Biochem. J. (2013) 455, 207–216 (Printed in Great Britain)

207

doi:10.1042/BJ20130760

*Department of Medicine, University of Hong Kong, L8-39, 21 Sassoon Road, Hong Kong, †State Key Laboratory of Pharmaceutical Biotechnology, University of Hong Kong, 21 Sassoon Road, Hong Kong, ‡Department of Pharmacology & Pharmacy, University of Hong Kong, L2-53, 21 Sassoon Road, Hong Kong, §The Key Laboratory of Regenerative Biology, Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou 510530, China, and Department of Biology, York University, Toronto, Ontario, Canada M3J 1P3

Insulin inhibits hepatic glucose production through activation of the protein kinase Akt, and any defect in this pathway causes fasting hyperglycaemia in Type 2 diabetes. APPL1 [adaptor protein, phosphotyrosine interaction, PH (pleckstrin homology) domain and leucine zipper containing 1] sensitizes hepatic insulin action on suppression of gluconeogenesis by binding to Akt. However, the mechanisms underlying the insulin-sensitizing actions of APPL1 remain elusive. In the present study we show that insulin induces Lys63 -linked ubiquitination of APPL1 in primary hepatocytes and in the livers of C57 mice. Lys160 located within the BAR (Bin/amphiphysin/Rvs) domain of APPL1 is the major site for its ubiquitination. Replacement of Lys160 with arginine abolishes insulin-dependent ubiquitination and membrane localization of APPL1, and also diminishes membrane recruitment and activation of Akt, thereby abrogating the effects

of APPL1 on alleviation of hepatic insulin resistance and glucose intolerance in obese mice. Further analysis identified TRAF6 (tumour-necrosis-factor-receptor-associated factor 6) as an E3 ubiquitin ligase for APPL1 ubiquitination. Suppression of TRAF6 expression attenuates insulin-mediated ubiquitination and membrane targeting of APPL1, leading to an impairment of insulinstimulated Akt activation and inhibition of gluconeogenesis in hepatocytes. Thus TRAF6-mediated ubiquitination of APPL1 is a vital step for the hepatic actions of insulin through modulation of membrane trafficking and activity of Akt.

INTRODUCTION

N-terminal BAR (Bin/amphiphysin/Rvs) domain, a central PH domain and a C-terminal PTB (phosphotyrosine-binding) domain, was initially identified as a binding partner of Akt2 in a yeast twohybrid assay [8]. Subsequent studies demonstrated that APPL1 regulates both activity and substrate specificity of Akt [9,10]. In particular, APPL1 potentiates insulin actions in hepatocytes [7], skeletal muscle [11], adipocytes [12], endothelial cells [13] and cardiomyocytes [14], and also enhances insulin secretion in pancreatic β-cells [15]. Genetic disruption of APPL1 results in insulin resistance and defective insulin secretion, leading to glucose intolerance in mice [15]. In contrast, transgenic expression of APPL1 protects mice from HFD (high-fat diet)induced insulin resistance and β-cell dysfunction [15]. Furthermore, APPL1 interacts with the adiponectin receptors, thereby mediating the insulin-sensitizing effects of this adipokine [16,17]. In several target cells of insulin (hepatocytes, skeletal muscle, endothelium and β cells), APPL1 potentiates insulin-stimulated Akt activation by competing with TRB3 (tribble 3) for binding to Akt [7,13,15]. TRB3 acts as an endogenous inhibitor of Akt, by trapping Akt within the cytosol and preventing its membrane translocation [7]. In contrast, the interaction of APPL1 with Akt enables the release of Akt trapped by TRB3 and promotes the trafficking of the APPL1–Akt complex to the endosomal and plasma membrane for further activation in response to insulin stimulation [7]. APPL1 serves as a dynamic scaffold that

Insulin suppresses hepatic glucose production by activating the protein kinase Akt, which in turn inhibits the expression of the gluconeogenic enzymes PEPCK (phosphoenolpyruvate carboxykinase) and G6Pase (glucose 6-phosphatase) [1]. Upon insulin stimulation, Akt is recruited to the plasma membrane via the interaction between its PH (pleckstrin homology) domain and membrane PtdIns(3,4,5)P3 produced by PI3K (phosphoinositide 3-kinase) [1]. The membrane–Akt interaction results in conformational changes of Akt, enabling its activation through phosphorylation at Thr308 and Ser473 by its upstream kinase PDK1 (3-phosphoinositide-dependent protein kinase 1) [2] and mTORC2 (mammalian target of rapamycin complex 2) [3] respectively. The impairment in insulin-stimulated Akt activation is a key feature of hepatic insulin resistance in Type 2 diabetic patients [1]. On the other hand, an activating mutation of Akt, which is constitutively recruited to the plasma membrane, causes hypoglycaemia in humans [4]. The recruitment of Akt to plasma membrane is tightly controlled by various post-translational modifications and/or protein–protein interactions [5–7]. However, whether these regulatory mechanisms are also engaged in insulin signalling cascades remains largely unknown. APPL1 (adaptor protein, phosphotyrosine interaction, PH domain and leucine zipper containing 1), which contains an

Key words: adaptor protein, phosphotyrosine interaction, PH domain and leucine zipper containing 1 (APPL1), gluconeogenesis, insulin resistance, signal transduction, ubiquitination.

Abbreviations used: APPL1, adaptor protein, phosphotyrosine interaction, PH domain and leucine zipper containing 1; BAR, Bin/amphiphysin/Rvs; DMEM, Dulbecco’s modified Eagle’s medium; EGF, epidermal growth factor; FOXO1, forkhead box protein O1; G6Pase, glucose 6-phosphatase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GSK3β, glycogen synthase kinase 3β; HA, haemagglutinin; HEK, human embryonic kidney; HFD, high-fat diet; IGF-I, insulin-like growth factor 1; PEPCK, phosphoenolpyruvate carboxykinase; PH, pleckstrin homology; PI3K, phosphoinositide 3-kinase; PTB, phosphotyrosine-binding; qPCR, quantitative real-time PCR; STC, standard chow; TRAF6, tumour-necrosis-factor-receptor-associated factor 6; TRB3, tribble 3; Ub, ubiquitin. 1 To whom correspondence should be addressed (email [email protected]).  c The Authors Journal compilation  c 2013 Biochemical Society

Biochemical Journal

Kenneth K. Y. CHENG*†, Karen S. L. LAM*†, Yu WANG†‡, Donghai WU§, Mingliang ZHANG*, Baile WANG*†, Xiaomu LI*†, Ruby L. C. HOO*†, Zhe HUANG*†, Gary SWEENEY and Aimin XU*†‡1

www.biochemj.org

TRAF6-mediated ubiquitination of APPL1 enhances hepatic actions of insulin by promoting the membrane translocation of Akt

208

K. K. Y. Cheng and others

modulates RAB5-associated signalling in endosomal membranes by its ability to undergo domain-mediated oligomerization, membrane targeting and phosphoinositide binding [9,10,18]. However, the mechanism that governs the membrane targeting of APPL1 and Akt in insulin-responsive target cells remains enigmatic. APPL1 is post-translationally modified by phosphorylation and ubiquitination [19,20]. Phosphorylation of APPL1 at Ser430 has been shown to impair insulin-stimulated Akt activation in hepatocytes [20]. APPL1 undergoes Lys63 -linked polyubiquitination [19], however the physiological roles of this post-translational modification are currently unknown. In the present study, we provide evidence that APPL1 ubiquitination is a crucial step of insulin action in hepatocytes, by modulating the membrane targeting of Akt. Furthermore, we have identified TRAF6 (tumour-necrosis-factor-receptor-associated factor 6) as an upstream ligase for APPL1 ubiquitination and a regulator of insulin sensitivity in hepatocytes.

