Phosphorylation of casein kinase II by p34cdc2. Identification of phosphorylation sites using phosphorylation site mutants in vitro.

The α and β subunits of casein kinase II are dramatically phosphorylated in cells that are arrested in mitosis (Litchfield, D. W., Lüscher, B., Lozeman, F. J., Eisenman, R. N., and Krebs, E. G.(1992) J. Biol. Chem. 267, 13943-13951). Comparative phosphopeptide mapping experiments indicated that the mitotic phosphorylation sites on the α subunit of casein kinase II can be phosphorylated in vitro by p34cdc2. In the present study, we have demonstrated that a glutathione S-transferase fusion protein encoding the C-terminal 126 amino acids of the α subunit is phosphorylated by p34cdc2 at the same sites as intact casein kinase II, indicating that the mitotic phosphorylation sites are localized within the C-terminal domain of α. Four residues within this domain, Thr-344, Thr-360, Ser-362, and Ser-370, conform to the minimal consensus sequence for p34cdc2 phosphorylation. Synthetic peptides corresponding to regions of α that contain each of these residues are phosphorylated by p34cdc2 at these sites. Furthermore, alterations in the phosphorylation of the glutathione S-transferase proteins encoding the C-terminal domain of α are observed when any of the four residues are mutated to alanine. When all four residues are mutated to alanine, the fusion protein is no longer phosphorylated by p34cdc2 at any of the sites that are phosphorylated in mitotic cells. These results indicate that Thr-344, Thr-360, Ser-362, and Ser-370 are the sites on the α subunit of casein kinase II that are phosphorylated in mitotic cells.

The ␣ and ␤ subunits of casein kinase II are dramatically phosphorylated in cells that are arrested in mitosis (Litchfield, D. W., Lü scher, B., Lozeman, F. J., Eisenman, R. N., and Krebs, E. G. (1992) J. Biol. Chem. 267, 13943-13951). Comparative phosphopeptide mapping experiments indicated that the mitotic phosphorylation sites on the ␣ subunit of casein kinase II can be phosphorylated in vitro by p34 cdc2 . In the present study, we have demonstrated that a glutathione S-transferase fusion protein encoding the C-terminal 126 amino acids of the ␣ subunit is phosphorylated by p34 cdc2 at the same sites as intact casein kinase II, indicating that the mitotic phosphorylation sites are localized within the C-terminal domain of ␣. Four residues within this domain, Thr-344, Thr-360, Ser-362, and Ser-370, conform to the minimal consensus sequence for p34 cdc2 phosphorylation. Synthetic peptides corresponding to regions of ␣ that contain each of these residues are phosphorylated by p34 cdc2 at these sites. Furthermore, alterations in the phosphorylation of the glutathione S-transferase proteins encoding the C-terminal domain of ␣ are observed when any of the four residues are mutated to alanine. When all four residues are mutated to alanine, the fusion protein is no longer phosphorylated by p34 cdc2 at any of the sites that are phosphorylated in mitotic cells. These results indicate that Thr-344, Thr-360, Ser-362, and Ser-370 are the sites on the ␣ subunit of casein kinase II that are phosphorylated in mitotic cells.
Biochemical and genetic studies have demonstrated that the p34 cdc2 protein kinase is an indispensable regulator of events leading to the division of eukaryotic cells (for reviews see Refs. [1][2][3][4]. To ensure that the division of cells is very precisely regulated, the activity of this protein serine/threonine kinase is exquisitely controlled through its interactions with regulatory cyclins and through phosphorylation of p34 cdc2 itself. p34 cdc2 is defined as a cyclin-dependent kinase since it is inactive unless it is associated with a regulatory cyclin subunit. Furthermore, p34 cdc2 is inhibited by phosphorylation of Tyr-15 and/or Thr-14 (5, 6) but requires phosphorylation of Thr-161 to be activated (7,8). CAK (the p34 cdc2 activating kinase) is responsible for phosphorylation of Thr-161 (9 -11), while the phosphorylation state of Thr-14 and/or Tyr-15 is at least in part controlled by the relative activities of the Wee1 protein kinase (12,13) and cdc25 protein phosphatase (14,15).
