Targeting of PED/PEA-15 Molecular Interaction with Phospholipase D1 Enhances Insulin Sensitivity in Skeletal Muscle Cells*

Phosphoprotein enriched in diabetes/phosphoprotein enriched in astrocytes (PED/PEA-15) is overexpressed in several tissues of individuals affected by type 2 diabetes. In intact cells and in transgenic animal models, PED/PEA-15 overexpression impairs insulin regulation of glucose transport, and this is mediated by its interaction with the C-terminal D4 domain of phospholipase D1 (PLD1) and the consequent increase of protein kinase C-α activity. Here we show that interfering with the interaction of PED/PEA-15 with PLD1 in L6 skeletal muscle cells overexpressing PED/PEA-15 (L6PED/PEA-15) restores insulin sensitivity. Surface plasmon resonance and ELISA-like assays show that PED/PEA-15 binds in vitro the D4 domain with high affinity (KD = 0.37 ± 0.13 μm), and a PED/PEA-15 peptide, spanning residues 1-24, PED-(1-24), is able to compete with the PED/PEA-15-D4 recognition. When loaded into L6PED/PEA-15 cells and in myocytes derived from PED/PEA-15-overexpressing transgenic mice, PED-(1-24) abrogates the PED/PEA-15-PLD1 interaction and reduces protein kinase C-α activity to levels similar to controls. Importantly, the peptide restores insulin-stimulated glucose uptake by ∼70%. Similar results are obtained by expression of D4 in L6PED/PEA-15. All these findings suggest that disruption of the PED/PEA-15-PLD1 molecular interaction enhances insulin sensitivity in skeletal muscle cells and indicate that PED/PEA-15 as an important target for type 2 diabetes.


Phosphoprotein enriched in diabetes/phosphoprotein enriched in astrocytes (PED/PEA-15) is overexpressed in several tissues of individuals affected by type 2 diabetes. In intact cells and in transgenic animal models, PED/PEA-15 overexpression impairs insulin regulation of glucose transport, and this is mediated by its interaction with the C-terminal
PED/PEA-15 (1) is a ubiquitously expressed protein that controls cell proliferation and death (2)(3)(4)(5)(6). It has been found that PED/PEA-15 is overexpressed 2-to 3-fold in skeletal muscle, adipose tissue, fibroblasts, and white blood cells from a large population of type 2 diabetic individuals and their first degree relatives (7,8). In cellular and animal models, PED/PEA-15 overexpression affects both insulin-stimulated glucose transport and glucose-stimulated insulin secretion (7, 9 -11). In particular, forced expression of PED/PEA-15 in muscle and adipose cells to levels comparable to those occurring in type 2 diabetes severely impairs insulin-stimulated glucose transport (11) and cell-surface recruitment of GLUT4, a major insulinsensitive glucose transporter (7). Furthermore, transgenic mice for PED/PEA-15 display impaired glucose tolerance and develop diabetes, if fed a high fat diet (9). All these observations suggest that PED/PEA-15 overexpression could be involved in the complex series of events ultimately leading to type 2 diabetes, one of the most common disorder in the world associated with impaired insulin action and secretion and for which no single defect has been so far unequivocally determined (12)(13)(14).
PED/PEA-15 is a 130-amino acid protein containing an ␣-helical-rich DED domain at its N terminus, whereas the C-terminal 40 residues appear largely unstructured (15). Two consensus serine phosphorylation sites have been identified at the C terminus of the protein (Ser 104 and Ser 116 ), and phosphorylation by protein kinase C (PKC), 3 calmodulin kinase II, and AKT/protein kinase B has been shown to occur in different cells types and to contribute to the regulation of PED/PEA-15 protein stability (16 -18). However, mutation experiments show that PED/PEA-15 phosphorylation, which is responsible for PED/PEA-15 antiapoptotic function (17,18), is not directly involved in changes to its gluco-regulatory functions (7,11), suggesting the implication of a different mechanism. PED/ PEA-15 has been found to be an interactor of the human phospholipase D1 (PLD1), an interaction which promotes PLD1 activity. Although the mechanism by which this occurs is still unknown, it has been observed that increasing PED/PEA-15 cellular abundance lengthens PLD1 persistence in the cell rather than increasing its enzymatic activity (9,19).
