Multiple Signal Transduction Pathways Induce Phosphorylation of Serines 16,25, and 38 of Oncoprotein 18 in T Lymphocytes*

A multitude of external signals induce extensive phos- phorylation of Oncoprotein 18 (OplS), which suggesta a putative role for this protein in signal transduction. We have recently identified two distinct proline-directed kinase families that phosphorylates Opl8 with overlap- ping but distinct site preference. These two kinase families, mitogen-activated protein ( M A P ) kinases and cyclin-dependent cdc2 kinases, are involved in receptor-and cell cycle-regulated phosphorylation events, respec- tively. In the present study, site-specific phosphorylation of Op18 in response to stimulation of the antigen receptor-associated CD3 complex was analyzed in the Jurkat T cell-line. The results show that CD3-induced phosphorylation of Ser-25 of Op18, which is the primary MAP kinase phosphorylation site, can be induced by an apparently protein kinase C (PKC)-independent signal transduction pathway. We also demonstrate that Ser-16 of OplS is specifically phosphorylated in response to the Caz+ signal generated by CD3 stimulation or by the Caz+ ionophore ionomycin. Ser-16 phosphorylation occurs independently of both PKC and MAP kinase activation. Using site-specific OplS mutants and tryptic phosphopeptide mapping, we show that phosphorylation of Ser-16 of Op18 together with Ser-25, or Ser-25 and Ser-38, generates two OplS phosphoisomers showing a dra- matic electrophoretic retardation. In conclusion, site-mapping studies of Opl8 reveal that CD3 stimulation results in an apparently PKC-independent activation of both the MAP kinase and a Caz+-regulated kinase path- way, which results in phosphorylation of distinct sites of OplS. The data also pinpoints the specific phosphoryla- tion events that result in electrophoretic retardation of Op18.


I "
A multitude of external signals induce extensive phosphorylation of Oncoprotein 18 (OplS), which suggesta a putative role for this protein in signal transduction. We have recently identified two distinct proline-directed kinase families that phosphorylates Opl8 with overlapping but distinct site preference. These two kinase families, mitogen-activated protein ( M A P ) kinases and cyclin-dependent cdc2 kinases, are involved in receptorand cell cycle-regulated phosphorylation events, respectively. In the present study, site-specific phosphorylation of Op18 in response to stimulation of the antigen receptor-associated CD3 complex was analyzed in the Jurkat T cell-line. The results show that CD3-induced phosphorylation of Ser-25 of Op18, which is the primary MAP kinase phosphorylation site, can be induced by an apparently protein kinase C (PKC)-independent signal transduction pathway. We also demonstrate that Ser-16 of OplS is specifically phosphorylated in response to the Caz+ signal generated by CD3 stimulation or by the Caz+ ionophore ionomycin. Ser-16 phosphorylation occurs independently of both PKC and MAP kinase activation. Using site-specific OplS mutants and tryptic phosphopeptide mapping, we show that phosphorylation of Ser-16 of Op18 together with Ser-25, or Ser-25 and Ser-38, generates two OplS phosphoisomers showing a dramatic electrophoretic retardation. In conclusion, sitemapping studies of Opl8 reveal that CD3 stimulation results in an apparently PKC-independent activation of both the MAP kinase and a Caz+-regulated kinase pathway, which results in phosphorylation of distinct sites of OplS. The data also pinpoints the specific phosphorylation events that result in electrophoretic retardation of Op18.
Oncoprotein 18 ( 0~1 8 ) ' was identified by its up-regulated expression in acute leukemia cells (1) and extensive phosphorylation in response to stimulation of various receptor systems (2-4). This protein, previously termed 19K by us, has been purified, and we have isolated a cDNA copy of the gene (5,6). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 8 These authors have contributed equally to this work.
Sequence comparison of the 19K cDNA revealed that this protein had previously been identified in other cellular systems under different names, such as p19, Stathmin, and Op 18 (7-10). The specific function of this protein is still unknown, but since a dramatic up-regulation is observed in many neoplasms (1,6, lo), we have adopted the designation Op18.
Stimulation of T cells through the antigen receptor elicits a cascade of phosphorylation events that ultimately results in cell activation (for review, see Ref. 11). An early consequence of antigen receptor stimulation is tyrosine phosphorylation of phospholipase C y l that results in activation of the enzyme and hydrolysis of inositol phospholipids (12)(13)(14). This in turn leads to elevation of intracellular calcium and activation of protein kinase C (PKC). The calcium signal seems important for both the regulation of PKC as well as activation of other protein kinase systems, such as Ca2+/calmodulin-dependent protein kinases (15,16).
The mitogen-activated protein ( M A P ) kinases, which have also been termed extracellular signal-regulated kinases (17), are thought to be involved in growth promotion after triggering of growth factor receptors (for review, see Ref. 18). The MAP kinase pathway is regulated by extracellular signals that trigger a kinase cascade system (19)(20)(21). This kinase cascade system appears to be regulated by tyrosine kinases that act upstream of the ~21"" and ~7 4~f -I protooncogenes (22)(23)(24)(25). Studies on T cells have shown that stimulation of the T cell antigen receptor, as well as direct activation of PKC by phorbol ester treatment, activates the p21"" protooncogene product and the MAP kinase (26,271. Moreover, T cell antigen receptor stimulation may activate p21" through at least two distinct regulatory pathways. Thus, while PKC activation is clearly sufficient to activate p21ms in T cells, it has also been reported that a PKC"5ndependent" route may regulate the activity of this regulatory protein in T cells (28,29).
