Unique Roles of β-Arrestin in GPCR Trafficking Revealed by Photoinducible Dimerizers

Intracellular trafficking of G protein-coupled receptors (GPCRs) controls their localization and degradation, which affects a cell’s ability to adapt to extracellular stimuli. Although the perturbation of trafficking induces important diseases, these trafficking mechanisms are poorly understood. Herein, we demonstrate an optogenetic method using an optical dimerizer, cryptochrome (CRY) and its partner protein (CIB), to analyze the trafficking mechanisms of GPCRs and their regulatory proteins. Temporally controlling the interaction between β-arrestin and β2-adrenergic receptor (ADRB2) reveals that the duration of the β-arrestin-ADRB2 interaction determines the trafficking pathway of ADRB2. Remarkably, the phosphorylation of ADRB2 by G protein-coupled receptor kinases is unnecessary to trigger clathrin-mediated endocytosis, and β-arrestin interacting with unphosphorylated ADRB2 fails to activate mitogen-activated protein kinase (MAPK) signaling, in contrast to the ADRB2 agonist isoproterenol. Temporal control of β-arrestin-GPCR interactions will enable the investigation of the unique roles of β-arrestin and the mechanism by which it regulates β-arrestin-specific trafficking pathways of different GPCRs.


Light-Induced Endocytosis of ADRB2 CIB .
To determine whether light irradiation controls the interaction between ADRB2 and β-arrestin, HEK293 cells stably expressing ADRB2 CIB and Arrestin CRY (HEK293 opt ) were irradiated with blue light under a confocal fluorescence microscope (Fig. 2a, Movie S1). Arrestin CRY was translocated from the cytosol to the cell membrane a few min after the start of irradiation and then moved back into the cytosol to form a dot-like structure. ADRB2 CIB fluorescent spots appeared in the cytosol at 15 min, and their number gradually increased until 30 min of irradiation. The fluorescent spots mostly colocalized with Arrestin CRY . In a control experiment, a fusion protein comprising SNAP-tag and ADRB2 without CIB was expressed in HEK293 cells. The localization of Arrestin CRY was unchanged upon stimulation with blue light due to the lack of the CIB portion ( Supplementary Fig. 1a). Furthermore, fluorescent spots in the HEK293 cells expressing Arrestin CRY and cell membrane-anchored CIB (Myr-Venus-CIB) were not observed during irradiation, despite Arrestin CRY being translocated to the cell membrane. To analyze the results quantitatively, we counted the number of spots containing ADRB2 CIB , ADRB2, and Myr-Venus-CIB in the obtained images ( Supplementary Fig. 1b). A significant increase in the number of spots was observed after light irradiation in case of ADRB2 CIB . In contrast, the number of spots containing ADRB2 or Myr-Venus-CIB did not change after light stimulation. The results indicate that recruitment of β-arrestin to ADRB2 is necessary for appearance of the spots. In addition, the cells expressing Myr-Venus-CIB and Arrestin CRY were irradiated for 120 min (Supplementary Fig. 1c). The number of spots containing Myr-Venus-CIB did not increase under the prolonged irradiation. Light-based manipulation of reversible interaction between β-arrestin and ADRB2 using photodimerizers. (a) Schematic structures of the optogenetic manipulation system for the reversible interaction of β-arrestin with ADRB2. ADRB2 CIB is composed of a signal peptide (SP) derived from 5HT3A serotonin receptor that targets the protein to the cell membrane, a SNAP-tag, a β2-adrenergic receptor (ADRB2), a CIB, and a V5 epitope tag. Arrestin CRY consists of a cryptochrome (CRY), an mCherry, a β-arrestin2, and a myc epitope tag. (b) Basic principle of light-induced interaction of β-arrestin2 with ADRB2. Before light stimulation, ADRB2 CIB localizes on the cell membrane, and Arrestin CRY distributes throughout the cytosol. Upon stimulation with blue light, Arrestin CRY interacts with ADRB2 CIB , which triggers the endocytosis of the ADRB2 CIB -Arrestin CRY complex. Under dark conditions, Arrestin CRY dissociates from ADRB2 CIB and redistributes into the cytosol.
