LCPTP–MAP kinase interaction: permanent partners or transient associates?

Molecular Immunology 39 (2002) 475–483

LCPTP–MAP kinase interaction: permanent partners or transient associates? Isabelle Brodeur a , Angela Boyhan a , Nikol Heinrichs a , Christopher Plumpton a , Benjamin Chain b,∗ , Wendy C. Rowan a a

Immunology Platform, GlaxoSmithKline Research and Development, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2NY, UK b Department of Immunology and Immunopathology, Windeyer Institute of Medical Sciences, 46 Cleveland Street, University College London, W1T 4JF London, UK Received 15 April 2002; received in revised form 7 May 2002; accepted 20 May 2002

Abstract LCPTP (leucocyte-phosphotyrosine phosphatase) is a 42 kDa protein tyrosine phosphatase expressed predominantly in haematopoietic cells which has been implicated in the early stages of the T cell receptor signalling pathway. The substrates of LCPTP have been shown to include MAP kinase family members, but it remains unclear whether LCPTP is found in stable constitutive association with these enzymes, or associates transiently during dephosphorylation. Here we report on LCPTP/MAP kinase interactions in CD3-stimulated Jurkat T cells. Pull-downs from Jurkat T cells using a recombinant GST-LCPTP substrate-trap protein, but not wild-type LCPTP show a clear specific association with both ERK1 and ERK2. In Jurkat cells overexpressing LCPTP, a small fraction of cell ERK1 can be immunoprecipitated in stable association with LCPTP. However, in both unstimulated and anti-CD3 antibody stimulated Jurkat T cells, we were unable to demonstrate any constitutive interaction between endogenous LCPTP and any MAP kinase family members. We propose that both ERK1 and ERK2 interact transiently with LCPTP as substrates for the phosphatase rather than as constitutive protein partners. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Phosphatase; T cell; Signalling

1. Introduction T cell activation is initiated by ligation of the T cell receptor (TCR) by an antigen presented in the context of an MHC molecule expressed on the surface of an antigen presenting cell (APC). Alternatively, T cell activation can be achieved by an antibody mimicking the antigen/APC interaction by binding to invariable chains of the TCR, such as an anti-CD3 antibody. TCR stimulation triggers a cascade of signalling events in the cell. Depending on a number of factors, for instance, the cytokine environment, engagement of costimulatory molecules, the strength of signal delivered via the TCR, the T cell can proliferate to produce effector cells and cytokines, which activate other cell types, or, alternatively can become anergic, or enter into apoptosis. One

Abbreviations: LCPTP, leucocyte phosphotyrosine phosphatase; kDa, kiloDalton; MAP, mitogen-activated protein; ERK, extracellular signal-regulated kinase; JNK, c-Jun kinase; IP, immunoprecipitation ∗ Corresponding author. Tel.: +44-20-7679-9402; fax: +44-20-7679-9357. E-mail address: [email protected] (B. Chain).

of the earliest measurable ways indicating that a T cell has been activated is by visualising increased levels of phosphotyrosine proteins (Isakov et al., 1994). Protein tyrosine phosphorylation is an important mechanism, which regulates T cell proliferation and differentiation. This phenomenon is orchestrated by the coordinated actions of kinases and phosphatases. In the early stages of the TCR response, several proteins including the TCR zeta chain, CD3, LAT and CD28 become tyrosine phosphorylated by the action of kinases such as ZAP-70, Fyn and Lck (Mustelin et al., 1998). This increase in total protein phosphorylation is regulated by the action of phosphatases, such as CD45 (Mustelin et al., 1998). To date, relatively few phosphatases have been identified and their role in T cell signalling is less well understood than the corresponding kinases. Our studies have focussed on the role of the haematopoietic restricted phosphotyrosine phosphatase LCPTP and its role in signalling following T cell receptor ligation. LCPTP is a 42 kDa protein (Zanke et al., 1992; Adachi et al., 1992). It lacks protein recognition binding motifs such as SH2, SH3 and PH domains but includes a kinase interactive motif (KIM) which is 75% homologous to the

0161-5890/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 1 6 1 - 5 8 9 0 ( 0 2 ) 0 0 1 1 8 - 9