EXPERIMENTAL

Ubiquitination assay, immunoblotting and co-immunoprecipitation

For the ubiquitination assay, HEK (human embryonic kidney)293 cells or primary mouse hepatocytes with ectopic expression of FLAG-tagged APPL1 or K160R and HA-tagged Ub or its mutants were solubilized in a lysis buffer (20 mM Tris/HCl, 1 % SDS and 150 mM NaCl, pH 7.4) plus 5 mM of the Ub isopeptidase inhibitor N-ethylmaleimide, followed by boiling at 95 ◦ C for 10 min. The cell lysate was diluted with an immunoprecipitation buffer [20 mM Tris/HCl, 150 mM NaCl, 1 mM Na2 EDTA, 1 mM EGTA, 1 % Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM 2-glycerophosphate, 1 mM Na3 VO4 and protease inhibitor cocktail (Roche), pH 7.4] at a 1:10 ratio. The lysate (1 mg) was immunoprecipitated with an anti-FLAG or anti-APPL1 antibody coupled with Protein A–agarose beads, and the immunocomplexes were subjected to immunoblotting analysis with different antibodies as specified in each Figure legend. For co-immunoprecipitation experiments, HEK-293 cells or primary mouse hepatocytes with ectopic expression of FLAG-tagged APPL1 or K160R mutant and HA-tagged Akt2 or Myc-tagged TRAF6 were lysed in the immunoprecipitation buffer, and were performed as previously described [17].

Materials

Rabbit polyclonal antibodies against total Akt, total GSK3β (glycogen synthase kinase 3β), phospho-GSK3β (Ser9 ), total FOXO1 (forkhead box protein O1), phospho-FOXO1, βtubulin and pan-cadherin were purchased from Cell Signaling Technology, and antibodies against Ub (ubiquitin), Ub-Lys48 specific (Ub-K48) and Ub-Lys63 -specific (Ub-K63) were obtained from Merck Millipore. Horseradish peroxidase-conjugated anti-HA (haemagglutinin) and anti-FLAG mouse monoclonal antibodies were from Sigma. Rabbit polyclonal antibodies against TRAF6 and PEPCK were from Abcam. Rabbit polyclonal antibodies against APPL1 and phospho-Akt (Ser473 ) were obtained from Antibody and Immunoassay Services (The University of Hong Kong, Hong Kong). Plasmids encoding HA– Ub and its mutants were gifts from Edward Yeh (Department of Cardiology, University of Texas M.D. Anderson Cancer Center, Houston, TX, U.S.A.) [21]. Dexamethasone, cAMP, type Icollagen and collagenase-I were from Sigma. Human recombinant insulin was from Novo Nordisk.

Animals and metabolic studies

Male C57 BL/6N mice (4-weeks old) were fed with a STC (standard chow) (Purina Mills) composed of 20 kcal % protein, 10 kcal % fat and 70 kcal % combined simple carbohydrates or a HFD (Research Diets) composed of 20 kcal % protein, 45 kcal % fat and 35 kcal % carbohydrates for 12 weeks, followed by administration with adenovirus encoding luciferase control, APPL1 or the K160R mutant by tail vein injection [1×108 pfu (plaque-forming units)/mouse]. Insulin tolerance, glucose tolerance and pyruvate tolerance tests were performed at day 5, day 9 and day 12 after adenoviral injection respectively [7]. For the glucose tolerance and pyruvate tolerance tests, mice were fasted overnight, followed by intraperitoneal injection with 2 g glucose/kg of body weight and 1 g pyruvate/kg of body weight respectively. For the insulin tolerance test, mice fasted for 6 h were intraperitoneally injected with insulin (0.75 units/kg). Serum levels of insulin and glucose were determined by an insulin ELISA (Antibody and Immunoassay Services, the University of Hong Kong) and a glucose meter (Roche Diagnostics). All animal experimental protocols were approved by the Animal Ethics Committee of the University of Hong Kong.  c The Authors Journal compilation  c 2013 Biochemical Society

Mutagenesis, generation and purification of adenoviruses

The recombinant adenoviruses encoding human APPL1 and luciferase were generated in our previous study [7]. The K6R, K63R, K70R or K160R mutation in APPL1 gene was introduced into pshuttle-CMV-APPL1 plasmid (as described in our previous study [7]) following PCR-mediated site-directed mutagenesis using the mutagenic primers specified in Supplementary Table S1 (at http://www.biochemj.org/bj/455/bj4550207add.htm). cDNA encoding human full-length TRAF6 with N-terminal Myc epitope was inserted into the pshuttle-CMV vector. The pshuttle vector encoding the K160R mutant or TRAF6 was subcloned into pAdeasy-1 adenoviral backbone vector (Agilent Technologies) through recombination as previously described [7]. For construction of the adenoviral vector encoding RNAi specific to mouse TRAF6 or scramble control, the oligonucleotides (Supplementary Table S1) were ligated into pENTR/U6 entry vector, and then subcloned into pAd/BLOCK-it DEST vector (Invitrogen) by recombination. All of the recombinant adenoviruses were packaged and amplified in HEK-293T cells, followed by purification using affinity column chromatography (Agilent Technologies). The adenoviral titer was determined by plaque assay [7]. Isolation of primary hepatocytes, glucose production and qPCR (quantitative real-time PCR) analysis

Primary mouse hepatocytes were isolated by collagenase digestion as previously described [7]. The isolated cells were seeded on to type I collagen-coated 12-well plates in DMEM (Dulbecco’s modified Eagle’s medium) with 10 % FBS for 24 h, and then infected with various adenoviruses (multiplicity of infection = 20) for 24 h, followed by serum starvation for 12 h. The cells were then washed twice with PBS and incubated with Phenol Red- and glucose-free DMEM containing cAMP (100 μM), dexamethasone (100 nM) and/or insulin (10 nM) for 6 h. Glucose concentration in the conditioned medium was measured by a Amplex® Red glucose assay kit (Invitrogen) and normalized to total protein content. Gene expression levels of G6Pase and PEPCK were determined by qPCR analysis and normalized against GAPDH (glyceraldehyde 3-phosphate dehydrogenase) as previously described [7].

Ubiquitination of APPL1 and insulin actions

Figure 1

209

Ubiquitination of APPL1 is Lys63 -linked and is enhanced by insulin stimulation

(A–C) HEK-293 cells were co-transfected with plasmids encoding FLAG-tagged APPL1 and HA-tagged wild-type Ub, HA–Ub-K48R mutant, HA–Ub-K63R mutant, HA–Ub-K48 mutant or HA–Ub-K63 mutant for 48 h. The transfected cells were subjected to immunoprecipitation (IP) with an anti-FLAG antibody, followed by immunoblotting with the indicated antibodies. (D) Twelve-week-old male C57BL/N lean mice were fasted overnight and intraperitoneally injected with insulin (Ins) (0.5 unit/kg of body weight) for the indicated time points. Liver tissues were rapidly homogenized and immunoprecipitated with an anti-APPL1 antibody or rabbit IgG as control, followed by immunoblotting with an anti-Ub or anti-APPL1 antibody. (E) Primary hepatocytes were infected with adenovirus encoding FLAG–APPL1 and HA–Ub for 24 h, followed by serum starvation for 12 h. The cells treated with or without insulin (10 nM) for 5 min were subjected to immunoprecipitation using anti-FLAG antibody, followed by immunoblotting with anti-FLAG, anti-Ub, anti-Ub-K63 and anti-Ub-K48 antibodies as specified. All experiments were repeated at least three times and representative images are shown.

