Location-biased activation of the proton-sensor GPR65 is uncoupled from receptor trafficking.
Free full text in Europe PMC
Abstract
Free full text
Location-biased activation of the proton-sensor GPR65 is uncoupled from receptor trafficking
GPR65 is a prototype for proton-sensing G protein–coupled receptors (GPCRs), which are highly overexpressed in many solid tumors and are emerging as attractive targets to treat cancer as they are thought to respond to acidic microenvironments and modify tumor signaling and the immune response. Despite their physiological importance, how acidic environments change the signaling patterns of proton-sensing GPCRs is still not known. This study shows that acidic extracellular environments, rather than simply switching GPR65 on, change the spatial pattern of GPR65 signaling by biasing signaling to the cell surface. These findings provide insights on physiological proton-sensing that are important to understand the function of this highly relevant but understudied family of receptors.
The canonical view of G protein–coupled receptor (GPCR) function is that receptor trafficking is tightly coupled to signaling. GPCRs remain on the plasma membrane (PM) at the cell surface until they are activated, after which they are desensitized and internalized into endosomal compartments. This canonical view presents an interesting context for proton-sensing GPCRs because they are more likely to be activated in acidic endosomal compartments than at the PM. Here, we show that the trafficking of the prototypical proton-sensor GPR65 is fully uncoupled from signaling, unlike that of other known mammalian GPCRs. GPR65 internalizes and localizes to early and late endosomes, from where they signal at steady state, irrespective of extracellular pH. Acidic extracellular environments stimulate receptor signaling at the PM in a dose-dependent manner, although endosomal GPR65 is still required for a full signaling response. Receptor mutants that were incapable of activating cAMP trafficked normally, internalize and localize to endosomal compartments. Our results show that GPR65 is constitutively active in endosomes, and suggest a model where changes in extracellular pH reprograms the spatial pattern of receptor signaling and biases the location of signaling to the cell surface.
Activation and membrane trafficking are tightly coupled for all known members of the physiologically and clinically important G protein–coupled receptor (GPCR) family (1–4). GPCRs activated at the plasma membrane (PM) on the cell surface are rapidly desensitized and internalized into endosomal compartments, from where they can either recycle back to the PM or be degraded in the lysosome. This sorting determines the further responsiveness of cells to ligands (5, 6). Endosomes also serve as signaling stations for many GPCRs. The same signals originating from GPCR endosomes can produce distinct downstream consequences from those originating in the PM (4, 7–10). These observations have led to an emerging model that the GPCR signaling is spatially encoded, where the integrated GPCR response is a balance of both surface and internal signals, determined by rates of trafficking of receptors to and from the PM.
Proton-sensing GPCRs present an interesting family of physiologically important GPCRs (11). The canonical view is that these receptors sense acidic extracellular environments by protonation and coordination of several amino acids, including extracellular histidines and buried amino acid triads containing aspartic acid and two glutamic acids (11–13). There are three GPCRs—GPR4, GPR65, GPR68—currently thought of as being the primary receptors that sense pH changes (11–13). Members of this family of GPCRs are highly overexpressed in many cancers (14). Acidic environments are defining hallmarks of cancer and inflammation as well as many physiological processes (12, 15–17), and these GPCRs have generated interest as potential therapeutic targets.
The relationship between trafficking and signaling is especially interesting for these proton-sensing GPCRs. Endocytosis of proton-sensing GPCRs transports them to endosomal compartments that are acidic and more likely to activate these receptors. Considering the intrinsic signaling potential of proton sensing GPCRs in intracellular compartments, whether these receptors follow the tight coupling of trafficking and signaling that have been described for most known GPCRs, or whether they can be selectively activated in endosomal compartments, are unanswered questions that are fundamental to understanding how cells respond to pH.
Here, we addressed these questions using the proton-sensing receptor GPR65 as a prototype. We show that GPR65 internalizes from the PM irrespective of extracellular pH, and that receptor internalization is required for a full cellular signaling response. Together, our findings show that GPR65 dynamically traffics to and signals from multiple cellular compartments, and that, unlike for most known GPCRs, activation of GPR65 is uncoupled from receptor trafficking. Characterizing proton-sensing receptor trafficking and signaling will improve our understanding of how these receptors integrate responses from multiple locations in the cell, and how these responses influence physiology and disease states.
GPR65 Stimulates cAMP Accumulation at Neutral and Acidic pH.
We first determined the pH-dependent activation of GPR65, by measuring levels of the second messenger cAMP in HEK293 cells stably expressing GPR65. We used the cAMP biosensor GloSensor, which exhibits increased luminescence when bound to cAMP, to quantitatively measure cAMP levels (Fig. 1A) (18, 19). Upon exposure to proton concentrations from pH 6.0 to 8.0, GPR65-expressing cells displayed a dose-responsive elevation of cAMP levels (Fig. 1 B and C). This elevation in cAMP saturated at pH 6.4. Interestingly, cAMP levels were elevated at both acidic and neutral pH ranges. As negative controls, HEK293 cells not expressing GPR65 did not show a response (Fig. 1 D and E). Similarly, a GPR65 mutant where three key histidines —H10, H14, and H243—were mutated, which has been shown to not increase cAMP levels, did not show a cAMP response (Fig. 1 D and E) (12, 20). This GPR65-mediated cAMP increase was similar to the prototypical Gs-coupled GPCR beta-2 adrenergic receptor (B2AR), which we used as a positive control for cAMP activation (Fig. 1 D and E). When cells expressing GPR65 were treated with high pH (i.e., a low concentration of protons), GloSensor luminescence rapidly decreased below that of vehicle media at pH 7.0 (Fig. 1 B and C and SI Appendix, Fig. S1B). Both acute and chronic basic pH treatments decreased GPR65-mediated GloSensor luminescence (Fig. 1B and SI Appendix, Fig. S1B). Both the absolute cAMP response and the changes were reduced substantially, with levels staying close to baseline, in HEK293 cells not expressing GPR65 (SI Appendix, Fig. S1 A and B), indicating that the cAMP response we observed in GPR65-expressing cells were a result of GPR65 activation. In HEK293 cells, forskolin-induced cAMP responses did not change until pH 8.0, indicating that the dose response observed in GPR65-expressing cells was not a direct effect of pH on the sensor (SI Appendix, Fig. S1C). Together, these data demonstrate that GPR65 increases cAMP levels at both acidic and neutral pH.
