The lipid phosphatase-like protein PLPPR1 associates with RhoGDI1 to modulate RhoA activation in response to axon growth inhibitory molecules

Abstract

Phospholipid Phosphatase-Related Protein Type 1 (PLPPR1) is a member of a family of lipid phosphatase related proteins, integral membrane proteins characterized by six transmembrane domains. This family of proteins is enriched in the brain and recent data indicate potential pleiotropic functions in several different contexts. An inherent ability of this family of proteins is to induce morphological changes, and we have previously reported that members of this family interact with each other and may function co-operatively. However, the function of PLPPR1 is not yet understood. Here we show that expression of PLPPR1 reduces the inhibition of neurite outgrowth of cultured mouse hippocampal neurons by chondroitin sulfate proteoglycans and the retraction of neurites of Neuro-2a cells by lysophosphatidic acid (LPA). Further, we show that PLPPR1 reduces the activation of Ras homolog family member A (RhoA) by LPA in Neuro-2a cells, and that this due to an association of PLPPR1with the Rho-specific guanine nucleotide dissociation inhibitor (RhoGDI1). These results establish a novel signaling pathway for the PLPPR1 protein.

1. INTRODUCTION

The Phospholipid Lipid Phosphate Phosphatase-Related (PLPRR) family of proteins were originally named “Plasticity-Related Genes” because PLPPR4, the first member to be discovered, was upregulated in sprouting axons following hippocampal deafferentation (Brauer et al. 2003). A total of five members, PLPPR1–5, have been identified to date. These proteins are characterized by six-transmembrane domains and share high sequence homology, with the exception of their C-terminal domain which is unique to each protein. They are highly enriched in the brain and expression is differentially regulated during development (Wang & Molnar 2005) and decreases with age in humans (Wallen et al. 2018).

We and others have previously shown that exogenous expression of PLPPR1 induced actin-rich membrane protrusions in many different cell types in culture (Sigal et al. 2007; Velmans et al. 2013; Yu et al. 2015; Broggini et al. 2016). Overexpression of PLPPR5, the closest relative of PLPPR1 in this family, produced a similar phenotype (Broggini et al. 2010), and we have shown that these two family members interact (Yu et al. 2015). In vivo, studies have shown that expression of PLPPR1 mRNA correlates with sprouting corticospinal axons after injury (Fink et al. 2017) and neuronal remodeling in the hippocampus after kainic acid treatment (Savaskan et al. 2004). Furthermore, dysregulated PLPPR1 expression correlated with perturbed neurogenesis and neuronal migration (Khalaf-Nazzal et al. 2017; Pfurr et al. 2017) and was provoked by psychological stress (Lyons et al. 2010). A recent GWAS study implicated mutations in PLPPR1 with earlier diagnosis of Parkinson’s Disease (Wallen et al. 2018), which often begins as diffuse axonal injury. Altogether, these studies support a role for PLPPR1 in neural development and axon growth and maintenance. However, the exact mechanisms by which the PLPPR1 protein mediates these changes are still unknown.

We independently identified PLPPR1 in a phosphoproteomic screen of proteins that responded to chondroitin sulfate proteoglycans (CSPGs), a potent inhibitor of axon growth (Yu et al. 2013). Many alterations of cytoskeletal dynamics, including those by CSPGs, are through the modulation of the Rho family of small GTPases, known to orchestrate actin cytoskeleton remodeling to influence cell migration, cell adhesion, morphology and neurite growth (Lawson & Burridge 2014; Ridley & Hall 1992; Fujita & Yamashita 2014). Rho GTPases act as molecular switches by cycling between their active membrane-associated GTP-bound form and inactive soluble cytosolic GDP-bound form. This GTPase cycling is regulated in part by the Rho-specific guanine nucleotide dissociation inhibitors (RhoGDIs), by solubilizing and maintaining the Rho GTPases in their inactive GDP-bound state (Boulter et al. 2010; Garcia-Mata et al. 2011; Golding et al. 2019). One study has shown that PLPPR1 modulates the RhoA-ROCK-PIP5K signaling pathway and impedes RhoA-mediated axon collapse in neurons (Broggini et al. 2016). We have recently demonstrated that PLPPR1 influences cell adhesion by modulating Rac1 activity (Tilve et al. 2020).

CSPGs and lysophosphatidic acid (LPA) are agents which reduce neurite outgrowth and cause neurite retraction through the activation of Rho kinase (Monnier et al. 2003; Jalink et al. 1993). We now show that PLPPR1 modifies the biological response to these agents by reducing Rho activation. We then searched the data from our previous proteomic study on PLPPR1-interacting proteins (Yu et al. 2015) for any that might modulate Rho and identified RhoGDI1 as one and confirmed this association. We propose a novel mechanism by which PLPPR1 modulates RhoA and Rac1 activation through its association with RhoGDI1 and alters cytoskeletal dynamics by decreasing the phosphorylation of downstream ROCK targets.

2. MATERIALS AND METHODS

This study has not been pre-registered.

2.1. DNA, Cell culture and transfection

All animal procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the National Heart, Lung, and Blood Institute, NIH, Protocol H-0076. Mice were chosen for the source of primary neurons due to the ability to compare our results with the extensive literature. A total of 6 timed pregnant C57BL/6J mice, ~ 20 g and at 6 – 8 weeks of age, were purchased from Charles River Laboratories (RRID:SCR_003792) and single-housed in ventilated cages with cage-level filtration for rodents in a pathogen free facility with free access to food and water under a standard 12 h light/dark cycle. To obtain embryonic pups, a pregnant mouse was anesthetized by CO₂ and subsequently killed by cervical dislocation. One pregnant mouse yielding 6 – 8 pups was used for each experiment.