Subcellular fractionation

Primary mouse hepatocytes infected with various adenoviruses were serum-starved for 24 h and then stimulated with insulin (10 nM) for different time periods. The cells were washed twice with cold PBS and resuspended in a buffer consisting of 250 mM sucrose, 20 mM Hepes, pH 7.4, 10 mM KCl, 1.5 mM MgCl2 , 1 mM EDTA, 1 mM EGTA and 1 mM DTT plus protease inhibitor cocktails. The cytoplasmic and plasma membrane fractions were isolated by differential centrifugation as described previously [7]. Protein concentration in each fraction was quantified by BCA protein assay (Pierce Thermo Scientific). Statistical analysis

Data are presented as means + − S.E.M. All experiments were repeated three times with representative data shown. For animal experiments, five mice were included in each group. Statistical significance was determined by Student’s two-tailed t test. A P value of less than 0.05 represented a significant difference in all statistical comparisons. RESULTS Insulin induces Lys63 -linked ubiquitination of APPL1

Consistent with a previous study [19], the anti-HA antibody recognizing HA-tagged Ub detected a high molecular smear of ectopically expressed APPL1 in Ub-transfected HEK-293 cells (Figure 1A), suggesting that APPL1 is modified by

polyubiquitination. To ascertain whether ubiquitination of APPL1 is Lys48 - or Lys63 -linked, FLAG–APPL1 was co-expressed with HA–wild-type Ub (HA–Ub) or with either HA–Ub-K48R or HA–Ub-K63R, in which lysine residues of Ub were mutated to arginine at position 48 or 63 respectively. Immunoprecipitation analysis showed that only HA–Ub and HA–Ub-K48R mediated ubiquitination of APPL1, whereas ubiquitination of APPL1 was hardly detectable in cells co-expressing the K63R mutant (Figure 1B). To confirm further the above findings, we transfected HEK-293 cells with another two Ub mutants, HA– Ub-K63 and HA–Ub-K48, in which all lysine residues of Ub were mutated into arginine except the one at position 63 and at position 48 respectively. Wild-type Ub and the Lys63 mutant, but not the Lys48 mutant, displayed ubiquitination of APPL1 (Figure 1C), suggesting that the ubiquitination of APPL1 is mainly Lys63 -linked. Treatment with the proteasome inhibitor MG-132 did not further enhance APPL1 ubiquitination (Supplementary Figure S1 at http://www.biochemj.org/bj/455/ bj4550207add.htm), indicating that this ubiquitination did not cause proteasomal degradation of APPL1. As APPL1 is an important component of insulin signalling by promoting Akt activation [7,11–13], we next investigated whether ubiquitination of APPL1 is responsive to insulin stimulation. In both the liver tissues of C57/BL6N lean mice and primary hepatocytes, insulin induced a marked elevation in both ubiquitination and membrane localization of APPL1 as early as 5 min after stimulation (Figures 1D and 1E), and this ubiquitination was Lys63 - but not Lys48 -linked (Figure 1E). The dynamic changes in insulin-induced ubiquitination of endogenous  c The Authors Journal compilation  c 2013 Biochemical Society

210

K. K. Y. Cheng and others

APPL1 in hepatocytes were similar to those of APPL1 membrane localization and Akt phosphorylation (Supplementary Figure S2 at http://www.biochemj.org/bj/455/bj4550207add.htm), suggesting that these events are closely related. Insulin-induced APPL1 ubiquitination was completely blocked by the PI3K inhibitor LY294002 (Supplementary Figure S3 at http://www. biochemj.org/bj/455/bj4550207add.htm). Furthermore, subcellular fractionation followed by immunoprecipitation assay demonstrated that the plasma membrane fraction of APPL1 was more efficiently ubiquitinated when compared with cytoplasmic fractions under basal conditions (Figure 2). Noticeably, insulin treatment promoted APPL1 ubiquitination in the plasma membrane fraction, but not in cytoplasmic fraction (Figure 2). Figure 2 Membrane-bound APPL1 is abundantly ubiquitinated and further enhanced by insulin stimulation Primary mouse hepatocytes were infected with adenovirus encoding FLAG-tagged APPL1 and HA-tagged Ub for 24 h, followed by serum starvation for 12 h. The starved cells treated with or without insulin (Ins; 10 nM) for 5 min were subjected to subcellular fractionation to isolate plasma membrane (PM) and cytoplasmic (C) proteins. Note that successful subcellular fractionation was confirmed by the immunoblotting of a plasma membrane marker (anti-pan-cadherin) and cytoplasmic marker (anti-β-tubulin). As the expression levels of APPL1 in different subcellular locations are influenced by insulin stimulation, the level of the APPL1 protein in each fraction was first measured by immunoblotting with an anti-FLAG antibody (left-hand panel). Equal amounts of APPL1 proteins from each fraction were subjected to immunoprecipitation (IP) using anti-FLAG antibody, followed by immunoblotting with an anti-HA antibody for analysis of APPL1 ubiquitination (right-hand panel). The experiments were repeated three times and representative images are shown.

Figure 3

Identification of Lys160 as a major site of APPL1 ubiquitination

To determine which lysine residue(s) of APPL1 is the Ub acceptor site(s), we next performed a series of deletion and mutation analysis. Full-length APPL1, but not a mutant lacking the BAR domain of APPL1, was ubiquitinated (Figures 3A and 3B). Prediction of the potential Ub acceptor site(s) using Ubsite (http://www.ubpred.org/) showed that Lys6 , Lys63 , Lys70 and Lys160 within the BAR domain were possibly ubiquitinated. Replacement of Lys160 with arginine (K160R) largely abrogated both basal and insulin-stimulated ubiquitination of APPL1, whereas mutation of either Lys6 , Lys63 or Lys70 to arginine had no obvious effect (Figures 3C and 3D). Furthermore, blunted ubiquitination of the K160R mutant was accompanied by decreased plasma membrane localization of APPL1 under both basal and insulin-stimulated

Lys160 in the BAR domain is the major acceptor site for ubiquitination of APPL1 and is important for its membrane localization

(A) Schematic diagram of FLAG-tagged APPL1 and its truncated mutant containing PH and PTB domains (FLAG-PH-PTB) used for ubiquitination assay. (B and C) HEK-293 cells were co-transfected with plasmids encoding HA–Ub and FLAG–APPL1, FLAG–PH-PTB, FLAG–APPL1 mutants (K160R, K6R, K63R and K70R) or an empty vector as negative control (-ve) for 48 h, followed by immunoprecipitation (IP) with an anti-FLAG antibody and immunoblotting to detect APPL1 ubiquitination as in Figure 1(A). (D) Primary mouse hepatocytes were infected with adenovirus expressing HA–Ub and FLAG–APPL1 or the FLAG–K160R mutant for 24 h, followed by starvation for 12 h. The starved cells were treated with or without insulin (Ins; 10 nM) for 5 min, followed by analysis of APPL1 ubiquitination. (E) Primary mouse hepatocytes infected with adenovirus encoding FLAG–APPL1, FLAG–K160R or luciferase (Luci) treated with or without insulin (10 nM) for 15 min were subjected to subcellular fractionation. Total lysate (T) and plasma membrane (PM) proteins were subjected to immunoblotting with an anti-FLAG or anti-Akt antibody. (F) Fold change of membrane-bound Akt relative to basal levels as quantified by densitometry of (E). (G) HEK-293 cells were co-transfected with HA-tagged Akt2 along with FLAG–APPL1 or its FLAG–K160R mutant or an empty vector as negative control (-ve) for 48 h, followed by IP using anti-FLAG antibody. The total lysates and immunocomplex were subjected to immunoblotting using the antibodies as indicated. All experiments were repeated at least three times and representative images are shown.  c The Authors Journal compilation  c 2013 Biochemical Society

