SEARCH

SEARCH BY CITATION

Keywords:

  • arrestin;
  • D2 dopamine receptor;
  • desensitization;
  • G-protein coupled receptor kinase;
  • protein kinase C;
  • uncoupling

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
Thumbnail image of graphical abstract

We reconstituted D2 like dopamine receptor (D2R) and the delta opioid receptor (DOR) coupling to G-protein gated inwardly rectifying potassium channels (Kir3) and directly compared the effects of co-expression of G-protein coupled receptor kinase (GRK) and arrestin on agonist-dependent desensitization of the receptor response. We found, as described previously, that co-expression of a GRK and an arrestin synergistically increased the rate of agonist-dependent desensitization of DOR. In contrast, only arrestin expression was required to produce desensitization of D2R responses. Furthermore, arrestin-dependent GRK-independent desensitization of D2R-Kir3 coupling could be transferred to DOR by substituting the third cytoplasmic loop of DOR with that of D2R. The arrestin-dependent GRK-independent desensitization of D2R desensitization was inhibited by staurosporine treatment, and blocked by alanine substitution of putative protein kinase C phosphorylation sites in the third cytoplasmic loop of D2R. Finally, the D2R construct in which putative protein kinase C phosphorylation sites were mutated did not undergo significant agonist-dependent desensitization even after GRK co-expression, suggesting that GRK phosphorylation of D2R does not play an important role in uncoupling of the receptor.

The classical model for the uncoupling and desensitization of G-protein coupled receptors (GPCRs) involves the phosphorylation of the agonist-bound receptor by G protein coupled receptor kinases (GRK), followed by the binding of arrestin to the GRK phosphorylated agonist-activated receptor. We reconstituted D2-dopamine receptor (D2R) signaling in Xenopus oocytes to show that arrestin-mediated uncoupling of D2R from associated G proteins (Gαβγ) occurs independently of GRKs.

Abbreviations used
ARRB2

β-arrestin2

D2R

D2-like dopamine receptor

DOR

delta opioid receptor

DPDPE

D-pen2–5-encephelan

GPCR

G-protein coupled receptor

GRK3

G-protein coupled receptor kinase3

Kir3

G-protein-gated inwardly rectifying potassium channel

The D2 dopamine receptor (D2R) is a clinically important G-protein coupled receptor (GPCR). It is a major target of drugs used to alleviate symptoms of Parkinson's disease and depression (Missale et al. 1998; Neve et al. 2004) and a common property of all currently available anti-psychotic drugs is that they block D2R at therapeutic concentrations (Remington 2003). There is evidence that the cellular regulation of D2R plays an important role in the physiological responses to long-term treatment with these drugs as well as in the underlying disease pathophysiology. For example, schizophrenia patients have increased levels of a ‘high affinity’ form of striatal D2R (Seeman et al. 2006), and D2R desensitization or down-regulation could explain the loss of efficacy of receptor-targeted anti-Parkinsonian after prolonged use (Rascol and Fabre 2001).

The classical model for agonist-dependent regulation of GPCRs begins with phosphorylation of the agonist-bound receptor by a G-protein coupled receptor kinase (GRK) followed by the binding of an arrestin molecule to the receptor. Arrestin-binding can terminate receptor-mediated activation of G-protein signals and link the receptor to the cellular endocytotic machinery to facilitate receptor endocytosis (DeWire et al. 2007; Gurevich et al. 2008). Interestingly, the formation of the arrestin-GPCR complex can also initiate cellular signals that are distinct from those produced by the receptor-mediated activation of G proteins (Gurevich et al. 2008).

GRK-phosphorylated receptor residues help to trigger the conformational change in arrestin, which enables it to recognize and bind the activated receptor (Gurevich and Gurevich 2004, 2006). However, for some GPCRs, negatively charged amino acids or residues phosphorylated by other kinases can substitute for GRK phosphorylation and promote arrestin binding to the agonist-activated receptor (Xiang et al. 2001; Mukherjee et al. 2002; Galliera et al. 2004).

For D2R, the molecular events which trigger arrestin activation and binding are not clear. Serine and threonine residues within D2R, which are phosphorylated in an agonist-dependent manner, have been previously mapped. However, alanine substitution of these residues in D2R did not block D2R internalization in mammalian cells and these agonist-dependent D2R phosphorylation sites were not required for agonist-dependent recruitment of arrestin to D2R (Namkung et al. 2009a). Even more, recent study examining the role of receptor-attached phosphates in binding of arrestins to GPCRs also suggests that the arrestin interactions with D2R are largely independent of GRK phosphorylation (Gimenez et al. 2012).