The expression plasmids for GFP-tagged PLPPR1 and its mutant were made in-house and previously described (Yu et al., 2015) and will be shared upon request.

Primary hippocampal neurons were isolated from embryonic day 18 – 20 C57BL/6J pups euthanized by decapitation (Seibenhener & Wooten 2012; Kaech & Banker 2006). Hippocampal neurons were pooled from 6 – 8 pups and transfected with 1 μg of either EGFP, EGFP-PLPPR1 or EGFP-PLPPR1ΔC43 plasmid DNA per well using the Amaxa Nucleofector™ with the 96-well shuttle (Lonza, Walkersville, MD, cat # AAM-1001S), with the V4XP-3032 reagent kit, program CA-138. Neurons were plated onto 12 mm glass coverslips coated with either 10 μg/ml poly-l-lysine (PLL) (Sigma Aldrich, St. Louis, MO) or 2.5 μg soluble CSPG mix containing neurocan, phosphacan, versican, and aggrecan isolated from embryonic chicken brain (Millipore Sigma, Billerica, MA, cat # CC117) in 24-well culture plates (Jin et al. 2018; See et al. 2010). Neurons were grown in Neurobasal™ Medium (ThermoFisher cat # 21103049) supplemented with penicillin/streptomycin (ThermoFisher cat #15070063), glutamax (ThermoFisher cat # 35050061), B27 (ThermoFisher cat #17504044) and 2 mM potassium chloride (KCl, Sigma Aldrich, St. Louis, MO cat # P5405) for 48 h.

Neuro-2a cells (ATCC, Manassas, VA, USA, cat #:CCL-131; RRID:CVCL_0470) were maintained in Dulbecco’s Modified Eagle Medium (DMEM, ThermoFisher cat #11995065) supplemented with penicillin/streptomycin antibiotics (ThermoFisher cat # 15070063) and 10% fetal bovine serum (FBS, R&D Systems, Minneapolis, MN, cat # S12450). Neuro-2a cells were used for this study because this cell line has been validated in studies investigating neurite outgrowth and cytoskeletal dynamics (Roisin & Barbin 1997; Dehmelt et al. 2006) and respond to LPA (Sun et al. 2011) and CSPGs (Ohtake et al. 2016). The Neuro-2a cell line is not listed as commonly misidentified cell lines by the International Cell Line Authentication Committee (ICLAC; http://iclac.org/databases/crosscontaminations/) and were routinely authenticated through a STR profile analysis (ATCC). Experiments were performed within the first 15 passages after receipt from ATCC, and the morphology of cells in each passage was assessed to ensure that there were no changes. Twenty-four hours after seeding, cells were transfected at 70% confluency with 1 μg (24-well plate) or 2.5 μg (6-well plate) of either EGFP or EGFP-PLPPR1 or EGFP-PLPPR1ΔC43, using Avalanche® Omni transfection reagent (EZ Biosystems, College Park, MD cat # EZT-OMNI-1).

All cell culture plates (Corning, Corning NY), 6-well (cat # 3506), 12-well (cat # 3512), 24-well (cat # 3527), MatTek dishes (35mm petri dish with 14 mm microwells, No. 1.5, MatTek Corporation, Ashland, MA) and coverslips (No. 1.5, 12 mm microscope cover glasses, Deckgläser cat # 41001112) were coated with 10 μg/ml PLL, unless otherwise indicated, and incubated overnight at 4˚C. Excess PLL was rinsed off by washing 3 times with deionized culture grade water.

2.2. Immunocytochemistry

Transfected Neuro-2a cells were fixed in 4% PFA 48 h after plating and 24 h after transfection. Cells were briefly rinsed in PBS and incubated in blocking buffer (0.3% Triton X-100 in PBS (T-PBS), supplemented with 10% normal goat serum, NGS) for 1 h and then for 1 h with primary antibody diluted in 0.3% T-PBS, supplemented with 2.5% NGS. Cells were washed in PBS and then incubated for 1 h with secondary antibody diluted in 0.3% T-PBS, supplemented with 2.5% NGS. All antibodies used in this study, along with dilutions and source catalog number and RRID numbers are included in Table S1. Cells were washed in PBS and then mounted with DAPI diluted in Fluoromount (Sigma Aldrich cat # F4680). All phalloidin (ThermoFisher cat # P1951) staining was performed as previously described (Yu et al., 2015).

Neurons were immunostained with anti-βIII tubulin (Sigma Aldrich) and imaged using a Nikon Eclipse Microscope with a 60×/1.4 NA oil immersion objective lens. Neurite outgrowth analysis was performed using the NeuronJ plugin of ImageJ by an unblinded observer.

2.3. Live cell imaging

All live cell imaging was conducted on a microscope equipped with a heated stage at 37˚C and 5% CO2 on a Nikon A1R microscope (Nikon Instruments Inc., Melville, NY). Confocal images were acquired on a Zeiss 780 LSM confocal microscope (Carl Zeiss, Thornwood, NY).

