Effects of PEA-15 phosphorylation on cellular functions. (A) Unphosphorylated PEA-15 binds activated ERK1/2 and blocks cell proliferation and reverses integrin suppression. Unphosphorylated PEA-15 binds activated ERK1/2 and prevents their nuclear accumulation, thereby preventing cell proliferation. Additionally, PEA-15 reverses activated ERK1/2-mediated integrin suppression. (B) Phosphorylated PEA-15 does not block cell proliferation or reverse integrin suppression, but it is recruited to the DISC to block apoptosis. Kinases such as PKC, CamKII, or Akt can phosphorylate PEA-15. Phosphorylated PEA-15 does not bind activated ERK1/2 and therefore has no effect on cell proliferation or integrin activation. In contrast, phosphorylated PEA-15 is recruited to the DISC, where it can block TNF receptor- or Fas receptor-mediated apoptosis.

PEA-15 phosphorylation prevents inhibition of nuclear translocation of ERK1/2. (A) Phosphomimetic mutations (S104D, and S116D) in PEA-15 do not block nuclear import of ERK2. CHO cells were cotransfected with HA tagged PEA-15 wild-type or mutants and HA-ERK2, grown in 10% serum for 24 h, followed by serum-free media for 24 h. The cells were then stimulated with 10% serum for 3 h, disrupted by osmotic shock, and fractionated into nuclear and cytoplasmic fractions. Each fraction or the total cell lysate was separated by SDS-PAGE and immunoblotted with anti-HA to detect ERK2 and PEA-15. Wild-type PEA-15 blocked nuclear accumulation of ERK2, whereas phosphomimetic PEA-15 (S104D and S116D) did not (top). Whole cell lysates also were assayed for HA-ERK and PEA-15 expression (middle and bottom). (B) Validation of subcellular fractionation. Whole cell lysates or nuclear and cytosolic fraction from A were separated by SDS-PAGE and immunoblotted with antibodies reactive for the cytosolic marker RhoGDI (bottom left) and the nuclear marker lamin A/C (bottom right).

PEA-15 phosphomimetic mutants do not block ERK-dependent transcription. (A) Phosphomimetic mutations (S104D and S116D) in PEA-15 do not block SRE luciferase reporter activity. CHO cells were cotransfected with PEA-15 or the indicated mutants, a SRE-dependent luciferase reporter construct (pSRE-Luc), and a constitutive Renilla reporter (pRL-TK). Cells were grown in 10% serum for 24 h, serum-free media for 24 h, and then stimulated with 10% serum for 3 h. ERK nuclear activity was assayed as a function of SRE luciferase reporter activity. SRE reporter activity was corrected for transfection efficiency by the cotransfected Renilla reporter activity. Results are reported as the percentage of increase in SRE reporter activity compared with vector-transfected serum-starved cells. Wild-type PEA-15 and both nonphosphorylatable mutants (S0104A, S116A) markedly suppressed SRE-dependent transcription. In contrast, the phosphomimetic PEA-15 mutants (S104D, S116D) did not inhibit SRE reporter activity. Data depicted are the mean ± SE of four determinations. (B) Expression of transfected PEA-15. Lysates of cells described in A were fractionated by SDS-PAGE and immunoblotted with anti-PEA-15 antibody to detect transfected PEA-15.