PLD1 is widely distributed in animals, plants, fungi, and bacteria and is implicated in several cellular processes, including receptor signaling, control of vesicular trafficking, and glucose transport (20). Furthermore, phospholipase D catalyzes the hydrolysis of the phosphodiester bond of glycerophospholipids to generate phosphatidic acid, an intracellular messenger implicated in a wide range of cellular processes. Phosphatidic acid can also be converted to other mediators, such as lysophosphatidic acid and diacylglycerol (21), with the latter being a major activator of the conventional PKC isoforms. Indeed, in PED/PEA-15 overexpressors, both diacylglycerol concentration (9) and PKC-␣ activity (11) are constitutively increased. It has also been shown that overactivation of PKC-␣ negatively regulates the activity of PKC-(11), a major controller of insulin-stimulated glucose transport. Therefore, disrupting the interaction between PED/PEA-15 and PLD1 by a cell-penetrating compound may represent a novel strategy for improving insulin sensitivity in target cells.
As a proof of principle, using L6 skeletal muscle cells stably overexpressing PED/PEA-15 (L6 PED/PEA-15 ), we show here that, by disrupting PED/PEA-15-PLD1 binding with protein fragments involved in the protein-protein interaction, the PED/PEA-15 downstream signaling and the negative effects on glucose uptake are largely rescued. Indeed, D4 expression in L6 PED/PEA-15 reduces the interaction between PLD1 and PED/PEA-15, lowers PKC-␣ activation, and restores insulin effect on glucose uptake. Consistently, the PED/PEA-15 region entailed in PLD1 recognition and identified following an approach of protein fragmentation and fractionation, when incorporated into L6 PED/PEA-15 cells, is capable of displacing the PED/PEA-15-PLD1 interaction and increasing insulin-stimulated glucose transport.
A deletion mutant of D4 (named dmD4), spanning residues 929 -1030 and with the two cysteines replaced by serines, was used as negative control in all experiments. Site-directed mutagenesis was performed using the QuikChange site-directed mutagenesis kit from Stratagene, according to the producer's protocol. Mutant oligonucleotides complementary to the single-stranded DNA and encoding the specific mutation were as follows: 5Ј-GGACTTCG-GCTACAGTCCTTTAGGGTTGTCC-3Ј and 5Ј-GGACAACC-CTAAAGGACTGTAGCCGAAGTCC-3Ј, and 5Ј-GACAAGG-TTTTCCGGTCCCTTCCCAATGATGAAG-3Ј and 5Ј-CTTCATCATTGGGAAGGGACCGGAAAACCTTGTC-3Ј. Mutations were confirmed by nucleotide sequencing. dmD4 was then subcloned into the pcDNA3-HA expression vector.
Cell Culture, Transfections, Western Blot Analysis, Co-immunoprecipitation Assay, and 2-Deoxy-D-glucose Uptake-L6 rat skeletal muscle cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 10,000 units/ml penicillin, 10 mg/ml streptomycin, and 2% L-glutamine (Invitrogen), in a 5% humidified CO 2 incubator. Transient transfections were performed using the Lipofectamine Plus according to the manufacturer's instructions (Invitrogen). Stable transfections were performed by selecting positive clones with G418 (Calbiochem) at the effective dose of 0.8 mg/ml. Cellular loading of FITC-conjugated peptides was performed with cationic lipid mixture Pro-Ject TM Protein Transfection Reagent kit according to the manufacturer's instructions (Pierce). Production of rabbit polyclonal PED/ PEA-15 antiserum, Western blot and co-immunoprecipitation analyses, and 2-deoxy-D-glucose uptake measurement were performed as previously reported (7,23,24,11).