It has been demonstrated that D2R that is expressed in HEK293 cells is basally phosphorylated at residues that are distinct from those that are phosphorylated after agonist treatment. Furthermore, the level of phosphorylation at these sites is enhanced after treatment with protein kinase C (PKC) activators such as phorbol esters and phorbol ester treatment can induce D2R internalization in the absence of agonist-treatment (Namkung and Sibley 2004; Thibault et al. 2011). Finally, the mutation of sites that are basally phosphorylated or phosphorylated after phorbal ester-treatment significantly decreased agonist dependent internalization of D2R (Cho et al. 2010). These studies suggest that it is phosphorylation of D2R by PKC, and not GRK, that serves as the trigger for arrestin activation and the formation of the D2R/arrestin complex.

It has been previously demonstrated that the mutation of one set of serine and threonine residues in the mu opioid receptor or the cannabinoid receptor could block internalization of the GPCRs while allowing for arrestin-mediated uncoupling of the GPCRs from coupled G proteins. Conversely, mutation of another set of serine and threonine residues blocked arrestin-mediated uncoupling while allowing for agonist-mediated internalization of the GPCR to occur (Jin et al. 1999; Celver et al. 2004). A related finding—that arrestins can form distinct complexes with the same receptor depending on the position of phosphorylation sites on the receptor was also established for N-formyl peptide receptor (Key et al. 2003). Thus, phosphorylation at distinct sites may determine whether arrestin binding uncouples the receptor from G protein or promotes internalization of the receptor instead. Desensitization of D2R-elicited cellular responses could result from either GPCR-G protein uncoupling and/or from receptor internalization. As these processes occur concurrently in mammalian cells it has not been established if the formation of the arrestin-D2R complex can uncouple D2R from associated G proteins in addition to allowing for receptor internalization.

GPCRs expressed in mammalian cells can simultaneously undergo uncoupling, internalization, and recycling that could either be dependent on, or independent of endogenously expressed GRKs and arrestins (Pak et al. 1999; Shapira et al. 2001; Qiu et al. 2003; Rasmussen et al. 2004). Thus, a multitude of processes acts concurrently in mammalian cells to regulate GPCRs. Furthermore, the endogenous expression, in mammalian cells, of numerous regulatory molecules including arrestins and kinases such as GRKs make it difficult to clearly define and study the separate cellular processes that are involved in regulating D2R.

Electrophysiological recordings of GPCR-mediated activation of co-expressed G-protein gated inwardly rectifying potassium channels (Kir3) reconstituted in the Xenopus oocyte expression system have been used extensively to specifically examine the effects of GRKs and arrestins on agonist-dependent GPCR-G protein uncoupling (Kovoor et al. 1997; Appleyard et al. 1999; Jin et al. 1999; Celver et al. 2004). A convenient property of the Xenopus oocyte expression system is that it lacks functional expression of endogenous GRKs (see Figure S2) and arrestins, and furthermore, exogenous expression of GRKs and arrestins can specifically reconstitute agonist-mediated GPCR uncoupling without producing significant receptor internalization (Kovoor et al. 1997; Appleyard et al. 1999; Jin et al. 1999; Celver et al. 2004)(see Figure S1). In this study, we have utilized the Xenopus oocyte expression system to examine the effects of GRK and arrestin co-expression on the agonist-mediated uncoupling of D2R and the delta opioid receptor (DOR) with the goals of (i) more clearly identifying the molecular triggers required for functional interaction between D2R and arrestin, and (ii) asking if the determinants implicated in the arrestin-mediated internalization of D2R in mammalian cells are the same as those required for D2R uncoupling.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Chemicals and reagents

Chemicals and reagents were purchased from Sigma Aldrich (St. Louis, MO, USA), from Fisher Scientific (Waltham, MA, USA) or from the suppliers specifically identified below.

cDNA clones and cRNA synthesis

Extracellular N-terminal FLAG tagged clones of the D2R and the DOR were obtained from Dr Ken Mackie and have been previously described (Kearn et al. 2005). The G-protein gated inwardly rectifying potassium channel subunits, GIRK1 (Kir3.1) and GIRK4 (Kir3.4), have been previously described (Kovoor et al. 1997). mMESSAGE MACHINE kits (Ambion, Austin, TX, USA) were used to generate capped cRNA from cDNA templates, which contained either T7 or T3 promoters to direct synthesis of sense transcripts. cDNA templates for DOR, D2R, the DORD2R3rd-loop chimera (described below), Kir3.1, and GRK3 were generated by linearizing the corresponding cDNA-containing plasmid vectors, while the cDNA templates for β-arrestin2 (ARRB2), Kir3.4, the D2RT225,S228,S229[RIGHTWARDS ARROW]A and D2R212–215(IYIV)[RIGHTWARDS ARROW]A mutant constructs (described below) were generated by PCR using oligonucleotide primers that added a 5′ T7 promoter and a 3′ poly-A (25 nucleotides long) tail.

