The authors have declared that no competing interests exist.
Receptor tyrosine phosphatase sigma (RPTPσ) plays an important role in the regulation of axonal outgrowth and neural regeneration. Recent studies have identified two RPTPσ ligands, chondroitin sulfate proteoglycans (CSPGs) and heparan sulfate proteoglycans (HSPG), which can modulate RPTPσ activity by affecting its dimerization status. Here, we developed a split luciferase assay to monitor RPTPσ dimerization in living cells. Using this system, we demonstrate that heparin, an analog of heparan sulfate, induced the dimerization of RPTPσ, whereas chondroitin sulfate increased RPTPσ activity by inhibiting RPTPσ dimerization. Also, we generated several novel RPTPσ IgG monoclonal antibodies, to identify one that modulates its activity by inducing/stabilizing dimerization in living cells. Lastly, we demonstrate that this antibody promotes neurite outgrowth in SH-SY5Y cells. In summary, we demonstrated that the split luciferase RPTPσ activity assay is a novel high-throughput approach for discovering novel RPTPσ modulators that can promote axonal outgrowth and neural regeneration.
Deregulation of protein tyrosine phosphorylation impacts a broad spectrum of human diseases including obesity, vascular diseases, cancers and neural degeneration [
In contrast to receptor tyrosine kinases, receptor tyrosine phosphatases (RPTPs) are inactive as homodimers. An initial model proposed that a motif of the RPTPs located between the transmembrane segment and the first catalytic domain (D1), named the “wedge domain”, mediates dimer formation and causes RPTP inactivation [
Here, we developed a split firefly luciferase-based sensor system to monitor in real time RPTPσ dimerization/activity in response to various ligands. Using this system, we have validated new anti-RPTPσ antibodies that are capable of modulating axonal outgrowth, and which could be developed as potential therapeutics for the treatment of nerve injury and other neurodegenerative diseases.
293T/17, COS7, SH-SY5Y cells were routinely maintained in DMEM containing 10% Fetal Bovine Serum and 1% penicillin/streptomycin. The following antibodies were used: Anti-PTPRS clone 1H6 (Abnova), Anti-RPTPσ clone 17G7.2 (MediMabs), Anti-tubulin (Sigma). The following reagents were used in this study: Chondroitin Sulfate (Sigma), Heparin (Sigma), Aggrecan (Sigma) and Chondroitin Sulfate Proteoglycan (CSPG) (Millipore), D-luciferin (ThermoFisher Scientific).
The murine (NM_011218) and human (BC143287) RPTPσ full-length cDNAs were sub-cloned using PCR-based strategies. The PCR products were digested by EcoRI and SalI restriction enzymes and inserted into pRK5-Nluc, pRK5-Cluc, pRK5-FN and PRK5-CF vectors (gift from David Piwnica-Worms) [
293T/17 cells were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. 5000 cells were plated in a 96 well plate for 24 hrs followed by transient transfection (0.1 μg/well) of the various constructs. Following transfection, 1 mM D-luciferin in phenol red free medium was added for 30 min at 37°C. Quantification of the luminescence was carried out using a luminometer (FLUOstar Omega) analysed in real-time for a total of 15 min with data collected at 3 min interval.
Two micrograms of plasmid DNA was transfected by lipofectamine 2000 in 293T/17 cells. 48 hours post-transfection, cells were treated with purified CSPG then lysed with NP-40 lysis buffer (1% NP-40, 20 mM HEPES, pH 7.4, 150 mM NaCl, 5mM NaF, 1 mM NaPO4, 10% Glycerol) containing 1X protease inhibitor (Roche) and 1 mM sodium orthovanadate. Cleared lysates were collected after centrifugation at 12000 rpm for 10 minutes. Protein concentration was determined by BCA (bicinchoninic acid) assay and the lysate supernatants were divided into two portions. One portion was added to 2X SDS-sample buffer (100 mM Tris-Cl (pH 6.8), 4% (w/v) SDS (sodium dodecyl sulfate), 0.2% (w/v) bromophenol blue, 20% (v/v) glycerol, 200 mM β-mecaptoethnol) and the other portion was added to coomassie blue sample buffer (100 mM Tris-Cl, pH6.8, 20% glycerol, 5% Coomassie Briliant Blue g-250). The protein sample with Coomassie blue sample buffer was subjected to native or SDS PAGE (Polyacrylamide gel electrophoresis) followed by standard western immunoblotting procedures as described below.
