The Sam-Sam interaction between Ship2 and the EphA2 receptor: design and analysis of peptide inhibitors

The lipid phosphatase Ship2 represents a drug discovery target for the treatment of different diseases, including cancer. Its C-terminal sterile alpha motif domain (Ship2-Sam) associates with the Sam domain from the EphA2 receptor (EphA2-Sam). This interaction is expected to mainly induce pro-oncogenic effects in cells therefore, inhibition of the Ship2-Sam/EphA2-Sam complex may represent an innovative route to discover anti-cancer therapeutics. In the present work, we designed and analyzed several peptide sequences encompassing the interaction interface of EphA2-Sam for Ship2-Sam. Peptide conformational analyses and interaction assays with Ship2-Sam conducted through diverse techniques (CD, NMR, SPR and MST), identified a positively charged penta-amino acid native motif in EphA2-Sam, that once repeated three times in tandem, binds Ship2-Sam. NMR experiments show that the peptide targets the negatively charged binding site of Ship2-Sam for EphA2-Sam. Preliminary in vitro cell-based assays indicate that -at 50 µM concentration- it induces necrosis of PC-3 prostate cancer cells with more cytotoxic effect on cancer cells than on normal dermal fibroblasts. This work represents a pioneering study that opens further opportunities for the development of inhibitors of the Ship2-Sam/EphA2-Sam complex for therapeutic applications.

Several signaling proteins are engaged to the plasma membrane by the lipid second messenger PIP3 (phosphatidylinositol 3,4,5-triphosphate) whose intracellular levels are regulated by phosphoinositide phosphatases. Among them, Ship2 (phosphatidylinositol 3,4,5-triphosphate 5-phosphatase 2) catalyses the dephosphorylation of PIP3 in position 5 to generate phosphatidylinositol (3,4) P2 and thus, downregulates different processes that are activated by PI3K (Phosphatidyl-Inositol 3 Kinase) 1,2 . In addition to this enzymatic activity, a prominent feature of Ship2 is the presence within its primary sequence of several regions able to mediate protein-protein interactions. In details, Ship2 includes from the N-to the C-termini a SH2 (Src homology 2) domain, followed by the catalytic 5-phosphatase domain, -NPXY-motifs, that generally recognize phosphotyrosine binding (PTB) domains, a proline-rich domain (PRD) with consensus sequences for SH3 modules and a sterile alpha-motif (Sam) domain 3 .
Ship2 interacts with several other proteins and acts in different processes like receptor internalization, cell spreading and adhesion, actin cytoskeletal reorganization 4 .
Ship2 is a well known target in drug discovery for type 2 diabetes as it modulates insulin sensitivity and obesity 5 . Nevertheless, Ship2 has been linked to other diseases such as neurodegenerative pathologies, atherosclerosis, as well as cancer 4 . The role of Ship2 in cancer needs to be further elucidated and is controversial. Ship2 is over-expressed in colorectal cancer where it indicates poor survival 6 . Ship2 inhibition and consequent Akt activation in gastric cancer cells contribute to improved tumorigenesis and proliferation 7 . Interestingly, Ship2 modulates EGFR (Epidermal Growth Factor Receptor) signaling: down-regulation of Ship2 in breast cancer cell lines improves EGFR internalization and degradation and arrests cell proliferation 8 .
Ship2 is also an inhibitor of EphA2 receptor endocytosis 9 ; EphA2 is a tyrosine kinase receptor that plays a complex role in cancer and is a known target in anticancer drug discovery 10 . Structural Genomics/Proteomics Initiative). Residues in the α1α2 loop and in the α5 helix belong to the EH region and are colored red. The analyzed peptide segments are delineated by arrows. (b) NMR solution structure of EphA2-Sam (first conformer, pdb entry 2E8N) in a ribbon representation: regions encompassing the S13-wt peptide are colored blue with the "KRI" and "KRIAY" sequences highlighted in orange; the additional fragment contributing to the S13-SS peptide is colored violet. The side chains of positively charged residues are shown in cyan. (c) NMR solution structure of Ship2-Sam (pdb entry 2K4P structure n.1 11 ). The ML interface is colored yellow and side chains of acidic and aromatic amino acids are reported in red and green respectively.
