Lysophosphatidic acid (LPA) is a serum-borne phospholipid that exerts a pleiotropic range of effects on cells through activation of three closely related G-protein-coupled receptors termed LPA1/EDG-2, LPA2/EDG-4 and LPA3/EDG-7. Of these receptors, the LPA1 receptor is the most widely expressed. In this study, we investigated the agonist-induced endocytosis of the human LPA1 receptor, bearing an N-terminal FLAG epitope tag, in stably transfected HeLa cells. Treatment with LPA induced the rapid endocytosis of approximately 40% of surface LPA1 within 15 minutes. Internalization was both dose dependent and LPA specific since neither lysophophatidylcholine nor sphingosine-1-phosphate induced LPA1 endocytosis. Removal of agonist following 30 minutes incubation resulted in recycling of LPA1 back to the cell surface. LPA1 internalization was strongly inhibited by dominant-inhibitory mutants of both dynamin2 (K44A) and Rab5a (S34N). In addition, both dynamin2 K44A and Rab5 S34N mildly inhibited LPA1-dependent activation of serum response factor. Finally, our results also indicate that LPA1 exhibits basal, LPA-dependent internalization in the presence of serum-containing medium.
LPA is a major serum phospholipid that exhibits growth factor-like properties towards a variety of cells (Moolenaar, 1999). Some of the pleiotropic cellular effects exerted by LPA include the stimulation of cell migration (Mukai et al., 2000), tumor cell invasion (Stam et al., 1998), neurite retraction (Jalink et al., 1994), as well as growth stimulation of a variety of normal and tumorigenic cells (van Corven et al., 1993). Most of these effects are mediated by the binding of LPA to cell-surface serpentine receptors that couple to and activate heterotrimeric G proteins of the Gi, Gq and G12/13 families (Chun et al., 1999). LPA stimulation of cells via these G-protein pathways has been shown to inhibit adenylyl cyclase, induce intracellular calcium release, activate rho GTPases, stimulate transcription of serum-responsive genes, and activate the ERK1/2 mitogen-activated protein kinases (Hill et al., 1995; Ishii et al., 2000; Moolenaar et al., 1997; van Corven et al., 1993).
Molecular cloning studies have identified three mammalian receptors that belong to the endothelial differentiation gene (EDG) family of G-protein-coupled receptors (GPCRs) that are activated by LPA: LPA1/EDG-2, LPA2/EDG-4 and LPA3/EDG-7 (Chun et al., 1999). These receptors were initially termed EDG receptors since they share sequence homology with the sphingosine-1-phosphate (S1P)-specific S1P1/EDG-1 receptor (Hla et al., 2001). Heterologous expression studies have shown that all three receptors can activate Gi- and Gq-coupled signaling pathways (Ishii et al., 2000). LPA1 and LPA2, but not LPA3, can additionally stimulate G12/13-coupled pathways. Recent studies have also indicated that LPA is a potent mitogen for ovarian cancer epithelial cells and that increased LPA concentrations in the serum and ascites might serve as a useful biomarker for ovarian cancer (Fang et al., 2000; Moolenaar, 1999; Xu et al., 1998). These studies also suggest that expression of LPA2 or LPA3, which are not expressed in normal ovarian epithelial cells, is upregulated in ovarian cancer epithelial cells (Fang et al., 2000). Interestingly, LPA1 has been shown to be a negative regulator of ovarian cancer cell growth (Furui et al., 1999). Of the three known LPA receptors, LPA1 shows the widest tissue distribution. Human LPA1 is expressed in adult organs such as brain, heart, ovary, testes, colon, prostate and spleen, but is not detectably expressed in liver, thymus or lung (Contos and Chun, 2001; Hecht et al., 1996).
Agonist binding and activation of most GPCRs usually results in the rapid phosphorylation and endocytosis of the receptor (Ferguson, 2001). GPCR endocytosis serves as an entry point for targeting activated GPCRs into a variety of intracellular compartments including endosomes and lysosomes. Dephosphorylation of receptors in endosomes and subsequent recycling back to the cell surface constitutes GPCR resensitization, whereas targeting receptors to lysosomes for degradation is used for GPCR downregulation. Thus far, nothing is known about the trafficking or intracellular destinations of any LPA-coupled receptor.
To gain further insight into how cells regulate the activity of specific LPA receptors, we investigated the agonist-induced trafficking of the human LPA1 receptor in HeLa cells. Our results indicate that LPA1 is rapidly internalized into cells via dynamin2- and Rab5-dependent mechanisms in an LPA-specific and LPA dose-dependent manner. Interestingly, we find that LPA1 is internalized and recycled at a low basal level when cells are cultured in medium that contained 10% FBS, which suggested that LPA levels in serum are sufficient to induce LPA1 activation and endocytosis.