Concomitant with the activation of p34 cdc2 at the G 2 -M transition of eukaryotic cells is a massive burst of protein phosphorylation. Many of the events that are associated with entry into mitosis including nuclear envelope breakdown, transcriptional termination, nucleolar disassembly, cytoskeletal reorganization, and chromosome condensation appear to be associated with protein phosphorylation. While it is evident that p34 cdc2 directly phosphorylates a number of proteins at the G 2 -M transition, there are also indications that p34 cdc2 could indirectly regulate phosphorylation events through its phosphorylation of other protein kinases (16 -23).
One protein serine/threonine kinase that could be regulated by p34 cdc2 is casein kinase II (CKII), 1 which has been shown to be dramatically phosphorylated in mitotic cells (19 -21). CKII is a tetrameric enzyme composed of two catalytic (␣ and/or ␣Ј-subunits) and two additional subunits (␤ subunits) (for reviews, see Refs. 24 -26). Our previous studies demonstrated that p34 cdc2 phosphorylates the ␤ subunit of CKII at Ser-209, a site that is maximally phosphorylated in mitotic cells (20). Interestingly, the ␣ subunit (but not the ␣Ј-subunit) of CKII is also dramatically phosphorylated in mitotic avian and mammalian cells (21). This result suggests that there may be differences in the functional or regulatory properties of the isozymic forms of the catalytic subunit of CKII. Our analyses demonstrated that the mitotic phosphorylation sites on ␣ can be phosphorylated in vitro by p34 cdc2 . To facilitate efforts to examine the functions of CKII during mitosis and how phosphorylation may affect these functions, the objective of the present study was directed toward identification of the sites on the ␣ subunit of CKII that are phosphorylated by p34 cdc2 so that non-phosphorylatable forms of CKII could be prepared by mutagenesis. Utilizing synthetic peptides and glutathione Stransferase (GST) fusion proteins containing the C-terminal domain of the human CKII ␣ subunit, p34 cdc2 phosphorylation sites were identified as Thr-344, Thr-360, Ser-362, and Ser-370. Furthermore, following mutation of each of these residues to alanine residues, fusion proteins containing the C-terminal domain of CKII ␣ were no longer phosphorylated by p34 cdc2 at any of the sites that are phosphorylated in mitotic cells.

Materials
Synthetic peptides (peptide 1, Pro-Gly-Gly-Ser-Thr-Pro-Val-Ser-Ser-Ala; peptide 2, Ile-Ser-Ser-Val-Pro-Thr-Pro-Ser-Pro-Leu-Gly-Pro-Leu-Ala-Gly; peptide 3, Ile-Ser-Ser-Val-Pro-Thr-Pro-Ala-Pro-Leu-Gly-Pro-Leu-Ala-Gly; peptide 4, Leu-Gly-Pro-Leu-Ala-Gly-Ser-Pro-Val-Ile-Ala-Ala) were synthesized with an Applied Biosystems model 431A peptide * This work was supported by grants from the Manitoba Health Research Council and the National Cancer Institute of Canada with funds from the Canadian Cancer Society and from the Terry Fox Foundation (to D. W. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Recipient of a Studentship from the George H. Sellers Endowment Fund.
§ Research Scientist of the National Cancer Institute of Canada. To whom correspondence should be addressed. Tel.: 204-787-2177; Fax: 204-787-2190; E-mail: litchfi@cc.umanitoba.ca. synthesizer using Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry and were purified by reverse phase high pressure liquid chromatography using a C18 column as described previously (20). Polyclonal antipeptide antibodies directed against the C terminus of CKII ␣ (anti-␣ 376 -391 ) and the C terminus of CKII ␤ (anti-␤ 198 -215 ) were previously described (20,21). p34 cdc2 was purified from nocodazole-arrested MANCA cells as described previously (27) and had a specific activity of approximately 400 nmol/min/mg when assayed at 30°C using histone H1 as substrate or approximately 200 nmol/min/mg using the Ser-209 peptide (modeled after CKII ␤) as substrate (27). [␥-32 P]ATP was obtained from DuPont NEN. Thermolysin was obtained from Boehringer Mannheim. The Muta-Gene M13 in vitro mutagenesis kit (version 2) and the Bst DNA sequencing kit were obtained from Bio-Rad. Restriction enzymes and T4 DNA ligase were from Pharmacia Biotech Inc. Protein A-Sepharose, glutathione-agarose, glutathione, and nocodazole were obtained from Sigma. Thin layer cellulose plates (0.1 mm) for phosphopeptide mapping and phosphoamino acid analysis were from Merck. Other chemicals and reagents were of reagent grade.