Scientific RePoRTS | 7: 17474 | DOI:10.1038/s41598-017-17684-5 Our work sheds further light on possible routes to target Sam-Sam interactions mediated by EphA2 and opens a window of opportunities for the design of novel compounds with different therapeutic applications.

Results and Discussion
Peptide Design. To identify peptide ligands of Ship2-Sam, we analyzed isolated EphA2-Sam regions in or close to the EH interface 11 (Fig. 1). Thus, the linear S13-wt peptide, corresponding to the 945-969 fragment from EphA2-Sam and, including the C-terminal α5 helix, the α4α5 loop and partially the C-terminal end of α4 helix (Fig. 1a,b), was first investigated. The S13-SS peptide was next conceived to better mimic the discontinuous epitope characterizing the EH interface (Fig. 1a). S13-SS includes, in addition to identical S13-wt regions, an EphA2-Sam portion covering mostly the α1α2 loop (a.a. 915-920). Moreover, two Isoleucine residues (i.e., Ile916 and Ile958) in the native sequence were mutated to cysteines to allow disulphide linkage of the two segments (915-920 and 945-969) (Fig. 1a,b and Supplemental Table S1).
Since the ML Interface in Ship2-Sam is negatively charged and also contains both a tryptophan and a phenylalanine 11 (Fig. 1c) we chose to analyze alone the short peptide motif -GHQKRIAY-(hereof named "KRI" peptide, Table S1), that is present at the N-terminal side of α5 helix in EphA2-Sam (Fig. 1a,b). It has been speculated that an intermolecular H-bond involving the backbone amide proton of the Gly residue within the "KRI" peptide sequence, may be crucial for the association between Sam domains 15,16 . Interestingly, the KRI peptide includes both aromatic residues and a stretch of positively charged amino acids which could all interact with the ML interface of Ship2-Sam (Fig. 1c). To maximize potential electrostatic and aromatic interactions with Ship2-Sam, two additional peptides containing the -KRIAY-penta-amino acid motif, repeated two (=(KRI) 2 peptide) and three (=(KRI) 3 peptide) times in tandem were studied as well ( Fig. 1a and Supplemental Table S1).

Conformational analysis by CD.
Peptide conformational studies were first performed in solution by far-UV CD spectroscopy. CD spectra in aqueous buffer clearly indicate for all peptides a prevalent random coil content for the minimum at ~200 nm (Supplemental Fig. S1). However, S13-wt and (KRI) 3 show a certain propensity to assume an α-helical conformation as indicated by TFE (2,2,2-trifluoroethanol) titration (Supplemental Fig. S2).

Structural insights by NMR spectroscopy.
Additional conformational studies were carried out by NMR spectroscopy. In aqueous solution all peptides appear unstructured in agreement with CD data. Almost complete proton resonance assignments in PBS were obtained for S13-wt (Supplemental Table S2). However, S13-wt and (KRI) 3 could be better studied in presence of TFE (2,2,2-trifluoroethanol) (Supplemental Figs S3 and S4, Tables S3, S4 and S5). TFE can reveal the intrinsic tendency of an amino-acid sequence to assume a specific secondary structure that may characterize its conformation when interacting with membranes or different peptides/ proteins 17,18 . Our recent work on several fragments from different Sam domains (Ship2 19 and the adaptor protein Odin 20 ) further shows, in agreement with earliest studies 21 , that native-like helices are well reproduced in solution containing TFE. However, propagation of helical content to the more disordered protein regions may occur at high TFE concentration 20 .
