G protein-coupled receptors (GPCRs) have the potential to play a role as molecular sensors responsive to luminal dietary contents. Although such a role for GPCRs has been implicated in the intestinal response to protein hydrolysate, no GPCR directly involved in this process has been previously identified. In the present study, for the first time, we identified GPR93 expression in enterocytes and demonstrated its activation in these cells by protein hydrolysate with EC50 of 10.6 mg/ml as determined by the induction of intracellular free Ca2+. In enterocytes, GPR93 was synergistically activated by protein hydrolysate in combination with an agonist, oleoyl-l-α-lysophosphatidic acid (LPA), which activated the receptor in these enterocytes with EC50 of 7.9 nM. The increased intracellular Ca2+ by GPR93 activation was observed without the addition of a promiscuous Gα protein and was pertussis toxin sensitive, which suggests Gαq- and Gαi-mediated pathways. Activated GPR93 also induced pertussis toxin-sensitive ERK1/2 phosphorylation. Both nuclear factor of activated T cells and 12-O-tetradecanoylphorbol 13-acetate responsive elements reporter activities were induced by protein hydrolysate in cells exogenously expressing GPR93. The peptidomimetic cefaclor by itself did not activate GPR93 but potentiated the protein hydrolysate response and further amplified the synergistic enhancement of GPR93 activation by protein hydrolysate and LPA. These data suggest that, physiologically, the composition of stimuli might determine GPR93 activity or its sensitivity toward a given activator and suggest a new mechanism of the regulation of mucosal cell proliferation and differentiation and hormonal secretion by dietary products in the lumen.
- extracellular signal-regulated protein kinase 1/2
- lysophosphatidic acid
intestinal mucosal homeostasis necessitates the coordination of physiological and chemical factors that regulate functions from cellular renewal and differentiation along the crypt-to-villus axis to the regulation of hormone secretion, immune response, and nutrient assimilation along the proximal-to-distal intestine. This coordination is partly facilitated by cellular factors that initiate intracellular signals in response to the luminal content. The importance of the presence of luminal dietary nutrients is well exemplified in patients undergoing total parenteral nutrition (TPN). Chronic TPN resulting in a dramatic mucosal remodeling can lead to a compromised absorptive capacity and intestinal immune function and the development of intestinal and liver diseases.
Studies investigating nutrient-induced signaling and gene regulation have mostly centered on the effects of fatty acids and glucose on the expression of proteins involved in transport and metabolism in adipocytes and pancreatic β-cells, respectively (15, 19, 33, 52, 65). Signaling events initiated by luminal protein hydrolysate have been a focus of studies for defining the mechanisms leading to the release of CCK (48, 49). In the enteroendocrine STC-1 cells, protein hydrolysate activates ERK1/2, CaMK pathways, as well as the PKA pathways (19). Several studies have demonstrated that the uptake of protein hydrolysate in the intestine is through the proton-coupled oligopeptide transporter PepT1 (1, 13, 14). Oligopeptides uptake through this transporter is linked to events that subsequently lead to an induced transcription of CCK (19) and the release of CCK (14). There are numerous examples of nutrient-sensing events in the lumen that may be mediated by transporters or receptors. The indirect involvement of G protein-coupled receptors (GPCRs) in these sensing events in which the release of their ligands is directly induced by luminal nutrients is well characterized. For example, free fatty acids directly induce the release of peptide YY (3), which in turn activates neuropeptide Y (NPY) receptors, or protein hydrolysate induces the release of CCK (12, 42, 48), which in turn activates the receptor CCK1R (63). The role of GPCRs as sensors that can directly respond to changes in luminal contents is not well defined. The characteristics of their seven transmembrane configuration not only allow for a given GPCR to recognize a wide range of molecular structures but also activate multiple pathways depending on the conditions of the stimuli.
Some GPCRs have been reported to be directly activated by basic l-amino acids such as the Ca2+-sensing receptor CaR (9); the GPRC6A, which has wide tissue distribution outside the intestine (61); and the heterodimeric taste receptor T1R1/T1R3 (47). The identification of Gα gustducin and transducin, which associate with taste receptors, in the gastrointestinal (GI) mucosa as well as the presence of T2R in the GI tract suggests this family of GPCRs can act as sensors that may respond to luminal contents such amino acids and toxins (27, 66).
Using the enterocyte-like hybrid Berkeley Rat Intestine Epithelial 380 (hBRIE 380i) cells, which do not express PepT1, we explored the possibility that signaling cascades in the enterocytes initiated by protein hydrolysate could be directly mediated by the activation of a GPCR. In the present study, we identified for the first time a GPCR, GPR93, in the rat enterocytes that is directly responsive to protein hydrolysate. GPR93 activation by protein hydrolysate mobilized intracellular Ca2+ concentration ([Ca2+]i) and activated ERK1/2 through both Gαq- and Gαi-mediated pathways. Protein hydrolysate and oleoyl-l-α-lysophosphatidic acid (LPA) synergistically activated GPR93. Our data suggest that GPR93 could be partly responsible for luminal protein hydrolysate-induced ERK1/2 activation and for the subsequent effects of this activation, such as alterations in cell proliferation and differentiation.
MATERIALS AND METHODS
All reagents, including meat protein hydrolysate (peptone) type I and LPA, which were used for GPR93 characterization in the entire study, were purchased from Sigma-Aldrich, unless indicated differently.
Cell culture condition and transfection.