Mouse Primary Fibroskeletal Muscle Cell Culture-Skeletal muscle biopsies were obtained after Wt and Tg PED/PEA-15 mice were sacrificed by pentobarbitone overdose, as previously described (25). The biopsies were collected in cold phosphatebuffered saline (PBS) supplemented with 1% PeSt (100 units/ml penicillin and 100 g/ml streptomycin), dissected, finely minced, and transferred to a digestion solution (0.015 g of Collagenase IV, 8% 10ϫ trypsin, 0.015 g of bovine serum albumin, 1% PeSt, in DMEM supplemented with 10% fetal calf serum, 10,000 units/ml penicillin, 10 mg/ml streptomycin, and 2% L-glutamine), then and incubated with gentle agitation at 37°C for 15-20 min. Thereafter, undigested tissue was allowed to settle, and the supernatant was collected and mixed with DMEM supplemented with 20% fetal calf serum and 1% PeSt. The remaining tissue was digested for a further 15-20 min at 37°C with fresh digestion solution. The resultant supernatant was then pooled with the previous cells and centrifuged for 10 at 350 ϫ g. The cell pellet was resuspended in DMEM supplemented with 20% fetal calf serum and 1% PeSt and was then seeded and grown in culture flask. After this, medium was again changed to DMEM supplemented with 10% fetal calf serum and 1% PeSt. Before any experiment, cells were serumstarved for 16 h and stimulated with insulin at specific times and concentrations as indicated.
PKC-␣ Activity Assay-PKC-␣ activation was measured by evaluating PKC-␣ Ser-657 phosphorylation. Western blot assay was performed with a specific pPKC-␣ antibody purchased from Upstate Biotechnology Inc. (Lake Placid, NY), and the intensity of the spots was evaluated by densitometric analysis, using Scion Image Analyzer software. All data were expressed as mean Ϯ S.D.
Statistical Analysis-Data were analyzed with StatView software (Abacus Concepts) by one-factor analysis of variance. p values of Ͻ0.05 were considered statistically significant.
PLD Assay in Intact Cells-PLD activity was evaluated by measuring phosphatidic acid and phosphatidylbutanol levels (26). Cells were labeled with 5 Ci of [ 14 C]palmitic acid for a 60-mm dish and incubated for 16 h at 37°C in a 5% CO 2 -enriched, humidified atmosphere. Cells were then washed twice in HHBG buffer (10 mM HEPES, 1.26 mM CaCl 2 , 0.5 mM MgCl 2 , 0.4 mM MgSO 4 , 5.37 mM KCl, 137 mM NaCl, 4.2 mM NaH 2 PO 4 , 1% (w/v) bovine serum albumin, 10 mM glucose, pH 7.4) and incubated in 0.3% (v/v) butan-1-ol in HHBG buffer for 20 min at 37°C. Subsequently, cells were treated with or without 100 nM phorbol myristate acetate for 30 min. After incubation, buffer was removed and 0.5 ml of ice-cold methanol was added to each dish. Cell debris was scraped into a glass vial and kept on ice. Cellular lipids were obtained using the Bligh and Dyer procedure and spotted onto Whatman TLC plates. Labeled products were separated by TLC using the upper phase of a mixture of ethyl acetate/2,2,4-trimethylpentane/acetic acid/H 2 O in the ratio of 13:2:3:10 (v/v). Positions of the spots corresponding to [ 14 C]phosphatidylbutanol were determined by autoradiography. The area containing phosphatidylbutanol was scraped, and radioactivity was counted.
Trx-His 6 Expression and Purification-100 ng of pETM-20 vector were chemically transformed in E. coli Bl21(DE3) (Novagen), and expression of Trx-His 6 , fused to a stuffer protein, was induced with 1 mM isopropyl 1-thio-␤-D-galactopyranoside for 3 h at 37°C. The entire fusion protein was purified on 1 ml of Chelating Sepharose Fast Flow resin (GE Healthcare), and Trx-His 6 was then separated from the stuffer protein after digestion with TEV protease. A further passage on Chelating Sepharose Fast Flow resin allowed us to obtain Trx-His 6 as the only protein unbound to the resin.