GPR65 increases cAMP levels at both neutral and acidic pH. (A) Schematic diagram of the GPR65-GloSensor assay. Proton-dependent activation of GPR65 stimulates production of intracellular cAMP. cAMP binding to the GloSensor luciferase produces luminescence directly correlated to increased cAMP levels. (B) GloSensor luminescence over time in GPR65-expressing cells treated with pH 6.0, 6.4, 6.8, 7.0, 7.2, 7.6 or 8.0 (n = 3 biological replicates). GPR65 displays elevated cAMP levels at neutral and acidic pH. (C) Proton-dependent changes in intracellular cAMP levels were measured by a concentration–response curve (pH 6.0–8.0) in HEK293 cells stably expressing FLAG-GPR65, and GloSensor (n = 3 biological replicates, Nonlinear regression fit [agonist] vs. response–Variable slope). GPR65-mediated cAMP increase is saturable at pH 6.4. (D) GloSensor luminescence over time in HEK293 cells and HEK293 cells stably expressing FLAG-GPR65, FLAG-B2AR, and FLAG-H10F, H14F, H243F-GPR65 mutant after addition of either pH 6.4 or iso (n = 3 biological replicates). Activation of GPR65 by pH 6.4 increases cAMP levels similar to positive control B2AR. (E) Bar graph of the area-under-curve (AUC) from D (n = 3 biological replicates) (Ordinary one-way ANOVA, P < 0.05). AUC of WT GPR65 is similar to positive control B2AR and significantly higher than negative controls.
GPR65 Internalizes from the PM Irrespective of Extracellular pH.
Because GPR65 exhibited increased cAMP levels at a physiologically relevant pH range, we asked whether GPR65 was constitutively internalizing from the PM, and how the internalization compared to the prototypical B2AR. To test this, HEK293 cells expressing epitope-tagged FLAG-GPR65 were immunolabeled live (Fig. 2A) and imaged by confocal microscopy after 10 min of labeling at 37 °C. At neutral and basic pH, FLAG-GPR65 localizes to the PM and internal compartments (Fig. 2B), suggesting that surface receptors are internalized rapidly at these pH levels. When exposed to acidic pH, FLAG-GPR65 localization did not change noticeably and is again observed at the PM and internal compartments (Fig. 2B). This agonist-independent internalization pattern was distinct from the prototypical GPCR B2AR. At baseline, B2AR was localized primarily to the PM (Fig. 2C). When exposed to a saturating concentration (10 μM) of the agonist isoproterenol, B2AR internalized and localized almost entirely to internal compartments (Fig. 2C). Quantification of the number of internal receptor spots shows a similar degree of internalization of FLAG-GPR65 exposed to a range of pH values from 6.4 to 8.0, comparable to B2AR after iso (Fig. 2D). The expression levels of receptors in all cell lines were comparable, as measured by fluorescence levels (SI Appendix, Fig. S2A), confirming that at similar expression levels, GPR65 and B2AR localized to different compartments at baseline. Additionally, GPR65 tagged with the SNAP-tag at the N terminus, labeled with a membrane-impermeable SNAP dye, displayed a similar localization pattern as FLAG-tagged GPR65 (Fig. 2 E and F). Together, these data demonstrate that GPR65 constitutively internalizes from the PM and localizes to internal compartments irrespective of extracellular pH.
GPR65 internalization and endosomal localization is independent of extracellular pH. (A) Schematic diagram of the FLAG internalization assay. A fluorescently tagged antibody will bind the FLAG tag (peptide sequence DYKDDDDK) at the N terminus of the receptor. Immunolabeling live cells with the tagged antibody will allow for visualization of receptors that started at the PM when cells were labeled. (B) FLAG-tagged GPR65-expressing cells exposed to pH 7.2 (Left) and pH 6.4 (Right) and immunolabeled live and imaged by confocal microscopy. Localization of FLAG-GPR65 does not change noticeably. (C) FLAG-B2AR expressing cells immunolabeled live and imaged by confocal microscopy. After the addition of 10 μM iso, B2AR internalizes and localizes to internal compartments. (D) Quantification of the number of internal receptor spots over time (FLAG-GPR65, n = 15–25 cells/condition; FLAG-B2AR, n = 16 cells) (line indicates mean, shading is SEM). FLAG-GPR65 localization remains constant even after pH 6.4 addition. (E) SNAP-tagged GPR65-expressing cells exposed to pH 7.2 (Top) and pH 6.4 (Bottom) were immunolabeled live and imaged by confocal microscopy. Localization of SNAP-GPR65 is similar to FLAG-GPR65 (Scale bars, 10 µm). (F) Quantification of the number of internal receptor spots over time. The FLAG-GPR65 internalization pattern is conserved in SNAP-GPR65 (SNAP-GPR65, n = 15 cells; FLAG-GPR65, n = 15 cells) (line indicates mean, shading is SEM). (G) FLAG-GPR65 cells expressing either dominant-negative Dyn K44A or WT Dyn were immunolabeled live and imaged by confocal microscopy. FLAG-GPR65 cells expressing Dyn K44A exhibit decreased endocytosis while cells expressing WT Dyn displayed increased internalization and endocytosis. (H) Quantification of the ratio of internal fluorescence over total cell fluorescence over time (WT Dyn, n = 20 cells each; Dyn K44A, n = 20 cells each) (line indicates mean, shading is SEM). FLAG-GPR65 cells expressing Dyn K44A exhibit a lower internal fluorescence while cells expressing WT Dyn display increased internal fluorescence. (I) Schematic of the endocytic pathway that allows for trafficking and transfer of cargoes between membrane-bound compartments. Once a membrane protein is internalized, the protein is transferred from very-early (APPL1+) endosomes to early (EEA1+) endosomes. From EEA1+ endosomes, the protein could either be inserted back into the PM via recycling endosomes (like Rab11+ endosomes) or it could be targeted for degradation via late endosomes (like LAMP1+ endosomes). (J) Representative confocal images of HEK293 cells expressing FLAG-GPR65 exposed to pH 7.0 or pH 6.4 for 20 min prior to fixing cells and staining for EEA1 with a 488 secondary antibody (Scale bar, 6 µm). Yellow arrows denote GPR65 endosomes that colocalize with EEA1. (K) Fluorescence linear profile plots of GPR65 and EEA1, measured by lines drawn across regions of the cell with GPR65 endosomes after treatment with pH 7.0 or pH 6.4 for 20 min. Plots show EEA1 immunofluorescence increases along with GPR65 in both pH 7.0 and pH 6.4. (L) Quantitation of the percentage of GPR65-containing endosomes that colocalize with each of the endosomal markers. For this, FLAG-GPR65 cells treated with either pH 7.0 or pH 6.4 for 20 min were fixed and processed for immunofluorescence with the noted markers. GPR65 localizes in APPL1-, EEA1- Rab11-, and Lamp1-positive endosomes irrespective of pH (n = 21–23, 19–20, 13–16, and 17–28 fields for APPL1, EEA1, Rab11, and Lamp1, respectively).
