Critical Role of Ser-520 Phosphorylation for Membrane Recruitment and Activation of the ZAP-70 Tyrosine Kinase in T Cells

Abstract

Regulation of protein tyrosine kinases (PTKs) by tyrosine phosphorylation is well recognized; in fact, nearly all PTKs require phosphorylation of tyrosine residues in their “activation loop” for catalytic activity. In contrast, the phosphorylation of PTKs on serine and threonine residues has not been studied nearly as much. We report that the ZAP-70 PTK contains predominately phosphoserine in normal T lymphocytes as well as in Jurkat T leukemia cells. We have identified one site of phosphorylation as Ser-520 and find this site to be important for the recruitment and activation of ZAP-70 in T cells. Mutant ZAP-70-S520A had reduced ability to autophosphorylate and to mediate antigen receptor-induced interleukin 2 gene activation and was not enriched at the plasma membrane. These defects were rescued by addition of a myristylation signal to the N terminus of ZAP-70-S520A to force its plasma membrane and lipid raft localization. We conclude that phosphorylation of ZAP-70 at Ser-520 plays an important role in the correct localization of ZAP-70 and in priming ZAP-70 for its acute recruitment and activation upon antigen receptor ligation.

The ZAP-70 protein tyrosine kinase (PTK) plays a critical role in T-cell-antigen-receptor (TCR) signaling (5, 6, 28). The lack of ZAP-70 expression causes a severe immunodeficiency (2, 14), characterized by the absence of CD8⁺ T cells and TCR-unresponsive mature CD4⁺ T cells. Mice lacking ZAP-70 are also deficient in the production of CD4⁺ T cells, while the natural killer cells are unaffected (30). Perhaps the most important feature of ZAP-70 is its recruitment and high-affinity association with the phosphorylated immune receptor tyrosine-based activation motifs (ITAMs) of the TCR receptor complex. The crystal structure of the complex of the N terminus of ZAP-70 bound to a doubly phosphorylated ITAM peptide (18) showed that the second SH2 domain binds the first phosphorylated tyrosine of the peptide in the usual SH2-ligand manner, while the second phosphorylated tyrosine interacts with both SH2 domains in a unique manner due to the presence of an incomplete phosphotyrosine (PTyr)-binding pocket in the N-terminal SH2 domain, which is made functional by the close proximity of the other SH2 domain (17). This feature of the solved structure explains the strongly synergistic binding of ZAP-70 to doubly phosphorylated ITAMs (18, 20, 29, 31). Binding of ZAP-70 to ITAMs alone is insufficient to activate the kinase (20, 30), and inactive ZAP-70 can be associated with TCR-ζ, as observed with thymocytes (29).

Activation of the receptor-associated ZAP-70 is accomplished by the phosphorylation of Tyr-493 of ZAP-70, a reaction catalyzed by the Src family kinase Lck (4, 26, 28, 36). Tyr-493 corresponds to the positive regulatory phosphorylation site in the activation loop in all protein kinases, except Csk. In contrast to Src family PTKs and many other kinases, ZAP-70 is unable to autophosphorylate at this site. However, once ZAP-70 has been activated by phosphorylation at Tyr-493, the kinase becomes able to autophosphorylate at a number of additional tyrosine residues, which subsequently function as docking sites for SH2 domain-containing enzymes or adapters, including Lck (13), Abl (30), Vav (22, 39), Cbl (15, 24, 25, 27), and PLCγ1 (12, 32).

We have found that ZAP-70, in addition to being phosphorylated on tyrosine, is highly phosphorylated on serine. One of the targets of this phosphorylation was identified as Ser-520, located in the conserved SDVWS motif immediately downstream of the activation loop of ZAP-70. This phosphorylation site proved to be important for the normal function of ZAP-70 in TCR signaling and interleukin-2 gene activation.

MATERIALS AND METHODS

Antibodies and synthetic peptides.

Anti-ZAP-70 and anti-PTyr monoclonal antibodies (MAbs) were from Upstate Biotechnology (Lake Placid, N.Y.), anti-CD3ɛ MAb (OKT3) was from Ortho Diagnostic Systems Inc. (Raritan, N.J.), anti-HA.11 MAb was from Bio/Can Scientific (Montreal, Canada), horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG) and goat anti-rabbit IgG were from Zymed (San Francisco, Calif.), anti-Lck MAb (3A5) and polyclonal anti-Lck (2102) were from Santa Cruz Biotechnology (Santa Cruz, Calif.), anti-human TCR-ζ was from BD Pharmingen (San Diego, Calif.), and agarose-conjugated goat anti-rabbit IgG and anti-mouse IgG were from Sigma (St. Louis, Mo.). The rabbit polyclonal antibodies against ZAP-70 were generated as previously described (8).

Plasmids and site-directed mutagenesis.