Phosphomimetic PEA-15 mutations block its association with ERK2 in cells. CHO cells were cotransfected with cDNAs encoding FLAG-tagged ERK2, and HA-tagged wild-type PEA-15 or the indicated PEA-15 mutants. Lysates of these cells were immunoprecipitated with anti-HA antibody, and the immunoprecipitates were fractionated by SDS-PAGE and immunoblotted with anti-FLAG antibody to detect coprecipitated ERK2. These immunoprecipitates also were blotted with anti-PEA-15 to verify consistent immunoprecipitation. The whole cell lysates were blotted with anti-FLAG to verify expression of the recombinant ERK2. The blots were quantified by densitometric scanning and the quantity of coprecipitated ERK2 was divided by ERK2 precipitated with wild-type PEA-15 to obtain the indicated ratios. (A) Phosphomimetic mutations (S104D and S116D) in PEA-15 reduce interaction with ERK2. Note that the S116D mutant blocked coprecipitation of ERK2 with PEA-15 to a greater extent than the L123R mutant, previously reported to disrupt the PEA-15-ERK interaction (Hill et al., 2002). The S104D mutant also inhibited the PEA-15–ERK2 interaction but to a lesser extent than either of the other two mutants. (B) Phosphorylation of both Ser¹⁰⁴ and Ser¹¹⁶ is required to block interaction with ERK1/2. The free Ser on each Asp mutant was mutated to a nonphosphorylatable Ala, creating double mutants PEA-15(S104A, S116D) and PEA-15(S104D, S116A). Each of these double mutants restored the interaction of PEA-15 with ERK1/2.

PEA-15 phosphorylation prevents its inhibition of cell proliferation. (A) Nonphosphorylatable PEA-15 mutants suppress cell proliferation more than wild type; phosphomimetic mutants fail to suppress. CHO cells were transfected with cDNAs encoding HA-PEA-15 or the indicated PEA-15 mutants. After 24 h in growth media, the cells were placed in 0.5% serum for an additional 24 h. BrdU (10 μM) was added to the medium, followed by 10% serum to stimulate cell proliferation. After 1 h, the cells were harvested, and BrdU incorporation was assayed by staining with anti-BrdU and quantified using flow cytometry. Data were acquired as geometric mean fluorescence intensity of BrdU staining and are expressed as percentage of BrdU incorporation, where 100% is basal BrdU incorporation of vector-transfected serum-starved cells. Transfection with PEA-15 blocked serum stimulation of proliferation. In contrast, both phosphomimetic mutants (S104D, S116D) did not block cell proliferation. Both nonphosphorylatable mutants (S104A, S116A) reduced proliferation less than that seen in the absence of serum. (B) Expression of transfected PEA-15 variants. Lysates of the cells described in A were fractionated by SDS-PAGE and immunoblotted with PEA-15 antibody to detect transfected PEA-15.

Phosphomimetic PEA-15 mutants do not reverse integrin suppression. (A) Effect of PEA-15 wild-type and mutants on Raf-CAAX–induced integrin suppression. CHO cells were cotransfected with cDNA encoding a constitutively active Raf kinase (Raf-CAAX) in combination with PEA-15 or the indicated mutants. After 24 h, the cells were detached, and the binding of soluble cell binding domain of Fn [3Fn(9-11)] binding was measured to assess the affinity state of integrin α5β1. This was quantified as an AI, where AI = 100 * (F –Fo)/(Fm –Fo). F represents the geometric mean fluorescence (GMF) of 3Fn-(9-11) binding alone, Fo is the GMF of 3Fn-(9-11) binding in the presence of 10 mM EDTA, and Fm is the GMF of 3Fn-(9-11) binding in the presence of 9EG7. The percentage of suppression was calculated as 100 * (AIM –AIT)/AIM, where AIM is the activation index of the vector transfected cells, and AIT is the activation index in the presence of a Raf-CAAX or Raf-CAXX + PEA-15 WT or mutants. The percentage of reversal of suppression was calculated as 100 * (AI –AIR)/AI, in which AI is the activation index of the control cells, and AIR is the activation index in the presence of a transfected PEA-15 WT or mutants. Wild-type PEA-15 reversed Raf-CAAX–induced integrin suppression >50%, whereas PEA-15 phosphomimetic mutants (S104D, S116D) failed to do so at all. (B) Expression of transfected PEA-15 variants. Lysates of cells described in A were fractionated by SDS-PAGE and immunoblotted with PEA-15 antibody to detect transfected PEA-15.

