Neuronal Cdc2-like Protein Kinase (Cdk5/p25) Is Associated with Protein Phosphatase 1 and Phosphorylates Inhibitor-2*

Protein phosphatase 1 (PP1) is complexed with inhibitor 2 (I-2) in the cytosol. In rabbit muscle extract PP1·I-2 is activated upon preincubation with ATP/Mg. This activation is caused by phosphorylation of I-2 on Thr72 by glycogen synthase kinase 3 (GSK3). We have found that PP1·I-2 in bovine brain extract is also activated upon preincubation with ATP/Mg. However, blocking GSK3 action by LiCl inhibited only ∼29% of PP1 activity and indicated that GSK3 is not the sole PP1·I-2 activator in the brain. When bovine brain extract was analyzed by gel filtration PP1·I-2 and neuronal Cdc2-like protein kinase (NCLK), a heterodimer of Cdk5 and the regulatory p25 subunit, co-eluted as a ∼450-kDa size species. The NCLK from the eluted column fractions bound to PP1-specific microcystin-Sepharose and glutathioneS-transferase (GST)-I-2-coated glutathione-agarose beads. Similarly, PP1 from the eluted column fractions was pulled down with GST-Cdk5-coated glutathione-agarose beads. In vitro, NCLK phosphorylated I-2 on Thr72 and activated PP1·I-2 in an ATP/Mg-dependent manner. NCLK bound to PP1 through its Cdk5 subunit and the PP1 binding region was localized to Cdk5 residues 28–41. Our data demonstrate that in brain extract PP1·I-2 and NCLK are associated within a complex of ∼450 kDa and suggest that NCLK is one of the PP1·I-2-activating kinases in the mammalian brain.

Protein phosphatase 1 (PP1) 1 is a major Ser/Thr phosphatase involved in the regulation of metabolism, cell cycle, cell signaling, muscle contraction, and gene expression (for reviews see Refs. 1,2). PP1 is a 37-kDa catalytic subunit bound to two types of regulatory subunits: a targeting subunit and an inhibitory subunit. Targeting subunits confer substrate specificity and localize PP1 to various subcellular compartments. Inhibitory subunits suppress PP1 activity. There are three PP1 inhibitory subunits: inhibitor 1 (I-1), DARPP-32, and inhibitor 2 (I-2) (1-3). I-1 and DARPP-32 require phosphorylation for PP1 inhibitory activity, whereas nonphosphorylated I-2 inhibits PP1. These inhibitors are phosphorylated in response to many extracellular stimuli and allow PP1 to respond to various growth factors and hormones (3).
In rabbit skeletal muscle extract PP1 is found in both particulate and cytosolic fractions. PP1 in the particulate fraction is active, whereas in the cytosolic fraction it is inactive (1,2). The inactive cytosolic enzyme, a PP1⅐I-2 complex, is activated upon incubation with ATP/Mg and is hence called ATP/Mg-dependent PP1 (1). An activating factor named Fa is necessary for ATP/Mg-dependent activation of PP1⅐I-2. Fa has been identified to be glycogen synthase kinase 3 (GSK3) (4 -6). The ATP/ Mg-dependent activation is due to the phosphorylation of I-2 within the PP1⅐I-2 complex by GSK3. Nonphosphorylated I-2 suppresses PP1 activity within the PP1⅐I-2 complex. GSK3 phosphorylates I-2 on Thr 72 and relieves PP1 from I-2 inhibition (4 -9). Even though GSK3 is a well-characterized PP1⅐I-2activating kinase, several reports suggest that other kinases also phosphorylate I-2 and activate PP1⅐I-2 (10 -12).
PP1 is highly expressed in brain (1). An earlier study found that most of the PP1 in brain extract is inactive and requires incubation with ATP/Mg to become active (13). The purified enzyme is a PP1⅐I-2 complex, which is activated upon incubation with ATP/Mg in the presence of muscle GSK3. It was suggested that brain ATP/Mg-dependent PP1 is regulated in a manner similar to its muscle counterpart, via phosphorylation of I-2 (13). A type of Fa activity was partially purified from porcine brain extract (13), but the identity of this activity has remained unknown. Thus, until now it has not been clear as to which kinase activates PP1⅐I-2 in the brain.
