A soluble derivative of PrP C activates cell-signaling and regulates cell physiology through LRP1 and the NMDA Receptor

Cellular prion protein (PrP C ) is a widely expressed, glycosyl-phosphatidylinositol-anchored membrane protein. PrP Sc is a misfolded and aggregated form of PrP C responsible for prion-induced neurodegenerative diseases. Understanding the function of the nonpathogenic PrP C monomer is an important ob-jective. PrP C may be shed from the cell surface to generate soluble derivatives. Herein, we studied a recombinant derivative of PrP C (S-PrP), which cor-responds closely in sequence to a soluble form of PrP C shed from the cell surface by ADAM family proteases. S-PrP activated cell-signaling in PC12 and N2a cells. TrkA was transactivated by Src family kinases (SFKs) and ERK1/2 was activated downstream of Trk receptors. These cell-signaling events were dependent on the N-methyl-D-aspar-tate Receptor (NMDA-R) and Low Density Lipoprotein Receptor-related Protein-1 (LRP1), which functioned as a cell-signaling receptor system in lipid rafts. Membrane-anchored PrP C and NCAM were not required for S-PrP-initiated cell-signaling. S-PrP promoted PC12 cell neurite outgrowth. This response required the NMDA-R, LRP1, SFKs, and Trk receptors. In Schwann cells, S-PrP interacted with the LRP1/NMDA-R receptor system to activate ERK1/2 and promote cell migration. The effects of S-PrP on PC12 cell neurite outgrowth and SC migration were similar to those caused by other proteins that engage the LRP1/NMDA-R receptor system, including activated α 2 -macroglobulin and tissue-type plasminogen activator. Collectively, these results demonstrate that shed forms of PrP C may exhibit important biological activities in the CNS and the PNS by serving as ligands for the LRP1/NMDA-R receptor system.


Introduction
Host-encoded cellular prion protein (PrP C ) is a widely expressed glycosyl-phosphatidylinositol (GPI)-anchored membrane protein (1,2). In prion diseases, PrP Sc , a misfolded conformer of PrP C , functions as an infectious agent, recruiting PrP C monomers into multi-molecular assemblies, which accumulate to high levels in the CNS. Whereas PrP C is monomeric and predominantly alpha-helical in structure, PrP Sc is rich in β-sheets and insoluble in non-denaturing detergents (3,4). PrP Sc aggregates include oligomers and insoluble fibrils, which continuously accumulate to cause the onset and progression of neurodegeneration and gliosis (2,5).
Understanding the activity of naturally occurring PrP C is an important objective. Because PrP C is GPI-anchored (6), it localizes largely to lipid rafts, which are known to be important for assembly of molecular scaffolds involved in signal transduction (7). PrP C also may be shed from cell surfaces to generate soluble derivatives. Various ADAMs have been implicated in PrP C shedding, including ADAM8, ADAM9, and ADAM10 (8)(9)(10).
There is considerable evidence that membraneanchored and soluble PrP C derivatives participate in signal transduction to regulate processes such as neuronal differentiation and synapse formation (11)(12)(13)(14). Neural cell adhesion molecule (NCAM) has been implicated in PrP C -activated cell-signaling (11,13,15). Metabotropic glutamate receptor type 5 (mGluR5) and the N-methyl-D-aspartate receptor (NMDA-R) function with membrane-anchored PrP C to activate signal transduction in response to α-synuclein oligomers (14). Irrespective of the receptors involved, activation of Src family kinases (SFKs) appears to be a key step in the pathway by which PrP C regulates cell physiology (11,13,14).
LDL Receptor-related protein (LRP1) is an endocytic and cell-signaling receptor expressed by diverse cell types (16). In the CNS, LRP1 is expressed by neurons and concentrated in postsynaptic densities, where it regulates synaptic plasticity (17)(18)(19). In response to ligands, including activated α2-macroglobulin (α2M) and tissue-type plasminogen activator (tPA), LRP1 activates SFKs in neurons and neuron-like cell lines (20). This pathway requires the NMDA-R in addition to LRP1; together these receptors form a cell-signaling receptor complex that functions in lipid rafts (21)(22)(23). Other receptors, which may be recruited into the LRP1/NMDA-R receptor complex to activate signal transduction include Trk receptors and p75 NTR (22,24). Important consequences of SFK activation, downstream of LRP1 and the NMDA-R, include Trk receptor transactivation and ERK1/2 activation (20). Activation of the LRP1/NMDA-R signaling receptor complex induces neurite outgrowth in neurons and migration of Schwann cells (SCs) (20,25).
In addition to LRP1 and the NMDA-R, activation of cell-signaling by tPA is reported to require membrane-anchored PrP C (22,26,27). Herein, we studied the cell-signaling activity of a recombinant product, S-PrP, corresponding closely in structure to the soluble form of PrP C released by ADAM10 (8). We demonstrate that S-PrP activates ERK1/2 by a pathway that requires LRP1, the NMDA-R, SFKs, and Trk receptors. Membrane-anchored PrP C and NCAM were not required. S-PrP induced neurite outgrowth in PC12 cells and SC migration.
Our results support a model in which the LRP1/ NMDA-R signaling receptor complex is capable of mediating biological activities of soluble PrP C derivatives and possibly membrane-anchored PrP C expressed by adjacent cells.