Materials and Methods
Cells, reagents and antibodies
HeLa cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS, 100 I.U./ml penicillin, and 100 mg/ml streptomycin (complete medium) at 37°C with 5% CO2. Mouse monoclonal antibodies against the FLAG epitope tag were purchased from Sigma, mouse antibodies against the early endosomal marker EEA1 were obtained from Transduction Laboratories (Lexington, KY), mouse antibodies to the human transferrin receptor B3/25 were from Roche Molecular Biochemicals (Indianapolis, IN), and mouse anti-LAMP-2 IgG (H4B4), developed by J. T. August and J. E. K. Hildreth, was obtained from the Developmental Hybridoma Bank developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences (Iowa City, IA). Alexa594- and Alexa488-conjugated goat anti-mouse and goat anti-rabbit IgG was purchased from Molecular Probes (Eugene, OR).
Lysophosphatidic acid (1-Oleoyl-2-hydroxy-sn-glycerol-3-phosphate; LPA) and D-erythro sphingosine-1-phosphate (S1P) were purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA). L-alpha-lysophosphatidylcholine (LPC), fatty acid-free BSA, and all other chemicals were purchased from Sigma. Stock solutions of LPA and LPC were prepared by dissolving in dH2O and sonication, whereas S1P was prepared by dissolving in methanol, followed by evaporation under a stream of nitrogen gas. The dried S1P was then dissolved in 4 mg/ml fatty acid-free BSA (Sigma) in dH2O. For lipid stimulation, cells were grown on glass coverslips for 16-24 hours at 37°C in complete medium and then incubated in serum-free DMEM (SF-DMEM) for an additional 16 hours at 37°C prior to incubation with the appropriate lipid in SF-DMEM.
DNA manipulations and transfections
An expression plasmid encoding the human LPA1 receptor containing an amino terminal FLAG epitope tag (Bandoh et al., 1999) was the kind gift of Junken Aoki (University of Tokyo, Japan). The FLAG epitope tag in this receptor is exposed to the extracellular environment when LPA1 is at the cell surface. To enhance cell-surface expression of LPA1, PCR was used to attach a signal leader sequence from the influenza hemaglutinin protein onto the amino terminus of FLAG-tagged LPA1 cDNA using the following primers: 5′-ATCATGAAGACCATCATCGCCCTGAGCTACATCTTCTGCCTGGTGTTCGCCGACTACAAAGACGATGACGATAAA-3′ and 5′-GATCTCAAACCACAGAGTGATC-3′. Following PCR amplification, the cDNA product was subcloned into the eukaryotic expression vector pcDNA 3.1/V5-His using a TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA) All DNA sequences were confirmed by DNA sequencing (Emory DNA Sequencing Core Facility, Atlanta, GA).
To generate stable HeLa cell transfectants, wild-type (WT) LPA1 was transfected into HeLa cells using the calcium phosphate co-precipitation method (Radhakrishna and Donaldson, 1997). At 36 hours after transfection, cells were detached and re-plated at a 1:25 dilution into complete medium containing 600 μg/ml G418 (Life Technologies, Gaithersburg, MD). Approximately two weeks later, G418-resistant clones were amplified and tested for LPA1 expression by indirect immunofluorescence microscopy. For immunolocalization studies, HeLa cells were grown on glass coverslips and transfected in six-well dishes using the calcium phosphate method. WT and mutant plasmids encoding green fluorescent protein (GFP)-Rab5 were transiently co-transfected along with plasmids for FLAG-tagged LPA1 into six-well dishes using 5 μg of Rab5 DNA. WT and mutant dynamin plasmids were co-transfected with FLAG-tagged LPA1 transfected using 10 μg of dynamin plasmid per well.
At 22 hours after transfection, the cells were rinsed with SF-DMEM and incubated in the same medium for 16-24 hours before further treatments. Cells were treated as described in the figure legends, fixed in 2% formaldehyde in PBS for 10 minutes, and rinsed with 10% FBS and 0.02% azide in PBS (PBS-serum). The cells were permeabilized by treating with ice-cold methanol for 30 seconds at –20°C, rinsing with ice-cold PBS twice and incubating in PBS-serum for 5 minutes. Fixed cells were incubated with primary antibodies diluted in PBS-serum containing 0.2% saponin for 45 minutes, and then washed (three times, 5 minutes each) with PBS-serum. The cells were then incubated in secondary antibodies diluted in PBS-serum plus 0.2% saponin for 45 minutes, washed with PBS-serum (three times, 5 minutes each) and once with PBS, and mounted on glass slides. Samples were observed using an Olympus BX40 epifluorescence microscope equipped with a 60× Plan pro lens and photomicrographs were prepared using a Spot RT monochrome `C' digital camera (Diagnostic Instruments, Sterling Hts, MI). The fluorescence images were photographed using the same exposure time and processed identically using Adobe Photoshop 5.0.