Neuronal Cdc2-like protein kinase (NCLK) is a heterodimer of cyclin-dependent protein kinase 5 (Cdk5) and a neuronalspecific p25 regulatory subunit (reviewed in Ref. 14). Cdk5, a member of the cyclin-dependent protein kinase family, is widely expressed in various tissues and cell lines (15). However, its kinase activity is detected only in terminally differentiated neurons where it is associated with a p25 subunit (16). p25 is a proteolytic fragment of a 35-kDa protein and is expressed only in neurons (17). NCLK is involved in brain development, neurite outgrowth, cell migration, cell signaling, microtubule dynamics regulation, and Alzheimer's disease pathogenesis (18 -24). Herein we show that NCLK is complexed with PP1⅐I-2 in brain extract, phosphorylates I-2 on Thr 72 , and activates PP1⅐I-2. Our data suggest that NCLK is one of the kinases that activate PP1⅐I-2 in the central nervous system.

MATERIALS AND METHODS
cDNA Cloning-Human I-2 cDNA plasmid in pT7T3D-pac vector (American Type Culture Collection, Manassas, VA) was subcloned into two different bacterial expression vectors: pET-9a (Promega, Madison, WI) and pGEX-6P-1 (Amersham Pharmacia Biotech, Baied'Urfe, Quebec, Canada). To subclone into the pET-9a vector Pfu DNA polymerase catalyzed 30 cycles polymerase chain reaction (PCR) was carried out * This work was supported by a grant from Canadian Institute for Health Research.
using I-2 cDNA as the template and forward (5Ј-AAA AAA CAT ATG GCG GCC TCG ACG G-3Ј) and reverse (5Ј-AAA AAA GGA TCC CTA  TGA ACT TCG TAA TTT GTT TTG-3Ј) primers. The PCR condition was 95°C for 1 min, 60°C for 1 min, and 72°C for 2 min. A final 7-min extension at 72°C was followed by a 10-min incubation with Taq polymerase (Promega) at 72°C. The PCR product was ligated into a pGEM-T Easy vector (Promega) and amplified. The I-2 cDNA insert from the pGEM-T Easy vector was excised by NdeI/BamHI and ligated into a NdeI/BamHI cloning site of the pET-9a vector. To subclone into pGEX-6P-1 vector, PCR was carried out using forward primer 5Ј-AAA AAA GGA TCC ATG GCG GCC TCG ACG G-3Ј containing a BamHI site (underlined) and reverse primer 5Ј-AAA AAA GAA TTC CTA TGA ACT TCG TAA TTT GTT TTG-3Ј containing an EcoRI site (underlined) as above, except Taq polymerase and pGEM-T Easy vector steps were excluded. The PCR product was excised with BamHI/EcoRI and ligated into a BamHI/EcoRI cloning site of the pGEX-6P-1 vector.
Proteins and Peptides-I-2 was purified from bacterial culture medium as described previously (25) with some modifications. Overnight bacterial culture (10 ml) was diluted 100-fold in fresh medium and incubated with vigorous shaking at 37°C. When the A 600 of the medium reached ϳ0.6, isopropyl-␤-D-thiogalactoside was added to a final concentration of 1 mM. Shaking then continued for another 3 h at 37°C. The medium was centrifuged, and the bacterial pellet was suspended in 100 ml of cold buffer A (20 mM Tris-HCl (pH 7.5), 0.2 mM EDTA, and 0.1% ␤-mercaptoethanol) and protease inhibitor mixture (2 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml pepstatin, and 5 mg/ml benzamidin). The suspension was subjected to freezing and thawing three times and then was sonicated three times, for 30 s each, using a probe sonicator. The sonicated suspension was centrifuged at 2.7 ϫ 10 4 ϫ g for 20 min at 4°C, and the supernatant was heated in a boiling water bath for 15 min. The heated sample was centrifuged, and the supernatant was loaded onto a DEAE-Sephacel column (2.3 ϫ 23 cm) pre-equilibrated with buffer B (20 mM Tris-HCl (pH 7.5), 0.2 mM EDTA, and 1 mM DTT). The column was washed with buffer B until the effluent A 280 was less than 0.1 and then eluted with a linear gradient of 0 -0.5 M NaCl in buffer B. Fractions containing I-2 were pooled, dialyzed against cold buffer B for 4 h, and then loaded onto an Affi-Gel blue (Bio-Rad, Mississauga, Ontario, Canada) column (3 ϫ 10 cm) previously equilibrated in buffer B. The column was washed and then eluted with a 250-ml linear gradient of NaCl (0 -1 M) in buffer B. Fractions containing I-2 were combined and dialyzed against buffer A for 4 h and loaded onto a Q-Sepharose (Amersham Pharmacia Biotech) column (1.5 ϫ 5 cm) pre-equilibrated in buffer A. After washing the column with ϳ50 ml of buffer A, the column-bound I-2 was eluted with a 50-ml linear gradient of NaCl (0 -0.8 M) in buffer A. Fractions containing I-2 were pooled and concentrated by dialysis against Aquacide III (Calbiochem, San Diego, CA). The concentrated sample was dialyzed against buffer A, and stored frozen at Ϫ80°C until used.
