Phosphoproteomic analysis identifies the tumor suppressor PDCD4 as a RSK substrate negatively regulated by 14-3-3

Abstract

The Ras/MAPK signaling cascade regulates various biological functions, including cell growth and proliferation. As such, this pathway is frequently deregulated in several types of cancer, including most cases of melanoma. RSK (p90 ribosomal S6 kinase) is a MAPK-activated protein kinase required for melanoma growth and proliferation, but relatively little is known about its exact function and the nature of its substrates. Herein, we used a quantitative phosphoproteomics approach to define the signaling networks regulated by RSK in melanoma. To more accurately predict direct phosphorylation substrates, we defined the RSK consensus phosphorylation motif and found significant overlap with the binding consensus of 14-3-3 proteins. We thus characterized the phospho-dependent 14-3-3 interactome in melanoma cells and found that a large proportion of 14-3-3 binding proteins are also potential RSK substrates. Our results show that RSK phosphorylates the tumor suppressor PDCD4 (programmed cell death protein 4) on two serine residues (Ser76 and Ser457) that regulate its subcellular localization and interaction with 14-3-3 proteins. We found that 14-3-3 binding promotes PDCD4 degradation, suggesting an important role for RSK in the inactivation of PDCD4 in melanoma. In addition to this tumor suppressor, our results suggest the involvement of RSK in a vast array of unexplored biological functions with relevance in oncogenesis.

Fig. 1.

Proteomic strategy for the characterization of the RSK-dependent phosphoproteome. (A) Schematic representation of the agonists and pharmacological inhibitors used in this study. (B) HEK293 and A375 cells were serum-starved for 24 h before incubation with PD184352 (10 µM) or BI-D1870 (10 µM) for 30 min in HEK293 cells and 2 h in A375 cells, respectively. HEK293 cells were stimulated with PMA (50 ng/mL) for 30 min or left unstimulated. Protein lysates were resolved by SDS/PAGE and analyzed by immunoblotting with the indicated antibodies. (C) HEK293 and A375 cells were infected with lentiviral shRNA constructs targeted against a scrambled sequence (Scr) or RSK1/2. After selection, cells were serum-starved and stimulated with either PMA (50 ng/mL) or left unstimulated. Protein lysates were resolved by SDS-PAGE and analyzed by immunoblotting with the indicated antibodies. (D) Schematic representation of the different conditions analyzed using SILAC and LC-MS/MS. The relative abundance in phosphopeptides was compared between SILAC pairs, which comprised HEK293 and A375 cells treated with MEK1/2 (PD184352) or RSK (BI-D1870) inhibitors, or subjected to a nontarget or RSK1/2 shRNAs.

Fig. 2.

Characterization of the RSK-dependent phosphoproteome. (A–C) Log₂ ratios of phosphopeptides identified comparing MEK1/2 (PD184352) and RSK (BI-D1870) inhibition, or RSK1/2 depletion by RNAi in HEK293 and A375 cells. Log₂ ratios below −1.5-fold were considered as significantly down-regulated. (D and E) Representative MS spectra of light and heavy peptides from Chk1 (S280) and rpS6 (Ser235). (Insets) Representative Western blots using corresponding phosphospecific antibodies. (F and G) MS quantification of the phosphopeptides containing Chk1 Ser280 or rpS6 Ser235 phosphorylation sites. (H) IPA of Gene Ontologies (GO) enriched within down-regulated phosphopeptides.

Fig. 3.

Peptide library profiling of the optimal substrate motif for RSK. (A) A spatially arrayed PSPL was subjected to in vitro phosphorylation with active RSK1 and radiolabeled ATP. Aliquots of each reaction were spotted onto a membrane and exposed to a phosphor storage screen. (B) Matrix of intensities derived from results shown in A. (C) Web logo representation of the RSK consensus phosphorylation motif. (D) Schematic representation of our global proteomic data from all experimental conditions. The data highlight the number of peptides and proteins affected by the MEK1/2 and RSK inhibitors, as well as the RSK1/2 RNAi. The proportions of phosphopeptides that fit the RSK consensus motif are indicated.

Fig. 4.

