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*

Background: Adaptor protein APPL1 plays a critical role in regulating both adiponectin and insulin signaling pathways. Results: ER stress-induced APPL1 phosphorylation at Ser430 blocks the insulin-sensitizing effect of APPL1 in a PKCα-dependent manner. Conclusion: APPL1 phosphorylation at Ser430 mediates ER stress-induced insulin resistance. Significance: PKCα-mediated APPL1 phosphorylation could be a potential drug target for the treatment of insulin resistance. APPL1 is an adaptor protein that plays a critical role in regulating adiponectin and insulin signaling. However, how APPL1 is regulated under normal and pathological conditions remains largely unknown. In this study, we show that APPL1 undergoes phosphorylation at Ser430 and that this phosphorylation is enhanced in the liver of obese mice displaying insulin resistance. In cultured mouse hepatocytes, APPL1 phosphorylation at Ser430 is stimulated by phorbol 12-myristate 13-acetate, an activator of classic PKC isoforms, and by the endoplasmic reticulum (ER) stress inducer, thapsigargin. Overexpression of wild-type but not dominant negative PKCα increases APPL1 phosphorylation at Ser430 in mouse hepatocytes. In addition, suppressing PKCα expression by shRNA in hepatocytes reduces ER stress-induced APPL1 phosphorylation at Ser430 as well as the inhibitory effect of ER stress on insulin-stimulated Akt phosphorylation. Consistent with a negative regulatory role of APPL1 phosphorylation at Ser430 in insulin signaling, overexpression of APPL1S430D but not APPL1S430A impairs the potentiating effect of APPL1 on insulin-stimulated Akt phosphorylation at Thr308. Taken together, our results identify APPL1 as a novel target in ER stress-induced insulin resistance and PKCα as the kinase mediating ER stress-induced phosphorylation of APPL1 at Ser430.

APPL1 is an adaptor protein that plays a critical role in regulating adiponectin and insulin signaling. However, how APPL1 is regulated under normal and pathological conditions remains largely unknown. In this study, we show that APPL1 undergoes phosphorylation at Ser 430 and that this phosphorylation is enhanced in the liver of obese mice displaying insulin resistance. In cultured mouse hepatocytes, APPL1 phosphorylation at Ser 430 is stimulated by phorbol 12-myristate 13-acetate, an activator of classic PKC isoforms, and by the endoplasmic reticulum (ER) stress inducer, thapsigargin. Overexpression of wild-type but not dominant negative PKC␣ increases APPL1 phosphorylation at Ser 430 in mouse hepatocytes. In addition, suppressing PKC␣ expression by shRNA in hepatocytes reduces ER stressinduced APPL1 phosphorylation at Ser 430 as well as the inhibitory effect of ER stress on insulin-stimulated Akt phosphorylation. Consistent with a negative regulatory role of APPL1 phosphorylation at Ser 430 in insulin signaling, overexpression of APPL1 S430D but not APPL1 S430A impairs the potentiating effect of APPL1 on insulin-stimulated Akt phosphorylation at Thr 308 . Taken together, our results identify APPL1 as a novel target in ER stress-induced insulin resistance and PKC␣ as the kinase mediating ER stress-induced phosphorylation of APPL1 at Ser 430 .
Adaptor protein containing pleckstrin homology domain, phosphotyrosine binding domain, and leucine zipper motif 1 (APPL1) is the first adiponectin receptor-binding protein identified to play a critical role in regulating adiponectin signaling including activation of AMPK, 3 p38 MAPK, and endothelial NOS pathways in muscle and endothelial cells (1)(2)(3)(4)(5). In addition to binding to adiponectin receptors, APPL1 has been found to interact with the catalytic subunit of phosphatidylinositol 3-kinase (p110) and Akt (6) and to control the substrate specificity of Akt toward GSK-3␤ but not TSC2 (7). The association of APPL1 and Akt appears to play a positive role in insulin signaling in liver tissue by blocking the association of Akt with its endogenous inhibitor tribble 3 (TRB3), thus promoting Akt translocation to the plasma membrane and the endosomes for further activation (8). Consistent with this, insulin-stimulated Akt activation and GLUT4 translocation were significantly reduced in APPL1-suppressed C2C12 myocytes and 3T3-L1 adipocytes (2,9).