The expression plasmids for ZAP-70 in pEF/HA and Lck in pEF-neo have been described previously (9). Luciferase reporter plasmids were described previously (1, 18, 21). Renilla luciferase as well as the Dual-Luciferase reporter assay system kit were purchased from Promega (Madison, Wis.). Site-directed mutagenesis was done using the QuikChange kit (Stratagene, La Jolla, Calif.), following the manufacturer's instructions. The cDNAs encoding myristylated versions of ZAP-70 were generated by addition of nucleotides encoding the first 16 amino acids from Lck upstream of the hemagglutinin (HA) tag.

Cells and transfections.

The Jurkat human T leukemia cell line, Jurkat TAg (a clone stably transfected with simian virus 40 large T antigen), and the ZAP-70-deficient Jurkat clone P116 (38) were kept at logarithmic growth in RPMI 1640 with l-glutamine, containing 5% fetal calf serum and antibiotics. Jurkat TAg cells were grown in the presence of 0.5 mg of G418/ml. COS cells were grown in Dulbecco's modified Eagle medium with 5% fetal calf serum and antibiotics.

COS cells were transfected at 10⁶ cells per sample by lipofection with 5 to 10 μg of plasmid DNA as previously described (8, 9). Jurkat cells were transfected with 10 to 30 μg of plasmid DNA at 10⁷ cells/sample by electroporation at 280 V, 950 μF, in a Gene Pulser II (Bio-Rad). Cells were used 48 h after transfection.

Immunoprecipitation, gel electrophoresis, and immunoblotting.

Immunoprecipitations were carried out as before (8-11). Proteins were resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis on 10 to 12% gels, transferred to nitrocellulose membranes, and immunoblotted with anti-ZAP-70 MAb (1:3,000), anti-HA MAb (10 μg/ml), or 4G10 antiphosphotyrosine (anti-pTyr) MAb (1:5,000). The blots were developed by the enhanced chemiluminescence technique (ECL kit; Amersham) according to the manufacturer's instructions.

Metabolic ³²P labeling, tryptic peptide mapping, and phosphoamino acid analysis.

Transfected cells were phosphate starved in phosphate-free RPMI 1640 medium containing 20 mM HEPES for 1 h at 37°C and labeled for 4 h with 4 mCi of ³²Pi/ml in phosphate-free medium. Cells were incubated with 5 μg of OKT3 MAb/ml on ice for 15 min, washed, and incubated with 10 μg of goat anti-mouse IgG/ml at 37°C (for the times indicated in the figures) or treated with 0.5 μM okadaic acid for 1 h at 37°C. The cells were lysed under denaturing conditions (11), and HA-tagged ZAP-70 was immunoprecipitated. Proteins were resolved by SDS-PAGE, transferred to nitrocellulose filters, localized by autoradiography, and excised. Tryptic peptide mapping was carried out by the method of Luo and colleagues (23) as members of our group have done before (1, 3, 9-11, 33). Briefly, the filter pieces were blocked with polyvinylpyrrolidone 360, washed, and incubated with two additions of 10 μg of tosylphenylalanyl chloromethyl ketone-treated trypsin in 50 mM ammonium bicarbonate. The released phosphopeptides were lyophilized twice and then separated by electrophoresis on cellulose thin-layer plates (Merck, Darstadt, Germany) at 1,000 V in pH 1.9 buffer for 27 min, followed by ascending chromatography in n-butanol-pyridine-acetic acid-water (75:50:15:60) for 16 h. The resulting peptide maps were then dried and exposed to film.

For secondary digests, phosphopeptides 1, 2, and 3 were eluted from the phosphopeptide map of wild-type ZAP-70 and divided into three samples. A first sample was redigested with trypsin to ensure that the first digest was complete, and a second sample was digested with 10 μg of chymotrypsin. The third sample was used for phosphoamino acid analysis (see below). Sensitivity to chymotrypsin was assessed by comigration of the tryptic and chymotryptic peptides originating from peptides 1 and 3. Peptide 2 was processed as described in Results. Phosphoamino acid analysis was performed as described previously (7). Briefly, the ³²P-labeled proteins or peptides were eluted from either polyvinylidene difluoride membranes or from peptide maps and hydrolyzed by incubation in 6 N HCl for 2 h at 110°C. The material was lyophilized and separated by electrophoresis in pH 1.9-pH 3.5 buffer (1:1 ratio [7]). ³²P-labeled amino acids were detected by autoradiography, and identified by comigration of standard phosphoamino acids that were added to each sample and detected by ninhydrin staining.

Manual Edman degradation.