To directly test whether the intracellular localization represented GPR65 that was internalized from the PM, we inhibited endocytosis prior to labeling surface receptors and tested whether this inhibition reduced GPR65 intracellular localization. To inhibit endocytosis, we expressed a dominant-negative dynamin mutant (Dyn K44A), which inhibits dynamin-mediated endocytosis (21). FLAG-GPR65 and FLAG-B2AR stable cells expressing either Dyn K44A or wild-type dynamin (WT Dyn) were labeled live using fluorescent anti-FLAG antibodies as above, and imaged using confocal microscopy. FLAG-GPR65 cells expressing Dyn K44A showed almost no intracellular puncta, while cells expressing WT Dyn showed multiple puncta similar to cells in Fig. 2 B and E (Fig. 2G). Quantification of the ratio of intracellular fluorescence over total cell fluorescence over time revealed that the fraction of intracellular fluorescence was substantially lower in GPR65 cells expressing Dyn K44A than in cells expressing WT Dyn (Fig. 2H). As controls, FLAG-B2AR cells expressing Dyn K44A, but not WT Dyn, showed a loss of intracellular fluorescence after isoproterenol (Fig. 2H). Together, these data demonstrate GPR65 internalizes from the PM in a dynamin-dependent manner irrespective of pH, and localizes to intracellular endosomal structures.
We next asked whether GPR65 was localized to a biochemically specific endosomal compartment, and whether the localization pattern of GPR65 changed between neutral and acidic extracellular pH. We treated HEK293 cells stably expressing FLAG-GPR65 with pH 7.0 or pH 6.4 for 20 min, immunolabeled live, fixed, and stained cells with markers for distinct compartments along the endosomal pathway: APPL1 (very early endosomes), EEA1 (early endosomes), Rab11 (recycling endosomes), and Lamp1 (lysosomes) (Fig. 2I). By using spot detection and colocalization analysis, we quantified the fraction of GPR65 puncta that colocalized with each endosomal marker. At pH 7.0, GPR65 localized to multiple compartments, including EEA1, APPL1, Rab11, and Lamp1 endosomes, and this localization pattern did not change after cells were exposed to pH 6.4 (Fig. 2 J–L). These results show that GPR65 is distributed across early and late endosomal compartments across neutral and acidic pH ranges, which is surprising considering that GPCR activation and trafficking are usually highly integrated (4, 7, 22, 23).
Proton-Dependent Activation of GPR65 is Uncoupled from Receptor Trafficking.
Because GPR65 was localized in internal compartments across basic and acidic pH ranges, we next asked whether internalization and endosomal localization of surface-labeled GPR65 required the receptor to be able to signal. To do this, we compared the localization of WT GPR65 to that of GPR65 mutants that were deficient in activating cAMP. In addition to the histidine mutant (H10F, H14F, and H243F) described in Fig. 1, which did not stimulate cAMP, we generated an independent GPR65 mutant deficient in cAMP signaling by mutating an arginine 112 in the DRY Motif, a conserved stretch of amino acids that governs GPCR activation and G protein coupling (24). When this mutant (R112A) was transiently expressed in HEK293 cells expressing the GloSensor cAMP sensor, at neutral pH, the mutant showed low levels of luminescence, comparable to that of HEK293 cells not transfected with the receptor (Fig. 3A). This low level was comparable to the GPR65 histidine mutant GPR65 described in Fig. 1. In contrast, cells expressing WT GPR65 showed a significantly higher level of luminescence (Fig. 3A). Exposure to pH 6.4 did not increase cAMP levels in cells expressing GPR65 R112A, or in cells not transfected with the receptor, in contrast to cells expressing WT GPR65 (Fig. 3 B and C), showing that the R112A mutant GPR65 was not capable of stimulating cAMP in response to pH (Fig. 3 A–C). Receptor expression were comparable across all cells analyzed, as measured by fluorescence levels (SI Appendix, Fig. S2B), confirming that the differences in cAMP levels were due to intrinsic differences in the ability of expressed receptors to stimulate cAMP and not due to differences in expression levels. Strikingly, under the same conditions, FLAG-R112A GPR65 and FLAG-H10, 14, 243F GPR65 mutants were both localized to intracellular endosomal structures at steady state, identical to WT FLAG-GPR65, when visualized via confocal microscopy (Fig. 3 D and E). These results show that the ability of GPR65 to be activated was not required for receptor internalization, and that GPR65 trafficking and signaling were uncoupled.
