Abstract
A widely accepted model for regulation of the Lck tyrosine kinase is that it is activated by CD45-mediated dephosphorylation of its COOH-terminal negative regulatory tyrosine (Tyr505). Previous work from our laboratory, however, found that despite hyperphosphorylation of Tyr505, the activity of Lck from CD45− T cell lines was actually increased due to hyperphosphorylation of the positive regulatory tyrosine, residue 394. To avoid potential complications introduced by transformed cells, in this study we have characterized the effect of CD45 on Lck activity in normal cells. Lck in thymocytes from CD45−/− mice was hyperphosphorylated on tyrosine residues. Importantly, and in disagreement with the model that CD45 only activates Lck in vivo, the kinase activity of Lck from cells lacking CD45 was substantially increased. These results support a model in which CD45 dephosphorylates both Tyr505 and Tyr394, the net effect in normal thymocytes being a decrease in enzymatic activity.
Activation of protein tyrosine kinases (PTKs)2 is an early and critical step in the signal transduction cascade originating from the TCR, and their inhibition is sufficient to prevent induction of almost all T cell responses 1, 2, 3 . Among the PTKs that have been implicated in this process are two members of the src family of tyrosine kinases: Fyn and Lck. The activity of src family kinases is regulated by the phosphorylation and dephosphorylation of two key tyrosine residues. One, in the COOH-terminal region (Tyr505 for murine Lck), is the major site of negative regulation 4, 5 . The other, in the kinase domain, enhances kinase activity when phophorylated. This tyrosine (residue 394 for murine Lck) is the major site of autophosphorylation 6 . The kinase activity of Lck and Fyn is therefore subject to regulation by enzymes that dephosphorylate these crucial tyrosines. The most abundant of these protein tyrosine phosphatases in hemopoietic cells is the transmembrane molecule CD45. Indeed, CD45− T cells have profound defects in TCR signaling 7, 8, 9 , and reconstitution of these cells with the catalytically active and membrane-associated CD45 phosphatase domain is sufficient to restore normal signaling 10, 11 . Since the COOH-terminal tyrosines of Lck and Fyn are good substrates for CD45 in vitro 12, 13 and phosphorylation of these residues is inhibitory for kinase activity, CD45 has been considered as an activator of src family kinases. An analysis of the published data, however, suggests that the relationship between CD45 and src family kinase activity is not so simple. While there are many reports showing that in CD45− T cell lines Lck is hyperphosphorylated on the COOH-terminal tyrosine 14, 15, 16, 17 , there is a lack of consistency between the studies on the influence of this hyperphosphorylation on PTK activity. Work from our laboratory using CD45− T cell lines found that Lck kinase activity was increased despite hyperphosphorylation of the COOH-terminal tyrosine residues 18 and that this was due to hyperphosphorylation of Tyr394 19 .
It has been suggested that some of the differences between studies on the role of CD45 on src family kinase activity may be due to the nature of the cells, which are typically retrovirally transformed and therefore potentially subject to the effects of viral oncogenes 20 . To address this, we now analyze the enzymatic activity of Lck in thymocytes from CD45-deficient mice 21 . Although Lck is hyperphosphorylated on tyrosine residues, its kinase activity is substantially increased in the absence of CD45, demonstrating that in normal lymphoid cells CD45 inhibits the activity of this src family kinase.
Materials and Methods
Cells and Abs
Mice with disruption of CD45 exon 9 on the C57BL/6J background 21 and normal C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Individual thymi were harvested, and single-cell suspensions were prepared in RPMI 1640 (Biofluids, Gaithersburg, MD) supplemented with 10% FCS. Cells were washed with PBS, counted, and used for flow cytometry analysis or biochemical studies. Anti-CD4-PE (RM4-5) and anti-CD8-FITC (53-6.7) were purchased from PharMingen (San Diego, CA). Anti-CD45 (M1/89) was used as culture supernatant and detected with FITC-goat anti-rat (Jackson ImmunoResearch, West Grove, PA). Anti-Lck mAb 3A5 bound to agarose beads and anti-phosphotyrosine Ab PY-99 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Lck antiserum 688 was kindly provided by Larry Samelson, National Institutes of Health, Bethesda, MD.