The hBRIE 380i cells used in this study were from a well-characterized subclone expressing enterocyte phenotypes and protein markers (2, 24, 25). Experiments were performed using hBRIE 380i cells of the same passage number. The hBRIE 380i and Chinese hamster ovary (CHO) cells were maintained in Iscove's modified Dulbecco's medium (IMDM; Invitrogen) with 10% bovine calf serum (BCS; Hyclone), 100 U/ml penicillin, and 100 μg/ml streptomycin as additional supplements at 37°C in 5% CO2-95% air. For experiments, cells were plated 24 h before experiments and were at ∼80% confluency on the day of use. The hBRIE 380i cells used for Western blot analysis were laid down in 35-mm tissue culture dishes (Corning), coated with rat-tail collagen type I (2). For transfection, cells were trypsinized, resuspended in IMDM (7.5 × 106 cells/ml), and incubated with the plasmid DNA in a volume between 0.2 and 0.7 ml at room temperature for 5 min. The total amount of DNA in each transfection was 8 μg/106 of cells, unless otherwise noted. Electroporation was carried out in a 0.4-mm cuvette at 0.25 kV and 960 μF, using a Gene Pulser (Bio-Rad). One milliliter of IMDM-10% BCS was added to the cuvette immediately after electroporation. The cells were then plated and allowed to recover for 20 h. The hBRIE 380i cells, stably expressing mitochondria-targeted aequorin (mtAEQ), were prepared as follows: 36 h after electroporation with 2 μg mtAEQ construct/106 hBRIE 380i cells, resistant clones were selected in the presence of 800 μg/ml G418 (Invitrogen). Expression of functional mtAEQ was verified using [Ca2+]i mobilization assay.
The open reading frames of GPR93, Gα15, Gαq, NPY receptor subtype 1 (NPY1R), β2-adrenergic receptor (β2AR), GPR103, P2Y5, P2Y9, and P2Y10 were PCR amplified using Pfu DNA polymerase (Stratagene) from rat intestine or brain cDNA. The sequences of oligonucleotide primers (Integrated DNA Technology) and GenBank accession numbers are listed in Table 1. Each receptor cDNA product was ligated into the multiple cloning site of pCI-neo expression vector (Promega), and each G protein cDNA was ligated into the multiple cloning site of pcDNA3.0 (Invitrogen). Gαq cDNA was used as a PCR template for the construction of GαΔ6qi5myr (10, 35, 36). GPR93 tagged at the COOH-terminus with the enhanced green fluorescent protein (GPR93-EGFP fusion), was constructed by ligating the GPR93 open reading frame (its stop codon removed using PCR) at the BamHI site of the enhanced green fluorescent protein (EGFP) expression vector (Clontech). The mtAEQ (53, 58), i.e., Ca2+-sensitive AEQ cDNA (AY604000) with the first 3 bases replaced with a 99-base fragment encoding the NH2-terminal 33 amino acids of human cytochrome c oxidase subunit VIII (including the 25 amino acid mitochondria signal sequence) (J04823), was constructed using overlapping synthetic oligonucleotide primers and ligated into pcDNA3.0. TPA response element (TRE) and nuclear factor of activated T cells (NFAT) luciferase reporter constructs are pBV-luc (a generous gift from Dr. Bert Vogelstein, The Johns Hopkins Kimmel Cancer Center) with 9× TPA response elements (TGACTAA) or 8× NFAT response elements (GGAGGAAAAACTGTTTCATACAGAAGGCGT), respectively, inserted in its multiple cloning site. All constructs were verified by DNA sequencing (DNA Sequencing Facility, University of California at Berkeley).
Tissue preparation and RNA isolation for GPR93 expression profile.
Overnight fasted male Sprague-Dawley rats (14 wk old) were used as tissues sources (n = 4). The protocol for animal use was reviewed and approved by the Animal Care and Use Committee of the University of California at Berkeley. GI tissue samples were prepared as follows: intestines were extracted and cut into segments (each ∼5 cm long). Residual luminal contents were removed by running ice-cold PBS through the intestinal segments. The mucosal layer of the intestine was obtained by gentle scraping of the exposed luminal surface. The muscle layer was obtained after further scraping to remove the residual mucosal layer. The stomach mucosa was prepared in a similar fashion. The purity of the mucosa was verified by the relative expression of villin and intestinal fatty acid binding proteins (I-FABP), differentiation markers of intestinal enterocytes, as determined by RT-PCR.
Non-GI tissue samples were obtained by removing the organs, followed by rinsing with ice-cold PBS twice and mincing with surgical scissors. Immediately, the tissue samples were immersed in ice-cold TRIzol (Invitrogen), homogenized, and frozen in liquid nitrogen. All tissue preparation steps were done on ice. RNA was isolated from the TRIzol tissue homogenate according to the manufacturer's protocol.
Reverse transcription was as described previously (40). PCR was carried out using Taq DNA polymerase (New England Biolab). GPR93 specific primers amplified a cDNA fragment of 249 bp. PCR parameters for GPR93 were as follows: 20 s at 94°C, 15 s at 55°C, and 30 s at 72°C for 32 cycles. The primers and PCR condition for villin, I-FABP, and 18S RNA were as previously described (40). Amplified cDNA fragments were analyzed by agarose gel electrophoresis followed by densitometry. The specificity of the PCR products was confirmed by DNA sequencing.
Laser microscopy dissection.
Rat duodenum tissue sections were prepared for cryostat as previously described (24). Briefly, the duodenum section was removed from overnight fasted male Sprague-Dawley rats (12 wk old), and the residual luminal contents were washed out by running ice-cold PBS through the duodenum segment. The duodenum was further cut into 2-mm sections in the horizontal direction after short fixation in 70% ethanol in PBS (pH 7.4). The tissue sections were briefly rinsed with ice-cold PBS and immersed in ice-cold 30% (wt/vol) sucrose in PBS overnight at 4°C. The sucrose-equilibrated sections were embedded into optimum cutting temperature compound (TissueTeK), frozen on dry ice, cryosectioned at 10 μm thickness, and then stored at −80°C.
