase activity was found with coordinate phosphorylation of threonine 456 by CaM kinase II and phosphorylation of serine 440 by protein kinase C. This phosphorylation combination was observed in choline acetyltransferase from A␤-(1-42)-treated cells. Treatment of cells with A␤-(1-42) resulted in two phases of activation of choline acetyltransferase, the first within 30 min and associated with phosphorylation by protein kinase C and the second by 10 h and associated with phosphorylation by both CaM kinase II and protein kinase C. We also show that choline acetyltransferase from A␤-(1-42)-treated cells co-immunoprecipitates with valosin-containing protein, and mutation of threonine 456 to alanine abolished the A␤-(1-42)-induced effects. These studies demonstrate that A␤-(1-42) can acutely regulate the function of choline acetyltransferase, thus potentially altering cholinergic neurotransmission.
Cholinergic neurons in brain are especially vulnerable to effects of both soluble/oligomeric and deposited/fibrillar forms of ␤-amyloid (A␤) 1 peptides released from amyloid precursor protein (APP). Shifts in production of soluble APP␣ by ␣-secretase to production of A␤-  and A␤-(1-42) with activation of ␤and ␥-secretase in Alzheimer's disease and following traumatic head injury are associated with decreased function and communication by cholinergic neurons (1)(2)(3). A complex relationship exists between cholinergic neuron function and APP processing and A␤ peptide production (4 -6). Short term exposure to low (picomolar or nanomolar) concentrations of soluble/ oligomeric A␤ peptide leads to presynaptic cholinergic dysfunction with a reduction in the availability of acetylcholine (ACh) precursors choline (7) and acetyl-coenzyme A (8,9) coupled to decreased ACh synthesis and release from hippocampal slices or neuronal cultures (9 -13). These acute effects of A␤ peptides on neurotransmission and synaptic efficacy probably differ from the neurotoxicity produced by long term exposure and high (micromolar) concentrations of the peptides that cause death of cholinergic neurons. Mechanisms underlying acute and long term effects of A␤ peptides on cholinergic function have not been resolved.
Choline acetyltransferase (ChAT; EC 2.3.1.6) produces the neurotransmitter ACh in cholinergic neurons. ChAT undergoes covalent modification post-translationally by protein kinasemediated phosphorylation (14 -18), and we showed previously that it is a substrate for a number of protein kinases (17). Catalytic activity of this enzyme, its subcellular distribution, and potentially its interaction with other cellular proteins can be regulated in a phosphorylation-dependent manner. For example, phosphorylation of ChAT by protein kinase C (PKC) on Serine 440 led to a significant increase in its activity and ionic binding to plasma membrane in cells (18).
Phosphorylation of ChAT could be altered by changes in activity or subcellular redistribution of protein kinases brought about by neuronal perturbations or pathology such as Alzheimer's disease and traumatic brain injury. This could alter ACh biosynthesis and cholinergic neurotransmission and cause dysfunction of cholinergic neurons. A␤ peptides modulate a range of cellular signal transduction pathways and protein kinases (19 -21). Whereas a number of potential cell surface receptors for A␤ peptides have been identified (20 -23), it is unclear how these peptides mediate their cellular actions either acutely or in the longer term. It is known, however, that A␤ peptides can alter cell calcium homeostasis, leading to increased cytosolic free calcium levels (24,25). Within a certain concentration range, this could activate a number of calcium-dependent processes, including calcium-dependent protein kinases such as PKC and ␣-calcium/calmodulin-dependent protein kinase II (CaM kinase II). Since ChAT is known to be a substrate for both of these protein kinases (17), it is likely that A␤ peptides could affect cholinergic neurotransmission through regulation of function of this enzyme.
In the present study, we tested the hypothesis that short term exposure of IMR32 neuroblastoma cells stably expressing human 69-kDa ChAT to A␤ peptides would lead to altered function of the enzyme. Interestingly, we observed that treatment of cells with A␤-(1-42), but not A␤- , changed the state of phosphorylation of ChAT, revealing a new putative CaM kinase II phosphorylation site. Furthermore, phosphorylation at this site, when coordinated with phosphorylation of Ser 440 by PKC, leads to a hierarchical activation of ChAT and phosphorylation-dependent association of the enzyme with other cellular proteins, including valosin-containing protein (VCP; p97, Cdc48).

EXPERIMENTAL PROCEDURES
Preparation of ChAT Constructs-The cDNA for human 69-kDa ChAT (N1-ChAT) in pcDNA3 was kindly provided by Dr. H. Misawa (Tokyo Metropolitan Institute for Neuroscience). The mutant ChAT-T456A was prepared by site-directed mutagenesis of Thr 456 3 Ala in wild-type 69-kDa ChAT using the QuikChange kit (Stratagene), with the forward primer 5Ј-CAGATCGGCCGCTCCAGAGGC-3Ј and the reverse primer 5Ј-GCCTCTGGAGCGGCCGATCTG-3Ј. The presence of the mutation was verified at the nucleotide level by automated DNA sequencing and at the protein level by ESI-MS/MS sequencing. Mutant ChAT-S440A was prepared previously (18).