Methods
Expression of GST Fusion Proteins-The cDNAs encoding the human CKII ␣ and human CKII ␣Ј-subunits (28) were obtained as clones, which had been cloned into the BamHI site of pBluescript SK ϩ . To express GST fusion proteins containing the C-terminal domain of CKII ␣ (GST-␣C126), a 629-base pair Sau3A1 fragment of the human CKII ␣ cDNA, encoding the C-terminal 126 amino acids of CKII ␣, was cloned into the BamHI site of the pGEX-1 vector (29) to make the pGEX-1/ CKII␣ construct. To express a fusion protein containing the C-terminal domain of CKII ␣Ј (GST-␣ЈC51), a 185-base pair BglII-BamHI fragment of the human CKII ␣Ј-cDNA, encoding the C-terminal 51 amino acids of CKII ␣Ј, was similarly cloned into the BamHI site of pGEX-1 to make the pGEX-1/CKII␣Ј construct. Site-directed mutants of GST-␣C126 were prepared by the method of Kunkel (30) using a Muta-Gene M13 in vitro mutagenesis kit (Bio-Rad) according to manufacturer's recommendations. Briefly, a 426-base pair fragment encoding the C-terminal region of CKII ␣ was obtained by digestion of the pBluescript SK ϩ / CKII␣ plasmid with SphI (at nucleotide 1174 within the CKII ␣ cDNA) and with EcoRI (within the multiple cloning site of pBluescript SK ϩ ) and cloned into M13 mp18. Single-stranded phage DNA was prepared and used as template for oligonucleotide-directed mutagenesis using the following oligonucleotides: 5Ј-AGGGGGCAGTGCGCCCGTCA-3Ј (for T344A), 5Ј-CAGTGCCAGCCCCTTCA-3Ј (for T360A), 5Ј-CAAC-CCCTGCACCCCTT-3Ј (for S362A), 5Ј-TGGCAGGCGCACCAGTG-3Ј (for S370A), and 5Ј-CAGTGCCAGCCCCTGCA-3Ј (for T360A and S362A). Mutants were identified and completely confirmed by DNA sequencing of single-stranded templates using the dideoxy method of Sanger et al. (31) with a Bst sequencing kit (Bio-Rad). Subsequent rounds of mutagenesis were performed to obtain constructs with mutations at multiple sites. Mutant SphI-EcoRI fragments were cloned into the pGEX-1/CKII ␣ construct encoding GST-␣C126 to express fusion proteins harboring mutations at the putative phosphorylation sites.
Fusion proteins were expressed in Escherichia coli JM109 and purified using glutathione-agarose as described previously (29). Fusion proteins were eluted from glutathione-agarose using buffer (20 mM Tris-Cl pH 8.0, 100 mM NaCl, 1 mM EDTA) containing 5 mM reduced glutathione. Protein determinations were by the method of Bradford (38) using ␥-globulin as standard.
