Ceramide disables 3-phosphoinositide binding to the pleckstrin homology domain of protein kinase B (PKB)/Akt by a PKCzeta-dependent mechanism

Abstract

Ceramide is generated in response to numerous stress-inducing stimuli and has been implicated in the regulation of diverse cellular responses, including cell death, differentiation, and insulin sensitivity. Recent evidence indicates that ceramide may regulate these responses by inhibiting the stimulus-mediated activation of protein kinase B (PKB), a key determinant of cell fate and insulin action. Here we show that inhibition of this kinase involves atypical PKCzeta, which physically interacts with PKB in unstimulated cells. Insulin reduces the PKB-PKCzeta interaction and stimulates PKB. However, dissociation of the kinase complex and the attendant hormonal activation of PKB were prevented by ceramide. Under these circumstances, ceramide activated PKCzeta, leading to phosphorylation of the PKB-PH domain on Thr(34). This phosphorylation inhibited phosphatidylinositol 3,4,5-trisphosphate (PIP(3)) binding to PKB, thereby preventing activation of the kinase by insulin. In contrast, a PKB-PH domain with a T34A mutation retained the ability to bind PIP(3) even in the presence of a ceramide-activated PKCzeta and, as such, expression of PKB T34A mutant in L6 cells was resistant to inhibition by ceramide treatment. Inhibitors of PKCzeta and a kinase-dead PKCzeta both antagonized the inhibitory effect of ceramide on PKB. Since PKB confers a prosurvival signal and regulates numerous pathways in response to insulin, suppressing its activation by a PKCzeta-dependent process may be one mechanism by which ceramide promotes cell death and induces insulin resistance.

FIG. 1.

Representative blots showing the effects of ceramide and kinase inhibitors on the insulin-mediated phosphorylation of PKB and GSK3. L6 myotubes were incubated in the absence or presence of C₂-ceramide (Cer, 100 μM) for 2 h. In some experiments, cells were preincubated with Ro 31.8220 (Ro, at the concentrations indicated) for 30 min prior to incubation with ceramide. At the end of this incubation period cells were incubated with insulin (Ins, 100 nM) for a further 10 min before being lysed. (A and B) Cell lysates were then immunoblotted with a phospho-specific antibody directed against PKB-Ser⁴⁷³, PKB-Thr³⁰⁸, or PKB (A) and used for assaying PKB activity as described in the text (B). (C) L6 myotubes were incubated in the absence or presence (singularly or in combination) of kinase inhibitors at the indicated concentrations for 30 min prior to incubation with C₂-ceramide (100 μM) for 2 h and with insulin (100 nM) for 10 min. Cell lysates were immunoblotted with a phospho-specific antibody directed against PKB-Ser⁴⁷³, with blots being reprobed with an antibody against p38 MAPK (which was used as a marker for protein loading). (D) L6 myotubes were pretreated with Ro 31.8220 (Ro, 5 μM) for 30 min prior to incubation with C₂-ceramide (100 μM) for 2 h and insulin (Ins, 100 nM) for 10 min. Cell lysates were immunoblotted with phospho-specific antibodies directed against either GSK3α-Ser²¹, GSK3β-Ser⁷, PKB-Ser⁴⁷³, or PKB.

FIG. 2.

Effects of insulin, ceramide, and PMA on PKB and PKCζ phosphorylation in L6 myotubes. (A) L6 myotubes were pretreated with 1 μM PMA for the times indicated prior to incubation with C₂-ceramide (Cer, 100 μM) for 2 h and with insulin (Ins, 100 nM) for 10 min. Cells lysates were immunoblotted with a phospho-specific antibody directed against PKB-Ser⁴⁷³ or PKB-Thr³⁰⁸. (B) L6 myotubes were incubated in the absence or presence of Ro 31.8220 (Ro, 5 μM) for 30 min prior to treatment with C₂-ceramide (Cer, 100 μM) for 10 min. Myotubes were harvested, and plasma membranes were isolated by subcellular fractionation as described in the text. Plasma membranes (20 μg of protein) were subjected to SDS-PAGE and immunoblotted with antibodies against the α1 subunit of the Na/K-ATPase (a plasma membrane marker) and PKCζ. (C) In vitro activation of PKCζ was assessed by incubating 0.1 μg of recombinant PKCζ protein with C₂-ceramide (Cer, 100 μM) and Ro 31.8220 (Ro, 5 μM) in the presence of [γ-³²P]ATP, PS (4 pg/ml), and 5 mM MgCl₂ at 30°C for 20 min as described in the text. PKCζ was subsequently resolved by SDS-PAGE and subjected to analysis by autoradiography. Loading of PKCζ on SDS gels was assessed by probing the transfer membrane with an antibody directed against PKCζ. (D) L6 myotubes were incubated with C₂-ceramide (Cer, 100 μM) for the times indicated and then lysed. Cells lysates were immunoblotted with a phospho-specific antibody directed against PKCζ-Thr⁴¹⁰ or PKCζ.