S-PrP initiates cell-signaling in neuron-like cells
A recombinant derivative of PrP C (S-PrP), corresponding roughly in sequence (residues 23-231) to the PrP C ectodomain fragment shed by ADAM10 (8), was expressed in Escherichia coli. SDS-PAGE and silver staining confirmed that a single major product with the anticipated molecular mass was purified (Fig. 1A). To authenticate this recombinant protein, the only detected band was excised from the gel and, following reduction and alkylation, proteolyzed with trypsin. Tryptic peptides were identified by LC-MS/MS. Over 98% of the identified peptides corresponded in sequence to mouse PrP C . Fig. 1B shows that the identified peptides covered the PrP C sequence from aa 24-228, closely matching the anticipated structure of S-PrP and excluding the known N-terminal PrP C signal peptide and C-terminal GPI-anchor signal sequence (28).
Neuron-like PC12 pheochromocytoma cells were treated with S-PrP at increasing concentrations for 10 min. ERK1/2 phosphorylation was determined by immunoblot analysis and evident in cells treated with PrP C at concentrations of 30 nM or higher (Fig.  1C). To confirm that the effects of S-PrP on ERK1/2 phosphorylation were not due to endotoxin contamination, S-PrP was boiled at 100 o C for 5 min before addition to PC12 cell cultures. This procedure inactivates most proteins but fails to affect the bioactivity of endotoxin (29). Fig. 1D shows that boiling completely eliminated the ability of S-PrP to activate ERK1/2.
Next we tested the ability of four PrP C -specific monoclonal antibodies to inhibit ERK1/2 activation in response to S-PrP (30). POM2 recognizes the N-terminal octarepeat region and thus, has the greatest potential to sterically interfere with interactions involving the absolute N-terminus, the octarepeat region, or the central unstructured region. POM3 recognizes an epitope C-terminal to the POM2 epitope and near the center of PrP C . POM1 and POM19 recognize epitopes in the C-terminal globular region of PrP C (30). PC12 cells were treated with 40 nM S-PrP in the presence of each antibody (10 µg/mL) for 10 min. Only POM2 blocked ERK1/2 phosphorylation in response to S-PrP (Fig. 1E). The activity of POM2 confirms that the effects of S-PrP are not due to a contaminant such as endotoxin. Furthermore, the selective activity of POM2 suggests that the N-terminal "non-globular" region of S-PrP is most likely involved in triggering signal transduction.