Previous studies, employing two-dimensional gel electrophoresis, have suggested a very complex pattern of Op18 phosphorylation (30)(31)(32). In attempts to identify the protein kinase systems involved, we have expressed specific Op18 cDNA mutant constructs and performed phosphopeptide-mapping analysis of Op18 (33). These experiments have identified two distinct proline-directed kinase families that phosphorylate Op18 with overlapping but distinct site preference. These two kinase families, MAP kinases and cyclin-dependent cdc2 kinases, are involved in receptor-and cell cycle-regulated phosphorylation events, respectively. The results suggested that the MAP kinase has a 20-fold preference for Ser-25 as opposed to Ser-38 of Op18, while cdc2 kinases have a 5-fold preference for the Ser-38 residue.
In the present study, we report on data concerning the activation-induced phosphorylation of Op18 in the Jurkat T cell line in the presence or absence of the PKC inhibitor Ro 31-7549.
The results provide evidence that multiple signal transduction 25671 T Cells pathways are involved in phosphorylation of distinct sites of 0~1 8 .
First, two distinct pathways appear to transduce signals from the T cell antigen receptor that results in phosphorylation of Op18 on Ser-25, one of which is PKC-mediated and one of which is not. Second, we observed a specific calcium-induced phosphorylation event that results in phosphorylation ofser-16 of Op18. This phosphorylation event was found to be independent of PKC and MAP kinase activation. Finally, it is also shown that phosphorylation of Ser-16, in specific combinations with Ser-25 and Ser-38, results in Op18 phosphoisomers with distinct and characteristic electrophoretic migration.
EXPERIMENTAL PROCEDURES Materials and Enzymes-MAP kinase (pp44 mpk) (catalog no. 14-102) was purchased from Upstate Biotechnology, Inc. According to the supplier's specifications, this kinase preparation was purified to greater than 95% homogeneity from the sea star Pisaster ochmceus. The catalytic subunit of CAMP-dependent protein kinase (catalog no. P 2645) was purchased from Sigma. This protein kinase preparation was, according to the supplier's specifications, purified from bovine heart to a specific activity of 50 d p g protein. Alkaline phosphatase was from Boehringer Mannheim and was used as recommended. The Ro 31-7549 PKC inhibitor is designed by Roche Products Ltd., and its specificity has been described previously (34,35). The Ro 31-7549 preparation used in the present study (D7502) was synthesized b y h t r a Draco AB, Sweden, and kindly provided by Prof. H. Bergstrand. Herbimycin A was from Life Technologies, Inc. Protein A-Sepharose, CM-Sepharose CL-GB, and molecular weight calibration kit (17-0446-01) were from Pharmacia LKB Biotechnology Inc. The anti-CD3 monoclonal antibody UCHT-1 was used at 8 pglml (33).
The mutagenized 181-base pair EcoRI fragments were subcloned into pBluescript SK+ (Strategene), to regenerate the entire coding sequence of Op18 in its native configuration. The introduction of the mutated bases and the junctions of ligated fragments were confirmed by nucleotide sequence analysis using Pharmacia's DNA sequencing kit and custom-designed oligonucleotides. Cell Cultures and Electrophoration-The leukemic T cell line Jurkat was maintained in RPMI 1640 supplemented with 5% fetal calf serum. A double thymidine block was used for synchronization of Jurkat cells at the G,/S phase of the cell cycle. Cells were incubated for 5 h with 1 rn thymidine, washed, and recultured for 17 h. After this period, about half of the cell population had progressed to the G&I phase of the cell cycle, and the other half had completed the cell cycle. At this time point, cells were subjected to a second 6-h thymidine block and labeled with 32Pi during the last 4 h. Analysis of DNA content in the resulting cell population showed that more then 80% of all cells were contained within the GI peak of the DNA histogram. The murine lymphoma T cell line EL-4 (lo7 cells in 300 pl) were subjected to electrophoration at 960 microfarads and 370 V with a Gene Pulser apparatus (Bio-Rad) in the presence of 20 of plasmid DNA as previously described (33).
32Pi Labeling of Cells, Immunoprecipitation, and Western Blotting of Human Opl8"Jurkat cells were washed and recultured for 4 h in phosphate-free RPMI 1640 (106/ml, 1 ml) containing 32Pi (100 pCi/ml). Thereafter, cells were stimulated as indicated in the text, solubilized in Triton X-100-containing lysis buffer, and Op18 immunoprecipitated using affinity-purified rabbit antibodies raised against Escherichia eoliproduced Opl8 protein as described in detail elsewhere (6). Western blot analysis of Op18 expressed in Jurkat cells was performed as previously described (5), using affinity-purified anti-Opl8. Western blot analysis of transfected murine EL-4 cells was performed by probing nitrocellulose filters with anti-Op18:2-33 (6) This anti-Opl8 preparation allows specific detection of human Op18 without interference of the murine Op18 analogue (33). Bound antibodies were revealed by incubation with horseradish peroxidase-labeled goat anti-rabbit antibody for 1 h, and blots were developed using the ECL detection system (Amersham Corp. p~ [y32PIATP (specific activity 2500 cpdpmol). The reactions were terminated after 2 h of incubation at 30 "C by addition of SDS-PAGE sample b e e r and were subsequently analyzed by 1620% SDS-PAGE.