Quantitation of LPA1 internalization
HeLa cells expressing FLAG-tagged LPA1 were treated as described in the figure legends and fixed as described above. The fixed cells were labeled with 10 μg/ml concentration of Alexa488-labeled conconavalin A (ConA), which was obtained from Molecular Probes in the absence of detergent permeabilization to label the plasma membrane uniformly. The cells were then washed with PBS/serum and labeled with mouse anti-FLAG antibodies (M1) and Alexa594-labeled goat anti-mouse antibodies in the presence of 0.2% saponin as described above. Photomicrographs of 24 cells per time point or experimental treatment were obtained from a total of three independent experiments using a Zeiss (Heidelberg, Germany) LSM 510 laser scanning confocal microscope equipped with a 63× PlanApochromat oil immersion lens. The percentage of cell-surface receptors was determined by measuring the extent of LPA1 co-localization with the cell-surface marker Alexa594-labeled ConA. Quantitation of co-localization was performed as described previously using Metamorph Imaging System Software (Universal Imaging Corporation, West Chester, PA) (Volpicelli et al., 2001). Briefly, background was substracted from unprocessed images and the percentage of LPA1 pixels (red) overlapping ConA pixels (green) was measured. The data was normalized to untreated cells (time=0) and the percentage of internalized receptors was calculated by subtracting the percentage of cell-surface receptors from 100%. The data is presented as mean (± s.e.m.) and statistical analysis was performed using ANOVA followed by a Dunnett's post-hoc test.
At 30-36 hours after plating, cells were detached from a T-75 flask with trypsin/EDTA or scraped from culture dishes after the indicated treatment, washed twice with ice-cold PBS, and pelleted by centrifugation at 300 g for 5 minutes at 4°C. The pellets were resuspended in 100-200 μl of cell lysis buffer (1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M sodium phosphate pH 7.2, 2 mM EDTA, 50 mM NaF, 0.2 M sodium orthovanadate, 0.02% azide, 100 μg/ml leupeptin and 0.1 mM PMSF) and incubated on ice for 15 minutes. Detergent-insoluble material was removed by centrifugation at 13,000 g for 10 minutes at 4°C. The samples (30 μg of protein per lane) were then separated by SDS-PAGE on 10% gels and transferred to nitrocellulose paper. MAP kinase activation was detected using the PhosphoPlus p44/42 MAP Kinase antibody kit (Cell Signaling, Beverly, MA) and LPA1 was detected using a polyclonal rabbit anti-FLAG antibody (Sigma). The binding of primary antibodies was detected by enhanced chemifluorescence detection (Amersham Biosciences, Piscataway, NJ).
Measurement of serum response factor (SRF) activity
A transcriptional reporter gene assay (Clontech) was used to monitor the activity of SRF. For these studies, we used the HepG2 human hepatoma cell line since this cell does not contain functional LPA receptors (Fischer et al., 1998). Approximately 7×104 HepG2 cells were plated in 96-well dishes and transfected with 0.2 μg plasmid encoding FLAG-LPA1, 0.2μ g pSRE-luc, 0.05 μg pRL-TK and either 0.2 μg of pBluescript KS+ or 0.2 μg of the GTPase construct. Cells were transfected in SF-DMEM using lipofectamine (Invitrogen) at 1 μl lipofectamine per 0.4μ g DNA. pSRE-luc encodes firefly luciferase and contains three tandem copies of the serum response element upstream of a basal promoter; luciferase expression is strongly stimulated by SRF. The pRL-TK construct constitutively encodes Renilla reniformis luciferase whose expression is controlled by a thymidine kinase promoter; Renilla luciferase expression serves to monitor transfection efficiency. After incubation with the DNA complexes for 24 hours, the cells were rinsed with SF-DMEM and incubated in the same medium with either no additions or 1 μM LPA for an additional 16 hours. Both firefly luciferase and Renilla luciferase activity was measured using the Dual Luciferase Reporter Assay System (Promega, Madison, WI) and data were collected with a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA). Normalized luciferase activity was calculated by dividing the firefly luciferase activity by the Renilla luciferase activity. Statistical analysis was performed using a single-factor ANOVA followed by a Tukey's statistical test.
Expression and functional analysis of epitope-tagged LPA1 receptors in HeLa cells
To investigate the consequences of agonist stimulation on the intracellular trafficking of human LPA1, we established a stably transfected HeLa cell line expressing human LPA1 containing an amino-terminal FLAG epitope tag. Western blotting showed that FLAG-tagged LPA1 was expressed as a protein of approximately 43 kDa in cell extracts prepared from stably transfected HeLa cells (Fig. 1A, E2), but was not detected in extracts from untransfected HeLa cells (Fig. 1A, H). This is consistent with a molecular mass of approximately 41 kDa that was previously reported for human LPA1 (Fukushima et al., 1998).