Stephane Richard (Lady Davis Institute, Montreal) provided purified CK2. Various GST fusion proteins were purified from respective bacterial lysates as described (21). The making of the NCLK synthetic peptide substrate has been described previously (22). cAMP-dependent protein kinase (PKA) substrate Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide) and inhibitory peptide PKI were obtained from Sigma-Aldrich Ltd. The GSK3 peptide substrate KRREILSRRPSYR (28) was synthesized at the peptide synthesis facility of the University of Calgary. This peptide was phosphorylated by PKA, and the phosphorylated peptide was purified by HPLC.
The PP1⅐I-2 complex was reconstituted from PP1␣ and I-2 as described previously (8) in a 0.1-ml reconstitution mixture containing 50 mM Tris-HCl (pH 7.4), 0.1 mM EDTA, 0.1 mM DTT, 0.1 mg/ml PP1, and 0.1 mg/ml I-2. The mixture was incubated at 30°C for 60 min and loaded onto an Affi-Gel Blue column (ϳ1 ml) pre-equilibrated with 50 mM Tris-HCl (pH 7.4), 0.1 mM DTT, and 0.1 mM EDTA. The column was washed with the equilibration buffer and fractions (0.5 ml each) were collected. Fractions were analyzed by SDS-PAGE. The reconstituted PP1⅐I-2 complex was recovered in flow-through fractions.
Protein and Peptide Concentrations-I-2, PP1, GSK3, and CK2 concentrations were determined by Bio-Rad protein assay (Bio-Rad laboratories, Mississauga, Ontario, Canada) using bovine serum albumin as the standard. Concentrations of phosphorylase kinase and phosphorylase b were determined spectrophotometrically as described previously (26). The concentration of NCLK is based on its activity (22). Concentration of PKI was based on dry weight. Concentrations of NCLK and GSK3 peptide substrates were determined by amino acid analysis.
Microcystin and GST Pull-down Assays-Microcystin pull-down assay was carried out by mixing 50 l of microcystin-Sepharose (Upstate Biotechnology) beads pre-equilibrated in 50 mM Tris-HCl (pH 7.5), 0.1 mM EDTA, 5 mM MnCl 2 , 15 mM ␤-mercaptoethanol, 0.05% Nonidet P-40 with 0.2 ml of combined gel filtration fractions from Fig. 1A. The mixture was incubated at 4°C overnight with end-over-end shaking. The incubated mixture was centrifuged, and the recovered beads were washed three times with 0.5 ml of 20 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, 0.2 M NaCl, and 0.05% of Tween 20. The washed beads were mixed with 50 l of SDS-PAGE sample buffer, boiled, and centrifuged, and 10 l of the supernatant was analyzed by immunoblot analysis using the indicated antibody. To perform the GST pull-down assay 50 l of glutathione-agarose beads (Sigma-Aldrich Ltd.) coated with the indicated GST fusion protein were mixed with 150 l of the indicated protein solution and incubated overnight with end-over-end shaking at 4°C. Incubated beads were washed four times with 20 mM Tris-HCl (pH 7.5), 0.1 mM EDTA, 0.2 M NaCl, and 0.05% Tween 20. The washed beads were mixed with 50 l of SDS-PAGE sample buffer, boiled, and centrifuged, and 10 l of the supernatant was analyzed by immunoblot analysis using the indicated antibody.