Neuro-2a cells transfected with either EGFP or EGFP-PLPPR1 were serum starved for 16 h after serial washing with warmed DMEM by incubating cells in DMEM without FBS and then treated with fatty-acid free bovine serum albumin (FAFBSA, Sigma Aldrich cat # 126609), or lysophosphatidic acid (LPA, Cayman Chemical, Ann Arbor, MI cat # 10010093) diluted in FAFBSA. Time lapse images were acquired on a Nikon A1R microscope with a 20×/0.75 dry objective lens with no time delay between frames for 20 min. Cell response was calculated as percentage of the total transfected cell count that retracted their neurites or underwent cell rounding as assessed by an unblinded observer.

2.4. Immunoblotting assay

Neuro-2a cells seeded at a density of 4 × 10⁵ cells/well in a 6-well culture dish were transfected with either EGFP or EGFP-PLPPR1. Cells were either serum-starved for 16 – 18 h or maintained in DMEM supplemented with 10% FBS. Serum starved cells were then treated with either FAFBSA, 10% serum or increasing concentrations of LPA: 4, 8 and 16 μM for 2 min in order to establish a dose-response for LPA in Neuro-2a cells and also to capture the effect of LPA on different signaling pathways. The concentrations of LPA were based on previous experiments in Neuro-2a cells (Sun et al. 2011; Yuasa et al. 2012) and are consistent with the level of LPA in brain following injury (Yung et al. 2015). The treatment time of 2 min was established in a preliminary time-course experiment that demonstrated complete cellular detachment at the highest concentration of LPA after this time. Media was gently but completely removed from all cells and rinsed once with warmed PBS. Cell lysates were prepared in 2 × SDS cell lysis buffer and clarified by centrifugation at 12,000 rpm for 20 min. Protein quantitation was performed using Pierce™ Rapid Gold BCA Assay Reagent (cat # A53225). Equal concentration of proteins (20 μg/lane) were separated on a 4 – 20% gradient gel (Invitrogen Novex™, cat # EC6026BOX) and transferred onto a polyvinylidene difluoridemembrane (Immobilon-P, Millipore Sigma, cat # IPVH304F0). Membranes were blocked for 1 h in 5% nonfat dried milk in either 0.1% Tween20 in PBS (PBST) or 0.1% Tween20 in 10 mM Tris–HCl pH 7.5 (TBST) or blocked in 5% BSA in TBST. Membranes were incubated with primary antibodies overnight at 4˚C, washed in either PBST or TBST and subsequently incubated with secondary antibody for 30 min. Antigen–antibody complexes were detected with KPL LumiGlo® peroxidase chemiluminescent substrate (Seracare, Milford, MA cat # 5430–0051).

2.5. Co-immunoprecipitation assay

Neuro-2a cells were seeded at a density of 4 × 10⁵ cells/well in a 6-well culture dish and transfected with either EGFP, EGFP-PLPPR1 or EGFP`-PLPPR1ΔC43. Cells were harvested, washed twice in ice-cold PBS and then resuspended in ice-cold cell lysis buffer (10 mM Tris pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1% NP-40 supplemented with protease inhibitor cocktail set V EDTA free (Millipore-Sigma cat # 539137) and phosphatase inhibitors (Thermo Fisher cat # 78428). Cell lysates were cleared at 15,100 ×g for 20 min at 4˚C and supernatant incubated with GFP-Trap MA beads as instructed in ChromoTek GFP-Trap^® MA immunoprecipitation kit (ChromoTek Inc., Hauppauge, NY, cat # gtma-20). The protein – bead complex was subsequently washed twice in dilution buffer (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40 supplemented with protease and phosphatase inhibitors) and then once in wash buffer (10 mM Tris-HCl pH 7.5, 500 mM NaCl, 0.5 mM EDTA, 0.5% NP-40 supplemented with protease and phosphatase inhibitors). Proteins were eluted and subjected to SDS-PAGE as described above.

To immunoprecipitate Rho-specific guanine nucleotide dissociation inhibitor (RhoGDI1), transfected cells were rinsed in warmed PBS, harvested in cell lysis buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1% NP-40 supplemented with protease and phosphatase inhibitors) and precleared with rabbit IgG serum and Protein G agarose beads (Seracare cat # 5720–0002). Samples were subsequently incubated with primary antibody overnight at 4˚C and then incubated with Protein G agarose beads for 1 h at 4˚C. Samples were washed twice in dilution buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40 supplemented with protease and phosphatase inhibitors) and once in wash buffer (20 mM Tris pH 7.5, 500 mM NaCl, 0.5 mM EDTA, 0.5% NP-40 supplemented with protease and phosphatase inhibitors). Protein was eluted and subjected to SDS-PAGE as described above.

2.6. RhoA pull-down assay

Serum starved Neuro-2a cells transfected with either EGFP or EGFP-PLPPR1 were treated with either FAFBSA, 10% serum or 16 μM LPA. Briefly, cells were rinsed with warmed PBS and cell lysates prepared in cell lysis buffer as instructed and provided by the RhoA or Rac1 pull-down instruction manual and kit (Cytoskeleton, Inc. Denver, CO, cat #s BK-036 and BK-035, respectively). Protein quantitation was performed by the BCA method and 600 μg of protein was incubated with 50 μg Rhotekin-RBD beads for RhoA pull-down or with 10 μg of PAK-PBD beads for Rac1 pull-down at 4˚C for 1 h. Protein-bead complexes were washed in wash buffer provided in the pull-down kit and protein was eluted in 2× SDS sample buffer and SDS-PAGE performed as described above.