FIG. 3.

A myristoylated PKCζ pseudosubstrate peptide inhibitor and expression of a kinase-inactive PKCζ mutant attenuate the inhibitory effects of ceramide on the insulin-induced phosphorylation of PKB in muscle cells. (A) L6 myotubes were pretreated with PKCζ myr-pseudosubstrate peptide (40 μM) or (B) with PKCα myr-pseudosubstrate peptide (50 μM) for 30 min prior to incubation with C₂-ceramide (100 μM) for 2 h and with insulin (100 nM) in the penultimate 10-min period prior to cell lysis. Cells lysates were immunoblotted with a phospho-specific antibody directed against PKB-Ser⁴⁷³ or PKB-Thr³⁰⁸. (C) L6 myoblasts were transfected with pCMV5 lacking or containing cDNA encoding kd-PKCζ or wild-type PKCζ as described in the text. Cells were then incubated with C₂-ceramide (100 μM) for 2 h prior to treatment with insulin (100 nM) for 10 min. Cell lysates were subsequently immunoblotted with a phospho-specific antibody to PKB-Ser⁴⁷³ or with an antibody to PKB.

FIG. 4.

PKB and PKCζ physically interact with each other both in vivo and in vitro, and this interaction requires the PH domain of PKB. (A) L6 myotubes were incubated in the absence or presence of C₂-ceramide (100 μM) for 2 h and/or with insulin (100 nM) in the penultimate 10-min period prior to cell lysis. PKBα was immunoprecipitated from cell lysates and resolved by SDS-PAGE prior to immunoblotting with antibodies to PKB or PKCζ. (B) Far-Western analysis was performed by resolving GST-tagged PKCζ (0.1 μg of protein) and whole-cell lysates from L6 cells (WCL, 50 μg of protein) by SDS-PAGE, followed by immobilization on polyvinylidene difluoride (PVDF) membranes. Membranes weresubsequently incubated overnight at 4°C with either 0.1 μg of recombinant PKB or 250 μg of whole-cell lysates prior to immunoblotting with antibodies directed against PKCζ (lane 1) or PKB (lanes 2 and 3). As a negative control, membranes retaining PKCζ were also probed with an antibody directed to c-myc (lane 4) after overnight incubation with whole-cell lysates. Expression of c-myc in L6 lysates was confirmed by probing whole-cell lysates with a c-myc antibody (lane 5). (C) To assess the importance of the PKB-PH domain for kinase interaction, far-Western analysis was performed by resolving 50 μg of whole-cell lysates and 0.1 μg of GST-PKCζ protein by SDS-PAGE, followed by immobilization of PKCζ on PVDF membranes. Membranes were subsequently incubated overnight at 4°C with either 0.1 μg of recombinant PKB or 0.1 μg of recombinant PKB lacking its PH domain (PKBΔPH) prior to immunoblotting with antibodies to PKCζ (lanes 1 and 2) or PKB (lanes 3 to 6). To confirm the presence of PKCζ on the membranes, lanes 3 to 6 were subsequently stripped and probed with an antibody to PKCζ.

FIG. 5.

Protein-lipid overlay assay showing that in vitro activation of PKCζ by ceramide leads to a loss in PIP₃ binding to PKB. (A) Nitrocellulose membranes were spotted with 1 μl of 500 pM PIP₃. Membranes were then incubated overnight at 4°C in TBST buffer containing 1 μM ATP, 4 pg of PS/ml, and 5 mM MgCl₂. This buffer also contained or lacked GST-PKB (0.5 μg/ml), GST-PKCζ (0.5 μg/ml), and/or C₂-ceramide (Cer, 100 μM) as indicated. Membranes were subsequently washed, and bound PKB was detected by probing with an anti-GST antibody. (B) To assess the importance of PKCζ activation in influencing the binding of PIP₃ to PKB, the experiment in panel A was repeated but with TBST buffer that either lacked ATP, PS, and MgCl₂ or which had been supplemented with Ro 31.8220 (Ro, 5 μM), as indicated.

FIG. 6.

