|Home | About | Journals | Submit | Contact Us | Français|
G protein-coupled receptors (GPCRs), including the dopamine receptors, represent a group of important pharmacological targets. Upon agonist binding, GPCRs frequently undergo internalization, a process that is known to attenuate functional responses upon prolonged exposure to agonists. In this study, internalization was visualized by means of total internal reflection fluorescence (TIRF) microscopy at a level of discrete single events near the plasma membrane with high spatial resolution. A novel method has been developed to determine the relative extent of internalized fluorescent receptor-ligand complexes by comparative fluorescence quantification in living CHO cells. The procedure entails treatment with the reducing agent sodium borohydride, which converts cyanine-based fluorescent ligands on the membrane surface to a long-lived reduced form. Because the highly polar reducing agent is not able to pass the cell membrane, the fluorescent receptor-ligand complexes located in internalized compartments remain fluorescent under TIRF illumination. We applied the method to investigate differences of the short (D2S) and the long (D2L) isoforms of dopamine D2 receptors in their ability to undergo agonist-induced internalization.
G protein-coupled receptors (GPCRs) represent a large family of integral membrane proteins and their primary function is to transduce extracellular stimuli (e.g ligand binding) into intracellular signals1. These receptors play fundamental roles in many physiological and pharmacological processes and therefore serve as important drug targets which are addressed by more than 30% of the current drugs on the market2. Dopamine D2 receptors, which belong to the family of class A GPCRs, mediate various physiological functions and are known as valuable targets for the treatment of neuropsychiatric disorders including schizophrenia, Parkinson’s disease and drug addiction3–8.
D2 receptors exist in two splice variants, the short (D2S) and the long (D2L) isoform, which have been suggested to be localized pre- and postsynaptically, respectively9. Although different spatial distributions and functions have been suggested, the individual contributions of the two isoforms to (patho)-physiological processes remain unclear.
Following agonist stimulation, GPCRs undergo conformational changes that allow binding and activation of heterotrimeric G proteins. Subsequently, the Gα and Gβγ subunits modulate a variety of signaling events within the cell. Importantly, GPCRs undergo a highly dynamic regulation upon activation. Multiple mechanisms provide control over the specificity and extent of the cellular response. For instance, agonist stimulation of receptors at the cell surface can induce receptor desensitization and, consequently, receptor internalization through different endocytotic pathways10, 11. Agonist activation of most GPCRs leads to phosphorylation of the receptors by G protein-coupled receptor kinases (GRKs). GKRs and β-arrestins, together orchestrate receptor uncoupling from the G protein and thereby terminate G protein signaling12, 13. In many cases, β-arrestin recruitment leads to receptor internalization, which generally proceeds by clathrin-coated pits or other mechanisms of endocytosis14–16, in which the receptors are translocated away from the plasma membrane to lager intracellular vesicular or endosomal structures. Once internalized, GPCRs are either dephosphorylated and recycled back to the plasma membrane, or targeted to lysosomes for proteolysis, which results, in part, in receptor downregulation17. Interestingly, recent reports have demonstrated that canonical GPCR signaling may also occur from internalized GPCRs in endosomes18–21.
Various experimental approaches have been utilized to measure GPCR ligand binding and internalization. Among them are, in particular, radioligand binding studies, enzyme-linked immunoabsorbent assays (ELISA) and fluorescent immunocytochemistry experiments22–24. However, as a principal limitation these protocols generally require fixation and/or permeabilization of cells. Other approaches for directly measuring receptor internalization include microplate-based functional cell assays build upon enzyme fragment complementation, bioluminescence- and fluorescence-resonance energy transfer technologies (BRET and FRET)25–27. Moreover, alternative strategies based on fluorescence microscopy using fluorescent protein tags have been followed to monitor internalization in living cells28, 29.
Fluorescence microscopy facilitates a novel way to study ligand binding and the subsequent internalization process of GPCRs in living cells, using fluorescence-based probes such as fluorescent ligands30. In particular, total internal reflection fluorescence (TIRF) microscopy has proved useful to study internalization because of its ability to selectively detect both fluorescent molecules situated within or in close proximity (~100nm) to the plasma membrane with high spatial and temporal resolution18, 31, 32.