The precise mechanisms by which APPL1 mediates adiponectin and insulin signaling remain largely unknown. Available evidence reveals that APPL1 contains multiple potential phosphorylation sites (Ser 151 , Ser 401 , Ser 427 , Ser 430 , and Ser 691/693/696 ) (10). However, the roles of the phosphorylation and the upstream kinases of APPL1 in regulating insulin sensitivity are unknown. In the current study, we demonstrate that APPL1 undergoes ER stress-induced and PKC␣-dependent phosphorylation at Ser 430 in hepatocytes. In addition, phosphorylation at Ser 430 impairs the insulin-sensitizing effect of APPL1 in hepatocytes. Our study reveals a novel mechanism by which ER stress induces insulin resistance and demonstrates that PKC␣ is a direct kinase for APPL1 phosphorylation, making PKC␣ a potential drug target for the treatment of obesity and its related diseases.

MATERIALS AND METHODS
Plasmids, Chemicals, and Antibodies-The plasmids encoding full-length human APPL1 were described previously (2). The plasmids encoding S430A and S430D mutated APPL1 were generated by site-directed mutagenesis from wild-type APPL1 plasmid. The RNAi-resistant APPL1 constructs (wildtype APPL1, S430A, and S430D plasmids) were generated by site-directed mutagenesis with a primer (5Ј-GTG GAT ATG CAC AAT AAA C-3Ј) that targets the APPL1 cDNA nucleotide sequence from 1154 to 1173 to abolish the RNAi target nucleotide sequence without changing the encoded protein sequence. Plasmids encoding wild-type and dominant negative PKC␣ were kindly provided by Dr. Jae-Won Soh (11). The PKC␣ scramble and shRNA constructs were from Santa Cruz Biotechnology (Santa Cruz, CA).
The following compounds were used in our studies. Gö 6976 and Gö 6983 were purchased from Calbiochem. Thapsigargin, recombinant PKC enzyme kits, PMA, AICAR, and H 2 O 2 were from Sigma.
Transfection of Hepatocytes-DNA constructs were introduced into mouse hepatocytes by electroporation using the Neon transfection system (Invitrogen). In brief, HepIR cells were washed with PBS and suspended in R buffer from the same company. Cell mixtures (100 l) containing about 10 6 cells and 5 g of DNA were loaded into a particular tip. Voltage was adjusted to 1,230 V, and pulse length was adjusted to 20 ms (two pulses). Thirty-six hours after transfection, the cells were serum-starved for 4 h followed with chemical treatment.
In Vivo Labeling-Mouse hepatocytes were transfected with cDNA encoding Myc-tagged APPL1. Thirty-six hours after transfection, cells were in vivo labeled with [ 32 P[orthophosphate according to a protocol described previously (12) and treated with (ϩ) or without (Ϫ) 2 M PMA for 30 min. APPL1 was immunoprecipitated with an antibody to the Myc tag, and APPL1 phosphorylation was visualized by autoradiography and quantified by using a PhosphorImager. The expression of APPL1 in these cells was confirmed by Western blot analysis using the anti-Myc antibody.
In Vitro Phosphorylation of APPL1 by PKC␣ or Other PKC Isoforms-GST-fused APPL1 was incubated with different recombinant PKC isoforms to determine the in vitro phosphorylation of APPL1 as described previously (12). Wild type and S430A mutant of APPL1 were purified by immunoprecipitation from C2C12 cells transiently expressing these proteins. Immunoprecipitated APPL1 proteins were incubated in 30 l of kinase assay buffer (50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM Na 3 VO 4 , 1 mM sodium pyrophosphate, 1 mM NaF, and 1 mM phenylmethylsulfonyl fluoride) containing purified PKC␣ and 2 Ci of [␥-32 P[ATP. In vitro phosphorylation reaction was carried out for 30 min at 30°C, and in vitro phosphorylated proteins were separated by SDS-PAGE and visualized by autoradiography.