Activity Assays-ATP/Mg-dependent PP1 activity was assayed as described previously (9) except that the assay mixture also contained okadaic acid (Sigma-Aldrich Ltd.) (to inhibit PP2A) and PP2B inhibitor cypermethrin (Calbiochem). Samples were preincubated at 30°C for 15 min in a mixture containing 50 mM Tris-HCl (pH 7.4), 0.1 mM EDTA, 1 mM EGTA, 0.2% ␤-mercaptoethanol, 0.2 mM ATP, and 5 mM MgCl 2 . The assay was initiated by the addition of 10 l of the preincubated sample to a vial containing 30 l of the rest of the assay components. The final concentrations of the various components in the assay were 50 mM Tris-HCl (pH 7.4), 0.2% ␤-mercaptoethanol, 5 mM caffeine, 0.5 mM MnCl 2 , 10 M [ 32 P]phosphorylase, 5 nM okadaic acid, 40 pM cypermethrin, 1 mM EGTA, and carryover preincubation mixture components. After 20 min at 30°C, 10 l of 50% trichloroacetic acid was added to the assay mixture. The assay mixture was cooled on ice for 10 min and centrifuged for 5 min using a bench top centrifuge. The supernatant (20 l) was withdrawn, spotted onto a filter paper, and counted in a liquid scintillation counter to determine the amount of 32 P i released. NCLK, GSK3, and PKA activities were assayed by using their respective peptide substrates (27). I-2 phosphorylation by NCLK was carried out in a reaction mixture containing 25 mM Hepes (pH 7.2), 0.1 mM EDTA, 1 mM DTT, 0.5 mM [␥-32 P]ATP, 10 mM MgCl 2 , 0.5 mg/ml I-2, and 400 units/ml NCLK. Phosphorylation was initiated by the addition of NCLK to the mixture containing the rest of the assay components. After indicated time points at 30°C aliquots were withdrawn, mixed with an equal volume of SDS-PAGE sample buffer, and electrophoresed on a 12% SDS-gel. The gel was stained and destained, and the I-2 band in the gel was sliced out and counted in a liquid scintillation counter to determine the amount of radioactivity that was incorporated. I-2 phosphorylation by CK2 was carried out as above by using 10 g/ml CK2 in the phosphorylation mixture. Over 4 h, CK2 incorporated 2.8 mol of phosphate/mol of I-2. CK2-phosphorylated I-2 was desalted by a Sephadex G-25 column and used to generate the data shown in Fig. 4.
Partial Purification of PP1 from Brain Extract-All procedures were performed at 4°C. Fresh bovine brain (0.5 kg) was homogenized for 1 min in 1 liter of buffer C (20 mM MOPS (pH 7.4), 50 mM ␤-glycerophosphate, 1 mM EDTA, 1 mM DTT, and 15 mM MgCl 2 ), which contained protease inhibitor mixture. The homogenate was centrifuged at 10 4 ϫ g for 30 min, and the supernatant was centrifuged at 10 5 ϫ g for 45 min. The resulting clear supernatant was loaded onto a DEAE-Sephacel column (2.5 ϫ 45 cm) previously equilibrated with buffer C. The flowthrough fraction containing PP1 activity was loaded onto an SP-Sepharose (Amersham Pharmacia Biotech) column (1 ϫ 60 cm) pre-equilibrated with buffer C. The column was washed with buffer C and was eluted with 400 ml of a linear NaCl gradient (0 -0.5 M) in buffer C. Fractions containing PP1 activity were pooled, dialyzed against buffer C, concentrated by dialysis against Aquacide III to ϳ8 ml, and chromatographed through an FPLC Superose 12 gel filtration column (Amersham Pharmacia Biotech) column (2 ϫ 70 cm) equilibrated and eluted with 20 mM MOPS (pH 7.4), 1 mM EDTA, 1 mM DTT, and 0.15 M NaCl. Fractions 41-43, containing PP1 activity, were combined. A portion of the combined fractions was used to generate Figs. 1B and 2, and the rest was loaded onto a hydroxylapatite column (1 ϫ 15 cm) pre-equilibrated with buffer C. The column was washed with ϳ40 ml of buffer C and eluted with a 50-ml linear gradient (0 -0.5 M) of K 2 HPO 4 (pH 7.4) in buffer C.
Phosphopeptide Purification and Peptide Sequencing-I-2 was phosphorylated for 16 h by NCLK in a 0.7-ml reaction mixture containing 25 mM Hepes (pH 7.2), 0.1 mM EDTA, 1 mM EGTA, 10 mM NaF, 0.5 mM [␥-32 P]ATP, 10 mM MgCl 2 , 1 mg/ml I-2, and 400 units/ml NCLK. Phosphorylated I-2 was desalted through a Sephadex G-25 column, lyophilized, dissolved in 0.2 ml of 50 mM NH 4 HCO 3 (pH 8.0) containing 50 g/ml trypsin, and incubated at 37°C for 16 h. The incubated tryptic digest was separated by a HPLC reverse phase column previously equilibrated with 0.1% trifluoroacetic acid as described (26) using 0 -40% acetonitrile gradient in 50 min. Radioactive fractions were lyophilized, dissolved in 50 mM NH 4 HCO 3 containing 50 g/ml of chymotrypsin, and incubated for 16 h at 37°C. The chymotryptic digest was lyophilized, dissolved in 25 mM Hepes (pH 7.2), 1 mM EDTA, 1 mM EGTA, and loaded onto a DEAE-Sephacel column (ϳ2 ml) previously equilibrated in 25 mM Hepes (pH 7.2). The column was washed, and bound peptide was eluted with 0.3 M NaCl in the equilibration buffer. Fractions containing radioactivity were lyophilized, dissolved in 25 mM Hepes (pH 7.2), 1 mM EDTA, 1 mM EGTA, and chromatographed through an HPLC column as described above. Only one radioactive peptide was recovered. This peptide was sequenced at the Department of Biochemistry and Microbiology, University of Victoria by using a gas-phase micro-sequencer.