Activation of GPR65 is uncoupled from receptor trafficking. (A) Luminescence at pH 7.0 of GPR65-expressing cells compared to R112A-GPR65 DRY Motif mutant and HEK293 cells (n = 4 biological replicates) (Ordinary one-way ANOVA, P < 0.0001). At neutral pH, WT GPR65 exhibits increased luminescence while R112A-GPR65 displays lower luminescence similar to HEK293. (B) Corrected luminescence trace of GPR65-expressing cells compared to R112A-GPR65 DRY Motif mutant and HEK293 cells exposed to pH 6.4 (n = 3 biological replicates). After exposure to pH 6.4, WT GPR65 cells display increased luminescence while R112A-GPR65 DRY Motif mutant and HEK293 cells did not increase cAMP levels in response to acidic pH. (C) Bar graph of the AUC from B (n = 3 biological replicates) (Ordinary one-way ANOVA, P < 0.05). AUC of WT GPR65 is significantly higher than R112A-GPR65 DRY Motif mutant and the negative control, HEK293 not expressing GPR65. (D) FLAG-R112A and FLAG-H10,14,243F GPR65-expressing cells exposed to pH 7.2 (Top) and pH 6.4 (Bottom) were immunolabeled live and imaged by confocal microscopy. Localization of FLAG-R112A GPR65 and FLAG-H10, 14, 243F GPR65 do not change noticeably upon pH 6.4 addition. (E) Quantification of the number of internal receptor spots over time (FLAG-H10, 14, 243F GPR65, n = 25–32 cells; FLAG-R112A GPR65, n = 22–29 cells; GPR65, n = 15–25 cells) (error bars are SEM). FLAG-R112A and H10,14,243F GPR65 localization is similar to WT GPR65.
Acidic Endosomal Environments Activate Endosomal GPR65 Irrespective of Extracellular pH.
Localization of GPR65 to distinct intracellular compartments at steady-state raised the possibility that GPR65 was persistently active at acidic endosomes. To test this possibility, we first asked whether internalization of GPR65 was required for the full cAMP response. We pretreated cells with the endocytosis inhibitor Dyngo-4a (Fig. 4A) for 15 min, and measured cAMP signaling via GloSensor luminescence after exposing cells to pH 6.4. Inhibition of GPR65 endocytosis reduced the total cAMP response when compared to vehicle-treated cells (Fig. 4 B and D). This reduction in cAMP response of Dyngo-4a-treated GPR65-expressing cells was similar to iso-activated B2AR cells exposed to Dyngo-4a (Fig. 4 B–D), where endosomal B2AR was required for the full cAMP response as previously described (7, 25). As a control, Dyngo-4a pretreatment did not change forskolin-induced cAMP activation in HEK293 cells not expressing GPR65 (Fig. 4E), indicating that Dyngo-4a had no direct effect on cAMP activation.
Internalized GPR65 contributes to cAMP response. (A) Proton-dependent activation of GPR65 stimulates production of intracellular cAMP from multiple cellular compartments. Treatment of GPR65-expressing cells with endocytosis inhibitor Dyngo-4a will prevent internalization of the receptor leading to a decreased whole-cell cAMP output. (B) Corrected luminescence trace of GPR65-expressing cells pretreated with vehicle or Dyngo-4a for 15 min before addition of pH 6.4 (n = 4 biological replicates). After exposure to pH 6.4, WT GPR65 cells treated with Dyngo-4a display lower luminescence and intracellular cAMP levels when compared to vehicle condition. (C) Corrected luminescence trace of B2AR-expressing cells pretreated with vehicle or Dyngo-4a for 15 min before addition of 10 μM iso (n = 4 biological replicates). Positive control B2AR cells treated with Dyngo-4a display lower luminescence and intracellular cAMP levels when compared to vehicle condition. (D) Peak luminescence response of GPR65 and B2AR in cells pretreated with Dyngo-4a, normalized to cells without Dyngo-4a (n = 4 biological replicates) (Welch’s t test, P < 0.05). WT GPR65 cells treated with Dyngo-4a display a significantly lower whole-cell cAMP output than vehicle. This decrease in GPR65 cAMP output is similar to B2AR. (E) Forskolin (Fsk)-stimulated cAMP accumulation in HEK293 pretreated with vehicle or Dyngo-4a (n = 3 biological replicates), showing that Dyngo-4a does not directly affect cAMP accumulation. (F) Luminescence of GPR65 cells at pH 8.0 pretreated with vehicle or increasing concentrations of Dyngo-4a for 15 min, showing dose-dependent decrease in cAMP levels (n = 3 biological replicates). (G) Peak cAMP luminescence response of GPR65-expressing HEK293 cells untreated or pretreated with chloroquine (CQ) for 30 min prior to exposure to pH 6.4 (n = 3 biological replicates). CQ-treated GPR65-expressing cells display similar cAMP levels as untreated cells (Welch’s t test, P > 0.05). (H) Peak luminescence response of GPR65-expressing HEK293 cells untreated or pretreated with CQ for 30 min prior to exposure to pH 7.2 (n = 4 biological replicates). GPR65 cells pre-treated with CQ display lower luminescence and intracellular cAMP levels when compared to untreated cells (Welch’s t test, P < 0.05).
To test whether endosomal GPR65 was active and signaling even when surface receptors are inactive, we incubated cells for 2 h with basic pH (pH 8.0), treated cells with increasing concentrations of Dyngo-4a (40, 80, and 100 μM) for 15 min, and measured cAMP levels using GloSensor luminescence. At pH 8.0, GPR65-expressing cells showed basal cAMP levels that decreased as the concentration of Dyngo-4a increased, suggesting that endosomal GPR65 was active even in basic extracellular pH, when surface GPR65 was inactive (Fig. 4F). Together, these data suggest that endosomal GPR65 contributes to the persistent cAMP response observed in GPR65-expressing cells.