Proteins were separated by SDS-PAGE and transferred onto polyvinylidende difluoride (PVDF) membrane (Immobilon-P Millipore). The membrane was blocked with 5% BSA in TBST at room temperature then was incubated with primary antibody. For the immunoprecipitation experiment, 500 μg of cell lysates were incubated with the indicated RPTPσ antibodies conjugated to agarose beads for 4 hr. Immunoprecipitated complex were washed with NP-40 lysis buffer for 3 times followed by western blot analysis.
Six micrograms of RPTPσ expression constructs were transfected into 297T/17 cells in a 10 cm dish using lipofectamine 2000 according to the manufacturer protocol. Ten dishes of transfected cells were rinsed with cold PBS then scraped in 1 ml PBS and collected by centrifugation. Cells were resuspended in membrane wash buffer (0.32 M sucrose, 10 mM Tris-HCl, pH7.5, 5 mM MgCl2, 50 mM NaCl) and homogenized using a glass Dounce homogenizer using 25 strokes. Different cellular fractions were separated by ultracentrifugation at 20,000 rpm for 10 min at 4°C. Membrane protein fractions were pelleted and washed 3 more times with the same wash buffer and the concentration measured by BCA assay.
50 μg of the membrane protein fraction overexpressing GFP or RPTPσ N/C was incubated with different amount of Chondroitin sulfate or Heparin at 37°C for 30 minutes followed by the addition of 1 mM p-nitrophenyl phosphate for 30 min. The assay was conducted at 25°C in 96 well plates and the absorbance of p-nitrophenol was monitored at 405 nm every minute using a Varioskan plate reader (Thermo Electron).
Specific sequences in the extracellular domain of human RPTPσ were selected to generate peptides. The KLH-linked RPTPσ peptides were injected into 2 months old RPTPσ KO mice in order to induce antibody-dependent immune response. Four weeks after post injection, mice’s blood was withdrawn for ELISA analysis. Mice positive to ELISA further injected KLH-linked RPTPσ peptides for the second boost of immune response. Mice reached to 3 months old then were sacrificed to take out blood and spleen for hybridoma generation. Mice were euthanized using isofluorane and carbon dioxide followed by cervical dislocation. All animal procedures were approved by the McGill Animal Care Committee and were conducted according to the Canadian Council of Animal Care ethical guidelines for animal experiments.
Epitope peptides from human RPTPσ extracellular region were coated onto 96 well plates overnight. The coated plates were washed with PBS 3 times to remove the excess peptides and 3% BSA solution in PBS was added to the peptide coated plate to block non-specific binding. Supernatant from individual hybridoma medium was added and incubated for 1 hr. Followed by 3 washes, a secondary antibody (1:10000, anti-mouse IgG-HRP) was added, incubated for 1 hr. After 3 washes with PBS, ABTS (2,2'-Azinobis [3-ethylbenzothiazoline-6-sulfonic acid]-diammonium salt) was added and the absorbance read at 410 nm using a Varioskan plate reader (Thermo Electron).
Six micrograms pRK5-hRPTPσ was transfected in COS7 cells using lipofectamine 2000. 24 hrs following transfection, cells were detached using 2.5 mM EDTA treatment and blocked by “Fc blocker” (BD) for 30 min on ice. Primary RPTPσ antibody (1H6) (1:100) or hybridomas supernatant was added to label the epitope on the cell surface for one hour on ice. Cells were then washed with cold PBS containing 2% FBS and labeled with anti-mouse FITC (1:100) for one hour on ice. Labeled cells were subjected to FACS analysis. The result was analysed using Flowjo software (Flowjo).