Taking advantage of our recently developed high affinity fluorescent dopamine receptor agonists and antagonists33, we visualize internalization by means of TIRF microscopy at a level of discrete single events near the plasma membrane. Employing the reducing agent sodium borohydride (NaBH4), which converts cyanine-based fluorescent dyes on the membrane surface to a long-lived reduced form, we herein present a novel method to determine the relative proportion of internalized fluorescent receptor-ligand complexes by comparative fluorescence quantification under living cell conditions. Using this method, we are able to investigate differences of the two dopamine D2 receptor isoforms, D2S and D2L, in their ability to undergo agonist-induced internalization.
Very recently, we reported on the pharmacological characterization of Cy3B-conjugated fluorescent antagonists (1a,b) and agonists (2a,b) targeting dopamine D2/D3 receptor subtypes (Fig. 1a)33. These four fluorescent ligands were shown to possess binding affinities in the low nanomolar range (Supplementary Table S1) and their suitability for live-cell fluorescence microscopy with excellent signal to noise ratios and high temporal and spatial resolution could be demonstrated33.
Initial TIRF microscopy experiments in this study showed that receptor-ligand complexes of the fluorescent antagonists 1a and 1b bound to dopamine D2S or D2L receptors stably expressed in CHO cells (10nM, for 1h at 37°C) were visible under TIRF illumination as individual, freely diffusing, diffraction limited spots. These spots were homogeneously distributed over the membrane of the living CHO cells (representative for 1b, see Fig. 1b,c, Supplementary Movie S1). Interestingly, cells incubated with the fluorescent agonists 2a,b showed an inhomogeneous labeling distribution (representative for 2b, see Fig. 1d,e, Supplementary Movie S2). In addition to the labeled receptors on the membrane surface, fluorescent puncta were observed, which were predominantly clustered. In contrast to the diffraction-limited spots of single ligand-receptor complexes (Fig. 1f), these fluorescent puncta could not be treated as diffraction-limited spots and showed confined diffusive behavior and increased intensities (Fig. 1g). Moreover, the TIRF experiments revealed that the short and long isoforms of the D2 receptors (D2S and D2L) differ in their ability to form the clustered fluorescent puncta (Fig. 1d,e). These findings could not be attributed to differences in receptor expression as the experiments were performed with monoclonal CHO cell lines with nearly identical expression levels of the D2S or D2L receptors, respectively (Bmax of 970±60 fmol mg-1 protein for D2S and 1060±60 fmol mg−1 protein for D2L)33.
To further investigate this initial observation, CHO cells stably expressing D2S receptors were pretreated with an excess of the unlabeled antagonist spiperone (10µM, 2h at 37°C), followed by the incubation with the fluorescent agonist 2b (10nM, 1h at 37°C). When these cells were imaged under TIRF illumination, neither clustered fluorescent puncta nor labeling of the receptors at the cell surface were observed, indicating that unspecific binding and uptake of ligand 2b were negligible (Supplementary Fig. S1). Thus, we hypothesized that the fluorescent agonists 2a,b were internalized into the living CHO cells in a dopamine D2 receptor specific manner.
In order to associate our initial findings, the significant agonist-induced relocalization of D2 receptors, with the GPCR internalization process, we used a classical ELISA (enzyme-linked immunosorbent assay) approach34. Therefore, the decrease of cell surface expression of FLAG-tagged D2S or D2L receptors was quantified in response to treatment with fluorescent agonists (2a,b) in transiently transfected HEK cells (Fig. 2a,b). At a concentration of 1µM, both fluorescent agonists induced internalization of dopamine D2S and D2L receptors, respectively, highly similar to a reference agonist quinpirole (10µM) (remaining surface expression for D2S: 68±5% quinpirole, 67±3% 2a, 67±4% 2b; and D2L: 71±5% quinpirole, 67±4% 2a, 64±6% 2b (mean±s.e.m)). Comparable results were also obtained when concentrations identical with those used within TIRF microscopy experiments (10-fold of K i concentration) were applied in the internalization assay (remaining surface expression for D2S: 69±7% quinpirole, 68±3% 2a, 62±5% 2b; and D2L: 69±8% quinpirole, 76±8% 2a, 66±7% 2b (mean±s.e.m), Fig. 2a,b).