The stoichiometry values of MAP kinase-or cAPK-phosphorylated Opl8 were routinely -1.2 or -1.5 moYmol Op18, respectively. Quantification of Op18 phosphorylation was obtained by excision of the relevant 3zPlabeled band of the gel, followed by liquid scintillation counting. Sequence Analysis of 32P-Labeled, in Vitro Phosphorylated Opl8 "Recombinant Op18 was phosphorylated in the presence of [yS2PIATP and cAPK, as described above, diluted 4-fold in 50 m~ sodium acetate (pH 4.7), and thereafter absorbed onto CM-Sepharose CL6B. After elution in the presence of 1 M NaCl, Op18 was desalted on a Sephadex G-50 column. In order to analyze phosphorylation sites, Opl8 was covalently attached (30 pmol, 240,000 cpm) to a Sequelon arylamine polyvinylidene difluoride membrane from Millipore. Subsequent N-terminal sequence analysis was performed using an Applied Biosystems 477N 120A pulsed liquid phase sequenator. Instead of butyl chloride, the anilinothiazolinone amino acids and any inorganic phosphate were extracted from the arylamine disc with 1% trifluoroacetic acid in methanol. About 33% of the sample were used for analysis of derivatized amino acids and 66% for analysis of 32P radioactivity.
Zkyptic P2PlPhosphopeptide Analysis-Phosphorylated proteins were separated by SDS-PAGE and electrotransferred to nitrocellulose filters. Radioactive bands were localized by autoradiography and excised. Proteins immobilized on nitrocellulose was digested with tosylphenylalanyl chloromethyl ketone-treated trypsin (Worthington, 4 x 7.5 pg of trypsin was added at 2-h intervals) and processed as described by Lou et al. (38). Phosphopeptide mapping was performed by two-dimensional separation on cellulose thin layer plates as described by Boyle et al. (39). The first dimension was electrophoresis in ammonium carbonate (pH 8.9, 10 g/liter of deionized water) at 1000 V for 60 min at 4 "C. The second dimension was chromatography for 9 h in a mixture of 750 ml of 1-butanol, 500 ml of pyridine, 150 ml of glacial acetic acid, and 600 ml of deionized water. Phosphorimager analysis of radioactive spots was used for quantification of phosphopeptides.

Evidence That Multiple Protein Kinase Pathways Are Involved in Opl8
Phosphorylation-Jurkat is a T cell line that has been extensively used in studies of T cell antigen receptor signaling. In the present study, we used G1/S-synchronized Jurkat cells, since this cell population shows a low level of "basal" phosphorylation of Op18 compared to proliferating cells (data not shown). Activation of synchronized Jurkat cells with phor-bo1 ester results in a 10-20-fold increased phosphorylation of Ser-25, and a previous study has suggested that this phos-  shows that ionomycin alone only stimulates a small, but still significant, increase in phosphorylation of the 19-kDa form of Op18, while ionomycin together with PDBu induces the appearance of both the 21-and 23-kDa phosphoisomers of Op18. Thus, it seems that generation of these phosphoisomers is dependent on a PKC signal in combination with a Ca2+-induced protein kinase pathway.
To analyze the role of PKC activation during anti-CD3-induced phosphorylation of Op18, we employed the PKC inhibitor Ro 31-7549. As could be expected, this inhibitor completely blocked PDBu-induced increases in Op18 phosphorylation (Fig.

lA, middle panel). It is also clear that Ro 31-7549 decreased
Op18 phosphorylation in the presence of PDBu and ionomycin to the level observed in the presence of ionomycin alone. However, anti-CD3-induced phosphorylation of the 19-, 21-, and 23-kDa phosphoisomers was essentially unaltered. These results suggest that the complex pattern of anti-CD3-induced phosphorylation of Op18 occurs independently of PKC activation.
To assess the involvement of protein tyrosine kinases in the regulation of Op18 phosphorylation, we employed the tyrosine kinase inhibitor herbimycin A (Fig. lA, lower panel ). As could be expected, the result demonstrates that anti-CD3-induced Opl8 phosphorylation is markedly inhibited, while the phosphorylation observed in the presence of PDBu alone, or PDBu and ionomycin, was essentially unaltered.
Our previous study suggests that PDBu and anti-CD3 induce the MAP kinase in T cells to phosphorylate primarily Ser-25 of Opl8 (33). Anti-CD3-dependent phosphorylation of Opl8 in the presence of the PKC inhibitor Ro 31-7549 suggest, therefore, that a PKC-independent pathway of MAP kinase activation also exists in T cells. To further characterize the specificity of the Ro 31-7549 compound, we determined the dose response of inhibition after stimulation with anti-CD3, PDBu, or PDBu together with ionomycin (Fig. 2). Accordingly, '"P-labeled Jurkat cells were preincubated with Ro 31-7549 for 10 min, stimulated as indicated for 9 min, and Op18 phosphorylation evaluated by immunoprecipitation. The data presented in Fig. 2 clearly shows that anti-CD3-induced Op18 phosphorylation is only slightly inhibited by high doses of Ro 31-7549, while the PDBu-induced response is completely blocked even a t low doses of the inhibitor. Op18 phosphorylation in the presence of both PDBu and ionomycin is partially blocked at low doses of Ro 31-7549, and the response is completely blocked a t high doses. Taken altogether, the results shown in Figs. 1 and 2 suggest that triggering of the CD3 complex induce a PKC-independent pathway of Op18 phosphorylation and, furthermore, that the associated Ca2+ response induces a specific phosphorylation event(s) that, in combination with Ser-25 phosphorylation, results in electrophoretic retardation of Op18.
