| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 18, 18623-18632, April 30, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
*
¶||
From the
Cell Signalling Group, Diabetes and Obesity Program and the Protein Analysis Laboratory, Garvan Institute of Medical Research, 384 Victoria Street, Darlinghurst, Sydney, New South Wales 2010, Australia and the **Laboratory of Developmental Biology, Graduate School of Frontier Bioscience, Osaka University, 1-3 Yamadaoka, Suita, Osaka 565-0871, Japan
Received for publication, February 11, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
and PKC
are activated in skeletal muscle of insulin-resistant high fat-fed rats, and to identify potential substrates for these kinases, we incubated recombinant PKC isoforms with rat muscle fractions in vitro. PKC specifically phosphorylated a 48-kDa protein that was subsequently identified by mass spectrometry as Ndrg2. Ndrg2 is highly related to N-Myc downstream-regulated protein 1, which has been linked to stress responses, cell proliferation, and differentiation, although Ndrg2 itself is not repressed by N-Myc. Ndrg2 contains several potential phosphorylation sites, including three Akt consensus sequences. Ndrg2 phosphorylation was enhanced in [32P]orthophosphate-labeled C2C12 muscle cells co-overexpressing either PKC or Akt. Phosphorylation of Ndrg2 was examined further using a phospho (Ser/Thr) Akt substrate antibody. Insulin increased Ndrg2 phosphorylation in C2C12 cells in a wortmannin- and palmitate-inhibitable manner, whereas rapamycin, PD98059, and bisindoylmaleimide I had no effect, supporting a direct role for Akt. Mutation of Ndrg2 indicated that Thr-348 is the major phosphorylation site detected by the antibody and that Akt stimulates phosphorylation of this site, whereas PKC phosphorylates Ser-332. PKC overexpression, however, diminished the effect of insulin on Thr-348 phosphorylation without reducing Akt activation, suggesting that this is mediated through phosphorylation of Ndrg2 at Ser-332. Our data identify Ndrg2 as a novel insulin-dependent phosphoprotein and suggest that PKC may inhibit insulin action in part by reducing its phosphorylation by Akt. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
, 
, and 
, which are activated by calcium and DAG; the nPKCs, e.g. 
, , and , which require DAG; and the aPKCs, 
and 
/
, which are independent of both calcium and DAG (1). The essential role of specific binding partners such as the RACKS and A-kinase anchoring proteins in the spatial co-ordination of PKC isoforms is also becoming apparent (2). Similarly, a number of substrates, which are directly phosphorylated by PKCs in particular cell types and which mediate PKC effects on cellular regulation, is emerging.
Specific PKC isoforms have been implicated in the inhibition of insulin action. Insulin resistance of target tissues such as skeletal muscle is a major contributing factor to the development of type 2 diabetes, and activation of nPKCs, including PKC, have been correlated with insulin resistance in a number of studies, especially in association with increased lipid availability (3, 4). The mechanisms involved, however, including the targets for PKC-mediated phosphorylation, remain unclear. PKCs could interfere at one or more steps in insulin signal transduction, and the possibility that they may mediate increased Ser/Thr phosphorylation of IRS-1 is currently receiving much attention. IRS-1 is a docking protein that is a substrate for the insulin receptor tyrosine kinase: Tyr phosphorylation of the protein promotes recruitment of PI3K and subsequent activation of downstream components of the insulin signaling cascade, including Akt and aPKC, which in part mediate the metabolic actions of the hormone. Ser/Thr phosphorylation of IRS-1 at specific sites inhibits its subsequent insulin-stimulated Tyr phosphorylation and PI3K recruitment and promotes IRS-1 degradation (5, 6). It is possible, however, that chronic activation of specific PKC isoforms also leads to phosphorylation of more distal components of insulin signaling to induce insulin resistance.
Ndrg1–4 (which comprises the family of proteins homologous to N-Myc downstream-regulated protein 1) have been implicated in stress responses and in the regulation of cell proliferation and differentiation. Ndrg1 (also called Drg1/Cap43/rit42/TDD5/Ndr1), which is the most widely expressed and best characterized member, is mutated in hereditary motor and sensory neuropathy-Lom, a form of Charcot-Marie Tooth disease (7). The expression of Ndrg1 is inhibited by N-Myc (8), and the protein is repressed in transformed cells and up-regulated in growth-arrested differentiating cells (9). Although Ndrg2 is in fact N-Myc-independent (10), it also appears to inhibit cell proliferation (11) and promote differentiation (12). Ndrg2 is highly expressed in skeletal muscle and brain, Ndrg3 is highly expressed in brain and testes, and Ndrg4 is specifically expressed in brain and heart (13). Ndrg2 cDNAs, corresponding to four different isoforms (Ndrg2a1, Ndrg2a2, Ndrg2b1, and Ndrg2a2), have been isolated from rat renal tissue, differing in their 5'-untranslated and N-terminal coding regions (14). The significance of these isoforms of rat Ndrg2 is unknown, and only one form of the gene has been described in mice (10) and humans (11). Phosphorylation of Ndrg proteins has been little studied, although protein kinase A-dependent phosphorylation of Ndrg1 has been described previously (15).