Since HeLa cells are known to express endogenous LPA1 and LPA2 receptors, we wanted to determine the time course of LPA-induced activation of signaling for later comparison with the time course of agonist-induced LPA1 endocytosis. Stimulation of many cell types with LPA induces a rapid, but transient, activation of the mitogen-activated protein kinase (MAPK) pathway (van Corven et al., 1992). Thus, we examined the time course of LPA-induced activation of the endogenous MAP kinases ERK1/2 (Fig. 1B) both in stably transfected HeLa cells expressing FLAG-tagged LPA1 and in untransfected HeLa cells. ERK1/2 activation was assessed using commercially available antibodies that recognize the dually phosphorylated, active form of ERK1/2. In cells stably expressing FLAG-LPA1, the levels of activated ERK1/2 increased rapidly from 1 to 5 minutes following treatment with 10 μM LPA, with peak activation occurring between 5 and 10 minutes, and then steadily decreased such that very little activated ERK1/2 could be detected after 30 minutes of LPA treatment. In the absence of LPA treatment, ERK phosphorylation was not detected. Untransfected HeLa cells also exhibited a rapid increase in activated ERK1/2; however, we consistently observed that the peak ERK activation occurred between 10 and 30 minutes. This response was slightly slower than that observed in cells over-expressing FLAG-LPA1 and was most probably due to enhanced ERK activation through the elevated levels of LPA1 present in the FLAG-LPA1-expressing cells. Taken together, these results indicated that LPA stimulation induced a rapid but transient activation of MAPK in HeLa cells.
Agonist-dependent internalization and recycling of LPA1
We next determined the effects of LPA stimulation on the cellular distribution of LPA1 (Fig. 2) using indirect immunofluorescence. Treatment with 10 μM LPA resulted in a time-dependent redistribution of LPA1 from a predominantly plasma membrane (PM) localization, observed in unstimulated cells (Fig. 2, 0 min), to small punctate intracellular structures. These structures are likely to be intracellular endosomal compartments since they were not observed if immunofluorescence labeling was performed without detergent permeabilization (data not shown). In the absence of permeabilization, anti-FLAG antibodies only labeled the LPA1 receptors at the cell surface by binding to the externally oriented FLAG epitope. Furthermore, the anti-FLAG antibodies did not label untransfected HeLa cells (data not shown). Endosomal staining was first observed within 10 minutes after LPA treatment and increased in fluorescence intensity such that, after 30 minutes of stimulation, LPA1 localized predominantly to these vesicular structures. There was also a noticeable decrease in plasma membrane labeling after 30 minutes of LPA treatment (Fig. 2, 30 min). This pattern of localization was the same after 60 minutes of LPA treatment (Fig. 2, 60 min).
To quantify LPA1 internalization, we used a quantitative immunocytochemical approach that was recently used to analyze muscarinic acetylcholine receptor internalization in PC12 cells (Volpicelli et al., 2001). This approach takes advantage of the observation that unstimulated receptors show greater co-localization with a plasma membrane marker than internalized receptors. Internalization is quantified by measuring the extent of fluorescence overlap between the receptor and plasma membrane marker using fluorescence imaging software (see Materials and Methods). In the case of the M4 muscarinic acetylcholine receptor, data obtained from the fluorescence quantitation of receptor endocytosis was indistinguishable from data obtained through radioactive ligand binding experiments (Volpicelli et al., 2001). We measured the effects of LPA treatment on the extent of fluorescent pixel overlap between a plasma membrane marker, Alexa488-labeled ConA, and LPA1, which was stained with mouse anti-FLAG antibodies and Alexa594-labeled secondary antibodies. Fig. 3A shows a representative panel of images, obtained by confocal microscopy, which compares the distribution of LPA1 and Alexa488-ConA in stably transfected HeLa cells either before or after treatment with 10 μM LPA. Both Alexa488-ConA and LPA1 are extensively co-localized at the PM in untreated cells. Following LPA treatment, there is a significant reduction in the extent of co-localization between LPA1 and Alexa488-ConA. Quantification of the overlap in fluorescence showed that approximately 40% of surface LPA1 receptors are internalized within 15 minutes after LPA treatment and that there is no further increase in internalization (Fig. 3B). This is comparable with the extent of internalization of b2-adrenergic receptors (b2ARs) (Oakley et al., 1999; Seachrist et al., 2000).
We next sought to determine the identity of the LPA1+ endosomal structures. We co-localized the internalized LPA1 with different endocytic organelle markers using double-label immunofluorescence staining (Fig. 4). The internalized LPA1 showed extensive overlap with both transferrin receptor (TfR) and the early endosomal marker EEA1. Interestingly, LPA1 appeared to coincide more with TfR+ compartments than with EEA1+ compartments. Since TfR labeling includes small transport vesicles, sorting endosomes, as well as juxtanuclear recycling endosomes, these observations are consistent with the possibility that internalized LPA1 traverses the same endocytic pathway as the TfR. By contrast, LPA1 did not co-localize with the lysosomal marker LAMP-2, indicating that following short-term exposure to LPA, these receptors are not transported to lysosomes. This raised the possibility that internalized LPA1 might recycle back to the cell surface.