476

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

consensus sequence (GLQERRGSNVSLTLDM). The KIM motifs were first identified in the phosphotyrosine phosphatases, PTP-SL and STEP, which are involved in regulating MAP kinase activity (Pulido et al., 1998). This suggested that likewise, LCPTP could bind and dephosphorylate MAP kinases. Two models have been put forward for the interaction of LCPTP with its potential substrate proteins in cells. In the first model, LCPTP interacts constitutively with ERK1, ERK2 and p38 kinases, via the non-catalytic N-terminal portion of the enzyme. Dephosphorylation and hence, deactivation of these kinases then occurs in a regulated manner. This model is based on data showing that recombinant LCPTP binds constitutively to ERK1, ERK2 and p38 but not to the c-Jun kinases, JNK1 and JNK2 and dephosphorylates ERK in vitro in 293T cells (Oh-hora et al., 1999; Saxena et al., 1999a). The second model proposed, based on studies using an LCPTP recombinant substrate-trap protein (C200S) suggests that LCPTP interacts with ERK2 predominantly via an active site/substrate interaction (although the N-terminal portion may contribute to overall binding)(Pettiford and Herbst, 2000). This interaction was dependent on the MAP kinase being tyrosine phosphorylated and the interaction with the MAP kinase was not constitutive (Pettiford and Herbst, 2000). Further studies have shown that cAMP-dependent protein kinase (PKA) mediates the phosphorylation of serine-34 (positioned within the KIM peptide) in LCPTP. This phosphorylation induces the release of ERK by LCPTP and the subsequent induction of transcription from the c-fos promoter (Saxena et al., 1999b). Thus, in over-expression systems using recombinant protein, LCPTP can inhibit ERK activity and also mediates a cross talk between the cAMP system and the MAP kinase cascade (Saxena et al., 1999b). Previous experiments demonstrating an interaction of LCPTP with one or more members of the MAP kinase family have all used systems in which recombinant LCPTP was over-expressed or added at levels much higher than those found normally in cells. In order to address the physiological significance of these findings, the present study included looking at protein associations of endogenous LCPTP, in T cells activated via receptor ligation.

for 2 min (time 0). The anti-CD3 antibody (Clone OKT3, A.T.C.C.) was added at a final concentration of 10 ␮g/ml and reactions stopped at suitable intervals by addition of SDS-sample loading buffer (125 mM Tris pH 6.8, 4% (w/v) SDS, 20% (w/v) glycerol, 0.2 M DTT, 0.013% bromophenol blue (Sigma, UK)). 2.2. LCPTP constructs Full length LCPTP was cloned by PCR into TA vector (Invitrogen, UK) from an H9 cell library, and the sequence confirmed by sequencing (GlaxoSmithKline sequencing service). In order to generate the substrate-trap mutant, two mutations were introduced sequentially (using the QuickChangeTM Site-directed mutagenesis kit, Stratagene, UK as per instructions) to convert cysteine at position 291 to a serine, and aspartic acid at position 257 to an alanine. The resulting doubling mutation was confirmed by sequencing. Both the wild-type and the substrate-trap gene were cloned in frame into the pcDNA 3.1myc vector (Invitrogen, UK) and into the pGEX-GST vector (Amersham Biosciences UK). E. coli (DH5␣) were transformed with the vectors containing the GST-fusion proteins, grown up and the proteins purified on GST-Sepharose (Amersham Pharmacia Biotech, UK) according to manufacturer’s protocols. 2.3. Phosphotyrosine Western blotting Samples were boiled for 5 min at 100 ◦ C, prior to loading the equivalent of 1 × 106 cells/track onto 4–20% gradient SDS-PAGE gels (Novex, UK). Proteins were transferred to polyvinylidene difluoride (PVDF) membrane (Millipore, UK) by semi-dry electroblotting; membranes were blocked to prevent non-specific binding of detecting reagents using 1% BSA (Sigma, UK) in PBS. Phosphotyrosine proteins were detected using the 4G10 mouse mAb (Upstate Biotechnology, USA) followed by goat anti-mouse horseradish peroxidase-conjugated antibody (Sigma, UK) and then exposure to enhanced chemiluminescence (ECL) reagents (Amersham, UK). To verify that equal amounts of protein were loaded per track, gels were stained with Coomassie blue (0.1% R-250 Coomassie blue (Fluka, UK), 45% methanol, 45% acetic acid) and then destained in 20% methanol, 7% acetic acid solution.