Culture and Treatment of IMR32 Cells-Human neuroblastoma IMR32 cells were transfected with plasmids containing wild-type 69-kDa human ChAT or mutants ChAT-T456A or ChAT-S440A in pcDNA3.1 using LipofectAMINE 2000 (Invitrogen). G418-resistant stable transformants were selected and tested for ChAT enzyme activity by radioenzymatic assay and ChAT protein by immunoblot. Cells were maintained in modified Eagle's medium containing 10% fetal calf serum, 50 g/ml gentamycin, and 0.5% G418 in humidified 5% CO 2 at 37°C. For experiments, monolayers of cells were treated at ϳ70% confluence. At 4 h before treatment, fresh medium was added to cells, and then A␤ peptides (1-40 or 1-42) or reverse peptides (40 -1 or 42-1) used as negative controls were diluted in culture medium to final concentrations of 100 nM from 100 M stocks and added to cells for varying times up to 18 h. For protein kinase inhibition studies, cellpermeable inhibitors (Calbiochem) of PKC (H7; 50 M), CaM kinase II (KN-93; 5 M), p38-mitogen-activated protein kinase (SB202190; 10 M), or MEK-1/MEK-2 (U0126; 50 M) were added to the media 2 h before the addition of A␤ peptides. A␤ peptides (Bachem) were dissolved in double-distilled H 2 O (1-40 and 40 -1) or 0.1% ammonium hydroxide (1-42 and 42-1) at 100 M and incubated at 37°C for 4 days (26); aliquots were stored at Ϫ80°C until use. For protein phosphorylation studies in cells, culture medium was changed to phosphate-free modified Eagle's medium (Sigma) containing [ 32 P]orthophosphate (200 Ci/ ml) at 3 h before the end of the treatment with A␤ peptides; in the case of treatment times shorter than 3 h, phosphate-free modified Eagle's medium and [ 32 P]orthophosphate were added with A␤ peptides at the beginning of the incubation interval. Following treatment, lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, leupeptin/aprotinin/pepstatin at 10/25/10 g/ml, 500 M sodium orthovanadate, 10 mM sodium fluoride, and 700 units/ml DNase I) was added to cells and incubated for 30 min on ice. Lysates were centrifuged (15,000 ϫ g for 10 min), and supernatants were used for analysis of activity or phosphorylation of ChAT or for analysis of proteins that co-immunoprecipitate with ChAT. ChAT activity was measured radioenzymatically using a modification of the method of Fonnum (27), as published previously (28).
Analysis of ␤-Amyloid Peptides-CD spectra of A␤-(1-40), A␤-(1-42), and the corresponding reverse peptides were recorded on a Jasco spectopolarimeter, model J-810, at 25°C in a 0.1-cm path length cell at 0.2-nm intervals over the wavelength range 190 -260 nm. Peptides were analyzed at a concentration of 50 M in double-distilled H 2 O (1-40 and 40 -1) or 0.1% ammonium hydroxide (1-42 and 42-1). To assess structural composition of A␤-(1-40) and A␤-(1-42) peptide solutions by electron microscopy, carbon-formvar-coated grids were floated on a drop of each sample to allow peptides to adhere. After 3 min, grids were blotted lightly and then floated on either 1% phosphotungstic acid or 2% uranyl acetate. Following staining, grids were blotted and air-dried, and then representative images were acquired by examining grids in a Philips EM300 electron microscope operated at 60 kV.
Western Blot Analysis-Proteins were transferred from SDS-PAGE gels to nitrocellulose membranes in a semidry electroblotting apparatus using transfer buffer (48 mM Tris, 39 mM glycine) containing 20% methanol. For detection of ChAT or VCP, blotting membranes were saturated with 8% nonfat milk powder in phosphate-buffered saline and probed with anti-ChAT CTab antibody (1:2,000) or anti-VCP antibody (1:2,000) for 1 h at room temperature. Membranes were washed with phosphate-buffered saline containing 0.5% Triton X-100, and bound antibodies were detected by incubation for 1 h with peroxidasecoupled secondary antibodies (1:5,000; Amersham Biosciences) and the ECL kit (Amersham Biosciences).
Phosphorylation of Purified ChAT by CaM Kinase II and PKC-Bacterially expressed recombinant 69-kDa human ChAT was immunoaffinity-purified as described previously (18). To identify amino acid residues phosphorylated by CaM kinase II, 5 g (0.5 l) of purified ChAT protein was mixed with 20 l of kinase buffer (50 mM Tris-HCl, pH 7.2, 0.4 mM dithiothreitol, 0.5 mM CaCl 2 , 5 mM MgCl 2 , 1 M calmodulin, 100 M ATP) and 2 milliunits of purified CaM kinase II (a gift from Dr. H. Schulman, Stanford University) for 30 min at 30°C. Phosphorylation reactions were stopped by the addition of electrophoresis sample buffer. To determine effects of hierarchical phosphorylation of ChAT by PKC and CaM kinase II on ChAT enzymatic activity, a PKC phosphorylation reaction was carried out for 15 min at 30°C as described previously (17), followed by phosphorylation by CaM kinase II for an additional 15 min. ChAT activity was measured immediately at 37°C.
CaM Kinase II Assay in Permeabilized Cells-IMR32 cells expressing 69-kDa human ChAT were used to assess whether CaM kinase II is activated by treatment with A␤-(1-40), A␤-(1-42), and reverse control peptides A␤-(40 -1) and A␤-(42-1) using the method of Heasley and Johnson (31). Briefly, synthetic CaM kinase II-specific peptide KKAL-RRQETVDAL was used as a substrate to monitor CaM kinase II activity in digitonin-permeabilized cells. Cells grown in 24-well plates were treated in the presence or absence of A␤ peptides and then rinsed twice with Dulbecco's modified Eagle's medium buffered with 20 mM HEPES, pH 7.2. Permeabilization solution (100 l composed of 137 mM NaCl, 5.4 mM KCl, 1 mg/ml glucose, 20 mM HEPES, pH 7.2, 50 g/ml digitonin, 10 mM MgCl 2 , 5 mM EDTA, 2.5 mM CaCl 2 , 25 mM ␤-glycerophosphate, 100 M cold ATP, 0.4 nM [␥-32 P]ATP) (10 Ci/well) supplemented with 1 mM CaM kinase substrate peptide was added to cells for 10 min at 30°C. A parallel set of control cells were treated in the same manner but without substrate peptide added to obtain a measure of background [ 32 P]phosphate incorporation. KN-93, a cell-permeable inhibitor of CaM kinase II, or its inactive analog KN-92 (5 M) was added as an additional control. Incubation was terminated by the addition of 10 l of ice-cold 25% trichloroacetic acid, and then 40 l of acidified cellular lysate was spotted onto 25-mm phosphocellulose discs (Whatman P-81) (32). Filter discs were washed three times with 75 mM phosphoric acid and once with 75 mM Tris-HCl buffer, pH 7.5, and then dried and placed in scintillation mixture for determination of radioactivity. Proteins were digested in 0.2 M NaOH and assayed as described by Bradford (33) to determine specific activity of peptide phosphorylation expressed as pmol/min/mg protein.