PED/PEA-15 Tryptic Digestion-PED/PEA-15 proteolysis was performed by digestion with L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Sigma-Aldrich) at an enzyme:substrate ratio of 1:100 (w/w) in 50 mM Tris-HCl, pH 7.5, for 16 h. Tryptic digests were separated into seven fractions on a Phenomenex Jupiter C18 column (1 ϫ 25 cm, 10 m) and analyzed by using a 50-ϫ 2-mm inner diameter C18 BioBasic column (ThermoFisher). LC-MS was performed using a LCQ DCA XP Ion Trap spectrometer (ThermoElectron). This was equipped with an Opton electrospray ionization source (operating at a needle voltage of 4.2 kV and at a temperature of 320°C) and a complete Surveyor HPLC system (including an MS pump, an autosampler, and a photo diode array). Mass spectra were recorded continuously between the mass interval 400 -2000 atomic mass units in positive mode and data-dependent analysis to fragment the eluted peptides and to obtain sequence information. Fragmentation was induced on selected ions from 400 to 1600 atomic mass units, with a fixed 35% of total energy. Multicharge spectra were then deconvoluted using the BioMass program implemented in the Bioworks 3.1 package provided by the manufacturer. Mass calibration was performed automatically by means of selected multiple charged ions, in the presence of a calibrant (Ultramark, ThermoElectron). All masses were reported as average values.
SPR Analysis-PED/PEA-15 was immobilized on a CM5 sensor chip using standard amine coupling procedures, as described by the manufacturer's instructions (Pharmacia Biosensor AB), to obtain a final 11,000 RU immobilization level. All assays were carried out at 25°C, at a flow rate of 30 l/min in 50 mM sodium phosphate, 150 mM NaCl, 1 mM TCEP, pH 7.4, buffer. Data were processed using BIAevaluation software, version 4.1 (BIAcore Technologies, Inc.).
For competition experiments with fractions derived from trypsintreated PED/PEA-15, 20 l of each fraction was preincubated with 1 M Trx-His 6 -D4 in a final volume of 100 l (assuming that 0.13 mol of each peptide was recovered after fractionation, the theoretical peptide concentration for these experiments was 260 M) at 25°C for 30 min and then injected. In the co-injection method, 15 l of each pool (diluted 1:1 in running buffer) was injected after injection of 60 l of 1 M Trx-His 6 -D4. For competition experiments with the synthetic PED/ PEA-15 peptides, 1 M Trx-His 6 -D4 was preincubated with 30 M of each peptide at 25°C for 30 min before injection.
Enzyme-linked Immunosorbent Assays-All assays were performed using an integrated platform for High-Throughput Screening (Hamilton Robotics) comprising a fully equipped Starlet 8 channel liquid handler, a robotic arm, a washer, and a multiwavelength plate reader. PED/PEA-15 was biotin-labeled using EZ-Link NHS-Biotin reagents, following the manufacturer's protocol (Pierce); LC-MS confirmed biotinylation. 100 l of a 0.5 M solution of Trx-His 6 -D4 was coated onto the wells for 16 h at 4°C; wells filled with 100 l of 0.5 M Trx-His 6 were used as blank control. Each data point was in triplicate. After washing with PBS-NaCl buffer (300 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 , 0.004% Tween 20, pH 7.4), wells were blocked with a 1% w/v solution of bovine serum albumin in PBS. Biotinylated-PED/PEA-15 at a fixed concentration of 1.5 M was preincubated at 25°C with 7.5 M peptide competitors for 30 min in 50 mM sodium phosphate, NaCl 300 mM, TCEP 1 mM, 0.004% Tween 20, pH 7.5, buffer and added to the washed wells. Following 1-h incubation in the dark at 37°C, solutions were removed, and wells were again washed with PBS-NaCl buffer. 100 l of a 1 mg/ml solution of streptavidin-horseradish peroxidase (Sigma-Aldrich) was added to each well, and the plate was left to incubate for 1 h at 37°C in the dark. After removal of the enzyme solution, Sigma-fast o-phenylenediamine dihydrochloride Tablet Set (Sigma-Aldrich) was used, and absorbance at 490 nm was monitored. Values were properly averaged, subtracting the corresponding blank lanes.