2. Experimental procedures

2.4. Immunoprecipitations

2.1. Cells and cell activation

Jurkat T cells (1 × 108 ) were activated as described above for 1 min in the presence of anti-CD3. Cells were snap frozen in liquid nitrogen and then lysed in 1% n-octyl glucoside, 10 mM Tris–HCl pH 8.0, 50 mM NaCl, 10 mM iodacetamide, 10 mM sodium orthovanadate, 2 mM EDTA, 0.076 TIU/ml aprotinin, 10 ␮g/ml pepstatin, 10 ␮g/ml leupeptin and 10 ␮g/ml 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF) (Sigma, UK). The lysate was separated from insoluble material by centrifugation for 5 min at 13,000 rpm. LCPTP protein was immunoprecipitated using an antibody

Jurkat T E6-1 cell line (ATCC) was maintained in medium RPMI 1640 (Life Technologies, UK) supplemented with 10% fetal calf serum, 2 mM l-glutamine, 100 IU/ml penicillin and 100 ␮g/ml streptomycin (Life Technologies, UK). To activate the Jurkat cells, cells were resuspended in 30 mM Hepes-buffered RPMI and prewarmed at 37 ◦ C for 10 min. Anti-mouse IgG (Sigma, UK) was added at a final concentration of 10 ␮g/ml and incubated with the cells

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

coupled to Aminolink Plus Coupling Gel (Pierce, UK). Following a 1 h incubation at room temperature with constant mixing, the immunoprecipitates were washed with lysis buffer five times. Bound proteins were eluted by heating the Coupling gel in the presence of SDS-sample loading buffer for 5 min at 100 ◦ C. To control for non-specific binding, lysate was incubated with Coupling gel in the absence of antibody. The MAPK members (ERK1, ERK2 and p38) were immunoprecipitated by mixing 3 ␮g of monoclonal antibody (ERK1 (Affinity Research Products, UK), ERK2 (Upstate Biotechnology, USA) or p38 (Upstate Biotechnology, USA)) with 40 ␮l 10% suspension of Protein A Sepharose (Sigma, UK) added to the cell lysate. Samples were separated on a 4–20% gradient SDS-PAGE and subjected to Western blotting using ERK1, ERK2 or p38 antibodies. Membranes were blocked in 3% milk in PBS (Sigma, UK) and probed for the respective antibody, followed by a goat anti-mouse horseradish peroxidase-conjugated antibody (Sigma, UK). Proteins were revealed by ECL as described above. Membranes to be reprobed were rehydrated in PBS 0.1% Tween-20 and stripped with Immuno Pure IgG Elution Buffer (Pierce, UK) for 1 h. The membranes were then washed with PBS 0.1% Tween-20 for 30 min. The membranes were blocked with 3% milk in PBS and probed for the presence of LCPTP (rabbit polyclonal antibody, GlaxoSmithKline Wellcome) followed by a goat anti-rabbit horseradish peroxidase-conjugated antibody (Sigma, UK). Proteins were detected as described above. Pull-downs with recombinant proteins were done by mixing either 3 ␮g of the GST fused protein (LCPTP wild-type or substrate-trap (C291S, D257A)) or the GST protein alone, coupled to Glutathione Sepharose 4B beads (Amersham Pharmacia Biotech, UK) with the cell lysate. The pull-downs were performed as described for the immunoprecipitations and the Western blotting was done as described above.

477

Kvanta et al., 1990, material and methods). After each treatment, cells were incubated for 2 min with anti-mouse IgG (10 ␮g/ml final concentration) followed by the addition of anti-CD3 (10 ␮g/ml final concentration) and incubated for 1 min at 37 ◦ C. cAMP measurements were done using a cAMP enzyme immunoassay (EIA) system (Amersham Pharmacia Biotech, UK) following the manufacturer’s instructions. 3. Results 3.1. Activation of Jurkat T cell line The majority of previous studies of LCPTP binding have used pervanadate to induce phosphorylation of multiple cell proteins, via inactivation of endogenous phosphatases. In order to use a more physiologically relevant stimulus, and also in order to allow study of interactions with endogenous LCPTP without interference by pervanadate (which modifies the catalytic sulfhydryl group at the active site), T cells were activated via cross-linking the CD3 complex with antibody.

2.5. Transient transfections A total of 2 × 107 Jurkat cells were resuspended in 0.3 ml RPMI 1640 medium (Life Technologies, UK), supplemented with 10% heat inactivated fetal calf serum, 2 mM l-glutamine, 100 IU/ml penicillin and 100 ␮g/ml streptomycin (Life Technologies, UK), in 0.4 cm electrocuvette (Bio-Rad, UK) incubated with 30 ␮g of vector pcDNA 3.1 (Invitrogen, UK) containing c-myc tag or plasmids pcDNA 3.1 c-myc-LCPTP WT or pcDNA 3.1 c-myc-LCPTP C291S-D257A for 5 min. Cells were electroporated at 260 V, 960 ◦ F, ∞  (Bio-Rad Gene Pulser, UK), washed, resuspended in complete RPMI medium and incubated at 37 ◦ C. 48 h post-transfection, LCPTP immunoprecipitations were carried out. 2.6. Intracellular cAMP measurement A total of 1 × 105 cells were stimulated in the presence, or absence of 0.05 mM forskolin for 7 min (following