In-gel Tryptic Digestion of Proteins and Sample Preparation-Following SDS-PAGE, proteins were stained briefly with Coomassie Blue, and then gels were destained and washed for 3 h with at least five solvent changes (50% methanol, 10% acetic acid) to ensure adequate removal of SDS. Gels were subsequently washed with three changes of H 2 O, and then bands corresponding to ChAT or co-immunoprecipitated proteins were excised from gels and washed with two changes of acetonitrile. Gel slices were reduced with 10 mM dithiothreitol at 50°C for 30 min and alkylated by 55 mM iodoacetamide at room temperature for 20 min, followed by washing three times with 100 mM ammonium bicarbonate. After two changes of acetonitrile, gel pieces were dried by vacuum centrifuge and rehydrated in trypsin digestion buffer (50 mM ammonium bicarbonate, 5 mM CaCl 2 ) containing 12.5 ng/l trypsin (Roche Molecular Biochemicals) (34). After a 45-min incubation on ice, excess trypsin solution was removed, 15 l of digestion buffer without trypsin was added, and samples were incubated for 18 h at 37°C. Tryptic peptides were extracted from the gel pieces with two changes of 100 l of ammonium carbonate buffer by shaking in an orbital shaker for 45 min. After a brief centrifugation, the supernatants with eluted peptides were pooled and concentrated by vacuum centrifugation to a final volume of 15 l. Pooled extracts were acidified with glacial acetic acid at a final concentration of 1%.
Two-dimensional Tryptic Phosphopeptide Mapping and Identification of Phosphorylated Residues-Two-dimensional thin layer phosphopeptide maps of ChAT were prepared as described previously (35). Following in-gel tryptic digestion of proteins, samples were applied to cellulose TLC plates by sequential spotting of 0.5-l droplets and electrophoresed in the first dimension in water/acetic acid/formic acid (89.7: 7.8:2.5, v/v/v, pH 1.9) at 1000 V for 45 min. Plates were air-dried and developed in the second dimension in water/n-butyl alcohol/pyridine/ acetic acid (30:37.5:25:7.5, v/v/v/v). Phosphopeptides were visualized by autoradiography using Eastman Kodak Co. XAR-5 film at Ϫ70°C.
Phosphopeptides required for further analysis were eluted from TLC plates with water/acetonitrile (4:1, v/v) (36) and then reduced to dryness in a vacuum centrifuge and reconstituted in 2% acetonitrile and 1% acetic acid. This solution of peptides was used directly for MALDI-TOF mass spectrometric analysis. For ESI-MS/MS sequencing, peptides were purified on ZipTip C18 TM according to the manufacturer's instructions (Millipore Corp.) and eluted from the tip resin with 65% acetonitrile and 1% acetic acid.
Phosphoamino acid analysis was also performed on phosphopeptides eluted from cellulose plates or directly on mixtures of phosphopeptides recovered after in-gel tryptic digestion. Tryptic peptides were lyophilized, resuspended in 70 l of 6 M HCl, and boiled at 110°C for 1 h. One-  . Enzyme activity returned to control levels at longer times and then increased transiently about 2-fold between 10 and 14 h (mean Ϯ S.E., n ϭ 5). B, incorporation of [ 32 P]orthophosphate into ChAT essentially paralleled the changes in ChAT activity induced by A␤-(1-42) treatment. Intensities of bands on ChAT immunoblots were quantified by densitometry for normalization of [ 32 P]phosphate incorporation data. For this latter measure, pieces of nitrocellulose membrane corresponding to the location of ChAT were excised, and radioactivity was determined by Cerenkov counting, or alternatively membranes was exposed to film for autoradiography. Data are expressed as the mean Ϯ S.E. of five independent experiments. Statistical differences at the level of p Ͻ 0.05 were determined by one-way analysis of variance with post hoc Tukey's multiple comparison test. *, differences relative to untreated controls; #, differences between treatment of cells for 30 min and 10 h. C, representative immunoblots for ChAT (upper panel) and autoradiography for [ 32 P]phosphate incorporation into ChAT (lower panel) are shown, with similar data obtained in at least four separate experiments.
or two-dimensional phosphoamino acid mapping were performed as described by Boyle et al. (37).
In Vitro Parent Ion Scanning-Experiments were performed on a Q-TOF2 mass spectrometer (Micromass), equipped with a nanoflow source. The instrument was calibrated with [Glu 1 ]Fibrinopeptide B (Sigma), and following desalting on a ZipTip C18 TM (4 l), the concentrated protein digest sample was loaded into a borosilicate capillary (type F; Micromass). A voltage between 600 and 1000 V was applied to the capillary in order to produce nanomolar flow. Parent ion scanning was performed over the m/z range of 300 -1500 monitoring neutral loss of 97.9769 and 49.9885 with a collision energy of 32 V. Masses that resulted in such neutral losses were subsequently further fragmented to obtain sequence information. Resultant spectra were backgroundsubtracted and deconvoluted using MaxEnt3 software provided in the Micromass MassLynx version 3.5 software package. The sequence of the peptides was determined using PepSeq version 3.3 software also provided in the MassLynx 3.5 software.