Phosphorylation Reactions-Purified fusion proteins (typically 4 g/ assay) were incubated with purified p34 cdc2 in kinase buffer (50 mM Tris-Cl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol, 100 M ATP) containing [␥-32 P]ATP (30 -60 Ci/nmol) in a total volume of 30 l. Kinase reactions were initiated by the addition of purified p34 cdc2 (typically 50 pM units of enzyme), and the assay performed at 30°C with constant agitation for the indicated length of time. Reactions were terminated by the addition of EDTA to a final concentration of 20 mM. Phosphorylated fusion proteins were immunoprecipitated by the addition of anti-␣ 376 -391 antiserum (0.5 l/g fusion protein) and protein A-Sepharose beads in antibody buffer (20 mM Tris-Cl, pH 7.5, 50 mM NaCl, 0.5% Nonidet P-40, 0.5% deoxycholate, 0.5% SDS) supplemented with phosphatase inhibitors (20 mM NaF, 20 mM ␤-glycerophosphate). After incubation for 60 min on ice, the beads were collected by centrifugation and washed four times with L buffer (phosphate-buffered saline, pH 7.5, 1% Nonidet P-40, 0.5% deoxycholate) containing 1% aprotinin and phosphatase inhibitors. The washed beads were then resuspended in Laemmli sample buffer, and the samples were boiled for 3-5 min prior to analysis by SDS-polyacrylamide gel electrophoresis (32). Alterna-tively, fusion proteins were recovered from phosphorylation reactions by glutathione-agarose bead purification (29). Bovine CKII was purified (37) and phosphorylated using purified p34 cdc2 as described previously (20,21). Synthetic peptides (1 mM, final concentration) were similarly phosphorylated with purified p34 cdc2 in kinase buffer containing [␥-32 P]ATP (30 -60 Ci/nmol). After terminating the reaction by the addition of EDTA, phosphopeptides were separated from the free ␥-32 P using thin layer cellulose chromatography as described below using Scheidtmann buffer (33) and visualized with autoradiography.
To determine the stoichiometry of phosphorylation of fusion proteins, the following procedure was followed. Known amounts (2, 4, 6 g) of each fusion protein were subjected to SDS-polyacrylamide gel electrophoresis and visualized by staining with Coomassie Blue. Typically, each fusion protein displayed one major band corresponding to undegraded GST-fusion protein since it reacted on immunoblots with antibodies directed against the C terminus of CKII ␣ (anti-␣ 376 -391 ) as well as bands representing degradation products that were not reactive with anti-␣ 376 -391 antibodies. Densitometry of the Coomassie Blue-stained gel was performed to determine the proportion of intact GST fusion protein. The phosphate incorporation into the undegraded GST fusion proteins were subsequently determined by analysis on a Phosphorimager (Molecular Dynamics). Stoichiometry of phosphorylation was calculated by the following formula: pmol of phosphate incorporated into intact fusion protein/pmol of intact fusion protein.
Two-dimensional Phosphopeptide Mapping and Phosphoamino Acid Analysis-The 32 P-labeled fusion proteins were visualized by autoradiography of unfixed dried SDS-polyacrylamide gels and excised. Proteins were recovered from homogenized gel slices by trichloroacetic acid precipitation and oxidized with performic acid as described previously (34 -36). The oxidized proteins were then resuspended in 25 l of 50 mM NH 4 HCO 3 and digested overnight (16 h) at 56°C with 10 g of thermolysin in the presence of 1 mM CaCl 2 . An additional 5 g of thermolysin was then added, and digestion continued for at least another 4 h. Following digestion, samples were repeatedly lyophilized (four times) using a Speedvac concentrator. Thermolytic digests of the phosphorylated fusion proteins were subjected to electrophoresis using pH 1.9 buffer (15% acetic acid, 5% formic acid) followed by ascending chromatography with Scheidtmann buffer (isobutyric acid:pyridine:acetic acid: butanol:water (65:5:3:2:29)) as described previously (21,33). Phosphopeptides were visualized by autoradiography. For phosphoamino acid analysis, individual spots were scraped from cellulose plates, and phosphopeptides were eluted from the cellulose with pH 1.9 buffer. Following removal of the buffer using a SpeedVac, phosphopeptides were subjected to partial acid hydrolysis, and partial hydrolysis products were subjected to two-dimensional electrophoresis as described (36).

RESULTS AND DISCUSSION
Previous work had shown that sites within the C-terminal domain of the ␣ subunit of CKII were phosphorylated in cells arrested in mitosis with nocodazole and that immunopurified p34 cdc2 was capable of phosphorylating the same sites in vitro (21). To provide additional evidence that the p34 cdc2 phosphorylation sites are localized to the C-terminal domain of CKII ␣, a GST fusion protein encoding the C-terminal 126 amino acids of CKII ␣ (GST-␣C126) was tested as a substrate for purified p34 cdc2 . As shown in Fig. 1, GST-␣C126 is effectively phosphorylated by p34 cdc2 . By comparison, neither GST nor a GST fusion protein encoding the C-terminal 51 amino acids of CKII-␣Ј (GST-␣ЈC51) is phosphorylated. In addition, no phosphorylation of GST-␣C126 was observed when p34 cdc2 was omitted from the reaction mixture.