Fig. 1. Time course of the activation of Jurkat T cells by anti-CD3 antibody. (A) Phosphotyrosine blot of cell lysates at different times. (B) Coomassie blue staining showing equivalent protein loading. Lanes 1–10 represent activation times of 0, 0.25, 0.5, 0.75, 1, 2, 5, 10, 20 and 30 min at 37 ◦ C, respectively. Each lane is the equivalent of 1 × 105 cells. The data shown are representative of four separate experiments.

478

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

The time point for peak protein tyrosine phosphorylation following receptor ligation of Jurkat cells was determined by time course studies. Cells were activated with an anti-CD3 antibody at a concentration which elicited the maximum phosphorylation signal as detected by Western blotting (data not shown). The time course shows a rapid rise in protein tyrosine phosphorylation peaking at 1 min (Fig. 1A, lanes 1–5) and a gradual reduction of tyrosine phosphorylation after this time point (Fig. 1A, lanes 5–10). To demonstrate equal protein loading, the gel was stained with Coomassie blue (Fig. 1B). Resting and cells activated maximally for 1 min (corresponding to the peak of protein phosphorylation) were used for subsequent immunoprecipitation experiments. The nature of the interaction between LCPTP and its potential substrates in resting or activated T cells was analysed in three different ways.

3.2. ERK1, ERK2 and p38 form a stable interaction with recombinant GST-LCPTP substrate-trap protein (D257A-C291S), but not with wild-type GST-LCPTP Lysates of resting and activated Jurkat T cells were incubated with immobilised GST-LCPTP, GST-LCPTP substrate trap (D257A-C291S), or GST alone (Fig. 2). Pull-down complexes were separated by SDS-PAGE and analysed by Western blot using monoclonal ERK1, ERK2 and p38 antibodies. As shown in Fig. 2A, lanes 1 and 2, ERK1 did not interact constitutively with GST-LCPTP wild-type protein. When the GST-LCPTP substrate-trap protein was used to immunoprecipitate from cell lysates (Fig. 2A, lanes 3 and 4), ERK1 was brought down from both the non-activated and anti-CD3 activated cell lysates, with a preference for the activated sample. ERK1 protein was not brought down by GST

Fig. 2. ERK1, ERK2 and p38 associate with recombinant GST-LCPTP substrate-trap protein in Jurkat T cells. Unstimulated or anti-CD3 antibody-stimulated Jurkat T cell lysates corresponding to 1 × 108 cells were used in pull-downs using recombinant GST-LCPTP wild-type and substrate-trap (C291S, D257A) protein. GST fusion proteins in the absence of lysates were also run to control for non-specific binding of detection reagents, but gave no detectable signal and are not shown. All panels: lanes 1 and 2 pull-downs with GST-LCPTP wild-type protein; lanes 3 and 4 pull-downs with GST-LCPTP substrate-trap (C291S D257A); lanes 5 and 6 pull-downs with GST-beads and lane 7 whole cell lysate equivalent to 1 × 106 cells. Alternate lanes show T cells either unstimulated or stimulated for 1 min with anti-CD3 as shown. Panel (A), blotted with anti-ERK1; panel (B), blotted with anti-ERK2; panel (C), blotted with anti-p38; and panel (D), blotted with anti-phosphotyrosine.