Peptide Sequencing by Mass Spectrometry-Amino acid sequences of tryptic phosphopeptides of ChAT isolated from control and A␤-treated cells or for identification of unknown proteins that co-immunoprecipitate with ChAT were obtained by mass spectrometry performed on a Micromass Q-TOF2 mass spectrometer equipped with a nanospray source and an online Waters CapLC (Waters). In all cases, 1 l of sample was injected from the autosampler. The instrument was calibrated with [Glu 1 ]Fibrinopeptide B (Sigma). A gradient consisting of 5-65% B in 15.5 min (A ϭ 0.1% formic acid, B ϭ acetonitrile with 0.1% formic acid) flowing at 1 l/min was used to elute peptides from a 300-m inner diameter reversed-phase precolumn (LC-packings, San Francisco, CA) to the mass spectrometer. Survey spectra were acquired in the m/z range of 400 -2000. Doubly or triply charged precursor ions were automatically selected for fragmentation by the quadrupole mass filter. In some cases, specific masses were identified for fragmentation. Fragmentation was achieved by collision with argon gas in the collision cell. The collision energy was automatically varied depending on the charge state of the parent peptide. Resultant spectra were backgroundsubtracted and deconvoluted using MaxEnt3 software provided in the Micromass MassLynx 3.5 software package. The sequence of the peptides was determined using PepSeq version 3.3 software also provided in the MassLynx 3.5 software.
MALDI-TOF MS-Proteins co-immunoprecipitating with ChAT were identified initially by MALDI-TOF mass spectrometry. Samples were mixed 1:1 (v/v) with matrix solution containing 1:1 ethanol/acetonitrile saturated with ␣-cyano-4-hydroxycinnamic acid. Each sample (1 l) was spotted onto the MALDI target plate in triplicate. MALDI-TOF MS was performed on a MALDI-R mass spectrometer (Micromass). Calibration was performed externally using angiotensin I (Sigma), renin substrate (Sigma), adenocorticotrophic hormone clip 18 -39 (Sigma) for a three-point calibration. In addition, for each sample, the lockmass method was used as additional calibration with the standard adenocorticotrophic hormone clip 18 -39. The peptide mass fingerprint spectra were matched to the NCBI nonredundant data base entries by using the following programs, available on the World Wide Web: Profound (www.proteometrics.com) and Mascot (www.matrixscience.com). The mass tolerance was set to 60 ppm, and two missed cleavage sites were tolerated with the search restricted to human proteins.

RESULTS
Characterization of A␤ Peptides-Stock solutions of A␤ peptide used to treat cells in the present studies were analyzed by CD and electron microscopy to obtain information that would allow relative comparisons to be made about peptide conformations. As shown in Fig. 1A, CD analysis revealed that the secondary structure of A␤-(1-40) and A␤-(1-42) were qualitatively identical in that they were both composed predominantly of ␤-sheet conformation indicated by minimum molar ellipticity at 215 nm. Based on the shape of the CD spectra, negligible random coil or ␣-helical content was present in these two peptides. To obtain quantitatively similar signal strength on CD spectra, the concentration of A␤-(1-40) and A␤-(1-42) used were 50 and 30 M, respectively. By comparison, the reverse Following in-gel tryptic digestion, tryptic peptides of ChAT were analyzed by two-dimensional phosphopeptide mapping. Under all conditions tested, a phosphopeptide(s) with phosphorylation on serine residue(s) was observed. By comparison, an additional phosphopeptide corresponding to phosphorylation on threonine residue(s) was observed only in cells treated with A␤- (1-42). B, tryptic peptides were also subjected to acid hydrolysis, and the phosphoamino acids produced were separated by two-dimensional electrophoresis. Circles indicate the migration of ninhydrinstained phosphoamino acid standards, with closed circles corresponding to phosphoserine and broken circles corresponding to phosphothreonine. ϫ, sample origins; P i , free [ 32 P]phosphate liberated during acid hydrolysis of phosphopeptides. Positive spots remaining near sample origins represent incompletely hydrolyzed phosphopeptides. Data illustrated are representative of between three and five independent experiments. peptides A␤-(40 -1) and A␤-(42-1) had a predominantly random coil conformation. In support of the CD data, as illustrated in Fig. 1B, electron microscopic analysis confirmed that solutions of both A␤-(1-40) and A␤-(1-42) contained characteristic A␤-fibrils.
Short Term Exposure to A␤-(1-42) Enhances Activity and Phosphorylation of ChAT-IMR32 cells expressing 69-kDa human ChAT were incubated with A␤-(1-42) for varying times, and then incorporation of [ 32 P]phosphate and enzyme activity were monitored. As illustrated in Fig. 2A, catalytic activity of ChAT was increased within 30 min of the addition of A␤-  to cultures, with this effect being maximal at 10 h (2-fold increase in ChAT-specific activity). The effect of A␤-(1-42) on ChAT activity followed a biphasic time course, with the response diminishing beyond 10 h. Immunoblots shown in Fig.  2C (upper panel) demonstrate that cellular ChAT concentration was unchanged over the treatment interval. Phosphorylation of ChAT was also increased up to 3-fold by treatment of cells with A␤-(1-42) in a manner that paralleled the time course for change in activity of the enzyme, as shown in Fig. 2B; the corresponding autoradiography data are provided in Fig.  2C (lower panel). Control cells were treated with inactive peptide A␤-(42-1); ChAT activity measured in cells with the addition of A␤-(42-1) did not differ from cells with no A␤ peptide added. ChAT activity was not altered in cells treated for up to 18 h with A␤-(1-40) or its reverse peptide A␤-(40 -1) (data not shown). However, a 2-fold increase in [ 32 P]phosphate incorporation into ChAT was found in IMR32 cells treated with A␤-(1-40) between 2 and 6 h (data not shown).