2.7. Real-time PCR

RNA was prepared from P1 brain of C57BL/6J mice and cultured Neuro-2a cells using TRIzol™ (Thermo Fisher, cat # 15596026). Reverse transcriptase reaction was performed with ReverTra Ace® qPCR RT Master Mix with gDNA Remover (TOYOBO, Japan, cat # FSQ-301). Real-time PCR was performed in a Mic qPCR (Bio Molecular Systems, Australia) with SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad Laboratories, Inc. Hercules, CA, cat # 1725271). Primers used for real-time PCR were listed in Table S2. To have absolute quantitation, DNA standards for PLPPR1–5 were synthesized by PCR with individual primer pairs and cloned into pCR2.1 (Addgene, Watertown, MA). The serially diluted standards of known concentrations were used to generate a standard curve. All experiments were done in triplicate.

2.8. Proteomics data analysis

Neuro-2a cells growing on 10-cm dishes were transfected with EGFP, EGFP-PLPPR1 or EGFP-PLPPR1ΔC43 were underwent affinity purification by GFP-Trap agarose beads (Chromotek, Martinsreid, Germany, cat # gtak-20), and eluted proteins were sepasrated by SDS-PAGE and subjected to in-gel digestion (Yu et al. 2015).

LC-MS/MS was performed as described previously (Yu et al. 2015), using an Eksigent nanoLC-Ultra 2D system (Dublin, CA) coupled to an Orbitrap Elite mass spectrometer (Thermo Scientific, San Jose, CA). The peptide sample was first loaded onto a Zorbax 300SB-C18 trap column (Agilent, Palo Alto, CA), and then separated on a reversed-phase BetaBasic C18 PicoFrit® analytical column (New Objective, Woburn, MA) using a linear gradient of 5–35% B (buffer A, 0.1% formic acid in water; buffer B, 0.1% formic acid in acetonitrile). Eluted peptides were sprayed into the Orbitrap Elite equipped with a nano-spray ionization source. Survey mass spectrometry spectra were acquired in the Orbitrap, and data-dependent MS/MS scans were performed in the linear ion trap with dynamic exclusion.

Scaffold (version Scaffold_4.11.0, Proteome Software Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Inclusion criteria: peptide identifications were accepted if they could be established at greater than 78.0% probability to achieve an FDR less than 5.0% by the Scaffold Local FDR algorithm. Protein identifications were accepted if they could be established at greater than 7.0% probability to achieve an FDR less than 1.0% and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (Nesvizhskii et al., 2003). Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters.

2.9. Statistical analysis

No sample size calculations were performed in advance. Experiments were repeated at least three times to ensure statistical reliability to detect significant changes based on previous experiments. No exclusion criteria were pre-determined. All analyses were performed using GraphPad Prism software version 7 (GraphPad Software, La Jolla, CA). Data was analyzed by unpaired Student’s t tests, one-way ANOVA or two-way ANOVA with Welch’s correction or Tukey’s post-hoc multiple comparison test to determine significance. Normality of the distribution of the data was tested with Kolmogorov–Smirnov normality tests using the column statistics function of GraphPad Software. All tests were two-tailed with significance indicated as follows: *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. No test for outliers were conducted. Unless otherwise specified, values represent the means ± SEM. All experiments were repeated independently three times.

A flow-chart has been included to illustrate experimental design for this study (Figure 1).

Figure. 1. Experimental design and timeline of the study.

(A) Schematic diagram of in vitro neurite outgrowth analysis for Fig. 2a data. (B) Schematic diagram of neurite retraction analysis for Fig. 2b data. (C) Schematic diagram of procedure for assessing Rho activation, phosphorylation of Rho downstream targets and interaction of PLPPR1 with RhoGDI for Fig. 3–5 data.

3. RESULTS

3.1. PLPPR1 modulates inhibition of neurite outgrowth and neurite retraction

CSPGs are significantly upregulated after CNS injury (Mckeon et al. 1991; Lemons et al. 1999) and are potent negative inhibitors of axonal growth, both in vivo (Davies et al. 1997; Moon et al. 2001) and in vitro (Snow et al. 1990; Wang et al. 2008). PLPPR1 phosphorylation is altered by CSPG treatment (Yu et al. 2013) and PLPPR1 improves functional recovery after spinal cord injury (Fink et al. 2017). We therefore tested the effects of PLPPR1 on CSPG-mediated inhibition of neurite outgrowth in culture. Hippocampal neurons were transfected with either EGFP or EGFP-PLPPR1 and seeded onto glass coverslips coated with either poly-L-lysine (PLL) or CSPGs. After 48 h, cultures were fixed and stained for βIII-tubulin. Neurite outgrowth was analyzed by measuring both total neurite length (mean length of all neurites emanating from the cell body) and the length of the longest neurite (Figure 2A). CSPGs significantly reduced the total neurite length of EGFP-transfected neurons compared to PLL; the mean length on PLL was 86.1±6.5 μm while on CSPGs, the mean length was 53.8±4.4 μm, (Figure 2B). In contrast, there was no significant difference in total neurite length of PLPPR1-transfected neurons between PLL and CSPGs; the mean length on PLL was 90.4±6.1 μm whereas that on CSPGs was 94.8±6.0 μm (Figure 2B). We also observed a decrease in the length of the longest neurite of EGFP-transfected neurons exposed to CSPGs, while this inhibition was not observed in neurons overexpressing PLPPR1 (Figure 2C). Thus, PLPPR1 attenuates the inhibitory activity of CSPGs on axonal growth.

Figure 2. PLPPR1 attenuates CSPG and LPA induced neurite retraction.