Although different mechanisms ultimately leading to receptor internalization have been described, we were interested if our fluorescent agonists are able to engage the most prevalent and best characterized pathway for GPCR desensitization: the recruitment of β-arrestins to ligand-activated receptors35–37. Employing a commercially available test system based on enzyme-fragment complementation (DiscoverX PathHunter assay), we found that the agonists 2a and 2b stimulate substantial recruitment of β-arrestin-2 at D2S and D2L receptors(D2S: EC50=240±70nM, Emax=87±4% and EC50=230±60nM, Emax=75±7%, for 2a and 2b, D2L: EC50=400±60nM, Emax=83±7%, and EC50=230±30nM, Emax=96±4%, for 2a and 2b (mean±s.e.m)), while the antagonists 1a and 1b were devoid of intrinsic activity (Fig. 2c,d and Supplementary Table S1). Compared to the reference agonist quinpirole (D2S: EC50=79±9nM, Emax=100±1% and D2L: EC50=110±10nM, Emax=100±1% (mean±s.e.m)) both, efficacies and potencies were found to be slightly reduced. Together, the results from ELISA-based internalization assays and β-arrestin-recruitment studies indicate that the fluorescent agonists 2a,b are able to stimulate receptor desensitization and trafficking while the fluorescent antagonists 1a,b do not induce these processes. Thus, we hypothesized that the occurrence of fluoresccent puncta observed by TIRF microscopy in presence of the agonists 2a,b but not the antagonists 1a,b may be related to these processes.
Several studies have shown that endocytosis and trafficking mechanisms are blocked at reduced temperatures38, 39. Indeed when we decreased the incubation temperature from 37°C to ambient temperature (22–24°C), the number of intracellular fluorescent puncta was significantly reduced after treatment of D2S and D2L with the fluorescent agonist 2b (10nM, Supplementary Fig. S2). The spot density corresponding to 2b-D2L complexes at the cell surface (0.65±0.03 spots µm−2, mean±s.d. of 10 cells) was found to be highly comparable to that of antagonist 1b-D2L complexes (0.67±0.03 spots µm−2, mean±s.d. of 10 cells, Fig. 1c).
To further explore the origin of the differences in the ability of D2S and D2L receptors to form clustered fluorescent puncta upon fluorescent agonist-binding, we developed a new method allowing to distinguish receptors on the cell surface from receptors internalized to intracellular compartments in living cells. The procedure entails treatment with the reducing agent sodium borohydride (NaBH4), which converts the cyanine-based fluorescent dyes to a long-lived reduced form. Because the highly polar NaBH4 is not able to pass the cell membrane, ligands bound to internalized receptors should be not affected by the treatment. Thus, the internalized Cy3B-conjugated ligands located in internalized compartments (endosomes) should remain fluorescent under TIRF illumination (Fig. 3a).
In order to prove that the fluorescence intensity of the Cy3B-conjugated ligands is negligible after reduction with NaBH4, we performed in vitro control experiments. Thus, we added NaBH4 (final concentration 30mM) to an aqueous solution of the fluorescent antagonist 1a (1µM in dPBS) and measured an emission scan (λex=530nm) with a spectrofluorimeter after 5min. In agreement with previous reports on the reduction of cyanine dyes40, 41, the hydro-form of 1a displayed non-fluorescent properties (Fig. 3b,c).
To demonstrate the suitability of the reductive treatment procedure under living cell conditions, CHO cells stably expressing the D2S receptor were incubated with the fluorescent ligands 1a (antagonist) and 2b (agonist) at a concentration of 10nM for 1h, followed by the treatment with 30mM NaBH4 for 5min. Images of the cells were acquired before and after NaBH4-treatment under TIRF illumination (Fig. 4a). As expected, for the fluorescent antagonist 1a treatment with NaBH4 reduced the initial fluorescence corresponding to receptor surface-bound fluorescent ligands to a nearly undetectable level. However, TIRF images of 2b-labeled cells after NaBH4 treatment revealed fluorescent puncta, indicating that the fluorescence results from reduction-resistant 2b-receptor complexes localized in intracellular compartments (Fig. 4a).
The dynamics of these internalized compartments under TIRF illumination suggests that the cell treatment with NaBH4 has minor effects on the cell viability (Supplementary Movie S3). Bright field images of the NaBH4-treated cells also showed no significant changes in cell morphology or adhesive behavior.
To determine the relative extent of internalization, the reductive treatment approach was applied and evaluated by fluorescence quantification. Therefore, CHO cells stably expressing either D2S or D2L receptor were incubated for 1h with the corresponding fluorescent ligands (1a,b or 2a,b) at a concentration corresponding to the tenfold of the K i value at 37°C (Supplementary Table S1). Cells of each condition were imaged under TIRF illumination before and after NaBH4 treatment.
