The following primary antibodies were used: mouse monoclonal NEP (F-4; western immunoblotting [WB] 1∶1,000), mouse monoclonal CK2α (1AD9; immunofluorescence [IF] 1∶200), rabbit polyclonal calnexin (H-70; WB 1∶1,000; IF 1∶300) (all Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse monoclonal c-myc (9E10; WB 1∶1,000; IF 1∶500; immunoprecipitation [IP] 1∶200; Abcam, Cambridge, UK), rabbit polyclonal phospho-Akt (serine 473; #9271; WB 1∶1,000; Cell Signaling Technology, Danvers, MA, USA), rabbit monoclonal Akt (AW24; WB 1∶3,000; Millipore, Billerica, MA, USA), mouse monoclonal GFP (clones 7.1 and 13.1; WB 1∶3,000; IF 1∶1,000; Roche, Mannheim, Germany), rabbit polyclonal TGN46 (T7576; IF 1∶300; Sigma-Aldrich). HRP coupled streptavidin protein (ab7403; WB 1∶1,000) was obtained form Abcam, Cambridge, UK. Secondary HRP-conjugated antibodies were purchased form Sigma-Aldrich and used in a 1∶50,000 dilution. Secondary Alexa 594- and Alexa 488-conjugated antibodies were purchased form Invitrogen and used in a 1∶500 dilution. Peptides representing the N-terminal domain (aa 2–23) of NEP and conjugated with biotin on a C-terminal lysine were obtained from Peptide Specialty Labs (PSL, Heidelberg, Germany). Radiochemicals were purchased from Hartmann Analytic. 2-Dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazole (DMAT), 4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI), insulin-like growth factor-1 (IGF-1) and insulin were from Sigma-Aldrich.

The cDNAs encoding human NEP and Ezrin were kindly provided by Drs. Takaomi Saido and Monique Arpin. The cDNA of human PTEN was obtained from the German Resource Center for Genome research (RZPD, Berlin, Germany). A C-terminal myc/his-tagged variant of NEP was generated by PCR using the oligonucleotides 5′-tttggtaccatgggcaagtcagaaagtc-3′ and 5′-cccgcggccgcccaaacccggcac-3′. The resulting fragment was subcloned into the KpnI/NotI restriction sites of pcDNA4 Myc/His containing a Zeocin resistance gene (Invitrogen, Karlsruhe, Germany). The phosphorylation site mutants of NEP were generated by PCR using the following forward oligonucleotides: S6A: 5′-tttggtaccatgggcaagtcagaagctcagatg-3′; T15A: 5′-tttggtaccatgggcaagtcagaaagtcagatggatataactgatatcaacgctccaaag-3′. The oligonucleotide 5′-cccgcggccgcccaaacccggcac-3′ was used as reward primer for all constructs. The resulting PCR fragments were subcloned into the KpnI/NotI restriction sites of the expression vector pcDNA4 Myc/His Zeo.

NEP variants tagged with green fluorescent protein (GFP) were generated by PCR using the following primers: 5′-tttggtaccatgggcaagtcagaaagtc-3′, 5′-cccggatccaatgcatagagtgcgatc-3′. The resulting fragments were subcloned into the KpnI/BamHI restriction sites of pEGFP-N1 (Clontech, Mountain View, CA, USA). The serine 6 to aspartate mutant was generated by PCR using oligonucleotides 5′-tttggtaccatgggcaagtcagaagatcagatg-3′ and 5′- cccggatccaatgcatagagtgcgatc-3′.

Fusion proteins of the NEP N-terminus (aa 1-28) and glutathione S-transferase (GST) were generated by PCR using the primers 5′-ccgaattcatgggcaagtcag-3′ and 5′-cccgtcgacctactccagtggagtcc-3′. The resulting fragments were cloned into the EcoRI/SalI restriction sites of pGEX-5X-1 (GE Healthcare, Little Chalfont, UK). A fusion protein of the maltose binding protein (MBP) and the N-terminus of Ezrin (amino acids 1–310; MBP-Ezrin-NT) was generated using primers 5′-cccgaattcatgccgaaaccaatcaat-3′ and 5′-tttgtcgacttaccgggcctgggccttcat-3′. The resulting fragment was cloned into the EcoRI/SalI restriction sites of pMAL-c2 (Promega, Mannheim, Germany).