Immunoprecipitation and immunoblotting
Cells were washed in cold PBS and lysed in buffer containing 50 mM Tris, 300 mM NaCl, 0.5% Triton X-100, leupeptin, aprotinin, and phosphatase inhibitors 2 mM sodium o-vanadate, 0.4 mM EDTA, 10 mM sodium fluoride, and 10 mM sodium pyrophosphate (lysis buffer). Immunoprecipitation was performed on postnuclear fractions for 2 h at 4°C with protein A-Sepharose beads (Pierce, Rockford, IL) precoated with the indicated Abs. The precipitated proteins or total cell lysates were separated on reducing 8% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with the indicated Abs. Immunoblots were developed by 125I-labeled protein A and analyzed with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
In vitro tyrosine kinase assay
PTK activity was assayed using phosphorylation of the peptide KVEKIGEGTYGVVKK from p34cdc2 residues 6–20 (Pierce) as described 22 . Briefly, washed cells were lysed in lysis buffer, and postnuclear fractions were immunoprecipitated with Lck antiserum 688. After extensive washing, the immunoprecipitates were incubated with 21 μl of a reaction mixture containing the cdc2 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 peptide substrate and 10 μCi of [γ-32P]ATP (ICN) at 30°C for 10 min. Blank values were obtained for each sample in the absence of substrate. Reactions were stopped by adding 10 μl of 10% acetic acid, the samples were centrifuged in an Eppendorf microfuge (Hamburg, Germany) for 30 s, and 25 μl of supernatant from each sample were spotted on a SpinZyme basic separation unit (Pierce). After two washes with 75 mM phosphoric acid, bound radioactivity was counted with a β-scintillation counter. At the same time, the immunocomplexes used for the kinase reaction as well as for the blank reaction (with no peptide) were washed three times, resuspended in sample buffer, and separated on reducing 8% SDS-PAGE to analyze Lck autophosphorylation. Gels were transferred to nitrocellulose, and radioactivity was measured by PhosphorImager (Molecular Dynamics).
Results and Discussion
Disruption of CD45 exon 9 in mice leads to the complete absence of CD45 expression in hemopoietic cells 21 . Cytofluorometric analysis of these animals confirmed the absence of CD45 on thymocytes as well as a previously reported defect in the transition from CD4+CD8+ to the more mature CD4+CD8− and CD4−CD8+ stages of differentiation (Fig. 1⇓). To characterize Lck expression and activity, equal amounts of total protein lysates from the CD45−/− and CD45+/+ thymocytes were separated on SDS-PAGE and blotted with an Lck-specific antiserum. Thymocytes from these mice expressed Lck at comparable levels (Fig. 2⇓A). The tyrosine phosphorylation status of Lck was determined by immunoprecipitating it from equal amounts of cell lysates and immunoblotting with an anti-phosphotyrosine Ab. Consistent with data obtained with CD45− cell lines and a previous analysis of these mice 21 , Lck isolated from CD45−/− thymocytes had a much higher degree of tyrosine phosphorylation than Lck obtained from CD45+/+ thymocytes (Fig. 2⇓B).
Flow cytometry analysis of CD45−/− and CD45+/+ thymocytes. Thymocytes from CD45−/− and CD45+/+ mice were stained for CD4 and CD8 (top) or CD45 (bottom) and analyzed with a FACScan.
Lck in CD45−/− thymocytes is hyperphosphorylated on tyrosine residues. A, Lysates of CD45−/− and CD45+/+ thymocytes were separated on 8% SDS-PAGE and immunoblotted with anti-Lck. B, Lck was immunoprecipitated from CD45−/− and CD45+/+ thymocytes using Ab 3A5 covalently coupled to agarose beads. The material was separated by 8% SDS-PAGE and immunoblotted with anti-phosphotyrosine Ab PY-99. The lane labeled Ab alone is a sham immunoprecipitation with no lysate.