SH-SY5Y human glioma cells were treated with 3% FBS, 10 μM Retinoic acid for 96hr in order to induce cell differentiation. After 96 hr, cells were washed with PBS and fixed by 4% paraformaldehyde for 5 min on ice. Cells were incubated with 5% BSA for one hour at RT in order to prevent non-specific binding. Primary antibody (IgG, anti-hRPTPσ 1.3H12, anti-hRPTPσ 4.5H5 at 10 μg/ml each) was applied to cells and incubated for 2 hr at RT. Following PBS washes, cells were incubated with the secondary antibody (1:2000, anti-mouse FITC) for 2 hr at RT. Finally, cells were washed 3 times with PBS and stained with actin dye (Phalloidin) and nucleus dye (DAPI) for 30 min. Samples were mounted and subjected to immunofluorescence analysis.
SH-SY5Y cells were infected with lentivirus carrying GFP. The infection procedures were described in a previous study [
The split luciferase system is particularly well suited for evaluating real-time receptor/ligand interactions in a cellular context [
(A) Schematic of RPTPσ constructs used in this study. Different RPTPσ domains were fused with either Nluc or Cluc fragment independently as indicated. FN: Fibronectin domain; Ig: Immunoglobulin-like domain. (B) Western blot analysis following transfection in 293T/17 cells with various RPTPσ constructs. IB: Immunoblot. (C) 293T/17 cells were transfected with various RPTPσ N/C constructs as indicated and 48 hrs post-transfection, relative luminescence unit (RLU) was determined using a luminometer following the addition of 1mM of D-luciferin. (D) 293T/17 cells were co-transfected with RPTPσ FN/FC and 48 hrs post-transfection, rapamycin was added for 4 hr in order to induce FRB and FKBP association followed by luminescence measurement. Values represent fold changes vs control (DMSO) and the mean of three independent experiments performed in triplicates is given ± SD. * p < 0.05.
In order to confirm that RPTPσ dimerization can be induced in this assay, we used the FRB-rapamycin-FKBP dimerization system [
Since dimerization and monomerization of RPTPσ could be seen in live cells, it opens the possibility that other ligands could modulate the dimerization status of RPTPσ through ligand/receptor interactions. Here we initially tested chondroitin sulfate (CS) and heparin, an analog of heparan sulfate since they are reported ligands of RPTPσ [
(A-C) 293T/17 cells were co-transfected with RPTPσ N/C and 48 hr post-transfection, cells were treated with the indicated RPTPσ ligands for 5 min followed by luminescence measurement. Values are expressed as percentage vs control (without treatment) and the mean of three independent experiments performed in triplicates is given ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001. (D) RPTPσ-transfected 293T/17 cells were treated with the indicated concentration of CSPG for 30 min and protein lysates were separated on a native (
Inactivation of the RPTPs CD45 and RPTPα has been show to be mediated by receptor dimerization [
(A) 293T/17 cells were co-transfected with GFP control or RPTPσ N/C and 48 hrs post-transfection, cells were lysed and membrane fraction were isolated followed by western blot analysis. (B) The membrane fraction was incubated with the phosphatase substrate PNPP and release of PNP was measure using the Varioskan plate reader. Phosphatase activity values are expressed as percentage over GFP control. (C-D) 50 ug of isolated protein membranes fractions were incubated with different amount of Chondroitin Sulfate (CS) or heparin for 30 min at RT followed by the phosphatase activity assay. Values are expressed as percentage vs control (without treatment) and the mean of three independent experiments performed in triplicates is given ± SD. *, p <0.05.
In an effort to generate antibodies targeting the extracellular domain of RPTPσ, we generated and screened hybridoma clones to obtain 17 positive clones revealed by ELISA (
(A) ELISA assay using peptides from human RPTPσ extracellular region coated in 96 well plates and incubated with the supernatant of hybridomas. (B) COS7 cells were transfected with human RPTPσ (hRPTPσ) and FACS analysis were performed using the supernatant of 11 positives hybridomas identified by ELISA. IgG is the negative control whereas the commercially available clone1H6 is the positive control. hRPTPσ positive represent the percentage of transfected cells detected expressing hRPTPσ on the cell surface. All experiments were performed in triplicates for a total of three experiments, which gave similar results. One representative experiment is given ± SD. * p < 0.05, ** p < 0.01.