We next asked whether GPR65 signaling from endosomes was due to activation by protons in acidic environments of endosomes, or whether it was due to ligand-independent constitutive activity. To distinguish these possibilities, we tested whether the acidic environment in endosomes was required for GPR65-mediated cAMP activation, by pretreating cells with the endosomal deacidifying agent CQ and measuring cAMP at acidic and basic extracellular pH. CQ pretreatment for 30 min did not change cAMP levels when the extracellular pH was 6.4 (Fig. 4G), consistent with surface receptors contributing to the majority of cAMP response in this situation. In contrast, CQ pretreatment for 30 min significantly reduced cAMP levels when the extracellular pH was 7.2 (Fig. 4H), where surface receptors are inactive and endosomal receptors contribute to the majority of cAMP response.
Uncoupled Trafficking and Endosomal GPR65 Signaling is Conserved in Physiologically Relevant Jurkat T Cells.
We next asked whether the uncoupling of trafficking and signaling that we observed with expressed GPR65 in HEK293 cells was conserved in physiologically relevant cells. We focused on Jurkat T cells, an immortalized acute T cell leukemia line, which expressed endogenous GPR65 (26), as a physiologically relevant model cell line. FLAG-GPR65 expressed in these cells was localized to intracellular compartments at pH 7.2 (Fig. 5A). This internal localization of FLAG-GPR65 did not change noticeably upon exposing cells to pH 6.4 (Fig. 5A). Quantification of the number of internal receptor spots in Jurkat cells shows a similar level of endosomal localization of FLAG-GPR65 (Fig. 5B) as was observed in HEK293 cells (Fig. 2D).
Internalized GPR65 contributes to cAMP signaling in Jurkat T cells. (A) FLAG-tagged GPR65-expressing Jurkat cells exposed to pH 7.2 (Left) and pH 6.4 (Right) were immunolabeled live and imaged by confocal microscopy (Scale bar, 17 µm). Localization of FLAG-GPR65 does not change noticeably upon pH 6.4 addition. (B) Quantification of the number of internal FLAG-GPR65 spots at pH 7.2 and pH 6.4 in Jurkat cells, showing that the number of internal puncta do not differ between the two conditions (n = 16) (Welch’s t-test, P > 0.05). (C) Corrected luminescence trace of FLAG-GPR65-expressing cells pretreated with vehicle or Dyngo-4a for 15 min before the addition of pH 6.4 (n = 3 biological replicates). After exposure to pH 6.4, GPR65-expressing cells treated with Dyngo-4a display lower luminescence and intracellular cAMP levels when compared to vehicle conditions. (D) Corrected luminescence trace of Jurkat cells expressing endogenous GPR65 pretreated with vehicle or Dyngo-4a for 15 min before addition of pH 6.4 (n = 3 biological replicates). After exposure to pH 6.4, Jurkat cells treated with Dyngo-4a display lower luminescence and intracellular cAMP levels when compared to vehicle conditions. (E) Peak luminescence response of FLAG-GPR65 and endogenous GPR65-expressing Jurkat cells with or without Dyngo-4a pretreatment (n = 3 biological replicates) (Welch’s t test, P < 0.05). GPR65-expressing cells treated with Dyngo-4a display a significantly lower whole-cell cAMP output than vehicle. (F) A model for location-switching of GPR65 signaling triggered by extracellular pH. WT GPR65 is constitutively (i.e., in a ligand- and activation-independent manner) internalized and localized to internal compartments even when the extracellular environment is at a basic pH. In the acidic environment of the endosome, endosomal GPR65 becomes activated and tonically increases cAMP levels. An acidic extracellular environment activates surface GPR65 increasing total cAMP levels, and switching the majority of cAMP production to the PM.
To test whether the endosomal pool of GPR65 contributed to the cAMP response in Jurkat cells, we pretreated Jurkat cells expressing FLAG-GPR65 with the endocytosis inhibitor Dyngo-4a for 15 min, and measured GloSensor luminescence before and after shifting cells to pH 6.4. Inhibition of GPR65 endocytosis significantly reduced the total cAMP response in Jurkat cells expressing FLAG-GPR65 (Fig. 5 C and E). Exposing Jurkat cells expressing endogenous GPR65 to pH 6.4 showed a cAMP response with more rapid kinetics. Importantly, inhibiting GPR65 endocytosis significantly reduced this cAMP response (Fig. 5 D and E), suggesting that endogenous endosome-localized GPR65 contributes to the endogenous cAMP response in these cells.
Here, we show that GPR65 localizes to endosomal compartments and stimulates cAMP production from endosomes independently from extracellular pH changes. Surprisingly, GPR65 activity was not required for receptor endosomal localization. Our results show that endosomal GPR65 sets the basal cAMP tone, and with acidic activation of GPR65 at the PM further increasing cAMP levels (Fig. 5F). These two sources of cAMP likely result in distinct pools of cAMP with distinct cellular functions.
Our results suggest that, contrary to known examples of mammalian GPCRs, trafficking of GPR65 to endosomal compartments is fully uncoupled from receptor activation at the PM. Prototypical class A GPCRs, like B2AR, are activated by agonist at the PM inducing a cascade of events that cause receptors to internalize and localize to internal compartments. GPR65 localized to endosomes even at extracellular pH conditions that did not stimulate cAMP above baseline (Fig. 2 B–E), indicating that GPR65 internalization is ligand- and activation-independent. Ligand-independent internalization has been reported for the viral GPCR US28, which is constitutively active and is localized to internal compartments (27–29). Similarly, cannabinoid receptors or delta opioid receptors, show relatively high basal activity in the absence of added external ligands, and therefore show higher internalization in the absence of activating ligands. In these cases, receptor internalization is tightly coupled to its activation state. When these receptors are inactivated either by inverse agonists or by mutations, receptor endocytosis is substantially inhibited (30, 31). In contrast, mutations in the GPR65 DRY motif that completely block signaling (Fig. 3 A–C) have no effect on the endosomal distribution of GPR65 (Fig. 3 D and E).