Construction of D2R mutants

PCR based overlap-extension mutagenesis strategies were used to generate the mutant D2R cDNA constructs starting with the FLAG-tagged human D2R long-form construct. The mutant D2RT225,S228,S229[RIGHTWARDS ARROW]A construct was one in which the threonine and serine residues at position 225, 228, and 229 of the human D2R long-form were substituted with alanines. D2R212–215(IYIV)[RIGHTWARDS ARROW]A refers to a mutant construct in which amino acids from position 212 to 215, Ile, Tyr, Ile, and Val, respectively, were substituted with alanines. The cDNA templates for the synthesis of the above two mutant cRNAs were produced by amplification with primers designed to introduce a T7 promoter and a poly A tail. A similar PCR based approach was utilized to construct the chimeric DOR construct (DORD2R3rd-loop) in which the DOR 3rd cytoplasmic loop was replaced by the 3rd cytoplasmic loop from D2R. This construct was subcloned into pCDNA3.1 (Invitrogen, Carlsbad, CA, USA) which contains a T7 promoter for in vitro RNA synthesis. All the mutant D2R cDNA constructs were checked by sequencing.

Oocyte culture, injection, and electrophysiology and calculation of GPCR desensitization rates

The use of Xenopus oocytes for reconstituting coupling of GPCRs to Kir3 channels and for assessing the effects of GRKs and arrestins on such coupling has previously been described (Kovoor et al. 1997; Celver et al. 2004). Briefly, cRNA encoding the relevant proteins was injected into Xenopus oocytes at a volume of 50 nl/oocyte using a microinjector (Drummond Scientific, Broomall, PA, USA). Oocytes were maintained in a saline buffer (96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1 mM CaCl2, and 5 mM HEPES, pH 7.5) solution supplemented with 2% (v/v) bovine serum albumin, sodium pyruvate (2.5 mM) and gentamycin (20 μg/mL). Two-electrode voltage clamp recordings of the Kir3 currents expressed in the oocytes were performed 36–72 h after cRNA injection. Membrane potential was clamped at −80 mV using an AxoClamp 900A amplifier (Molecular Devices, Sunnyvale, CA, USA) and pCLAMP 10 software. Electrodes were filled with 3 M KCl and had resistances of 0.5–1.5 MΩ. To facilitate the inward potassium current flow through the channels, normal oocyte saline buffer was modified to increase KCl concentration to 16 mM with a corresponding decrease in NaCl concentration. All drug responses (i.e. drug-elicited inward currents through Kir3 channels) were evaluated in this high K+-containing solution. A valve system (AutoMate Scientific, Berkeley, CA, USA) was used to control solution changes and data were collected using pCLAMP 10 software (Molecular Devices).

Agonist/receptor evoked Kir3 responses were followed for 4 min and the average rate of receptor desensitization normalized as a percent of the peak response was calculated as previously described (Kovoor et al. 1997). Briefly, the rate of desensitization was calculated using the formula (amplitude of the peak agonist induced response – amplitude of the agonist response at time t) ÷ t, where ‘t’ is the time in minutes after the peak response at which the final response was measured. The rate was then normalized by expressing of the peak agonist-activated response obtained after agonist perfusion.

Statistical analysis

The Student's t-test was used with two-tailed p-values for comparisons of independent mean values. For multiple comparisons, one-way anova was applied and when applicable followed by the Tukey's post hoc analysis. GraphPad Prism 4 (GraphPad Software, Inc, La Jolla, CA, USA) was used to fit agonist dose response data to a simple Emax model, and to derive the different EC50 values.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Arrestin-dependent but GRK-independent uncoupling of D2R

To compare the actions of GRK and arrestin on D2R and DOR-elicited responses, we co-expressed D2R and DOR with Kir3.1 and Kir3.4 subunits, and with either GRK3 or ARRB2 alone, or with both GRK3 and ARRB2. As previously described (Kovoor et al. 1997; Celver et al. 2010), activation of D2R or DOR with their respective agonists, dopamine or D-pen2–5-encephelan, produced robust activation of the co-expressed Kir3 channels indicated by an inward current in oocytes clamped at −80 mV and perfused with 16 mM K+ containing buffer (Fig.  1). In the absence of GRK3 or ARRB2 co-expression, sustained agonist exposure produced only a modest reduction in the evoked current (Fig.  1), which has previously been shown to result from changes occurring downstream of the receptor (Kovoor et al. 1995, 1997). The desensitization rates of the DOR response were not significantly altered by the coexpression of GRK3 or ARRB2 alone but, as demonstrated previously (Kovoor et al. 1997), was synergistically enhanced by the coexpression of both GRK3 and ARRB2 (Figs 1 and 5). Interestingly, in contrast to what was observed with DOR, co-expression of ARRB2 alone promoted rapid desensitization of D2R responses and the rate of desensitization of the D2R-elicited response was not significantly enhanced by co-expression of GRK3 (Figs 1, 2 and 4).