³²P-labeled phosphopeptides were extracted from the thin-layer plate and dissolved in 20 μl of water, and 2 μl was taken as starting material. Twenty microliters of 5% phenyl isothiocyanate in pyridine was added and incubated for 30 min at 45°C. The sample was extracted at room temperature with heptane-ethyl acetate (10:1) and then with heptane-ethyl acetate (2:1). The final aqueous phase was frozen and lyophilized. The sample was cleaved in 50 μl of trifluoroacetic acid, incubated for 10 min at 45°C, and lyophilized. The residue was resuspended in 18 μl of water, and 2 μl was kept as first cycle product. The volume was then restored to 20 μl, and the whole procedure was repeated up to six times. The reaction products of each cycle were analyzed by thin-layer electrophoresis for 25 min at 1 kV at pH 1.9 followed by autoradiography with free ³²Pi as a marker. In this protocol, the release of free ³²Pi results from β elimination during cyclization and indicates the presence of a PSer or phosphothreonine (PThr) residue. In contrast, PTyr is stable to cyclization and is released as the anilinothiazolinone derivative of PTyr, which can be converted to the phenylthiohydantoin of PTyr by incubation of the reaction products in 0.1 N HCl for 20 min at 80°C and can be detected as a dark spot when the thin-layer chromatography plate is examined under UV light or stained with ninhydrin.

Kinase assays.

Immune complex kinase assays were carried out by incubation of immunoprecipitated ZAP-70 in a solution containing 25 μl of 10 mM Tris-HCl (pH 7.5), 10 mM MgC1₂, 10 mM MnCl₂, 0.1% Triton X-100, with 1 μM ATP and 10 μCi of [γ-³²P]ATP for 20 min at 30°C. The reaction was stopped by adding SDS sample buffer and heating at 95°C for 2 min. The phosphorylation of proteins was analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography. Incorporation of ³²P was quantitated by liquid scintillation in a β-counter.

Luciferase assays.

These assays were carried out essentially as described previously (1, 21). Briefly, ZAP-70-deficient P116 cells were transiently cotransfected with empty vector, ZAP-70 wild type, or mutant constructs, together with 0.5 μg of Renilla luciferase plasmid and 5 μg of reporter plasmid. Twenty-four hours later, cells were stimulated with 2 μg of anti-CD3ɛ MAb/ml plus 4 μg of goat anti-mouse antibody/ml alone or with 50 ng of phorbol myristate acetate (PMA)/ml and for maximum activity with 50 ng of PMA/ml plus 1.5 μM ionomycin for 6 h at 37°C. Cells were lysed, and luciferase activity was determined with a luminometer (EG&G Berthold Lumat LB 9507) according to the instructions for the Dual-Luciferase reporter assay system (Promega). Luciferase activities were normalized for transfection efficiency using the Renilla luciferase activity, and results are expressed as a percentage of the activity induced by PMA plus ionomycin. Each shown value is an average from triplicate determinations.

Confocal microscopy.

Immunofluorescence staining was done as before (1, 17, 37). Briefly, cells were washed in phosphate-buffered saline and fixed in freshly made 3.7% formaldehyde. Fixed cells were permeabilized with 0.1% saponin in phosphate-buffered saline, blocked in 2.5% normal goat serum in 0.1% saponin in phosphate-buffered saline for 30 min at room temperature, and then incubated with primary and secondary antibodies diluted in the same buffer for 1 h each at room temperature. After three washes with phosphate-buffered saline, the cells were mounted onto glass slides and viewed under a confocal laser scanning microscopy MRC-1024 (Bio-Rad). A differential interference contrast image was also taken of most cells.

Isolation of lipid rafts.

Detergent-soluble and -insoluble fractions of transfected Jurkat cells were prepared by lysing 2 × 10⁷ cells in a solution containing 1 ml of 1% Triton X-100, 25 mM 2-(N-morpholino)ethanesulfonic acid (pH 6.5), 150 mM NaCl, 5 mM EDTA, 30 mM sodium pyrophosphate, 1 mM Na₃VO₄ and 10 μg of leupeptin and aprotinin/ml for 20 min on ice and homogenized 10 to 15 times with a Dounce homogenizer. Samples were centrifuged at 1,000 × g for 10 min at 4°C. The supernatants were collected and mixed in 1 ml of lysis buffer containing 80% sucrose, transferred to ultracentrifuge tubes, and overlaid with 2 ml of lysis buffer containing 30% sucrose and 1 ml of buffer with 5% sucrose. Samples were centrifuged at 200,000 × g for 16 to 18 h at 4°C. Twelve fractions of 0.4 ml were collected from the top of the gradient. Fractions 2 to 4 and 8 to 12 were combined and referred to as Triton-insoluble glycolipid-enriched membranes (lipid rafts) and Triton-soluble fractions, respectively.

Subcellular fractionation.

Jurkat cells (10⁸) were suspended in a solution containing 1 ml of 20 mM Tricine (pH 7.8), 250 mM sucrose, 1 mM EDTA, 1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, and 1 μg of aprotinin/ml, homogenized 15 times with a Dounce homogenizer, and centrifuged at 1,000 × g for 10 min at 4°C. The supernatant was then centrifuged at 368,000 × g for 1 h at 4°C. The supernatant of this centrifugation is referred to as cytosol. Plasma membrane fractions were obtained by centrifugation (84,000 × g for 35 min at 4°C) of 1.5 ml of the postnuclear supernatant overlaid with homogenization buffer containing 30% Percoll. The opaque material at the interface was collected. The proteins from each fraction were precipitated by trichloroacetic acid in the presence of 20 μg of bovine serum albumin as a carrier, and the precipitate was washed with ice-cold ethanol and dissolved in SDS sample buffer.