Phosphorylation ofSer-16, Ser-25, a n d Ser-38 Is Required for Electrophoretic Retardation of Opl8-Wild type and mutated human Opl8 cDNAs were transiently expressed in the murine T cell line EL-4. Antibodies specific for human Opl8 were used in Western blot analyses to distinguish Opl8 expressed from the transfected cDNA from endogenous murine Op18. Fig. 3A

(upper panel) shows that a 19-kDa form of human Opl8 is expressed in non-stimulated cells transfected with wild type
Opl8 cDNA ( w t ) , but not in cells transfected with a rat CD2 cDNA (CD2). Moreover, Fig. 3A (lower panel) demonstrates that stimulation ofwild type Opl8-transfected cells with PDBu and ionomycin generates the additional 21-and 23-kDa forms of Opl8, which were observed after the same treatment of Jurkat cells (Fig. 1 Fig. 1 ( A and B) demonstrate that the 21-and 23-kDa forms of Opl8 are generated by phosphorylation. Therefore, expression of specific Opl8 mutants indicate that diphosphorylation of Ser-16 and Ser-25 is required to generate of the 21-kDa form of Opl8, and that triphosphorylation of Opl8 on Ser-16, Ser-25, and Ser-38 is required to generate the 23-kDa form of the protein.
To establish an in vitro phosphorylation system that generates electrophoretic retardation of Op18, we employed cAPK in combination with the MAP kinase. The autoradiograph presented in Fig. 38 shows that wild type recombinant Op18 phosphorylated using [y"P]ATP and either MAP kinase or cAPK migrates at 19 kDa (lanes a and b ) . However, simultaneous phosphylation of Opl8 by both these kinases results in a different migration pattern. As shown in Fig. 3B, prephosphorylation of Opl8 in the presence of unlabeled ATP and cAPK, followed by heat inactivation of cAPK and subsequent phosphorylation using the MAP kinase and [y"PIATP, results in Opl8 phosphoisomers migrating a t 19.21. and, as a faint band, 23 kDa (lane c). Previous studies have revealed that the MAP kinase primarily phosphorylates O p l 8 at Ser-25 (95%) and has Ser-38 a s a minor target (5%) (33). In agreement with these studies, it is shown in Moreover, the data also demonstrate that cAPK used in vitro can provide the additional phosphorylation event4 s) that is required to cause electrophoretic retardation of Op18. However, for reasons that will be discussed below, we do not find it likely that cAPK in T cells is involved in CD3-induced generation of Op18 was phosphorylated by cAPK in the presence of [y32P]ATP and subjected to N-terminal sequence analysis. A Burst" of 32P label was detected at a position corresponding to Ser-16 of Opl8, which demonstrates cAPK-dependent phosphorylation of this amino acid (Fig. 4 . 4 ) . The second phosphorylation target of cAPK, however, is as yet unidentified.
Earlier phosphopeptide-mapping studies, employing wild type and site-specific mutants of the Opl8 protein. have identified tryptic phosphopeptides corresponding to phosphorylation of Ser-25 and Ser-38 (33). These studies demonstrated that MAP kinase primarily phosphorylates OplS at Ser-25 and has Ser-38 as a minor target. The tryptic phosphopeptide map of the Op18 Ser-38 + Ala mutant protein phosphorylated using the MAP kinase is shown in Fig. 50. The result demonstrates the location of the Ser-25-phosphorylated peptides S25 and S25' (a prime indicates a partial cleavage product). As shown above (Fig. 3). prephosphorylation of the Op18 Ser-38 -. Ala  followed by heat inactivation of cAPK and subsequent phosphorylation using the MAP kinase and [y3'P1ATP, generates 19-and 21-kDa Op18 phosphoisomers. This in vitro phosphorylation protocol results in Op18 proteins that are 32P-labeled solely at Ser-25. As could be predicted, tryptic phosphopeptide maps of the 19-kDa phosphoisomer contain 32P-labeled S25 and S25' peptides (Fig. 5E). The two faint spots located below the S25 and S25' peptides are due to contamination with the 21-kDa phosphoisomer. However, the 21-kDa phosphoisomer does not contain 32P-labeled S25 and S25' peptides; instead, all the 32P label introduced by the MAP kinase on Ser-25 is present on two peptides designated b and a (Fig. 5F). As depicted in Fig.  4 B , trypsin does not cleave efficiently at Arg in the sequence kg-Ala-SedP) (39), and it follows that phosphorylation of Ser-16 can be predicted to interfere with trypsin cleavage between amino acids 14 and 15. Thus, Ser-16 phosphorylation will result in a tryptic peptide encompassing amino acid 14-27, while the S25 peptide encompasses amino acid 15-27. The peptide-spanning amino acids 14-27, phosphorylated at both Ser-16 and Ser-25, would have a predicted mobility consistent with peptide a, observed in the 21-kDa phosphoisomer of Op18. Since both peptides a and b contain 32P-labeled Ser-25, peptide b is likely to represent a partial cleavage product of peptide a. As shown in Fig. 4, the N-terminal Op18 sequence contains several cleavage sites that are inefficiently cleaved by trypsin.