Effects of PED/PEA-15 Overexpression on Phospholipase D1 in L6 Cells-Preparation of L6 skeletal muscle cells (L6)
stably transfected with PED/PEA-15 cDNA (L6 PED/PEA-15 ) was previously described (7). In these cells, overexpression of PED/PEA-15 was paralleled by a consistent increase of PLD1 cellular content (Fig. 1A). Similarly, detection of PLD1 in PED/ PEA-15 immunoprecipitates was increased more than 3-fold in L6 PED/PEA-15 , as compared with untransfected L6 controls (Fig.  1B). To further address if the larger amount of PLD1 was accompanied by an increase in its enzymatic activity, we measured the levels of transphosphatidylbutanol generated by the PLD1-mediated reaction. As shown by TLC analysis (Fig. 1C), both basal and phorbol myristate acetate-stimulated PLD1 activity was increased in L6 PED/PEA-15 compared with L6 cells (Fig. 1D).
Effects of D4 Expression on PED/PEA-15 Interaction with PLD1 and on Glucose Uptake-Following reverse transcription-PCR amplification of D4 from HeLa cells, the cDNA was inserted into the pcDNA3 vector in-frame with the HA epitope, and the entire construct (pcDNA3-HA-D4) was transfected in L6 and in L6 PED/PEA-15 cells. The HA-tagged D4 domain was expressed at comparable levels both in L6 and L6 PED/PEA-15 cells ( Fig. 2A). Interestingly, HA-D4 expression significantly reduced the amount of PLD1 co-precipitated with PED/PEA-15, as compared with the transfection of the empty pcDNA3-HA vector (Fig. 2B). In addition, HA-D4, but not the control vector, reduced the activity of PKC-␣ in L6 PED/PEA-15 by ϳ60% (Fig. 2C). The expression of HA-D4 in L6 PED/PEA-15 cells was also accompanied by a recovery of the effect of insulin on glucose uptake. Indeed, insulin treatment failed to increase glucose uptake in L6 PED/PEA-15 cells, which instead was largely rescued in cells expressing the HA-D4 construct, at levels comparable to those observed in L6 control cells (Fig. 2D). Transfection of a deletion mutant of D4 (dmD4, see also  Results were reported as (RU/RU 0 )*100, where RU is the maximum response (RU max ) for a given competitor and RU 0 is the RU max without competitors.

TABLE 1 PED/PEA-15 peptides identified by LC-MS/MS analysis of HPLC fractions derived from tryptic digestion of PED/PEA-15
The peptide in fraction 6 contains the extra dipeptide Gly-Ala at its N terminus derived from cloning. The relative amount within the fractions was calculated by comparing peak areas from the corresponding extracted ions. Targeting of the PED/PEA-15-PLD1 Interaction AUGUST 1, 2008 • VOLUME 283 • NUMBER 31 "Experimental Procedures") in L6 PED/PEA-15 cells ( Fig. 2A) did not affect either the interaction between PED/PEA-15 and PLD1 (Fig. 2B) or the constitutive activation of PKC-␣ (Fig. 2C) and the uptake of 2-deoxyglucose (Fig. 2D). It should be also noted that L6 PED/PEA-15 cells displayed a higher basal glucose uptake (7,11), which is decreased by the expression of HA-D4 (Fig. 2D). These findings led us to hypothesize that the D4 fragment could bind PED/PEA-15 preventing its interaction with the full-length PLD1 and restoring insulin action.
PED/PEA-15-D4 in Vitro Binding Assay-The in vitro interaction between PED/PEA-15 and D4 was analyzed by SPR using a BIAcore 3000 system. To set up this assay, both PED/PEA-15 and D4 were recombinantly expressed in an E. coli expression system. PED/PEA-15 was preliminarily expressed using the pETM30 vector; following purification and digestion with TEV protease to remove the fusion partners, ϳ20 mg of PED/ PEA-15 per liter of initial culture was purified. Instead, the D4 domain of PLD1 was expressed and fused at its N terminus to a thioredoxin A and a histidine tag for affinity purification (Trx-His 6 ); yield was low (ϳ1 mg of purified protein per liter of culture), but it was sufficient for our experiments. Notably, any attempt to cleave D4 from its fusion partner failed or led to a highly unstable product, so the whole protein (named Trx-His 6 -D4) was used in all subsequent experiments. Due to the high number of cysteines, purifications and subsequent manipulations were all performed in 1 mM TCEP-containing buffers. The protein was purified to homogeneity by two subsequent steps of affinity purification and ionic exchange chromatography.