The comparison of Hα chemical shifts deviations from random coil values in PBS and in TFE solution clearly indicates for S13-wt an increase in helical content, with the percentage of helical population, estimated by chemical shifts, raising from 10% in PBS to 29% in presence of TFE 22 Fig. S3a). NMR studies in TFE reveal a completely native-like fold for the S13-wt peptide (Supplemental Fig. S3b,c and Table S4).
Similarly, structure calculations performed for the (KRI) 3 peptide in presence of 50% TFE indicate the presence of an α-helix extending through the whole peptide sequence (Supplemental Fig. S4 and Table S5).
Binding studies with EphA2-Sam. SPR   Histograms of Response Unit (RUmax) values, evaluated at 30 and 100 µM concentrations, show that the KRI and (KRI) 2 peptides are not able to bind Ship2-Sam while (KRI) 3 and S13-SS generate significant dose-response signal variations (Fig. 2). SPR sensorgrams recorded for S13-SS in a more wide concentration range, indicate that this peptide does not reach saturation (Supplemental Fig. S5b). Sensorgrams of the (KRI) 3 peptide, recorded at concentrations of 500 and 700 µM, appear quite noisy suggesting partial insolubility at the Ship2-Sam interface (Fig. 3). However, (KRI) 3 curves were fitted according to a single binding site model with 1:1 stoichiometry and provided a K D value of 100 µM, that is in agreement with the one measured by the binding isotherm K D = 83 ± 8 µM (Table 1 and Fig. 3).
The interaction between the (KRI) 3 peptide and Ship2-Sam was further investigated by MST (Microscale Thermophoresis) experiments (Supplemental Fig. S6) 23 . MST studies clearly show a protein/peptide interaction but, they do not allow an accurate K D measurement due to precipitate formation at the highest (KRI) 3 peptide concentrations (Supplemental Fig. S6).
NMR and CD assays. To further analyze the interaction between Ship2-Sam and each peptide, an NMR-based screening was conducted by means of chemical shift perturbation experiments with HSQC spectra 24 and 15 N labeled Ship2-Sam (Figs 4 and S7). These experiments appear in perfect agreement with SPR data and show more significant chemical shifts variations in the spectrum of Ship2-Sam only upon addition of the (KRI) 3 peptide (Fig. 4). Moreover, as already observed in MST assays, at the highest peptide concentrations (values > 300 µM) precipitation occurs with loss of protein signal (Fig. 4a,b) and this avoids reaching saturating conditions and achieving accurate K D evaluation. Nevertheless, we evaluated CSP (Chemical Shifts Perturbations) induced in the protein spectrum by the peptide and mapped them on the 3D solution structure of Ship2-Sam (Fig. 4c,d). Interestingly, most of the changes affects the highly negatively charged ML region of Ship2-Sam and the external close regions (such as the C-terminal side of α5 helix), which may undergo "indirect" conformational movements. Indeed the pattern of variations closely resembles the ones observed for the interaction between Ship2-Sam and EphA2-Sam 11 (Fig. 4d).
To further demonstrate that EphA2-Sam and the (KRI) 3 peptide are (completely or partially) targeting the same Ship2-Sam regions, a displacement experiment 25,26 was conducted (Fig. 5). Upon addition of unlabelled Ship2-Sam to 15 N-labeled EphA2-Sam, the [ 1 H, 15 N] HSQC spectrum undergoes several changes canonical for the Sam-Sam association (Fig. 5a) 11 . By further adding increasing amounts of the (KRI) 3 peptide, many peaks within the spectrum of the complex restore the characteristic position of the unbound EphA2-Sam form (Fig. 5b,c,d). However, despite the presence of a large excess of (KRI) 3 , displacement of Ship2-Sam is only partial (Fig. 5b). These results are likely due to the weaker binding affinity of Ship2-Sam for the (KRI) 3 peptide (Table 1) with respect to EphA2-Sam (the K D value for the Ship2-Sam/EphA2-Sam complex is 0.75 ± 0.12 µM in ITC experiments 11 ). Interestingly, a patch of negatively charged residues is located in proximity of the α2 helix in Ship2-Sam (Supplemental Fig. S8). We cannot exclude that at the highest peptide concentrations, (KRI) 3 may also target this area, that is adjacent to the ML interface, and this would reduce the exposed polar surface causing precipitation of the peptide/protein complex.