Ubiquitination of APPL1 and insulin actions

Figure 4

211

Ubiquitination of APPL1 at Lys160 is required for its insulin-sensitizing actions in hepatocytes

Primary mouse hepatocytes were transduced with adenovirus encoding APPL1 or its mutant K160R, or luciferase (Luci) control for 24 h, followed by serum starvation for 12 h. (A) The starved cells were stimulated with insulin (Ins; 10 nM) for the indicated time points and subjected to immunoblotting using an anti-APPL1 or anti-[p-Akt (Ser473 )]/-(total Akt), or anti-[p-GSK3β (Ser9 )]/-(total GSK3β) or anti-APPL1 antibody. The histogram shows fold change of phosphorylation against total relative to basal levels as quantified by densitometry. (B) The starved cells were treated with dexamethasone (Dex)/cAMP and/or insulin (Ins) for 6 h as described in the Experimental section. Glucose levels in the conditioned medium are expressed as fold change relative to the basal levels. (C) The mRNA expression of PEPCK and G6Pase in hepatocytes treated with dexamethasone/cAMP and/or insulin was quantified by real-time PCR and normalized against GAPDH. *P < 0.05 (n = 5–6). N.S., not significant.

conditions (Figure 3E). Importantly, the promoting effect of APPL1 overexpression on membrane targeting of Akt under insulin-stimulated condition was abrogated by the K160R mutant (Figures 3E and 3F). On the other hand, APPL1 and its K160R mutant exhibited a comparable binding activity to Akt2 (Figure 3G). Taken together, these findings suggest that Lys160 is the major site of APPL1 ubiquitination and is indispensable for its membrane localization. Ubiquitination of APPL1 at Lys160 is required for its insulin-sensitizing actions

To investigate whether ubiquitination of APPL1 correlates with the ability of this adaptor protein to sensitize insulin signalling, we transduced primary mouse hepatocytes with adenovirus encoding APPL1, the K160R mutant and luciferase as control. The APPL1 protein levels in cells with ectopic expression of both APPL1 and K160R were increased by approximately 2.5-fold relative to endogenous APPL1 (Figure 4A). Consistent with previous reports [7,11], ectopic overexpression of APPL1 enhanced insulininduced phosphorylation of Akt and its downstream substrate GSK3β when compared with cells expressing luciferase control, whereas overexpression of the K160R mutant did not have such an insulin-sensitizing activity (Figure 4A). Likewise, the ability of APPL1 to potentiate the inhibitory effects of insulin on glucose production and expression of the two glucogenogenic genes (PEPCK and G6Pase) was also abrogated in the K160R mutant (Figures 4B and 4C).

To explore the physiological relevance of the above findings, we injected adenovirus encoding APPL1, the K160R mutant or luciferase control into C57 BL/6N mice fed with a HFD for 12 weeks. Hepatic expression of APPL1 was elevated by approximately 2.5-fold in mice administrated with APPL1and the K160R mutant-expressing adenovirus when compared with the luciferase control (Figure 5A). Consistent with our previous study [7], hepatic overexpression of APPL1 significantly attenuated HFD-induced elevation of plasma glucose and insulin levels in both fed and fasted states (Table 1). However, such glucose- and insulin-lowering effects were not observed in mice with hepatic overexpression of the K160R mutant (Table 1). Insulin and glucose tolerance tests showed a marked alleviation of HFD-induced systemic insulin resistance and glucose intolerance in mice with hepatic overexpression of APPL1, but not the K160R mutant (Figures 5B and 5C). In addition, insulin levels during glucose tolerance test were significantly lower in HFD-fed mice with overexpression of APPL1, but not the K160R mutant, when compared with luciferase controls (Figure 5D). Similarly, the suppressive effects of APPL1 on gluconeogenesis following intraperitoneal injection of pyruvate were also abrogated by the K160R mutation (Figure 5E). Consistent with the above findings in primary hepatocytes (Figure 4), HFD-fed mice with overexpression of APPL1, but not the K160R mutant, exhibited an obvious reduction in expression of the glucogenogenic genes (PEPCK and G6Pase) and a significant augmentation of insulinstimulated phosphorylation of Akt and GSK3β in the liver tissues when compared with mice with ectopic expression of luciferase  c The Authors Journal compilation  c 2013 Biochemical Society

212

Figure 5

K. K. Y. Cheng and others

The K160R mutant of APPL1 exhibits impaired insulin-sensitizing activities in mice

C57/BL6N mice (16 weeks old) fed with STC or HFD were injected with adenovirus encoding luciferase (Luci) control, APPL1 or its mutant K160R. (A) Immunoblotting analysis of the expression levels of APPL1 in liver tissues collected after 7 days of adenoviral injection. The histogram shows densitometric quantification of APPL1 expression levels. (B) Insulin tolerance test on day 5 after adenoviral injection. Data are expressed as the percentage of baseline blood glucose values. (C and D) Glucose tolerance test (C) and glucose-stimulated insulin secretion (D) on day 9 after adenoviral infection. (E) Pyruvate tolerance test on day 12. (F) The mRNA expression of PEPCK and G6Pase in the liver of the overnight-fasted mice was quantified by real-time qPCR and normalized against GAPDH. (G) Liver tissues from overnight-fasted mice injected with or without insulin (Ins) (0.75 unit/kg of body weight) for 10 min were subjected to immunoblotting using an anti-APPL1 or anti-[p-Akt (Ser473 )]/-(total Akt), or anti-[p-GSK3β (Ser9 )]/-(total GSK3β) antibody. The histogram shows fold change of phosphorylation against the basal levels as quantified by densitometry. *P < 0.05; against the HFD-Luci group (n = 5 in each group). N.S., not significant.

(Figures 5F and 5G), suggesting that ubiquitination of APPL1 at Lys160 is important for its insulin-sensitizing activity both in vitro and in vivo.

TRAF6 is an E3 ligase responsible for APPL1 ubiquitination

Global protein–protein interaction network analysis (http://www. genome.ucsc.edu/cgi-bin/hgNear) [22] showed a possible association of APPL1 with TRAF6, an E3 Ub ligase that catalyses the formation of non-canonical Lys63 -linked polyubiquitin chains [23]. We further confirmed the interaction between APPL1 and TRAF6 using co-immunoprecipitation analysis, and this interaction was enhanced by insulin stimulation in hepatocytes (Figure 6A). Overexpression of TRAF6 robustly enhanced ubiquitination of APPL1, whereas the expression level of this adaptor  c The Authors Journal compilation  c 2013 Biochemical Society

protein was not altered (Figure 6B). By contrast, overexpression of the C70A mutant of TRAF6, a mutant with defective ligase activity, did not enhance ubiquitination of APPL1 (Figure 6B), suggesting that Ub ligase activity of TRAF6 is essential for APPL1 ubiquitination. In hepatocytes, overexpression of TRAF6 promoted insulin-induced ubiquitination of endogenous APPL1 (Figure 6C), whereas reduction of TRAF6 expression by RNAi silencing attenuated this insulin action (Figure 6D), suggesting that APPL1 is an endogenous substrate of TRAF6 E3 ligase.