image

Figure 1. Arrestin-dependent, G-protein coupled receptor kinase (GRK)-independent uncoupling of D2 like dopamine receptor (D2R)-elicited Kir3 responses in Xenopus oocytes. (a) Control oocytes were injected with cRNA for D2R and delta opioid receptor (DOR) and the Kir3 channel subunits, Kir3.1 and Kir3.2. Some oocytes were injected, in addition, with cRNA for either GRK3 or β-arrestin2 (ARRB2) or both GRK3 and ARRB2 as indicated. Representative baseline subtracted current traces of Kir3 responses elicited by the DOR agonist, D-pen2–5-encephelan (DPDPE) (1 μM, top) or dopamine (1 μM, bottom). Vertical line marks on the traces indicated the time at which the agonist was washed out. (b) Quantification of the desensitization rates of dopamine (white bar) and DPDPE (black bar)-activated responses measured in the different oocyte groups indicated above (mean ± SEM, n = 4–5). The responses were all obtained on the same day and from the same oocyte donor. However, the effects of GRK3 and ARRB2 expression on agonist-dependent desensitization rates were confirmed at each receptor in oocytes harvested from multiple donors. Analysis of variances indicated significant differences in average desensitization rates between cRNA injection groups for both DPDPE (anova, p < 0.005) and dopamine-activated (anova, p < 0.005) responses. For the DPDPE responses, only desensitization rates from oocytes injected with cRNA for ‘GRK3 and ARRB2′ were significantly different from the ‘control’ oocytes (Tukey's post hoc test, p < 0.001). For the dopamine responses, desensitization rates from oocytes injected with cRNA for either ‘ARRB2′ or ‘GRK3 and ARRB2′ were significantly different from dopamine treated ‘control’ oocytes (Tukey's, p < 0.05), but the desensitization rates were not different between the latter two groups.

Download figure to PowerPoint

image

Figure 2. Residues 212–215 of D2 like dopamine receptor (D2R) are required for arrestin-mediated uncoupling of D2R responses in Xenopus oocytes. (a) Representative current traces of Kir3 responses elicited via dopamine activation of D2R (top) and the D2R212–215(IYIV)[RIGHTWARDS ARROW]A mutant (bottom). Control oocytes were injected with cRNA for the Kir3 subunits, Kir3.1 and Kir3.2, and either D2R or the D2R212–215(IYIV)[RIGHTWARDS ARROW]A mutant. Some oocytes in each of the groups injected with either D2R or D2R212–215(IYIV)[RIGHTWARDS ARROW]A cRNA were also injected with cRNA for β-arrestin2 (ARRB2) as indicated. The traces represent the baseline subtracted Kir3 responses elicited after perfusion of oocytes with dopamine (1 μM, bottom row). Vertical line marks on the traces indicate the time at which dopamine was washed out. (b) Quantification of the desensitization rates of D2R and the D2R212–215(IYIV)[RIGHTWARDS ARROW]A-activated Kir3 responses in control oocytes (white bar) or in oocytes injected with cRNA for ARRB2 (black bar, mean ± SEM, n = 4). ARRB2 co-expression significantly increased the desensitization rate of D2R responses (p < 0.0005, t-test) but not of D2R212–215(IYIV)[RIGHTWARDS ARROW]A. (c) Dopamine dose-response curves for D2R and D2R212–215(IYIV)[RIGHTWARDS ARROW]A-mediated Kir3 activation. The response at each dose is expressed as a percent of the average maximal response (n = 4). The EC50 (nM) derived for the dopamine-mediated activation of D2R and D2R212–215(IYIV)[RIGHTWARDS ARROW]A was 70.7 ± 10.6 and 99.4 ± 38.3 nM (mean ± SD), respectively, and were not significantly different between the two receptors.

Download figure to PowerPoint

Previously we had shown that agonist-mediated GPCR uncoupling can be reconstituted in Xenopus oocytes without producing significant receptor internalization (Appleyard et al. 1999; Jin et al. 1999; Celver et al. 2004). We confirmed that the observed desensitization of D2R and DOR reported here also occurred in the absence of significant receptor internalization (see Figure S1).

D2R residues in the third cytoplasmic loop required for internalization of D2R in mammalian cells and involved in mediating high-affinity interactions with arrestin and are also required for ARRB2-mediated uncoupling of D2R in Xenopus oocytes

Mutation of a four amino acid segment (Ile212, Tyr213, Ile214, and Val215), in the N-terminal region of the 3rd intracellular loop, of human D2R long form to alanine disrupts D2R-ARRB2 interactions in vitro and blocks agonist-dependent internalization of D2R in mammalian cells (Lan et al. 2009). We generated such a D2R mutant, D2R212-215(IYIV)[RIGHTWARDS ARROW]A in which amino acids 212–215 were substituted with alanine and expressed this receptor in Xenopus oocytes. Dose-response curves showed that the EC50 for dopamine at this D2R212-215(IYIV)[RIGHTWARDS ARROW]A mutant receptor was not significantly different from that of the receptor (Fig. 2a). ARRB2 co-expression, however, did not produce a significant increase in the agonist-dependent desensitization rate of the D2R212-215(IYIV)[RIGHTWARDS ARROW]A mutant (Fig. 2b and c).