The mean fluorescence intensity of a fluorescent ligand-labeled cell after NaBH4 treatment was used to quantify fluorescence in the intracellular TIRF illumination field. The ratio of the intracellular fluorescence and the total fluorescence of the same cell before NaBH4 treatment was used as a measure of receptor internalization. Importantly, TIRF microscopy only visualizes the peripheral cytoplasm (~100nm from the basolateral membrane) and thus cannot be used to quantify the total amount of internalized receptors within an entire cell. However, the relative ratio of intracellular to total fluorescence can be applied to compare the internalization behavior of two different systems (e.g. receptor subtypes or ligands).
Figure 4b summarizes the results of the experiments described above. The extent of agonist-induced D2S-mediated internalization was found to be significantly higher (2a: 35.3±4.2% and 2b: 32.9±1.9% (mean±s.e.m.)) than for the D2L receptor isoform (2a: 16.3±1.5% and 2b: 19.8±1.5% (mean±s.e.m.)). The fluorescent antagonists 1a and 1b were not able to promote D2S and D2L receptor internalization (D2S-1a 1.1±0.9%, D2S-1b 1.4±1.3%, D2L-1a 1.9±0.4% and D2L-1b 0.4±1.5% (mean±s.e.m)).
There are still many open questions regarding the regulatory mechanisms involved in the downregulation of activated GPCRs including receptor internalization and signal termination18–21. Some of these questions can be addressed by fluorescence single-molecule imaging of GPCRs and their signal transducers with suitable fluorescent probes18, 42. In particular, TIRF microscopy offers the possibility of studying membrane proteins with higher spatial and temporal resolution than conventional epifluorescence or confocal fluorescence microscopy. Moreover, events occurring within the plasma membrane like receptor-dimerization33, 43, 44 or in close proximity to the membrane such as internalization can be studied in living cells under nearly physiological conditions32, 45.
Very recently, we have demonstrated that the fluorescent dopamine receptor antagonists 1a,b and agonists 2a,b can be used to directly visualize ligand binding to GPCR monomers and dimers with single molecule resolution by TIRF microscopy33. In the present study, we found that D2 expressing CHO cells labeled with the antagonists 1a,b showed a homogenous spot density and distribution, while our experiments with the agonists 2a,b revealed the formation of clustered fluorescent puncta under TIRF illumination. Using a new approach based on the treatment with the mild reducing agent NaBH4, we were able to convert fluorescent ligands bound to receptors at the cell surface to a non-fluorescent dark state40, 41. In contrast, the clustered fluorescent puncta were inaccessible for the polar NaBH4 and remained fluorescent. Since the fluorescent agonists 2a,b stimulate substantial β-arrestin-2 recruitment and a comparable degree of receptor internalization in HEK cells transiently transfected with D2S or D2L receptors, we attributed these fluorescent clusters to internalized receptor ligand complexes.
Up to date, several methods have been employed to study receptor internalization. Very frequently, radioligand binding studies, ELISA- or immunofluorescence-based techniques and microplate-based functional whole cell assays have been used to quantify the proportion of internalized receptors22–29. Each of these methods has its own advantages and limitations. Classical radioligand binding studies require the availability of radioligands with distinct polarity and thus membrane permeability24, 46 to determine the fraction of extra- and intracellular receptors. In contrast, the surface-bound fraction of a radioligand or an antibody is removed under acidic conditions in the “acid wash” method. The remaining cell-bound fraction is considered to be internalized in acid-resistant compartments such as endosomes. Although this method has been widely used to investigate ligand-dependent GPCR internalization47–50, it is not particularly suitable for dynamic analyses and fluorescence quantification, since not all compounds are washed off easily, requiring harsh treatment of the cells and cell membrane permeabilization47, 51. Other methods such as ELISA or immunofluorescence require highly specific antibodies or adequately tagged receptors, complicating or even excluding their application to native cells or tissue. In addition, cells usually have to be fixed and/or permeabilized. Although intact living cells are used in microplate-based whole cell functional assays on first hand, the detection step often requires cell lysis. Moreover, these bulk measurements preclude single-cell or subcellular resolution.