A transferred-NOESY experiment 27 was conducted and failed to evidence formation of a rigid defined conformation in the (KRI) 3 peptide upon binding to Ship2-Sam (Supplemental Fig. S9). Since the relatively low    27 , CD spectra were as well recorded for (KRI) 3 in the presence of different equivalents of Ship2-Sam (Supplemental Fig. S10). The overlay of CD spectra indicates a certain increase of order upon complex formation but, in agreement with NMR data, allows excluding a (KRI) 3 bound conformation provided with a fully defined secondary structure (Supplemental Fig. S10).

Effects of (KRI) 3 peptide in cancer cells.
After conjugating the peptide (KRI) 3 to a cell penetrating sequence (i.e., fragment 48-60 from the HIV TAT sequence (Supplemental Table S1) 28,29 ), its effects in a cellular context were evaluated (Fig. 6). Upon assessing cellular uptake (Fig. 6a), cell viability was first investigated in absence and presence of the FITC (Fluorescein-Isothiocyanate)-TAT-(KRI) 3 peptide in the carcinoma prostate cancer cell line PC-3 (Fig. 6b). PC-3 cells were chosen due to their high endogenous levels of EphA2 30,31 and Ship2 32 . After treatment with the FITC-TAT-(KRI) 3 peptide (50 µM and 200 µM concentrations) a reduction of cell viability of approximately 30% and 90% respectively, compared to the untreated cells, was observed (Fig. 6b). NMR analysis of the FITC-TAT-(KRI) 3 peptide clearly indicates that the conformational freedom of the (KRI) 3 peptide is not influenced by the presence of the FITC-TAT unit (Supplemental Fig. S11). To exclude that the reduction of cell viability could be merely linked to the FITC or the TAT sequence, experiments were conducted with FITC-TAT-Pep1 (Supplemental Table S1 and Moreover, the activity of FITC-TAT-(KRI) 3 was also analyzed in Normal Human Dermal Fibroblasts (NHDF). The FITC-TAT-(KRI) 3 peptide, at the highest tested concentration (i.e., 200 µM), has less cytotoxic activity in NHDF than in PC-3 cells (Fig. 6b).
It has recently been pointed out that pro-necrotic peptides may represent a different route to overcome the limits of pro-apoptotic anti-cancer compounds and that their specificity towards cancer cells can be highly increased by fusion to tumor-homing motifs 34 .
To get further insights into the biological function of the (KRI) 3 peptide, its "KRIAY" motif was searched against a database of bioactive peptides (i.e., the EROP (Endogenous Regulatory OligoPeptide knowledgebase)-Moscow at http://erop.inbi.ras.ru/) 35   "KRI" sequence produced several positive results including antimicrobial and antibacterial peptides 36,37 . On the same time, no hits could be retrieved by searching the sequence "KRIAY" against a database of anticancer peptides and proteins (http://crdd.osdd.net/raghava/cancerppd/) whereas the search of the "KRI" sequence produced several results among antibacterial and antimicrobial peptides with anticancer activity. In major details, the "KRI" three amino acid motif is found in the peptide Cecropin P1, that has biological activity against diverse bacteria as well as mammalian leukemias and lymphomas 38 ; in the peptide LL-37 39,40 , that reduces cell viability in many cancer cell lines including ovarian (OVRCAR-3), breast (MCF-7) and prostate cancer cell lines (LNCaP); finally in the peptide NK-11 that is active against skin cancer 41 .