For the laser microscopy dissection (LMD) procedure, 10 μm cryosections were mounted on slides, fixed with 70% ethanol for 30 s, and stained with eosin, followed by a 5-s dehydration step in each of 70%, 95%, and 100% ethanol. After brief air drying, the sections were laser microdissected using a Leica AS LMD system with the following setting: aperture, 6–10; intensity, 45; and speed, 2–5. Total RNA from laser-captured villus and crypt regions (15 patches from the crypt and villus area, separately, were pooled) was isolated using RNeasy Micro Kit (Qiagen). Semiquantitative RT-PCR was then performed. The purity of harvested LMD samples was confirmed by comparing the expression level of villin and I-FABP, differentiation markers of intestinal epithelial cells.
Localization of GPR93 in CHO cells.
CHO cells were transfected with the GPR93-EGFP fusion construct by electroporation (4 μg plasmid DNA/106 cells). After 24 h of recovery incubation in IMDM-10% BCS under normal culture conditions, cells were trypsinized, resuspended in phenol red-free IMDM-10% BCS, plated on six-well slides coated with collagen type I at a density of 104/well, and incubated for 1 h in humidified petri dishes. Slides were then transferred into fresh media and further incubated for 16 h under normal culture conditions. The localization of EGFP-tagged GPR93 was visualized by using Laser Scanning Confocal Microscopy (Zeiss 510 UV/Vis Meta system).
AEQ-based [Ca2+]i mobilization assay.
[Ca2+]i mobilization assay was performed as previously described (54) with slight modifications. Briefly, mtAEQ expression vector was coelectroporated (2 μg/106 cells) with other plasmid constructs as indicated in the figures, and the cells were allowed to recover for 20 h in IMDM-10% BCS. For CHO cells, cells were dislodged with 5 mM EDTA-PBS and loaded with 5 μM coelenterazine-h (Promega)-300 μM glutathione in IMDM (2 × 106 cells/ml) at 37°C for 2 h with gentle rolling. For hBRIE 380i cells, cells were trypsinized and gently rolled for 1 h in IMDM-10% BCS, followed by loading in HBSS (Invitrogen) for 1 h under the same condition as used for CHO cells. A 100-μl aliquot (5 × 104 cells in HBSS) was then assayed in a luminometer equipped with an injector (Turner BioSystem). The stimulus was injected into the cell suspension in a 100-μl aliquot at a 2× final concentration in PBS. A 100-μl aliquot of lysis buffer (300 mM CaCl2 and 300 μM digitonin) was injected 40 s later to react with the remaining AEQ. Luminescence [as relative light units (RLU)] was recorded continuously. Fractional RLU is increased RLU due to a stimulus normalized to the total RLU, i.e., the integrated RLU value for 30 s after injection of the stimulus plus that for 20 s after the addition of the lysis buffer. All reagents tested were dissolved in PBS (pH 7.4).
To determine whether protons could activate GPR93, stimuli in the form of the buffer at varying pH were used. Phosphate buffer (PB; 10 mM Na2HPO4 and 2 mM KH2PO4) was pH adjusted with HCl or NaOH to 6.5 (acidic), 7.4 (neutral), or 8.5 (basic), and NaCl was added to bring the osmolality to 300 osmol/kgH2O, as determined by microosmometer (Precision Systems). To test whether osmotic pressure could stimulate GPR93, modified PB (PB at 100 osmol/kgH2O, adjusted with NaCl, pH. 7.4), 2% glycerol in modified PB (211 osmol/kgH2O), and 4% glycerol in modified PB (553 osmol/kgH2O) were used as stimuli.
Fatty acid-free BSA (Roche) was dissolved in PBS (pH 7.4) at a concentration of 10 mg/ml (wt/vol). Digested BSA was prepared by incubating a BSA solution (10 mg/ml) with proteinase K (5 μg enzyme/mg BSA, Invitrogen) at 37°C for 20 h with gentle agitation. The solution was heated at 80°C for 15 min to inactivate proteinase K and then centrifuged at 16,000 g at 4°C for 5 min.
Luciferase reporter assay.
Two micrograms of the reporter construct/106 of either CHO or hBRIE 380i cells were electroporated with other constructs as indicated in the figures, and cells were seeded into 12-well plates at 5 × 105 cells/well in IMDM-10% BCS. For the NFAT reporter study, cells were allowed to recover in IMDM-10% BCS for 20 h after transfection. On the day of each experiment, cells were first washed three times with PBS, serum starved for 2 h, and then treated with either 10 μM LPA or 50 mg/ml peptone in serum-free IMDM for 6 h. For the TRE reporter study, cells were allowed to recover in IMDM-10% BCS for 36 h. On the day of each experiment, cells were washed three times with PBS and treated with either 10 μM LPA or 50 mg/ml peptone in serum-free IMDM for 12 h. Fatty acid-free BSA, at a concentration of 0.1% (wt/vol), was added as a carrier in the treatments. Forty microliters of passive lysis buffer (Promega) were added to each well after the treatments. The luciferase activities of the samples were determined according to the manufacturer's protocol using a luminometer and normalized to the total protein concentration, determined by the Bio-Rad protein assay (Bio-Rad).
For [Ca2+]i mobilization assay, cells were incubated with 80 ng/ml pertussis toxin (PTX) for 24 h before the assay. U-73122, or its inactive analog U-73343, at 10 or 20 μM, was mixed with the stimulus without preincubation. Nifedipine (10 μM) and thapsigargin (20 nM) were added to the cells 5 min and 30 min, respectively, before a stimulus was added. For Western blot analysis, cells were preincubated with 100 ng/ml PTX for 20 h. The preincubation with 100 nM wortmannin, 5 μM U-73122 or U-73343, or 50 μM PD-98059, was for 30 min.
Western immunoblotting analysis.