(A) Representative images showing EGFP (left) or EGFP-PLPPR1 (right) transfected hippocampal neurons cultured on either PLL (top) or PLL and CSPGs (bottom). Scale bar, 50 μm. Representative analysis of (B) total neurite length and (C) longest neurite were assessed at 2 DIV using NeuronJ. EGFP/PLL, n = 112; EGFP-PLPPR1/PLL, n = 115; EGFP/CSPGs, n = 113; EGFP-PLPPR1, n = 134 neurons. (D) Time lapse imaging of Neuro-2a cells transfected with either EGFP (top) or EGFP-PLPPR1 (bottom) and exposed to 16 μM LPA. Black arrows indicate retracting neurites. Black asterisks indicate nonretracting neurites. (E) Cell response was calculated as the percentage of transfected cells that retracted their neurites at 20 min. Data represent the mean retracted for three independent cell cultures. p-values were calculated using Two-way ANOVA with Tukey’s posthoc analysis, **p<0.01, ***p<0.001, ****p<0.0001. *p<0.05, ****p<0.0001. Scale bar, 50 μm. All experiments were repeated three times.

The inhibitory activity of CSPGs has been attributed to the activation of RhoA (Monnier et al. 2003; Jain et al. 2004). Because LPA induces neurite retraction and cell rounding in several different cell types through RhoA (Jalink et al. 1993; Tigyi et al. 1996; Kranenburg et al. 1999; Sun et al. 2011), we investigated the effects of PLPPR1 on LPA-induced cell responses. We used live-cell imaging to evaluate neurite retraction after LPA treatment of the Neuro-2a neuroblastoma cell line, which does not express detectable levels of PLPPR1 (Figure S1). LPA induced significantly more events of neurite retraction and cell rounding compared to fatty acid-free bovine serum albumin (FAFBSA) (Figure S2; Movie S1). In contrast, fewer cells expressing PLPPR1 retracted upon LPA treatment. Upon LPA treatment, 59% of cells expressing EGFP underwent cell rounding and neurite retraction, while 42% of cells expressing EGFP-PLPPR1 responded (Figure 2E; Movie S2). This result demonstrates the inhibitory effect of PLPPR1 on LPA-induced cellular responses.

The C-terminal domain PLPPR1 has been deemed essential to produce protrusions (Sigal et al. 2007) as well as to increase neurite outgrowth (Broggini et al. 2016). We therefore investigated if the C-terminal was required to overcome these actions of CSPGs and LPA. Hippocampal neurons were transfected with either EGFP, PLPPR1, or PLPPR1 with a deletion of the 43 aa comprising the C-terminal domain (PLPPR1ΔC43). Neurite outgrowth on PLL was not affected by PLPPR1, though neurites of the neurons expressing PLPPR1ΔC43 were shorter. However, both PLPPR1 and PLPPR1ΔC43 were able to overcome the inhibition of CSPGs (Figure S3A–C). Similarly, the percentage of Neuro-2a cells that retracted in response to LPA was reduced in cells expressing either the full length or the truncated PLPPR1 (Figure S3D, E). Thus, while the C-terminal has been deemed essential for promoting protrusion formation and neurite outgrowth, it is dispensable for the response to inhibitory molecules.

3.2. PLPPR1 modulates RhoA activation

Because the inhibitory actions of both CSPGs and LPA are dependent upon RhoA activity, we next tested the hypothesis that PLPPR1 would reduce RhoA activation. Neuro-2a cells cultured under serum-free conditions were treated with FAFBSA, serum or LPA and active RhoA was enriched using rhotekin-RBD beads. Western blot analyses (Figure 3A) showed an increase in active RhoA after serum and LPA treatment in control cells. In contrast, no significant increase in active RhoA was observed in cells expressing PLPPR1 (Figure 3B). Thus, we conclude that PLPPR1 modulates RhoA activation.

Figure 3. PLPPR1 attenuates RhoA activation.

(A) RhoA-GTP was pulled down using rhotekin-RBD beads from cell lysates prepared from transfected Neuro-2a cells treated with either FAFBSA, 10% serum or 16 μM LPA for 2 min. Immunoblots were probed for RhoA. ß-actin was used as loading control for total RhoA. Membranes were stripped and reprobed for GFP. (B) Densitometric analysis was performed using Image Studio Lite. Data represent values from 3 independent immunoprecipitations. p-values were calculated using Two-way ANOVA with Tukey’s posthoc analysis, *p<0.05, ***p<0.001.