A S G Q A F E L I L S P R
To evaluate the predictions made above, we performed tryptic mapping of Op18 phosphorylated by cAPK. The result, shown in Fig. 5.4, reveals one major and two minor phosphopeptides. As will be argued below, the major peptide termed X corresponds to the unidentified C-terminal cAPK phosphorylation site, while the peptides termed Z and Y are the complete and partial cleavage products containing phosphorylated Ser-16. We also analyzed Op18 phosphoisomers that were generated as follows; Op18 was stoichiometrically prephosphorylated on Ser-25 in the presence of unlabeled ATP and MAP kinase, followed by heat inactivation of MAP kinase and subsequent phosphorylation using cAPK and [y3'P1ATP. Peptide mapping of the resulting 19-and 21-kDa phosphoisomers is shown in Fig. 5 ( B and C, respectively), and it should be noted that only cAPK-specific sites are 32P-labeled. The 19-kDa Op18 phosphoisomer generated in this manner, only exhibits peptide X, while the 21-kDa Op18 phosphoisomer exhibits peptides a and b as well as peptide X. Mixing experiments confirmed co-migration of phosphopeptides a and b shown in Fig. 5 ( C and F) (data not shown).
The predicted mobility of the peptides generated by phosphorylation of Ser-16 is consistent with the migration of peptide Z, and it seems likely that peptide Y represents a partial cleavage product of peptide Z. Moreover, calculations of predicted mobility are also consistent with the proposal that peptides Z and Y are monophosphorylated versions of the diphosphorylated peptides a and b, which, as shown in Fig. 5F, both contain phosphorylated Ser-25. Since one of the in vitro phosphorylation targets of cAPK was shown to be Ser-16, it can be predicted from the tryptic cleavage pattern that peptides a and b also harbor phosphorylated Ser-16 (Fig. 4). Hence, our results suggest that in vitro phosphorylation of Ser-16 by cAPK generates peptides Z and Y and, furthermore, that peptides a and b are diphosphorylated on Ser-16 and Ser-25. It may appear surprising that peptides Z and Y do not co-migrate with 525 and S25'. However, as argued above, phosphorylation of Ser-16 can be predicted to interfere with trypsin cleavage between amino acids 14 and 15 (39). This would result in a Ser-25containing peptide, encompassing amino acids 1P27, that harbors the positively charged Arg-15 residue and a predicted migration profile consistent with peptide Z.
Phosphopeptides a and b, in contrast to the X peptides, are only visible in the Op18 phosphoisomer migrating at 21 kDa. This observation suggests a causal relationship between diphosphorylation of Op18 on Ser-16 and Ser-25 and electrophoretic retardation of the 21-kDa phosphoisomer. The result in Fig. 3 8 demonstrated that generation of the 23-kDa phosphoisomer required phosphorylation on both Ser-25 and Ser-38. Hence, it seems likely that triphosphorylation of Op18 on Ser-16, 25, and 38 results in generation of the 23-kDa phosphoisomer. This interpretation agrees with the data in Fig. 3A, which shows that all 3 of these serine residues are required to generate the 23-kDa Op18 phosphoisomer in intact cells. Finally, phosphopeptide mapping of in vivo phosphorylated Op18, which will be presented below, shows that peptide X is not involved in the generation of Op18 phosphoisomers with retarded electrophoretic mobility.
A Ca2+ Signal Induces Phosphorylation of Ser-16 of Opl8 Zndependently of MAP Kinase Activation-To study Op18 phosphorylation in response to cellular signals, we performed twodimensional phosphopeptide mapping of the 19-, 21-, and 23-kDa phosphoisomer of Op18 prepared from 32P-labeled Jurkat cells. Tryptic phosphopeptides from the 19-kDa phosphoisomers of Op18 from unstimulated cells (Co), and cells stimulated with either ionomycin alone (iono) o r PDBu together with ionomycin (PDBu liono) are shown in Fig. 6 ( A -C ) . The results in Fig. 6A show that the main phosphopeptide in unstimulated cells corresponds to phosphorylated Ser-38, which has been identified and termed 538 in a previous study (S38' indicates the main partial cleavage product; Ref. 33). It is shown in Fig.  6B that stimulation with ionomycin alone results in a selective increase in the intensity of phosphopeptides Y and Z, which has been shown above to correspond to phosphorylation of Ser-16 (co-migration of phosphopeptides Y and Z generated by in vivo T Cells  Fig. 1, were used as a source for in uiuo phosphorylated Opl8 phosphoisomers. The 19-, 21-, and 23-kDa phosphoisomers were prepared from cells that had been subject to the treatment indicated in panels A-F, and were subsequently digested with trypsin as described under 'Experimental Procedures." Panel MXX2I-kDal23-kDa shows a phosphopeptide map obtained from mixing equal cpm of material derived from the 21-and 23-kDa phosphoisomer from cells stimulated with PDBu and ionomycin. Phosphopeptides were resolved by two-dimensional separation on thin layer cellulose plates and the designations of phosphopeptides are as described in Fig.   4B and 5. and in vitro phosphorylation was confirmed by mixing experiments; data not shown). Moreover, stimulation of cells with both PDBu and ionomycin results in a dramatic and selective increase in Ser-25 phosphorylation (Fig. 6C). Most interestingly, the Z and Y phosphopeptides, which are strongly labeled after stimulation of Jurkat cells with ionomycin alone (Fig. 6B 1, are not clearly visible in the 19-kDa phosphoisomer after stimulation of cells with ionomycin together with PDBu (Fig.  6 0 . -tic phosphopeptides of the 21-and 23-kDa phosphoisomers of Opl8 were also analyzed. The data shows that both the 21-kDa (Fig. 6D) and 23-kDa (Fig. 6E) phosphoisoners contain the two major a and b phosphopeptides. By mixing samples from the 21-and 23-kDa forms of Op18, comigration of phosphopeptides a and b was confirmed (Fig. 6F). Moreover, comi-gration of phosphopeptides a and b generated by in vivo and in uitro phosphorylation was also confirmed by mixing experimenta (data not shown). The result in Fig. 3 demonstrated that phosphorylation of both Ser-25 and Ser-38 is required to generate the 23-kDa phosphoisomer of Opl8. B a d on these data, it can be predicted that the S38 and S38' phosphopeptides should be present in the 23-kDa form of Opl8 shown in Fig. 6E. This is the case, but the position of the S38 peptide in thin experiment partially overlaps with that of peptide b, and they appear as a single diffise spot on the autoradiograms. However, the main partial cleavage product of the Ser-38cantaining peptide, S38', is present as a distinct spot in tryptic digests of the 23-kDa, but not the 21-kDa, phosphoisomer of Op18 (Fig.   6, compare panels D and E).