How GPR65 is constitutively internalized and distributed in endosomes in the absence of activation is still not known. The conventional view is that GPCRs are unable to interact with the clathrin endocytic machinery before being activated by ligands. Ligand-binding causes G protein activation, which initiates a cascade of events that activate GPCR kinase 2 (GRK2), which phosphorylates the GPCR, allowing receptors to recruit arrestins, which act as endocytic adapters that sequester receptors in endocytic domains (32). For activation-independent endocytosis, GPR65 could recruit a different member of the GPCR kinase family (33) which does not need to be activated by a GPCR-G protein pathway. Alternatively, GPR65 could constitutively interact with arrestins or other endocytic adapters, via sequence motifs in the cytoplasmic elements, that allows ligand-independent sequestration in endocytic domains and uncoupling of activation from endocytic trafficking.
Irrespective of the mechanism, this uncoupling of GPR65 activation and endosomal localization suggests that GPR65 signaling is regulated differently from other GPCRs. Activation of most GPCRs by ligand binding, typically at the PM, switches receptors from an “off” state to an “on” state. The initial G protein activation on the PM induced by ligand binding is rapidly desensitized by phosphorylation and arrestin binding. After internalization, a second phase of G protein–mediated signaling is initiated on endosomes. Importantly, G protein signaling from the PM and endosomes, even though they activate the same second messengers such as cAMP, activate separate sets of genes downstream of signals (7, 10, 34). For example, B2AR activates second-messenger cAMP from the cell surface and endosomes, but only endosomal cAMP induces the transcription of specific genes including phosphoenolpyruvate carboxykinase 1 (PCK1) and nuclear receptor subfamily 4 group A member 1 (NR4A1) (7, 25). Gi-coupled receptors, such as opioid receptors, can also be in active conformations on endosomes after ligand-dependent activation at the PM (22). However, the consequence of endosomal cAMP are likely to be receptor specific, as inhibition of cAMP by opioid receptors from endosomes does not have an opposite effect on the same genes (22). The physiological outcome of activating a receptor, therefore, is an integrated response of these multiple phases of signaling from the surface and along the endocytic pathway, separated by time and space (35, 36). Importantly, for these known examples, because the ligand is extracellular, mechanisms exist to transport the ligand to the endosomes, either by transporters that allow movement of small ligands such as catecholamines across the membranes, or by trafficking mechanisms that internalize larger ligands such as peptides (22, 36). Unlike these, GPR65 activation in endosomes is independent of specific transport or trafficking mechanisms for secreted ligands, as the high proton concentrations are generated as part of normal endosomal acidification.
Whether GPR65 requires consistently high concentrations of protons in endosomes for signaling is an interesting question. This question is unclear even for canonical GPCRs such as opioid or adrenergic receptors, as there is no strong evidence that ligands internalize with receptors into endosomes. In our experiments, inhibition of endosomal acidification by CQ inhibited the basal cAMP signaling seen under conditions where surface GPR65 was inactive, and endosomal GPR65 was presumably the main source of cAMP (Fig. 4H). Interestingly, at acidic extracellular pH, CQ had minimal inhibitory effect on cAMP signaling (Fig. 4G). Although this experiment is confounded by observations that acidic environments can decrease the efficiency of CQ (37), the data could reflect the fact that the magnitude of surface signaling is high enough to mask relative differences in endosomal signaling caused by neutralization, or that there is a pH-independent constitutive component to GPR65-mediated cAMP activation. The latter is consistent with the presence of residual GPR65 signaling in CQ-treated cells at high extracellular pH (Fig. 4H). Overall, however, our results suggest that endosomal GPR65 activation is primarily determined by constitutive internalization of the receptor to late endosomal compartments where the acidic environment results in receptor activation and cAMP production (Fig. 5F).
Based on this unique uncoupling of trafficking and signaling, we propose a model where GPR65 is active in endosomes at all times, and where acidic extracellular environments, rather than globally turning receptors on, instead switch or bias the intensity, timing, and location of signaling to the PM (Fig. 4). At a physiological pH of 7.4, GPR65 is inactive at the PM (Fig. 1C). The steep dose–response in the pH range of 7.2 to 6.8 allows cells to rapidly switch signaling to the PM, which could induce rapidly variable signaling outcomes that depend on the conformational biases induced by the membrane environment at the PM vs. endosomes (38). Therefore, acidic environments such as those observed in solid tumors could convert “tonic” endosomal signaling by GPR65, which is physiologically beneficial for immune cells, where GPR65 is highly expressed, to “spikes” of surface cAMP signaling, as seen in Jurkat cells expressing endogenous GPR65 upon switching to acidic media (Fig. 5D).
The model suggests interesting aspects of how GPR65 activation regulates signaling in physiological systems such as in immune cells. A role for cAMP in modulating immune cells is well established, mainly downstream of adrenergic receptors expressed in both innate and adaptive immune cells (39, 40). Activation of cAMP can drive the secretion of selected cytokines and reduce inflammatory responses and infiltration by innate immune cells. However, cAMP also inhibits immune cell activation and proliferation, chemokine-dependent cell migration, and secretion of other cytokines and interferons (40, 41), which could collectively inhibit the effectiveness of immune cells in tumor clearance. Immune cells, as they infiltrate different environments are exposed to different extracellular pH, like in the acidic environment in solid tumors. It is possible that the baseline level of tonic cAMP signaling, via constitutive GPR65 signaling from endosomes, is critical for maintaining normal function of immune cells, and that spikes of surface signaling via adrenergic agonists and GPR65 enable rapid and specific responses, depending on the precise cell type and immune environment. The model for proton-sensing, suggested by our results, opens an exciting area in understanding GPR65 and the understudied family of proton-sensing receptors (42). Whether GPR4 and GPR68 also follow a similar paradigm of uncoupled signaling and trafficking leading to location biased signaling, and how this uncoupling is important for the function of these receptors in physiology, are important and exciting areas to pursue in the future.
Cell Culture and Transfection.
