Extensive crosstalk between O-GlcNAcylation and phosphorylation regulates Akt signaling

Abstract

O-linked N-acetylglucosamine glycosylations (O-GlcNAc) and O-linked phosphorylations (O-phosphate), as two important types of post-translational modifications, often occur on the same protein and bear a reciprocal relationship. In addition to the well documented phosphorylations that control Akt activity, Akt also undergoes O-GlcNAcylation, but the interplay between these two modifications and the biological significance remain unclear, largely due to the technique challenges. Here, we applied a two-step analytic approach composed of the O-GlcNAc immunoenrichment and subsequent O-phosphate immunodetection. Such an easy method enabled us to visualize endogenous glycosylated and phosphorylated Akt subpopulations in parallel and observed the inhibitory effect of Akt O-GlcNAcylations on its phosphorylation. Further studies utilizing mass spectrometry and mutagenesis approaches showed that O-GlcNAcylations at Thr 305 and Thr 312 inhibited Akt phosphorylation at Thr 308 via disrupting the interaction between Akt and PDK1. The impaired Akt activation in turn resulted in the compromised biological functions of Akt, as evidenced by suppressed cell proliferation and migration capabilities. Together, this study revealed an extensive crosstalk between O-GlcNAcylations and phosphorylations of Akt and demonstrated O-GlcNAcylation as a new regulatory modification for Akt signaling.

Figure 1. O-GlcNAcylations of endogenous Akt inhibit Akt phosphorylations at Thr 308 and Ser 473.

(A) Schematic representation of the method for characterization of the interplay between O-GlcNAc and O-phosphate of Akt. The proteins in the cell are first divided into two pools by O-GlcNAc immunoprecipitation: the un-GlcNAcylated and O-GlcNAcylated proteins subpopulations. Next, the phosphorylation levels of Akt between the two subpopulations are measured by western blotting. This will directly reflect the effect of O-GlcNAcylations on the phosphorylations of Akt. Here, MCF-7 cells were serum-starved under PUGNAc treatment (B and C) or under normal conditions (D and E), followed by IGF-1 stimulation for the indicated times. The un-GlcNAcylated and O-GlcNAcylated proteins subpopulations were prepared. Akt O-GlcNAcylations under PUGNAc treatment: (B) Determination of the same level of total Akt between the O-GlcNAcylated and un-GlcNAcylated proteins subpopulations. The O-GlcNAcylated proteins subpopulation was compared to a series of known amounts of the un-GlcNAcylated proteins subpopulations by immunoblotting against total Akt. (C) Comparison of the phosphorylation levels of the un-GlcNAcylated and O-GlcNAcylated Akt. The un-GlcNAcylated and O-GlcNAcylated proteins subpopulations were immunoblotted against Akt, O-GlcNAc, and the phosphorylations at Thr 308 and Ser 473. The constitutive O-GlcNAcylations of Akt: (D) Determination of the same level of total Akt between the O-GlcNAcylated and un-GlcNAcylated proteins subpopulations. The O-GlcNAcylated proteins subpopulation was compared to a series of known amounts of the un-GlcNAcylated proteins subpopulations by immunoblotting against total Akt. (E) Comparison of the phosphorylation levels of the un-GlcNAcylated and O-GlcNAcylated Akt. The un-GlcNAcylated and O-GlcNAcylated proteins subpopulations were subjected to immunoblot assay of Akt, O-GlcNAc, and the phosphorylations at Thr 308 and Ser 473. One unit was equal to 0.01 µl of the un-GlcNAcylated proteins. U: un-GlcNAcylated proteins; O: O-GlcNAcylated proteins.

Figure 2. Identification of Akt O-GlcNAcylation sites at Thr 305, Thr 312, Ser 126 and Ser129.