Animal Studies-Male db/db mice and their lean controls (stock number 000642, 3-5 weeks old) were purchased from The Jackson Laboratory (Bar Harbor, ME). The mice were group-housed in a specific pathogen-free facility at 22-24°C on a 12-h light/12-h dark cycle with the lights on at 8:00 a.m. The mice were fed with standard rodent chow, and all animals had access to water ad libitum. At 12 weeks of age, mice were then sacrificed, and the tissues were isolated according to the procedure described in our recent studies (13). For high fat diet experiments, 5-week-old C57BL/6 male mice (10 mice per group) were fed with normal chow diet (Research Diets number D12328: 10.5 kcal% fat, 73.1 kcal% carbohydrate, and 16.4 kcal% protein) or high fat diet (Research Diets number D12330: 58.0 kcal% fat, 16.0 kcal% protein, and 26 kcal% carbohydrate) for 5 months. Mice were sacrificed, and the tissues were isolated as described (13). All animal procedures were approved by the University of Texas Health Science Center Animal Care and Use Committee.
Statistical Analysis-Quantification of APPL1 phosphorylation at Ser 430 or Akt phosphorylation at Thr 308 was performed by Western blot analysis using the National Institutes of Health Scion Image software and was normalized with the amount of protein expression of APPL1 or Akt, respectively, in each experiment. Results were expressed as the mean Ϯ S.E. Differences between the groups were examined for statistical significance using analysis of variance. * indicates p Ͻ 0.05; ** indicates p Ͻ 0.01.

Identification of Ser 430 as a PMA-stimulated APPL1 Phosphorylation
Site-APPL1 is involved in multiple intracellular cell signaling events including adiponectin and insulin signaling. However, how the protein is regulated in cells remains elusive. A recent mass spectrometry study revealed that APPL1 is phosphorylated at several sites including Ser 151 , Ser 401 Ser 427 , Ser 430 , and Ser 691/693/696 and that the protein could be a potential target for multiple important cellular kinases including PKC, AMPK, and MAPK family members Erk, p38 MAPK, and c-Jun N-terminal kinase (JNK) (10). To determine whether these kinases were involved in the phosphorylation of APPL1, we performed in vivo labeling experiments with C2C12 cells transiently expressing Myc-APPL1. Treating the labeled cells with the PKC activator PMA but not the AMPK activator AICAR or the MAPK activator H 2 O 2 (14) led to a marked increase in APPL1 phosphorylation (Fig. 1A), suggesting that PKC, but not AMPK and MAPK, is involved in the phosphorylation of APPL1.
To map the region of APPL1 involved in PMA-stimulated phosphorylation, C2C12 cells expressing Myc-tagged full-length or truncated forms of APPL1 were in vivo labeled with [ 32 P[orthophosphate and treated with or without PMA. We found that APPL1 (1-471) but not APPL1 (1-269) was phosphorylated in the cells in response to PMA treatment (Fig. 1B), suggesting that PMA stimulates phosphorylation of APPL1 in a region containing amino acids from 269 to 471. This region contains the pleckstrin homology (PH) domain and a fragment between the pleckstrin homology domain and the phosphotyrosine binding (PTB) domain, which contains three putative phosphorylation sites, Ser 401 , Ser 427 , and Ser 430 (10). Western blot analysis using the phosphospecific antibody to Ser 430 indicated that PMA stimulated the phosphorylation of endogenous APPL1 at Ser 430 in hepatocytes (Fig. 1C), which is consistent with a recent study showing that APPL1 is phosphorylated at Ser 430 in HEK293 cells (10).
Phosphorylation of APPL1 at Ser 430 Is Associated with Obesity and Insulin Resistance-To determine the physiological relevance of APPL1 phosphorylation at Ser 430 , we examined APPL1 phosphorylation in the liver of db/db mice. As shown in Fig. 2, A and B, APPL1 Ser 430 phosphorylation was dramatically elevated in the liver tissues of db/db mice as compared with those of lean control mice, suggesting that phosphorylation of APPL1 at Ser 430 is associated with obesity and insulin resistance. Consistent with this finding, APPL1 phosphorylation at Ser 430 was also enhanced in the liver of high fat diet-induced obese mice as compared with those of normal chow lean mice (Fig. 2, C and D).