To determine whether GST-␣C126 and intact CKII ␣ are phosphorylated at the same sites, we performed comparative phosphopeptide mapping. Thermolytic digestion of phosphorylated GST-␣C126 and purified CKII ␣ resulted in the production of four major phosphopeptides that comigrated when phosphopeptides from each of the two samples were mixed (Fig. 2). It is therefore apparent that all of the sites that are phosphorylated on CKII ␣ by p34 cdc2 are present in the C-terminal 126 amino acids of the protein. Furthermore, since comparative phosphopeptide mapping experiments had previously demonstrated that phosphopeptides a, b, and c obtained following in vitro phosphorylation of purified CKII comigrate with phosphopeptides obtained following 32 P labeling of mitotic Jurkat cells (21), these results indicate that GST-␣C126 is phosphorylated at sites that are phosphorylated in mitotic cells. We previously noted that purified CKII is phosphorylated at a site, represented by phosphopeptide d (Fig. 2), that was not detected from samples of CKII obtained from intact cells (21). It is apparent that this phosphopeptide, which is also present following phosphorylation of GST-␣C126, does not represent an in vivo phosphorylation site.
Inspection of the amino acid sequence of the C-terminal domain of CKII ␣ revealed the presence of four residues, Thr-344, Thr-360, Ser-362, and Ser-370, that conform to the minimal consensus sequence for p34 cdc2 phosphorylation (39). Synthetic peptides corresponding to portions of CKII ␣ containing each of the putative phosphorylation sites were synthesized and phosphorylated in vitro with purified p34 cdc2 (see Table I).
It is interesting to note that replacement of Ser-362 with alanine (compare peptide 3 with peptide 2) abolishes serine phosphorylation, suggesting that p34 cdc2 does not likely phosphorylate Ser-356 or Ser-357 within this peptide. In a similar vein, peptide 1 is exclusively threonine phosphorylated, demonstrating that p34 cdc2 does not likely phosphorylate Ser-343, Ser-348, or Ser-349 within the peptide.
To identify the p34 cdc2 phosphorylation sites on CKII ␣, a mutagenesis strategy was employed. We prepared fusion proteins in which one or more of the putative p34 cdc2 phosphorylation sites had been mutated to non-phosphorylatable alanine residues (see Table II for summary). Each of these fusion proteins was tested as an in vitro substrate for p34 cdc2 and analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography (see Fig. 3). Each of the fusion proteins displays at least one phosphorylated band with noticeably reduced electrophoretic mobility, except for the fusion protein in which all four putative phosphorylation sites had been mutated (Fig.  3K). This fusion protein yields a single phosphorylated band of unaltered electrophoretic mobility. In addition to differences in the locations or shift in electrophoretic mobility of each protein, differences in the extent of phosphorylation were observed for each fusion protein. Interestingly, fusion proteins containing both residues Thr-344 and Ser-370 (lanes A, B, C and I; TTSS, TTAS, TASS, and TAAS, respectively) exhibited the most significant shifts in electrophoretic mobility and also produced some of the most intense bands. Taken together, these results suggest that Thr-344 and Ser-370 are important for the optimal phosphorylation of the C-terminal domain of CKII ␣ by p34 cdc2 . Phosphorylation of the wild-type GST-␣C126 fusion protein achieved an approximate stoichiometry of 3 mol of phosphate/ mol of protein (Fig. 4), while the AAAA mutant was phosphorylated to a stoichiometry of approximately 0.3 mol of phosphate/mol of protein.