The 3rd cytoplasmic loop of D2R confers ARRB2-dependent but GRK-independent uncoupling to DOR

Previous studies have localized a majority of the agonist-dependent and phorbal ester-induced phosphorylation sites to the third cytoplasmic loop of D2R (Namkung and Sibley 2004; Namkung et al. 2009a; Cho et al. 2010). Therefore, we asked if this region was important in mediating the arrestin-dependent but GRK-independent uncoupling of D2R. We constructed a chimeric DOR receptor, DORD2R3rd-loop, in which the third cytoplasmic loop of DOR was substituted with the third cytoplasmic loop of D2R. DORD2R3rd-loop was able to couple to Kir3 channels when co-expressed in Xenopus oocytes and the dose-response curve produced with the DOR agonist, D-pen2–5-encephelan, at this mutant receptor was similar to that obtained with DOR (Fig. 3b). Significantly, however, co-expression of ARRB2 significantly increased the agonist-dependent desensitization rate of DORD2R3rd-loop but not that of DOR (Fig. 3c). These data suggest that the determinants for arrestin-dependent but GRK independent uncoupling of D2R are localized to its third cytoplasmic loop.

image

Figure 3. The third cytoplasmic loop of D2 like dopamine receptor (D2R) can confer G-protein coupled receptor kinase-independent and arrestin-dependent uncoupling to delta opioid receptor (DOR). (a) Schematic of DOR, D2R and the chimeric DOR receptor construct (DORD2R3rd-loop) in which the third cytoplasmic loop of DOR was substituted with the third cytoplasmic loop of D2R. (b) Average desensitization rates (± SEM) of the Kir3 responses elicited by the DOR agonist, D-pen2–5-encephelan (DPDPE), in oocytes injected with cRNA for the indicated receptor constructs (DOR or DORD2R3rd-loop) and Kir3 channels (control, white bars). The desensitization rates are compared to those oocytes that were also injected in addition with cRNA for β-arrestin2 (ARRB2, black bars). Analysis of variances indicated significant differences in the desensitization rates between the different oocyte groups (p < 0.001). Tukey's post hoc test indicated that the average rate of desensitization in oocytes injected with cRNA for the DORD2R3rd-loop construct and ARRB2 was significantly different from the corresponding control (p < 0.001) as well as from oocytes injected with cRNA for DOR and ARRB2 (p < 0.05). The desensitization rate observed in oocytes injected with cRNA for DOR and ARRB2 was not significantly different from the corresponding control. (c) Dose-response curves for DOR or DORD2R3rd-loop-mediated Kir3 activation produced with the DOR agonist, DPDPE. The response at each DPDPE dose is expressed as a percent of the average maximal response that was obtained (n = 3). The EC50 (in nM) for DPDPE activation of DOR and DORD2R3rd-loop was 22.4 ± 8.8 and 28.0 ± 5.6 nM (mean ± SD), respectively, and were not significantly different between the two receptors.

Download figure to PowerPoint

Arrestin-dependent but GRK-independent uncoupling of D2R is blocked by staurosporine

As GRK co-expression was not required for arrestin-dependent uncoupling of D2R in the Xenopus oocyte system, we investigated if phosphorylation of D2R by other kinases could be implicated in such uncoupling. We found that pre-treatment of Xenopus oocytes with the potent serine/threonine kinase inhibitor, staurosporine (10 μM, 10 min), blocked the ARRB2-dependent uncoupling of D2R (Fig. 4a).

image

Figure 4. G-protein coupled receptor kinase-independent arrestin-dependent uncoupling of D2 like dopamine receptor (D2R) is blocked by staurosporine treatment and by alanine substitution of putative protein kinase C phosphorylation sites. (a) Average desensitization rates (± SEM) of the Kir3 responses elicited by dopamine in oocytes injected with cRNA for D2R and Kir3 channel subunits (control, white bars) and in oocytes that were also injected with cRNA for β-arrestin2 (ARRB2, black bars). Staurosporine (10 μM, 10 min) treatment was performed immediately prior to measuring the dopamine response. ARRB2 cRNA injection significantly increased desensitization rate of D2R compared with control (*p < 0.005, t-test, n = 4–6). After staurosporine treatment, however, desensitization rates were not significantly different between control and ARRB2 injected groups. (b) Average desensitization rates (± SEM) of the Kir3 responses elicited by dopamine in oocytes injected with cRNA for the indicated receptor constructs (D2R or D2RT225,S228,S229[RIGHTWARDS ARROW]A) and Kir3 channels (control, white bars) and in oocytes that were also injected with cRNA for ARRB2 (black bars). Data represent two independent experiments from two separate oocyte donors, n = 8–10. ARRB2 cRNA injection significantly increased desensitization rate compared with control for the D2R group (*p < 0.005, t-test). However, for the oocyte groups that were injected with cRNA for the D2RT225,S228,S229[RIGHTWARDS ARROW]A construct, the desensitization rates were not significantly different between the ARRB2 cRNA injected group and the corresponding control. C. Dopamine dose-response curves for D2R or D2RT225,S228,S229[RIGHTWARDS ARROW]A-mediated Kir3 activation. The response at each dopamine dose is expressed as a percent of the average maximal response that was obtained (n = 4). The EC50 (nM) for dopamine activation of D2R and D2RT225,S228,S229[RIGHTWARDS ARROW]A were significantly different (p < 0.05, t-test)and were 61.5 ± 15.5, and 36.4 ± 10.4 nM, respectively.