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

479

protein alone (lanes 5 and 6) but could be readily detected in the total cell lysate control (lane 7). The same pattern was observed with ERK2. ERK2 interacted constitutively with GST-LCPTP substrate-trap protein, although in contrast to ERK1, more ERK2 was captured from the non-activated lysate compared to the activated lysate (Fig 2B, lanes 3 and 4). The presence of p38 protein was also detected in pull-downs using GST-LCPTP substrate-trap protein but the amount of p38 associating with LCPTP appeared to be much less compared to that seen with ERK1 and ERK2 (Fig 2C, lanes 3 and 4). Similarly to ERK2, more p38 bound to the phosphatase-trap when the cells were not activated. A phosphotyrosine protein blot revealed that there was an overall increase in the phosphorylation state of the MAPK family members associated with LCPTP substrate trap following CD3 activation, suggesting that ERK1 is the major phosphorylated form in activated cells (Fig 2D, lanes 3 and 4). 3.3. ERK1 is found in immunoprecipitates of overexpressed LCPTP In a second set of experiments, we overexpressed LCPTP in Jurkat cells and analysed which proteins could be isolated bound to LCPTP under these conditions. Jurkat T cells were transiently transfected with c-myc, c-myc-LCPTP WT or c-myc-LCPTP substrate-trap mutant proteins. Unstimulated cells and cells activated for 1 min by CD3 ligation, corresponding to the time point at which maximum total protein tyrosine phosphorylation was seen, were used for LCPTP immunoprecipitations. Immunoprecipitations were probed for the presence of the MAP kinase ERK1 (ERK2 was not probed in this case). ERK1 was not brought down when the myc tag protein alone was overexpressed (Fig. 3A and B lanes 1 and 2). A very low level of ERK1 was detected, however, associated with both LCPTP wild-type and substrate-trap proteins (Fig. 3A and B lanes 3–6). This small level of association was not seen when doing pull-downs with GST-fusions from cell lysates (Fig. 2), and perhaps reflects a subtle difference in folding between overexpressed endogenously synthesised c-myc-LCPTP shown in Fig. 3, and the bacterially expressed GST-LCPTP preparations used in Fig. 2. Equal levels of ERK1 were immunopreciptated from resting and anti-CD3 stimulated cell lysates. As expected, LCPTP-myc (in amounts much larger than the amounts from endogenous LCPTP) was present in the myc immunoprecipitates of the transfectants (Fig 3B, lanes 3–6). 3.4. MAP kinases ERK1, ERK2 and p38 are not found in immunoprecipitates of endogenous LCPTP in Jurkat T cells The experiments outlined above all used a large excess of LCPTP to try and detect the presence of bound partner proteins. In order to address the physiological significance of these findings, we performed experiments in which

Fig. 3. ERK1 is found in immunoprecipitates of overexpressed LCPTP. Lanes 1–6. Jurkat T cells were transfected with myc-tag alone (lanes 1 and 2), myc-LCPTP wildtype (lanes 3 and 4), myc-LCPTP substrate-trap (C291S;D257A, lanes 5 and 6). Alternate lanes contain either unstimulated T cells, or cells stimulated with anti-CD3 (1 min), as shown. Lysates from all groups (107 cells) were immunoprecipitated with anti-LCPTP antibody and blotted with anti-ERK1 (top panel A) or anti-LCPTP (lower panel B). Lane 7 contains whole lysate from untransfected Jurkat cells (106 cells equivalent) probed directly without immunoprecipitation. The data shown are representative of two separate experiments.

endogenous LCPTP was immunoprecipitated and protein complexes were probed with antibodies raised against various MAP kinase family members. As shown in Fig. 4A, ERK1 was not detected in LCPTP immunoprecipitates (lanes 1 and 2), but was readily detected in the total cell lysate control (lane 5), containing 100 times less total protein than the IPs. A similar pattern was seen when LCPTP IPs were probed for ERK2 protein and p38 (Fig. 4B and C, respectively). Some p38 was detected binding non-specifically to coupling gel (Fig 4C, lane 3). In order to confirm the presence of LCPTP in the immunoprecipitates, membranes were stripped and reprobed for the presence of LCPTP. Fig. 4 shows one representative example. LCPTP was brought down by the coupled antibody and the amount was enriched in the immunoprecipitation compared to the total

480

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

Fig. 4. LCPTP immunoprecipitates from Jurkat T cells do not contain ERK1, ERK2 or p38. Jurkat cells lysates (1 × 108 ) were immunoprecipitated with anti-LCPTP rabbit antiserum (lanes 1 and 2) or coupling gel alone (lanes 3 and 4). Alternate lanes contain lysates from unstimulated or stimulated (anti-CD3, 1 min) Jurkat cells as shown. Lane 5 contains whole lysate from unstimulated 106 cells. The gels were blotted for ERK1 (panel A), ERK2 (panel B), p38 (panel C) or LCPTP (panel D, using same rabbit antiserum as used for immunoprecipitates). The data shown are representative of two separate experiments.

lysate (Fig. 4D, lanes 1 and 2 and 5); LCPTP did not interact non-specifically with the coupling gel (lanes 3 and 4). 3.5. Endogenous LCPTP is not found in ERK1, ERK2 or p38 immunoprecipitates

the presence of each MAPK family member in the appropriate immunoprecipitation and in the total cell lysate (Fig. 5B, lanes 1–6 and 10) but not in control immunoprecipitations (lanes 7 and 8). 3.6. Intracellular cAMP measurement