Conclusions
Targeting the Ship2-Sam/EphA2-Sam complex with peptide antagonists may represent an alternative approach for the discovery of anticancer therapeutics. Thus, we analyzed isolated fragments of EphA2-Sam encompassing its binding site for Ship2-Sam (EH interface) and the close loop regions. We adopted a multidisciplinary methodology involving conformational analysis by CD and NMR spectroscopies, binding assays by NMR, SPR and MST techniques, in vitro cell-based experiments.
Structural studies point out that isolated EphA2-Sam fragments maintain a relatively poor native-like fold in aqueous buffer and are unable to bind Ship2-Sam with high affinity and specificity. However, careful analysis of obtained results along with inspection of the EphA2-Sam sequence, surrounding its binding interface for Ship2-Sam, led us to identify a "KRIAY" motif, made up of amino acids that could provide important interactions with the negatively charged ML surface. This penta-amino acid motif if repeated three times in tandem (=(KRI) 3 peptide in Table S1) produces appreciable binding to Ship2-Sam. Preliminary in cell studies show that the (KRI) 3 peptide is endowed with higher cytotoxicity against PC-3 prostate carcinoma cell line, in which endogenous levels of both EphA2 and Ship2 are elevated, compared to normal fibroblasts. The peptide contains a "KRI" three amino-acid sequence that is common to several antibacterial and antimicrobial peptides provided with anticancer activity 38,41 . Additional cell-based investigations are needed to clarify the mechanism of action of the (KRI) 3 peptide in cancer cells and to establish its correlation to the EphA2/Ship2 signaling. Although the (KRI) 3 peptide shows some cytotoxicity in normal cells we envision that an increase in selectivity towards cancer cells could be achieved through fusion with tumor-homing motifs 34 or peptide sequences specific for tumor cells over-expressing EphA2 receptor 42,43 .
In conclusion, the (KRI) 3 peptide represents a "pioneering peptide ligand" of Ship2-Sam, and opens additional opportunities for the rational design of original classes of active peptide antagonists of the Ship2-Sam/ EphA2-Sam complex which could modulate EphA2 receptor endocytosis and degradation, lowering its pro-oncogenic signaling.
Helical populations were obtained with the equation:

Displacement experiment.
To assess if (KRI) 3 could compete with EphA2-Sam for the same binding site on the surface of Ship2-Sam, we performed a displacement experiment through 2D [ 1 H, 15 N]-HSQC spectra 15,25,26 . We first recorded a spectrum of 15 N labeled EphA2-Sam in its free form then, unlabeled Ship2-Sam was added to the sample and consequent changes in the HSQC were revealed. Finally, we tried to restore the spectrum of unbound EphA2-Sam by addition of (KRI) 3 (EphA2-Sam concentration equal to 25 μM, Ship2-Sam concentration equal to 80 μM, peptide concentrations equal to 1.1 mM, 2.2 mM, 4.9 mM, and 7 mM). Circular Dichroism. CD spectra were recorded in a 0.1 cm path-length quartz cuvette on a Jasco J-810 spectropolarimeter (JASCO Corp, Milan, Italy) as already described 20,23 . Spectra derived from the average of three scans, the subtraction of blanks and the conversion of the signal to mean residue ellipticity (deg × cm 2 × dmol −1 × res −1 ). All peptides were analyzed at 100 μM, in 10 mM phosphate at pH = 7.2 and for (KRI) 3 and S13-wt TFE titrations were carried out as well. CD spectra of (KRI) 3  Reference channel signals were subtracted as blanks and the software BIAevaluation (version 4.1, GE Healthcare) was employed to estimate kinetic and thermodynamic parameters of the complexes. The Graph-Pad Prism software (version 7.00; GraphPad Software, San Diego, California) was used to fit RU max for (KRI) 3 concentrations by nonlinear regression analysis.

MST.