The TIRF imaging approach in combination with our high affinity fluorescent dopamine receptor agonists and antagonists is compatible with the physiological conditions of ligand-receptor interaction, allowing the investigation of D2 receptor internalization in living cells. Receptors are visualized in one single labeling step allowing for the monitoring of ligand binding and receptor trafficking. Treatment with NaBH4 represents a mild, fast and efficient method to distinguish receptors on the cell surface from those in intracellular compartments without causing obvious changes in cell morphology. The relative extent of receptor internalization can be determined by comparison of the overall fluorescence intensity of the same living cell before and after treatment with the reducing agent. Although we have used CHO cells stably expressing our receptors of interest (D2S and D2L), TIRF microscopy using fluorescent ligands is generally applicable to cells endogenously expressing GPCRs52. Limitations of the approach include the relatively low throughput and the requirement of suitable fluorescently labeled agonists. Further, it should be acknowledged that due to the low penetration depth of the evanescent field (~100nm) only a relative and no absolute quantification of receptors can be obtained.
Taking advantage of our new methodology, we have observed that the dopamine D2 receptor isoforms, D2S and D2L, stably expressed in CHO cells, show differences in their internalization behavior. While approximately 32–35% of the detected D2S receptors were found to be localized in intracellular compartments, only 16–20% of the observed D2L receptors were found to internalize upon incubation with the fluorescent agonists 2a,b for 1h at 37°C. Consistent with the present results, a number of previous studies have described that D2S and D2L receptors behave differently in their sensitivity to internalization/desensitization processes24, 34, 46. For instance, Itokawa et al. demonstrated in radioligand binding studies with stably transfected CHO cells that about 44% of surface expressed D2S receptors became unavailable for the hydrophilic radioligand [3H]sulpiride after incubation with the endogenous agonist dopamine. The internalization of D2L was not only lower (22%), but also proceeded significantly slower (half-life of decrease 19min for D2S versus 33min for D2L)46. Although to a lower overall extent, Kim et al. found a similar 2: 1 ratio for D2S (20%) versus D2L (10%) receptor internalization. In this case, internalization was determined as the decrease in [³H]spiperone binding on the cell surface of transiently transfected CHO cells and found to be dependent on the expression of β-arrestin24. Using confocal microscopy, the same group observed a higher overall formation of endocytotic vesicles in D2S-expressing cells compared to the D2L-expressing counterparts upon agonist stimulation24.
Interestingly, our ELISA-based internalization studies in transiently transfected HEK cells did not reveal significant differences between D2S and D2L receptor internalization. Similar observations have been previously described by Thibault et al. who employed a nearly identical protocol. However, this group still found enhanced D2S internalization compared to D2L upon heterologous desensitization in the same cell line34. Together these results suggest crucial influences of the employed cell types, transfection conditions and probably receptor expression levels on the highly regulated internalization process.
Importantly, D2S and D2L receptors share an extremely high sequence homology differing in only 29 additional amino acids within the third intracellular loop (ICL3) of D2L 53. Since the relatively long ICL3 has been previously reported to be important for D2 receptor desensitization and trafficking4, 34, 54, it is reasonable to assume that this region within ICL3 may attenuate receptor internalization and trafficking5, 24.
Previously, fluorescence microscopy based on immunohistochemistry staining has been used to measure the dynamics of agonist-induced D2 receptor internalization in vitro in intact cells55. Furthermore, receptor adaptations have been studied by means of simultaneous positron emission tomography (PET) and functional magnetic resonance imaging (fMRI), giving access to non-invasive methods of studying receptor internalization in vivo 56. The TIRF microscopy approach will allow to image native tissue with subcellular spatial resolution. Recent developments in the field of fluorescence microscopy will further promote the understanding of trafficking and regulation processes. As an example, multicolor imaging systems will facilitate a real time tracking of the dynamics of receptor-ligand complexes and their co-localization with various proteins involved in downstream signaling and internalization processes (e.g. agonist-receptor-G protein or agonist-receptor β-arrestin complexes) in living cells.
Chinese hamster ovary cells (CHO-K1) stably expressing human dopamine D2L 57or D2S 57 receptors were maintained in DMEM/F12 medium supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 1% penicillin-streptomycin, and 800µgmL−1 geneticin (all cell culture reagents purchased from Invitrogen/Thermo Fisher Scientific) at 37°C, 5% CO2.