Activation of cell-signaling by S-PrP requires LRP1 and lipid rafts
We showed previously that LRP1 localizes transiently to lipid rafts (31) and that this sub-population of LRP1 is essential for LRP1-initiated cellsignaling (23). Because LRP1 and membrane-anchored PrP C apparently function as co-receptors to trigger cell signaling events initiated by the LRP1 ligand, tPA (27), we tested whether LRP1 is necessary for S-PrP-initiated cell signaling. To begin, we treated PC12 cells with S-PrP together with Receptor-associated Protein (RAP), which binds to LRP1 and inhibits binding of all other known ligands that activate signal transduction (16,22,32). In the absence of RAP, 40 nM S-PrP induced ERK1/2 phosphorylation in PC12 cells; however, this response was entirely blocked by 150 nM RAP ( Fig. 2A). ERK1/2 phosphorylation in response to S-PrP also was blocked by methyl-β-cyclodextrin (MβCD), a cholesterol sequestration reagent known to disrupt lipid rafts (33). These results suggest that lipid raft-associated LRP1 is involved in S-PrP-initiated cell signaling.
To further test whether intact lipid rafts are required for S-PrP-initiated cell signaling, we pre-treated PC12 cells with Fumonisin B1 (FM, 25 μM) for 4 24 h. FM blocks synthesis of sphingolipids that are key components of lipid rafts (34). Fig. 2B shows that FM blocked ERK1/2 phosphorylation in response to enzymatically-inactive (EI) tPA, as previously demonstrated (23), and in response to S-PrP.
Similar results were obtained when we studied mouse N2a neuroblastoma cells (Fig. 2C). ERK1/2 was phosphorylated in N2a cells treated with 40 nM S-PrP. The response to S-PrP was blocked by 150 nM RAP and by pre-treating the cells with MβCD. Thus, S-PrP activates cell-signaling in two distinct neuron-like cell lines via a pathway that is inhibited by RAP and by disrupting lipid rafts.
In addition to LRP1, RAP blocks binding of ligands to other receptors in the LDL receptor gene family (35). To confirm the importance of LRP1 in S-PrPactivated signaling, we silenced LRP1 gene expression in PC12 cells with siRNA. Fig. 2D shows that LRP1 mRNA expression was significantly decreased in cells transfected with LRP1-specific siRNA, compared with cells transfected with nontargeting control (NTC) siRNA. Fig. 2E shows that LRP1 gene-silencing blocked activation of ERK1/2 in response to S-PrP. MβCD inhibited ERK1/2 phosphorylation in cells transfected with LRP1specific or NTC siRNA.

Activation of cell signaling by S-PrP requires the NMDA-R
LRP1 activates cell-signaling in response to diverse ligands by functioning as part of a system of receptors, including the NMDA-R, which plays an essential role (21, 22,36). To test whether S-PrPactivated cell-signaling requires the NMDA-R, we treated PC12 cells with dizocilpine (MK801), a non-competitive antagonist of the NMDA-R (37). MK801 blocked ERK1/2 phosphorylation in response to the known LRP1 ligand, tPA, which was introduced as an enzymatically-inactive (EI) derivative (Fig. 3A), as anticipated (22). MK801 also blocked ERK1/2 phosphorylation in response to S-PrP. To further test the activity of the NMDA-R in S-PrP-activated signaling, we treated PC12 cells with dextromethorphan hydrobromide (DXM), a non-competitive NMDA-R antagonist (38). DXM blocked ERK1/2 phosphorylation in response to S-PrP (Fig. 3B).
Finally, to confirm the importance of NMDA-R in S-PrP-activated signaling, we applied a gene silencing approach, targeting the GluN1 NMDA-R subunit with siRNA in PC12 cells, as previously described (22). The GluN1 subunit is essential for assembly of intact NMDA-R heterotetramers (39). Our gene-silencing approach decreased GluN1 mRNA expression by 80±3%, as determined by RT-qPCR (n=3). Fig. 3C shows that GluN1 genesilencing blocked ERK1/2 activation in response to S-PrP.