A Monolith NT 115 system (Nano Temper Technologies) with 100% LED and 40% IR-laser power was used to perform MST experiments. The His-Tag Labeling Kit RED-tris-NTA was implemented to achieve protein labeling. For Ship2-Sam labeling, the protein concentration was adjusted to 100 nM in labeling buffer (Nano Temper Technologies) 54 , while the dye concentration was set to 100 nM. Both protein and fluorescent dye solutions were incubated at room temperature in the dark for 30 min. A 16-step serial dilution (1:1) was prepared for (KRI) 3  Protein constructs along with expression and purification methodologies have been reported in our previous publications 11,20,23 . Proteins were either produced in LB (Luria-Bertani) broth or M9 minimal medium supplemented with 1 g/L of 15 NH 4 Cl. Bacteria were grown at 37 °C till a cell optical density OD 600nm equal to 0.6, then in the induction step isopropyl β-D-1-thiogalactopyranoside (IPTG) at 1 mM concentration (25 °C, overnight) was used. Purification of his-tagged Ship2-Sam and EphA2-Sam was achieved by affinity chromatography with an Akta Purifier FPLC System (GE Healthcare, Milano, Italy) and a Nickel column. Finally, proteins were dialyzed in PBS pH = 7.4.
Cell culture and cytotoxicity. Human prostate cancer cell lines (PC-3) were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) (GIBCO, USA), 2 mM L-glutamine (Lonza, Belgium) in 5% CO2 humidified atmosphere and harvested at approximately 90% confluence. The normal human dermal fibroblasts (NHDF) were purchased from Lonza, seeded on T-25 primary flasks (Beckton Dickinson) and maintained in fibroblast basal medium (FBM) supplemented with 2% FBS, 5 mg/ml rh insulin, 1 mg/ml hydrocortisone, 50 mg/ml ascorbic acid, 5 ng/ml rh FGF b. For cytotoxicity assay 4 × 10 3 and 3.5 × 10 3 were seeded for cancer and normal cells respectively in 50 µl medium per well in 96-well flat-bottom microplates and incubated overnight to allow cell adhesion. Subsequently, culture medium was removed and cells were incubated with 100 µl of their growth medium with different concentrations of peptide in quadruplicate for 4 hours. Reduction of cell viability was determined by the crystal violet assay. Shortly, after treatment, culture medium was removed, cells were washed with phosphate-buffered saline (PBS), then they were fixed and stained with 0.1% (w/v) crystal violet in 25% methanol for 30 minutes at dark; afterwards, crystal violet was removed and cells were washed twice with double distilled water and let dry. In the end, the cells were solubilized by adding 10% (v/v) acetic acid and the amount of dye taken up was quantified with a plate reader (Multiskan Fc 10094, Thermo) at 595 nm. Data were expressed as mean ± SD of three independent experiments. Flow cytometer analysis. PC-3 cells, plated at a density of 6 × 10 5 , were incubated with FITC-TAT-peptides for 4 hours at 37 °C. Then, the cells were gently washed twice with PBS, permeabilized with 0.1% Triton X-100 in 33 mM sodium citrate and analyzed with a flow cytometer equipped with a 488 nm argon laser (FACScan, Becton Dickinson, CA, USA). A total of 20,000 events per sample were collected. Values of fluorescence intensity were obtained from histogram statistic of CellQuest software. LDH assay. The release of LDH was used as a marker for cytotoxicity. After treatment with 50 µM FITC-TAT-(KRI) 3 for 4 hours, cleared supernatants were incubated with reaction buffer for 30 minutes at room temperature according to the manufacturer instructions (ab65393, LDH-Cytotoxicity Assay Kit II). Cells untreated and treated with 1% Triton-X 100 in PBS were used respectively as negative and positive control. Absorbance was measured at 490 nm and the cytotoxicity percentage was calculated as: [(Absorbance_test sample -Absorbance_low control)/ (Absorbance_high control -Absorbance_low control) × 100.
Data availability statement. S13-wt NMR structures in presence of TFE have been deposited in the Protein Data Bank (PDB entry code: 5NZ9). NMR structures calculated in presence of TFE for (KRI) 3 , whose amino acid sequence is short and unnatural, are not publicly available but will be provided by the corresponding author on reasonable request.