In vitro activation of PKCζ by ceramide results in the phosphorylation of the isolated PH domain of PKB on Thr³⁴ with important consequences for PIP₃ binding. (A) The isolated PH domain of PKB (1 μg) was incubated in the absence or presence of 30 ng of PKCζ and/or C₂-ceramide (Cer, 100 μM) in buffer containing [γ-³²P]ATP and cofactors required to support kinase activation for 20 min at 30°C as described in the text. Phosphorylated proteins were then resolved by SDS-PAGE and transferred to PVDF membranes prior to autoradiography and immunoblotting with anti-PH antibodies. (B) To identify putative PKCζ phosphorylation sites within the isolated PH domain of all three PKBisoforms, a web-based motif-scanning analysis tool was used (49). The aligned peptide sequences of all three PKB isoforms are shown and highlight the putative PKCζ phosphorylation site. (C) In vitro phosphorylation of wild-type (PH-wt) and T34A mutant PH (PH-T34A) domains was performed by incubating 30 ng of recombinant PKCζ with the appropriate PH peptide (1 μg of protein) in the presence of 1 μΜ [γ-³²P]ATP and cofactors required for supporting kinase activation but in the absence or presence of C₂-ceramide (Cer, 100 μM) for 20 min at 30°C as indicated. PH peptides were then resolved by SDS-PAGE and subjected to analysis by autoradiography. Protein loading was subsequently assessed with an antibody directed against the PKB-PH domain. (D) Far-Western analysis was performed by resolving recombinant PKCζ (0.1 μg of protein) by SDS-PAGE and transferring it onto PVDF membranes. The membranes were then incubated overnight at 4°C with either 0.1 μg of PH-wt (lane 2) or 0.1 μg of PH-T34A peptide (lane 3) prior to immunoblotting them with antibodies to PKCζ (lane 1) or the PH domain of PKB (lanes 2 to 4). (E) Protein-lipid overlay was performed to assess PIP₃ binding to the PH-wt and T34A-PH peptides. Nitrocellulose membranes were spotted with 1 μl of PIP₃ and subsequently incubated overnight at 4°C in TBST buffer containing 1 μM ATP, 4 pg of PS/ml, and 5 mM MgCl_2. The incubation buffer either contained or lacked 0.5 μg of the appropriate PH peptide/ml, 0.5 μg of PKCζ/ml, and C₂-ceramide (Cer, 100 μM) as indicated. Membranes were washed, and bound PH protein was detected by probing samples with an anti-PH domain antibody.

FIG. 7.

Ceramide induces phosphorylation of PKB on Thr³⁴ in intact cells and a PKB T34A mutant is ceramide resistant. (A) L6 myotubes were incubated in the absence or presence of C₂-ceramide (Cer, 100 μM) for 2 h prior to cell lysis. Cell lysates were resolved by SDS-PAGE prior to immunoblotting with an antibody to PKB, a phospho-specific antibody to PKB-Thr³⁴, or anti-PKB-Thr³⁴ that had been preadsorbed the antigenic phospho-peptide (100 μg/ml). (B) Myotubes were treated as in panel A but, in addition, were also exposed to insulin (Ins, 100 nM for 10 min) or pretreated with Ro 31.8220 (Ro, 5 μM) prior to incubation with Cer. Cells were lysed and immunoblotted with antibodies to PKB or PKBThr³⁴. (C) HA-tagged PKB (wild type) and HA-tagged PKB T34A were transiently transfected into L6 cells as described. Cells were exposed to C₂-ceramide (Cer, 100 μM, 2 h) and or insulin (Ins, 100 nM, 10 min) prior to cell lysis and immunoprecipitation with an anti-HA antibody. Precipitated kinases were resolved by SDS-PAGE and immunoblotted with antibodies to PKB-Ser⁴⁷³ or anti-HA. Phospho-PKB-Ser⁴⁷³- and HA-immunoreactive bands were quantified and are expressed as a ratio (lower panel).

FIG. 8.

Effects of ceramide on muscle cell viability. L6 myoblasts transiently transfected with pCMV5 lacking or containing cDNA encoding wild-type PKCζ, kd-PKCζ, or PKB T34A were incubated with either vehicle alone (dimethyl sulfoxide), C₂-ceramide (Cer, 100 μM, 2 h), or C₂-ceramide (100 μM, 2 h) plus Ro 31.8220 (5 μM, added 15 min prior to ceramide). Ceramide promotes detachment and death of myoblasts, and thus viable cells were those that remained adherent, displayed trypan blue exclusion, and stained positively for DAPI. (A and C) Representative images of DAPI-stained L6 cells; (B and D) quantitative analysis of viable cells from five randomly chosen visual fields. Cell loss was expressed as a percentage change in cell number relative to that from the appropriate untreated experimental cell population. Asterisks signify significant changes (P < 0.05) between the indicated bars or compared to an untreated cell population, as determined by one-way analysis of variance.

FIG. 9.

Structure of PKBαPH complexed to Ins(1,3,4,5)P₄. A ribbon structure of the PH domain of PKBα, depicting the localization of Thr³⁴ relative to that of Arg²⁵ and Arg²³, which play a critical role in the binding of the inositol head group, is shown. Also labeled are the seven β strands (labeled β1 to β7). The ribbon drawing is based on the crystal structure proposed by Thomas et al. (47).