In this study, we initially examined the phosphorylation of skeletal muscle proteins by individual recombinant PKC isoforms in vitro. PKC, but not PKC, was found to phosphorylate a 48-kDa protein, which was subsequently identified as Ndrg2. Because Ndrg2 possesses three Akt consensus phosphorylation sites, we examined its phosphorylation by both PKC and Akt in intact C2C12 skeletal muscle cells. Ndrg2 was found to be phosphorylated by Akt in response to insulin stimulation, and this was inhibited in the presence of PKC activity. Our data therefore indicate that Ndrg2 is a novel target for Akt and may represent a site for PKC-mediated inhibition of insulin signal transduction.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-32P]ATP, [32P]orthophosphate, and Promix L-[35S]methionine/cysteine (1000 Ci/mmol) were from Amersham Biosciences (Sydney, New South Wales, Australia). Macroprep High Q resin, ceramic hydroxyapatite, Ready Gels, nitrocellulose, and polyvinylidene difluoride membrane were from Bio-Rad (Sydney). The Filtron Omegacell stirred cell device was from Pall Gelman (Sydney). Slide-A-Lyzer cassettes were from Pierce (Rockford, IL). LipofectAMINE 2000 reagent was Invitrogen (Carlsbad, CA). Constitutively active myristoylated/palmitoylated Akt1 (16) was a gift from J. Tavaré (University of Bristol, UK). -Phosphatidyl-L-serine, 1,2-dioctanoyl-sn-glycerol, protein kinase A inhibitor peptide (rabbit sequence), anti-FLAG M2 affinity gel, and anti-FLAG and FITC-conjugated anti--actin antibodies were from Sigma (St. Louis, MO). PD98059 was from Promega (Sydney). Wortmannin and rapamycin were from BIOMOL (Plymouth Meeting, PA). Bisindoylmaleimide I was from Alexis (Carlsbad, CA). Anti-phospho (Ser/Thr) Akt substrate; anti-phospho-Ser-473 Akt; anti-Akt; anti-phospho-Ser-2448 mTOR; anti-phospho-Thr-389 p70S6K; anti-phospho-Thr-37/46 4E-BP1; and anti-phospho-Thr-202/Tyr-204 ERK1/2 and anti-ERK1/2 antibodies were from Cell Signaling Technology (Beverly, MA). Mouse monoclonal anti-PKC and anti-PKC antibodies were from BD Transduction Laboratories (San Diego, CA). Cy3-conjugated donkey anti-rabbit antibodies were from Jackson ImmunoResearch Laboratories (West Grove, PA). Immuno-fluore was from ICN Biomedicals Inc. (Sevenhills, New South Wales, Australia). Preparation of Crude Muscle Fractions—Solubilized membrane fractions were prepared from red quadriceps muscle of male Wistar rats. Muscles (5 g) were homogenized in 20 ml of ice-cold buffer A (20 mM MOPS, pH 7.5, 4% (v/v) glycerol, 250 mM mannitol, 1 mM EGTA, 1 mM EDTA, 1 mM DTT, 2 mM PMSF, 200 µg/ml leupeptin, 2 mM benzamidine) and centrifuged at 100,000 x g for 45 min. This and all subsequent purification procedures were carried out at 4 °C. The supernatant was retained as the cytosolic fraction. The pellet was washed by resuspension with buffer A and recentrifuged. The washed pellet was resuspended with 20 ml of buffer B (20 mM MOPS, pH 7.5, 4% (v/v) glycerol, 1 mM EGTA, 1 mM EDTA, 0.5% (w/v) decanoyl-N-methylglucamide, 1 mM DTT, 2 mM PMSF, 200 µg/ml leupeptin, 2 mM benzamidine). After incubation at 4 °C for 1 h, the suspension was again centrifuged, and the supernatant was retained as the crude membrane fraction.
Further Fractionation of Muscle Proteins—The crude membrane fraction was applied to a 1-ml Mono Q column equilibrated in buffer B, and proteins eluted with a 0.05–1 M NaCl gradient in buffer B to yield 30 fractions over 30 ml. Alternatively, the membrane fraction was applied to a 1-ml hydroxyapatite column equilibrated with buffer B containing 20 mM KH2PO4. Proteins were eluted with a 20–300 mM KH2PO4 gradient in buffer B to yield 30 fractions over 30 ml.
Preparation of Recombinant PKC Isoforms—cDNA constructs in the vector pRC/CMV, encoding wild-type 6-His-tagged PKC and PKC were kind gifts from G. Baier, University of Innsbruck, Austria. PKC was excised by BamH1 digestion and cloned into the baculovirus transfer vector pVL1393, which had been linearized with BamH1. PKC was excised by XhoI digestion and cloned into the baculovirus transfer vector pAcSG2, which had been linearized using XhoI. These constructs were used to generate high titer recombinant baculoviruses for the expression of these PKC isoforms in insect Sf9 cells using the BD Pharmingen BaculoGold System according to the manufacturer's instructions. Sf9 cells from three 150-cm2 culture flasks, which had been infected 3 days previously with baculoviruses for the overexpression of either PKC or PKC, were harvested and washed in PBS prior to sonication (15 pulses, using a Branson 250 Sonifier and microtip at 20% duty cycle, power setting 2) in 5 ml of buffer C (20 mM MOPS, pH 7.5, 150 mM NaCl, 4% (v/v) glycerol, 1% (v/v) Tween 80, 2 mM PMSF, 200 µg/ml leupeptin, 2 mM benzamidine). Extracts were centrifuged at 100,000 x g for 40 min, and supernatants were applied at 0.1 ml/min to a 1-ml nickel-nitrilotriacetic acid Superflow equilibrated with buffer C. The column was washed first with 10 ml of buffer C then with 10 ml of buffer C containing 50 mM imidazole. Finally, 6-His-tagged PKCs were eluted in 0.5-ml fractions with buffer C containing 250 mM imidazole. Peak PKC-containing fractions, identified by SDS-PAGE and Coomassie Blue staining, were diluted with glycerol (final concentration 50% (v/v)) and stored at -20 °C.