In a further experiment to look for a constitutive interaction of endogenous LCPTP with MAPK members, the reverse immunoprecipitations were carried out. LCPTP was not detected in MAPK immunoprecipitations (Fig. 5A, lanes 1–6), or control precipitations (lanes 7 and 8), but was seen in the total cell lysate (lane 10), containing 100 times less total protein than the IPs. To confirm the presence of the respective MAP kinases in the immunoprecipitations, the membranes were stripped and reprobed using ERK1, ERK2 and p38 monoclonal antibodies simultaneously. Fig. 5B shows

Previous studies have suggested that cAMP regulates LCPTP interactions and that increased levels of cAMP resulted in the dissociation of MAP kinases and LCPTP complexes (Saxena et al., 1999b). High endogenous levels of cAMP in Jurkat T cells could therefore explain why we cannot demonstrate the interaction between endogenous LCPTP and MAP kinases. To address whether Jurkat T cells have elevated cAMP levels thus favouring release of LCPTP from MAP kinase complexes, the levels of cAMP in

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

481

Fig. 6. Intracellular cAMP measurement in Jurkat T cells. Intracellular cAMP from resting (empty bars) or activated (1 min, anti-CD3, filled bars) Jurkat T cells was measured. cAMP levels are compared between cells non-stimulated (left) or stimulated (right) with forskolin (0.05 mM). The amount of cAMP is measured in fmol and the results are the mean of two experiments.

4. Discussion

Fig. 5. MAPK immunoprecipitates do not contain LCPTP. Jurkat T cell lysates (108 cells) were immunoprecipitated with anti-ERK1 (lanes 1 and 2), anti-ERK2 (lanes 3 and 4), anti-p38 (lanes 5 and 6), or Protein A-Sepharose beads without antibody (lane 7 and 8). Alternate lanes contain lysates from unstimulated or stimulated (anti-CD3, 1 min) Jurkat cells as shown. Lane 9 contains material eluted from protein A-Sepharose beads and anti-p38 antibody, but no cell lysate. Lane 10 contains whole cell lysate of 106 activated Jurkat cells. Gels were blotted either for LCPTP (panel A above), or a combination of anti-ERK1, anti-ERK2 and anti-p38 (panel B below). The data shown are representative of two separate experiments.

non-activated and anti-CD3 stimulated cells were measured. As a positive control, cells were activated in the presence or absence of forskolin which induces raised cAMP levels. The difference in the concentration of cAMP in resting cells and those activated with an anti-CD3 antibody for 1 min appeared negligible (Fig. 6). In contrast, cAMP levels were increased 10,000 fold after forskolin stimulation.

Following recognition by the TCR of an antigenic peptide presented by MHC molecules, a cascade of signal transduction events occurs. The early stages of this cascade are represented by a dramatic but transient increase in the total protein tyrosine phosphorylation. The Jurkat T lymphocyte cell line, activated with an anti-CD3 antibody, has been used extensively as a model system for studying TCR signal transduction (Landegren et al., 1985). This provided us with a good model system for looking at LCPTP function, since Jurkat T cells also express high levels of endogenous LCPTP. Some previous studies (Oh-hora et al., 1999; Saxena et al., 1999a), but not others (Pettiford and Herbst, 2000) have suggested a model in which members of the MAP kinase families (ERK1, ERK2 and p38) interact constitutively with recombinant LCPTP. Both wild-type LCPTP protein and an LCPTP substrate-trap (C200S) protein formed complexes with ERK1, ERK2 and p38 (Model 1) (Oh-hora et al., 1999; Saxena et al., 1999a), although no information was given on the stoichiometry of this interaction. We were not able to demonstrate any constitutive interaction with members of the ERK and p38 MAP kinase families using either pull-downs with immobilised LCPTP, or endogenous LCPTP. The only condition in which we were able to detect a small degree of association, was in immunoprecipitates of LCPTP, when this was highly overexpressed by transient transfection (Oh-hora et al., 1999; Saxena et al., 1999a). Even under these conditions, only a very small percentage of total LCPTP protein associates with the MAP kinases.