S-PrP induces neurite outgrowth in PC12 cells
ERK1/2 activation has been implicated in neuronal differentiation and in neurite outgrowth in PC12 cells (40). We examined the effects of S-PrP on PC12 cell neurite outgrowth. Fig. 4 shows that S-PrP promoted neurite outgrowth in PC12 cells maintained in serum-free media (SFM) for 48 h. The extent of neurite outgrowth was similar to that detected in cells treated with NGF-β. RAP and MK801 did not independently regulate neurite outgrowth: however, both reagents entirely blocked the response to S-PrP, suggesting that LRP1 and the NMDA-R are necessary for S-PrP-induced PC12 cell neurite outgrowth.

Trk is transactivated by SFKs upstream of ERK1/2 in S-PrP-treated PC12 cells
SFKs have been implicated in cell signaling events initiated by PrP C in association with NCAM and mGluR5 (11,13,14). SFKs also are activated in neurons and neuron-like cells when specific ligands such as tPA or activated α2-macroglobulin bind to LRP1 (20). Activated SFKs transactivate Trk receptors, which leads to ERK1/2 activation (20). To further test this model, PC12 cells were treated for 2 h with the Trk tyrosine kinase inhibitor, K252a (10 nM), and then with S-PrP. K252a blocked ERK1/2 phosphorylation in response to S-PrP (Fig.  5D).
To test whether S-PrP-activated cell-signaling in PC12 cells requires membrane-anchored PrP C , we silenced PRNP gene expression in PC12 cells with siRNA. Fig 5E shows that gene silencing was nearly complete at the protein level, as determined by immunoblot analysis. When PrP C -deficient PC12 cells were treated with S-PrP (40 nM), ERK1/2 was activated, suggesting that membraneassociated PrP C may not be necessary for this response.
Next, we studied whether NCAM is necessary for activation of cell-signaling by S-PrP in PC12 cells. NCAM1 gene expression was silenced in PC12 cells with siRNA. Fig. 5F shows that NCAM protein expression was nearly but not entirely eliminated by gene silencing. In PC12 cells in which NCAM1 was silenced, S-PrP (40 nM) still activated ERK1/2, suggesting that the LRP1/NMDA-R-dependent pathway by which S-PrP activates cell-signaling is probably distinct from the previously described NCAM pathway (11,13,15).
The results of our cell signaling experiments suggested a model in which SFKs and Trk receptors function downstream of LRP1 and the NMDA-R in the pathway by which S-PrP activates ERK1/2. To test whether SFKs and Trk receptors are necessary for S-PrP-induced PC12 cell neurite outgrowth, we treated PC12 cells with 40 nM S-PrP in the presence and absence of PP2 and K252a for 48 h. Both reagents blocked neurite outgrowth in response to S-PrP (Fig. 6).

S-PrP activates cell-signaling in SCs by a pathway that requires the NMDA-R and LRP1
SCs express LRP1 and the NMDA-R (41,42). In response to specific ligands that bind to the LRP1/ NMDA-R receptor complex, ERK1/2 is activated and increased SC migration is observed (25,(41)(42)(43). As a second model system to test whether the LRP1/NMDA-R receptor system activates cell signaling in response to S-PrP, we treated primary cultures of rat SCs with S-PrP (40 nM) for 10 min. Fig. 7A shows that ERK1/2 was activated and the response was blocked by RAP (150 nM) and MK801 (1.0 μM).
Next, we studied the effects of S-PrP on SC migration using Transwell cell migration devices. SCs were treated with S-PrP in the presence or absence of RAP or MK801 in the upper chamber and then allowed to migrate for 48 h. Representative images of cells that migrated to the underside surfaces of Transwell membranes are shown in Fig. 7B. The results of three separate experiments with different SC preparations are summarized in Fig. 7C. S-PrP significantly increased SC migration. RAP and MK801 blocked SC migration in response to S-PrP. Thus, S-PrP promotes SC migration by a pathway that is inhibited by antagonists of LRP1 or the NMDA-R.