Identification of a New Phosphorylation Site in ChAT in Cells Treated with A␤-(1-42)-IMR32 cells expressing ChAT were treated with A␤ peptides for 10 h, and then ChAT was recovered by immunoprecipitation and subjected to in-gel tryptic cleavage. In control cells and cells treated with A␤-  and the reverse sequence control peptides A␤-(40 -1) and A␤-(42-1), a single phosphopeptide was observed, as illustrated in Fig. 3A. In contrast, treatment of cells with A␤-(1-42) resulted in the appearance of a second phosphorylated ChAT peptide.  . Samples from untreated control samples and cells treated with reverse peptide A␤-(42-1) yielded spectra similar to A, indicating lack of phosphorylation of this ChAT peptide. All samples were prepared in duplicate, and multiple injections were examined with identical results. C and D, two-dimensional patterns for phosphoamino acids and phosphopeptides of ChAT-T456A, respectively, from cells treated with A␤- . This demonstrates the lack of threonine phosphorylation as compared with that observed in wild-type ChAT (shown in Fig. 3).
Further analysis of these phosphopeptides following acid hydrolysis showed that all samples contained phosphorylated serine residue(s). Interestingly, phosphorylated ChAT obtained from cells treated with A␤-(1-42) contained phosphothreonine as well as phosphoserine residues (Fig. 3B).
To identify the threonine residue phosphorylated in ChAT with A␤-(1-42) treatment, ChAT was immunoprecipitated from IMR32 cells after 10 h of treatment with A␤ peptides. Following isolation by SDS-PAGE, ChAT-containing bands were subjected to in-gel tryptic digestion, and samples were prepared for mass spectrometric analysis by ESI-MS/MS. Mass spectra revealed a doubly charged peak at 621.6 m/z present in ChAT recovered from cells treated with A␤-(1-42) (Fig. 4B) but not in cells treated with A␤-(1-40) (Fig. 4A) or A␤-(42-1) or in ChAT from untreated control cells (data not shown). Although other peaks were detected in tryptic digests of samples treated with A␤-(1-42), none of these corresponded to ChAT peptides with potential phosphorylation sites. The amplitude of the signal detected for this tryptic phosphopeptide from ChAT from A␤-(1-42)-treated cells was very low when compared with purified ChAT phosphorylated by incubation with protein kinases in vitro (see below). This suggests that a relatively low proportion of the enzyme is phosphorylated in situ. Consequently, sequence information was obtained by fragmenting at this m/z for extended periods. Subsequent analysis of the fragmentation pattern revealed the C-terminal sequence VR as well as the partial sequence EAL. This allowed identification of this peptide as amino acid residues 454 -464 of 69-kDa human ChAT with the sequence SATPEALAFVR and a mass of 1160.626. The immonium ion region also revealed the presence of the correct amino acids for the peptide encoding residues 454 -464 of 69-kDa human ChAT (Pro, Val, Thr, Leu, Glu, Phe, and Arg). This sequence contains a putative consensus sequence for CaM kinase II involving phosphorylation of Thr 456 ; the canonical consensus sequence for CaM kinase II is RXX(*S/*T) (38 -40), with the corresponding ChAT sequence of 453 RSA*T 456 . To confirm this assignment, Thr 456 was changed to an alanine residue (ChAT-T456A) by site-directed mutagenesis to provide a plasmid for use as an investigative tool; the presence of the mutation was verified by DNA sequencing of the plasmid and ESI-MS/MS sequencing of the tryptic peptide encoding amino acid residues 454 -464 in T456A-ChAT and absence of the tryptic peptide encoding the wild-type sequence. As illustrated in Fig. 4, C and D, analysis of tryptic peptides of ChAT-T456A obtained from IMR32 cells expressing the mutant enzyme and treated with A␤-(1-42) by phosphopeptide and phosphoamino acid analysis revealed a single phosphopeptide, with serine being the residue phosphorylated. This finding is in sharp contrast to the results obtained with cells expressing wild-type ChAT shown in Fig. 3, A and B. A doubly charged peak at 658.3 m/z (M ϭ 1314.6) was also found in samples. This corresponds to the phosphorylated form of the tryptic peptide 432-442 of 69-kDa human ChAT (LVP-TYESASIR). This peptide contains a serine residue (Ser 440 ) that was found previously to be phosphorylated by PKC (18).

ESI-MS/MS Identification of Threonine 456 as a Putative CaM Kinase II Phosphorylation Site in ChAT in
Vitro-Immunoaffinity-purified 69-kDa ChAT was incubated under phosphorylating conditions with CaM kinase II and then resolved by SDS-PAGE and digested with trypsin for analysis by ESI-MS to identify phosphorylated peptide(s) and amino acid residue(s). Using parent ion scanning, a doubly charged peak at 621.6 m/z was found to produce a neutral loss of 98 that is indicative of loss of a phosphate group under moderate fragmentation conditions (41). This m/z value was subjected to full fragmentation in order to obtain sequence information. This peptide sequenced to SApTPEALAFVR (where pT represents phosphothreonine), as can be seen in Fig. 5, indicating the presence of a phosphorylated threonine residue. This sequence corresponds to tryptic fragment 454 -465 of 69-kDa ChAT that contains a putative consensus sequence for CaM kinase II; this is the same threonine residue (Thr 456 ) found to be phosphorylated in IMR32 cells treated with A␤- .
