Phosphorylation of Bcl-associated death protein (Bad) by erythropoietin-activated c-Jun N-terminal protein kinase 1 contributes to survival of erythropoietin-dependent cells

Abstract

The glycoprotein erythropoietin (Epo) is a hematopoietic cytokine necessary for the survival of erythrocytes from immature erythroid cells. The mitogen-activated c-Jun N-terminal kinase 1 (JNK1) plays an important role in the proliferation and survival of erythroid cells in response to Epo. However, the precise mechanism of JNK1 activation promoting erythroid cell survival is incompletely understood. Here, we reported that JNK1 is required for Epo-mediated cell survival through phosphorylation and inactivation of the pro-apoptotic, Bcl-2 homology domain 3 (BH3)-only Bcl-associated death protein (Bad). Upon Epo withdrawal, HCD57 cells, a murine Epo-dependent cell line, displayed increased apoptotic cell death that was associated with decreased JNK1 activity. Epo withdrawal-induced apoptosis was promoted by inhibition of JNK1 activity but suppressed by expression of a constitutively active JNK1. Furthermore, Epo-activated JNK1 phosphorylated Bad at threonine 201, thereby inhibiting the association of Bad with the anti-apoptotic molecule B-cell lymphoma-extra large (Bcl-X(L)). Replacement of threonine 201 by alanine in Bad promoted Epo withdrawal-induced apoptosis. Thus, our results provide a molecular mechanism by which JNK1 contributes to the survival of erythroid cells.

Fig. 1. JNK1 is an Epo-activated protein kinase

A, HCD57 cells were deprived of Epo for 18 hrs (−Epo), followed by Epo readdition (+Epo). Apoptotic cell death was measured with PI staining. The results are presented as means ± standard error and represent three individual experiments (*p<0.01 compared to 0 h control; #p<0.05 compared to 0 h control). B, HCD57 cells were deprived of Epo for the indicated times (−Epo) and the apoptotic cell death was measured. The specific caspase inhibitor Z-VAD-FMK was added 30 min prior to Epo withdraw (*p<0.01 compared to Z-VAD-FMK untreated control). C, HCD57 cells were deprived of Epo for indicated times (−Epo), and the JNK1 activity was measured by immune complex kinase assays (KA) using GST-c-Jun as a substrate and the expression level of JNK was detected by immunoblotting (IB) using anti-JNK antibody. D, HCD57 cells were deprived of Epo for 18 hrs (−Epo), followed by Epo readdition (+Epo, 10 U/ml) for 5–120 min, and the JNK1 activity and expression level of JNK were measured as described in Fig. 1C.

Fig. 2. JNK1 activation is required for preventing Epo withdrawal-induced apoptosis

A, HCD57 cells were deprived of Epo for 18 hrs (−Epo), followed by Epo readdition (+Epo, 10U/ml) for 5 or 15 min. The specific JNK inhibitor SP600125 (10 μM) was added 30 min prior to Epo readdition. JNK1 activity and expression level of JNK were measured as described in Fig. 1C. B and C, HCD57 cells were deprived of Epo for 18 hrs, followed by Epo readdition (+Epo, B), or left untreated (−Epo, C) in the presence of SP600125 (5, 10, 15 μm, respectively). Apoptotic cell death was measured as described in Fig. 1A (#, p<0.05 compared to SP600125 untreated control; *, p<0.01 compared to SP600125 untreated control). D, HA-MKK7-JNK1 and MKK7 (KM)-JNK1 were immunoprecipitated with anti-HA antibody from HDC57 cells that stably express HA-MKK7-JNK1 or HA-MKK7 (KM)-JNK1, respectively. The activity and expression level of HA-MKK7-JNK1 or HA-MKK7 (KM)-JNK1 were measured as described (Yu et al., 2004). E, HCD57 cells stably express HA-MKK7-JNK1 or HA-MKK7 (KM)-JNK1 were deprived of Epo (−Epo) for the time indicated. Apoptotic cell death was analyzed as described in Fig. 1A (*p<0.01 compared to 0 h control).

Fig. 3. JNK1 is an Epo-activated Bad kinase

HCD57 cells were deprived of Epo for 6 or 18 hrs, followed by Epo readdition (10 U/ml) for 15 or 30 min. A, Ser136 phosphorylation and expression of Bad were detected by immunoblotting (IB) using anti-phospho-Bad-Ser136 antibody (upper panel) or anti-Bad antibody (lower panel). B, Phosphorylation of GST-Bad (upper panel) or GST-c-Jun (middle panel) by Epo-activated JNK1 was measured by immune complex kinase assays (KA). C. Phosphorylation of GST-Bad or GST-c-Jun by activated recombinant GST-JNK1 in vitro. Inactive GST-JNK1 (APF) was used as a control.

Fig. 4. Epo-activated JNK1 phosphorylation of Bad at Thr201

A, Purified wt GST-Bad and GST-Bad (T201A) mutant were phosphorylated by Epo-activated JNK1 in the presence of nonradioactive ATP (17 μM) and analyzed by immunoblotting using anti-phospho-Thr201 antibody (upper panel). The amount of GST-Bad and GST-Bad (T201A) mutant was examined by immunoblotting (IB) using anti-Bad antibody (lower panel). B, HCD57 cells were transfected with expression vector encoding M2-Bad along with HA-MKK7-JNK1, HA-MKK7 (K149M)-JNK1, or empty vector. After 24 hrs, cells were deprived of Epo for 18 hrs. Thr201 phosphorylation and expression of Bad were analyzed as described in Fig. 4A. C, HCD57 cells were deprived of Epo for 18hrs (−Epo), followed by Epo readdition (+Epo, 10 U/ml) for 15 min. Activation of JNK1 was determined by immune complex kinase assay (KA, upper panel). Thr201 phosphorylation and expression of Bad were analyzed as described in Fig. 4A. D, wt GST-Bad and GST-Bad (T201A) mutant proteins modified with nonradioactive ATP were mixed with in vitro translated ³⁵S-labeled Bcl-X_L for 12 hrs and extensively washed. GST-Bad-associated ³⁵S-Bcl-X_L was visualized by radio autography. One-tenth of the input ³⁵S-Bcl-X_L was analyzed by radio autography. The amount of GST-Bad proteins was examined by CBB staining. E, Immunoblotting (IB) analysis of Bad (wt or T201A) in cell extracts from corresponding stable cell lines with anti-M2 antibody. F, HCD57 cells stably express Bad (T201A) mutant were more sensitive to Epo withdrawal-induced apoptosis. Apoptotic cell death was analyzed as described in Fig. 1(*p<0.01 compared to 0 h control).