PED/PEA-15 was thus immobilized on the surface of a CM5 sensor and, to exclude any interaction between PED/ PEA-15 and the fusion partner of D4, solutions of Trx-His 6 (expressed and purified independently) at concentrations up to 3.6 M, were firstly injected on the PED/PEA-15-derivatized microchip. Under these conditions, no interaction was detected between PED/PEA-15 and Trx-His 6 . Then increasing amounts of Trx-His 6 -D4 were analyzed using a TCEP-containing buffer. Trx-His 6 -D4 exhibited dose-dependent association curves (Fig. 3), characterized by slow association and dissociation rates (dissociation constant of 0.37 Ϯ 0.13 M). As a further control, the deletion mutant dmD4, also fused to Trx-His 6 , was not able to associate to immobilized PED/PEA-15. Thus, in vitro the D4 domain of PLD1 selectively interacts with PED/ PEA-15. Identification of D4-binding PED/PEA-15 Region-To delineate PED/PEA-15 sub-domains involved in D4 binding, we extensively hydrolyzed PED/PEA-15 with trypsin to obtain short peptide fragments of known size and sequence. To this aim, a protein aliquot (0.13 mol, 2.0 mg) was treated with the enzyme at a 1:100 ratio for 16 h at 37°C. A small aliquot (500 ng) was analyzed by LC-MS/MS to assess the completion of the reaction and to identify fragments, before adding trifluoroacetic acid to stop the hydrolysis reac-

TABLE 2 List of synthesized peptides
Sequences reported in bold correspond to peptides selected after competition with tryptic derived fractions. The corresponding secondary structure as it appears within the full-length protein was indicated in the second column.

Targeting of the PED/PEA-15-PLD1 Interaction
tion. Tryptic fragments were then fractionated by RP-HPLC recovering six separate aliquots that were subsequently lyophilized, dissolved in 100 l of H 2 O, and characterized by LC-MS/MS (Table 1); hence PED/PEA-15 peptide fragments will be indicated as PED (fragment numeration). PED tryptic fragments PED-(25-28), -(84 -88), -(99 -107), and -(114 -122) were not recovered within the HPLC fractions, whereas fractions 0, 1, and 4 contained no peptides. All fractions were subsequently used in competition experiments of PED/PEA-15-D4 binding on the BIAcore 3000 system. At first, they were tested for their capacity to bind the immobilized protein to exclude possible false negative results originating from signal compensation effects. No evident interactions were recorded with all tested samples (data not shown). Competition was carried out using two different approaches. Initially, peptide competitors were preincubated with Trx-His 6 -D4 and injected on the chip (incubation method). As shown in Fig. 4, Fraction 6, containing the fragment Gly-Ala-PED-(1-24) (with the dipeptide Gly-Ala on the N terminus derived from the vector linker), reduced binding by 44%, suggesting an active role of this peptide in preventing protein-protein contact. Fraction 5, containing peptides spanning the region 36 -83, reduced the binding of Trx-His 6 -D4 to immobilized PED/PEA-15 by 30% . Fractions 2 and 3, containing peptides from the C-terminal region and the peptide 29 -35, had no significant effects on protein-protein interaction. In a second approach, the six fractions were instead injected at the end of the association phase with D4, and their relative capacity to displace the bound ligand were evaluated (co-injection method). Again, peptide Gly-Ala-PED-(1-24) confirmed its capacity to interfere with the PED/PEA-15-D4 recognition (Fig. 4), although this method appeared generally less sensitive than the previous one.