Discussion
PrP C is expressed at high levels by neurons and astrocytes in the central nervous system (CNS), by neurons in the peripheral nervous system (PNS), and by diverse cell types outside the nervous system (44,45). Nevertheless, mice in which the gene encoding PrP C , PRNP, is globally deleted develop normally (46). In neurons and non-neuronal cells, PrP C has been reported to inhibit apoptosis, although this activity may be context-specific (47)(48)(49)(50). In the PNS, PRNP gene deletion causes a chronic form of demyelinating polyneuropathy (51). Because conditional deletion of PRNP in SCs did not replicate the demyelinating disorder, it is assumed that SCs respond to neuronal PrP C , which may be membrane-anchored, due to the intimate association of axons with SCs, or present in a soluble form similar to the derivative studied here (S-PrP).
Membrane-anchored PrP C has been shown to function in concert with LRP1 to initiate cell-signaling in response to tPA (26). This is intriguing because there are similarities in the reported activities of tPA and PrP C relative to neuronal cell physiology. For example, tPA is reported to be neuroprotective and inhibit neuron apoptosis (52)(53)(54). The effects of tPA on cell survival in the CNS are independent of its activity as a protease and instead dependent on interaction with the LRP1/NMDA-R receptor system (52)(53)(54). The LRP1/NMDA-R receptor system also promotes survival of SCs in the PNS (42).
We studied a soluble recombinant derivative of PrP C , which is almost full-length. Due largely to the activity of ADAM gene family members, as much as 10-15% of membrane-anchored PrP C may be released as soluble derivatives (7-9, 54, 55). By studying recombinant S-PrP, we hoped to uncover activities attributable to soluble derivatives of PrP C . Our studies also may model interactions of membrane-anchored PrP C with receptors expressed by closely associated cells in trans configuration.
The results presented here show that S-PrP activates ERK1/2 in PC12 cells, N2A cells, and SCs. The response requires the NMDA-R in addition to LRP1 and causes changes in cell physiology that are anticipated for this receptor system, including neurite outgrowth in PC12 cells (20,57) and increased SC migration (25,42). SFKs played a critical role in S-PrP-initiated cell-signaling, transactivating Trk receptors, which are apparently upstream of ERK1/2, as previously reported (20). Inhibiting SFKs with PP2 blocked Trk receptor transactivation, ERK1/2 activation, and the effects of S-PrP on PC12 neurite outgrowth. Interestingly, membrane-anchored PrP C did not appear to be necessary for activation of cell-signaling by S-PrP, in contrast with the results obtained in a previous study of tPA-initiated cell signaling through LRP1 (26).
We confirmed the purity of S-PrP by SDS-PAGE, silver-staining, and LC-MS/MS. Because cell-signaling activity was entirely eliminated when S-PrP was boiled, it is highly unlikely that endotoxin contributed to the activity of S-PrP (29). Furthermore, a single monoclonal antibody that recognizes PrP C , POM2, entirely neutralized S-PrP activity, again indicating that S-PrP is the active agent. Nevertheless, we cannot be certain that our entire preparation of recombinant S-PrP, expressed in bacteria, re-folded equivalently to PrP C that is shed from eukaryotic cells. S-PrP lacks glycosylation which is present in PrP C expressed by mammalian cells (58). However, we are encouraged that S-PrP initiated cell-signaling at concentrations as low as 30 nM, which is similar to the concentration of tPA required to trigger cell signaling via the LRP1/ NMDA-R receptor system (22,59). If refolding of recombinant S-PrP was imperfect and a fraction of the preparation was incapable of binding to the LRP1/NMDA-R receptor system, then the remainder of the S-PrP preparation probably functions at concentrations lower than 30 nM.
Although LRP1 ligands are diverse in structure and function (16,35), the structural elements required for LRP1-binding appear to be conserved. Lys residues, frequently positioned in tandem, are major elements of the LRP1-binding sites in activated α2M, RAP, and plasminogen activator inhibitor-1 (60)(61)(62). Adjacent Lys residues are present in the unstructured N-terminal region of PrP C , close to the absolute N-terminus and between the octarepeat and hydrophobic regions (see Fig. 1B). The Lys residues that are more distant from the N-terminus are adjacent to His residues implicated in copperbinding, which regulates the activity of PrP C in neuritogenesis (63). The POM2 epitope is positioned in the primary structure of PrP C so that POM2 binding may sterically inhibit interaction of LRP1 with Lys residues either at the absolute N-terminus or between the octarepeat and hydrophobic regions (30).
Our studies demonstrating an essential role of the NMDA-R in mediating cell-signaling initiated by S-PrP are intriguing given prior studies demonstrating possible interactions between PrP C and the NMDA-R. PrP C has been reported to attenuate neuronal excitotoxicity by regulating GluN2D-containing NMDA-Rs and to co-immunoprecipitate with the NMDA-R (64). PrP C and the NMDA-R are both key components of the receptor system activated by α-synuclein oligomers (14). In this pathway, the G-protein coupled receptor, mGluR5, plays an essential role by activating Fyn, which targets the NMDA-R as a substrate.
Like the NMDA-R, Trk receptors are targeted as substrates by SFKs. This interaction results in Trk receptor transactivation (65). Because inhibiting the NMDA-R with MK801 blocked Trk receptor transactivation in response to tPA (22) and S-PrP (Fig. 5A), we hypothesize that SFKs are activated downstream of the NMDA-R in response to LRP1 ligands, an apparent difference compared with the pathway reported for α-synuclein oligomers (14). Nevertheless, assessing the role of mGluR5 as a component of the NMDA-R/LRP1 cell-signaling receptor system appears justified. LRP1 localizes primarily in clathrin-coated pits and undergoes rapid endocytosis and recycling, apparently in the presence or absence of ligands (66,67). A small fraction of LRP1 transiently localizes to lipid rafts before transferring to clathrin-coated pits: however, this subpopulation of LRP1 may be very important because lipid raft-associated LRP1 is apparently responsible for initiating cell-signaling in response to LRP1 ligands (22,30). The ability of S-PrP to stimulate cell-signaling was blocked when lipid rafts were disrupted, consistent with this model. It thus appears that lipid raft microdomains play an instrumental role in the cellsignaling activity of membrane-anchored PrP C and soluble derivatives of this protein.
Not all LRP1 ligands trigger cell-signaling equivalently. The pathway described herein, in which S-PrP activates ERK1/2 downstream of the NMDA-R and Trk, is observed with quite a few LRP1 ligands (20,22). However, other LRP1 ligands, such as myelin-associated glycoprotein (MAG), force assembly of a cell-signaling receptor complex that includes p75 NTR instead of Trk receptors and activate RhoA instead of ERK1/2 (24). RAP activates neither of these pathways but instead serves as an antagonist of cell-signaling responses mediated by LRP1 in conjunction with all other coreceptors, including the NMDA-R, Trk receptors, and p75 NTR (22). The ability of diverse LRP1 ligands to force assembly of different cell-signaling receptor complexes and thereby activate different cell-signaling pathways provides a novel form of specificity to a receptor system with over 100 structurally diverse ligands (35). Engaging LRP1 and associated co-receptors represents a novel mechanism by which S-PrP may regulate cell physiology in diverse cell types including neurons and glia in the CNS and PNS.