Download figure to PowerPoint

Mutation of putative protein kinase C phosphorylation sites within D2R blocks arrestin-dependent D2R uncoupling

Previous studies have identified specific serine and threonine residues within the third cytoplasmic loop of D2R that are phosphorylated after treatment of D2R-expressing cells with PKC activating phorbol esters, (Namkung and Sibley 2004; Namkung et al. 2009a; Cho et al. 2010). Furthermore, substitution of a cluster of these serine and threonine residues, Thr225, Ser228, and Ser229, to alanine significantly inhibited agonist dependent internalization of D2R (Cho et al. 2010). These residues were also implicated in PKC-dependent desensitization of D2R responses in primary neuron cultures although it was not determined whether the desensitization resulted from internalization or uncoupling (Thibault et al. 2011). We therefore, constructed a mutant D2R, D2RT225,S228,S229[RIGHTWARDS ARROW]A, in which these residues were mutated to alanines. When expressed in Xenopus oocytes with Kir3, the co-expression of ARRB2 failed to significantly enhance the desensitization rate of D2RT225,S228,S229[RIGHTWARDS ARROW]A elicited Kir3 responses (Fig. 4b), suggesting that these mutations blocked arrestin-dependent D2R uncoupling.

The dopamine dose-response curve for D2RT225,S228,S229[RIGHTWARDS ARROW]A indicates a small but significant difference in the EC50 from that of (Fig. 4c). As we were careful to adjust the expression levels of the D2R constructs in the oocyte so that there were no spare receptors and because D2R desensitization was induced using saturating concentrations of dopamine (1 μM) we expect that the small change in the EC50 of the mutant receptor does not directly affect its desensitization rate.

Agonist-dependent uncoupling of the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant is not observed even after GRK and arrestin co-expression

The data presented in Fig. 4 suggest that agonist-induced arrestin-dependent uncoupling of D2R observed in Xenopus oocytes was likely mediated by the phosphorylation of Thr225, Ser228, and Ser229 by a kinase endogenously expressed in Xenopus oocytes. The D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant provided the opportunity to investigate if, in the absence of the above described dominant mechanism for the arrestin-dependent uncoupling of D2R, GRK-mediated phosphorylation of other serine and threonine residues could produce ARRB2-dependent uncoupling of D2R. Therefore, we co-expressed the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant in Xenopus oocytes with Kir3 channels and then examined the effects of co-expression of either GRK3 or ARRB2 alone, or both GRK3 and ARRB2 (Fig. 5). DOR was also co-expressed and served as a positive control for confirming the functional expression of ARRB2 and GRK3. The expression of GRK3 produced a slight but significant increase in the desensitization rate of D2RT225,S228,S229[RIGHTWARDS ARROW]A-elicited responses. We have previously attributed such GRK3-dependent arrestin-independent desensitization to sequestration of the G protein Gβγ dimer by GRK (Kovoor et al. 1997). The Kir3 response is thought to be mediated by free Gβγ dimers that are released after receptor-mediated activation of the trimeric G proteins (Kovoor and Lester 2002), and both GRK2 and three can bind to and sequester the released Gβγ (Gurevich et al. 2012). Importantly, however, we observed that co-expression of ARRB2 and GRK3 expression did not further enhance the rate of D2RT225,S228,S229[RIGHTWARDS ARROW]A desensitization (Fig. 5). The DOR desensitization rate, in the same oocytes, was clearly enhanced by the co-expression of both GRK3 and ARRB2 confirming the functional expression of all the co-expressed constructs.