Internalization of other GPCRs, such as the b2AR, is thought to be required for receptor resensitization and subsequent recycling (Oakley et al., 1999). Internalized b2ARs have been shown to be dephosphorylated in an early endosomal compartment prior to recycling back to the cell surface (Pitcher et al., 1995; Seachrist et al., 2000). We investigated whether internalized LPA1 could recycle back to the PM upon removal of LPA (Fig. 5). Cells were first treated with 10 μM LPA for 30 minutes to induce internalization of LPA1 into endosomal compartments. The cells were rinsed to remove LPA and then incubated at 37°C for various times prior to fixation and immunofluorescence localization of LPA1. In the absence of LPA treatment, LPA1 was predominantly localized to the PM (Fig. 5A, Untreated). After 30 minutes treatment with 10 μM LPA, LPA1 localized to numerous endosomal structures (Fig. 5A, +LPA). Upon removal of agonist (Fig. 5B), LPA1 first localized to large juxtanuclear endosomes after 5 minutes and then began to appear at the PM after 15 minutes with a corresponding decrease in endosomal labeling. Within 30 to 60 minutes after removal of agonist, LPA1 was predominantly localized to the PM. These observations indicated that internalized LPA1 rapidly recycled back to the PM upon removal of LPA.
LPA1 internalization is both dose dependent and LPA specific
To determine whether LPA-induced internalization of LPA1 occurred at physiologically relevant concentrations of LPA, we determined the dose dependence of LPA treatment on LPA1 internalization (Fig. 6). Concentrations of LPA in the range of 1-10 μM have been reported to be required for growth stimulation of fibroblasts (van Corven et al., 1992). Following 30 minutes incubation with different concentrations of LPA, we observed that LPA1 internalization was dose dependent and that labeling of small punctate endosomal structures was first observed after treatment with 10 nM LPA. We observed a steady increase in the number and fluorescence intensity of these endosomal structures as the concentration of LPA was increased up to 100 μM.
To determine whether internalization of LPA1 was specific for LPA, we examined the effects of two related bioactive lipids, S1P and LPC. S1P (100 nM) has been shown to potently and specifically activate the closely related S1P1/EDG-1, S1P3/EDG-3, S1P2/EDG-5, S1P4/EDG-6 and S1P5/EDG-8 receptors (Hla et al., 2001). Treatment of LPA1-expressing HeLa cells with either S1P (0.1 μM or 10μ M) or LPC (1 μM) did not induce the internalization of LPA1. Rather, LPA1 remained at the cell surface, suggesting that neither of these related lipids stimulated LPA1 internalization (Fig. 7). Although cells treated with S1P appeared to have larger puncta, these were not internal structures since immunofluorescence labeling in the absence of detergent permeabilization was the same as that observed in permeabilized cells (data not shown). At higher concentrations of LPC, the cells became rounded and detached from the substratum (not shown). Taken together, these results indicated that LPA1 internalization was dependent upon LPA concentration and was specifically stimulated by LPA and not by other related lipids.
Agonist-induced internalization of LPA1 is dependent upon functional dynamin2 and Rab5 proteins
Since internalized LPA1 co-localized with endosomal markers of the clathrin-mediated endocytic pathway, we investigated whether LPA1 was perhaps internalized by clathrin-dependent mechanisms. To address this question, we examined the effects of either WT or dominant-inhibitory mutants of dynamin2 and Rab5, which are known regulators of clathrin-dependent endocytosis (Bucci et al., 1995; Cao et al., 1998; Damke et al., 1994). Stably transfected HeLa cells expressing LPA1 were transiently transfected with GFP-tagged mutants of dynamin2, K44A (Dyn2-GFP K44A) (Fig. 8), or Rab5a, S34N (GFP-Rab5a S34N) (Fig. 9), as well as GFP-tagged WT forms of these GTPases, and assessed for agonist-induced endocytosis.
The 100 kDa GTPase dynamin2 is ubiquitously expressed and is involved in the severing of deeply invaginated clathrin-coated pits to form clathrin-coated vesicles (Damke et al., 1994). In cells expressing Dyn2-GFP K44A, agonist-stimulated internalization of LPA1 was completely inhibited and LPA1 remained at the cell surface (Fig. 8B). In contrast to Dyn2-GFP K44A, cells transfected with WT Dyn2-GFP displayed agonist-induced internalization of LPA1 that was indistinguishable from cells expressing LPA1 alone. Both WT and mutant Dyn2 localized in a diffuse cytoplasmic pattern in the transfected cells. This suggested that LPA1 internalization followed a dynamin-dependent pathway.
The Ras-related Rab5 GTPase is another regulator of early endocytic traffic between the PM and early endosomes (Bucci et al., 1995). Rab5 is known to stimulate homotypic endosomal fusion following endocytosis. A recent study by Seachrist et al. (Seachrist et al., 2000) has shown the dominant-inhibitory GFP-tagged Rab5a S34N mutant potently inhibits agonist-induced internalization of b2-adrenergic receptors. Similarly, expression of GFP-Rab5a S34N in LPA1-expressing HeLa cells strongly inhibited agonist-induced internalization (Fig. 9B). In these cells, Rab5a S34N showed a diffuse cytosolic distribution throughout the cell. In these same cells, LPA1 was localized at the cell surface and showed no vesicular labeling as observed in cells that were not transfected with GFP-Rab5a S34N. Transfection with WT GFP-Rab5a did not alter agonist-induced internalization of LPA1, which localized to punctate internal structures as observed in cells expressing LPA1 alone. To quantify the phenotypic effects of Dyn2-GFP K44A and GFP-Rab5a S34N on LPA1 internalization, we scored the percentage of cells expressing these mutants for the presence of LPA1+ endocytic structures (Fig. 9C). In the absence of these mutant proteins, 71±4% of the cells contained LPA1+ endocytic structures following treatment with 10μ M LPA for 30 minutes. However, the results from three independent experiments indicated that expression of either Dyn2-GFP K44A (7±4%) or GFP-Rab5a S34N (2±1%) almost completely inhibited the appearance of LPA1+ endocytic structures following LPA treatment. Taken together, these results indicated that LPA-stimulated endocytosis of LPA1 is strongly dependent upon both dynamin2 and Rab5a.