Cell lines used were validated, and cells were purchased from ATCC. Cells in the lab were routinely tested for Mycoplasma contamination, and only uncontaminated cells were used and maintained at 37 °C with 5% CO2. Stable clonal HEK293 cells expressing either GPR65, H10,14,243F GPR65 or B2AR N-terminally tagged with FLAG were cultured in DMEM high glucose (Cytiva, SH3024301) supplemented with 10% fetal bovine serum (FBS; Gibco, 26140079). Stable cell lines (GPR65, H10,14,243F and B2AR cells) expressing one of the constructs were generated using Geneticin (Gibco, #10131035) as selection reagent. All stable cell line plasmid transfections were conducted with Effectene (Qiagen, #301425) as per the manufacturer’s instructions. HEK293 cells were also transiently transfected with GPR65, H10,14,243F or R112A GPR65 fused to Flag on its N terminus using Effectene as per manufacturer’s protocol. Jurkat cells, gifted by Dr. Adam Courtney and Yating (Christina) Zheng, were cultured in RPMI-1640 medium (Gibco, #A1049101) supplemented with 10% fetal bovine serum (FBS) and 2 mM glutamine. Jurkat cells were transiently transfected at 90% confluency according to the manufacturer’s guidelines with TransIT-LT1 (Mirus, #MIR2300) with 1.5 μg of each DNA construct to be expressed.
DNA Constructs.
FLAG-GPR65 construct consists of an N-terminal signal sequence followed by a FLAG tag followed by the human GPR65 sequence in a pcDNA3.1 vector backbone. To create SNAP-GPR65, the receptor sequence was amplified from the FLAG-GPR65 construct by PCR with compatible cut sites (BamHI and XbaI) and ligated into a pcDNA3.1 vector containing an N-terminal sequence, followed by a SNAP tag. FLAG-H10,14,243F GPR65 and FLAG-R112A GPR65 DRY mutant were created with a full-length receptor sequence gene block from Integrated DNA Technologies (IDT) with restriction sites (AgeI and XbaI) compatible to FLAG-GPR65 vector backbone and ligated into a pcDNA3.1 vector containing an N-terminal sequence, followed by a FLAG tag. FLAG-B2AR construct was described previously (7). WT Dyn and Dyn K44A were gifts from Adam Linstedt. pcDNA3.1 empty vector was a gift from Drs. Alan Smrcka and Hoa Phan.
Reagents.
Leibovitz L15 imaging medium (Gibco, #21083-027) was used as the vehicle (pH 7.0) to deliver the desired pH since the buffered medium covers a wide pH range. The pH was adjusted by adding either 0.1 M HCl (Fisher Scientific, A144S-500) or 1 M NaOH (Fisher Scientific, #S318-500) and measured using pH test strips (Fisher Scientific, #13-640-502). Isoproterenol (iso, #I5627) was purchased from Sigma Aldrich and used at 10 μM from a 10 mM frozen stock. Dyngo-4a was purchased from ApexBio (#B5997), dissolved in DMSO (Fisher Scientific, #BP231-100) and used at the noted concentrations. CQ was purchased from ApexBio (#B5997), dissolved in DMSO and used at 4 μM. Mouse anti-FLAG M1 monoclonal antibody (Sigma Aldrich, #F3040) conjugated to Alexa 647 (Invitrogen, #A20173) and SNAP-surface dye 647 (NEB, #S9102S) were purchased from Sigma Aldrich, Invitrogen and New England BioLabs, respectively. anti-APPL1 (1:200; CST, D83H4, #3858S), anti-EEA1 (1:50; CST, C45B10, #3288), anti-Rab11 (1:50; CST, C45B10, #3288) or anti-LAMP1 (1:100; CST, D2D11 XP, #9091) rabbit monoclonal antibodies were purchased from Cell Signaling Technology (CST). Alexa 488 goat anti-rabbit secondary antibody (1:1000; #A11008) was bought from Invitrogen.
GloSensor cAMP Assay.
HEK293 cells (5–7 × 104 cells per well) were plated in a 96-well plate (Costar Corning, #3917) coated with poly-D-lysine (Sigma Aldrich, #P6407) to allow for adherence of cells. The following amounts of DNA were used per well: 60 ng of pGloSensor-22F cAMP plasmid (Promega, E2301), 100 ng of receptor or empty vector (control, pCDNA3.1+). Reverse transfection was performed using Effectene. Twenty-four hours after transfection, cells were washed once with Leibovitz L15 medium (Gibco, #21083-027), and 100 μL of 500 μg/mL D-luciferin (Goldbio, LUCK-1G) in Leibovitz’s L-15 medium was added for 2 h at room temperature. Luminescence was measured for 30–50 min using a Varioskan LUX multimode microplate reader. For basic pH experiments, cells were treated with 5 μM Forskolin (Sigma Aldrich, #F3917). For Jurkat cells, wells in a 96-well plate were coated with Collagen IV (Sigma-Aldrich, #C5533) to allow for adherence. The following amounts of DNA were used per well: 60 ng of pGloSensor-22F cAMP plasmid, 100 ng of receptor or empty vector (control, pCDNA3.1+). Reverse transfection was performed using TransIT-LT1. Twenty-four hours after transfection, cells were washed once with Leibovitz L15 medium, and 100 μL of 500 μg/mL D-luciferin in Leibovitz’s L-15 medium was added for 2 h at room temperature. Luminescence was measured for 30–50 min using a Varioskan LUX multimode microplate reader. Raw luminescence values or the values corrected to baseline before acute changes in pH are noted as described.
Live-Cell Imaging.
All confocal live-cell imaging was conducted using an Andor Dragonfly multimodal microscopy system (Andor). HEK293 cells were plated onto 25-mm coverslips (Electron Microscopy Sciences, #50949050) coated with poly-D-lysine (Sigma Aldrich, #P6407) to allow for adherence. Two days later, cells were labeled with M1-647 antibody (1:1000) for 10 min and imaged in Leibovitz L15 imaging medium at 37 °C in a CO2-controlled imaging chamber, using a spinning disk confocal microscope (Andor, Belfast, UK) and a 60× objective. Confocal images were acquired using an iXon+ 897 electron-multiplying charge-coupled device camera (Andor, Belfast, UK) and solid-state lasers of 488 nm or 647 nm. Laser excitation levels, exposure, and other acquisition conditions were identical for each set of experiments where we directly compared cells.