Scientific REPORts | (2018) 8:677 | DOI: 10.1038/s41598-017-19130-y We next examined the endocytosis of ADRB2 after ADRB2 CIB -Arrestin CRY interaction. ADRB2 is endocytosed with β-arrestin in a clathrin-mediated manner 22 . Segregation of clathrin-coated pits from the cell membrane is regulated by dynamin GTPases. To show that the fluorescent spots resulted from clathrin-mediated endocytosis, we pretreated the HEK293 opt cells with a dynamin inhibitor, Dyngo4a. The fluorescent spots did not appear in the presence of Dyngo4a after the stimulation by light (Fig. 2b). To quantitate this inhibitory effect, we quantified the number of ADRB2 CIB fluorescent spots and the colocalization of Arrestin CRY with ADRB2 CIB using the quantitative parameter Manders' colocalization coefficient, which is proportional to the total amount of fluorescence intensity from Arrestin CRY in the pixels where it colocalizes with ADRB2 CIB (Supplementary Fig. 2a,b). The number of spots increased with prolonged irradiation time in the absence of Dyngo4a. In contrast, light irradiation did not induce an increase in the number of spots in the presence of Dyngo4a although the colocalization of Arrestin CRY with ADRB2 CIB was induced. Considering these results, we concluded that the endocytosis of ADRB2 CIB was triggered by light-induced interaction of β-arrestin with ADRB2, and the observed ADRB2 CIB fluorescent spots were ADRB2 CIB -containing vesicles produced by clathrin-mediated endocytosis.
We further investigated the time dependency of the light-induced endocytosis of ADRB2 CIB . After HEK293 opt cells were stimulated with blue light for different durations, the amount of ADRB2 CIB on the cell surface was quantified using an ELISA 23 (Supplementary Fig. 3a). The amount of ADRB2 CIB on the cell surface decreased with increasing irradiation time. We also examined the light intensity dependency of endocytosis ( Supplementary  Fig. 3b). Endocytosis was promoted by increased light intensity and plateaued at 3 mW/cm 2 . The amount of ADRB2 CIB endocytosed by 3 mW/cm 2 light was more than that induced by 0.01 μM ISO and less than that induced by 0.1 and 1.0 μM ISO. These results demonstrate that the endocytosis of ADRB2 CIB can be controlled by modulating light intensity and irradiation time. Localization of ADRB2 CIB after Light-Induced Endocytosis. Most GPCRs are transported to an endocytic pathway after internalization. To confirm that ADRB2 CIB is transported to endosomes after light-induced endocytosis, we immunostained early and late endosomes using antibodies specific to endosome marker proteins Rab5 and Rab7, respectively (Supplementary Fig. 4a,b). The internalized ADRB2 CIB was partially localized in Rab5-or Rab7-positive vesicles after 30 min of irradiation. The residual ADRB2 CIB did not colocalize with the endosomes, indicative of sorting to other endosomes or lysosomes. Based on these results, we concluded that part of the ADRB2 CIB is transported to endosomes after light-induced endocytosis.

Induction of the Recycling of ADRB2 CIB in Dark Conditions.
To show the reversibility of the light-induced interaction between ADRB2 CIB and Arrestin CRY , we irradiated the HEK293 opt cells with blue light for 30 min and then incubated them in the dark. After Arrestin CRY colocalized with ADRB2 CIB , Arrestin CRY redistributed uniformly in the cytosol under dark conditions (Fig. 3a). The fluorescence intensity of Arrestin CRY in the cytosol decreased 1 min after irradiation was started and then recovered to the basal level 10 min after irradiation ceased (Fig. 3b). The colocalization coefficient of Arrestin CRY with ADRB2 CIB decreased after stopping irradiation ( Supplementary Fig. 5). The results confirm that Arrestin CRY dissociated from ADRB2 CIB in the dark.
To demonstrate the recycling of ADRB2 CIB to the cell surface after the dissociation of Arrestin CRY , we quantified the amount of ADRB2 CIB on the cell surface using an ELISA after stopping irradiation. The HEK293 opt cells were stimulated with blue light for 60 min and then incubated in the dark. The amount of ADRB2 CIB on the cell surface decreased to 75% after light stimulation and recovered to 80% after 60 min of incubation in the dark (Fig. 3c, left). This result indicates that the recycling of the ADRB2 CIB was triggered by the dissociation of β-arrestin from ADRB2 in the dark. Furthermore, in HEK293 opt cells stimulated with ISO, the amount of ADRB2 CIB on the cell surface recovered from 63 to 77% after ISO was eliminated (Fig. 3c, right) The temporal changes in recycling after light irradiation ceased were similar to those in recycling when ISO was eliminated, demonstrating that temporal control of the dissociation between β-arrestin and ADRB2 is useful for investigating the intracellular dynamics of ADRB2 and regulatory proteins in the recycling pathway.
Sorting ADRB2 CIB into the Lysosome Pathway after Prolonged Light Irradiation. We next investigated the effects of a prolonged interaction of Arrestin CRY with ADRB2 CIB on the sorting of ADRB2 CIB . HEK293 opt cells were irradiated with blue light for 120 min. Lysosomes were immunostained with an antibody specific to the lysosome marker protein LAMP1 (Fig. 4a). Sustained light irradiation for 120 min resulted in the disappearance of the ADRB2 CIB from the cell membrane; instead, ADRB2 CIB predominantly localized in lysosomes. Furthermore, we examined the colocalization of ADRRB2 CIB with lysosomes (Fig. 4b). The Manders' coefficient increased with increased irradiation time, indicating that the prolonged interaction of Arrestin CRY with ADRB2 CIB drives ADRB2 CIB to lysosomes. We also examined the ubiquitination of ADRB2 CIB , which is a crucial signal that sorts GPCRs to lysosomes 24 . Furthermore, ADRB2 CIB was ubiquitinated to a greater degree when the irradiation time was increased to 120 min ( Fig. 4c,d). These results demonstrate that prolonged interaction of β-arrestin with ADRB2, stimulated by blue light, promotes the sorting of ADRB2 CIB to lysosomes.