In Vitro Phosphorylation of Muscle Proteins—Crude membrane and chromatography fractions (10 µl) were incubated at 30 °C in the absence or presence of purified recombinant 6-His-tagged PKC or PKC, in a final volume of 50 µl, containing in addition 20 mM MOPS, 0.04% (v/v) Triton X-100, 1 mM EGTA, 120 nM cAMP-dependent protein kinase inhibitor peptide, 120 µg/ml -phosphatidyl-L-serine, 2.5 µg/ml 1,2-dioctanoyl-sn-glycerol, and 100 µM [-32P]ATP (1000–2000 cpm/pmol). Reactions were terminated after 30 min by addition of 10 µl of 2x Laemmli sample buffer (17) and analyzed by SDS-PAGE, electroblotting to nitrocellulose, and phosphorimaging using an Amersham Biosciences 445 SI PhosphorImager.
Large-scale Purification of pp48 —Rat red quadriceps muscle (50 g) was homogenized in 200 ml of buffer A, and a solubilized membrane fraction was prepared as above. This was applied to a 25-ml High Q column equilibrated with buffer B containing 50 mM NaCl, and protein eluted with a non-linear gradient of 50–500 mM NaCl over 40 fractions of 12.5 ml, followed by 500–1000 mM over 10 fractions of 12.5 ml. Fractions were subjected to phosphorylation by PKC and those containing pp48 (fractions 13–17) were pooled and applied to a 25-ml ceramic hydroxyapatite column equilibrated in buffer B containing 2 mM KH2PO4, which was further washed with 125 ml of 2 mM KH2PO4. Proteins were eluted with a non-linear gradient of 2–75 mM KH2PO4 in buffer B over 25 fractions of 12.5 ml followed by 75–150 mM over 10 fractions of 12.5 ml. pp48 was found by in vitro phosphorylation by PKC to be present mainly in the initial flowthrough of this purification step, and this fraction (60 ml) was concentrated to 1.0 ml under vacuum using a Filtron Omegacell 10-ml stirred cell device with 30-kDa molecular mass cut-off membrane.
2-DE of pp48 —200 µl of concentrated pp48 was phosphorylated by incubation with PKC as detailed above, in a final volume of 1 ml. The reaction mixture was dialyzed against two changes of deionized water using a Slide-A-Lyzer cassette with 10-kDa cut-off membrane and freeze-dried. For first dimension isoelectron focusing, the sample was rehydrated with 250 µl of 8 M urea, 0.5% IPG buffer (pH 4–7), 2% (w/v) CHAPS, and 100 mM DTT, and the solution used to rehydrate a 13-cm, pH 4–7 IPG strip for 16 h. Proteins were focused on the IPG strip using an Amersham Biosciences IPGphor isoelectric focusing unit at 500 V for 1 h, followed by 1000 V for 1 h, and finally 8000 V for 2 h. The IPG strip was then equilibrated in 50 mM Tris-HCl, pH 8.8, 6 M urea, 30% (v/v) glycerol, 2% (w/v) SDS, and a trace amount of bromphenol blue and subjected to SDS-PAGE in the second dimension, using a 14 x 15 x 0.15 cm 10% (w/v) polyacrylamide gel. The gel was silver-stained (18) and subjected to phosphorimaging.
Protein Identification by Mass Spectrometry—The 32P-labeled protein spots were excised from silver-stained gels, vacuum-dried, and subjected to tryptic digestion as adapted from Shevchenko et al. (18). Briefly, gels chips were washed with CH3CN, dried, and reswollen with 50 µl of 50 mM NH4HCO3, pH 8.0, containing 10 ng/µl trypsin and digested for 16 h at 37 °C, and peptides were extracted by washing with 20 mM NH4HCO3 and three 20-min changes of 5% (v/v) formic acid in 50% (v/v) CH3CN. Peptides were lyophilized, then twice resuspended in water and lyophilized again to give a final volume of 5 µl. Protein identification was performed by µLC/ESI-MS/MS (19). A quaternary Hewlett-Packard 1100 series high-performance liquid chromatograph was directly coupled to a Finnigan MAT TSQ7000 mass spectrometer equipped with an in-house built microspray device for peptide ionization. MS/MS spectra were correlated with data base entries using SEQUEST (Thermo Finnigan, Sydney, Australia), a program that correlates uninterpreted peptide fragmentation patterns with amino acid sequences contained in databases and also capable of identifying phosphorylation sites from uninterpreted MS/MS spectra (20).