RESULTS

ZAP-70 is phosphorylated on serine and tyrosine in T cells.

Syk and ZAP-70 were immunoprecipitated from metabolically ³²P-labeled peripheral blood lymphocytes from healthy volunteers and subjected to phosphoamino acid analysis (Fig. 1A), which revealed that both enzymes were heavily phosphorylated on serines. Long exposures also revealed a trace of PTyr (data not shown). Similar results were obtained with a HA epitope-tagged ZAP-70 expressed in Jurkat T cells (Fig. 1B and C). As expected, anti-CD3 stimulation increased both ³²P labeling and tyrosine phosphorylation of ZAP-70 (Fig. 1B, compare lanes 2 and 4). Treatment of the cells with the serine phosphatase inhibitor okadaic acid increased both the overall ³²P labeling of ZAP-70 and its tyrosine phosphorylation (Fig. 1B, lanes 2 and 4). Tryptic peptide maps showed that both anti-CD3 and okadaic acid caused an increase in ³²P content of several peptides, anti-CD3 being more efficient (Fig. 1C). Phosphoamino acid analysis of several spots revealed the presence of mostly PSer, with PTyr in some peptides and even a trace of PThr in one peptide (Table 1).

FIG. 1.

Serine phosphorylation of Syk family PTKs in normal lymphocytes and in Jurkat T cells. (A) Phosphoamino acid analysis of Syk and ZAP-70 immunoprecipitated from metabolically ³²P-labeled human blood lymphocytes. The cells were grown for 48 h in the presence of IL-2 andphytohemagglutinin prior to labeling and immunoprecipitation. (B) Phosphorylation of ZAP-70 in ³²P-labeled Jurkat T cells treated with medium alone (lane 2), 0.5 μM okadaic acid for 1 h (lane 3), or 5 μg of anti-CD3 MAb/ml for 5 min (lane 4). Top panel shows the autoradiogram of the material precipitated using a control MAb (lane 1) or anti-ZAP-70 MAb (lanes 2 to 4). Middle and lower panels represent anti-PTyr and anti-ZAP-70 immunoblots, respectively, of the same samples. (C) Tryptic phosphopeptide maps of the ZAP-70 band from lanes 2 to 4 of panel B. The lower-right panel is a schematic representation of the peptides subjected to phosphoamino acid analysis (Table 1). The same peptide numbering system is used throughout the text and figures.

TABLE 1.

Phosphoamino acid analysis of tryptic peptides from ZAP-70

Peptide no.a	Phosphoamino acid(s) detected
1	PSer, trace of PTyr
2	PSer, PTyr, trace of PThr
3	PSer, trace of PTyr
4	PSer, trace of PTyr after CD3 stimulation
5	PSer, trace of PThr
6	PTyr, trace of PSer
7	PSer, trace of PTyr
8	PSer

^aSee Fig. 1C for location of each peptide on the tryptic peptide map.

Identification of one phosphorylation site as Ser-520.

To begin the process of phosphorylation site identification, we aligned the amino acid sequence of the catalytic domain of ZAP-70 with that of other PTKs. A well-conserved motif, SDVWS, found in all Src and Syk family PTKs, was chosen as a first possible phosphorylation site. Three ZAP-70 mutants were generated to test whether the S⁵²⁰DVWS⁵²⁴ motif in ZAP-70 contains phosphate: S520A, S524A, and a double S520A/S524A (SSAA) mutation. When these constructs were expressed in Jurkat T cells metabolically labeled with ³²Pi, immunoprecipitated, and analyzed by tryptic peptide mapping, it was clear that three peptides were missing from maps of the S520A and double SSAA mutants (Fig. 2A). Examination of the predicted tryptic peptide containing Ser-520 (Fig. 2B) revealed that this peptide is likely to yield multiple spots because it contains a poorly cleaved Lys-Pro bond and incompletely digested Lys-Lys in its C terminus. Together with possible phosphorylation of other residues within the same peptide, this may explain why more than one spot disappeared in the map of the S520A point mutant compared with the unmutated ZAP-70. Alternatively, the three spots represent additional sites, the phosphorylation of which depends on prior phosphorylation at Ser-520. To test this directly, we took advantage of the presence of several chymotrypsin cleavage sites in the predicted Ser-520-containing tryptic peptide (Fig. 2B). Indeed, when peptides 1, 2, and 3 were excised and digested with chymotrypsin, the migration of peptide 2 changed dramatically (Fig. 3A), while peptides 1 and 3 remained unchanged as demonstrated by the comigration of the trypsin- and chymotrypsin-treated peptides (“mix,” right panels). Peptide 2 not only changed migration upon chymotrypsin digestion but also fragmented into several spots. It is also worth noting that peptide 2 became elongated with a tail to the left, as usually seen with peptides containing tryptophan, which is partially oxidized in contact with air. Indeed, when the tailed spot was excised from the two maps, mixed, and rerun on a thin-layer plate, the tail resolved into a separate spot. Together, these results indicate that peptide 2 may represent phospho-Ser-520 with additional phosphates in the same peptide, while peptides 1 and 3 are likely other sites of phosphorylation.