482

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

Our data therefore, provides support for a more recent second model in which LCPTP interacts with MAP kinase predominantly in the context of an enzyme/substrate interaction (Pettiford and Herbst, 2000). This interaction could be readily demonstrated in this study, using substrate-trap mutants of the phosphatase (C200S), which interact with phosphorylated ERK, but cannot hydrolyse the phosphate group, and do not therefore, release the substrate. In contrast to the results of Pettiford and Herbst who demonstrated interaction only with ERK2, both ERK1 and ERK2 bound strongly to LCPTP substrate trap, while p38 showed a much weaker binding. Interestingly, LCPTP substrate-trap interacted with ERK2 and p38 better in resting cells than activated cells, but interacted with ERK1 much more in activated cells. These subtle differences require further study. One may speculate that they may reflect the different proportions of MAP kinase found in fully phosphorylated (active) form in unstimulated and stimulated cells, although no previous reports of such differences exist. Alternatively, it may reflect other modifications of the substrate proteins occurring after activation which alter their ability to interact with LCPTP. Such modifications, which may be cell specific, may also explain the discrepancy between our study and that of Pettiford and Herbst. In the light of the contradictory reports in the literature, it was important to consider the several reasons which could explain our inability to detect a constitutive interaction between endogenous LCPTP and the MAP kinase family members ERK1, ERK2 or p38. Insufficient sensitivity of detection does not seem likely as the amount of total protein present in each immunoprecipitation was 100 times higher compared to protein visualised by blotting (LCPTP or MAP kinase) in the total lysate. Since the MAP kinase members in the total cell lysate are still clearly detectable, the experimental system should detect interactions between LCPTP and MAP kinases unless the percentage of total LCPTP associating with MAP kinases is extremely low. A previous report suggested that elevated intracellular cAMP levels inhibit the interaction of LCPTP with ERK, by increasing the phosphorylation of LCPTP on Ser-34 and thereby releasing ERK from its association with the phosphatase (Saxena et al., 1999b). We therefore, measured the amount of intracellular cAMP in the Jurkat T cells used for immunoprecipitating LCPTP to see if the absence of interaction could be explained by the fact that the cells contained high levels of cAMP preventing this association. The amount of 8-CPT-cAMP added in the aforementioned study, to measure the release of ERK from recombinant LCPTP, was 1 mM (Saxena et al., 1999b). The amount of cAMP detected in our system was in the range of 1 fmol of cAMP per 105 cells. Adjusting the amount of cAMP for 108 cells, immunoprecipitations were performed in the presence of 1 pmol of cAMP per IP. This is 109 times less cAMP per IP than in the experiments described. The amount of cAMP is therefore, unlikely to be a significant factor inhibiting endogenous LCPTP from interacting with ERK in the cell.

A third explanation for the negative results obtained could be differences in lysis conditions. However, immunoprecipitations of the endogenous LCPTP were repeated using exactly the same experimental conditions as in previous studies (Oh-hora et al., 1999) but still did not reveal any MAP kinase association (data not shown). Furthermore, the lysis conditions used in this study did allow the interaction of ERK1, ERK2 and p38 proteins with a recombinant GST-LCPTP substrate-trap protein in Jurkat T cells. The only condition under which stable, and activationindependent interactions were observed, as mentioned above, was in Jurkat cells transiently transfected, and hence, overexpressing LCPTP. The high levels of enzyme in the cell, relative to the normal endogenous levels, may itself allow very weak associations to be detected as in our own (Fig. 3) and previous studies (Oh-hora et al., 1999; Saxena et al., 1999a). Alternatively, under these conditions of LCPTP excess, the ERK detected may represent the small proportion which is being dephosphorylated by LCPTP at the time the cells were lysed. A third possibility is that high levels of transient expression may result in expression of a proportion of misfolded and inactive enzyme, which may promote a low level of constitutive binding. We cannot, of course, completely rule out the possibility that the lysate conditions used in our studies destroy a weak constitutive interaction between LCPTP and MAP kinases. However, if this is the case, this applies also to previous studies. We believe it is more likely, taking our results together with the previous published data, that the main element which allows LCPTP to bind ERK1 and 2 (and to a lesser extent perhaps p38), is in the context of an enzyme/substrate interaction. Interaction between substrate and enzyme may also involve regions outside the active site (as suggested previously), but these alone are unlikely to lead to formation of a stable complex after catalysis has taken place. Finally, our study provides a cautionary note to the interpretation of protein/protein interactions seen under conditions when one partner is expressed at levels many times higher than is normally present. Such interactions have been documented for many other proteins, including other phosphatases including PTPBR7 which forms stable complexes via KIM motifs with ERK1 and ERK2 both in its wild-type or substrate-trap conformation (Haneda et al., 1999; Ogata et al., 1999), and PTP-SL, STEP protein phosphatase and the Drosophila PTP-ER protein phosphatase (Blanco-Aparicio et al., 1999; Karim and Rubin, 1999). More sophisticated approaches which allow measurement of such interactions under more physiological conditions will be important in determining the contribution of such interactions to the regulation of cellular activity in vivo.