3.3. PLPPR1 reduces the phosphorylation of ROCK targets upon RhoA activation

Rho-associated kinase (Rho kinase/ROCK) is a major RhoA effector and is implicated in the regulation of actinomyosin dynamics and inhibition of neurite outgrowth (Amano et al. 1997; Lingor et al. 2007; Monnier et al. 2003). ROCK mediates its effect by phosphorylating several substrates including cytoskeletal proteins, myosin light chains (MLC), myosin phosphatase 1 (MYPT1) and ezrin, radixin, and moesin (the ERM proteins) (Yoneda et al. 2005; Kaneko-Kawano et al. 2012; Sutherland et al. 2016; Matsui et al. 1998). We examined the modulatory effect of PLPPR1 on the RhoA-ROCK pathway by assessing the phosphorylation levels of several ROCK substrates in response to LPA (Figure 4). First, we measured the phosphorylation levels of MLC. Serum starved Neuro-2a cells expressing EGFP or EGFP-PLPPR1, were treated with either FAFBSA or LPA for 2 min and immunostained for phosphorylated MLC (pMLC). Confocal imaging showed basal levels of pMLC with a diffuse cytoplasmic distribution after FAFBSA treatment. Upon exposure to LPA, phosphorylation of MLC was increased and localized to the plasma membrane in EGFP-transfected cells. On the contrary, a negligible change in pMLC was observed in cells expressing PLPPR1, and the diffuse cytosolic distribution was retained with minimal localization to the plasma membrane (Figure 4A). Quantitative image analyses based on the mean fluorescent intensity showed an increase in pMLC in EGFP-transfected cells upon treatment with LPA (Figure 4B); mean fluorescence with FAFBSA was 918.3±72.9 (n = 61) and 1283.0±49.0 (n = 54) with LPA. Conversely, PLPPR1 expressing cells demonstrated little increase in pMLC upon LPA treatment; mean fluorescence with FAFBSA was 830.1±90.0 (n = 45) and with LPA, 879.9.0±71.3 (n = 41).

Figure 4. PLPPR1 decreases LPA induced translocation and phosphorylation of MLC, MYPT1 and ERM proteins.

(A) Representative images of serum starved Neuro-2a cells expressing either EGFP or EGFP-PLPPR1 that were exposed to FAFBSA or 16 μM LPA for 2 min and immunostained for pMLC. Arrows indicate the cell used for the profile plot in B. Scale bar, 50 μm. (B) Representative plot of fluorescence intensity of pMLC in cells after FAFBSA or LPA treatment using ImageJ. Representative immunoblots of cells transfected with either EGFP or EGFP-PLPPR1 and treated with FAFBSA, serum or LPA were probed for (C) pMLC and MLC (E) pMYPT1 and MYPT1 (G) pERM and ERM. (D, F, H) Densitometry analysis of phosphorylated protein versus total protein was performed using Image Studio Lite. Data represent values from 3 independent cell culture experiments. p-values were calculated using Two-way ANOVA with Tukey’s posthoc analysis, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 respectively. All cell culture experiments were repeated three times.

Changes in phosphorylation of ROCK substrates due to the presence of PLPPR1 in Neuro-2a cells were further examined by immunoblot analysis. Since myosin light chain phosphatase 1 (MYPT1) regulates MLC phosphorylation levels by targeting MLC S19, we similarly assessed phosphorylation of MYPT1 at threonine 853 (T853), which, when phosphorylated, decreases its phosphatase activity on MLC S19. Additionally, in a positive feedback manner, RhoA regulates the activation of the ERM proteins, that serve as crosslinkers for cytosolic proteins to the actin cytoskeleton (Niggli & Rossy 2008; Ivetic & Ridley 2004). We found that phosphorylation of MLC at serine19 (S19), myosin light chain phosphatase 1 (MYPT1) at threonine 853 (T853), and the ERM proteins were significantly increased by LPA in a dose-dependent manner in EGFP-transfected cells. However, cells expressing PLPPR1 displayed a minimal increase in phosphorylation of these substrates and dose-dependent phosphorylation was largely eliminated (Figure 4C–H). We also confirmed the alterations in pERM proteins by immunocytochemistry. Similarly to pMLC, the phosphorylation of pERM proteins was significantly increased in control LPA-treated cells, while LPA-induced phosphorylation was negligible in PLPPR1-transfected cells, (Figure S4A, B). Altogether, these results demonstrate that the presence of PLPPR1 attenuates the phosphorylation of MLC, MYPT1, and the ERM proteins after LPA stimulation.

LPA activates a variety of well-characterized signaling cascades and induces the phosphorylation of key proteins in these pathways, including GSK3ß through activation of PLCγ-PKC (Fang et al. 2002; Sun et al. 2011), phosphorylation of the serine/threonine kinase AKT through activation of the PI3K/PDK1 (Baudhuin et al. 2002; Weiner & Chun 1999), and phosphorylation of Tau through GSK3ß and PKA activation (Sayas et al. 2006). Consequently, we tested if the effects of PLPPR1 expression extend to these pathways. While LPA and serum increased phosphorylation levels of GSK3ß, Tau and AKT, we found no significant difference in the phosphorylation levels of GSK3ß at Y216 and S9, Akt at S473 and Tau at S396, between EGFP and EGFP-PLPPR1 transfected cells (Figure S5). Expression levels of neither EGFP or EGFP-PLPPR1 were altered under serum or LPA conditions compared to FAFBSA as indicated by GFP immunoblot analysis (Figure S5H).

3.4. PLPPR1 alters Rho dissociation from RhoGDI1

Based on the identification of the alteration in Rho signaling by PLPPR1, we revisited a proteomic data set of proteins that were immunoprecipitated with PLPPR1 (Yu et al. 2015) to identify proteins that interacted with Rho family GTPases. This analysis identified RhoGDI1 as a potentially interacting protein (Table S3). To validate the interaction between RhoGDI1 and PLPPR1, we expressed EGPP-PLPPR1 in Neuro-2a cells and immunoprecipitated EFGP-PLPPR1 with GFP magnetic agarose beads. We then assessed whether RhoGDI1 was among the associating proteins by immunoblotting. Figure 5A demonstrates that RhoGDI1 was pulled down with EGFP-PLPPR1. This interaction was maintained when EGFP-PLPPR1ΔC43 was used as a bait for RhoGDI1.

Figure 5. PLPPR1 associates with RhoGDI and maintains RhoGDI interaction with RhoA and Rac1 after LPA treatment.