Ndrg2 Phosphorylation Site Mutants—Ser-332 
Ala, Thr-348 Ala, and Ser-350 Ala Ndrg2 mutants were generated by PCR-based mutagenesis of wild type mouse Ndrg2 cDNA in the pCMV/SV-FLAG1 vector using 5'-CGGTCTCGCACAGCAGCTCTGACCAGTGCAGCA-3', 5'-CGGTCCCGATCCCGCGCCCTGTCGCAGAGTAGC-3', and 5'-GGGCCCATCGATGGCAGTCGGTCCCGATCCCGCACCCTGGCGCAGAGTAGCG-3' as primers, respectively. The Ser-332 Ala/Thr-348 Ala/Ser-350 Ala combined mutant was generated by PCR-based mutagenesis of the Ser-332 Ala Ndrg2 construct using 5'-GGGCCCATCGATGGCAGTCGGTCCCGATCCCGCGCCCTGGCGCAGAGTAGCG-3' as a primer. All mutations were confirmed by DNA sequence analysis.
Mouse C2C12 Skeletal Muscle Cell Culture and Transfection—Cells were maintained as myoblasts in minimum essential medium with Earles' salts supplemented with 10% (v/v) FBS, and where stated induced to differentiate into myotubes by lowering FBS concentration to 1% (v/v), as previously described (21). For transfection with cDNA constructs, myoblasts were seeded at a density of 4 x 105 cells/well in 6-well culture plates. After 24 h, cells were transfected using LipofectAMINE 2000 reagent according to the manufacturer's instructions, with a 2:1 ratio of reagent to DNA. Typically, 1 µg/well of a Ndrg2-pCMV/SV-FLAG1 DNA construct was used when Ndrg2 was overexpressed alone, whereas for co-expression experiments 0.3 µg of Ndrg2 was used with 3 µg of DNA constructs encoding protein kinases as indicated. Cells were used in experiments 48 h post-transfection. Palmitate pretreatment of cells was as described previously (21).
Adenovirus-mediated Overexpression of Proteins in C2C12 Myotubes—The use of recombinant adenoviruses for the overexpression of constitutively active Akt, wild type PKC, or wild type PKC, together with GFP in C2C12 myotubes, generated using the pAdEasy system (22), has been previously described (23, 24). The pAdEasy system was also used to generate adenoviruses for the expression of Ndrg2 constructs, enabling co-expression of GFP and Ndrg2. A pAdEasy-derived virus expressing GFP alone was used in control infections. The amount of virus required to give 90–100% infection efficiency was individually determined for each virus by visualizing GFP expression. All viruses were used to infect C2C12 myotubes in 12-well plates at day 2 of differentiation, i.e. 3 days before experiments.
[32P]Orthophosphate Labeling of C2C12 Cells—Myoblasts in 6-well plates were washed twice with a phosphate-free salt solution (25 mM PIPES-NaOH, pH 7.2, 119 mM NaCl, 5 mM KCl, 5.6 mM glucose, 0.4 mM MgCl2, 1 mM CaCl2, 0.1% (w/v) BSA, and 4 mM glutamine) (25), labeled with [32P]orthophosphoric acid in 500 µl of the same solution (500 µCi/well) for 2 h at 37 °C, and stimulated as indicated. Radioactive supernatant was removed, and cells washed three times with PBS before addition of 500 µl of lysis buffer (PBS containing 1% (v/v) Nonidet P-40, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, 1 mM sodium pyrophosphate, 10 mM NaF, 2 mM phenylmethylsulfonyl fluoride, 200 µg/ml leupeptin, 2 mM benzamidine, and 20 µM MG132). Cells were incubated in lysis buffer for 30 min on ice, and the resulting extracts were centrifuged for 10 min at 16,000 x g. The supernatant was retained as the cell lysate.
Immunoprecipitation and Immunoblotting—Except for [32P]orthophosphate-labeled cells, muscle cells were washed twice with PBS and harvested in lysis buffer (250 µl for cells grown in 12-well plates), sonicated (10 pulses), and centrifuged at 16,000 x g for 10 min at 4 °C. Ndrg2 was immunoprecipitated from all cell lysates by addition of 5 µl of anti-FLAG M2 affinity gel and 20 µl of CL-6B-Sepharose. Lysates were rocked gently for 16 h, and beads were washed three times in lysis buffer before addition of 50 µl of Laemmli sample buffer. Immunoprecipitates and lysates were subjected to SDS-PAGE using 4–15% gradient Ready Gels and electroblotted onto a polyvinylidene difluoride membrane. Immunoblotting, densitometry, and statistical analysis were carried out as described previously (3). Primary antibodies were diluted 1:1000 except for anti-PKC (1:250).
Two-dimensional Tryptic Phosphopeptide Mapping—Ndrg2 immunoprecipitated from [32P]orthophosphate-labeled C2C12 cells, or pp48 phosphorylated in vitro, was digested with trypsin as above. Peptides were resolved in two dimensions as described previously (26), except that the electrophoresis buffer (first dimension) used was 5% (v/v) acetic acid, 1% (v/v) pyridine, and the chromatography buffer (second dimension) used was 30% (v/v) butan-1-ol, 30% (v/v) pyridine, 6% (v/v) glacial acetic acid.