Acknowledgements We wish to thank Neil T. Thompson (GSK, Stevenage) and Jean-Yves Masson (ICRF, Clare Hall Laboratories) for

I. Brodeur et al. / Molecular Immunology 39 (2002) 475–483

their useful comments and manuscript reading. I.B. is supported by a PhD studentship from Fond de la Recherche en Santé du Québec and Fond pour la Formation de Chercheurs et l’Aide à la Recherche (FRSQ-FCAR) and from Glaxo Wellcome Research and Development. References Adachi, M., Sekiya, M., Isobe, M., Kumura, Y., Ogita, Z., Hinoda, Y., Imai, K., Yachi, A., 1992. Molecular cloning and chromosomal mapping of a human protein-tyrosine phosphatase LCPTP. Biochem. Biophys. Res. Commun. 186, 1607–1615. Blanco-Aparicio, C., Torres, J., Pulido, R., 1999. A novel regulatory mechanism of MAP kinases activation and nuclear translocation mediated by PKA and the PTP-SL tyrosine phosphatase. J. Cell Biol. 147, 1129–1136. Haneda, M., Sugimoto, T., Kikkawa, R., 1999. Mitogen-activated protein kinase phosphatase: a negative regulator of the mitogen-activated protein kinase cascade. Eur. J. Pharmacol. 365, 1–7. Isakov, N., Wange, R.L., Samelson, L.E., 1994. The role of tyrosine kinases and phosphotyrosine-containing recognition motifs in regulation of the T cell-antigen receptor-mediated signal transduction pathway. J. Leukocyte Biol. 55, 265–271. Karim, F.D., Rubin, G.M., 1999. PTP-ER, a novel tyrosine phosphatase, functions downstream of Ras1 to downregulate MAP kinase during Drosophila eye development. Mol. Cell 3, 741–750. Kvanta, A., Gerwins, P., Jondal, M., Fredholm, B.B., 1990. Stimulation of T cells with OKT3 antibodies increases forskolin binding and cyclic AMP accumulation. Cell Signal 2, 461–470.

483

Landegren, U., Andersson, J., Wigzell, H., 1985. Analysis of human T lymphocyte activation in a T cell tumor model system. Eur. J. Immunol. 15, 308–311. Mustelin, T., Brockdorff, J., Gjorloff-Wingren, A., Tailor, P., Han, S., Wang, X., Saxena, M., 1998. Lymphocyte activation: the coming of the protein tyrosine phosphatases. Front. Biosci. 3, D1060–D1096. Ogata, M., Oh-hora, M., Kosugi, A., Hamaoka, T., 1999. Inactivation of mitogen-activated protein kinases by a mammalian tyrosine-specific phosphatase, PTPBR7. Biochem. Biophys. Res. Commun. 256, 52–56. Oh-hora, M., Ogata, M., Mori, Y., Adachi, M., Imai, K., Kosugi, A., Hamaoka, T., 1999. Direct suppression of TCR-mediated activation of extracellular signal-regulated kinase by leukocyte protein tyrosine phosphatase, a tyrosine-specific phosphatase. J. Immunol. 163, 1282–1288. Pettiford, S.M., Herbst, R., 2000. The MAP-kinase ERK2 is a specific substrate of the protein tyrosine phosphatase HePTP. Oncogene 19, 858–869. Pulido, R., Zuniga, A., Ullrich, A., 1998. PTP-SL and STEP protein tyrosine phosphatases regulate the activation of the extracellular signalregulated kinases ERK1 and ERK2 by association through a kinase interaction motif. EMBO J. 17, 7337–7350. Saxena, M., Williams, S., Brockdorff, J., Gilman, J., Mustelin, T., 1999a. Inhibition of T cell signaling by mitogen-activated protein kinasetargeted hematopoietic tyrosine phosphatase (HePTP). J. Biol. Chem. 274, 11693–11700. Saxena, M., Williams, S., Tasken, K., Mustelin, T., 1999b. Crosstalk between cAMP-dependent kinase and MAP kinase through a protein tyrosine phosphatase. Nat. Cell Biol. 1, 305–311. Zanke, B., Suzuki, H., Kishihara, K., Mizzen, L., Minden, M., Pawson, A., Mak, T.W., 1992. Cloning and expression of an inducible lymphoidspecific, protein tyrosine phosphatase (HePTPase). Eur. J. Immunol. 22, 235–239.