Alterations in the intensity of phosphorylation and in the extent of the electrophoretic mobility shift of individual GST fusion proteins (GST-␣C126 and corresponding mutants) suggested that some of the p34 cdc2 sites had been eliminated by mutation. To directly examine the phosphorylation pattern of the individual mutants, two-dimensional phosphopeptide mapping procedures were utilized (Fig. 5). In all cases, the most highly phosphorylated form (i.e. the uppermost phosphorylated band observed in Fig. 3) was subjected to analysis. A number of observations are apparent from examination of these phosphopeptide maps. As evidenced by examination of maps D, F, G, J, K, and L (ATSS, ATSA, AASS, AASA, AAAA, and AAAS, respectively), mutation of Thr-344 to alanine results in loss of phosphopeptide a, indicating that this phosphopeptide contains Thr-344. Phosphoamino acid analysis demonstrated that phosphopeptide a contains exclusively phosphothreonine (data not shown) supporting this interpretation. Phosphopeptide b is absent on all phosphopeptide maps obtained from fusion proteins that are mutated at both Ser-362 and Ser-370 (panels H and K; TTAA and AAAA, respectively), suggesting that phosphopeptide b is derived from similar, if not identical, peptides that are phosphorylated at either Ser-362 or Ser-370. Phosphoamino acid analysis of phosphopeptide b shows that this peptide is composed exclusively of phosphoserine (data not shown).
The region of CKII ␣ that contains the putative p34 cdc2 phosphorylation sites does not contain any charged amino acids GST-␣C126 is composed of GST fused to the 126 C-terminal amino acids of the ␣ subunit of CKII (indicated by solid black rectangle). GST-␣ЈC51 encodes GST and the 51 C-terminal residues of CKII ␣Ј (indicated by striped rectangle). B, equal amounts (4 g) of purified GST, GST-␣C126, and GST-␣ЈC51 proteins were incubated with purified p34 cdc2 (lanes 1, 3, and 6) and without purified p34 cdc2 (lanes 2, 4, and 5, respectively) for 90 min using the in vitro phosphorylation conditions described under "Experimental Procedures." The proteins were recovered from the phosphorylation reactions using glutathione-agarose before analysis on a 12% SDS-polyacrylamide gel and were visualized by autoradiography.

FIG. 2.
Comparative phosphopeptide maps of GST-␣C126 and of the ␣ subunit of CKII phosphorylated in vitro by purified p34 cdc2 . GST-␣C126 and purified bovine CKII were phosphorylated in vitro using purified p34 cdc2 and were then immunoprecipitated from kinase reactions using anti-␣ 376 -391 antiserum. Immunoprecipitates were analyzed by autoradiography after separation of proteins on a 12% SDS-polyacrylamide gel. The phosphorylated fusion protein and the ␣ subunit of CKII were recovered from homogenized gel slices excised from the SDS-polyacrylamide gel. The samples were exhaustively digested with thermolysin and then separated by electrophoresis at pH 1.9 (horizontal dimension with anode to the left), followed by ascending chromatography as described under "Experimental Procedures." The MIX phosphopeptide map was obtained by mixing aliquots (equal cpm) of each phosphorylated sample (GST-␣C126 and the ␣ subunit of CKII) prior to two-dimensional separation. The positions of the origins are marked by arrows and the letter O. Individual phosphopeptides are identified with letters as in Ref. 21. (28). As a result, monophosphorylated peptides would be nearly neutral at pH 1.9 (36). Phosphopeptides containing more than one phosphate would have sufficient negative charge at pH 1.9 to migrate toward the positive electrode (to the left in Fig. 5). The minimal migration exhibited by phosphopeptides a and b in the electrophoretic dimension is consistent with the presence of only a single phosphate on each of these peptides. Phosphopeptide c, observed only in A (TTSS), exhibits the electrophoretic mobility of a peptide with a significant negative charge. Furthermore, this peptide contains both phosphothreonine and phosphoserine, suggesting that it is indeed a multiply phosphorylated peptide (data not shown). When Ser-362 (map B, TTAS) or Thr-360 (map C, TASS) were mutated to alanine residues, phosphopeptide c was not observed. Instead, phosphopeptide e was observed on map B (TTAS), and phosphopeptide f was observed on map C (TASS). Mixing experiments indicated that both phosphopeptides e and f were less negatively charged than phosphopeptide c (data not shown). These results indicate that phosphopeptides e and f are phosphorylated at more than one site but that they are phosphorylated to a lesser extent than phosphopeptide c. Furthermore, phosphoamino acid analysis indicated that spot f was phosphorylated exclusively on serine, whereas spot e contained a mixture of phosphoserine and phosphothreonine. These results suggest that the negatively charged phosphopeptide c is a triply phosphorylated peptide that had been phosphorylated at Thr-360, Ser-362, and Ser-370. Mutation of Thr-360 to alanine (map C, TASS) results in the disappearance of phosphopeptide c and a concomitant appearance of phosphopeptide f. Phosphopeptide f, which is exclusively serine phosphorylated, has a lesser negative charge than phosphopeptide c but still behaves as a multiply phosphorylated peptide, which is presumably phosphorylated at Ser-362 and Ser-370. Similarly, mutation of Ser-362 to alanine (map B, TTAS) results in the loss of phosphopeptide c with the gain of the less negatively charged phosphopeptide e,  b Residues in bold are the residues that conform to the minimal consensus for p34 cdc2 phosphorylation. c Synthetic peptides were assayed at a concentration of 1 mM using purified p34 cdc2 with an activity of 52 pmol/min/l with 1 mM Ser-209 peptide as substrate.