Indirect Immunofluorescence—C2C12 cells grown on acid-washed coverslips were transiently transfected with plasmid DNA encoding FLAG-tagged Ndrg2. The cells were treated as stated in the figure legends, fixed in 3.7% (w/v) paraformaldehyde in PBS, permeabilized in 0.2% (v/v) Triton X-100 in PBS, and blocked with 2% (w/v) BSA in PBS. Cells were subsequently stained with both rabbit polyclonal anti-FLAG and FITC-conjugated anti-actin antibodies for 2 h. After three washes in PBS, cells were incubated with Cy3-conjugated donkey anti-rabbit antibodies for 1 h. After washing in PBS and water, cells were mounted in Immuno-fluore and viewed using a Leica DM RBE TCS SP1 confocal microscope with Plan-Apochromat 63x/1.4 oil immersion objective and a helium/neon laser providing excitation wavelengths (Leica Microsystems, Wetzlar, Germany). The fluorochromes FITC and Cy3 were excited with 488- and 568-nm wavelengths, respectively, and individual channels were scanned in series. Images were recorded digitally with a Leica TCS software package and processed using Photoshop 3.05 software (Adobe Inc., San José, CA). Each image represents a single 
0.8-µm `Z' optical section.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-specific Substrate in Skeletal Muscle—The nPKC isoforms PKC and PKC have been strongly implicated in the generation of lipid-induced insulin resistance in skeletal muscle, although the precise mechanisms involved are unclear. An indication of potential substrates for nPKCs in skeletal muscle was obtained by employing purified recombinant PKC isoforms to phosphorylate muscle proteins in vitro. PKC was able to phosphorylate a 48-kDa protein (pp48) in crude preparations, which was best observed in membrane fractions (Fig. 1A), although it was also apparent to a lesser degree in cytosolic fractions (not shown). When a similar level of PKC activity was used in incubations, normalized by comparison of in vitro PKC phosphorylation of histone IIIS (27), PKC was found to be less effective in the phosphorylation of pp48 (Fig. 1A). Partial purification of a membrane fraction by Mono Q ion exchange chromatography resulted in protein fractions eluting with 330 mM NaCl in which pp48 was the major phosphorylated species (in addition to autophosphorylated PKC) (Fig. 1B). Similarly, we observed that pp48 eluted from Mono S resin at 200 mM NaCl (not shown) and hydroxyapatite resin equilibrated with 20 mM KH2PO4 (Fig. 1C) and was precipitated with NH4SO4 (saturation 25–40% (w/v), not shown). In the absence of muscle-derived fractions, only bands corresponding to the autophosphorylated PKC isoforms were observed in phosphorimaging analyses, demonstrating that pp48 was not co-purified with PKC from Sf9 insect cells (Fig. 1D).
|
, prior to 2-DE, followed by silver staining and phosphorimaging (Fig. 2A). The major spot observed in the phosphorimaging analysis coincided with a major 48-kDa silver-stained protein spot, which was excised from the gel and subjected to tryptic digestion. PKC (pI = 8.2) focuses outside the pH range employed and, therefore, is not observed here. Tryptic peptides from pp48 were analyzed by µLC/ESI-MS/MS, and three peptides corresponding to the amino acid sequence of rat Ndrg2 (14) were identified (Fig. 2B). The presence of the peptide corresponding to residues 26–32 (Fig. 2, B and C) indicated that the Ndrg2a1/Ndrg2b1 isoform (Swiss-Prot/TrEMBL accession number Q8VBU2) must be present (14). This protein has also been termed Anti-depressant-related protein ADRG123 (Q8VI01).
|
has not been defined. To determine whether PKC could specifically phosphorylate Ndrg2 in intact cells, we overexpressed PKC and PKC with FLAG-tagged mouse Ndrg2, which is highly homologous to rat Ndrg2 (Fig. 2C), in C2C12 mouse skeletal muscle cells. Cells were co-transfected with PKC isoforms followed by metabolic labeling with [32P]orthophosphate. Ndrg2 was immunoprecipitated and subjected to either phosphorimaging or immunoblotting with FLAG antibody (Fig. 3A). In agreement with our in vitro experiments, PKC overexpression caused increased phosphorylation of Ndrg2 (3.9 ± 1.1-fold over basal phosphorylation in 12 experiments), which was further enhanced by TPA treatment, whereas PKC had little effect. TPA had no effect in the absence of PKC overexpression, most likely due to the low levels of endogenous PKC in C2C12 cells (24, 29).
|
motif (where X is any amino acid, and is Leu or Phe) (30), Ser-350 lacks the hydrophobic residue at +1, and none of the potential sites has the optimal residues Gly at +2 or Arg at -7 (31). Because we were interested in PKC substrates in relation to insulin signaling, and Akt is activated by insulin in C2C12 cells, we also investigated whether Ndrg2 could be phosphorylated in an Akt-dependent manner in intact cells. Co-expression of constitutively active Akt increased Ndrg2 phosphorylation (8.6 ± 4.0-fold over basal phosphorylation in eight experiments) as determined by phosphorimaging (Fig. 3B, top panel). A similar increase was also observed by immunoblotting with the phospho-(Ser/Thr) Akt substrate antibody (Fig. 3B, second panel). Immunoblotting Ndrg2 immunoprecipitates with FLAG antibodies indicated that the recovery of Ndrg2 protein was not affected by overexpression of either Akt or PKC (Fig. 3, A and B). These results indicate that Ndrg2 is indeed a phosphoprotein in whole cells and that its phosphorylation can be mediated by both PKC and Akt.