Proteins and reagents
S-PrP (residues 23-231) was expressed in Escherichia coli BL21 cells using the T7 promoter. His-tagged S-PrP was isolated from inclusion bodies, denatured in guanidinium hydrochloride, purified by Ni 2+ -nitrilotriacetic acid agarose affinity chromatography, oxidized, and refolded out of denaturant. The N-terminal histidine tail was proteolytically dissociated with thrombin. The thrombin was then removed by ion exchange chromatography. S-PrP preparations were examined by nonreducing SDS-PAGE and visualized by silver staining (Pierce Silver Stain Kit).
The monoclonal antibodies POM1, POM2, POM3 and POM19, directed against different epitopes of mouse PrP C , were purified as previously described (30). Human EI-tPA, which carries the S478A mutation and is thus inactive as a protease, was from Molecular Innovations. EI-tPA carries a second mutation (R275E) that disallows conversion of the single-chain form into two-chain tPA. We previously demonstrated that human EI-tPA and mouse EI-tPA signal equivalently in rodent cells (59). Endotoxin-free, monomeric RAP was provided by Dr. Travis Stiles (Novoron Biosciences). NGF-β was from Invitrogen. MK801 was from Cayman Chemicals. MβCD and FM were from Sigma-Aldrich. PP2 and DXM were from Abcam. K252a was from Cell Signaling Technology.