d Residue underlined in peptide 3 represents a serine to alanine change.

T344A, S370A AASS
A***************A*********** which is most likely phosphorylated at both Thr-360 and Ser-370. The presence of phosphoserine and phosphothreonine in phosphopeptide e supports this conclusion. If phosphopeptides e and f are indeed diphosphorylated peptides, it would naturally follow that phosphopeptide c is a triphosphorylated peptide that is phosphorylated at Thr-360, Ser-362, and Ser-370. The greater negative charge and diminished chromatographic migration of phosphopeptide c in comparison to phosphopeptides e or f is consistent with the presence of an additional phosphate on the former peptide (36).
Phosphopeptide d, which is only observed on maps derived from CKII or GST fusion proteins that are phosphorylated in vitro, is present on all phosphopeptide maps, including the fusion protein in which all putative p34 cdc2 phosphorylation sites, Thr-344, Thr-360, Ser-362, and Ser-370, have been mutated to non-phosphorylatable alanine residues. In fusion proteins with alanine instead of Thr-360 (maps C, G, I, J, K, L), it was noticed that spot d (denoted as dЈ on maps) only contained phosphoserine (data not shown). By comparison, when threonine was present at residue 360, phosphoserine and phosphothreonine were detected (data not shown) at spot d (designated as spot d on the maps). This result suggests that spot d, which is not observed following the phosphorylation of CKII in cells (21), was a mixture of comigrating or identical mono-phosphorylated peptides arising either from phosphorylation of Thr-360 or phosphorylation of an unidentified serine. The result that the four-site mutant (i.e. AAAA) was phosphorylated by p34 cdc2 was somewhat unexpected since the most likely p34 cdc2 phosphorylation sites had been eliminated through mutagenesis. It would therefore appear that the additional phosphorylation site does not conform to the minimal consensus for p34 cdc2 phosphorylation since the C-terminal region of CKII ␣ does not contain any additional serine residues that are followed by a proline with the exception of Ser-295. Elimination of this site by expressing a GST fusion protein encoding residues 300 -391 of CKII ␣ did not abolish the ability of p34 cdc2 to phosphorylate the fusion protein (data not shown), suggesting that Ser-295 is not the unknown phosphorylation site. The phosphorylation of non-consensus residues by p34 cdc2 has been previously noted (39,40). It is important to emphasize that phosphopeptide d is not observed on phosphopeptide maps derived from CKII ␣ that had been isolated from 32 P-labeled human or chicken cells that had been arrested in mitosis.