image

Figure 5. G-protein coupled receptor kinase (GRK) co-expression with the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant does not produce agonist-mediated arrestin-dependent uncoupling that is observed with D2 like dopamine receptor (D2R). Control oocytes were injected with cRNA for delta opioid receptor, the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant and the Kir3 channel subunits, Kir3.1 and Kir3.2. Some oocytes were injected, in addition, with cRNA for either GRK3 or β-arrestin2 (ARRB2) or both GRK3 and ARRB2 as indicated. Quantification of the average desensitization rates (± SEM, n = 5–8) of D-pen2–5-encephelan (DPDPE) (open bar) and dopamine (closed bar)-activated Kir3 responses measured in the different oocyte groups indicated above. Analysis of variance indicated significant differences in desensitization rates for the DPDPE- and dopamine-elicited responses (p < 0.0005 and p < 0.01, respectively) between the different cRNA injection groups. For DPDPE-elicited responses, Tukey's post hoc analysis indicated significant difference between the following groups: ‘control’ and ‘GRK3 and ARRB2’ (*p < 0.001) and between ‘GRK3’ and ‘GRK3 and ARRB2’ (ϕp < 0.001). For dopamine-elicited responses significant differences (#p < 0.05) were observed between the following groups: ‘control’ and ‘GRK3’; and ‘control’ and ‘GRK3 and ARRB2’; but not between ‘GRK3’ and ‘GRK3 and ARRB2’.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In this study, we utilized the Xenopus oocytes expression system to identify the molecular triggers which enable arrestin to interact with D2R in an agonist-dependent manner and produce D2R uncoupling. We used cell surface biotinylation to confirm that the desensitization of the receptor-activated Kir3 response observed for D2R and DOR occurred without detectable receptor internalization (see Figure S1) and most likely occurred as a result of receptor uncoupling from associated G proteins. The novel findings of the study are as follows: we found that arrestin could produce dopamine-dependent uncoupling of D2R from G-protein pathways even in the absence of functional GRK expression. Unlike what was observed for DOR in the same cells GRK co-expression was not required for and did not enhance arrestin-mediated uncoupling of D2R. The novel agonist-induced ARRB2-dependent but GRK3 independent uncoupling observed with D2R could be conferred to DOR by substituting the third cytoplasmic loop of DOR with the third cytoplasmic loop of D2R. Arrestin-dependent GRK3-independent uncoupling of D2R was blocked by staurosporine treatment and by alanine substitution of putative PKC phosphorylation sites in the third cytoplasmic loop of D2R. Finally, phosphorylation by GRK could not substitute for PKC phosphorylation to produce arrestin-mediated uncoupling of D2R.

Arrestins can bind and uncouple GPCRs from G proteins, serve as adapters that link the GPCR to the cellular internalization machinery and function as molecular scaffolds to support the activation of alternate signaling pathways by the GPCR (Gurevich et al. 2008). Importantly, the complex of a GPCR with an arrestin molecule does not exist as a singular entity or in a single fixed conformation. For example, some GPCRs have distinct phosphorylation requirements that determine whether arrestin-binding will mediate uncoupling or internalization of the GPCR indicating the existence of distinct conformations of the GPCR-arrestin produced by phosphorylation of the GPCR at different sites (Jin et al. 1999; Key et al. 2003; Celver et al. 2004). Furthermore, biased ligands identified for some GPCRs preferentially activate distinct arrestin-dependent signal transduction mechanisms revealing further potential for variability in the nature of the ligand/GPCR/arrestin complex (Reiter et al. 2012).

The molecular trigger utilized by D2R for ‘activating’ arrestin is unclear. However, specific residues in the D2R third cytoplasmic loop (212–215) have previously been identified as being required for the high-affinity interaction between D2R and arrestin, and for agonist-dependent internalization of D2R in mammalian cells (Lan et al. 2009). In the Xenopus oocyte expression system, we showed that the mutation of these residues (i.e. the D2R212-215(IYIV)[RIGHTWARDS ARROW]A mutant) blocked the arrestin-mediated uncoupling of D2R. Thus, it appears that the arrestin-D2R complex that forms in Xenopus oocytes, to uncouple D2R, shares binding properties of the D2R-arrestin complex that forms in mammalian cells to mediate agonist-dependent internalization. Thus, our data suggest that an important structural interface of the D2R-arrestin that exists in mammalian cells is also established in the Xenopus oocyte system and supports the use of the oocyte expression system as a physiologically valid system for further investigations into the D2R-arrestin complex.

The stark difference in the requirement for GRK co-expression to produce agonist-dependent uncoupling between DOR and D2R (Fig. 1) support the suggestion that arrestin-mediated regulation of D2R diverges from classical models of arrestin action and does not require phosphorylation of D2R by GRKs. For example, these data suggest that another kinase phosphorylates D2R or a proximate substrate and such phosphorylation triggers arrestin activation and allows for binding to the agonist-activated D2R. Alternatively, arrestin can uncouple D2R in a manner that does not require D2R phosphorylation. Our observation that staurosporine can inhibit the arrestin-depended uncoupling of D2R suggests the former explanation is the correct one and indicates that phosphorylation of D2R by an endogenously expressed serine/threonine kinase can trigger arrestin activation. Furthermore, the finding that arrestin-dependent uncoupling of D2R was blocked by mutation of serine and threonine residues (i.e. the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant) suggests that it is phosphorylated at these D2R sites that trigger arrestin activation. Such phosphorylation then allows arrestin to bind to the agonist-bound receptor in a manner that uncouples D2R from associated G proteins.