APPL1 Phosphorylation at Ser 430 Is Mediated by PKC␣-The finding that APPL1 phosphorylation at Ser 430 is stimulated by PMA suggests that a classic PKC isoform is involved in the phosphorylation. To identify the PKC isoform involved in the phosphorylation of APPL1 at Ser 430 , different recombinant PKC isoforms were incubated with purified GST-APPL1 fusion protein in the presence of [␥-32 P[ATP. The in vitro kinase assays showed that incubation with PKC␣, but not PKC␤1 and  The phosphorylation of APPL1 at Ser 430 in liver tissue is associated with obesity and insulin resistance. A, the phosphorylation of APPL1 (p-APPL1) at Ser 430 was enhanced in the liver tissues of db/db mice. Threemonth-old db/db and the lean control mice were sacrificed, and the liver tissues were collected. The liver homogenates were used to detect the phosphorylation of APPL1 at Ser 430 and APPL1 protein level with specific antibodies. B, the ratio of phosphorylation of APPL1 at Ser 430 to APPL1 protein shown in A was quantified. The data represent mean Ϯ S.E. **, p Ͻ 0.01. C, the phosphorylation of APPL1 at Ser 430 was enhanced in the liver tissue of obese mice. High fat diet (HFD)-fed mice were sacrificed, and the liver tissues were collected. The phosphorylation of APPL1 at Ser 430 and APPL1 protein level in liver tissue were determined by Western blot with specific antibodies. ND, normal diet. D, the ratio of phosphorylation of APPL1 at Ser 430 to APPL1 protein shown in C was quantified. The data represent mean Ϯ S.E. **, p Ͻ 0.01. PKC␤2, led to marked phosphorylation of APPL1 (Fig. 3A). To determine whether PKC␣ directly catalyzes APPL1 phosphorylation at Ser 430 , in vitro kinase assays were performed using recombinant PKC␣ and immunoprecipitated wild type or S430A mutant of APPL1. We found that wild type but not the S430A mutant of APPL1 was greatly phosphorylated by PKC␣ in vitro (Fig. 3B). Consistent with these findings, pretreatment of hepatocytes with Gö 6976, a specific inhibitor of PKC␣ and -␤, or Gö 6983, a broad spectrum PKC inhibitor that inhibits PKC␣, -␤, -␥, -␦, and -, blocked PMA-stimulated phosphorylation of APPL1 at Ser 430 in mouse hepatocytes (Fig. 3C). To determine the effect of PKC␣ on phosphorylation of APPL1 at Ser 430 in intact cells, we generated stable hepatic cells in which the expression levels of PKC␣ were suppressed by shRNA. PMA-stimulated phosphorylation of APPL1 at Ser 430 was significantly reduced in the PKC␣-suppressed cells as compared with the scramble control cells (Fig. 3D), further indicating that PKC␣ is the kinase responsible for PMA-stimulated phosphorylation of APPL1 in mouse hepatocytes.

ER Stress-stimulated Phosphorylation of APPL1 at Ser 430 Is
Mediated via a PKC␣-dependent Mechanism-ER stress has been shown to contribute to obesity-induced insulin resistance by suppression of insulin receptor signaling (15) and inducing autophagy-dependent insulin receptor degradation (16). To understand the mechanisms underlying ER stress-induced insulin resistance, we treated mouse hepatocytes with thapsigargin (TG), an ER stress inducer, and PMA. TG treatment markedly enhanced the level of ER stress marker, phosphorylation of eIF2␣ at Ser 51 . In the meantime, treating the cells with TG or PMA led to a great increase in APPL1 phosphorylation at Ser 430 (Fig. 4A), and this phosphorylation was significantly reduced by pretreating the cells with PKC inhibitor Gö 6976 (Fig. 4B). Consistent with a role of PKC␣ in ER stress-induced phosphorylation of APPL1 at Ser 430 , overexpression of HAtagged wild-type but not the dominant negative PKC␣ enhanced TG-stimulated phosphorylation of APPL1 at Ser 430 (Fig. 4C), suggesting that PKC␣ positively regulates ER stressinduced phosphorylation of APPL1 at Ser 430 . Consistent with this, TG-stimulated phosphorylation of APPL1 at Ser 430 was reduced in PKC␣-suppressed hepatocytes, further indicating an important role of PKC␣ in mediating ER stress-induced insulin resistance in hepatocytes (Fig. 4D).