The phosphorylation of Ndrg2 by PKC and Akt in intact cells was further investigated by tryptic digestion and two-dimensional phosphopeptide mapping. Phosphorimaging analyses of maps obtained from C2C12 myoblasts overexpressing Ndrg2 in the absence or presence of either PKC or Akt indicated the presence of one major phosphopeptide that migrated close to the DNP-lysine marker, as well as minor species, migrating further toward the cathode (Fig. 4). The minor species were not always observed and may be due to incomplete digestion of Ndrg2. Likewise, pp48 purified from skeletal muscle and phosphorylated in vitro by incubation with PKC gave a similar map, strongly supporting the identification of the protein as Ndrg2, and indicating that phosphorylation by PKC in vitro occurred on the same tryptic peptide(s) to that in intact cells.
|
|
Determination of Ndrg2 Phosphorylation Sites—Because three putative Akt consensus phosphorylation sites were identified in Ndrg2, we wished to determine which of these were detected by the phospho-(Ser/Thr) Akt substrate antibody. Ser-332 had been identified as a phosphorylation site for PKC in partly purified Ndrg2 in vitro by µLC/ESI-MS/MS analysis (Fig. 2B), and phosphorylation at this site in intact C2C12 cells overexpressing Ndrg2 was confirmed by this method (not shown). For unknown reasons, however, the theoretical C-terminal tryptic peptide containing Thr-348 and Ser-350 (indicated in Fig. 2C) could not be detected by µLC/ESI-MS/MS, preventing further analysis by this method. Instead, we mutated the three putative Akt phosphorylation sites individually and in combination to Ala, to give four Ndrg2 mutants: Ser-332 Ala, Thr-348 Ala, Ser-350 Ala, and Ser-332 Ala/Thr-348 Ala/Ser-350 Ala. These were overexpressed in C2C12 myotubes, which were subsequently stimulated with insulin, and phosphorylation was again examined with the phospho-specific Akt consensus site antibody. This experiment indicated that Thr-348 is the major phosphorylation site detected by the antibody, whereas mutation of either Ser-332 or Ser-350 had little effect (Fig. 6A). The Ser-332 Ala and Thr-348 Ala Ndrg2 mutants were also used to examine the phosphorylation of the protein by phosphorimaging of immunoprecipitates of Ndrg2 from [32P]orthophosphate-labeled cells. Ndrg2 in C2C12 cells co-transfected with PKC was poorly phosphorylated when Ser-332 was mutated to Ala, whereas in cells overexpressing constitutively active Akt, only mutation of Thr-348 reduced Ndrg2 phosphorylation (Fig. 6B). Taken together, the results presented in Fig. 6 suggest that insulin stimulates phosphorylation of Ndrg2 primarily at Thr-348, which is mediated by Akt, whereas PKC phosphorylates the protein principally at Ser-332.
|
|
- and Akt-mediated Phosphorylation of Ndrg2—Finally, because PKC and Akt appear to phosphorylate Ndrg2 at distinct but closely located sites, we examined the effect of nPKCs on basal and insulin-stimulated Ndrg2 phosphorylation in C2C12 cells, detected using the phospho-(Ser/Thr) Akt substrate antibody. In control cells co-overexpressing GFP, the PKC activator TPA reduced insulin-stimulated phosphorylation of Ndrg2 (Fig. 8, first panel, lanes 1–4). This effect is probably mediated at least in part by the inhibition of insulin-stimulated Akt activation through activation of endogenous PKC, as indicated by the phospho-Ser-473 Akt immunoblot (third panel). Importantly, co-overexpression of PKC (Fig. 8, lanes 5–8) abolished the effect of insulin on Ndrg2 phosphorylation even in the absence of TPA (first panel, compare lanes 1 and 3 to lanes 5 and 7), even though a greater than 10-fold activation of Akt by the hormone could still be detected (third panel, compare lane 5 to lane 7). Densitometry and statistical analysis of three independent experiments indicated that there was a significant effect of PKC overexpression on insulin-stimulated Ndrg2 phosphorylation (p < 0.05), whereas there was no significant effect on Akt phosphorylation, in the absence of TPA. This strongly suggests a direct effect of PKC on Ndrg2, inhibiting its subsequent insulin-stimulated phosphorylation by Akt. Ndrg2 phosphorylation was further reduced by pretreatment of the cells with TPA (first panel, lanes 6 and 8), which again greatly reduced insulin-stimulated Akt activation (third panel, lane 8). Slightly different effects on Ndrg2 phosphorylation were obtained in cells co-overexpressing PKC rather than PKC, suggesting that different mechanisms are involved. Thus PKC reduced basal and insulin-stimulated Ndrg2 phosphorylation (p < 0.02) to an even greater extent (Fig. 8, first panel, lanes 9–12), and in a manner independent of TPA. Taken together with the data presented in Fig. 6, these results indicate that there is cross-talk between PKC and Akt signaling, which affects Ndrg2 phosphorylation at Thr-348, and the effect of PKC may be mediated by prior Ser-332 phosphorylation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
, and we present evidence that it is a site of cross-talk between the two signaling pathways, and hence of relevance to the development of PKC-mediated insulin resistance in this tissue.