PP1 in Bovine Brain
Extract-In a previous study PP1⅐I-2 was found to be the predominant form of PP1 in brain extract (13). To learn more about brain PP1⅐I-2, we chromatographed a fresh bovine brain extract through a DEAE-Sephacel column then an SP-Sepharose column. The effluent containing ATP/ Mg-dependent PP1 activity was then analyzed by an FPLC Superose 12 gel filtration column. PP1 activity eluted from the gel filtration column as a broad peak (Fig. 1A). Immunoblot analyses of various gel filtration fractions (by using anti-PP1 or anti-I-2 antibodies) indicated that both PP1 and I-2 were present in column fractions containing PP1 activity (data not shown). The gel filtration analysis indicated that the molecular size of PP1⅐I-2 in Fig. 1A (fraction 42) is ϳ450 kDa. Because the molecular size of PP1⅐I-2 is ϳ70 kDa, these observations corroborate previous reports (2) and indicate that PP1⅐I-2 is complexed with other protein (s) in the brain.
As observed previously (13), we found that PP1 has low activity in brain extract, is complexed with I-2, and is activated by preincubation with ATP/Mg (data not shown). These observations are consistent with a previous report (13) and indicate that the PP1⅐I-2-activating factor is present in brain. An immunoblot analysis using anti-GSK3 antibody showed that GSK3 is present in Fig. 1A column fractions (data not shown). To examine if GSK3 is responsible for ATP/Mg-dependent PP1⅐I-2 activation, we combined gel filtration fractions 41-43 containing peak PP1 activity from Fig. 1A. We preincubated an aliquot from the combined fractions with ATP/Mg the in the presence of the GSK3 inhibitor LiCl (29) to block GSK3 action and assayed PP1 activity. Surprisingly, PP1 activity was only ϳ29% inhibited (Fig. 1B), although a kinase assay confirmed that LiCl had completely suppressed GSK3 activity in our samples (data not shown). These data suggested that the brain contains a kinase other than GSK3 that phosphorylates I-2 and activates PP1.
To identify the PP1⅐I-2-activating kinase we included 2 mM Ca 2ϩ -chelator EGTA, or 50 M PKA inhibitory peptide (PKI) in the preincubation mixture and assayed the Fig. 1A combined column fractions. PP1 activity was approximately the same in samples preincubated in the presence of EGTA, PKI, or buffer control (data not shown). These data indicated that LiCl-insensitive ATP/Mg-dependent PP1 activation in Fig. 1B could not be due to the involvement of PKA or any Ca 2ϩ -dependent kinases (protein kinase C, calmodulin-dependent kinase 2, or phosphorylase kinase).

FIG. 1. Gel filtration of PP1 in bovine brain extract and inhibition of PP1 activity by LiCl and olomoucine.
Bovine brain extract was chromatographed through a DEAE-Sephacel column, and the flow-through fractions containing PP1 were chromatographed through an SP-Sepharose column. ATP/Mg-dependent PP1 activity recovered from the SP-Sepharose column was analyzed by a FPLC Superose 12 gel filtration column (2 ϫ 70 cm). Fractions 41-43 containing the PP1 activity peak from the gel filtration column were combined and analyzed for the effect of LiCl and olomoucine. A, gel filtration. Gel filtration was performed by using a FPLC system (Amersham Pharmacia Biotech) at 1 ml/min flow rate. Fractions (1 ml each) were collected, and 10 l from the indicated fractions was analyzed for PP1 and NCLK activities. B, inhibition of ATP/Mg-dependent PP1 activity. Aliquots from the combined gel filtration fractions were preincubated with ATP/Mg in the presence of LiCl (20 mM) and/or olomoucine (50 M) or buffer control. Each preincubated sample was assayed for PP1 activity. PP1 activity is expressed as a percentage of the control sample preincubated with buffer. Values are the average Ϯ S.D. of three independent determinations. (11). To determine if MAPK is a PP1⅐I-2-activating kinase, we analyzed Fig. 1A column fractions by immunoblot analysis using anti-MAPK antibody. Our antibody that detected nanograms of MAPKs (p43 erk1 and p42 erk2 ) in brain extract failed to show any immunoreactivity (data not shown). By similar immunoblot analyses we determined that CK1 and CK2 were also absent in Fig. 1A column fractions (data not shown). These observations indicated that MAPK, CK1, and CK2 are not responsible for activating PP1⅐I-2 in Fig. 1B.