RSK phosphorylates PDCD4 at S457 and regulates its subcellular localization. (A) HEK293 cells were transfected with PDCD4, serum-starved, and pretreated with PD184352 (10 µM), rapamycin (25 nM), or BI-D1870 (10 µM) for 30 min before PMA (50 ng/mL) stimulation. Phosphorylation was assayed with an anti-RXXpS/T motif antibody. (B) HEK293 cells were transfected with WT PDCD4 or the S67A and S457A mutants, serum-starved, and stimulated with PMA (50 ng/mL) for 30 min. Phosphorylation was assayed by immunoblotting using the phospho-Ser457 and anti-RXXpS/T motif antibodies. (C) Recombinant RSK1 was incubated with immunopurified PDCD4 in a kinase reaction with [γ-³²P]ATP. The resulting samples were subjected to SDS/PAGE and the gel autoradiographed. In parallel, samples were immunoblotted with phospho-Ser457 antibodies. (D) Normal human melanocytes and three melanoma cell lines were analyzed for PDCD4 levels and phosphorylation. (E) The phosphorylation status of PDCD4 at Ser457 was analyzed in A375 cells treated with PD184352 (10 µM) or BI-D1870 (10 µM) for 1 h. (F) A375 cells treated as in E were imaged using immunofluorescence microscopy. Cells were stained with anti-PDCD4 antibodies to visualize endogenous PDCD4, phalloidin to visualize F-actin, and DAPI to visualize nuclei.

Fig. 5.

Identification of PDCD4 as a 14-3-3 binding protein in melanoma. (A) Subtractive fractionation proteomic scheme for enrichment of phospho-dependent 14-3-3 binding proteins. (B) Eluates were resolved by SDS/PAGE and gels stained with Coomassie or subjected to immunoblotting using the 14-3-3–binding motif antibody. (C) Comparisons of proteomic datasets between predicted RSK substrates (from Dataset S4) and 14-3-3 interacting proteins identified from A375 melanoma cells. (D) HEK293 cells were transfected with WT PDCD4, serum-starved, and stimulated with PMA (50 ng/mL) for 30 min before being harvested; 14-3-3 binding was analyzed in a pull-down assay. (E) HEK293 cells were transfected with WT PDCD4, serum-starved, pretreated with PD184352 (10 µM) or BI-D1870 (10 µM) followed by PMA (50 ng/mL) stimulation for 30 min. PDCD4 interaction to GST-14-3-3 was assessed as in D.

Fig. 6.

RSK mediates site-specific 14-3-3 binding to PDCD4 and thereby promotes its degradation. (A) HEK293 cells were transfected with different 14-3-3 isoforms, serum-starved, and stimulated with PMA (50 ng/mL) for 30 min. PDCD4 was immunoprecipitated and 14-3-3 binding was analyzed by immunoblotting. (B) Scansite analysis of PDCD4 sequence for 14-3-3 binding sites, and bar graph representation of PDCD4 Ser76 phosphorylation changes observed in this study. (C) HEK293 cells were cotransfected with 14-3-3β and WT PDCD4 or the S76A, S457A, and S76/457A mutants, serum-starved, and stimulated with PMA (50 ng/mL). The association between 14-3-3β and the different PDCD4 alleles was verified by coimmunoprecipitation. (D) HEK293 cells were transfected with WT PDCD4 or the double phosphorylation mutant S76/457A and treated with PMA (50 ng/mL) during a CHX (100 µg/mL) time course. Extracts were prepared at each time points and analyzed by immunoblotting. (E) A375 cells were treated with vehicle (DMSO), PD184352 (10 µM), BI-D1870 (10 µM), or SL0101 (50 µM) during a time course of CHX treatment (100 µg/mL). Extracts were prepared at the indicated times and endogenous PDCD4 levels were analyzed by immunoblotting. (F) Densitometric analysis of PDCD4 was performed on CHX time course shown in E and normalized to actin band intensity. The data were then expressed relative to respective controls (t = 0).

Fig. 7.

Schematic representation of the role of RSK in the regulation of the tumor suppressor protein PDCD4. Proposed model whereby the Ras/MAPK pathway converges on PDCD4 to regulate its nuclear accumulation and proteasomal degradation via a mechanism that involves 14-3-3 proteins.