TRAF6 is required for the potentiating effects of APPL1 on hepatic actions of insulin

To investigate whether TRAF6 regulates the insulin-sensitizing actions of APPL1, we transduced primary hepatocytes with

Ubiquitination of APPL1 and insulin actions Table 1 Metabolic profiles of STC- and HFD-fed mice with adenovirusmediated expression of luciferase (Luci), APPL1 or its K160R mutant Blood samples were collected from 16-week-old STC- and HFD-fed mice after 5 days of adenoviral injection under fed and fasting (16 h) conditions. Glucose and insulin levels were measured using a glucose meter and insulin ELISA kit respectively. *P < 0.05 against the HFD-Luci group (n = 5 per each group). Treatment

STC-Luci

HFD-Luci

HFD-APPL1

HFD-K160R

Fed glucose (mM) Fasting glucose (mM) Fed insulin (ng/ml) Fasting insulin (ng/ml)

7.10 + − 0.43* 4.05 + − 0.10* 0.45 + − 0.031* 0.075 + − 0.0021*

8.65 + − 0.45 5.23 + − 0.45 2.58 + − 0.39 0.40 + − 0.063

7.2 + − 0.30* 4.55 + − 0.21* 0.82 + − 0.41* 0.095 + − 0.03*

8.15 + − 1.10 5.30 + − 0.39 3.14 + − 0.61 0.57 + − 0.12

adenovirus encoding APPL1 or luciferase plus TRAF6 RNAi (TRAF6-1) or scramble control RNAi (Supplementary Table S1). Subcellular fractionation analysis showed a dramatic reduction in insulin-evoked membrane recruitment of APPL1 and Akt in hepatocytes with RNAi-mediated reduction of TRAF6 expression as compared with scramble control (Figures 7A and 7B). In addition, knockdown of TRAF6 expression significantly attenuated insulin-stimulated phosphorylation of Akt and its downstream substrates GSK3β and FOXO1 (Figures 7C and 7D), thereby leading to impaired suppression of insulin on glucose production (Figure 7E) and mRNA and protein expressions of gluconeogenic genes (Figures 7F and 7G). Knockdown of TRAF6 with another RNAi targeting against different sequences of TRAF6 (TRAF6-2, Supplementary Table S1) also diminished insulin-elicited phosphorylation of Akt and its downstream substrates FOXO1 and GSK3β (results not shown). Furthermore, the potentiating effects of APPL1 overexpression on insulinevoked phosphorylation of Akt/GSK3β and on inhibition of hepatic gluconeogenesis were largely abolished by downregulation of TRAF6 (Supplementary Figure S4 at http://www. biochemj.org/bj/455/bj4550207add.htm).

Figure 6

213

DISCUSSION

A large body of evidence suggests that APPL1 is an important regulator of both insulin secretion and insulin actions by promoting Akt activation in multiple insulin-responsive tissues [7,12,13,15]. Decreased APPL1 expression and reduced interaction between APPL1 and Akt have been found to be causally associated with βcell dysfunction and insulin resistance [7,12,13,15]. However, the detailed mechanism whereby APPL1 potentiates insulin-evoked Akt activation remains obscure. The present study provides both ex vivo and in vivo evidence demonstrating that Lys63 linked ubiquitination of APPL1 is a crucial step for insulinevoked membrane recruitment and activation of Akt, which in turn suppresses hepatic gluconeogenesis. Furthermore, we have identified TRAF6 as a crucial E3 ligase for APPL1 ubiquitination and an important component of insulin actions in hepatocytes. Ubiquitination is the enzymatic post-translational modification process that mediates the covalent conjugation of Ub, a highly conserved polypeptide of 76 amino acids, to protein substrates [24]. Ubiquitination plays important roles in many fundamental biological processes, including cell cycle control, differentiation, metabolism and immune responses, by modulation of protein stability, activity and localization [24]. Although Lys48 -linked polyubiquitination mainly targets the substrate proteins for proteasome-mediated degradation, Ubmediated protein trafficking events are largely governed by Lys63 linked polyubiquitination [24]. The present study demonstrates that Lys63 -linked ubiquitination acts as a sorting signal directing APPL1 and Akt from the cytosol to the plasma membrane for further activation. Insulin stimulation leads to a marked induction of Lys63 -linked ubiquitination of APPL1 in both mouse livers and cultured hepatocytes. Ubiquitinated APPL1 is highly enriched in the membrane fraction, but is hardly detectable in the cytosol. On the other hand, impairment of APPL1 ubiquitination by either mutagenesis or knockdown of TRAF6 expression leads to diminished insulin-dependent plasma membrane recruitment of APPL1 and Akt. As ubiquitination is not required for the binding of APPL1 to Akt, it is likely that the primary role of APPL1 ubiquitination is to facilitate the membrane targeting of the APPL1–Akt complex. Notably, Akt itself is also constitutively

TRAF6 E3 ligase mediates APPL1 ubiquitination

(A) Primary mouse hepatocytes were infected with adenovirus expressing FLAG-tagged APPL1 and Myc-tagged TRAF6 for 24 h and then starved for 12 h, followed by insulin (Ins; 10 nM) stimulation for the indicated time points. Cells were subjected to immunoprecipitation (IP) and immunoblotting using the indicated antibodies. (B) HEK-293 cells were co-transfected with plasmids encoding HA–Ub and FLAG–APPL1 along with Myc–TRAF6 or its mutant C70A (Myc–C70A) or an empty vector as control (Ctrl) for 48 h, followed by IP and immunoblotting to detect APPL1 ubiquitination. (C) Primary mouse hepatocytes were infected with adenovirus expressing HA–Ub together with luciferase (Luci) or Myc–TRAF6 for 24 h. (D) Primary mouse hepatocytes were infected with adenovirus encoding HA–Ub along with scramble or TRAF6-1 RNAi for 24 h. (C and D) The infected cells were serum-starved for 12 h and then stimulated with or without insulin (Ins; 10 nM) for 5 min, followed by immunoprecipitation and immunoblotting with the antibodies indicated. All experiments were repeated at least three times and representative images are shown.  c The Authors Journal compilation  c 2013 Biochemical Society

214

Figure 7

K. K. Y. Cheng and others

TRAF6 regulates insulin sensitivity in hepatocytes

Primary mouse hepatocytes were transduced with adenovirus encoding scramble or TRAF6-1 RNAi (Supplementary Table S1 at http://www.biochemj.org/bj/455/bj4550207add.htm) for 24 h, followed by serum starvation for 12 h. (A) Cells treated with or without insulin (Ins; 10 nM) for 15 min were subjected to subcellular fractionation. The total lysates (T), cytoplasmic (C) and plasma membrane (PM) fractions were subjected to immunoblotting using an anti-APPL1, anti-Akt, anti-β-tubulin or anti-pan-cadherin antibody. (B) Fold change of membrane-bound APPL1 and Akt relative to basal levels as quantified by densitometry. (C) The infected hepatocytes were serum starved and treated with insulin (10 nM) for the indicated time points. The total lysates were subjected to immunoblotting using an anti-[p-Akt (Ser473 )]/-(total Akt), anti-[p-GSK3β (Ser9 )]/-(total GSK3β), anti-[p-FOXO1 (Ser256 )]/-(total FOXO1) or anti-TRAF6 antibody. (D) Fold change of phosphorylation against the total relative to basal levels as quantified by densitometry in (C). (E–G) The effects of insulin on dexamethasone (Dex)/cAMP-induced glucose production (E), mRNA expression of PEPCK and G6Pase (F), and protein expression of PEPCK (G) were analysed. *P < 0.05 (n = 5). N.S., not significant.