MAPK phosphorylates I-2 in vitro
I-2 Thr 72 is followed by a proline residue (9), and NCLK recognizes an (S/T)P motif (22). An immunoblot analysis using anti-NCLK antibody (data not shown) as well as an NCLK activity assay demonstrated that NCLK was present in almost all fractions containing PP1 and displayed a gel filtration profile similar to PP1 (Fig. 1A). We therefore included the NCLK inhibitor olomoucine (27,30) in the preincubation mixture, and assayed PP1 and NCLK activities in the combined column fractions from Fig. 1A. Interestingly, not only did olomoucine completely inhibit NCLK activity (data not shown), but it also suppressed ATP/Mg-dependent PP1 activity by ϳ30% (Fig. 1B). Inclusion of both olomoucine and LiCl in the preincubation mixture inhibited ϳ60% of ATP/Mg-dependent PP1 activity (Fig. 1B). Thus, blocking NCLK action partially suppressed ATP/Mg-dependent activation of PP1 in our samples. These data suggested that ATP/Mg-dependent PP1 activity in the brain might also be regulated by NCLK.
NCLK Is Associated with PP1 in Fig. 1A Column Fractions-To substantiate the above suggestion, we incubated combined column fractions from Fig. 1A with PP1-specific microcystin-Sepharose beads (31). The beads were washed and immunoblotted with either anti-NCLK or anti-PP1 antibody. PP1 was pulled down with microcystin-Sepharose beads as expected (Fig. 2B). Importantly, NCLK also was pulled down with microcystin-Sepharose beads ( Fig. 2A). We then performed two GST pull-down assays. In the first assay, glutathione-agarose beads coated with GST-I-2 were mixed with the combined column fractions, washed, and immunoblotted by using an anti-PP1 or anti-NCLK antibody. In the second assay, glutathione-agarose beads coated with GST-Cdk5 were mixed with the combined column fractions, washed, and immunoblotted with anti-PP1 antibody. As shown in Fig. 2, C and D, both PP1 and NCLK specifically were pulled down with the GST-I-2-coated beads from the column fractions. Likewise, PP1 was pulled down with the beads coated with GST-Cdk5 but not GST control (Fig. 2E). Finally, we chromatographed the combined column fractions from Fig. 1A through a hydroxylapatite column. An immunoblot analysis of effluent fractions, using anti-PP1 and anti-NCLK antibodies, indicated that PP1 and NCLK co-eluted from the column (data not shown). Thus, NCLK and PP1 from brain extract could not be separated from each other by DEAE-Sephacel, SP-Sepharose, FPLC gel filtration, or hydroxylapatite chromatographies. Taken together, these data indicated that NCLK is complexed with PP1⅐I-2 in the brain extract and is likely to be a PP1⅐I-2-activating kinase.
NCLK Phosphorylates Thr 72 of I-2-Because PP1⅐I-2 activation occurs via I-2 phosphorylation (5-9) we examined whether NCLK phosphorylates I-2. We incubated I-2 with NCLK in the presence of [␥-32 P]ATP/Mg 2ϩ and the product analyzed by SDS-PAGE/autoradiography. I-2 was phosphorylated in a time-dependent manner (Fig. 3, A and B). Over 16 h, NCLK incorporated ϳ1.02 mol of phosphate/mol of I-2. Because NCLK used in this study was purified from brain extract, we examined to see if our NCLK preparations were contaminated by any other kinase(s). Immunoblot analyses using antibodies directed to various kinases determined that our kinase preparations did not contain any detectable MAPK (p42 erk1 and p43 erk2 ), CK1, or CK2. Similarly, our kinase preparations contained neither PKA nor GSK3 activity (data not shown), and the phosphorylation of I-2 by our NCLK preparations was insensitive to 2 mM EGTA. Finally, we included NCLK inhibitor olomoucine in the assay mixture and monitored I-2 phosphorylation. Olomoucine inhibited I-2 phosphorylation by our NCLK preparation in a dose-dependent manner (Fig. 3, C and D). From these data we concluded that NCLK phosphorylates I-2.