Immunofluorescence of Endosomal Markers.
HEK293 cells stably expressing FLAG-GPR65 were plated to poly-D-lysine (Sigma Aldrich) coated 12-mm glass coverslips (Fisher Scientific, #1254580P) and grown for 24–48 h at 37 °C. Cells were labeled with M1 647 antibody (1:1000) for 10Â min, exposed to either pH 7.0 or pH6.4 for 20 min at 37 °C, then fixed with 4% formaldehyde (FB002, Invitrogen) for 20Â min at room temperature. Cells were rinsed with wash solution (PBS containing 1.25 mM calcium chloride, 1.25 mM magnesium chloride, with 5% FBS, 5% 1 M glycine) twice and then blocked in PBS containing 1.25 mM calcium chloride, 1.25 mM magnesium chloride, with 5% FBS, 5% 1 M glycine, and 0.75% Triton X-100. After, FLAG-GPR5 cells were incubated with either rabbit anti-APPL1 (1:200; CST, D83H4, #3858S), anti-EEA1 (1:50; CST, C45B10, #3288), anti-Rab11 (1:50; CST, C45B10, #3288) or anti-LAMP1 (1:100; CST, D2D11 XP, #9091) endosomal marker antibodies for 1 h. Cells were washed three times with PBS containing calcium and magnesium and then labeled with Alexa 488 goat anti-rabbit secondary antibody (1:1000; Invitrogen, #A11008) in a blocking buffer for 1 h. Cells were washed three times for 5 min and coverslips were mounted onto glass slides (Fisher Scientific, #12550123) using Prolong Diamond Antifade Mountant (Invitrogen, #P36961). Confocal imaging of cells was performed using a spinning disk confocal microscope (Andor) and 100× objective. Representative images were taken across 10–20 fields per condition. Laser excitation levels, exposure, and other acquisition conditions were identical for each set of experiments where we directly compared cells.
Image Analysis and Quantification.
Stacks and time-lapse images were collected and analyzed with either FIJI or Imaris. To measure receptor expression across cells, we analyzed images with FIJI and quantified mean cell fluorescence in a region of interest outlining each cell. Cells expressing comparable levels of receptors were selected for further analysis. We quantified intracellular receptors in two ways. We determined the total number of receptor spots in the cytoplasm of cells, using the Imaris software (Andor) spots function. To determine the internal receptor fluorescence, we also analyzed images with FIJI and quantified receptor fluorescence in a region of interest corresponding to the cytoplasm of the cell as a fraction of total receptor fluorescence. For the endosomal colocalization analysis, we acquired the percent colocalization of receptor spots with the endosomal marker over the total number of receptor-positive endosomes via the Imaris software (Andor) spots function and the MATLAB-based colocalize spots extension. Statistical tests and graphs were generated using Prism 9 (GraphPad Software). Schematics were made using BioRender.com (Toronto).
We would like to thank Dr. Hoa Phan, Dr. Alan Smrcka, Dr. Carole Parent, Dr. Maria Castro, and Dr. Wenjing Wang for their valuable feedback on this project. We also thank Yating (Christina) Zheng, Dr. Adam Courtney, Dr. Hoa Phan and Dr. Alan Smrcka for generously providing key reagents and equipment. L.M.M.R. was supported by the NSF Graduate Research Fellowship under Grant DGE 1256260. M.A.P. was supported by NIH GM117425, DA055026, and NSF Grant 1935926. D.G.I. was supported by NIH R35119518.
Author contributions
L.M.M.R. and M.A.P. designed research; L.M.M.R. and S.E.C. performed research; L.M.M.R., J.B.R., D.G.I., and M.A.P. contributed new reagents/analytic tools; L.M.M.R., S.E.C., and M.A.P. analyzed data; M.A.P. acquired funding and supervised project; and L.M.M.R., S.E.C., J.B.R., D.G.I., and M.A.P. wrote the paper.
Competing interests
The authors declare no competing interest.
This article is a PNAS Direct Submission.
All study data are included in the article and/or SI Appendix.
Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Discover the attention surrounding your research
https://www.altmetric.com/details/154404299
Article citations
Small-molecule probe for IBD risk variant GPR65 I231L alters cytokine signaling networks through positive allosteric modulation.
Sci Adv, 10(29):eadn2339, 19 Jul 2024
Cited by: 1 article | PMID: 39028811 | PMCID: PMC11259170
The ufmylation cascade controls COPII recruitment, anterograde transport, and sorting of nascent GPCRs at ER.
Sci Adv, 10(25):eadm9216, 21 Jun 2024
Cited by: 1 article | PMID: 38905340 | PMCID: PMC11192079
Beneath the surface: endosomal GPCR signaling.
Trends Biochem Sci, 49(6):520-531, 19 Apr 2024
Cited by: 1 article | PMID: 38643023
Review
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Ubiquitin-driven G protein-coupled receptor inflammatory signaling at the endosome.
Am J Physiol Cell Physiol, 326(6):C1605-C1610, 22 Apr 2024
Cited by: 0 articles | PMID: 38646783
Review
Rab family proteins regulate the endosomal trafficking and function of RGS4.
J Biol Chem, 288(30):21836-21849, 03 Jun 2013
Cited by: 20 articles | PMID: 23733193 | PMCID: PMC3724640
Funding
Funders who supported this work.
HHS | NIH | National Institute of General Medical Sciences (1)
Grant ID: GM117425
HHS | NIH | National Institute on Drug Abuse (1)
Grant ID: DA055026
NIDA NIH HHS (1)
Grant ID: R01 DA055026
NIGMS NIH HHS (1)
Grant ID: R01 GM117425
NSF | BIO | Division of Molecular and Cellular Biosciences (2)
Grant ID: 1256260
Grant ID: 1935926