Dynamic adipocyte phosphoproteome reveals that Akt directly regulates mTORC2

Abstract

A major challenge of the post-genomics era is to define the connectivity of protein phosphorylation networks. Here, we quantitatively delineate the insulin signaling network in adipocytes by high-resolution mass spectrometry-based proteomics. These data reveal the complexity of intracellular protein phosphorylation. We identified 37,248 phosphorylation sites on 5,705 proteins in this single-cell type, with approximately 15% responding to insulin. We integrated these large-scale phosphoproteomics data using a machine learning approach to predict physiological substrates of several diverse insulin-regulated kinases. This led to the identification of an Akt substrate, SIN1, a core component of the mTORC2 complex. The phosphorylation of SIN1 by Akt was found to regulate mTORC2 activity in response to growth factors, revealing topological insights into the Akt/mTOR signaling network. The dynamic phosphoproteome described here contains numerous phosphorylation sites on proteins involved in diverse molecular functions and should serve as a useful functional resource for cell biologists.

Graphical abstract

Figure 1

Quantification of the Insulin-Regulated Phosphoproteome using Tandem Mass Spectrometry (A) Experimental design of inhibitor screens. (B) Experimental design of temporal phosphoproteome screen. (C) Workflow for the proteome and phosphoproteome analysis. (D) Summary of the quantified phosphoproteome and proteome. (E) Efficiency of phosphopeptide enrichment for a representative experiment. Above each bar is the percent of the total identified peptides in each fraction that were phosphorylated. (F) Quantitative reproducibility for a representative experiment of one of the large-scale MS studies. See also Figures S1 and S2 and Tables S1, S2, and S3.

Figure 2

Dynamic Range and Enrichment of GO Terms in Adipocyte Proteome and Phosphoproteome (A) Abundance of detected proteome was estimated using the summed peptide intensities of each protein, and proteins were ranked by abundance and divided into four quartiles. Enrichment of protein GO terms (biological process and cellular component) in each protein abundance quartile was assessed by Fisher’s exact test (FDR < 0.01 following Benjamini-Hochberg correction). The position of GO terms along the horizontal axis represents enrichment of these terms within the respective protein abundance quartile. (B) Frequency and significance of enrichment or de-enrichment of GOCC terms in the proteome and phosphoproteome (Fisher’s exact test, FDR < 0.01 after Benjamini-Hochberg correction). See also Table S4.

Figure 3

Quantification of Insulin/PI3K/Akt-Sensitive Phosphoproteomes (A) Distribution of all quantified phosphopeptides in inhibitor screens and insulin-responsive sites (purple). Known Akt-mediated (blue box) and mTOR-mediated (pink box) phosphosites regulated by insulin. Phosphosites are annotated as PI3K/mTOR- (orange squares) or Akt (red triangles)-dependent if their insulin-mediated phosphorylation was reversed by >40%. (B) PI3K- and Akt-regulated components of the insulin-regulated phosphoproteome. A subset of these were considered PI3K regulated if they were inhibited by LY294002 by more than the indicated thresholds and Akt regulated if they were also inhibited by MK2206. (C) Sequence logos for phosphorylation sites that were insulin regulated.

Figure 4

Dynamic Quantitative Analysis of Akt/mTOR Networks (A) Immunoblot analysis of adipocytes following different insulin-stimulated time points for proteins known to belong to the Akt (blue) and mTOR (pink) pathways. (B and C) Temporal profiles generated from SILAC-MS data for known direct Akt (B) and mTOR (C) substrates. (D) Network model depicting the activation of Akt, mTORC1, and mTORC2 by growth factors. See also Figure S3.

Figure 5

Temporal Phosphorylation in Response to Insulin Reveals Signaling Network Topology Data from the literature were used to construct a cell signaling network. Proteins identified in this study were annotated with their respective insulin-dependent phosphorylation sites color coded according to the temporal patterns derived from unsupervised clustering (fuzzy c-means), shown at the right. Complete clusters (A–R) are shown in Figure S3 and listed in Table S2. See also Figures S4 and S5.

Figure 6

Akt Is the Physiological Kinase for SIN1 T86, and its Phosphorylation Directly Regulates mTORC2 Activity (A) SIN1 domain structure and sequence homology of the region surrounding T86. TORC, putative mTORC-binding domain; CC, coiled-coil domain; CRIM, conserved region in the middle domain; RBD, Raf-like Ras-binding domain; PH, pleckstrin homology domain. Enlarged is the region containing the insulin-responsive phosphorylation site, T86. Residues surrounding several other known Akt substrates (AS160 T642, FOXO1A S256, TSC2 S939 and BAD S99) are shown. (B) Endogenous SIN1 is rapidly phosphorylated in response to insulin and blocked by pretreatment with the Akt inhibitor MK2206. 3T3-L1 adipocytes were serum starved, treated with MK2206 (MK; 10 μM, 30 min), and stimulated with insulin (100 nM) for the indicated times, and samples were assessed by immunoblotting. (C) Insulin-stimulated phosphorylation of endogenous SIN1 T86 is blocked by MK2206 and GDC-0068, but not by rapamycin. HEK293 cells were serum starved overnight, treated with MK2206 (MK; 10 μM and 1 μM), GDC-0068 (GDC;10 μM), or rapamycin (50 nM) followed by insulin (200 nM, 10 min), and samples were analyzed by immunoblotting. (D) Akt in vitro kinase assay performed using recombinant GST-Akt results in specific phosphorylation of SIN1 at T86 and is blocked by GDC-0068 (GDC^*; 10 μM) added to the in vitro kinase reaction. (E) Expression of SIN1, but not SIN1 T86A mutant, in SIN1^−/− MEFs rescues mTORC2-dependent signaling. SIN1 WT or phosphomutants (T86A, T86E) were expressed in SIN1^−/− MEFs, and cells were selected by FACS. Cell lines were serum starved and stimulated with insulin, and samples were analyzed by immunoblotting. (F) In vitro kinase activity of endogenous mTORC2 isolated from cells is enhanced by insulin stimulation (200 nM, 10 min) and blocked by pretreatment of cells with MK2206 (MK; 10 μM), but not rapamycin (R; 50 nM). LY294002 (LY^*; 15 μM) was added directly to the in vitro kinase assay. (G) mTORC2 isolated from SIN1^−/− MEFs reconstituted with SIN1 WT or phosphomutants (T86A, T86E) displays differential growth factor-stimulated kinase activity in in vitro kinase assay, with enhanced mTORC2 activity isolated from T86E phosphomimetic mutants. (H) Model depicting growth factor-dependent activation of mTORC2 mediated by Akt phosphorylation of SIN1. See also Figure S6.