FIG. 2.

ZAP-70 is phosphorylated at Ser-520 in resting Jurkat T cells. (A) Tryptic peptide maps of ZAP-70, ZAP-70-S520A, or ZAP-70-SSAA, immunoprecipitated with the anti-HA MAb from metabolically ³²Pi-labeled Jurkat cells transfected with the corresponding expression plasmids. The lower-right panel depicts the major peptides obtained from ZAP-70. Note that peptides 1, 2, and 3 are missing in the two mutants. (B) Amino acid sequence of the predicted tryptic peptide containing Ser-520. Potentially phosphorylated serines and tyrosines are highlighted in bold font. Note that the peptide contains three incomplete internal tryptic cleavage sites and five predicted chymotrypsin cleavage sites.

FIG.3.

Ser-520 is located in peptide 2. (A) Phosphopeptide mapping of peptides 1, 2, and 3 digested again with trypsin (left panels) or chymotrypsin (middle panels). For peptides 1 and 3, insensitivity to chymotrypsin was shown by extraction, mixing, and rerunning the peptide (right panels). For peptide 2, which gave rise to multiple peptides (=chymotrypsin sensitive), only the diffuse major spot was extracted and rerun for better resolution (top-right panel). (B) Edman degradation of peptide 2. The release of free ³²Pi in cycle 1 is consistent with the presence of pSer at position 1 in the peptide, while the absence of free phosphate in cycle 5 indicates that Ser-524 is not phosphorylated. The last lane represents marker ³²Pi. (C) Amino acid sequence of the Ser-520-containing peptide to show how Edman degradation cycles correspond to residues in the peptide. Ser-520 and Ser-524 are highlighted in bold font.

To confirm that Ser-520 is phosphorylated and to determine whether Ser-524 may also contain phosphate, we subjected peptide 2 to manual Edman degradation (Fig. 3B), which showed that free ³²Pi was detected in the first cycle, but not in cycle 5. In this protocol, the release of free ³²Pi results from β elimination during cyclization and indicates the presence of a PSer or PThr residue. In contrast, PTyr is stable to cyclization and is released as the anilinothiazolinone derivative of PTyr, which can be converted to the phenylthiohydantoin of PTyr by incubation of the reaction products in 0.1 N HCl for 20 min at 80°C and can be detected as a dark spot when the thin-layer chromatography plate is examined under UV light or stained with ninhydrin. We did not detect PTyr within the first six cycles of peptide 2. These results confirm the identity of the peptide, since very few other tryptic peptides from ZAP-70 contain a Ser at position 1 and only one other is sensitive to chymotrypsin. Our data also demonstrate that Ser-524 and Tyr-525 were not phosphorylated (Fig. 3C). We conclude that ZAP-70 is phosphorylated at Ser-520 in intact T cells.

ZAP-70-S520A has an impaired ability to autophosphorylate.

To begin the process of determining whether phosphorylation of ZAP-70 at Ser-520 plays any role in the regulation or function of ZAP-70, we first expressed the Ser-to-Ala mutants in COS cells alone or together with the Src family PTK Lck, the upstream activator of ZAP-70 in T cells (4, 5, 28, 36). Subsequently, the expressed ZAP-70 was immunoprecipitated and analyzed for kinase activity by autophosphorylation and for PTyr content by immunoblotting. As shown in Fig. 4A, unmutated ZAP-70 and the ZAP-70-S524A mutant were phosphorylated on tyrosine and readily autophosphorylated in vitro. In contrast, ZAP-70-S520A and ZAP-70-SSAA did not contain PTyr and had little kinase activity, despite being expressed at very similar levels. A negative control, the kinase-inactive ZAP-70-K369S mutant, behaved like the S520A and SSAA mutants. Coexpression of Lck (Fig. 4B) resulted in some tyrosine phosphorylation of all ZAP-70 constructs but only a modest increase in the kinase activity of the S520A and SSAA mutant ZAP-70s, despite levels of PTyr-content in these proteins similar to that in the unmutated ZAP-70. As expected, inactive or “kinase-dead” ZAP-70-K369S remained inactive. When the ZAP-70 constructs were expressed in Jurkat TAg cells, which express endogenous Lck, the catalytic activity of ZAP-70-S520A was also found to be very low (Fig. 5). No activation was observed after anti-CD3 stimulation of the cells, and even pervanadate had only a small effect on this mutant. Taken together, these experiments demonstrate that mutation of Ser-520 impaired ZAP-70 activation by Src family kinases.