modified by Lys63 -linked ubiquitination at Lys8 and Lys14 of its PH domain, and this modification is obligatory for IGF-I (insulin-like growth factor 1)-induced membrane recruitment and activation of Akt [23]. The human cancer-associated Akt mutant displays an increase in Akt ubiquitination, in turn contributing to the enhancement of Akt membrane localization and phosphorylation [23]. In contrast, the ubiquitination-defective Akt mutant loses its ability to bind to PtdIns(3,4,5)P3 and membranes in vitro [23]. Since ubiquitination is not required for the binding of APPL1 to Akt, it appears that ubiquitination of APPL1 constitutes a vital step for membrane targeting of the APPL1–Akt complex upon insulin stimulation. A unique feature of APPL1 is that it shuttles between different subcellular compartments in response to various extracellular  c The Authors Journal compilation  c 2013 Biochemical Society

stimuli [7,12,18]. Although EGF (epidermal growth factor) promotes the translocation of APPL1 from the early endosomes to nuclei [18], insulin induces the recruitment of APPL1 from cytosol to endosomes and plasma membrane [7,12]. The membrane targeting properties of APPL1 are attributed to its phosphoinositide-binding ability via its BAR domain [25,26]. The BAR domain of APPL1 forms crescent-shaped homodimers with a cluster of positively charged residues at the surface [25], thereby enabling its interaction with negatively charged phospholipids in the membrane. Several mutations in the BAR domain of APPL1 abolishes its membrane-targeting ability, leading to impaired function of APPL1 [9]. Notably, we found that Lys160 , a major ubiquitination acceptor site that is responsive to insulin, is also located within the BAR domain. Although the

Ubiquitination of APPL1 and insulin actions

K160R mutant of APPL1 loses its membrane localization and fails to potentiate insulin-evoked Akt activation, it is still able to bind both TRAF6 (Supplementary Figure S5 at http://www. biochemj.org/bj/455/bj4550207add.htm) and Akt (Figure 3G), suggesting that the primary role of ubiquitination at Lys160 is to facilitate BAR-domain-mediated membrane-targeting events. E3 Ub ligases are key enzymes mediating the addition of Ub to substrates and thus determining the type and extent of Ub modifications [24]. Mounting evidence suggests that Akt activation triggered by diverse growth factors is mediated by different E3 ligases [23,27]. Although EGF-evoked Akt activation is dependent on Skp2-SCF (Skp1/cullin/F-box) E3 ligase [27], TRAF6 E3 ligase serves as an upstream regulator of IGF-I-induced Akt activation [23]. A deficiency of TRAF6 attenuates IGF-I-induced Akt membrane recruitment and activation in mouse embryonic fibroblasts [23]. Likewise, the present study demonstrates that insulin-induced APPL1 ubiquitination is accompanied by increased interaction between APPL1 and TRAF6. Furthermore, insulin-stimulated APPL1 ubiquitination, membrane recruitment of APPL1 and Akt and suppression of gluconeogenesis are all impaired in hepatocytes with reduced TRAF6 expression, suggesting that TRAF6 is an obligatory signalling component that mediates insulin actions in hepatocytes, by promoting ubiquitination-dependent membrane targeting of the APPL1–Akt complex. Although membrane targeting of Akt induced by IGF-I is regulated by TRAF6mediated ubiquitination of Akt [23], we did not observe any change in Akt ubiquitination upon insulin stimulation in primary mouse hepatocytes (results not shown). In addition to TRAF6, TRAF2 has been shown to mediate glucagon-induced hepatic glucose production by potentiating cAMP-responseelement-binding protein phosphorylation and gluconeogenic gene expression [28]. Therefore the members of the TRAF family may act in a concerted manner to maintain glucose homoeostasis by modulating hepatic signalling pathways of both catabolic and anabolic hormones. In summary, the present study identifies the TRAF6/APPL1 axis as a novel signalling component that mediates hepatic actions of insulin by promoting Lys63 -linked ubiquitination and membrane targeting of the APPL1–Akt complex (Supplementary Figure S6 at http://www.biochemj.org/bj/455/bj4550207add.htm). Further characterization of the detailed molecular and structural basis whereby TRAF6-mediated APPL1 ubiquitination potentiates insulin signalling cascades may help to understand the pathogenesis of hepatic insulin resistance, a major contributor to fasting hyperglycaemia in Type 2 diabetes. Although the present study uncovers the role of Lys63 -linked ubiquitination in insulin signalling by modulating the membrane targeting of APPL1 and Akt, E3 Ub ligase MG53-mediated Lys48 ubiquitination has recently been shown to induce insulin resistance and metabolic disorders by promoting the degradation of the insulin receptor and insulin receptor substrate 1 [29]. Therefore, ubiquitination may, similar to phosphorylation, act as a critical regulator of insulin actions by modulating both the stability and localization of the major insulin signalling components.

AUTHOR CONTRIBUTION Kenneth Cheng researched data and wrote the paper. Karen Lam supervised the study and edited the paper prior to submission. Yu Wang contributed to discussion and edited the paper prior to submission. Donghai Wu researched data and edited the paper prior to submission. Mingliang Zhang, Baile Wang, Xiaomu Li, Ruby Hoo and Zhe Huang researched data. Gary Sweeney reviewed the paper prior to submission. Aimin Xu designed and supervised the study, and wrote the paper.

215

FUNDING This work was supported by the National Basic Research Program of China [grant number 2011CB504004] and its matching fund from the University of Hong Kong, the General Research Fund [grant number HKU 783010M and 782612 (to A.X.)] and a Collaborative Research Fund [grant number HKU4/CRF/10] from the Research Grants Council of Hong Kong, and the National Science Foundation of China [grant number 81270881 (to K.K.Y.C.)].