Assay of CaM Kinase II Activity in IMR-32 Cells with A␤ Peptide Treatment-We tested whether CaM kinase II was activated in IMR32 cells treated with A␤ peptides. Cells expressing wild-type ChAT were grown with A␤-(1-42), A␤-(1-40), or reverse peptides A␤-(42-1) or A␤-(40 -1) and then digi-tonin-permeabilized and incubated with substrate peptide encoding a CaM kinase II phosphorylation consensus sequence. As illustrated in Fig. 6A, 10-h treatment with A␤-(1-42) selectively increased phosphorylation of the CaM kinase II substrate peptide by more than 2-fold, indicating activation of this protein kinase. Phosphorylation of the substrate peptide in cells treated with the other three A␤ peptides did not differ from untreated control cells. This corresponds to the time point when A␤-(1-42)-mediated increases in activity and phosphorylation of ChAT are maximal. Time course experiments revealed that phosphorylation of CaM kinase II substrate peptide was not different from control at 30 min or 2 h, significantly increased (160% compared with control) by 6 h, and returned to control levels at 14 and 18 h after the addition of A␤-(1-42) (data not shown). To further confirm that phosphorylation of the substrate peptide in A␤-(1-42)-treated cells was related to activation of CaM kinase II, we tested the effects of the CaM kinase II inhibitor KN-93 and its inactive analogue KN-92. The addition of 5 M KN-93, but not KN-92, to cells during treatment with A␤-(1-42) markedly reduced subsequent phosphorylation of the CaM kinase II substrate peptide (Fig. 6B).
Hierarchical Activation of ChAT with Phosphorylation by PKC and CaM Kinase II-We investigated the relationship between phosphorylation of ChAT by PKC and CaM kinase II and activation of the enzyme using both purified recombinant ChAT and IMR32 cells treated with A␤-(1-42). As demonstrated in Fig. 7A, purified ChAT activity was increased about 3-fold by CaM kinase II-mediated phosphorylation only when the enzyme was also phosphorylated by PKC. Phosphorylation of ChAT by PKC alone led to a 2-fold increase in enzyme activity, whereas phosphorylation by CaM kinase II alone did not alter ChAT activity. Moreover, inhibition of PKC by H7 (10 M) blocked activation of ChAT observed when the enzyme was sequentially incubated with PKC and CaM kinase II under phosphorylating conditions. Incubation of purified ChAT with PKC and CaM kinase II in the presence of KN-93 resulted in a 2-fold increase in ChAT activity similar to that observed for phosphorylation of ChAT by PKC alone.
As shown in Fig. 7B, H7 also inhibited activation of ChAT in IMR32 cells treated with A␤-(1-42) for either 30 min or 10 h. On the other hand, the CaM kinase II inhibitor KN-93 partially attenuated activation of ChAT by 10 h of A␤ treatment and had no effect on ChAT activation at 30 min of treatment. Whereas Thr 456 is situated in a consensus sequence that could be recognized by CaM kinase II, it is also positioned at Ϫ1 from a proline residue, creating the possibility that this proline-directed threonine residue could be phosphorylated by other protein kinases such as mitogen-activated protein kinase. To test this, we used inhibitors of MEK-1/MEK-2 (U0126) and p38mitogen-activated protein kinase (SB202190) to probe their involvement in phosphorylation of Thr 456 in A␤-(1-42)-treated IMR32 cells. The addition of U0126 or SB202190 to IMR32 cells during A␤ treatment also had no effect on A␤-mediated activation of ChAT at either time point.
To address this further, we investigated the temporal relationship of serine and threonine phosphorylation of ChAT in IMR32 cells treated with A␤-(1-42) and the effect of kinase inhibitors on phosphorylation of ChAT in A␤-treated cells. As shown in Fig. 7C, H7 reduced serine phosphorylation to control levels at both 30 min and 10 h, whereas KN-93 eliminated threonine phosphorylation in cells treated with A␤-(1-42) for 10 h. U0126 or SB202190 had no effect on phosphorylation at either 30-min or 10-h time points. We also tested the effect of A␤-(1-42) treatment on activity of ChAT-S440A and ChAT-T456A mutants in comparison with that of wild-type enzyme. As shown in Fig. 8, ChAT-specific activity does not differ in the A␤ Peptide Treatment Alters Interaction of ChAT with Other Cellular Proteins-ChAT was immunoprecipitated from IMR32 cells treated with A␤- , and co-immunoprecipitating proteins were separated and analyzed by one-dimensional SDS-PAGE. It is apparent from the Coomassie-stained gel of ChAT immunoprecipitates shown in Fig. 9A that a number of other   FIG. 7.

Hierarchical activation of purified ChAT and ChAT in A␤-(1-42)-treated IMR32 cells related to phosphorylation by PKC and CaM kinase II.
A, purified recombinant ChAT (1 g/sample) was phosphorylated by either PKC or CaM kinase II or by both kinases. Phosphorylation by CaM kinase II alone did not alter ChAT activity compared with control, whereas phosphorylation by PKC alone led to a 2-fold increase in ChAT activity that was blocked by H7. When ChAT was phosphorylated by both PKC and CaM kinase II, activity of the enzyme increased 3-fold. Importantly, inhibition of CaM kinase II by KN-93 attenuated enhancement of ChAT activity to the level obtained with phosphorylation by PKC alone. Moreover, inhibition of PKC by 10 M H7 blocked activation of ChAT when it was incubated under phosphorylating conditions with PKC and CaM kinase II. B, IMR32 cells were incubated with A␤-(1-42) for 30 min or 10 h in the absence or presence of protein kinase inhibitors to test involvement of these kinases in the two phases of activation of ChAT observed in these studies. None of the inhibitors altered basal ChAT activity. At 30 min of treatment with A␤-(1-42), ChAT activity was increased about 1.6-fold, with this effect blocked by H7. At 10 h of treatment with A␤-(1-42), ChAT activity was increased about 2-fold. This effect was attenuated to a 1.6-fold increase by KN-93 and completely blocked by H7. The MEK-1/MEK-2 and p38-kinase inhibitors U0126 and SB202190 did not alter A␤-(1-42)-mediated changes in ChAT activity. C, phosphoamino acid analysis revealed basal phosphorylation of ChAT on serine but not threonine residue(s) under control conditions. This was not altered by any of the kinase inhibitors. Treatment of cells with A␤-(1-42) for 30 min increased phosphoserine levels, but phosphothreonine was not observed. H7 decreased phosphoserine levels. At 10 h of A␤-(1-42) treatment, there was an increase in phosphoserine levels and phosphorylation on threonine residue(s). KN-93 blocked threonine phosphorylation, whereas H7 blocked serine phosphorylation. For A and B, results are expressed as mean Ϯ S.E. of four or five independent experiments with duplicate or triplicate determinations. Statistical differences at the level of p Ͻ 0.05 were determined by one-way analysis of variance with post hoc Tukey's multiple comparison test. *, differences relative to controls; #, differences relative to PKC alone (A) or between treatment of cells for 30 min and 10 h (B). Data shown in C are representative of at least three separate experiments.