(A) Exogenous Akt was immunoprecipitated from MCF-7 cells under PUGNAc treatment. The eluted proteins were further separated by SDS-PAGE. Akt bands were in-gel digested and analyzed by mass spectrometry. (B) Tandem mass spectrum of the triply charged precursor ion at m/z 1300.24 with Xcorr score 2.87. The spectrum corresponded to Akt peptide 298–328 bearing one oxidation at Met 306 and two O-GlcNAc moieties. The two possible O-GlcNAcylation sites were Thr 312 and one of Thr 305 and 308. (C) Tandem mass spectrum of the triply charged precursor ion at m/z 1178.97 with Xcorr score 3.29. This spectrum corresponded to Akt peptide 302–328 containing one O-phosphate and two O-GlcNAc moieties. The two potential O-GlcNAcylation sites were Thr 312 and one of Thr 305 and 308. (D) O-GlcNAc levels of wild type and mutants of Akt. Akt and its mutants were immunoprecipitated from MCF-7 cells, and then subjected to western blotting analysis of total Akt and O-GlcNAc level. (E) Tandem mass spectrum of the doubly charged precursor ion at m/z 981.8 with Xcorr score 2.64. The spectrum corresponded to Akt peptide 118–132 bearing two O-GlcNAc modifications at Ser 126 and Ser 129. (F) Tandem mass spectrum of the doubly charged precursor ion at m/z 989.7 with Xcorr score 2.70. This spectrum corresponded to Akt peptide 118–132 bearing one oxidation at Met 118 and two O-GlcNAc modifications at Ser 126 and Ser 129. g: O-GlcNAc; o: oxidation, *: possible site for O-GlcNAcylation or phosphorylation.

Figure 3. Molecular modeling analysis.

(A) Molecular modeling of T305Y mutant (pink) and Akt O-GlcNAcylated at Thr 305 (green). Tyr305, O-GlcNAcylated Thr305 and Asp325 are represented as sticks. (B) Molecular modeling of T312Y mutant (pink) and Akt O-GlcNAcylated at Thr 312 (green). Tyr312, O-GlcNAcylated Thr312, Asp274 and Lys276 are shown as sticks. Additionally, oxygen atoms are shown in red, nitrogen atoms are shown in blue, and the hydrogen bonds are shown in red dot lines.

Figure 4. O-GlcNAclyations at Thr 305/312 suppress the Thr308 phosphorylation via disrupting the interaction between Akt and PDK1.

(A) Immunoblot analysis of the phosphorylation levels of wild-type Akt and its mutants. MCF-7 cells were transfected by the indicated plasmids, followed by IGF-1 stimulation. The cell lysates were subjected to immunoblotting analysis of the phosphorylation levels of Akt. (B) Fluorescence images of the subcellular localization of Akt and its mutants in COS-7 cells. COS-7 cells were transfected by the indicated plasmids and serum-starved overnight, followed by IGF-1 stimulation. Cells were fixed and probed by primary Akt antibody. (C) Immunoblot analysis of subcellular fractions of COS-7 cells expressing wild-type Akt and its mutants in response to IGF-1 stimulation. Na/K ATPase and actin were used as the plasma membrane and cytoplasmic markers, respectively. PM, plasma membrane; CYT, cytosol. (D) Coimmunoprecipitation assay for the interaction between PDK1 and Akt or its mutants. MCF-7 cells were transfected by the indicated plasmids and treated by IGF-1. Akt and its mutants were immunoprecipitated with anti-Flag agarose and the precipitate were probed against Akt and PDK1. (E) In vitro Kinase assay of PDK1. Akt and its mutants were immunoprecipitated from overexpressed MCF-7 cells and incubated with PDK1 at room temperature. The reaction mixture were immunoblotted against the Thr308 phosphorylation and total Akt.

Figure 5. Abnormal O-GlcNAcylations down-regulate Akt activity and functions.

(A) In vitro Kinase assay of Akt. MCF-7 cells were transfected by the indicated plasmids and treated by IGF-1. The immunoprecipitated Akt and its mutants were incubated with substrate GSK-3 fusion protein, followed by immunoblotting against the GSK3α/β(Ser21/9) phosphorylation, Akt phosphorylations and total Akt. (B&C) Cell proliferation assay of Akt and its mutants. MCF-7 and COS-7 cells transfected by wild-type Akt and its mutants were plated and cultured for the indicated times. Then, the cells were trypsinized and counted relative to the initially seeded cell number. Representative result was shown from three independent experiments with similar tendency. (D) Cell migration assay of wild-type Akt and its mutants. COS-7 cells were transfected by wild-type Akt and its mutants, and were tested using Transwell cell migration assay. The migrated cells were stained by crystal violet. (E) Quantification analysis of cell migration effect of Akt and its mutants. The stained cells were extracted with a solution of 10% acetic acid, which absorbance at 600 nm was measured. Cell migration levels were presented as mean ± s.d. from three independent experiments.