REFERENCES 1 Saltiel, A. R. and Kahn, C. R. (2001) Insulin signalling and the regulation of glucose and lipid metabolism. Nature 414, 799–806 2 Stephens, L., Anderson, K., Stokoe, D., Erdjument-Bromage, H., Painter, G. F., Holmes, A. B., Gaffney, P. R., Reese, C. B., McCormick, F., Tempst, P. et al. (1998) Protein kinase B kinases that mediate phosphatidylinositol 3,4,5-trisphosphate-dependent activation of protein kinase B. Science 279, 710–714 3 Sarbassov, D. D., Guertin, D. A., Ali, S. M. and Sabatini, D. M. (2005) Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098–1101 4 Hussain, K., Challis, B., Rocha, N., Payne, F., Minic, M., Thompson, A., Daly, A., Scott, C., Harris, J., Smillie, B. J. et al. (2011) An activating mutation of AKT2 and human hypoglycemia. Science 334, 474 5 Yang, W. L., Wu, C. Y., Wu, J. and Lin, H. K. (2010) Regulation of Akt signaling activation by ubiquitination. Cell Cycle 9, 487–497 6 Sundaresan, N. R., Pillai, V. B., Wolfgeher, D., Samant, S., Vasudevan, P., Parekh, V., Raghuraman, H., Cunningham, J. M., Gupta, M. and Gupta, M. P. (2011) The deacetylase SIRT1 promotes membrane localization and activation of Akt and PDK1 during tumorigenesis and cardiac hypertrophy. Sci. Signal 4, ra46 7 Cheng, K. K., Iglesias, M. A., Lam, K. S., Wang, Y., Sweeney, G., Zhu, W., Vanhoutte, P. M., Kraegen, E. W. and Xu, A. (2009) APPL1 potentiates insulin-mediated inhibition of hepatic glucose production and alleviates diabetes via Akt activation in mice. Cell Metab. 9, 417–427 8 Mitsuuchi, Y., Johnson, S. W., Sonoda, G., Tanno, S., Golemis, E. A. and Testa, J. R. (1999) Identification of a chromosome 3p14.3-21.1 gene, APPL, encoding an adaptor molecule that interacts with the oncoprotein-serine/threonine kinase AKT2. Oncogene 18, 4891–4898 9 Schenck, A., Goto-Silva, L., Collinet, C., Rhinn, M., Giner, A., Habermann, B., Brand, M. and Zerial, M. (2008) The endosomal protein Appl1 mediates Akt substrate specificity and cell survival in vertebrate development. Cell 133, 486–497 10 Zoncu, R., Perera, R. M., Balkin, D. M., Pirruccello, M., Toomre, D. and De Camilli, P. (2009) A phosphoinositide switch controls the maturation and signaling properties of APPL endosomes. Cell 136, 1110–1121 11 Cleasby, M. E., Lau, Q., Polkinghorne, E., Patel, S. A., Leslie, S. J., Turner, N., Cooney, G. J., Xu, A. and Kraegen, E. W. (2011) The adaptor protein APPL1 increases glycogen accumulation in rat skeletal muscle through activation of the PI3-kinase signalling pathway. J. Endocrinol. 210, 81–92 12 Saito, T., Jones, C. C., Huang, S., Czech, M. P. and Pilch, P. F. (2007) The interaction of Akt with APPL1 is required for insulin-stimulated Glut4 translocation. J. Biol. Chem. 282, 32280–32287 13 Wang, Y., Cheng, K. K., Lam, K. S., Wu, D., Huang, Y., Vanhoutte, P. M., Sweeney, G., Li, Y. and Xu, A. (2011) APPL1 counteracts obesity-induced vascular insulin resistance and endothelial dysfunction by modulating the endothelial production of nitric oxide and endothelin-1 in mice. Diabetes 60, 3044–3054 14 Fang, X., Palanivel, R., Cresser, J., Schram, K., Ganguly, R., Thong, F. S., Tuinei, J., Xu, A., Abel, E. D. and Sweeney, G. (2010) An APPL1-AMPK signaling axis mediates beneficial metabolic effects of adiponectin in the heart. Am. J. Physiol. Endocrinol. Metab. 299, E721–E729 15 Cheng, K. K., Lam, K. S., Wu, D., Wang, Y., Sweeney, G., Hoo, R. L., Zhang, J. and Xu, A. (2012) APPL1 potentiates insulin secretion in pancreatic β cells by enhancing protein kinase Akt-dependent expression of SNARE proteins in mice. Proc. Natl. Acad. Sci. U.S.A. 109, 8919–8924 16 Mao, X., Kikani, C. K., Riojas, R. A., Langlais, P., Wang, L., Ramos, F. J., Fang, Q., Christ-Roberts, C. Y., Hong, J. Y., Kim, R. Y. et al. (2006) APPL1 binds to adiponectin receptors and mediates adiponectin signalling and function. Nat. Cell. Biol. 8, 516–523 17 Cheng, K. K., Lam, K. S., Wang, Y., Huang, Y., Carling, D., Wu, D., Wong, C. and Xu, A. (2007) Adiponectin-induced endothelial nitric oxide synthase activation and nitric oxide production are mediated by APPL1 in endothelial cells. Diabetes 56, 1387–1394 18 Miaczynska, M., Christoforidis, S., Giner, A., Shevchenko, A., Uttenweiler-Joseph, S., Habermann, B., Wilm, M., Parton, R. G. and Zerial, M. (2004) APPL proteins link Rab5 to nuclear signal transduction via an endosomal compartment. Cell 116, 445–456  c The Authors Journal compilation  c 2013 Biochemical Society

216

K. K. Y. Cheng and others

19 Pilecka, I., Sadowski, L., Kalaidzidis, Y. and Miaczynska, M. (2011) Recruitment of APPL1 to ubiquitin-rich aggresomes in response to proteasomal impairment. Exp. Cell Res. 317, 1093–1107 20 Liu, M., Zhou, L., Wei, L., Villarreal, R., Yang, X., Hu, D., Riojas, R. A., Holmes, B. M., Langlais, P. R., Lee, H. et al. (2012) Phosphorylation of adaptor protein containing pleckstrin homology domain, phosphotyrosine binding domain, and leucine zipper motif 1 (APPL1) at Ser430 mediates endoplasmic reticulum (ER) stress-induced insulin resistance in hepatocytes. J. Biol. Chem. 287, 26087–26093 21 Kamitani, T., Kito, K., Nguyen, H. P. and Yeh, E. T. (1997) Characterization of NEDD8, a developmentally down-regulated ubiquitin-like protein. J. Biol. Chem. 272, 28557–28562 22 Kent, W. J., Hsu, F., Karolchik, D., Kuhn, R. M., Clawson, H., Trumbower, H. and Haussler, D. (2005) Exploring relationships and mining data with the UCSC Gene Sorter. Genome Res. 15, 737–741 23 Yang, W. L., Wang, J., Chan, C. H., Lee, S. W., Campos, A. D., Lamothe, B., Hur, L., Grabiner, B. C., Lin, X., Darnay, B. G. et al. (2009) The E3 ligase TRAF6 regulates Akt ubiquitination and activation. Science 325, 1134–1138 Received 10 June 2013/5 August 2013; accepted 5 August 2013 Published as BJ Immediate Publication 5 August 2013, doi:10.1042/BJ20130760

 c The Authors Journal compilation  c 2013 Biochemical Society

24 MacGurn, J. A., Hsu, P. C. and Emr, S. D. (2012) Ubiquitin and membrane protein turnover: from cradle to grave. Annu. Rev. Biochem. 81, 231–259 25 Li, J., Mao, X., Dong, L. Q., Liu, F. and Tong, L. (2007) Crystal structures of the BAR-PH and PTB domains of human APPL1. Structure 15, 525–533 26 Chial, H. J., Wu, R., Ustach, C. V., McPhail, L. C., Mobley, W. C. and Chen, Y. Q. (2008) Membrane targeting by APPL1 and APPL2: dynamic scaffolds that oligomerize and bind phosphoinositides. Traffic 9, 215–229 27 Chan, C. H., Li, C. F., Yang, W. L., Gao, Y., Lee, S. W., Feng, Z., Huang, H. Y., Tsai, K. K., Flores, L. G., Shao, Y. et al. (2012) The Skp2-SCF E3 ligase regulates Akt ubiquitination, glycolysis, herceptin sensitivity, and tumorigenesis. Cell 149, 1098–1111 28 Chen, Z., Sheng, L., Shen, H., Zhao, Y., Wang, S., Brink, R. and Rui, L. (2012) Hepatic TRAF2 regulates glucose metabolism through enhancing glucagon responses. Diabetes 61, 566–573 29 Song, R., Peng, W., Zhang, Y., Lv, F., Wu, H. K., Guo, J., Cao, Y., Pi, Y., Zhang, X., Jin, L. et al. (2013) Central role of E3 ubiquitin ligase MG53 in insulin resistance and metabolic disorders. Nature 494, 375–379