(A) Neuro-2a cells expressing either EGFP, EGFP-PLPPR1 or EGFP-PLPPR1ΔC43 were subjected to immunoprecipitation with GFP-MATM beads and immunoblotted with anti-RhoGDI antibody. Input was probed for GFP. (B) Serum starved transfected Neuro-2a cells were treated with either FAFBSA, 10% serum or 16 μM LPA for 2 min and subjected to immunoprecipitation with RhoGDI antibody. Immunoblots of input and immunoprecipitates were probed for either RhoGDI, RhoA or Rac1. Inputs were stripped and reprobed for GFP and β-actin antibody for loading control. Densitometric analysis of RhoA (C) and Rac1 (D) was performed using Image Studio Lite. Data represent values from 3 independent immunoprecipitation experiments. p-values were calculated using Two-way ANOVA with Tukey’s posthoc analysis, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 respectively.

We then evaluated the potential role of PLPPR1 in RhoA signaling. Inactive RhoA family GTPases (bound to GDP) is tightly sequestered by RhoGDI1 in the cytosol, and their activation requires the displacement of RhoA from RhoGDI1(DerMardirossian et al. 2004; Dovas & Couchman 2005). This allows RhoA to disengage GDP, bind GTP and become activated (Garcia-Mata et al. 2011; Michaelson et al. 2001; Takahashi et al. 1997; Yamashita & Tohyama 2003). Cdc42 (Suzuki et al. 2000) and Rac1 (Boulter et al. 2010; Chuang et al. 1993) are reported to be regulated by RhoGDI1 in a similar manner. To determine the effect of PLPPR1 on RhoGDI1-RhoA and Rac1 interaction, we transfected Neuro-2a cells with either EGFP or EGFP-PLPPR1under serum-free conditions, and treated cells with FAFBSA, serum or LPA, followed by immunoprecipitation with an antibody directed against RhoGDI1. In EGFP-transfected cells treated with FAFBSA, both RhoA and Rac1 were found to be associated with RhoGDI1 (Figure 5B); this association was eliminated upon exposure to serum or LPA. However, in cells expressing PLPPR1, the dissociation of RhoA and Rac1 from RhoGDI1 was significantly inhibited in cells expressing PLPPR1 (Figure 5B–D). This suggests that exogenous PLPPR1 impedes the dissociation of RhoA and Rac1 from RhoGDI1. We did not investigate Cdc42, given that PLPPR1 has been shown to act independently of Cdc42 (Sigal et al. 2007).

4. DISCUSSION

Members of the family of PLPPR proteins are being increasingly recognized as playing roles in many nervous system functions, including neurogenesis and neuronal migration (Khalaf-Nazzal et al. 2017; Pfurr et al. 2017) as well as axonal sprouting after injury (Fink et al. 2017; Savaskan et al. 2004), and adult neurogenesis (Lyons et al. 2010). However, the exact mechanisms by which PLPPR causes these effects are not known. One common mechanism in all these functions is the control of cytoskeletal dynamics and cell adhesion. Exogenous expression of PLPPR1 induces the production of actin-rich membrane protrusions in many different cell types in culture (Sigal et al. 2007; Velmans et al. 2013; Yu et al. 2015; Broggini et al. 2016), while overexpression of PLPPR5, the closest relative of PLPPR1 in this family, produced a similar phenotype (Broggini et al. 2010). Our experiments have demonstrated that PLPPR family members interact (Yu et al. 2015), and we have recently shown that PLPPR1 can modulate cell adhesion through modulation of Rac1 activity (Tilve et al. 2020). Herein we demonstrate that PLPPR1 associates with RhoGDI1 to attenuate RhoA and Rac1 activity thereby modulating changes in actinomyosin dynamics (Figure 6).

Figure 6. PLPPR1 modulates RhoA-ROCK pathway.

Schematic diagram showing the modulatory effect of PLPPR1 on RhoA signaling. LPA and CSPG trigger the RhoA signaling cascade, followed by increased phosphorylation of ROCK effector proteins (MLC, MYPT1 and ERM) (green arrows). PLPPR1 associates with RhoGDI (black arrow), resulting in reduced RhoA activation and diminished neurite retraction and cell rounding (red lines and arrow indicate effects of PLPPR1).

The Rho GTPases are controlled by regulatory proteins such as the guanine nucleotide exchange factors that catalyze the exchange of GDP for GTP; as well as the GDIs, that sequester Rho GTPases in the cytosol, regulating their translocation to the plasma membrane and thus, activation (Dovas & Couchman 2005; Hodge & Ridley 2016; Bokoch et al. 1994). RhoGDI1 is also known to extract RhoA from the cell membrane to prevent further activation (Forget et al. 2002). We identified RhoGDI1 as a possible binding target for PLPPR1 in a proteomic screen for interacting proteins (Yu et al. 2015), and we have confirmed this interaction by co-immunoprecipitating PLPPR1 with RhoGDI1. Without stimulation, GDP-bound RhoA is tightly bound to RhoGDI1 through its N-terminal domain that binds the switch regions of the Rho GTPases and affects the GDP-GTP exchange; and by its C-terminal domain, which binds the isoprenyl moiety of Rho GTPases, regulating their cytosolic-membrane translocation (Dovas & Couchman 2005). Upon activation, this interaction is disrupted, enabling the translocation of RhoA to the plasma membrane and its subsequent activation. By using LPA and CSPGs to stimulate RhoA activation, our results suggest a novel signaling mechanism for PLPPR1, that by associating with RhoGDI1, PLPPR1 stabilizes the RhoGDI1-RhoA complex and thus, attenuates actinomyosin dynamics.