Agonist-dependent phosphorylation, presumably by GRKs, at a different set of D2R serine and threonine residues has previously been reported. However, the co-expression of both GRK and arrestin was unable to produce dopamine-dependent uncoupling of the D2RT225,S228,S229[RIGHTWARDS ARROW]A mutant, suggesting that GRK phosphorylation of D2R cannot substitute for the actions of the endogenous kinase; a subtle but important distinction which is difficult to discern in other systems. Our results support the conclusions from a recent study which examined the role of receptor-attached phosphates in binding of arrestins to GPCRs and suggested that arrestin interactions with D2R are largely independent of GRK phosphorylation (Gimenez et al. 2012).

From this study it is not clear whether the D2R is basally phosphorylated, thus priming arrestin for binding to D2R once the receptor is activated by dopamine, or whether the phosphorylation events occur in an agonist-dependent manner through an endogenously expressed GRK. However, Xenopus oocytes do not endogenously express detectable levels of GRK (see Figure S2) and phosphorylation studies in mammalian cells suggest the former: phosphorylation of the residues mutated in D2RT225,S228,S229[RIGHTWARDS ARROW]A may be blocked by PKC inhibitors and enhanced by PKC activation but is not affected by agonist activation of D2R (Cho et al. 2010). Thus, our data, taken together with previous studies (Namkung et al. 2009a,b; Cho et al. 2010; Thibault et al. 2011), suggest a scenario where alterations in PKC activity in D2R-expressing cells can have marked effects on the rates of both dopamine-mediating uncoupling and internalization of D2R. In this study, while we have not identified the kinase responsible for phosphorylating D2R in Xenopus oocytes, we believe we have more specifically linked the putative PKC phosphorylation sites, Thr225, T228, and S229, to the formation of an arrestin/D2R complex which can uncouple D2R from G protein in an agonist-dependent manner.

Mechanisms for the regulation of D2R remain poorly understood. For example, agonist-induced phosphorylation has recently been implicated in the recycling of D2R; that is, mutation of agonist-dependent phosphorylation sites has been reported as causing the D2R mutant to recycle at a slower rate to the plasma membrane. Consequently, by increasing the rate of reinsertion of internalized D2R back to the plasma membrane, GRK phosphorylation should decrease the net loss of D2R from the cell surface after agonist treatment. Instead, it has been found that GRK over-expression enhances the amount of D2R removed from the cell surface following agonist activation of D2R in mammalian cells (Ito et al. 1999; Celver et al. 2010; Cho et al. 2010). In another example, Cho and colleagues reported that a D2R-elicited cellular response, thought to be mediated via the second messenger, cAMP, was not desensitized after 40 min of dopamine pre-treatment even though such pre-treatment caused more than 40% of the receptor to be internalized. These apparently paradoxical observations were made in mammalian cell systems where GPCR uncoupling, internalization, recycling, and arrestin-mediated signaling can all occur simultaneously and in parallel with multiple other regulatory processes (Pak et al. 1999; Shapira et al. 2001). Mammalian cells also support GPCR internalization that occurs through mechanisms that do not involve GRK and arrestin (Shapira et al. 2001; Qiu et al. 2003; Namkung and Sibley 2004; Rasmussen et al. 2004). The concurrent occurrence of many of these processes in mammalian cells poses an impediment to untangling the multiple molecular determinants that may regulate D2Rs. The goal of fully untangling these processes may be supported by reconstitution experiments in multiple expression systems, each of which expresses a different set of regulatory molecules or which lack the full complement of regulatory elements found in vivo.

We believe this study is one such example. Agonist-induced internalization is not observed in Xenopus oocytes (Appleyard et al. 1999; Jin et al. 1999; Celver et al. 2004) and Xenopus oocytes do not express functional GRKs (see Figure S2) or arrestins (Kovoor et al. 1997; Appleyard et al. 1999; Jin et al. 1999; Celver et al. 2004). Here, in the absence of receptor internalization (see Figure S1) the effects of phosphorylation at specific D2R residues and the action of arrestin were uniquely associated with the functional uncoupling of D2R.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors have no potential conflict of interest with the publication of this manuscript. The project described was supported by grants from the National Institute of Mental Health (R15MH091639-01 to AK and MS), from the National Center for Research Resources (5P20RR016457-11) and the National Institute for General Medical Science (8 P20 GM103430-11), components of the National Institutes of Health (NIH), and its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIMH, NCRR, NIGMS or the NIH. Additional support includes work conducted using the Rhode Island Genomics and Sequencing Center which is supported in part by the National Science Foundation under EPSCoR Grants Nos. 0554548 & EPS-1004057.

The authors declare that they have reviewed the ARRIVE guidelines and these guidelines have been followed. Furthermore, the authors have no conflict of interest to declare.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
jnc12359-sup-0001-FigS1-S2.pdfapplication/PDF180K

Figure S1. Agonist treatment does not reduce cell surface levels of either D2R or DOR when the receptors are expressed in Xenopus oocytes.

Figure S2. Xenopus oocytes lack detectable endogenous GRK activity.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.