The fusion protein of GST and the C-terminal domain (aa 367–403) of PTEN (GST PTEN-CT) was generated by PCR using primers 5′-cccgaattcgatgttagtgacaatg-3′ and 5′-cccctcgagatcagacttttgtaatttgtg-3′. The resulting fragment was cloned into the EcoRI/XhoI restriction sites of pGEX-5X-1 (GE Healthcare, Little Chalfont, UK). All fusion proteins were expressed in Escherichia coli DH5α. GST and MBP fusion proteins were precipitated by GSH-resin and Amylose-resin (GE Healthcare, Little Chalfont, UK) respectively according to the supplier's instructions. All constructs were verified by sequencing of both strands.

Human embryonic kidney (HEK) 293 and human HeLa cells (both ATCC, Manassas, VA, USA) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum at 37°C in a 5% CO₂ atmosphere. DMAT was dissolved in DMSO at concentrations of 10 mM and applied to cells at a final concentration of 20 µM. Control cells were incubated with the carrier alone. Transfections of the cells with cDNA constructs were carried out using Lipofectamine 2000 (Invitrogen, Karlsruhe, Germany) according to supplier's instructions. Cells were analyzed 36–48 hrs after transfection. For the activation of Akt, HEK293 cells were starved in DMEM without serum for 16 hrs and then treated with 100 nM insulin or 1 ng/ml (0.13 nM) IGF-1 (Sigma-Aldrich, Munich, Germany) for 20 min.

Cells were washed twice with PBS, collected by scraping from the culture dish and lysed for 30 min on ice (lysis buffer: 50 mM Tris/HCl, pH 7.6, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 1% Igepal, Complete protease inhibitor mixture, PhosSTOP phosphatase inhibitor (Roche, Mannheim, Germany)). The lysate was then cleared by centrifugation (16,000× g, 4°C, 30 min). Proteins were immunoprecipitated from cleared lysates using primary antibodies (5 µg/ml) and protein G-conjugated sepharose (GE Healthcare, Little Chalfont, UK) for 3 hrs at 4°C. Precipitates were washed three times with washing buffer (50 mM Tris/HCl, pH 7.4, 500 mM NaCl, 2 mM EDTA, 0.2% Igepal) for 10 min at 4°C. Precipitates were collected by centrifugation (4,000× g, 4°C, 5 min) and eluted by incubation with SDS sample buffer (25 mM Tris/HCl, pH 6,8, 10% Glycerin, 1,5% SDS, 20 mM DTT) for 10 min at 95°C.

Proteins were then separated by SDS-PAGE, electro-transferred to nitrocellulose membranes (Whatman, Dassel, Germany), and detected with appropriate antibodies by enhanced chemiluminescence (ECL)-imaging (ChemiDoc XRS, Bio-Rad, Munich, Germany).

Phosphorylation of NEP in cultured cells was carried out as described earlier [37]. Briefly, cells were grown on 28 cm² dishes to subconfluency and then incubated for 1 h in phosphate-free media (Sigma-Aldrich, Munich, Germany). The media was aspirated, and phosphate-free media was added, together with 0.25 mCi of [³²P] orthophosphate (Hartmann Analytic, Braunschweig, Germany) for 2 hrs at 37°C. The conditioned media were aspirated, cells washed three times with PBS and immediately lysed on ice with lysis buffer. Cell lysates were centrifuged for 10 min at 16,000× g, and supernatants were immunoprecipitated with anti-myc antibodies. Radiolabeled proteins were separated by SDS-PAGE as described above and detected by autoradiography and western immunoblotting. In vitro phosphorylation assays with recombinant CK2 (alpha and beta subunit; New England Biolabs) were carried out according to the manufacturer's instructions. Phosphorylation reactions were started by addition of 10 µM [γ-³²P] ATP and allowed to proceed for 10 min at 32°C. Reactions were stopped by the addition of 5 mM EDTA. GST-fusion proteins were precipitated by GSH-sepharose (GE Healthcare, Little Chalfont, UK) and separated by SDS-PAGE.