The initial experiments described here indicate that recombinant PKC, but not PKC, is able to phosphorylate a 48-kDa protein in crude and partly purified skeletal muscle fractions in vitro (Fig. 1), which was identified as Ndrg2 (Fig. 2). This approach to identifying new PKC substrates has a relatively low sensitivity compared with protein expression library and yeast two-hybrid interaction screens and may detect only highly expressed substrates (27). In addition, it has been argued that PKC substrate specificity in vivo is regulated largely by association with binding partners and hence cellular localization (2). Despite these limitations, we have been able to identify a substrate from crude extracts showing isoform selectivity, indicating that the interaction between PKC and Ndrg2 has a degree of specificity. It is interesting to note that both of these proteins are in fact highly expressed in skeletal muscle (11, 14, 32), suggesting that phosphorylation of Ndrg2 by PKC in this tissue has physiological importance.
Our subsequent experiments concerning phosphorylation of overexpressed Ndrg2 in 32P-labeled muscle cells indicated that even in untreated cells, the protein exists in a partly phosphorylated form (Fig. 3B, top panel), which is not surprising given the large number of potential phosphorylation sites for basophilic kinases near the C terminus. Phosphorylation was augmented in phorbol ester-stimulated cells co-transfected with PKC but not PKC (Fig. 3A), in agreement with the in vitro data and suggesting that Ndrg2 is indeed a specific substrate for PKC in intact cells. Furthermore, we demonstrated that phosphorylation of the protein could also be increased by overexpression of Akt (Fig. 3B), as expected from the presence of three potential C-terminal Akt phosphorylation sites. The phospho-(Ser/Thr) Akt substrate antibody was therefore employed in further experiments, because this should specifically detect changes in phosphorylation at these sites. This antibody has been used similarly in the identification and characterization of a number of novel Akt substrates including a Rab-GAP (33) and ATP-citrate lyase (34).
From our data obtained using pharmacological kinase inhibitors (Fig. 5A), it is likely that Akt itself phosphorylates Ndrg2 at the sites detected by the antibody, rather than through the activation of additional kinases, such as p70S6K, which has a similar phosphorylation consensus sequence (30). We have not, however, excluded the possibility that serum- and glucocorticoid-inducible kinase, activated in a similar fashion to Akt and exhibiting a similar consensus sequence (35), may also phosphorylate Ndrg2. This may have physiological relevance in kidney cells, in which both serum- and glucocorticoid-inducible kinase and Ndrg2 are induced rapidly by aldosterone treatment (14, 36). Although also wortmannin-sensitive, aPKC or / are unlikely to account for the observed insulin-stimulated phosphorylation of Ndrg2, because the aPKC consensus sequence does not fit an Arg-X-Arg-X-X-Ser/Thr motif (37). In addition, recombinant aPKC/ was unable to phosphorylate pp48 in vitro (not shown).
Use of Ndrg2 mutants indicated that Thr-348 is the major phosphorylation site recognized by the phospho-(Ser/Thr) Akt substrate antibody (Fig. 6A). This is probably the major site of phosphorylation of Ndrg2 by Akt, because mutation of Thr-348 to Ala also markedly reduced the Akt-dependent [32P]orthophosphate labeling of the protein in C2C12 cells (Fig. 6B). In contrast, mutation of Ser-332 to Ala demonstrated that this residue was the major site phosphorylated by PKC (Fig. 6B), in agreement with MS data obtained from Ndrg2 phosphorylated in vitro by recombinant PKC (Fig. 2B) or in intact cells overexpressing the kinase (not shown). Although Ser-332 resides in an Arg-X-Arg-X-X-Ser motif, its phosphorylation by PKC does not appear to be recognized by the phospho-(Ser/Thr) Akt substrate antibody, because PKC overexpression in fact decreased Ndrg2 phosphorylation detected by the antibody (Fig. 8). The two phosphorylation sites identified are predicted to reside in separate tryptic peptides (Fig. 2C). This is inconsistent with the results of two-dimensional phosphopeptide mapping, which indicated the presence of only one major phosphopeptide (Fig. 4). The reason for this discrepancy is unclear, but it is most likely explained by a high lability of the tryptic peptide containing Thr-348 upon digestion, which is supported by the inability to detect this peptide by µLC/ESI-MS/MS. Thus, in cells overexpressing Akt, the major phosphopeptide observed (Fig. 4) probably corresponds to the peptide containing Ser-332, phosphorylated by endogenous kinases, whereas the expected phosphopeptide containing Thr-348 was not detected.
The study of Ndrg protein phosphorylation has been limited to Ndrg1 in intact endothelial cells or lysates (15). Ndrg1 is also Ser-enriched in its C-terminal region and may be phosphorylated at up to eight sites according to 2-DE (15). The purified protein could be phosphorylated in vitro by protein kinase A, whereas incubation of whole cells either with forskolin or 8-bromo-cAMP, or under conditions causing oxidative stress, also increased its phosphorylation. Similar to the data presented here for Ndrg2, altered phosphorylation of Ndrg1 did not alter its predominantly cytosolic localization, and because the biological role of the protein is unclear, it was not possible to determine an effect of phosphorylation on Ndrg1 function. Computational analysis of Ndrg protein structure has suggested that these proteins comprise a novel class in the / hydrolase superfamily, which consists of proteins exhibiting the / hydrolase fold, mostly enzymes that catalyze a diverse range of reactions (38). Despite the strong likelihood that Ndrg proteins possess the / hydrolase fold, they do not, however, have the expected catalytic motif and do not appear to be hydrolases (38).