To determine the effect of protein hydrolysate on ERK1/2 activation, hBRIE 380i cells were transfected with 6 μg GPR93 cDNA/106 cells and laid down in 60-mm dishes at a density of 2 × 105 per dish. After a 24-h incubation in IMDM-10% BCS, cells were serum starved in IMDM-0.1% BCS for 12 h, followed by a 1-h incubation in HBSS. The cells were treated with 20 mg/ml of peptone or 100 nM LPA for 4 min, unless otherwise noted in the figures. After the treatment, cells were immediately placed on ice, rinsed twice with ice-cold PBS, and scraped with 40 μl of 2× Laemmli sample buffer per 60-mm dish. Western blot analysis on polyvinylidene difluoride membrane was carried out as previously described (24). Protein concentration was determined by the Bio-Rad protein assay. The amount of protein per lane was 0.2 μg for ERK1 detection and 10 μg for phosphorylated ERK1/2 (pERK1/2). The primary antibody against ERK1 (Santa Cruz Biotechnology) was used at 1:3,000 and against pERK1/2 (Cell Signaling) at 1:2,000. The horseradish peroxidase-conjugated secondary antibody was used at 1:10,000.
Identity score between different GPCRs was determined by the ClustalW alignment method.
Data are expressed as means ± SD. Statistical difference between multiple groups was determined by one-way ANOVA with Tukey's post hoc test performed using SPSS version 11. Significance was accepted at P < 0.05. Dose-response curves were generated using the curve-fitting software GraphPad Prism version 4.
Our objective was to study the GPCRs that transduce external dietary signals into the intestinal epithelial cells. We chose protein hydrolysate (peptone), a mixture of enzymatically derived peptide fragments (mostly between 120 and 1,200 Da) and free amino acids, as a representative dietary component since peptone mimics dietary proteins digest in the luminal chyme (12). Protein hydrolysate stimulates mucosal brush border ERK1/2 (6), but a GPCR that mediates this event has yet to be identified. The significance of ERK1/2 activity in the intestinal mucosa has been illustrated by the prominent subcellular localization of ERK1/2 in the brush border and by the responsiveness of ERK1/2 signaling to changes in the luminal content in response to feeding (6).
We determined that peptone also stimulated the phosphorylation of ERK1/2 in enterocyte-derived hBRIE 380i cells (Fig. 1). The oligopeptide transporter PepT1, a reported mediator of peptone responses in intestinal cells (14, 44), is not expressed in hBRIE 380i cells as determined by RT-PCR (data not shown). We took advantage of this fact to search for other mediators of peptone response using hBRIE 380i cells.
We focused on orphan GPCRs because the activation of many GPCRs often leads to increased ERK1/2 (28, 57), and GPCRs have a wide variety of activators: from large glycoprotein hormones, neurotransmitters, and metabolites to external stimuli, such as light and odorant molecules. Besides being activated by their high affinity ligands, GPCRs can also be stimulated by activators with different molecular characteristics; for example, the Ca2+-sensing GPCR can also be activated by l-amino acids (9), and the proton-sensing GPCR can also be activated by lysolipids (59).
The GenBank and Ensembl genomic DNA databases were searched for putative GPCRs with non-olfactory-like sequences. RNAs from small intestinal epithelial cells isolated by dispersion (2) and from hBRIE 380i cells were used to screen candidate GPCRs by RT-PCR (unpublished data). The selected candidate GPCRs were tested for their responsiveness to peptone in a [Ca2+]i mobilization assay by cotransfecting the promiscuous Gα15 (unpublished data).
We determined GPR93 was a candidate GPCR that was responsive to peptone. O'Dowd and coworkers (39) reported its expression in several mouse tissues. GPR93 is a family A GPCR that belongs to a group of purinoreceptor-like GPCRs (29), i.e., P2Y5, P2Y9, and P2Y10. The aim of this study was to characterize GPR93 activation by peptone in hBRIE 380i cells, focusing on the collective net [Ca2+]i flux and ERK1/2 phosphorylation.
Peptone and LPA induce [Ca2+]i in CHO cells overexpressing GPR93.
The first step of the characterization of GPR93 activation was performed in CHO cells using the mtAEQ-based [Ca2+]i mobilization assay. The proper localization of the transfected GPR93 to the cell surface plasmalemma was confirmed by fluorescence confocal microscopy using GPR93-EGFP fusion (Fig. 2A, inset). Although in the initial screening process Gα15 was cotransfected with GPR93, cells transiently transfected only with GPR93 also showed a significant, transient [Ca2+]i flux in response to peptone. There was no [Ca2+]i in the empty vector transfectant in response to peptone, with or without a promiscuous Gα (data not shown). Thus all experiments were carried out with GPR93-transfected cells without cotransfecting a promiscuous Gα (Gα6qi5myr or Gα15).
We tested the possibility that peptone might activate other GPCRs nonspecifically. Peptone-treated CHO cells overexpressing GPR103, NPY1R (cotransfected with Gα6qi5myr), or β2AR (cotransfected with Gα15) did not show an increase in [Ca2+]i. P2Y5, P2Y9, and P2Y10, with or without cotransfection with Gα15, were also not stimulated by peptone (data not shown). The possibility of GPR93 activation by compounds other than peptone, most of which could be in the diet, was also explored. Soy protein hydrolysate stimulated GPR93 in a similar extent to peptone. The compounds that did not induce [Ca2+]i are listed in Table 2. The possibility of the activation of GPR93 by proton/pH or hypo/hypertonicity was also explored. pH from 6.5 to 8.5, as well as osmolality from 100 to 553 osmol/kgH2O, did not induce [Ca2+]i in GPR93-overexpressing CHO cells.
Another compound that was tested was LPA. External LPA can be present in the diet as it is or as a product of PLA2 digestion in the lumen of the intestine. P2Y9, which has a 30% amino acid identity with GPR93, was recently reported to be activated by LPA (37). We treated GPR93-transfected cells with 5 μM LPA and observed that [Ca2+]i was significantly induced, but there was no [Ca2+]i release observed in the empty vector transfectant (data not shown). Therefore, LPA was included as a stimulus to further characterize GPR93 activation.