The glass slide cleaning procedure was carried out as described previously33. In brief, 18 mm No. 1 glass slides (Assistent) were extensively cleaned to remove any background fluorescence. First, they were sonicated in a solution containing 12% Decon 90 for 1h. After three washes with Milli-Q filtered water, they were further sonicated in a solution of 5M NaOH for 1h and washed again three times with Milli-Q filtered water. Glass slides were then dried followed by sonication in chloroform for 1h. Cleaned glass slides were dried and stored in 100% ethanol until usage.
24h before the TIRF experiment, glass slides were placed in a 12-well plate and coated with 20µgml−1 fibronectin (Sigma-Aldrich) in sterile dPBS (without Ca2+/Mg2+), for 1h at 37°C. After coating, fibronectin was aspirated and the glass surface was rinsed once with sterile dPBS (without Ca2+/Mg2+). CHO cells stably expressing dopamine D2L or D2S receptors were seeded on coated glass slides in phenol red-free DMEM/F12 supplemented with 10% FBS and were allowed to adhere overnight at 37°C, 5% CO2.
Cells were washed twice with phenol red-free DMEM/F12 supplemented with 10% FBS, labeled with a tenfold K i value concentration of the corresponding fluorescent ligand and incubated unless otherwise stated at 37°C and 5% CO2 for 1h. Specific labeling of the fluorescent ligands was confirmed by preincubation with 10µM spiperone (a potent non-fluorescent dopamine D2 receptor antagonist) for 2h, followed by incubation with the fluorescent ligand as described above. Subsequently after labeling, cells were washed three times with phenol red-free DMEM/F12 supplemented with 10% FBS. Glass slides with labeled cells were placed in a custom-built imaging chamber (volume=500μL), washed twice with imaging buffer (137mM NaCl, 5.4mM KCl, 2.0mM CaCl2, 1.0mM MgCl2, and 10mM HEPES, pH 7.4). Finally, the imaging chamber was refilled with fresh imaging buffer and mounted immediately on a microscope stage for TIRF microscopy imaging.
TIRF microscopy was carried out as described previously33. In brief, experiments were performed at room temperature (24.0±0.3°C) on a motorized Nikon TI-Eclipse inverted microscope equipped with a 100x, 1.49 NA oil-immersion objective. Fluorescent dyes were excited using a Nikon D-Eclipse C1 laser box with 561nm laser for TIRF microscopy. The laser light was filtered by an excitation filter 561/14nm and directed by a dichroic long-pass mirror (cut-off wavelength 561nm). The emitted light was passed through an emission bandpass filter 609/54nm (Semrock Rochester), and was projected onto a water-cooled (Polar Series Accel 250 LC, Thermo Scientific) EM-CCD camera at −98°C (512×512 FT, DU-897, Andor, UK). The microscope control and image acquisition was performed by the NIS Elements software (Nikon Instruments). To ensure homogenous illumination, only the central quarter of the chip (300×300 pixel) was used for imaging analysis. The gain of the EM-CCD camera was kept constant at 300, binning at 1×1, BitDepth at 14 bits, readout speed 10MHz. Image sequences (5–500 frames) were acquired with an exposure time of 50 ms, resulting in the frame rate of 19.32 fps (frames per second).
An automated single particle tracking (ASPT) algorithm58 implemented in custom-written software, GMimPro (www.mashanov.uk), was used to identify and track individual fluorescent spots on sequential video frames to further determine their diffusive behavior. The procedure has been described previously in detail33, 43, 59.
Calculation of the mean background corrected fluorescence intensity I(t) (arbitrary units) at time t of single cells for internalization experiments were performed as described previously33. In brief, regions of interest (ROI) were drawn around the membrane of an individual fluorescent cell (ROIcell) and the background (ROIbackground) outside the cell using Fiji software57. The total mean intensity over the entire cell area (I(t) total cell) and the mean intensity over the background area (I(t) bg) were measured and I(t) were calculated as I(t)=I (t) total cell−I (t) bg.
Samples of CHO cells stably expressing D2L and D2S receptors were prepared and labeled with the corresponding fluorescent ligands as described above. Labeled cells were treated with freshly prepared 30mM NaBH4 solution for 5min and washed once with dPBS (with Ca2+/Mg2+). Finally the imaging chamber was refilled with fresh imaging buffer and image acquisition was performed under TIRF illumination at 561nm. The entire process was conducted on the microscope stage so that the same cells could be imaged before and after the reductive treatment.