FIG. 4.

Ser-520 is required for the activation of ZAP-70 in transfected COS-1 cells. (A) In vitro kinase assay of ZAP-70 obtained by anti-HA immunoprecipitation from COS-1 cells transfected with vector alone or with the indicated constructs. Top panel is an autoradiogram of the radiolabeled material. The middle and lower panels are anti-PTyr and anti-HA blots of the same membrane, respectively. Similar results were obtained in at least three independent experiments. (B) Same as panel A, except that the COS-1 cells were cotransfected with a plasmid encoding Lck. Note that the presence of Lck induced similar levels of tyrosine phosphorylation of each ZAP-70 construct (middle panel).

FIG. 5.

Ser-520 is required for the activity of ZAP-70 in transfected Jurkat T cells. (A) In vitro kinase assay of ZAP-70 immunoprecipitated with anti-HA MAbs from resting (−) or anti-CD3-stimulated (+) Jurkat cells transfected with vector alone (lane 1), ZAP-70 (lanes 2 and 3), ZAP-70-S520A (lanes 4 and 5), ZAP-70-S524A (lanes 6 and 7), or ZAP-70-K369S (kinase-deficient enzyme; lanes 8 and 9). The upper panel is the autoradiogram showing the ability of ZAP-70 to autophosphorylate, and the middle and lower panels represent anti-PTyr or anti-HA immunoblots, respectively, of the same samples. (B) Similar experiment as in panel A, except that the cells were treated for 5 min with pervanadate instead of anti-CD3. Note that ZAP-70-S520A was inducibly tyrosine phosphorylated (middle panels) in both experiments but remained unable to autophosphorylate (upper panels).

Phosphorylation at Ser-520 is required for TCR-induced transactivation of the interleukin 2 (IL-2) promoter.

To test if phosphorylation at Ser-520 is important for the function of ZAP-70 in TCR-mediated signal transduction, we expressed the S520A mutant ZAP-70 in the ZAP-70-deficient P116 subline of Jurkat (38) together with a luciferase reporter gene driven by the entire 5′ IL-2 promoter. While cells transfected with the reporter gene alone did not respond to stimulation, cells cotransfected with unmutated ZAP-70 exhibited a robust activation of the IL-2 promoter in response to anti-CD3 plus phorbol ester (Fig. 6). In contrast, the ZAP-70-S520A or SSAA mutants failed to reconstitute IL-2 promoter activation. Kinase-dead ZAP-70-K369S was equally unable to reconstitute signaling. All ZAP-70 proteins were expressed at levels similar to or higher than wild-type protein levels, as shown by anti-HA blots of equal amounts of total cell lysates (not shown). These results indicate that phosphorylation of Ser-520 is required for the proper function of ZAP-70 in TCR-mediated signal transduction.

FIG. 6.

ZAP-70-S520A cannot mediate TCR-induced transactivation of the IL-2 promoter in ZAP-70-negative T cells. (A) Luciferase activity in lysates of ZAP-70-deficient P116 cells cotransfected with the indicated ZAP-70 expression constructs, IL-2-luciferase reporter plasmid, and control Renilla luciferase, and either left unstimulated (gray columns) or stimulated with anti-CD3 MAb plus goat anti-mouse MAb and PMA for 6 h at 37°C. The measured luciferase activity was normalized for transfection efficiency using the Renilla luciferase activity, and results are expressed relative to that induced by ionomycin plus PMA. Values represent the average of triplicate determinations. (B) A similar independent experiment.

Phosphorylation at Ser-520 does not affect binding to TCR-ζ.

To exclude the possibility that mutation of Ser-520 had more global effects on the folding or function of ZAP-70, we tested whether ZAP-70-S520A could still be recruited and bind phosphorylated TCR-ζ. COS-18 cells, which stably express a chimeric CD8α/TCR-ζ protein (5), were transfected with ZAP-70 constructs, and CD8α/TCR-ζ was immunoprecipitated and analyzed for bound ZAP-70. These experiments did not reveal any differences between constructs (not shown). This was not surprising, given that Ser-520 is in the catalytic domain of ZAP-70, which is not required for interaction of the tandem SH2 domains in the N terminus of ZAP-70 with phospho-ITAMs. It should also be noted that all ZAP-70 mutants were expressed at similar levels, suggesting that they all folded properly in cells.

ZAP-70-S520A is not enriched at the plasma membrane.