To further investigate the ability of these fragments to compete with PED/PEA-15 binding to D4, we prepared synthetic peptides corresponding to some of the tryptic fractions. These peptides are listed in Table 2 along with the secondary structure they correspond to (15). In particular, we synthesized peptide PED-(1-24), found in Pool 6 (without the N-terminal dipeptide Gly-Ala), and peptides PED-(72-83) and PED-(36 -54), corresponding to the major components of Pool 5. Furthermore, we synthesized a panel of new peptides corresponding to most PED/PEA-15 helices that were only partially represented by the tryptic fragments (PED-(1-15), PED-(16 -30), PED-(71-92), PED-(33-56), and PED-(40 -56)). We also prepared peptides PED-(53-77) and PED-(79 -112), which cover the PED/PEA-15 region 53-112 previously hypothesized (21) as the shortest sequence interacting with D4, and PED-(114 -122), which is the largest fragment not recovered after PED/PEA-15 digestion. Again, to exclude possible false negative results, these peptides were preliminarily assayed for their capacity to associate with the on-chip immobilized protein. For this purpose, 30 M solutions of each peptide were injected on the chip and the binding signal recorded. Of note, peptide PED-(71-92) was able to interact with the full-length PED/PEA-15 protein; therefore, this peptide was not considered in subsequent experiments. The remaining peptides were instead utilized in the SPR-based competition assay using the incubation method. Data obtained by this approach confirmed the ability of peptide PED-(1-24) to interfere with PED/PEA-15-D4 recognition.
To confirm these data, we set up a competition experiment based on an ELISA-like assay whereby Trx-His 6 -D4 was immobilized on the microwell surface, and biotinylated PED/PEA-15 (biotin-PED) was instead used as the soluble binder. Initially dose-dependent binding of PED/PEA-15 to immobilized Trx-His 6 -D4 (0.5 M) was assessed, observing signal saturation for concentrations of biotin-PED higher than 2 M (not shown). Peptides at a 5-fold molar excess over biotin-PED were then used to disrupt the binding. As shown in Fig. 5A, synthetic PED-(1-24) was again the most efficacious competitor, whereas PED-(1-15) and PED-(16 -30) were only slightly active and were not further investigated. To further assess the capacity of PED-(1-24) to interfere with the interaction of the two proteins, this peptide was utilized in a dose-dependent competition assay at concentrations ranging between 0.5 and 7.5 M. Importantly, the synthetic peptide totally blocked PED/PEA-15 binding to immobilized D4 at the highest concentration (ϳ7 M) and at ϳ3 M, 50% reduction was detected (Fig. 5B).
Effect of PED-  on Primary Myocytes from PED/PEA-15 Transgenic Mice-The effects of the PED-(1-24) peptide were finally tested in primary myocytes derived from quadriceps muscles of transgenic (Tg) mice overexpressing PED/PEA-15 (9) and their wild-type littermates. Interestingly, peptide delivery in Tg myocytes led to a drastic reduction of PLD1 interaction with PED/PEA-15 (Fig. 8A), paralleled by a similar reduction of PKC-␣ activity. However, no change in PED/PEA-15-PLD1 co-precipitation was observed in wild-type (Wt) myocytes (Fig. 8A). Consistent with the observations made in L6 PED/PEA-15 , treatment of Tg myocytes with PED-(1-24) almost completely restored insulin-stimulated PKCactivation. Again, no effects were detected in Wt myocytes.

DISCUSSION
Diabetes is a widespread disease with more than 150 million individuals affected worldwide. Of these, most are affected by the type 2 form (source: American Diabetes association, see www.diabetes.org), which is characterized by resistance to insulin action on glucose metabolism. No single defects have been identified for this disease, rather a multitude of concurrent alterations contribute to disease onset and progression  (12)(13)(14). Among these defects, the PED/PEA-15 gene has been found to be overexpressed in tissues from type 2 diabetic individuals (7). This is paralleled by increased cellular abundance of the PED/PEA-15 gene product, a 15-kDa cytoplasmic protein.