If CD45 activates Lck by dephosphorylating Tyr505, it would be expected that the hyperphosphorylated Lck from the CD45−/− mice would have decreased enzymatic activity. To determine whether this is so, kinase activity of Lck from CD45−/− and CD45+/+ thymocytes was compared. Two methods were used to quantitate kinase activity: autophosphorylation of immunoprecipitated Lck, and phosphorylation of an exogenous peptide substrate. As shown in Fig. 3⇓A, immunoprecipitated Lck from CD45−/− thymocytes was phosphorylated in an in vitro kinase assay to a much greater extent than Lck from CD45+/+ thymocytes. Phosphorimager measurement of the autophosphorylated Lck in independent experiments with three pairs of CD45−/− and CD45+/+ thymi revealed that there was approximately a 3.7-fold increase in kinase activity for Lck from CD45−/− compared with CD45+/+ thymocytes (Fig. 3⇓B). To measure phosphorylation of an exogenous substrate, kinase assays were performed in which a peptide from p34cdc2 (residues 6–20), which is specifically phosphorylated by src family kinases 22 and which we have previously used to determine the kinase activity of Lck 19 , was added to the beads coated with immunoprecipitated Lck (Fig. 4⇓A). As for autophosphorylation, the phosphorylation of an exogenous substrate by Lck was substantially higher when the Lck was obtained from CD45−/− compared with CD45+/+ animals. In five independent experiments, the kinase activity of Lck from CD45−/− thymocytes was on average more than twofold higher than the activity of Lck from CD45+/+ thymocytes (Fig. 4⇓B). Therefore, while Lck from thymocytes lacking CD45 is hyperphosphorylated on tyrosine residues, its kinase activity is elevated.
Measurement of Lck autophosphorylation activity. A, Lck was immunoprecipitated from CD45−/− and CD45+/+ thymocytes and the material subjected to immune complex kinase assay (top) or immunoblot with anti-Lck antiserum (bottom). The reactants were separated by 8% SDS-PAGE. B, PhosphorImager analysis of three independent autophosphorylation experiments using three pairs of CD45−/− and CD45+/+ mice. The geometric mean and the SEM (x/÷) are shown.
Assessment of Lck activity by measuring phosphorylation of an independent substrate. A, Lck was immunoprecipitated from CD45−/− and CD45+/+ thymocytes and was used to phosphorylate p34cdc2 (6–20). B, Results of 5 independent experiments in which the phosphorylation of p34cdc2 (6–20) by Lck was measured. The geometric mean and the SEM (x/÷) are shown.
There has been a great deal of work on the influence of CD45 on Lck and, to a lesser extent, Fyn activity. One experimental approach has been to mix anti-CD45 and anti-CD4 immunoprecipitates in vitro, and it was found that under these conditions there was dephosphorylation of the COOH-terminal tyrosine of the CD4-associated Lck and an increase in kinase activity 12, 23 . Another common approach has been to compare src family kinase activity from CD45− and CD45+ cell lines. Numerous studies found that the loss of CD45 resulted in increased tyrosine phosphorylation of Lck and Fyn and, when mapping analyses were performed, that the COOH-terminal tyrosine was the major phosphorylation site 14, 15, 16, 17 . Since mutational studies revealed that phosphorylation of the COOH-terminal tyrosine is an inhibitory event 4, 24 , the inference from this work was that CD45 activates Lck and Fyn by dephosphorylating the COOH-terminal tyrosine.
Despite the commonly held assumption that CD45 activates Lck in vivo 25, 26 , based largely on the analyses of Tyr505 phosphorylation, there is surprisingly little direct evidence for this model, and in fact quite a bit of data indicating that the relationship is not so straightforward. In favor of a direct positive effect of CD45 on the activation of Lck is a report in which Lck and Fyn from a CD45− CD8+ IL-2-dependent T cell clone were hyperphosphorylated in the COOH-terminal region and had reduced kinase activity 14 . In another report using a CD45− human leukemic T cell line, the Lck coprecipitated with anti-CD4 was reported to have reduced kinase activity 27 . However, in the same CD45− cells, the activity of Lck immunoprecipitated from whole cell lysates was clearly elevated. Moreover, another analysis of CD45− variants of the T cell line HPB-ALL found an increase in the kinase activity of CD4-associated Lck 28 . Cross-linking CD4 further enhanced Lck activity in these CD45− cells as well as in CD45− SAKRTLS cells 29 , demonstrating that CD45 is not required for induction of Lck activity. Other data in the literature also challenge the simple model that CD45 as a pure activator of src family kinases. For example, coclustering of CD45 with CD4 inhibits both Lck phosphorylation and increases in kinase activity induced by CD4 engagement 29 , suggesting that CD45 dephosphorylation of Lck decreases rather than increases Lck activity. In agreement with this is the finding that inhibition of tyrosine phosphatase activity by pervanadate is associated with cellular hyperphosphorylation and activation of Lck and Fyn and that these events depend on inhibition of CD45 since they do not occur in CD45− cells 30, 31 . Furthermore, in one set of CD45− Jurkat T cells, the baseline activities of Lck and Fyn have been shown to be elevated 31 . We have demonstrated in three independent T cell lines that in the absence of CD45 src family kinases are hyperphosphorylated, predominantly on Tyr505 18 . Despite this, Lck and Fyn activity were substantially elevated, and this activity was reduced to baseline levels by in vitro exposure of the hyperphosphorylated Lck to CD45. Phosphopeptide mapping and the characterization of Lck mutants with substitutions of critical tyrosine residues with phenylalanine demonstrated that loss of CD45 loss resulted in increased phosphorylation not only of Tyr505 but also of Tyr394 and that the latter was responsible for the increased kinase activity 19 . These results offer one reason for the apparent discrepancies between some of the studies, namely, that depending on the balance of phosphorylation between Tyr394 and Tyr505 one might obtain either inhibition or activation of kinase activity.