PP1⅐I-2 activation requires phosphorylation of I-2 on Thr 72 (5)(6)(7)(8)(9). To determine if NCLK phosphorylates I-2 Thr 72 , NCLKphosphorylated 32 P-labeled I-2 was trypsinized and the product was fractionated over an HPLC reverse phase column. Only one radioactive peak eluted from the column (data not shown). Radioactive fractions were combined and digested with chymotrypsin. The chymotryptic digest was chromatographed sequentially through a DEAE-Sephacel column and an HPLC reverse phase column. Only one radioactive peptide was recovered. Purified peptide was sequenced by using a gas-phase micro-sequencer. Ile, Asp, Glu, Pro, Ser, and Tyr were identified as the first, second, third, fourth, fifth, and seventh residues, respectively, of this peptide (Table I). The sixth residue (indicated by Xaa) could not be identified, indicating that this residue was phosphorylated. This notion was confirmed by the release of radioactivity during the sixth sequencing cycle. Based on these data and the amino acid sequence of human I-2 (32), we concluded that this phosphopeptide extends from I-2 residues 67-73 and that Thr 72 is the phosphorylation site. Thus NCLK indeed phosphorylates Thr 72 of I-2.
Phosphorylation of I-2 by NCLK and GSK3-GSK3 phosphorylates I-2 on Thr 72 . However, GSK3 phosphorylates CK2phosphorylated I-2 better than nonphosphorylated I-2 (8,9). Because we found that NCLK also phosphorylates I-2 on Thr 72 , we further evaluated phosphorylation of I-2 by NCLK and GSK3. Over 16 h, NCLK and GSK3 incorporated ϳ1 and 0.6 mol of phosphate/mol of I-2, respectively. We then phosphorylated I-2 and CK2-phosphorylated I-2 by NCLK or GSK3 under identical conditions and compared the results (Fig. 4). NCLK incorporated 0.47 mol of phosphate/mol of I-2 and 0.51 mol of phosphate/mol of CK2-phosphorylated I-2. GSK3 as expected, incorporated 0.25 mol of phosphate/mol of I-2 and 0.61 mol of phosphate/mol of CK2-phosphorylated I-2. Thus, NCLK phosphorylated I-2 and CK2-phosphorylated I-2 in a similar manner. GSK3, on the other hand, phosphorylated CK2-phosphorylated I-2 2-fold better than the nonphosphorylated counterpart. These data indicate that GSK3 and NCLK display different substrate specificity for I-2 phosphorylation.
NCLK Directly Binds to PP1-We found that NCLK is complexed with PP1 in brain extract (Fig. 2). To examine if NCLK binds to PP1, glutathione-agarose beads coated with GST, GST-Cdk5, or GST-p25 were mixed with recombinant PP1, washed, and immunoblotted with anti-PP1 antibody. PP1 was found to be associated with GST-Cdk5 but not with GST-p25 or GST (Fig. 5). These data indicate that NCLK binds to PP1 through its Cdk5 subunit.
In a previous study almost all PP1 in brain extract was found to exist as PP1⅐I-2 and required preincubation with ATP/Mg to become active (13). In this study we partially purified brain PP1⅐I-2 and found that GSK3 was present in our preparation. Importantly, a complete inhibition of GSK3 activity by LiCl suppressed only ϳ29% of ATP/Mg-dependent PP1⅐I-2 activation in our preparation (Fig. 1B). These data indicate that GSK3 is not the sole PP1⅐I-2 activator in the brain.
We found that the ATP/Mg-dependent activation of our partially purified PP1⅐I-2 preparation was sensitive to the NCLK inhibitor olomoucine (Fig. 1B). NCLK was pulled down from our preparation with microcystin-Sepharose and GST-I-2coated glutathione-agarose beads (Fig. 2, A and C). Similarly, PP1 in our preparation bound to glutathione-agarose beads coated with GST-Cdk5 (Fig. 2E). In vitro, NCLK phosphorylated I-2 on Thr 72 and activated PP1⅐I-2 in an ATP/Mg-dependent manner (Fig. 7). Taken together, these data strongly argue that NCLK is one of the PP1⅐I-2-activating kinases in the brain. PP1 displays a broad substrate specificity and dephosphorylates targets of many different protein kinases (1, 2). As dis-cussed above, NCLK phosphorylates I-1 and DARPP-32 (27,33). In this study we showed that NCLK also phosphorylates I-2. These observations suggest that NCLK plays a central role in neuronal signaling by phosphorylating PP1 inhibitory subunits I-1, DARPP-32, and I-2.