Based on ELISA and FACS analysis, 1.3H12 and 4.5H5 antibodies were selected because they displayed the highest sensitivity for RPTPσ. They were purified for further characterization. The two antibodies reacted only with the human form of RPTPσ but not the mouse isoform (
COS7 cells were transfected with either human RPTPσ (hRPTPσ) or mouse RPTPσ (HA-mRPTPs) and 24 hrs post-transfection, western blot (A) and immunoprecipitation (B) analysis was performed using the indicated purified monoclonal antibodies. Antibody 1H6 reacted with both forms was used as positive control. Actin was the loading controls.
We tested whether antibodies modulate the dimerization of RPTPσ using the split luciferase assay. Our result shows that the 4.5H5 antibody but not the 1.3H12 antibody significantly increased luminescence, indicating that 4.5H5 was able to significantly induce the dimerization of RPTPσ (
(A) 297T/17 cells were transfected with hRPTPσ N/C and 48 hrs post-transfection, cells were treated with various amount of the indicated RPTPσ monoclonal antibody followed by luminescence measurement. IgG was used as negative control. Values are expressed as fold changes vs respective IgG control. All experiments were performed in triplicates for a total of three experiments, which gave similar results. One representative experiment is given ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001 (B) SH-SY5Y neuroblastoma cells were treated with human RPTPσ antibodies (10μg/ml) in the differentiation medium. 96 hr post-treatment cells were subjected to image analysis using the NeuronJ program and neurite length and the distribution of neurite length (C) were determined. A.U.: arbitrary unit. One representative experiment of a total of three experiments is shown. ***, p < 0.001.
RPTPσ is one of the three members of the type II sub-family of receptor PTPs [
We selected the split luciferase assay, which can be used to monitor protein-protein interactions in intact cells [
CSPGs and HSPGs are newly identified ligands of RPTPσ [
Using our screening system, we confirmed that CSPG is a ligand of RPTPσ and that CSPGs prevent RPTPσ from dimerizing, resulting in its activation. However, HSPG-mediated RPTPσ dimerization seems moderate in our system. One possible explanation is that RPTPσ exists primarily as a dimer in cells and it is more difficult to induce further dimer formation and inhibition of phosphatase activity.
Our results indicate that the ligand-dependent switch between the monomeric and the dimeric forms of RPTPσ provides an important means of regulating RPTPσ function. There is still much to explore regarding the interaction of CSPGs and HSPGs with RPTPσ from a structural point of view. The crystal structure of the extracellular domain of RPTPσ has been resolved and gives us some structural insights into ligands/receptor interactions [
The development of small molecule inhibitors has been extremely difficult for researchers studying protein tyrosine phosphatases. Traditional PTP inhibitors mostly target the catalytic domain of PTPs, which is highly conserved among all PTPs; hence they generally suffer from lack of specificity toward its targeted PTPs. Recently, Silver’s group has developed a wedge domain peptide-based RPTPσ inhibitor and showed its effectiveness in a spinal cord injury animal model [
Antibodies of the IgG class generally mediate receptor dimerization since they are bivalent in nature, facilitating the non-covalent crosslinking of two receptors. However, cumulative evidence indicates that the structures of the receptor ectodomains have to be considered when designing therapeutic antibodies. Indeed, taking lessons from therapeutic antibodies directed against EGFR [
Collectively, data collected using our RPTPσ split luciferase system not only provide a cell-based platform to examine the genuine properties of RPTPσ in live cells, but also reveals its potential for the screening of ligands and potential therapeutic antibodies. This system could ultimately lead to discovery of novel RPTPσ inhibitors designed for the treatment of patients suffering from spinal cord injury or a broad array of neurodegenerative disorders.
We thank David Piwnica-Worms for providing the plasmids for the split luciferase system.