To test if either Rab5 S34N or Dyn2 K44A affected LPA1-mediated signaling, we examined the effects of these mutants on LPA1-mediated stimulation of transcription via SRF activation (An, 2000). These experiments were performed in HepG2 human hepatoma cells since these cells are nonresponsive to LPA (Fig. 10A) and do not express any known LPA receptors (Fischer et al., 1998). SRF activity was monitored using a firefly luciferase reporter gene plasmid that contains three tandem copies of the serum response element upstream of a basal promoter (see Materials and Methods). HepG2 cells were transiently transfected in serum-free medium with plasmids encoding the firefly luciferase reporter plasmid, the Renilla luciferase reference plasmid (to normalize for transfection efficiency), and the FLAG-tagged LPA1 expression plasmid. In addition, the cells were also transfected with either WT or mutant Rab5 or Dyn2 expression plasmids. At 24 hours following transfection, the cells were incubated either in the presence or absence of 1 μM LPA for 16 hours prior to determination of luciferase activity. In cells expressing the SRE-luciferase plasmid alone, LPA treatment did not induce luciferase expression, which is consistent with the absence of LPA receptors in these cells (Fig. 10A, SRE-Luc Alone). By contrast, cells that co-expressed LPA1 and the SRE-luciferase construct exhibited a 1.5- to 2-fold increase in firefly luciferase activity when treated with 1 μM LPA (Fig. 10A, SRE-Luc + LPA1). The data in Fig. 10B and 10C show that neither expression of WT Rab5 nor WT Dyn2 significantly affected the LPA1-mediated induction of firefly luciferase activity in response to agonist treatment. Induction of SRF activity was mildly inhibited in cells expressing dominant-inhibitory Rab5 S34N. Co-expression of dominant-inhibitory Dyn2 K44A greatly elevated SRF activity in both untreated and LPA-treated cells; however, this increase was observed in cells expressing Dyn2 K44A alone and thus was independent of LPA1 expression (data not shown). Analysis of the LPA-dependent increase in SRF activity (Fig. 10C) showed that both Rab5 S34N (28% inhibition) and Dyn2 K44A (26% inhibition) slightly diminished LPA1-dependent SRF activation (P<0.05). The fold increase in LPA-stimulated SRF activity was 72% (cells co-expressing LPA1 and Rab5 S34N) and 74% (cells co-expressing LPA1 and Dyn2 K44A) of that observed in cells expressing LPA1 alone. These data indicate that Rab5 and Dyn2 are critical for the agonist-induced internalization of LPA1 and can also influence LPA1-dependent SRF activation.
Basal internalization and recycling of LPA1 in serum-containing medium
Given that cells in culture are constantly bathed in serum-containing medium, we examined whether the LPA present in medium containing 10% FBS was sufficient to trigger LPA1 internalization. It has been estimated that normal serum levels of LPA range from 0.1 to 10 μM (Xu et al., 1998). Immunofluorescence localization of LPA1 in cells grown in 10% serum showed that it was primarily localized to the PM with little vesicular labeling (Fig. 11A, No treatment). To investigate whether LPA1 was internalized at a low level, we determined the localization of LPA1 in the presence of 10% serum in the presence of the proton ionophore monensin Monensin has been shown to disrupt the low pH environment of endosomal compartments and, as a consequence, disrupt receptor recycling to the PM (Basu et al., 1981). Incubation of the LPA1 stable HeLa transfectants with 25 μM monensin resulted in a time-dependent accumulation of LPA1 in endosomal structures (Fig. 11A, +25μ M monensin, 15 min and 30 min). Labeling of these structures was first observed after 5 minutes of treatment and then steadily increased such that, after 30 minutes of treatment, the pattern of LPA1 labeling was similar to that of cells treated with 10 μM LPA in serum-free medium for 30 minutes (Fig. 2, 30 min). Monensin treatment itself did not induce LPA1 internalization since monensin treatment in serum-free medium did not stimulate LPA1 internalization; internalization in serum-free medium required the addition of LPA (Fig. 11B). Furthermore, monensin treatment inhibited LPA1 recycling in serum-free medium upon removal of LPA (data not shown). These results suggest that LPA1 undergoes a low basal internalization and most probably recycles back to the cell surface when cells are cultured in serum-containing medium.