Studies in cellular and animal models have also shown a causeeffect relationship between the overexpression of PED/PEA-15 and impairment of insulin action (7)(8)(9)(10)(11). Therefore, elevated expression of PED/PEA-15 may represent a risk for the progression toward type 2 diabetes. To further support this hypothesis, it has recently been reported that ϳ30% of individuals affected by type 2 diabetes and their first degree relatives have elevated PED/PEA-15 mRNA and protein levels (7,8). The molecular bases of PED/PEA-15 overexpression have not yet been fully elucidated; however, its interaction with PLD1 may represent a molecular link between high PED/PEA-15 levels and the impairment of insulin action. Thus, the elucidation of the PED/PEA-15 pathological signaling pathway is of utmost importance for the development of novel drugs to restore insulin sensitivity.
We have now shown that PED/PEA-15 and the D4 domain of PLD1 bind in vitro with high affinity. The N-terminal region of PED/PEA-15, encompassing residues 1-24 (PED-(1-24)), is involved in D4 (and thus PLD1) recognition. A previous report indicated the region 53-112 as forming the minimal PLD1binding site (19). However, although we cannot exclude a direct involvement of the entire region in D4 binding, peptides 53-77, 71-83, and 79 -112 were ineffective in our peptide-based assays. Of note, the region 1-24 encompasses residues matching the first helix, the second loop, and part of the second helix of PED/PEA-15 (the latter terminating at residue 30) (15).
Either D4 transfection or PED-(1-24) loading into intact L6 PED/PEA-15 cells and in myocytes from PED/PEA-15 Tg mice, disrupt PED/PEA-15-PLD1 interaction. Noticeably, PED/ PEA-15 Tg mice are insulin-resistant and highly susceptible to diabetes (9). In addition, as previously reported, overexpression of PED/PEA-15 in skeletal muscle cells, impairs insulin action on glucose uptake without affecting the early steps of insulin signaling (7,11). We now provide the first experimental evidence that blocking the protein-protein interaction between PED/PEA-15 and PLD1 is sufficient to impair the molecular mechanisms triggered and maintained by PED/PEA-15 overexpression. Indeed, ectopic expression of the PLD1 D4 domain abolishes binding of the full-length parental protein to PED/ PEA-15 and restores basal PKC-␣ activity and normal insulinstimulated glucose uptake. Consistently, loading of the L6 PED/PEA-15 cells with the PED/PEA-15 N-terminal peptide responsible of the D4 (PLD1) recognition also rescues insulin action. It must be underscored that, although targeting a protein-protein interaction might generally be largely more difficult than finding kinase or phospholipase inhibitors, directly blocking these other players can provide several deleterious outcomes. Indeed, PKC-␣ plays important roles in many different cellular processes, including cell proliferation, cell cycle checkpoint, cell adhesion, and cell volume control (28). For example, impairment of PKC-␣ activity can affect cardiac function, because it has been identified as a major regulator of cardiac contractility and calcium handling in myocytes (29). Likewise, PLD1 has pleiotropic roles in many aspects of cell regulation, including proliferation, survival, and vesicular transport exocytosis (20). Thus, by disrupting the interaction with PED/PEA-15, one might expect to reduce only the pathologically high cellular PLD1 and PKC-␣ activities, without affecting regulation by other factors. To support this hypothesis, no significant change of PLD1 and PKC-␣ activity has been detected following D4 transfection or peptide loading of wildtype L6 cells and primary myocytes. These observations indicate that disruption of the PED/PEA-15-PLD1 interaction selectively ameliorates insulin signaling in cells bearing high levels of PED/PEA-15. In addition, forced D4 expression induced no variation of growth profiles in both L6 and L6 PED/PEA-15 cells (data not shown), suggesting that no gross impairment of the cell cycle was induced by blocking of PED/ PEA-15-PLD1 interaction. Because PLD1 and PLD2 isoforms display a very high sequence homology in the D4 region (ϳ90%), one would expect that PLD2 may contribute as well to the effects mediated by the D4 or by the antagonist peptide.
In conclusion, we show that the PED/PEA-15-PLD1 interaction may represent a privileged molecular target for insulin resistance, particularly in those individuals with high PED/ PEA-15 levels. Indeed, selective disruption of this interaction with cell-penetrating agents does not affect constitutive PKC-␣ and PLD1 functions, which may turn out to be of great physiological relevance, but rescues insulin action on glucose uptake in skeletal muscle cells.