Peptone and LPA activation of GPR93 in CHO and hBRIE 380i cells is PTX sensitive and mediated by PLC-β.
GPR93-mediated [Ca2+]i flux in response to peptone or LPA was dose responsive (Fig. 2, A and C). The EC50 was 4 mg/ml for peptone and 3.6 nM for LPA (Fig. 2C). For comparison, known GPCRs were activated by their agonists, i.e., NPY1R (a Gαi-GPCR) with/without chimeric Gα6qi5myr (a chimeric G protein to enhance signals from Gαi-coupled receptor) by NPY, β2AR (a Gαs-GPCR) with/without Gα15 (to enhance signals from Gαs-coupled receptor) by isoproterenol, and GPR103 [a Gαq-GPCR (31)] by RFamide P518 (Fig. 2B). Cotransfecting Gα6qi5myr was needed to observe a significant increase in [Ca2+]i flux elicited through NPY1R. This was similar to β2AR, for which a cotransfection with Gα15 was required to yield an observable [Ca2+]i peak. In contrast, upon its activation, GPR103 yielded an increased [Ca2+]i similar to that of GPR93 without the addition of a promiscuous Gα protein. This suggests that a Gαq-mediated pathway is downstream of GPR93 activation. A [Ca2+]i flux due to a GPCR activation can be caused by Gαq or Gβγ dissociated from Gαi. PTX (80 ng/ml), a specific inhibitor of Gαi, reduced peptone stimulation by 50% and LPA stimulation by 60% (Fig. 2D), suggesting that GPR93 activation partially leads to a pathway involving Gαi. GPCRs often exert intracellular signaling events by coupling with more than one kind of G proteins; as an example, LPA receptor 1, 2, and 3 can couple with both Gαi and Gαq (8). The possible involvement of other Gα proteins in GPR93-mediated [Ca2+]i increase in our system remains to be determined. A possible downstream effector of Gαi- or Gαq-coupled receptors mediating the increase of [Ca2+]i is PLC-β. A treatment with PLC-β inhibitor U-73122, but not its inactive analog U-73343, almost abolished peptone induction and inhibited LPA stimulation by more than 50% (Fig. 2E). PLC-β liberates inositol triphosphate, which activates the Ca2+ channel on the endoplasmic reticulum membrane resulting in a [Ca2+]i flux. Nifedipine (10 μM), an inhibitor of plasma membrane L-type voltage-gated Ca2+ channel, did not affect the [Ca2+]i flux induced by peptone or LPA stimulation of GPR93 (data not shown). Thapsigargin (20 nM), an inhibitor of sarco(endo)plasmic reticulum Ca2+ ATPase, completely eliminated the increase in [Ca2+]i, suggesting that GPR93 activation induces Ca2+ release from intracellular stores (data not shown).
We repeated the characterization of GPR93 activation in hBRIE 380i cells. In hBRIE 380i cells transiently transfected with GPR93, peptone or LPA induced a [Ca2+]i flux without cotransfecting a promiscuous Gα protein. The increase in [Ca2+]i levels was concentration dependent (Fig. 3A). The EC50 for peptone was 10.6 mg/ ml and 7.9 nM for LPA (Fig. 3A). PTX reduced the peptone stimulation by 68% and LPA stimulation by 50% (Fig. 3B). U-73122 decreased peptone stimulation by 83%, whereas U-73343 had no effects (Fig. 3C). The effect of U-73122 was dose responsive (data not shown). Similarly, U-73122, but not U-73343, inhibited LPA stimulation by 52% (Fig. 3C). These data are consistent with what was observed in CHO cells.
GPR93 activation leads to induced NFAT- and TRE-luciferase reporter activities.
We tested whether the increase in [Ca2+]i levels due to GPR93 activation could lead to downstream gene responses by the luciferase reporter assay. We examined the luciferase activity of reporter constructs containing response elements to NFAT and TPA (TRE) as indicators for increased [Ca2+]i. GPR93 activation by peptone or LPA in CHO cells induced luciferase activity from a construct containing NFAT responsive elements (Fig. 4A). The NFAT luciferase reporter could not be used in hBRIE 380i cells because even the positive control (TPA plus ionomycin) did not induce the luciferase activity, which might be due to the lack of expression of the transcription factor (unpublished data). Therefore, we used a luciferase reporter construct containing TRE. The activation of GPR93 by peptone or LPA induced luciferase expression (Fig. 4B). Interestingly, the effect of peptone on NFAT-driven reporter was weaker than that of LPA in CHO cells, but it was LPA that had a weaker effect on the TRE luciferase reporter in hBRIE 380i cells. Peptone and LPA also induced luciferase expression from the TRE-luciferase reporter in GPR93-transfected CHO cells with LPA having a stronger effect (data not shown). The difference in peptone and LPA effects on the reporter gene suggests that activation of GPR93 by peptone and LPA is distinct and cell type specific.
GPR93 activation mediates enhanced ERK1/2 phosphorylation in hBRIE 380i cells that is sensitive to PTX, mediated by PLC-β, and dependent on MAPKK.
Cells transfected with GPR93 exhibited enhanced ERK1/2 activation in response to peptone or LPA compared with cells transfected with the empty vector. The level of pERK1/2 was induced rapidly in response to peptone or LPA, reaching a maximum at 4 min and then gradually decreasing (data not shown). Peptone at a dose between 20 and 50 mg/ml maximally induced ERK1/2 phosphorylation in GPR93-overexpressing cells, whereas specific LPA stimulation was only observed at 100 nM (data not shown).