proteins co-immunoprecipitated with both wild-type ChAT and mutant ChAT-S440A from cells treated with A␤-(1-42). These were not observed in lanes corresponding to treatment of cells with A␤-  or A␤-(42-1) or from cells expressing mutant ChAT-T456A treated with A␤- .
A combined mass spectrometry and immunoblot approach was taken to begin to identify proteins co-immunoprecipitating with ChAT. For mass spectrometric analysis, bands were excised from SDS-PAGE gels and digested with trypsin. Based on MALDI-TOF peptide mass fingerprint data of the tryptically digested protein, we identified one co-immunoprecipitating protein with an apparent molecular mass of about 90 kDa to be human VCP. Sequence coverage of 40% was obtained, and 38 peptides were matched to VCP in two independent analysis. Partial sequence data of tryptic peptides obtained by ESI-MS/MS also matched to VCP sequences. Fig. 9B confirms the identity of VCP by immunoblot with anti-VCP antibody in ChAT-immunoprecipitates only from cells expressing wild-type 69-kDa ChAT that were treated with A␤-(1-42) but not other A␤ peptides; this finding was obtained in three independent experiments. To further verify this interaction, VCP was immunoprecipitated from lysates of IMR32 cells expressing wildtype ChAT treated with A␤ peptides using an antibody to VCP. As illustrated in Fig. 9C, a ChAT immunopositive band was observed in the anti-VCP immunoprecipitate from cells treated with A␤- . DISCUSSION Increased concentrations of A␤ peptides are released into the brain by cleavage of APP in neurodegenerative diseases such as Alzheimer's disease and following traumatic head injury. In the present study, we tested the hypothesis that exposure of IMR32 neuroblastoma cells stably expressing ChAT to A␤ peptides would alter function of the enzyme. We demonstrate for the first time that 1) phosphorylation patterns and activity of ChAT are changed by short-term exposure of cells to A␤-(1-42), 2) exposure of cells to A␤-(1-42) leads to activation of CaM kinase II and phosphorylation on a threonine residue in ChAT, 3) Thr 456 in ChAT is phosphorylated by CaM kinase II in vitro and in cells following treatment with A␤-(1-42), 4) the increase in ChAT activity observed with A␤-(1-42)-treatment is hierarchically organized with phosphorylation of Ser 440 by PKC being required, and this effect is amplified by CaM kinase II-mediated phosphorylation of Thr 456 at 10 h; 5) treatment of cells with A␤-(1-42) leads to Thr 456 phosphorylation-dependent protein-protein interactions between ChAT and other cellular proteins, with one of these identified as VCP.
Although mechanisms by which A␤ peptides mediate either acute or long term actions on neurons are not resolved, it is known that they interact with several cell surface receptors including ␣-7 nicotinic ACh receptor (21,42), p75 nerve growth factor receptor (43), G-protein-linked formyl-peptide receptor (44), and advanced glycosylation end product receptor (45). Signal transduction pathways and cellular responses recruited by binding of A␤ peptides to these receptors have not been worked out, but some responses are mediated by pertussis toxin-sensitive G-protein-coupled mechanisms (22). Interaction of A␤ peptides with some receptors alters cellular calcium homeostasis with enhanced influx of extracellular calcium or release from intracellular stores (24,48). Also, A␤ peptides can form calcium ionophores in plasma membrane, leading to calcium conductances across plasma membrane (48,49). In addition, intracellular accumulation A␤-(1-42) but not other A␤ peptides, possibly mediated by interaction with the ␣-7 nicotinic ACh receptor (50), led to neuronal toxicity through a p53-Bax cell death pathway (51). In regard to the present studies, it is known that IMR32 cells express ␣-7 nicotinic ACh receptors (50) and may express low levels of p75 nerve growth factor receptors (51); it is not known whether other putative binding sites for A␤ peptides are expressed by these cells. Increased cytosolic free calcium levels caused by A␤ peptide treatment could activate Ca 2ϩ -dependent protein kinases such as PKC and CaM kinase II. ChAT is a substrate for both of these kinases (17). The present study extends these findings by demonstrating a link between activation of CaM kinase II and PKC and regulation of ChAT activity and phosphorylation of ChAT and its interaction with other proteins. We also showed that A␤-(1-42) increased CaM kinase II activity in permeabi-lized IMR32 cells, and, although we did not measure PKC activation in permeabilized cells, we found increased PKC activity in cell lysates (data not shown). Activation of PKC isoforms by A␤ peptides was reported previously (52)(53)(54). A key observation in the present study is that only A␤-(1-42) caused changes in ChAT phosphorylation, activity, and protein interactions. Mechanisms underlying this selectivity are unclear, but other studies have demonstrated actions mediated specifically by A␤-(1-42) and not other A␤ peptides. To confirm that this did not relate to differences in conformation between A␤-(1-42) and A␤-(1-40), we analyzed samples of peptides by CD and electron microscopy and found both preparations to be comprised of A␤-fibrils and to have similar ␤-sheet structure. It is likely that the differences in cellular responses to A␤-(1-42) and A␤-(1-40) observed in the present study relate to differences in ability of the peptides to stimulate receptors that initiate the cellular events.