Cells were grown on poly-L-lysine-coated glass coverslips to 50–80% confluency and fixed with 4% paraformaldehyde/PBS at room temperature and processed for immunofluorescence as described previously [38]. Bound primary antibodies were detected by Alexa 488- or Alexa 594-conjugated secondary antibodies (Invitrogen, Karlsruhe, Germany). GFP-tagged proteins were detected directly by fluorescence microscopy. The samples were analyzed with an Axiovert 200 Fluorescence Microscope (Carl Zeiss, Jena, Germany).

Surface plasmon resonance (SPR) measurements were carried out on a Biacore3000 (GE Healthcare, Little Chalfont, UK) as described previously [39], [40]. Synthetic peptides with a C-terminal biotin-label that represent amino acids 2–23 of the human NEP cytosolic domain with serine 6 in phosphorylated or non-phosphorylated state, were coupled to streptavidine (SA) chip (GE Healthcare, Little Chalfont, UK) until ∼1000 response units (RU) were reached. The chip was washed with HEPES buffer (10 mM, pH 7.4) containing 150 mM NaCl and 0.005% Surfactant P20 (HBS-P buffer; GE Healthcare, Little Chalfont, UK) at 25°C and a constant flow rate of 30 µl/min. To avoid bulk effects, a reference surface (no coupled peptide) was continuously subtracted from the signal of the active surface. To calculate the association and dissociation rates of the respective sensograms and the resulting equilibrium dissociation constant (K_D) the BIAevaluation 3.1 software was used [41], [42].

The above described biotin-labeled NEP NT peptides were coupled to Streptavidin (SA)-agarose beads (GE Healthcare, Little Chalfont, UK) and incubated with GST PTEN-CT (2 µM), MBP Ezrin-NT (100 nM) or CK2 (30 nM; alpha and beta subunit; New England Biolabs, Frankfurt, Germany) in binding buffer (10 mM Tris/HCl, pH 7.4, 150 mM NaCl, 0.4% Igepal) for 2 hrs. After washing with the same buffer, bound proteins were eluted with SDS sample buffer, separated by SDS-PAGE, and detected by Western immunoblotting with polyclonal antibodies against MBP, (MBP Ezrin-NT), PTEN (GST PTEN-CT) or monoclonal antibody against CK2α.

Cells were washed three times with ice-cold PBS and incubated on ice with PBS containing 0.5 mg/ml EZ-Link sulfo-N-hydroxysuccinimide-biotin (Thermo, Bonn, Germany) for 30 min. Cells were then washed three times with ice-cold PBS supplemented with 20 mM glycine and finally lysed on ice with lysis buffer. Biotinylated proteins were precipitated from cleared lysates with SA-agarose beads (GE Healthcare, Little Chalfont, UK) and separated by SDS-PAGE. The respective proteins were then detected by Western immunoblotting.

For enhanced chemiluminescence detection, signals were measured and analyzed using an ECL imager (ChemiDoc XRS; Bio-Rad, Munich, Germany) and the Quantity One software package (Bio-Rad, Munich, Germany). Statistical analysis was carried out using Student's t test. Significance value: ^*, p<0.05.