To determine the correlation between GPR93 activation and the increase in pERK, PTX and U-73122 were used. PTX (100 ng/ml) almost abolished peptone (Fig. 5A) as well as LPA (Fig. 5A) induction. Likewise, U-73122 decreased pERK induction by peptone (Fig. 5B) and LPA (Fig. 5B) to the basal levels. The treatment with phosphatidylinositol 3-kinase inhibitor wortmannin (100 nM) did not change the effect of peptone on pERK (data not shown). These results suggest an involvement of Gβγ subunits from Gαi and PLC-β activation in the induction of pERK. This does not exclude the possibility of other pathways, which were not explored in this study, contributing to ERK1/2 activation (including those that are Gαq mediated). The observed induction of pERK due to GPR93 activation requires MAPKK (MEK1/2), since PD-98059 (MEK1/2 inhibitor) at 50 μM abolished induced pERK by peptone (Fig. 5C) and LPA (Fig. 5C).
Peptone and LPA synergistically activate GPR93.
Whether peptone and LPA could act in concert to activate GPR93 was tested by treating GPR93-transfected hBRIE 380i cells with varying doses of peptone and LPA simultaneously in a [Ca2+]i mobilization assay (Fig. 6A). Peptone or LPA individually activated GPR93. However, [Ca2+]i increase was more than additive when both were combined and used as a stimulus (Fig. 6A, bar 4 vs. 2 and 3; and bar 7 vs. 5 and 6). Fatty acid-free BSA at 0.1% and peptone were added together in this particular experiment because LPA solutions used in all our experiments contained 0.1% fatty acid-free BSA as a carrier. We observed that GPR93 did not significantly respond to BSA stimulation up to 100 mg/ml (data not shown); however, BSA reduced [Ca2+]i flux due to peptone (data not shown) or LPA stimulation (Fig. 7A, bar 6 vs. 5; and bar 8 vs. 7). Therefore, [Ca2+]i induction by peptone in Fig. 6 was to a less extent than what was shown in Fig. 3C.
GPR93 could be activated by peptone and LPA as shown in Fig. 3. We tested whether endogenous GPR93 activation in hBRIE 380i cells could also be determined by [Ca2+]i mobilization assay using hBRIE 380i cells stably expressing mtAEQ. At the cell stage before confluency, peptone or LPA did not increase [Ca2+]i (data not shown). However, when the cells were allowed to become differentiated at 8-d postconfluency, peptone and LPA induced [Ca2+]i (Fig. 6B). Induction by peptone was much weaker than LPA compared with what was observed in proliferating cells in the transient transfection system. This might be due to increased expression of other endogenous LPA receptors in our cell culture system that, unlike GPR93, are not peptone responsive. The induction of [Ca2+]i was synergistically enhanced when peptone and LPA were combined (Fig. 6B, bar 4 vs. 2 and 3; and bar 6 vs. 2 and 5).
We used two approaches to verify that the peptide components in peptone were indeed responsible for GPR93 activation and for synergistic effects of peptone plus LPA on GPR93 activation. We tested the effect of proteinase K-digested fatty acid-free BSA on the [Ca2+]i flux in the GPR93-overexpressing hBRIE 380i cells. In contrast to undigested BSA, which did not induce a change in [Ca2+]i, digested BSA (50 mg/ml) showed a significant increase over basal (Fig. 7A, bar 4 vs. 1). The change in [Ca2+]i was less than that observed with peptone at a similar dose. This is likely due to the fact that the peptide components in peptone are derived from more than one kind of protein. Digested BSA synergistically induced [Ca2+]i release when it was combined with LPA (Fig. 7A, bar 9 vs. 2 and 5; and bar 10 vs. 3 and 5). We also utilized cefaclor, a peptidomimetic substrate of the oligopeptide transporter PepT1 (7), which did not induce [Ca2+]i release in hBRIE 380i cells (Fig. 7B, bar 5; and tested up to 5 mM; data not shown). Cefaclor synergistically enhanced GPR93 activation by peptone or LPA (Fig. 7B, bar 6 vs. 2 and 5; and bar 7 vs. 3 and 5). Enhanced activation of GPR93 by a combination of peptone and LPA was also further amplified in the presence of cefaclor (Fig. 7B, bar 8 vs. 4).
The implication of this result is that LPA and peptone might act on different sites of the GPR93 molecule. Although at the doses of peptone and LPA that were used, the effects were synergistic whether the GPR93 activation by saturated doses of peptone could be amplified by LPA and vice versa could not be determined due to the nature of the assay. The physiological significance of the synergistic effect of peptone plus LPA on GPR93 activation is that the sensitivity of endogenous GPR93 toward its activators is likely modulated depending on the composition of the dietary components that are in contact with enterocytes in situ. The effects of digested BSA or cefaclor on GPR93 activation illustrate a modulation of GPCR activities by compounds that by themselves result in lower or no activity and suggest that various components of peptone might act in concert, leading to the observed net activation of GPR93.
GPR93 is expressed the most in the small intestinal mucosa and in postconfluent hBRIE 380i cells.