Biochem. J. (2013) 455, 207–216 (Printed in Great Britain)

doi:10.1042/BJ20130760

SUPPLEMENTARY ONLINE DATA

TRAF6-mediated ubiquitination of APPL1 enhances hepatic actions of insulin by promoting the membrane translocation of Akt Kenneth K. Y. CHENG*†, Karen S. L. LAM*†, Yu WANG†‡, Donghai WU§, Mingliang ZHANG*, Baile WANG*†, Xiaomu LI*†, Ruby L. C. HOO*†, Zhe HUANG*†, Gary SWEENEY and Aimin XU*†‡1 *Department of Medicine, University of Hong Kong, L8-39, 21 Sassoon Road, Hong Kong, †State Key Laboratory of Pharmaceutical Biotechnology, University of Hong Kong, 21 Sassoon Road, Hong Kong, ‡Department of Pharmacology & Pharmacy, University of Hong Kong, L2-53, 21 Sassoon Road, Hong Kong, §The Key Laboratory of Regenerative Biology, Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou 510530, China, and Department of Biology, York University, Toronto, Ontario, Canada M3J 1P3

Figure S1 Effect of the proteasome inhibitor MG-132 on APPL1 ubiquitination HEK-293 cells were transfected with plasmids encoding HA-tagged Ub and/or FLAG-tagged APPL1 for 36 h, followed by serum starvation for 12 h. The starved cells were treated with or without MG-132 (10 μM) for 4 h. Total lysates were subjected to immunoprecipitation (IP) with an anti-FLAG antibody and immunoblotting with an anti-HA antibody for analysis of APPL1 ubiquitination. Representative immunoblots from three independent experiments are shown.

Figure S2 Insulin-elicited ubiquitination and membrane recruitment of APPL1 are closely associated with Akt activation (A) Primary mouse hepatocytes were infected with adenovirus encoding HA-tagged Ub for 36 h, followed by serum starvation for 12 h. The cells were treated with insulin (10 nM) for the indicated time points. The treated cells were collected for immunoprecipitation (IP) using anti-APPL1 antibody, followed by immunoblotting with anti-HA antibody to detect APPL1 ubiquitination. (B) Primary mouse hepatocytes were serum-starved for 12 h and treated with insulin for the indicated time points. The treated cells were subjected to subcellular fractionation as described in Figure 2 of the main text. The total lysate (T), membrane (M) and cytoplasmic (C) fractions were subjected to immunoblotting with an anti-APPL1 or anti-Akt antibody. Representative immunoblots from three independent experiments are shown.

1

To whom correspondence should be addressed (email [email protected]).  c The Authors Journal compilation  c 2013 Biochemical Society

K. K. Y. Cheng and others

Figure S3

Insulin-induced APPL1 ubiquitination is PI3K dependent

Primary mouse hepatocytes were infected with adenovirus encoding HA-tagged Ub for 36 h, followed by serum starvation for 12 h. The starved cells were pre-treated with or without the PI3K inhibitor LY294002 (50 μm) for 30 min, followed by insulin stimulation (Ins; 10 nM) for 5 min. Total lysates were subjected to immunoprecipitation (IP) with an anti-APPL1 antibody and immunoblotting with an anti-HA antibody for analysis of APPL1 ubiquitination. Representative immunoblots from three independent experiments are shown.

 c The Authors Journal compilation  c 2013 Biochemical Society

Ubiquitination of APPL1 and insulin actions

Figure S4

Suppression of TRAF6 abolishes the insulin-sensitizing actions of APPL1

Primary mouse hepatocytes were infected with adenovirus encoding luciferase (Luci), APPL1, TRAF6-1 RNAi or scramble control RNAi for 24 h, followed by serum starvation for 12 h. (A) Immunoblotting analysis for total lysates of the infected cells treated with insulin (Ins; 10 nM) for the indicated time points, using an anti-APPL1, anti-TRAF6, anti-[p-Akt (Ser473 )]/-(total Akt), or anti-[p-GSK3β (Ser9 )]/-(total GSK3β) antibody respectively. Note that ectopic expression of APPL1 was increased by ∼ 2.5-fold relative to endogenous protein. (B) Relative fold change of phosphorylation in (A) as quantified by densitometry. (C) Glucose production in the infected hepatocytes treated with dexamethasone (Dex)/cAMP/insulin as measured in Figure 4 of the main text. (D) The gene expression levels of PEPCK and G6Pase normalized against GAPDH in the infected hepatocytes treated with dexamethasone/cAMP/insulin for 6 h (fold change over the group of scramble RNAi + luciferase). *P < 0.05 (n = 5). N.S., not significant.

Figure S5 TRAF6

APPL1 and its K160R mutant exhibit a similar binding ability to

HEK-293 cells were co-transfected with plasmids encoding Myc-tagged TRAF6 and with FLAG-tagged APPL1 or its FLAG-tagged K160R mutant (FLAG-K160R) or an empty vector as negative control (-ve) for 48 h, followed by immunoprecipitation (IP) with an anti-FLAG antibody. The immunocomplex or total lysate was subjected to immunoblotting using indicated antibodies. Representative images are shown from three independent experiments.

 c The Authors Journal compilation  c 2013 Biochemical Society

K. K. Y. Cheng and others

Figure S6 Schematic diagram showing how ubiquitination of APPL1 potentiates insulin actions by facilitating the membrane localization and activation of Akt in hepatocytes Insulin stimulation induces the association of TRAF6 with APPL1 and Lys63 -linked ubiquitination of APPL1 at Lys160 , which in turn promotes the membrane targeting of the APPL1–Akt complex to the membrane for further activation by its upstream kinases PDK1 and mTORC1, consequently leading to suppression of gluconeogenic gene expression. IR, insulin receptor; IRS, insulin receptor substrate.

Table S1

The oligonucleotides used for site-directed mutagenesis and RNAi in the present study Name

Oligonucleotide sequence (5 →3 )

Scramble (forward) Scramble (reverse) TRAF6-1 (forward) TRAF6-1 (reverse) TRAF6-2 (forward) TRAF6-2 (reverse) K6R mutant (forward) K6R mutant (reverse) K63R mutant (forward) K63R mutant (reverse) K70R mutant (forward) K70R mutant (reverse) K160R mutant (forward) K160R mutant (reverse)

CACCGAATATTATTAAGGCGACAGAGCGAACTCTGTCGCCTTAATAATATT AAAAAATATTATTAAGGCGACAGAGTTCGCTCTGTCGCCTTAATAATATTC CACCGCTACTATGAGTCTCTTAAACCGAAGTTTAAGAGACTCATAGTAGC AAAAGCTACTATGAGTCTCTTAAACTTCGGTTTAAGAGACTCATAGTAGC CACCGCAACTTTGGGATGCACTTGACGAATCAAGTGCATCCCAAAGTTGC AAAAGCAACTTTGGGATGCACTTGATTCGTCAAGTGCATCCCAAAGTTGC ATGCCGGGGATCGACAGGCTGCCCATCGAGGAG CTCCTCGATGGGCAGCCTGTCGATCCCCGGCAT CAACACACCTGACCTCAAGGCTTTTAAAAGAATATG CATATTCTTTTAAAAGCCTTGAGGTCAGGTGTGTTG CTTTTAAAAGAATATGAAAGGCAGCGTTTTCCATTGGGAGG CCTCCCAATGGAAAACGCTGCCTTTCATATTCTTTTAAAAG GAAAATGACAAGGTGAGGTATGAAGTAACAGAAG CTTCTGTTACTTCATACCTCACCTTGTCATTTTC

Received 10 June 2013/5 August 2013; accepted 5 August 2013 Published as BJ Immediate Publication 5 August 2013, doi:10.1042/BJ20130760

 c The Authors Journal compilation  c 2013 Biochemical Society

Copyright of Biochemical Journal is the property of Portland Press Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.