PEA-15 is phosphorylated in cells and tissues. (A) Phospho-specific anti-PEA-15 (anti-p-PEA-15) is PEA-15 specific. CHO cells were transfected with cDNAs encoding PEA-15 or the indicated mutants. After 48 h, the cells were lysed, fractionated by SDS-PAGE, and analyzed by immunoblotting with either anti-phospho-PEA-15 (p-PEA-15) or an anti-PEA-15 antibody (PEA-15) that reacts with both phosphorylated and unphosphorylated PEA-15. Note that both antibodies reacted with a polypeptide of molecular weight of ∼15 k and no other cellular polypeptides. This polypeptide was absent from immunoblots of untransfected CHO cells (our unpublished data). Mutation of Ser¹¹⁶ to Ala abolished reactivity of the p-PEA-15 antibody. (B) Anti-p-PEA-15 is phospho-specific. Purified recombinant PEA-15 or the indicated mutants were phosphorylated in vitro with purified CamKII, fractionated by SDS-PAGE, and immunoblotted with anti-p-PEA-15 or anti-PEA-15. Anti-p-PEA-15 reacted with phosphorylated but not unphosphorylated PEA-15. Reactivity was abolished by mutating the CamKII phosphorylation site, Ser¹¹⁶, to Ala. (C) Phosphorylation of PEA-15 in cultured cells. The indicated tumor cell lines were lysed in sample buffer, fractionated by SDS-PAGE, and immunoblotted with either anti-PEA-15 or anti-p-PEA-15. Note that PEA-15 is expressed and phosphorylated in both cell lines. (D) PEA-15 is phosphorylated in situ in normal breast epithelium. Sections of normal mammary tissue were immuno peroxidase stained for PEA-15 (left) or phospho-PEA-15 (middle) and counterstained with Mayer's hematoxylin. Preimmune rabbit IgG was used as a staining control (right).

Phosphorylation of PEA-15 blocks binding to ERK1/2. Recombinant GST-PEA-15 was immobilized on glutathione-Sepharose beads and incubated with lysates of NIH 3T3 cells in kinase active buffer containing 10 μCi of [γ-³²P]ATP in the presence or absence of 50 mg/ml phosphatidyl serine and 100 nM phorbol myristate acetate (lipids) at 37°C for 2 h. The beads were washed, and bound proteins were eluted in sample buffer, separated by SDS-PAGE, and immunoblotted with anti-ERK1/2 to assess ERK binding and with anti-PEA-15 to assess the retention of PEA-15 on the beads. Incorporation of ³²P into PEA-15 was detected by autoradiography of the dried blots. The ERK1/2 blots were quantified by densitometric scanning, and the ERK bound was divided by that bound to wild-type PEA-15 to obtain the indicated ratios. (A) Phosphorylation of wild-type PEA-15 in the presence of lipid activators blocked the interaction of ERK1/2 with PEA-15. Treatment with PKC (1 μM Bis) or CamKII (10 μM KN-62) inhibitors blocked phosphorylation of wild-type PEA-15 and restored the interaction of ERK1/2 with PEA-15. (B) Mutation of PEA-15 Ser¹⁰⁴ or Ser¹¹⁶ phosphorylation sites to Ala blocked phosphorylation and preserved the ability of PEA-15 to bind ERK1/2. (C) Phosphorylation of PEA-15 is necessary but not sufficient to inhibit ERK1/2 binding. Recombinant GST-PEA-15 was incubated with lysates of NIH 3T3 cells in the presence of absence of lipid activators, as described in A. Lysate and PEA-15 were then separated by centrifugation and subsequently recombined for binding assays. When lipid activator-treated lysate was added to modified PEA-15 (L*+P*) ERK1/2 binding was strongly inhibited. In contrast, combinations in which either the lysate or the PEA-15 had been incubated in the absence of activators (L*+P, L+P*) bound ERK1/2 to the same extent as combinations in which no lipid activator was added (L+P).