GPR93 mRNA levels in rat tissues were determined by semiquantitative RT-PCR (Fig. 8A). GPR93 is highly expressed in the intestine. The level of GPR93 in the hypothalamic region of the brain was ∼10% of that in the duodenum mucosa. The GPR93 was also detected in kidney, spleen, and heart. Like other GPCRs found in the intestine, GPR93 was found in both neuronal as well as nonneuronal tissues. The significance of GPR93 expression in muscle tissues will be more readily elucidated when the characterization of endogenous ligands or activators is resolved. Given that the distribution of GPR93 in the intestine might be similar to that of other GPCRs found in the intestine and that they are present in both neuronal and nonneuronal tissues, it also remains to be determined whether GPR93 is present in the myenteric plexus or vagal nerves where it could also serve as a signal transducer of luminal activators. The level of GPR93 in the mucosal layer was three times higher than that in the muscle layer of the respective section; the most abundant was in the duodenal mucosa with a slight progressive decrease toward the colon. GPR93 in the stomach mucosa was at a lower level than in the intestinal mucosa. We further characterized the pattern of GPR93 expression in the intestinal mucosa by LMD in both the area of cell proliferation (crypt) and the area of cells that are nonproliferative and terminally differentiated (villus) (Fig. 8B). The growth stage of the laser microdissected cells from tissue was confirmed by the expression of villin (an early differentiation marker) and I-FABP (a late differentiation marker). GPR93 mRNA level was the highest in the villar tip in an area corresponding to the brush border, which is composed mostly of enterocytes (Fig. 8C). This in situ distribution corresponded with the pattern of expression in the hBRIE 380i cell line. GPR93 expression pattern in hBRIE 380i cells was determined by RT-PCR when the cells were at the state of near/early confluency and postconfluency (Fig. 8D). Villin expression was used to verify the state of differentiation. Consistent with the result of the in vivo tissue distribution, GPR93 transcript was detected at near confluency (similar to crypt cells), and its level continued to rise after confluency in parallel to the increase in the degree of cell differentiation (equivalent to cells toward the villar tips of microvilli). The hBRIE 380i cells, which are nontumorigenic small intestinal cells, after formation of a monolayer, form differentiated clusters of cells that resemble the polarized columnar epithelia of the intestinal villi (2, 25). The high level of GPR93 expression in differentiated hBRIE 380i cells explains the observed induction of [Ca2+]i in the presence of peptone or LPA (Fig. 6B) and also suggests that, in the intestine, the activation of GPR93 by exogenous agents would be the highest in the cells of the brush border or villar tips.
In the present study we identified GPR93 as a highly expressed GPCR in the rat intestinal mucosa and demonstrated its activation by protein hydrolysate. GPR93 was initially identified by O'Dowd and coworkers (39) in customized searches of the high-throughput GenBank genomic sequences databases using previously known GPCR-encoding sequences as queries. GPR93 shares amino acid identity with purinergic-like receptors, 31% identity with P2Y5, 28% with P2Y9, and 25% with P2Y10 (29). Despite being classified as purinoreceptor-like, in this study GPR93 could not be activated by nucleotides and their derivatives. GPR93 also shares 32% amino acid identity with free fatty acid receptor GPR40 but could not be activated by free fatty acids. Protein hydrolysate activated GPR93, which in turn induced [Ca2+]i through Gαq- and Gαi-mediated pathways.
The best characterized intestinal epithelial cell response to protein hydrolysate is that of the CCK endocrine cells (43). Protein hydrolysate, which mimics dietary proteins digest in the intestinal lumen (12), induces CCK expression (11) and stimulates transcription and the release of CCK (48) in the STC-1 enteroendocrine cell line. In STC-1 cells, peptone stimulates CCK secretion through a PTX-sensitive pathway, involving the Rab3A small G protein and intracellular Ca2+ (20, 48). PTX decreases the CCK release induced by peptones or peptidomimetic cephalosporins, indicating a Gαi-mediated pathway (48), thus raising the possibility that a protein hydrolysate responsive GPCR such as, or similar to, GPR93 might be involved.
PepT1, which is predominantly located in the apical membrane of intestinal epithelial cells (22), is a proton/oligopeptide cotransporter that takes up peptidomimetic cephalosporins (48) coupled to the entry of extracellular Ca2+ to the cells (14). Unlike the peptone response through PepT1, which is not expressed in hBRIE 380i cells, induced [Ca2+]i as a result of the peptone response through GPR93 was from intracellular Ca2+ stores, indicated by nifedipine insensitivity and blockage by thapsigargin. Although PepT1 is not expressed in hBRIE 380i cells, the possibility remains for a transporter linked to GPR93 activation in response to protein hydrolysate.
In STC-1 cells, peptidomimetics increase c-fos expression via the p42/44 mitogen-activated protein kinase pathway resulting in increased expression of the luciferase gene reporter linked to AP-1 responsive elements (TRE) (45). Protein hydrolysate activation of GPR93 in hBRIE 380i cells induced ERK1/2 phosphorylation. We also observed increased activity of luciferase expressed from a reporter construct linked to TRE. Whether the induced reporter activity was due to the increased synthesis or activation of the transcription factor(s) remains to be determined.
The activation of the ERK1/2 pathway is a common characteristic of many activated GPCRs. The Gαi subtype of GPCRs can activate ERK1/2 through the direct interaction of the Gβγ subunits with p21ras and the Raf-1 kinase complex, which act upstream of ERK1/2 (21, 34, 46). Pretreatment of hBRIE 380i cells with the Gαi inhibitor PTX inhibited protein hydrolysate-induced ERK1/2 phosphorylation. The PLC-β inhibitor U-73122 also inhibited the response of GPR93 activation in hBRIE 380i cells, indicating the role of [Ca2+]i flux in the activation of ERK1/2 in response to GPR93 activation. These data suggest the involvement of not only the βγ dimer from the Gαiβγ trimer but also Gαq in increased ERK1/2 phosphorylation due to GPR93 activation.
Like P2Y9, which was reported to be an LPA receptor (50), GPR93 was also activated by LPA. This activation led to induced [Ca2+]i as well as ERK1/2 phosphorylation. The possibility that the pathways resulted from GPR93 activation by protein hydrolysate and that LPA might not be exactly the same was suggested by the synergistic enhancement of GPR93 activation in the presence of protein hydrolysate and LPA as a mixture and also by differential activation of the NFAT and TPA response element-linked reporter gene.
Recently, during the preparation of this article, Kotarsky et al. (37) reported that GPR92 (the same as GPR93) binds LPA and increases cAMP levels. More recently, during the submission process of this article, Lee et al. (38) reported that GPR92/93 binds LPA in rat neuroblastoma B103 cells. Their data also suggest that GPR92/93 response is mediated by Gα12/13. It is possible that, in our cell system, GPR93 response could also lead to Gαs- and Gα12/13-mediated pathways. In contrast to these two studies, since significant expression of GPR93 was observed in the small intestine, we examined the activation of GPR93 in a nontumorigenic enterocyte cell line.