Phosphorylation of APPL1 at Ser 430 Mediates ER Stress-induced Insulin Resistance-To characterize the functional role of APPL1 phosphorylation at Ser 430 , we examined adiponectin signaling in hepatocytes expressing wild type and S430A mutant of APPL1. Treating the cells with adiponectin led to a great increase in AMPK phosphorylation at Thr 172 , but had no significant effect on the phosphorylation of APPL1 at Ser 430 in hepatocytes (data not shown). To examine whether the phosphorylation of APPL1 modulates the insulin-sensitizing effect of APPL1, we examined insulin signaling in hepatocytes expressing wild type or S430A or S430D mutant of APPL1. Overexpression of wild-type APPL1 enhanced insulin-stimulated phosphorylation of Akt at Thr 308 (Fig. 5A), and the phosphorylation was further enhanced in cells expressing S430A but not S430D mutant of APPL1 (Fig. 5A). Taken together, these results show that phosphorylation of APPL1 at Ser 430 plays a negative role in regulating APPL1 action in insulin signaling. In agreement with this, the ER stress-induced APPL1 phosphorylation at Ser 430 and inhibition of insulin-stimulated Akt phosphorylation at Thr 308 was greatly reduced in PKC␣-suppressed cells (Fig. 5, B and C), further demonstrating a negative regulatory role of APPL1 phosphorylation at Ser 430 in insulin signaling.

DISCUSSION
APPL1, the first signaling molecule identified downstream of adiponectin receptors, has been shown to play an important role in adiponectin signaling and action (2-5, 17, 19). In addition to binding to adiponectin receptors, APPL1 has also been shown to interact with PI3K and Akt (6) and facilitate Akt activation and substrate specificity (7). APPL1 potentiates insulinstimulated Akt activation and subsequent suppression of hepatic glucose production through counteracting the inhibition of Akt activation by TRB3 in the liver (8). However, how APPL1 is regulated in cells remains largely unknown. A recent study predicted that APPL1 could be phosphorylated at multiple residues, suggesting that post-modification could be a potential mechanism regulating APPL1 action in cells (10). However, the kinases that mediate APPL1 phosphorylation and the functional roles of the phosphorylation are unknown. In the present study, we demonstrate that APPL1 undergoes phosphorylation at Ser 430 under obesity and ER stress conditions. In addition, we demonstrate that phosphorylation at this site impairs the insulin-sensitizing effect of APPL1. Furthermore, we have identified PKC␣ as the kinase mediating ER stressinduced phosphorylation of APPL1 at Ser 430 in hepatocytes. However, how APPL1 phosphorylation at Ser 430 negatively regulates the insulin-sensitizing effect of APPL1 remains elusive. Because APPL1 stimulates AMPK and p38 MAPK phosphorylation by interacting with AdipoR1, LKB1, PP2A, PKC, and Rab5 in myotubes (1)(2)(3)(4), it is possible that phosphorylation of APPL1 at Ser 430 down-regulates adiponectin signaling by inhibiting APPL1 from binding to AdipoR1, LKB1, PP2A, PKC, and/or Rab5. It is also possible that phosphorylation of APPL1 at Ser 430 affects the association between APPL1 and PI3K or Akt (6,7). Future studies are needed to investigate the mechanisms.