stimulation of cells with pharmacological activators. -tic phosphopeptide mapping of Op18 was also performed after stimulation of cells with anti-CD3. The result in Fig. 7 demonstrates that stimulation of the CD3 complex generates an identical pattern of Op18 phosphopeptides (note that peptides b and S38, shown in Fig. 7C, are better resolved than the corresponding peptides shown in Fig. 6E). Thus, these results confirm the data in Fig. 1, namely that PDBu-induced PKC activation in combination with a Ca2+ signal mimics anti-CD3 stimulation with respect to phosphorylation of Ser-16, Ser-25, and Ser-38 of Op18.
The result shown in Fig. 6B suggests that ionomycin treatment alone induce phoaphorylation of Ser-16, as indicated by a selective increase in the 32P labeling of peptides Z and Y. To confirm that this Ca2+-induced phosphorylation event causes electrophoretic retardation in combination with Ser-25 phosphorylation introduced in vitro, the following experiment was performed. Op18 was purified from control Jurkat cells (Co) or Jurkat cells stimulated with ionomycin (iono), PDBu (PDBu), or ionomycin and PDBu (iono/PDBu). Western blot analysis of these purified Op18 preparations confirmed that stimulation of cells with ionomycin or PDBu alone generates only minor appearance of the 21-and 23-kDa phosphoisomers (Fig. 8A 1. The same Op18 preparations were also phosphorylated in vitro using the MAP kinase, which primarily phosphorylates Ser-25. The data shown in Fig. 8B reveals that a significant fraction of Op18 prepared from ionomycin-treated cells and, to a lesser extent, ionomycin and PDBu-treated cells, change electrophoretic mobility in response to MAP kinase-dependent phosphorylation. When comparing the relative intensities of the 19-, 21-, and 23-kDa phosphoisomers shown in Fig. 8B with the in uiuo result shown in Fig. 1, it should be kept in mind that phosphorylated Ser-16 is not 32P-labeled in the experiment shown in Fig. 8B, which results in lower specific 32P activity in the 21-and 23-kDa phosphoisomers. The results presented in Fig. 6 and 7 show that a Ca2+ signal alone induce increased phosphorylation of Ser-16 of Opl8 and that subsequent phosphorylation of Ser-25 results in electrophoretic retardation. Phosphopeptide X, which corresponds to the unidentified in vitro target for cAPK (Fig. 5, A X ) , does not show increased phosphorylation in response to ionomycin treatment (Fig. 6B 1. Moreover, phosphopeptide X is present in very low levels in the 21-and 23-kDa phosphoisomers of Opl8 observed in uiuo (Fig. 6, D-F). These results suggest that phosphorylation of Ser-16 in combination with Ser-25 or with Ser-25 and Ser-38 is sufficient to cause electrophoretic retardation of Op18.
The data presented above pinpoint the combinations of phosphorylation events that results in electrophoretic retardation.
As depicted in Fig. 9 our data suggest that the migration of Opl8 on SDSPAGE is diagnostic for certain phosphorylation events.

DISCUSSION
Many previous studies have suggested a putative role of Op18 in signal transduction and growth control. Evidence for such a role of Op18 includes: (i) phosphorylation of Op18 in response to a multitude of external signals, (ii) differentiationspecific regulation of Opl8 expression levels, and (iii) profound up-regulation of Op18 in acute leukemia cells (1-10. 3 M 3 , 43-46). Analysis of Opl8 by two-dimensional PAGE has suggested a very complex pattern of phosphorylation (30)(31)(32). and elucidation of these phosphorylation events may facilitate the search for the function of this phosphoprotein.