A previous study with a different CD45 knockout strain reported that the activity of the src family kinases Hck and Lyn were in fact elevated, although the activity of Lck in these mice was not determined 32 . Another study with the mice used in the present report demonstrated that Lck and Fyn were hyperphosphorylated on tyrosine residues and used Lck precipitation by a peptide containing phosphorylated Tyr505 that cannot bind the “closed form” of the molecule (i.e., phosphorylated at Tyr505) to suggest that there was more kinase in the inactive form 33 . This result is not surprising, as loss of CD45 is always associated with increased phosphorylation of Tyr505. However, since loss of CD45 also results in increased phosphorylation of Tyr394 19 , this does not mean that the net kinase activity in the population of Lck molecules will in fact be decreased. No direct assay of overall Lck kinase activity was performed in that study 33 . In the current report, we have directly quantitated the enzymatic function of Lck. In accord with our previous data with CD45− T cell lines, despite tyrosine hyperphosphorylation, Lck from CD45−/− thymocytes was 2–4 times more active than Lck from CD45+/+ thymocytes. This is similar to the degree to which kinase activity is elevated in the “constitutively active” form of Lck in which Tyr505 is replaced with a Phe 19 . These findings, together with previous observations 18, 19, 30, 31 , support an alternative to the simple model that CD45 merely activates Lck by dephosphorylation of Tyr505 (Fig. 5⇓). In this scheme, CD45 is able to dephosphorylate both Tyr505 and Tyr394, and the consequence of these activities in normal thymocytes is a decrease in Lck kinase activity. Therefore, one direct role for CD45 in normal T cell activation would be to reset Lck to its hypophosphorylated status at the end of the TCR signaling cascade, bringing the kinase back to its preactivation status. Together, these results indicate a more complex role for CD45 on src family kinase activity than is typically held and demonstrate that in normal nontransformed thymocytes CD45 down-regulates Lck activity.
Schematic of how CD45 influences Lck activity by dephosphorylating regulatory tyrosine residues. Lck can be phosphorylated at Tyr505 (left) or Tyr394 (right), or possibly both (not shown). CD45 dephosphorylates both tyrosine residues, which can either enhance or diminish Lck activity, depending on the relative abundance of Tyr505 vs Tyr394 phosphorylation. In thymocytes, the net effect is to diminish Lck activity. Lck in which Tyr394 has been substituted with a Phe is inactive, even when Tyr505 is substituted with a Phe, and therefore cannot be phosphorylated (19); this implies that this phosphorylation of Tyr394 is a prerequisite for Lck activity. The issue of whether there are Lck molecules phosphorylated on both Tyr394 and Tyr505, and if so what would the kinase activity of such a molecule be, has not been experimentally addressed.
Acknowledgments
We thank Allan Weissman for critical review of the manuscript.
Footnotes
↵1 Address correspondence and reprint requests to Dr. Jonathan Ashwell, Laboratory of Immune Cell Biology, National Cancer Institute, National Institutes of Health, 9000 Rockville Pike, Building 10, Room 1B40, Bethesda, MD 20892-1152. E-mail address: jda{at}box-j.nih.gov
↵2 Abbreviation used in this paper: PTKs, protein tyrosine kinases.
- Received November 4, 1998.
- Accepted December 10, 1998.
- Copyright © 1999 by The American Association of Immunologists