Cell culture
Rat PC12 cells were from the ATCC (CRL-1721) and subjected to quality control tests by the ATCC. PC12 cells were cultured in Dulbecco's modified Eagle's medium (DMEM, high glucose; Gibco) containing 10% heat-inactivated fetal bovine serum (FBS; Gibco), 5% heat-inactivated horse serum (HyClone), penicillin (100 units/ml) and streptomycin (1 mg/ml), in plates that were pre-coated with 0.01 mg/ml type IV collagen (Sigma-Aldrich). Cells were passaged no more than eight times. Mouse N2a neuroblastoma cells were a generous gift from Dr. Katerina Akassoglou (Gladstone Institute of Neurological Disease, University of California, San Francisco, CA). N2a cells were cultured in DMEM containing 10% FBS, penicillin, and streptomycin.
SCs were isolated from sciatic nerves of 1-day-old Sprague Dawley rats. The SCs were enriched and separated from fibroblasts using fibronectinspecific antibody and rabbit complement cytolysis. Final preparations consisted of 98% SCs, as determined by immunofluorescence microscopy by guest on August 12, 2020 http://www.jbc.org/ Downloaded from for S100β, which is a specific SC marker. Primary SC cultures were maintained in plates coated with 1.0 µg/ml poly-D-lysine in low-glucose DMEM containing 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, 21 μg/ml bovine pituitary extract (Sigma-Aldrich) and 4 μM forskolin (Cell Signaling Technology), and were passaged no more than six times before conducting experiments. Our protocol for isolating SCs from animals was approved by the University of California San Diego Institutional Animal Care and Use Committee.

Analysis of S-PrP by LC-MS/MS
S-PrP (10 µg) was reduced and denatured and subjected to SDS-PAGE. A single band was observed. The band was excised, reduced, alkylated, and trypsin-digested using the In-Gel Tryptic Digestion Kit (Thermo Fisher Scientific). Digested peptides were passed through C18 spin tips (Thermo Fisher Scientific) and eluted in 80% acetonitrile and 0.1% formic acid. Samples were then vacuum-dried, equilibrated in 1% acetonitrile and 0.1% formic acid, bomb-packed into 70 µm C18 infused capillaries, and eluted in a positive ion nanospray with a 1-90% acetonitrile gradient using an Agilent 1200 series liquid chromatography injection system. Peptides were detected using an LTQ OrbiTrap XL mass spectrometer (Thermo Fisher Scientific) using Xcalibur 2.1 (Thermo Fisher Scientific). For assignment, raw files were searched against the Mus musculus proteome (UniProt Taxonomy ID#10090), containing 21986 entries, using Proteome Discoverer Software 2.0 with SEQUEST HT and MS Amanda search engines (Thermo Fisher Scientific). Our search parameters identified fixed modifications, including cysteine carbamidomethylation, variable methionine oxidation, lysine carbamylation, and N-terminal acetylation and oxidation. The maximum number of missed cleavages permitted was two. Mass tolerance for precursor ions was set to 50 ppm and for fragment ions, 0.6 Da. Peptides were not considered if they had an Xcorr threshold ≤1%. A strict peptide false positive rate of 5% was used to accept proteins based on spectral match. Protein samples were analysed in technical duplicates.