The presence of nutrients in the intestinal lumen is necessary for the structural and functional integrity of the mucosa. Genes that are expressed in intestinal epithelial cells, which are in direct contact with the luminal content, are potential targets for dietary control of transcription. Most studies have been centered on the extraluminal effects of nutrient compounds, such fatty acids and glucose, in the transcriptional control of genes encoding proteins that play significant roles in lipid and glucose transport or metabolism in hepatocytes, adipocytes, or pancreatic β-cells (15, 19, 60), rather than these nutrient molecules as ligands of transmembrane receptors. Mucosal alterations resulting from the exclusion of dietary components (23, 41, 64) are exemplified in patients that receive TPN, for whom food is supplied intravenously, thus completely bypassing the intestinal lumen. TPN in animals induces an atrophy of the mucosa in which both villus and crypt are affected (5, 30). TPN in human adults leads to mucosal remodeling of extracellular matrix, such as the distribution of laminin, fibronectin, and tenascin, and expression of collagen IV along the crypt-villus axis. These changes in ECM are also associated with a decrease in both villar cell shedding and apoptosis (23). ERK1/2 might play a role in this process since elevated ERK1/2 activities stimulate proliferation of intestinal cells, whereas low sustained levels of ERK1/2 activities correlate with G1 arrest and enterocyte differentiation (6).
The identification of GPR93 in the intestinal mucosa, which is activated by protein hydrolysate followed by the subsequent activation of ERK1/2, suggests that GPR93 might be a link between the presence of luminal dietary protein hydrolysate and the ERK1/2 signaling. Prominent subcellular localization of ERK1/2 with upstream modulators, such as Ras, p85 (phosphatidylinositol 3-kinase), Rac1, and MEK1, in the brush border of differentiated human enterocytes suggests that one of the sites of action of ERK1/2 signaling is at the apex of these epithelial cells in the brush border where ERK1/2 activity could also be responsive to changes in luminal contents in response to feeding. Although it remains to be determined whether GPR93 colocalizes with microvillar ERK1/2, the identification of this receptor in the brush border of the intestinal mucosal epithelial cells raises the possibility that GPR93 might be partly responsible for the activation of ERK1/2 observed in response to luminal factors, such as dietary protein hydrolysate, which results in the ERK1/2-dependent changes along crypt-to-villar length and in the cellular absorptive capacity (6).
The physiological significance of the synergistic effect of peptone plus LPA on GPR93 activation is that the sensitivity of endogenous GPR93 toward its activators is likely modulated by the composition of dietary components that are in contact with enterocytes in situ. This is further exemplified by the action of the peptide mimic cefaclor, which by itself did not activate GPR93 as determined by the lack of Ca2+ response but significantly enhanced GPR93 activation by protein hydrolysate and LPA. Unlike the receptors expressed in enterocytes under cell culture conditions, receptors expressed in enterocytes in the mucosa are not likely exposed to a signal activator in exclusion of other extracellular molecules that could alter its activity. Our observation that the activation of GPR93 by LPA was significantly attenuated in the presence of intact BSA is consistent with what was reported for the LPA endothelial cell differentiation gene 7 receptor, the activation of which by LPA is completely blocked with 1% BSA (26). The presence of factors that could contribute to the negative or positive regulation of GPR93 in response to activators, such as LPA or protein hydrolysate, likely varies depending on the physiological state of the organism or the receptor expressing tissue. The effects of cefaclor also illustrate that, physiologically, the activation of a GPCR may not be determined by the presence of an isolated agonist since the cell in which that GPCR is expressed encounters various compounds, both agonists and nonagonists. This is particularly important for enterocytes that are exposed to external luminal molecules.
Although protein hydrolysate stimulates GPR93, protein hydrolysate may also act as a receptor modulator in response to a specific agonist. The mechanism for the activation of GPR93 by the components of protein hydrolysate remains to be determined. GPCR activation involves the relaxation of constraining intramolecular interactions and the formation of new interactions involving specific movements of the transmembrane helixes (4, 17, 32). In response to different interactions with agonists, antagonists, and inverse agonists, GPCRs have the potential to adopt multiple activated conformational states (18, 51). Each of these states may have different affinities for ligands, G proteins, and the proteins that control receptor internalization (16, 62). Therefore, the components of protein hydrolysate provide many possible receptor-protein hydrolysate interactions that might lead to the observed net activation of GPR93. The ability of GPCRs to act as a toggle switch when activated (55), as well as their ability to integrate its response to a multitude of extracellular and intracellular effectors, and to deliver a graded signaling response make these seven transmembrane receptors suitable candidates for molecular sensors that respond to the luminal content as a cocktail of activators.
The present study is the first to identify a GPCR in the enterocytes that could be partly responsible for the reported mucosal responses to luminal protein hydrolysate, such as ERK1/2 activation, endocrine cell secretions, and cell proliferation and differentiation. Future studies may reveal the possibility that luminal molecules derived from the digestion of dietary proteins act to stabilize an active receptor conformation by binding to segments of the extracellular loops or act analogous to ago-allosteric modulators, i.e., exogenous compounds acting both as agonists themselves and as enhancers for endogenous agonist(s), thereby influencing the potency of that endogenous agonist activity (56).
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-58592 (to G. W. Aponte).
We thank Drs. Steven E. Ruzin and Denise E. Schichnes (Biological Imaging Facility, University of California at Berkeley) for generously providing expert technical help with LMD and confocal microscopy. We acknowledge the use of the GPCRDB information system (http://www.gpcr.org/7tm/).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society