It is established that PP1 binds to I-2 (1-3, 8, 9). We found that PP1 also binds to NCLK (Fig. 5). These observations suggest that PP1 is the central molecule that holds I-2 and NCLK together within a PP1⅐I-2⅐NCLK complex. Our GST pulldown assay demonstrated that, from the gel filtration column fractions containing PP1⅐I-2⅐NCLK complex, PP1 and NCLK are pulled down with GST-I-2 (Fig. 2, C and D) and PP1 is pulled down with GST-Cdk5 (Fig. 2E). These observations raise a question as to how the PP1⅐I-2⅐NCLK complex could bind to an exogenous GST-I-2 or GST-Cdk5.
It has been suggested that, in vivo, free PP1 and I-2 are in a dynamic equilibrium with PP1⅐I-2 and an excess of I-2 can replace PP1-bound I-2 (34,35). As shown in Fig. 1A, PP1 in brain extract elutes form a gel filtration column as a component of various species with sizes of ϳ40 to ϳ450 kDa. Some of these species may represent PP1, PP1⅐I-2, PP1⅐NCLK, or PP1⅐I-2⅐NCLK. Thus, it is possible that PP1, I-2, and NCLK may also be in a dynamic equilibrium with the PP1⅐I-2⅐NCLK complex in the brain. Because I-2 within PP1⅐I-2 may be displaced by exogenous I-2 (34,35), GST-I-2 could similarly displace I-2 from PP1⅐I-2⅐NCLK and form a PP1⅐GST-I-2⅐NCLK species. Likewise, GST-Cdk5 may compete with Cdk5 within the PP1⅐I-2⅐NCLK complex and displace NCLK to result in the formation of a PP1⅐I-2⅐GST-Cdk5 complex.
I-2 is bound to PP1 in tissue extracts and inhibits PP1 in its nonphosphorylated state. GSK3 phosphorylates I-2 on Thr 72 within the PP1⅐I-2 complex. This phosphorylation causes a conformational change in I-2 and also within the PP1⅐I-2 complex, leading to PP1 activation without complex dissociation (3,8,9,34). Activated PP1 rapidly dephosphorylates I-2. This dephosphorylation, however, does not cause immediate loss of PP1 activity and only after some time the complex returns to its inactive conformation (3,8,9,34). GSK3 phosphorylation of I-2 within PP1⅐I-2 has been suggested to be short-lived (3,8,9).
To further investigate the phosphorylation of I-2 we incubated PP1⅐I-2 with NCLK in the presence of [␥-32 P]ATP/Mg 2ϩ . We found that the incubation robustly activated PP1 within PP1⅐I-2. However, when we analyzed the product of the incubation by SDS-PAGE/autoradiography, we could not detect any significant I-2 phosphorylation (data not shown). These observations suggest that phosphorylation of I-2 by NCLK within the PP1⅐I-2 complex is also a transient event.
I-2 is phosphorylated on Ser 86 , Ser 120 , and Ser 121 in vivo (36), and these sites are phosphorylated by CK2 in vitro (1,37). CK2 phosphorylation does not activate PP1⅐I-2 but enhances Thr 72 phosphorylation by GSK3 (8,9). We found that NCLK also phosphorylates I-2 on Thr 72 . However, NCLK phosphorylation of I-2 is insensitive to a prior phosphorylation by CK2 (Fig. 4). These data indicate that NCLK and GSK3 display different substrate specificity for I-2 phosphorylation. In neurons CK2/GSK3 and NCLK may activate PP1⅐I-2 in response to different cellular stimuli. They may also function in different regions of the brain or in different subcellular compartments.
In this study we showed that a major PP1-binding site is located within a 14-residue-long Cdk5 region (Fig. 6). Within this region there is a 34 RVRL 37 sequence (Fig. 6A) similar to an RVXF sequence motif found in many PP1-binding proteins (2,38,39). This putative PP1-binding site is located in between the ATP binding region and the PSSLAIR helix of Cdk5. This site is identical in Cdk5 from human, bovine, mouse, rat, and Xenopus (15, 40 -43) and has a conserved substitution (Asp 40 3 Glu) in Drosophila (44). In crystal structure, this sequence is an exposed loop available for interaction with other proteins (45,46). It should also be noted that deletion mutant GST-Cdk5-(42-292) also displays a weak PP1 binding (Fig. 6B), indicating that a low affinity PP1 binding region is located within Cdk5 residues 42-292. More studies will be required to identify this low affinity PP1 binding region.