Phosphorylation of Op18 has previously been shown in response to cAPK activation in neurons and neuroendocrine cells (47,48). Computer searches identify two putative cAPK sites within the Op18 sequence, namely Ser-16 (Lys-Arg-Ala-Se6P)) and Ser-63 (Arg-Arg-Lye-SedP)). The present study shows that Ser-16 and an unidentified site located on peptide X, possibly Ser-63, are in vitro targeta for cAPK (Figs. 4 and 5). The sequence surrounding Ser-16 does not strictly conform to the optimal consensus recognition sequence of cAPK, which is Lys-T Cells Arg-X-X-Ser(P)/"hr or Arg-Arg-X-Ser(P)/Thr (where X represents any amino acid) (49). Nevertheless, Ser-16 of Op18 still appears to provide a reasonably efficient in vitro target for cAPK, but it is evident from phosphopeptide analysis that cAPK phosphorylates the unidentified site located on peptide X more efficiently than that on Ser-16 (Fig. SA).

It still remains to be established if cAPK phosphorylates
Ser-16 and the unidentified site harbored on peptide X after activation of the kinase in intact cells. A previous study has suggested that stimulation of T cells with agents that increase cAPK activity does not result in increased phosphorylation of Op18 (31). These results are in contrast to previous studies performed on neurons and neuroendocrine cells (47,48). cAPK target located on peptide X are constitutively phosphorylated in Jurkat cells (Fig. 6). Hence, it is at present uncertain which kinase(s) is involved in maintaining basal phosphorylation level of these sites.
In the present study, we show that a Ca2+ signal provided by ionomycin induces a selective increase in Ser-16 phosphorylation. This phosphorylation event appears independent of both PKC and MAP kinase activation, since it is resistant to the RQ 31-7549 PKC inhibitor and is induced by ionomycin alone (Figs.  1,6, and 7). On the basis of the arguments outlined above, we find it highly unlikely that cAPK is involved in Ca2+-induced Ser-16 phosphorylation. It is also clear from our result that a Ca2+ signal results in a selective increase in Ser-16 phosphorylation and the phosphorylation level of the major CAPK in vitro target, located on peptide X, is essentially unaltered (Fig.  6). In addition, treatments of Jurkat cells with PDBu, in the presence of agents that activates cAPK, do not result in generation of the 21-and 23-kDa Op18 phosphoisomers (data not shown), which are characteristic of phosphorylation of Ser-16 in specific combinations with Ser-25 and Ser-38. Thus, although Op18 may be a physiological substrate for cAPK, the result outlined above suggest that this protein kinase is not involved in activation-induced phosphorylation of Op18 in T cells.
Activation of Ca2+/calmodulin-dependent protein kinases in T cells has, to our knowledge, not been studied in detail, but these multisubstrate kinases seem to be ubiquitously expressed (16,50). One obvious possibility is that a Ca 2+ signal generated by ionomycin, or by anti-CD3 stimulation, may induce Ser-16 phosphorylation via this protein kinase pathway. There are some interesting reports demonstrating that the Ca2+/calmodulin-dependent protein kinase I1 and cAPK share specific phosphorylation sites. These cases include the transcription factors CREB and CIEBPP, as well as synapsin I (51)(52)(53). The sequence surrounding Ser-16 (Lys-Arg-Ala-SedP)) does not strictly conform to the optimal consensus recognition sequence of the Ca2+/calmodulin-dependent protein kinase 11, which is Arg-X-X-SedP) (where X represents any amino acid) (54). Thus, future in vitro phosphorylation experiments, using kinase preparations purified from T cells, are required to evaluate the role of Ca2+/calmodulin-dependent protein kinases in Ser-16 phosphorylation.
We have recently identified two distinct proline-directed kinase families that phosphorylates Op18 with overlapping but distinct site preference (33). These two kinase families, MAP kinases and cyclin-dependent cdc2 kinases, are involved in receptor-and cell cycle-regulated phosphorylation events, respectively (for review, see Refs. 18 and 55). Our previous study demonstrated that the MAP kinase has a 20-fold preference for Ser-25, as opposed to Ser-38 of Op18, while cdc2 kinases have a &fold preference for the Ser-38 residue. The Jurkat T cell line expresses approximately 4.5 x lo6 Op18 moleculedcell, and antigen receptor stimulation has been shown to result in a rapid conversion of 3545% of all Opl8 molecules to the Ser-25 phosphorylated form (33). Hence, the result suggested that Ser-25 of Op18 may be a major cytoplasmic substrate for the MAP kinase. In the present study, we have identified a phosphorylation site located on Ser-16 of Op18. The data suggest that CD3 triggering induces phosphorylation of Ser-16, via a Ca2+-activated pathway, concomitant with MAP kinase-dependent phosphorylation on primarily Ser-25. Since Op18 contains a significant "basal" level of Ser-38 phosphorylation, it follows that CD3-induced phosphorylation of primarily Ser-16 and Ser-25 will result in generation of mono-, di-, and triphosphorylated forms of Op18 that are phosphorylated on Ser-16, Ser-25, and Ser-38 in various combinations. It is shown in the present study that certain combinations of Ser-16, Ser-25, and Ser-38 result in generation of 19-, 21-, and 23-kDa Opl8 phosphoisomers, as depicted in Fig. 9. Hence, our results may be useful in the interpretation of some previous reports, which are based on the analysis of proteins separated by two-dimensional PAGE and suggest a very complex pattern of Op18 phosphorylation (30)(31)(32).