The mean fluorescence intensity (background corrected) of a fluorescent ligand labeled cell after NaBH4 treatment was used to estimate internalized fluorescence as a percentage of the mean fluorescence intensity (background corrected) of a labeled cell before NaBH4 reduction. Mean values and s.e.m were calculated from 9–24 cells from at least three independent experiments using Prism 6.0 (GraphPad Software, Inc.). Unpaired two-tailed Student’s t-tests were used to determine statistical significance.
To characterize the emission behaviour of the fluorescent ligand 1a in the absence and presence of NaBH4 (30mM, 5min) emission spectra were recorded on a CLARIOstar multimode microplate reader (BMG Labtech, λex of 530nm) using black, clear bottom 96-well plates (Greiner Bio-One) and the ligand diluted in dPBS (with Ca2+/Mg2+, pH 7.4) to a concentration of 1μM.
Ligand-stimulated receptor internalization was quantified in analogy to a previously described procedure34. Briefly, HEK293T cells were transiently transfected in suspension with plasmids (pcDNA 3.1) encoding D2SR or D2LR fused to a FLAG-epitope at their N-terminus together with a plasmid encoding GRK2 (3:1 receptor to GRK2 ratio), using polyethyleneimine as transfection reagent60. Transfected cells were transferred into 48-well plates (65,000 cells/well) pretreated with Poly-D-Lysine and maintained at 37°C, 5% CO2 for 48h. After stimulation with the ligands dissolved in complete growth medium for 1h at 37°C, incubations were terminated by removal of the medium and fixation with 4% PFA (10min, room temperature). Subsequently, cells were washed with washing buffer (150mM NaCl, 25mM Tris, pH 7.5), blocked for 1h (3% skim milk powder in washing buffer) and incubated with the primary antibody (mouse anti-FLAG M2, 1:4,000, Sigma-Aldrich) for 1h at room temperature. Cells were washed twice and blocked again before incubation with the horseradish peroxidase-conjugated secondary antibody for 1h (HRP-rabbit anti mouse IgG, 1:20,000, Sigma-Aldrich). After three washing steps, 200µL of a peroxidase substrate-containing solution (2.8mM o-phenylenediamine in 35mM citric acid, 66mM Na2HPO4, pH 5.0) were added. Reactions were terminated after 30min by addition of 200µL 1M H2SO4. From each well, 200µL were transferred into a 96-well plate for absorbance readings at 492nm on a CLARIOstar microplate reader. Data were analyzed by subtraction of the background (nontransfected cells) and normalization to control conditions (growth medium with 0.1% DMSO).
The measurement of β -arrestin-2 recruitment to activated-receptors was performed utilizing the PathHunter assay (DiscoverX, Birmingham, UK) as described previously60. HEK293 cells stably expressing the EA-tagged β-arrestin-2 fusion protein (provided by DiscoverX) were transiently transfected with the ProLink(ARMS2-PK2)-tagged dopamine D2S or D2L receptors, respectively, using Mirus TransIT-293 (Mobitech, Göttingen, Germany) transfection reagent. 24h after transfection, cells were detached using Versene (Invitrogen) and 5,000 cells per well were seeded into white, clear bottom 384-well plates (Greiner Bio-One) and maintained at 37°C, 5% CO2 for 24h in assay medium. After incubation with different concentrations of test compounds (from 10−12 to 10−5M final concentration) in duplicates for 5h, the detection mix was added and incubation was continued for further 60min. Chemiluminescence was determined on a CLARIOstar microplate reader. Resulting responses were normalized to the maximum effect obtained with quinpirole (100%) and the basal response (vehicle, 0%). Dose–response curves were calculated by nonlinear regression using the algorithms of Prism 6.0.
The data that support the findings of this study are available within the Supplementary Information files and/or from the corresponding authors upon request.
This work was supported by the German Research Foundation (DFG 13/8-2, GRK1910). We thank Prof. Dr. Michel Bouvier (IRIC, Montréal, Canada) for providing the cDNA encoding the human isoform of GRK2.
A.T. designed, conducted and analyzed the TIRF experiments and wrote the manuscript. D.M. performed and analysed ELISA internalization experiments and wrote the manuscript. H.H. performed and analyzed β-arrestin recruitment experiments. J.K. contributed to the methodology. P.G. was responsible for the overall project strategy, provided project supervision and wrote the manuscript.
The authors declare that they have no competing interests.
Electronic supplementary material
Supplementary information accompanies this paper at doi:10.1038/s41598-017-11436-1
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.