It has been reported that ZAP-70 is enriched at the plasma membrane in resting T cells in a manner that is independent of its SH2 domains but requires an intact kinase domain (19). Indeed, unmutated ZAP-70 was found to be enriched at the plasma membrane when expressed in Jurkat T cells, stained with anti-HA MAbs, and viewed under a confocal microscope (Fig. 7A). ZAP-70-S524A had a similar subcellular localization (not shown). In contrast, the S520A mutant displayed a more diffuse distribution throughout the cytosol (Fig. 7A). This result was obtained in many independent experiments and indicates that the S520A mutation impaired the normal targeting of ZAP-70 to docking sites at the plasma membrane, an event that presumably prepares the kinase for rapid mobilization upon TCR ligation (19, 34).

FIG. 7.

Ser-520 is required for membrane localization of ZAP-70 in T cells. (A) Confocal images of Jurkat T cells transfected with empty vector, HA-tagged ZAP-70, or HA-tagged ZAP-70-S520A, as indicated. The left panels represent anti-HA staining, the middle panels are Nomarski differential contrast images of the same cells, and the right panels are overlays of the two. (B) Anti-HA immunoblots of subcellular fractions of transfected Jurkat T cells, as indicated. Left panels represent Triton X-100-soluble (Ts) fractions and insoluble buoyant fractions (rafts), respectively. Right panels represent cytosolic (C) and membrane (M) fractions. Middle and lower panels represent control anti-Lck and anti-LAT immunoblots of the same samples. Note that ZAP-70-S520A is present in Ts and C fractions much more than unmutated ZAP-70.

To verify this interpretation by biochemical means, we first fractionated anti-CD3-stimulated ZAP-70 transfected cells by sucrose gradient centrifugation to separate the buoyant, detergent-insoluble, glycolipid-enriched lipid rafts from the detergent-soluble fractions of the cells. Immunoblotting of these fractions for the HA epitope tag demonstrated that ZAP-70-S520A was considerably less concentrated in the raft fractions than unmutated ZAP-70 (Fig. 7B, left panels). Similar results were obtained by a simple fractionation of cells into cytosolic and particulate fractions (Fig. 7B, right panels). The S520A mutant ZAP-70 was clearly more cytosolic than unmutated ZAP-70. The lack of Lck and LAT, two firmly raft-associated proteins, in our cytosolic preparations indicates that ZAP-70-S520A was indeed freely cytosolic. We repeatedly observed that more Lck and LAT could be recovered from ZAP-70 S520A-transfected cells than from cells transfected with the wild-type cDNA. The reason for this is unclear. It is unlikely, however, that the differences in cellular localization observed here are due to overexpression of the mutated enzyme over the wild-type construct, since the expression levels of both forms of ZAP-70 in unfractionated cell lysates were identical (not shown; see Fig. 4, 5, and 8). Based on all these experiments, we conclude that phosphorylation of ZAP-70 at Ser-520 plays a role in proper intracellular targeting of ZAP-70. This could well explain the impaired recruitment and activation of ZAP-70, particularly in COS cells, which have a much larger cytoplasm than T cells.

FIG. 8.

Rescue of signaling defects of ZAP-70-S520A by addition of a raft-targeting motif. (A) In vitro kinase assay of ZAP-70 obtained by anti-HA immunoprecipitation from resting (−) or anti-CD3-stimulated (+) Jurkat T cells transfected with the indicated constructs. The top panel is an autoradiogram of the kinase assay, while the middle and lower panels are immunoblots of the precipitates with anti-PTyr or anti-HA, respectively. (B) Luciferase activity in lysates of ZAP-70-deficient P116 cells cotransfected with the indicated ZAP-70 expression constructs. The experiments were performed exactly as for Fig. 6. The result presented here is representative of four independent experiments.

Effects of forced membrane localization of ZAP-70-S520A.

To directly address the notion that faulty targeting of ZAP-70-S520A to a submembranous location may be responsible for the impaired recruitment and activation of the kinase upon TCR stimulation, we created constitutively plasma membrane-targeted forms of ZAP-70 and its mutants by adding the first 16 amino acids of Lck to their N termini. This short sequence contains the necessary motifs to drive myristylation and palmitylation of the proteins. As shown in Fig. 8, the membrane-targeted ZAP-70-S520A (myrZAP-70-S520A) now responded to TCR stimulation by increased kinase activity, as did the double mutant, myrZAP-70-SSAA. However, their kinase activity was still lower than the activity of unmutated or S524A-mutated ZAP-70. More importantly, when all these constructs were expressed in the ZAP-70-negative P116 cell line together with the IL-2 promoter reporter gene, it was clear that myrZAP-70-S520A had regained the ability to support TCR signaling. Thus, it appears that the impaired activation and function of ZAP-70 caused by replacement of Ser-520 by an alanine can be, at least partly, overcome by forced plasma membrane targeting of the mutant ZAP-70. Since expression of the myrZAP-70-K369S kinase-deficient mutant failed to restore IL-2 promoter activation in this experiment, it is clear that the ZAP-70-S520A mutant is not a kinase-dead enzyme, despite the fact that it failed to autophosphorylate in vitro (Fig. 4 and 5). It is possible that the nonmyristylated, nonpalmitylated ZAP-70-S520A mutant may require more than 5 min to respond to TCR stimulation or may not be activated to a sufficient level to show autophosphorylation activity in vitro.