ER stress has been shown to contribute to obesity-induced insulin resistance (15,20,21). ER stress impairs insulin signaling through hyperactivation of JNK and subsequent stimula- . ER stress-stimulated phosphorylation of APPL1 at Ser 430 is in a PKC␣-dependent manner. A, ER stress induced the phosphorylation of APPL1 (p-APPL1) at Ser 430 in hepatocytes. Starved hepatocytes were treated with or without 1 M thapsigargin (TG), an ER stress inducer, for 1 h or 1 M PMA for 30 min. P-eIF2␣, phosphorylation of eIF2␣. B, ER stress-stimulated phosphorylation of APPL1 at Ser 430 was reduced by inhibition of PKC in hepatocytes. Starved hepatocytes were pretreated with or without Gö 6976, a PKC-specific inhibitor, for 1 h and then treated with or without 1 M TG for 1 h. C, the enzymatic activity of PKC␣ is required for ER stress-induced phosphorylation of APPL1 at Ser 430 in hepatocytes. Hepatocytes transiently expressing HA-tagged wild-type or dominant negative (DN) PKC␣ were starved and then treated with or without 1 M TG for 1 h. D, ER stress-induced phosphorylation of APPL1 at Ser 430 was reduced in PKC␣-suppressed hepatocytes. Starved control or PKC␣-suppressed hepatocytes were treated with or without 1 M TG for 1 h and then treated with or without 10 nM insulin for 3 min. The phosphorylation of APPL1 at Ser 430 , phosphorylation of eIF2␣ at Ser 51 , and protein levels of APPL1, eIF2␣, PKC␣, and tubulin were determined by Western blot using specific antibodies as indicated. FIGURE 5. The phosphorylation of APPL1 at Ser 430 plays a negative role in regulating insulin signaling. A, the phosphorylation of APPL1 at Ser 430 negatively regulates insulin-stimulated Akt phosphorylation (P-Akt). The hepatocytes transiently expressing Myc-tagged wild type or S430A or S430D mutant were starved for 4 h and then treated with or without 10 nM insulin (Ins) for 3 min. The phosphorylation at Thr 308 and protein levels of Akt, APPL1, and tubulin were determined by using the specific antibodies as indicated. B, ER stress inhibited insulin-stimulated Akt phosphorylation at Thr 308 via PKC␣dependent mechanism. Starved PKC␣ scramble and shRNA-suppressed hepatocytes were treated with or without 1 M TG for 1 h followed with insulin treatment (10 nM for 3 min). P-APPL1, phosphorylation of APPL1. C, the ratio of insulin-stimulated phosphorylation of Akt at Thr 308 to Akt protein shown in B was quantified. The data represent mean Ϯ S.E. *, p Ͻ 0.05. tion of phosphorylation of the insulin receptor substrate-1 (IRS-1) at Ser 307 (15). Serine phosphorylation of IRS-1 suppresses its tyrosine phosphorylation and thus the functional roles of this adaptor protein, which contributes to ER stressinduced insulin resistance (15). Our previous study revealed that ER stress induces autophagy-dependent insulin receptor degradation (16). Our current study indicates that APPL1 phosphorylation at Ser 430 is also involved in ER stress-induced insulin resistance.
Activation of PKC isoforms has been shown to be associated with obesity and insulin resistance (22,23). PKC␣, one classic PKC isoform, was found to mediate the suppressing effect of TNF␣ on insulin action and signaling by promoting IRS-1 serine/threonine phosphorylation (24) and inhibiting IR tyrosine phosphorylation (17,25). PKC␣ has also been shown to mediate the inhibitory effect of advanced glycation end products on insulin action in the muscle through the formation of a multimolecular complex including RAGE-IRS-1-Src (18). In our present study, we found that induction of ER stress stimulated the phosphorylation of APPL1 at Ser 430 via a PKC␣-dependent mechanism. Overexpression of wild-type APPL1 and the S430A mutant, but not the S430D mutant, increased insulinstimulated Akt phosphorylation. These data further suggest that the phosphorylation of APPL1 at Ser 430 reduces the insulin-sensitizing effect of APPL1, which is a novel mechanism underlying ER stress-stimulated insulin resistance. In summary, our study has demonstrated for the first time that APPL1 undergoes ER stress-induced and PKC␣-mediated phosphorylation at Ser 430 and that the phosphorylation negatively regulates APPL1 function in hepatocytes. This result suggests that APPL1 phosphorylation at Ser 430 could be an important mechanism underlying obesity-and ER stress-induced insulin resistance.