Gene silencing
Rat specific ON-TARGETplus SMARTpool siRNA, which targets LRP1, the GluN1 subunit of the NMDA-R (GRIN1), NCAM1, and membraneanchored PrP C (PRNP) were from Horizon Discovery, as was NTC siRNA. PC12 cells (2×10 6 ) were transfected with siRNA by electroporation using the Cell Line Nucleofector Kit V (Lonza). Gene silencing was determined at the mRNA level by RT-qPCR or at the protein level by immunoblot analysis. Experiments were performed 36-48 h after transfection.

RT-qPCR
RNA was isolated using the NucleoSpin RNA kit (Macherey-Nagel) and reverse-transcribed using the iScript cDNA synthesis kit (Bio-Rad). qPCR was performed using TaqMan gene expression products (Thermo Fisher Scientific). The relative change in mRNA expression was calculated using the 2 ΔΔCt method with GAPDH mRNA as an internal normalizer.

Neurite outgrowth
PC12 cells were plated at 1×10 5 cells/well and maintained in serum-containing medium for 24 h. The medium was then replaced with SFM supplemented with S-PrP (40 nM), RAP (150 nM), MK801 (1.0 μM), PP2 (1.0 μM), K252a (10 nM), NGF-β (50 ng/ml), or vehicle, as indicated, for 48 h. The cells were imaged by phase contrast microscopy, using a Leica DMi8 microscope (Leica Microsystems) equipped with a Leica DFC3000 G digital camera and Leica Application Suite X (LAS X) software. Neurite length was determined in 50 cells per replicate using ImageJ software (NIH). Results were subjected to statistical analysis using GraphPad Prism.

SC migration
SC migration was studied using 6.5 mm Transwell chambers with 8 µm pores (Corning), as previously described (25,42). The bottom surface of each membrane was coated with 10 µg/ml fibronectin. SCs in Sato medium (68) supplemented with 1 mg/ml BSA were pre-treated with S-PrP (40 nM), in presence or absence of RAP (150 nM) or MK801 (1 μM), for 10 min at 37°C. Cells (10 4 ) were then transferred to upper Transwell chambers. The bottom chamber contained Sato Medium with 1 mg/ml BSA, 10% FBS, and the same proteins that were added to the top chamber. Cell migration was allowed to occur for 4 h at 37°C in 5% CO2. The upper surface of each membrane was cleaned with a cotton swab. The membranes were then stained with PROTOCOL Hema 3. The number of cells on the bottom surface of each membrane was quantified using ImageJ. Four fields were examined per filter. Each condition was studied in triplicate.

Data availability
The        Immunoblot analysis was performed to detect phosphorylated ERK1/2 (p-ERK) and total ERK1/2 (T-ERK). The blots shown are representative of at least three independent studies. B. Representative images of Schwann cells that migrated through Transwell membranes to the lower surfaces. The cells were treated with S-PrP (40 nM), RAP (150 nM), MK801 (1.0 μM) or vehicle, as indicated. All reagents were added to both Transwell chambers. Migration was allowed to proceed for 4 h at 37°C (scale bar, 10 µm). C. Quantification of cell migration results. Data are expressed as the mean ± SEM (n=3, one-way ANOVA with Tukey's post hoc; ****p<0.0001 compared with the vehicle control).