DISCUSSION

While it is well recognized that PTKs are regulated by phosphorylation of tyrosine residues, the role of serine phosphorylation in the physiology of PTKs remains to be assessed. Nevertheless, there are many examples of regulatory serine phosphorylation of PTKs, such as the serine phosphorylation of Csk (35) or the phosphorylation of Syk by mitogen-activated protein (MAP) kinases (40). We expect that most, if not all, PTKs will be found to contain abundant PSer and that many serine phosphorylation events will be found to regulate a number of different aspects of PTK function.

A key event in the activation of most kinases is the phosphorylation of one or several amino acid residues in their “activation loop,” a flexible protrusion that participates in blocking substrate access in its unphosphorylated state but swings out and stabilizes an active conformation of the catalytic domain upon phosphorylation. In ZAP-70 the critical amino acid residue in this loop is Tyr-493 (4). The serine phosphorylation site that we have identified, Ser-520, is only 27 residues downstream of Tyr-493 and lies at the edge of the catalytic cleft. Ser-520 lies in a sequence motif (S⁵²⁰DVWS), which is well conserved in the beginning of subdomain IX of almost all protein kinases. Interestingly, the Src family PTK Blk is unique among kinases in that the serine corresponding to Ser-520 in ZAP-70 is replaced by an alanine, as in our ZAP-70-S520A mutant. This supports our notion that the hydroxyl group of Ser-520 is not necessary for proper folding of the kinase domain.

It is unclear at this point whether Ser-520 phosphorylation is constitutive or is inducible by TCR stimulation. Since the experiments leading to the identification of this site were conducted with resting Jurkat cells, we speculate that Ser-520 is phosphorylated prior to receptor engagement. Unfortunately, comparative phosphopeptide mapping of material isolated from different samples (here resting versus activated cells) is semiquantitative at best. Until we develop a reagent capable of detecting the phosphorylated form of Ser-520, it will be difficult to ascertain the inducible nature of Ser-520 phosphorylation.

Our data suggest that phosphorylation of ZAP-70 at Ser-520 may be involved in localizing the kinase to a submembranous location from which it can be rapidly recruited and activated following receptor ligation. Indeed, experiments with Jurkat T cells indicate that ZAP-70 activation occurs within seconds of TCR engagement. We speculate that prior membrane attachment of ZAP-70, perhaps through Ser-520, allows this nearly instantaneous activation. Such a mechanism may have evolved with the need for ZAP-70 to interact with phosphorylated ITAMs before the phosphorylated tyrosines of the motif are dephosphorylated by pTyr phosphatases, such as CD45 (16) or SHP-1 (32). It is also possible that Ser-520-mediated membrane localization allows ZAP-70 to compete more efficiently with other SH2-containing molecules for ITAM binding.

We contend that the primary role of Ser-520 phosphorylation may be to target the enzyme to the membrane, rather than merely activating ZAP-70. This possibility is supported by our finding that a forced plasma membrane and lipid raft location improves the ability of the mutant ZAP-70-S520A to mediate TCR signaling. This putative role of Ser-520 phosphorylation is in agreement with the earlier finding that ZAP-70 is enriched in the submembranous space of resting T cells in a kinase domain-dependent, but SH2-independent, manner (19). Phosphorylation at Ser-520 may promote binding to another cellular protein at the plasma membrane, which then serves as an anchor in this location. A possible candidate would be a member of the 14-3-3 family of proteins, which bind other proteins in a serine phosphorylation-dependent manner. It is also possible that phosphorylation of Ser-520 promotes the adoption of the enzymatically active conformation of ZAP-70, analogous with the “priming” phosphorylation that protein kinase C and many kinases require for subsequent activation by a variety of mechanisms. This might explain why the S520A-mutated ZAP-70 has a lower enzymatic activity than the wild-type enzyme in vitro.

Finally, it should be noted that we have identified only one site of serine phosphorylation in ZAP-70 and that several additional sites presumably remain to be recognized. A recent study showed that the Erk2 MAP kinase can phosphorylate and activate Syk (40). Since MAP kinases are proline-directed kinases, we find it very unlikely that Ser-520 is the site targeted by Erk2. There are several serine and threonine residues followed by prolines in Syk and ZAP-70 that constitute much more likely candidate sites for Erk2, such as Thr-293, Ser-301, and Ser-317, all of which lie close to important regulatory tyrosine residues. We also detected the presence of PSer in several other peptides in our tryptic peptide maps of metabolically labeled ZAP-70. It remains to be determined how many of these sites have a regulatory influence on ZAP-70, what aspects of ZAP-70 function they affect, and which Ser/Thr kinases phosphorylate these sites.