The N-terminal cytoplasmic domain of NEP contains several potential phosphorylation sites (Fig. 1A). To assess whether NEP undergoes phosphorylation, HEK293 cells transiently expressing myc-tagged variants of full-length NEP were labeled with ³²P-orthophosphate. Phosphate incorporation into non-glycosylated immature (∼90 kDa) and N-glycosylated mature (∼110 kDa) forms of NEP was detected [5]. Addition of okadaic acid (OA) during labeling markedly increased the phosphate incorporation into NEP, indicating that phosphoprotein phosphatases (PP) 1 and/or 2 are involved in its dephosphorylation (Fig. 1B). In order to identify the phosphorylation sites of NEP we first analyzed its cytoplasmic domain by the NetPhos 2.0 server (Fig. 1A). cDNA constructs with mutations of the potential phosphorylation sites (Fig. 1A) within the cytoplasmic domain of NEP were generated and expressed in HEK293 cells. ³²P-labeling revealed that substitution of serine residues 6 to alanine (S6A) strongly decreased phosphate incorporation. As compared to NEP wt, the phosphorylation of the Thr15Ala mutant appeared to be increased in independent experiments and might be due to effects of the mutation on substrate recognition (Fig. 1C). Together, these data indicate that serine 6 within the N-terminus is the main phosphorylation site of NEP.

Ser6 is localized within a consensus sequence for protein kinase CK2 (S-X-X-E/D) [43]. To test whether NEP is phosphorylated by CK2 in vivo, cells were labeled with ³²P-orthophosphate in the presence or absence of the selective CK2 inhibitor DMAT. Cell treatment with DMAT markedly decreased the phosphorylation of NEP in cultured cells (Fig. 1D). Next we tested the phosphorylation of NEP by CK2 in vitro. CK2 efficiently phosphorylated the GST-NEP fusion protein, but not GST alone. Importantly, CK2-mediated phosphorylation of GST-NEP was strongly decreased in a Ser6Ala variant (Fig. 1E), demonstrating that CK2 predominantly phosphorylates serine 6 within the cytoplasmic domain of NEP.

Because cell clones that stably express full-length catalytically active NEP could not be established (not shown), we generated NEP variants with substitutions of the extracellular catalytic domain by GFP (NEP-NT GFP; Fig. 2A). These constructs allowed the selection of cell clones that stably express NEP variants. In addition, the substitution of the extracellular domain of NEP by GFP allows to specifically assess functions of NEP that are independent of its proteolytic activity. To assess potential effects of the substitution of the ectodomain by GFP on the subcellular localization of NEP, we compared the distribution of NEP-NT GFP and full-length NEP by fluorescence microscopy. Both variants showed very similar distribution and localized to the endoplasmic reticulum and Golgi compartments (Fig. S1). These data demonstrate that substitution of the NEP ectodomain with GFP did not affect its subcellular trafficking and localization.

In order to mimic phosphorylated NEP we also generated a (pseudophosphorylated) Ser6Asp variant of NEP. As indicated in Fig. 2B, GFP signals were observed in juxtanuclear structures indicative for Golgi compartments for both NEP-NT GFP wild type (wt) and NEP-NT GFP Ser6Asp. Both variants were also detected at the plasma membrane indicating transport to the cell surface (Fig. 2B). The localization of NEP-NT GFP variants as well as the mature form of the full length NEP variants at the plasma membrane was also evident upon specific labeling of cell surface proteins with biotin (Fig. 2C). Thus, these NEP-NT GFP constructs proved suitable to study the role of the NEP N-terminus and its phosphorylation in cellular models.

We next analyzed the interaction of CK2 and NEP by pull-down experiments. As demonstrated in Fig. 3A, recombinant CK2 efficiently co-precipitated with synthetic peptides that represent the cytosolic N-terminal domain of NEP (NEP-NT). Notably, this interaction was strongly decreased with a synthetic peptide containing Ser6 in a phosphorylated state (pNEP-NT; Fig. 3A). These data suggested that binding of CK2 decreased after phosphorylation of NEP. The subcellular localization of NEP and CK2 was then investigated by fluorescence microscopy. Consistent with previous data [44], CK2 showed broad distribution in the cytoplasm and also localization in the nucleus (Fig. 3A). NEP-NT GFP was mainly localized in juxtanuclear compartments and at the cell surface. However, there was no significant alteration in the localization of CK2 upon expression of NEP-NT GFP. These combined data indicate that CK2 transiently interacts with the N-terminus of NEP to mediate its phosphorylation.