In addition to attenuating CSPG inhibition of axon growth in hippocampal neurons, we show that expression of PLPPR1 reduced LPA-induced neurite retraction. This is consistent with the results observed from the axon spreading experiment of neuroblastoma cells expressing PLPPR1, on myelin (Broggini et al. 2016). Each of these negative guidance cues are known to signal through activation of the RhoA GTPase (Alabed et al. 2006; Ohtake et al. 2016). Studies have demonstrated reducing RhoA activity with C3 exoenzyme or inhibition of the downstream effector, Rho kinase, ROCK, attenuates its inhibitory activity on neurite outgrowth and enhances axon regeneration in the injured CNS (Dergham et al. 2002; Fournier et al. 2003).

Activation of RhoA leads to increased ROCK activity on several different substrates involved in regulating actin-filament assembly and contractility. These substrates include MLC, MYPT1 and the ERM proteins (Yoneda et al. 2005; Kaneko-Kawano et al. 2012; Sutherland et al. 2016; Matsui et al. 1998). Direct phosphorylation of MLC at T18/S19 by ROCK and myosin light chain kinase (MLCK) results in actinomyosin contraction leading to neurite/axon retraction and growth cone collapse (Yoneda et al. 2005; Jalink et al. 1994). Interestingly, while MLCK activation occurs through calcium mobilization and release in a pathway distinct from RhoA-ROCK signaling, MLCK is also directly phosphorylated by ROCK, leading to increased phosphorylation levels of MLC; at the same time, phosphorylation of MLC is modulated by MYPT1 phosphatase (Vicente-Manzanares et al. 2009). Phosphorylation of MYPT1 at T853 decreases its phosphatase activity on MLC and antagonizes the effects of direct phosphorylation of MLC by ROCK and MLCK (Khromov et al. 2009; Riddick et al. 2008; Kaneko-Kawano et al. 2012). Our results show reduced phosphorylation of MLC and MYPT1 as a consequence of diminished RhoA activation in cells expressing PLPPR1. This is consistent with a previous study that showed PLPPR1 modulates PIP5K activity, a protein kinase downstream of RhoA (Broggini et al. 2016).

The PLPPR proteins are similar in their N-terminal domains, with significant divergence in their intracellular C-terminal regions (Brauer & Nitsch 2008). The C-terminal domain of PLPPR1 has been demonstrated to be important for the induction of membrane protrusions (Broggini et al. 2016). This domain is also the site of the change in phosphorylation by CSPGs (Yu et al. 2013) and has been reported to modulate Ras activity by interacting with RasGRF1 (Broggini et al. 2016). We previously showed that PLPPR1 interacts with other PLPPR proteins and this interaction is independent of the C-terminal domain (Yu et al. 2015). Here, our data show that the PLPPR1 C-terminus is also dispensable for its ability to overcome the inhibitory activity of CSPGs and LPA and is not necessary for its association with RhoGDI1. Interestingly, the C-terminal of PLPPR5, the closest family member to PLPPR1, has also been reported to be essential for the formation of protrusions (Broggini et al. 2010). Whether there is a similar interaction of PLPPR5 with RhoGDI1 remains an open question.

Neuro-2a cells are commonly used to study neuronal-associated actin and microtubule dynamics and GTP kinetics (Xiao et al. 2013; Yuasa et al. 2012; Liu et al. 2014). Among the PLPPR family, endogenous levels of PLPPR1 in Neuro-2a cells were under the detection level in a real-time RT-PCR assay. Additionally, low levels of PLPPR3 and PLPPR5 were detected in Neuro-2a cells while PLPPR2 exhibited the highest expression. While PLPPR1 and PLPPR4 were not detected in Neuro-2a cells, brain from postnatal day 1 C57BL/6J mice expressed all members of the PLPPR family. Given the lack of endogenous expression of PLPPR1, Neuro-2a cells are an ideal culture model to investigate the function of PLPPR1.

As its original name, Plasticity-Related Gene 3 protein, infers, PLPPR1 plays a role in mediating plasticity, as shown in our results demonstrating modulation of the response to CSPGs. This occurs by altering the RhoA – ROCK signaling pathway, likely via the PLPPR1 association with RhoGDI1. This novel mechanism serves as a potential explanation for the ability of PLPPR1 to overcome LPA, CSPG, and myelin-mediated inhibition of neurite outgrowth. These results, in addition to the role PLPPR1 has been shown to play in neurogenesis, neuronal migration as well as improving functional recovery after spinal cord injury, provide support to investigate and fully characterize PLPPR1 signaling mechanisms as they are potential therapeutic targets for CNS injury.

Abbreviations:

CSPG

Chondroitin Sulfate Proteoglycan

ERM

ezrin-radixin-moesin

FAFBSA

Fatty acid-free bovine serum albumin

GSK3β

Glycogen synthase kinase 3β

LPA

lysophosphatidic acid

MLC

myosin light chains

MYPT

myosin phosphatase

PLPPR

Phospholipid Lipid Phosphate Phosphatase-Related

PLL

poly-L-lysine

RhoA

Ras homolog family member A

RRID

Research Resource Identifier

RhoGDIs

Rho-specific guanine nucleotide dissociation inhibitor

ROCK

Rho-associated kinase