Akt is activated by insulin in its target tissues, including skeletal muscle, and mediates both metabolic and mitogenic effects of the hormone (39, 40). Because Ndrg2 has been linked to inhibition of proliferation (11), it is possible that insulin-stimulated Akt-dependent phosphorylation modulates effects of the protein on cell growth. Indeed, insulin causes proliferation in C2C12 myoblasts (41), but also promotes myogenesis in these cells in an Akt-dependent manner (42). In preliminary experiments, however, we were unable to demonstrate any major effects of Ndrg2 overexpression on myoblast proliferation or differentiation (not shown).
Although we were unable to gain insights into Ndrg2 function, we did find evidence that Ndrg2 phosphorylation is diminished under conditions that promote insulin resistance. First, pretreatment of C2C12 myotubes with the saturated FFA palmitate reduced the effect of insulin on phosphorylation of the protein as detected by the phospho-specific antibody (Fig. 5B). This inhibition is most likely due to the inhibition of Akt through the de novo synthesis of the inhibitory sphingolipid ceramide from palmitate (21, 23). Second, and in contrast to the effects of palmitate, overexpression of PKC abolished the insulin-stimulated increase in Ndrg2 phosphorylation, whereas the activation of Akt was unaffected (Fig. 8). This is consistent with direct phosphorylation of Ndrg2 by PKC at Ser-332 and inhibition of subsequent phosphorylation by Akt at Thr-348. Although PKC overexpression also reduced Ndrg2 phosphorylation, this is unlikely to involve such sequential phosphorylation, because PKC phosphorylates Ndrg2 relatively weakly both in vitro (Fig. 1) and in intact cells (Fig. 3A). Because PKC had a greater effect on basal Ndrg2 phosphorylation at Thr-348 (Fig. 8), it is possible that the inhibitory action of this PKC is not specific to insulin signaling.
Interestingly, both PKC and PKC undergo chronic activation in insulin-resistant skeletal muscle from rats fed a safflower oil diet, rich in the unsaturated FFA linoleate (3), and are also activated in C2C12 cells by pretreatment with unsaturated FFAs, which promote insulin resistance (24). Thus it appears that inhibition of insulin action in skeletal muscle cells by either saturated or unsaturated FFAs, acting through distinct mechanisms, will be associated with diminished Ndrg2 phosphorylation by Akt. Although activation of PKC, especially PKC, has been correlated with insulin resistance in many studies (43), the specific pathways involved are unclear. We suggest that alteration in Ndrg2 phosphorylation plays a role and could have effects on muscle growth or metabolism (Fig. 9).
|
and Akt, and have presented evidence that Ser-332 and Thr-348 are the respective phosphorylation sites for these kinases. The protein is probably directly phosphorylated by endogenous Akt upon stimulation of muscle cells with insulin, and although the effect of these phosphorylations on the function of the protein remains to be determined, we have demonstrated that the Akt-mediated phosphorylation of Ndrg2 at Thr-348 is inhibited by PKC and propose that this cross-talk might represent one mechanism by which lipid-activated PKCs interfere with insulin action. |
FOOTNOTES |
|---|
¶ Current address: Bioanalytical Mass Spectrometry Facility, University of New South Wales, Sydney, Australia.
|| Current address: Biomedical Proteomics Research Group, Geneva University Hospital, Switzerland.
Current address: National Cardiovascular Center, Research Institute, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan.
To whom correspondence should be addressed. Tel.: 61-2-9295-8212; Fax: 61-2-9295-8201; E-mail: c.schmitz-peiffer{at}garvan.org.au.
1 The abbreviations used are: PKC, protein kinase C; 2-DE, two-dimensional electrophoresis; 4E-BP1, eukaryote initiation factor 4E-binding protein 1; aPKC, atypical protein kinase C; BSA, bovine serum albumin; cPKC, classic protein kinase C; DAG, diacyglycerol; ERK1/2, extracellular signal-regulated kinases 1/2; FBS, fetal bovine serum; FFA, free fatty acid; FITC, fluorescein isothiocyanate; IRS-1, insulin receptor substrate-1; µLC/ESI-MS/MS, microcapillary liquid chromatography electrospray ionization tandem mass spectrometry; mTOR, mammalian target of rapamycin; nPKC, novel protein kinase C; p70S6K, 70-kDa S6 protein kinase; PBS, phosphate-buffered saline; PI3K, phosphatidylinositol 3-kinase; RACKS, receptors for activated C-kinase; TPA, 12-O-tetradecanoylphorbol-13-acetate; MOPS, 4-morpholinepropanesulfonic acid; DTT, dithiothreitol; PMSF, phenylmethylsulfonyl fluoride; CMV, cytomegalovirus; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; GFP, green fluorescent protein; PIPES, 1,4-piperazinediethanesulfonic acid.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