Overall, the results obtained by phosphorylation of mutant fusion proteins and resultant phosphopeptide maps support the conclusion that the preferred sites of phosphorylation by p34 cdc2 in cells are Thr-344, Thr-360, Ser-362, and Ser-370 on the CKII ␣ subunit (summarized in Table III). In fact, mutation of each of these residues to alanine results in elimination of all phosphopeptides that are detected following the phosphoryla-  Fig. 3) were phosphorylated in vitro using purified p34 cdc2 followed by immunoprecipitation using anti-␣ 376 -391 antibodies as described under "Experimental Procedures." Following separation by SDS-polyacrylamide gel electrophoresis and visualization by autoradiography, the most heavily phosphorylated form (i.e. the band with the least electrophoretic mobility as in Fig. 3) of each fusion protein was recovered from the SDS-polyacrylamide gel, oxidized with performic acid, and exhaustively digested with thermolysin. Thermolytic peptides were then separated in two dimensions as described in the legend to Fig. 2. Phosphopeptide maps were visualized by autoradiography. Capital letters over each map designate the identity of the amino acid residues present at positions 344, 360, 362, and 370, respectively (see Table II). Individual phosphopeptides are indicated by small letters on each map as in Fig. 3. tion of CKII ␣ in mitotic cells. Since p34 cdc2 also phosphorylates Ser-209 on CKII ␤ in mitotic cells, the present results indicate that p34 cdc2 could phosphorylate the CKII holoenzyme to a stoichiometry of up to 10 mol of phosphate/mol of ␣ 2 ␤ 2 tetramer in mitotic cells. The high stoichiometry of phosphorylation suggests that phosphorylation could regulate functional properties of CKII and that it could in some way participate in the burst of phosphorylation that accompanies the activation of p34 cdc2 at the G 2 -M transition (1)(2)(3)(4)(41)(42)(43).
At present, the role of phosphorylation in regulating the functions of CKII in mitotic cells remains speculative. There have been a number of studies in amphibian (44 -46) or starfish oocytes (47) and in mammalian cells (48) that have suggested that the activity of CKII is regulated at different stages in the cell cycle. However, a direct link between these changes in the catalytic activity of CKII and the phosphorylation state of CKII has not been demonstrated. In vitro studies by Mulner-Lorillon et al. (19) demonstrated that CKII isolated from Xenopus laevis can be phosphorylated and activated by p34 cdc2 . By comparison, when we examined the activity of CKII that had been isolated in its fully phosphorylated state by immunoprecipitation from mitotic cells, the catalytic properties of the enzyme were not significantly different from the unphosphorylated enzyme (21). The latter results suggest that phosphorylation of CKII in mammalian or avian cells may not have direct effects on the enzymatic properties of CKII. Despite the lack of evidence that directly links the phosphorylation of CKII to changes in its catalytic activity during mitosis, there are suggestions that the functions of CKII could be altered during mitosis. For example, independent studies using immunofluorescence have demonstrated that CKII is associated with the mitotic spindle in dividing cells (49,50). The factors that control the interaction of CKII with the mitotic spindle remain uncharacterized. However, it should be noted that alterations in the ability of CKII to phosphorylate substrate proteins could be mediated by affects on its intracellular distribution or on its interaction with specific substrates without obvious effects on its enzymatic activity. In this regard, there are indications from the studies of Cardenas et al. (51) that the phosphorylation of topoisomerase II is increased in mitosis and that the major mitotic phosphorylation sites are CKII sites. At the restrictive temperature in yeast harboring a temperature-sensitive form of CKII, topoisomerase II is hypophosphorylated, and the cells fail to divide. In mammalian cells, the phosphorylation of topoisomerase II is also elevated in mitosis (52). With the exception of topoisomerase II, very little is known regarding the stage in the cell cycle when CKII phosphorylates its substrates. It will certainly be of interest to examine the phosphorylation of other CKII substrates to determine if any of these proteins are phosphorylated at the G 2 -M transition. It is also noteworthy that the mitotic phosphorylation sites that have been identified on the ␣ subunit of human CKII are highly conserved between mammalian species and are even present in the ␣ subunit of chicken CKII (28,53). Interestingly, none of these sites are present of the ␣Ј-subunit of CKII from either species. The latter observation suggests that the regulation, and perhaps other functional properties, of the ␣ and ␣Ј-isozymic forms of CKII could be distinct. Identification of the mitotic phosphorylation sites on the ␣ subunit of CKII will undoubtedly facilitate efforts to define the role of CKII and its phosphorylation during cell division.