In this study, we investigated the agonist-induced trafficking of the LPA-coupled LPA1/EDG-2 receptor in stably transfected HeLa cells expressing FLAG-tagged human LPA1. LPA1 was rapidly internalized from the PM in response to LPA stimulation in both a time-dependent and dose-dependent manner (Figs 2, 3 and 6). LPA1 internalization was specific for LPA, since neither S1P nor LPC, which are structurally similar to LPA, stimulated internalization (Fig. 7). Removal of agonist stimulates recycling of LPA1 back to the PM (Fig. 5). Internalized LPA1 co-localized with the endosomal markers EEA1 and TfR (Fig. 4), which label early endosomal compartments of the clathrin-dependent endocytic pathway. Dominant-inhibitory mutants of dynamin2 and Rab5a potently inhibited LPA1 internalization and also slightly diminished LPA1-dependent stimulation of SRF (Figs 8, 9, 10). These results are consistent with the agonist-induced internalization of LPA1 following a clathrin- or caveolae-dependent process. Finally, our results indicate that LPA1 cycles between the PM and endosomes at a low basal level in cells that are cultured in serum-containing medium.
LPA1 internalization is a consequence of receptor activation
Several lines of evidence indicate that LPA1 internalization is a consequence of agonist-induced receptor activation. First, LPA1 internalization was dependent upon LPA concentration. We observed that concentrations as low as 10 nM LPA could induce modest LPA1 internalization (Fig. 6). Internalization continued to increase as the LPA concentration was increased up to 100 μM LPA. This is consistent with published reports that have shown LPA concentrations in this range (i.e. 0.1-20 μM) potently induce intracellular signaling pathways such as stress fiber formation, inhibition of forskolin-stimulated adenylate cyclase activity and growth stimulation (Fukushima et al., 1998; Ishii et al., 2000; van Corven et al., 1992).
Second, comparison of the time course of LPA1 internalization with that of LPA-induced MAPK activation showed that LPA1 internalization coincides with signal desensitization. Analysis of the time course of LPA stimulation of MAPK activity (Fig. 1B) showed that maximal MAPK activation occurred after approximately 5 minutes of LPA treatment and MAPK activity then decreased between 10 to 30 minutes of LPA treatment. By contrast, LPA1 was primarily localized to the PM after 5 minutes of LPA treatment (Fig. 2). LPA1 was first observed in endosomal structures after 10 minutes of LPA stimulation and this endosomal labeling steadily increased thereafter. This is consistent with the internalization of LPA1 occurring after signal desensitization. Finally, LPA1 internalization was specific for LPA treatment. Neither S1P (10 μM) nor LPC (0.1 μM) stimulated the internalization of LPA1. Thus, these observations suggest that LPA1 internalization is a consequence of receptor activation, similar to other GPCRs that undergo agonist-induced internalization.
LPA1 is likely to be internalized via clathrin-dependent endocytosis
Several observations from this study are consistent with LPA1 internalization occurring via clathrin-dependent endocytosis. First, we observed that internalized LPA1 showed extensive co-localization with the clathrin-dependent endosomal markers EEA1 and TfR (Fig. 4). TfRs are internalized via clathrin-dependent endocytosis and EEA1 is a Rab5 effector that is recruited to early endosomal membranes by activated Rab5 (Bucci et al., 1995).
Second, our findings that LPA1 internalization is dependent upon the function of dynamin2 and Rab5a suggest that LPA1 might be internalized via clathrin-dependent endocytosis. Both dynamin2 and Rab5 GTPases are known regulators of clathrin-dependent endocytosis (Bucci et al., 1995; Damke et al., 1994). Dynamin2 is ubiquitously expressed and has been shown to be required for the severing of deeply invaginated clathrin-coated pits to form coated vesicles and also for the severing of invaginated caveolae (Damke et al., 1994; Henley et al., 1998). Following coated vesicle formation, the clathrin coats rapidly dissociate from coated vesicles in an ATP-dependent fashion. The Rab5a GTPase then stimulates the homotypic fusion of these uncoated vesicles by regulating the formation of the proper v-SNARE/t-SNARE associations and by recruiting the components of the vesicle fusion machinery (Miaczynska and Zerial, 2002).
Dominant-inhibitory mutants of both dynamin2 (Dyn2 K44A) and Rab5a (Rab5 S34N) both strongly inhibited the LPA-induced internalization of LPA1 (Figs 8 and 9). In cells expressing these GTPase mutants, LPA1 was confined to the cell surface. Given that Rab5a and dynamin2 are known regulators of clathrin-dependent endocytosis, these results suggest that LPA1 is likely to be internalized via clathrin-dependent endocytosis. However, since Dyn2 K44A inhibits both clathrin-coated vesicle formation as well as formation of caveolae-dependent transport structures; it remains possible that LPA1 is internalized by either clathrin- or caveolae-dependent mechanisms.
Interestingly, we observed that LPA1 was confined to the cell surface in cells expressing GFP-Rab5a S34N. The best-described role for Rab5 is in mediating the homotypic fusion of early endosomes (Miaczynska and Zerial, 2002). However, several recent studies have also shown a role for Rab5 in the sequestration of receptor–ligand complexes into clathrin-coated pits (McLauchlan et al., 1998; Seachrist et al., 2000). A complex of Rab5 and Rab guanine nucleotide dissociation inhibitor (Rab-GDI) has been shown to be a necessary cytosolic component for the sequestration of TfRs into coated pits (McLauchlan et al., 1998). Thus, failure to internalize LPA1 in cells expressing GFP-Rab5a S34N may be a consequence of a defect in receptor localization to coated pits.