The cytosolic domain of NEP interacts with the scaffolding proteins Ezrin/Radixin/Moesin (ERM) [7], [11] as well as the tumor suppressor PTEN [10]. To assess whether NEP phosphorylation affects these interactions, we first performed pull down assays with peptides representing the N-terminus of NEP in a non-phosphorylated and phosphorylated.state. Ezrin co-precipitated with both the phosphorylated and non-phosphorylated variants of NEP-NT (Fig. 4A). Surface plasmon resonance (SPR) spectroscopy revealed that Ezrin binds with very high affinity to the non-phosphorylated NEP-NT (K_D∼5.8 nM). However, no significant difference in the binding of Ezrin to phosphorylated NEP-NT was observed (K_D∼5.9 nM; Fig. 4B, C), indicating that the binding of Ezrin is independent of the phosphorylation state of NEP at serine 6.

Consistent with previous data [10], the NEP-NT also bound to the PTEN C-terminus in pull-down assays. Interestingly, this interaction was strongly decreased with the NEP-NT phosphorylated at Ser6 (Fig. 4D). SPR measurements fully supported these findings. While PTEN efficiently bound the non-phosphorylated NEP-NT (K_D∼0,71 µM), the binding to the phosphorylated NEP-NT was strongly decreased (K_D could not be calculated; Fig. 4E, F). Together, these data demonstrate that the phosphorylation of NEP negatively regulates its interaction with PTEN.

To further investigate the effect of phosphorylation on the interaction of NEP and PTEN in cultured cells, we next analyzed the co-localization of endogenous PTEN with NEP-NT GFP wt and pseudophosphorylated NEP-NT GFP Ser6Asp variants. In non-transfected cells, PTEN revealed a broad distribution in the cytosol and the nucleus (Fig. 5, upper panels). In cells expressing NEP-NT GFP wt, PTEN was found to co-localize with NEP in juxtanuclear compartments. More importantly, there was also significant co-localization of NEP-NT GFP wt and PTEN at the plasma membrane (Fig. 5, middle panels and enlarged area). In contrast, very little if any co-localization of PTEN with the NEP-NT GFP Ser6Asp variant was observed at the plasma membrane (Fig. 5, lower panels and enlarged area). These data are consistent with our biochemical experiments and demonstrate that phosphorylation of NEP decreases the interaction with PTEN and its localization at the plasma membrane.

Recruitment of PTEN to the plasma membrane leads to increased dephosphorylation of PIP₃ and attenuates the activation of Akt. To investigate whether the phosphorylation-dependent interaction of NEP and PTEN could affect the activation of Akt we used HEK293 cells stably expressing NEP-NT GFP wt and Ser6Asp variants (see also Fig. 2A).

Cells were starved over night in media without serum and then treated with insulin to activate PI3K via insulin receptors. Incubation with insulin led to a time-dependent increase in the phosphorylation of Akt indicating rapid activation of PI3K. After 20 min, the phosphorylation of Akt reached a maximum in untrasfected cells (data not shown). Expression of NEP-NT GFP wt significantly attenuated the phosphorylation of Akt by 23%, indicating recruitment of PTEN to the cytoplasmic domain of NEP and increased dephosphorylation of PIP₃ (Fig. 6A, B). Interestingly, the pseudophosphorylated NEP-NT GFP Ser6Asp variant had no inhibitory effect on insulin-dependent activation of Akt (Fig. 6A, B). Very similar data were also obtained when IGF-1 was used to stimulate the phosphorylation of Akt (Fig. 6C, D).