It is well established that stimulation of the T cell antigen receptor result in a prominent Ca2+ response and activation of PKC (11). Recent studies have also demonstrated that anti-CD3, as well as phorbol esters, activates the MAP kinase in Jurkat cells (56)(57)(58). It does not seem surprising, therefore, that treatment of Jurkat cells with PDBu, in combination with the Ca2+ ionophore ionomycin, mimics CD3 triggering with respect to increased phosphorylation of Ser-16, Ser-25, and Ser-38 of Op18. Thus, in the present study we showed phosphopeptide analysis of the 19-, 21-, and 23-kDa phosphoisomers of Op18 that were generated after stimulation of cells with either PDBu and ionomycin or anti-CD3 (Figs. 6 and 7). The results revealed that anti-CD3 induces exactly the same pattern of phosphopeptides, in each phosphoisomer of Op18, as PDBu together with ionomycin. Thus, it seems reasonable to suggest that the pharmacological activation of cells with ionomycin together with PDBu,mimics the response elicited via the T cell antigen receptor-associated CD3 complex with respect to phosphorylation of Ser-16, Ser-25, and Ser-38 of Op18.
Activation of PKC has been shown to activate the MAP kinase in several cellular systems (59, 601, including T cells (57,58). In previous reports concerning T cell activation, it has been suggested that CD3-induced activation of the MAP kinase is at least partially dependent on PKC activation (57, 58). By employing the Ro 31-7549 PKC inhibitor (34, 3 9 , which completely blocked PDBu-induced phosphorylation of Op18, we obtained data suggesting a PKC-independent pathway of Op18 Ser-25 phosphorylation (Figs. 1 and 2). Since this particular phosphorylation event seems to be catalyzed by the MAP kinase, these results suggest that CD3-induced activation of the W kinase pathway is independent of concomitant activation of PKC. Such a possibility is strongly supported by a recent report, demonstrating that CD3 triggering can induce activation of the p2lnS oncoprotein independently of PKC activation (28,29). Thus, since p21n8 has been shown to be involved in the regulation of the MAP kinase in both PC12 cells and fibroblast cell lines (23)(24)(25), it seems possible that PKC-independent activation of p2lnS in CD3-induced T cells would also be manifested on the level of MAP kinase activation. Thus, although PKC activation clearly regulates the MAP kinase in T cells, it appears that the MAP kinase can also be regulated via the antigen receptor independently of PKC. Moreover, the present study also suggests that Ca2+-induced phosphorylation of Ser-16 of Op18 is regulated independently of PKC activation.
These results indicate that the T cell antigen receptor may regulate phosphorylation of several sites of Op18 independently of the PKC pathway (Figs. 1, 6, and 7).
A recent report has demonstrated that a Ca2+ signal alone, induced by ionomycin or thapsigargin, activates the MAP kinase in foreskin fibroblasts or the A431 epidermal carcinoma (61). The present study shows that ionomycin treatment of the Jurkat T cell line results in a selective phosphorylation of Ser-16, while the level of the MAP kinase targets Ser-25 and Ser-38 were unaltered (Fig. 6B ). Hence, there appears to exist a clearcut tissue-specific difference with respect to Ca2+-dependent regulation of the MAP kinase pathway. Moreover, there are other examples of differences between T cells and fibroblast systems on the level of signal transduction (62). The most striking difference between these cellular systems is on the level of p2lns regulation. Thus, while p21" in T cells can be regulated by both a PKC-independent and PKC-dependent pathway, only a PKC-independent pathway operates in fibroblasts (63).
The present study employed Jurkat cells synchronized in the G I B phase of the cell cycle by a double thymidine block. This procedure decreases the "basal" phosphorylation of Op18, which is primarily located on Ser-38, in unstimulated cells (data not shown). Removal of thymidine from synchronized cells results in S phase progression and a concomitant increase in Op18 phosphorylation, primarily on Ser-38, which reaches its peak during mitosis.2 In vitro phosphorylation experiments, employing either the ~3 4~~~ kinase or the ~3 3 '~~~ kinase, revealed that Ser-38 is a major and Ser-25 is a minor target for cyclin-dependent protein kinases (33).2 Thus, in addition to the receptor-stimulated phosphorylation events discussed above, Op18 is also a target for cell cycle-regulated protein kinases. It seems intriguing that the MAP kinases and the cdc2 kinases show overlapping but distinct site preference (see Fig. 9). Based on site-mapping studies in intact cells and in vitro phosphorylation experiments, it appears that the Op18 protein resides at a junction where receptor-and cell cycle-regulated protein kinases interact. A future challenge will be to elucidate the putative function(s) of these phosphorylation events.
Establishment of transfected cell lines that overexpressed Op18 resulted in a somewhat surprising finding (6). Our initial expectation, based on overexpression of Op18 in lymphoid malignancies, was that ectopic expression of Op18 in a non-transformed cell line would result in less stringent growth control. However, the result showed that overexpression of Op18 results in a partial growth arrest. It seems possible that the phenotypic outcome of ectopic expression of Op18 is governed by the phosphorylation status of this cytoplasmic protein. As outlined in Fig. 9, distinct Op18 Ser residues are targets for MAP kinases, cdc2 kinases, and an as yet unidentified Ca2+regulated kinase(s). The identification of Ser-16, Ser-25, and Ser-38 as targets for distinct kinase systems will allow future experiments where the putative function(s) of the observed phosphorylation events may be evaluated by transfection of site-specific Op18 mutants.
ing of the manuscript, Prof. H&n Bergstrand for providing D7502 (Ro