Using phosphopeptide mapping, phosphoamino acid analysis, and mass spectrometry, we identified Thr 456 as a novel phosphorylation site in 69-kDa ChAT following short term treatment of IMR32 cells with A␤- . Residue Thr 456 in ChAT is phosphorylated in A␤-(1-42)-treated cells over a time course coinciding with increased activity of CaM kinase II. Since other protein kinases including members of the mitogenactivated protein kinase family could be activated by A␤-treatment (21), we used specific inhibitors to determine that these are probably not mediating phosphorylation of Thr 456 in this situation. Also, kinase inhibitors and site-directed mutagenesis of critical residues in ChAT were used to establish a relationship between phosphorylation of Ser 440 by PKC and phosphorylation of Thr 456 by CaM kinase II in enhanced catalytic activity of ChAT. Phosphorylation of ChAT on Ser 440 by PKC alone increased ChAT activity by 2-fold, with this increased to about 3-fold when Thr 456 is also phosphorylated by CaM kinase II (Fig. 2B). Importantly, phosphorylation by CaM kinase II alone did not alter ChAT activity. This suggests that phosphorylation of ChAT on Ser 440 leads to the initial activation of ChAT by 30 min. This increased ChAT activity is not maintained over the next few hours, but a second phase of activation occurs by 10 -14 h after the addition of A␤-(1-42) paralleling increased CaM kinase II activity, phosphorylation of Thr 456 , and increased serine phosphorylation. Mechanisms underlying this delayed increase in ChAT activity and activation of CaM kinase II are unclear. One possibility, however, is that cellular events initiated with acute addition of A␤-(1-42) result in production/release of cellular mediators producing effects several hours later. These delayed effects include activation of CaM kinase II, enhanced activity of ChAT, and interaction of ChAT with VCP and other cellular proteins.
To date, there have been no reports identifying proteins that ChAT interacts with in the cell. Previously, it was suggested that neurons contain endogenous modulator(s) of ChAT that may be proteins (55,56). Other compounds that may regulate ChAT appear to be small molecules or lipid products, such as phosphomonoesters (57) and dihydrolipoic acid (58). In the present study, a number of proteins immunoprecipitated with ChAT following treatment of cells expressing ChAT with A␤-  and associated with phosphorylation of Thr 456 . We identified one of these by mass spectrometry, with confirmation by immunoblotting, to be VCP. VCP, a member of the AAA-ATPase family of proteins, is a multifunctional protein in yeast and mammalian cells (59 -61) with roles in diverse cellular functions including cell cycle regulation, clathrin-mediated receptor endocytosis, protein ubiquitination, and proteasome function. Other cellular proteins with which VCP is known to interact include BRCA1 (62) and histone deacetylase-6 (46). Lysates were prepared from cells expressing wild-type or mutant ChAT-S440A or ChAT-T456A treated for 10 h with A␤ peptides and then incubated with CTab anti-ChAT antibody to immunoprecipitate ChAT. A, proteins in samples obtained by immunoprecipitating ChAT from cell lysates were separated by SDS-PAGE, and then gels were stained with Coomassie Blue to observe patterns of co-immunoprecipitating proteins. Selected bands were excised from gels and subjected to in-gel tryptic digestion, with the resultant peptides analyzed by mass spectrometry. MALDI-TOF mass fingerprint analysis and partial ESI-MS/MS sequence analysis of samples from two independent samples identified a band with apparent molecular mass of ϳ90 kDa to be VCP. B, the identity of VCP as a protein that co-immunoprecipitates with ChAT in A␤-(1-42)-treated IMR32 cells expressing wild-type ChAT was confirmed by immunoblot using an anti-VCP antibody. C, in separate samples, VCP was immunoprecipitated from cell lysates using the anti-VCP antibody, and the presence of ChAT was identified as a protein that co-immunoprecipitates with VCP by immunoblot using the CTab anti-ChAT antibody.
Interestingly, it appears that association of VCP with ChAT is phosphorylation-dependent, occurring under conditions where both Thr 456 and Ser 440 are phosphorylated. VCP does not coimmunoprecipitate with ChAT in A␤-(1-42)-treated cells expressing mutant ChAT-S440A. The functional significance of interaction of ChAT with VCP is unclear. A predominant cellular function of VCP is its role in the link between acetylation/ deacetylation/ubiquitination and proteosomal degradation of proteins (46,61). One function for an acetyltransferase is to add acetyl-residues to proteins to protect them from ubiquitination and targeting for degradation (46,47). Although a function for the interaction between ChAT and VCP remains to be determined, two hypothesis could be tested. In the first, ubiquitinated-ChAT binds to VCP, which then serves as a chaperone to target the enzyme to the proteosome. In the second, ChAT may serve as an acetylase, transferring acetate to other proteins to alter their function or delay their degradation. As an acetylase, ChAT catalyzes the transfer of acetyl groups to hydroxyl moieties of choline and potentially other small molecules through an O-acetylation reaction. It has never been demonstrated that ChAT can catalyze the N-acetylation reaction that would normally be found with acetylation of proteins on lysine residues.
The findings reported in the present study identify a new action of A␤-(1-42) on cholinergic neuron function. Although loss of ChAT activity as a consequence of the toxic effects of A␤ peptides on cholinergic neurons has been demonstrated, acute effects of A␤ peptides on function of the enzyme ChAT have not been examined. Exposure of primary cultures of rat septal neurons to A␤-(1-42), but not A␤- , at concentrations used in the present study for 12-24 h suppressed ACh synthesis but did not alter ChAT activity (9). Also, exposure of SN56 cells to 100 nM A␤-(1-42) for 48 h significantly reduced ChAT activity (10). As demonstrated in the present study, activation of cell signaling pathways and alteration of the intracellular milieu in response to exposure to A␤-(1-42) leads to short term changes in the phosphorylation state of ChAT that could result in acute changes in cholinergic neurotransmission.