Liu et al. (Liu et al., 1999), have previously shown that the S1P-coupled receptor, S1P1/EDG-1, also undergoes agonist-stimulated internalization and extensively co-localizes with internalized transferrin and also partially co-localizes with lysosomal markers suggesting that S1P1/EDG-1 is internalized via clathrin-dependent endocytosis. Together with our results on LPA1 trafficking, these observations suggest that perhaps other lysophospholipid receptors may also undergo agonist-induced internalization. Whether or not internalization of these other family members occurs via clathrin-mediated mechanisms or perhaps non-clathrin-dependent pathways remains to be determined.
Role of endocytosis in regulation of LPA1 function
GPCR endocytosis in many instances occurs subsequently to ligand-induced G-protein activation and involves receptor phosphorylation and the binding of arrestin proteins (Ferguson, 2001). Internalization is thought to contribute to either signal desensitization and/or resensitization once the internalized GPCR is dephosphorylated in an endosomal compartment. Thus, one role for LPA1 internalization might be to facilitate its dephosphorylation and subsequent resensitization.
In addition to receptor resensitization, several observations suggest a broader role for GPCR endocytosis in receptormediated signaling events. Several recent studies suggest that activated GPCRs can assemble multi-protein signaling complexes to initiate secondary signaling events from endosomal compartments within cells. Studies of the thrombin receptor, PAR1, the neurokinin-1 receptor, and the angiotensin 1a receptor have shown that following agonist treatment, these internalized GPCRs form complexes, via b-arrestin, with downstream components of the MAPK signaling pathway including Raf1, MEK1 and ERK2 (DeFea et al., 2000; Luttrell et al., 2001). Interestingly, these MAPK components co-localize with the internalized GPCRs on endosomal structures. It has been suggested that this may provide a G-protein-independent mechanism to target activated ERKs to specific intracellular compartments to phosphorylate cytoplasmic targets selectively.
The data in Fig. 10 indicate that inhibition of LPA1 internalization slightly decreased LPA-dependent induction of SRF-mediated transcription; dominant-inhibitory Rab5a S34N and Dyn2 K44A reduced LPA1-dependent activation of SRF by 28% and 26%, respectively. However, these mutants strongly inhibited LPA1 internalization (Fig. 9), suggesting that the primary effect of these mutants was to impede LPA1 endocytosis. LPA-dependent activation of SRF is mediated through the stimulation of Ras- and Rho-dependent signaling (Hill et al., 1995; van Corven et al., 1993) through Gbg and G12/13 signaling pathways. Others have shown that dynamin mutants can inhibit LPA-induced ERK activation via the Ras pathway (Daaka et al., 1998; Kranenburg et al., 1999). Thus, one possible explanation for the slight reduction in LPA1-dependent SRF activation is that dyn2 K44A and Rab5a S34N might inhibit the Ras/ERK-dependent component of SRF activation.
It is also possible that LPA1 internalization may be important for other LPA-dependent signaling processes. The data in Fig. 11 suggests that LPA1 is internalized and most probably recycled at a low basal level in cells cultured in serum-containing medium. Given that serum contains LPA, this basal internalization is likely to represent agonist-induced uptake. If so, then this raises the question of what the long-term signaling consequence of such basal uptake is on cells. Further studies of the role of LPA1 localization in the regulation of LPA-stimulated signaling, as well as other LPA-coupled receptors, is likely to provide important information about the role of endocytosis in regulating LPA-induced cellular responses.
Finally, an important implication of our finding that LPA1 is internalized in serum-containing medium is that LPA1 internalization may be a useful diagnostic measure of the relative levels of LPA present in clinical serum samples. Recent observations indicate that serum LPA levels are increased in patients with ovarian cancer even at early stages (Xu et al., 1998). Measurement of LPA1 internalization could be adapted into a simple bioassay for screening patient serum and/or ascites samples for LPA.
We thank Junken Aoki (University of Tokyo, Japan) for kindly providing an expression plasmid encoding FLAG-tagged human LPA1, Mark McNiven (Mayo Clinic, Rochester, MN) for an expression plasmid encoding a green fluorescent protein (GFP)-tagged dynamin2 K44A mutant (Dyn-GFP2 K44A) and Stephen Ferguson (Robarts Research Institute, London, Ontario) for providing expression vectors encoding GFP-Rab 5 S34N and GFP-Rab 5 Q79L. We also thank Julie Donaldson, Nael McCarty and members of the Radhakrishna lab for critically reviewing the manuscript. This work was supported in part through an American Heart Association Beginning Grant-in-aid 00602758 and National Institutes of Health Grant HL 67134 to H.R.
↵* These authors contributed equally to this work
- Accepted January 28, 2003.
- © The Company of Biologists Limited 2003