Monitoring a Protein that Regulates ADRB2-β-arrestin Complexes. Some regulatory proteins are recruited to ADRB2-β-arrestin complexes; monitoring these regulatory proteins is important to determine their dynamics and functions in trafficking pathways. Mdm2 is a protein that regulates the endocytosis of ADRB2 25 . Although recent studies suggested that Mdm2 has several important roles in the regulation of other membrane receptors, such as the induction of endocytosis of opioid receptors and the activation of a signaling pathway regulated by an IGFR, the dynamics for Mdm2 recruitment to these receptors is unclear 26,27 . To monitor the intracellular dynamics of Mdm2, we established a cell line that stably expresses ADRB2 CIB and Arrestin CRY , but we replaced mCherry with a CLIP tag (HEK293 optCLIP ). HEK293 optCLIP cells were transfected with cDNA coding mCherry-fused Mdm2 and were then stimulated with blue light under a confocal microscope (Fig. 5a). Before stimulation, most Mdm2 localized on the nucleus, and the residual Mdm2 distributed throughout the cytosol. Several fluorescent spots of Mdm2 were observed in the cytosol after 15 min of light irradiation, and they colocalized with the ADRB2 CIB -containing vesicles. This result suggests that the cytosolic Mdm2 is recruited to the light-induced ADRB2-β-arrestin complexes. The colocalization of Mdm2 with ADRB2 CIB -containing vesicles was sustained during continuous irradiation ( Fig. 5a upper, Movie S2a). In contrast, the Mdm2 dissociated from ADRB2 CIB 10 min after irradiation ceased ( Fig. 5a lower, Movie S2b). We counted the number of Mdm2-recruiting vesicles in the cytosol (Fig. 5b). The number of Mdm2-recruiting vesicles increased with increasing irradiation time but decreased with increasing incubation in the dark. These results indicate that the interaction of β-arrestin with ADRB2 promotes the recruitment of Mdm2 to the ADRB2-β-arrestin complex and that the interaction of Mdm2 with the ADRB2 CIB complex continues until Arrestin CRY dissociates from ADRB2 CIB .
Examination of the Ability of Arrestin CRY to Desensitize ISO-activated ADRB2 CIB . Desensitization of ligand-activated GPCRs is one of the important functions of β-arrestin. To examine whether light-induced recruitment of Arrestin CRY desensitizes the ISO-activated ADRB2 CIB , we investigated cAMP production after ISO stimulation under blue light irradiation using cAMP biosensor ( Supplementary Fig. 6). The light irradiation did not affect luminescence intensity changes, suggesting that light-induced recruitment of Arrestin CRY does not promote desensitization of ADRB2 CIB .

Investigation of ADRB2 CIB Activity during Light Stimulation.
To confirm the activity state of ADRB2 CIB during the light stimulation, HEK293 opt cells were transfected with cDNA coding a GFP-fused nanobody (Nb80), which binds specifically to the active state of ADRB2 interacting to G proteins 28 . The cells were stimulated with light or isoproterenol for 30 min under a confocal microscope. The GFP-Nb80 colocalized with ADRB2 CIB -containing endosomes after ISO stimulation, indicating that the ADRB2 CIB has an ability to activate the downstream signaling upon ISO stimulation ( Supplementary Fig. 7a). In contrast, the colocalization of GFP-Nb80 with ADRB2 CIB was not observed after light stimulation. To demonstrate the results quantitatively, we quantified the number of spots containing GFP-Nb80 and the colocalization of GFP-Nb80 with ADRB2 CIB . The spots containing GFP-Nb80 were not observed 30 min after light irradiation, although ISO stimulation induced increases in the number of spots ( Supplementary Fig. 7b). The Manders' colocalization coefficient did not change after light stimulation, but increased after ISO stimulation ( Supplementary Fig. 7c). These results suggest that endocytosed ADRB2 CIB by light stimulation does not form the active state required for G protein signaling.
Investigating the Ability of Arrestin CRY to Induce Phosphorylation of ERK1/2. β-Arrestin, after interacting with ADRB2, functions as a scaffold for the activation of MAPK signaling 29 . Ligand stimulation induces phosphorylation of ADRB2 via G protein-coupled receptor kinases. Phosphorylation of ADRB2 increases the affinity of ADRB2 for β-arrestin and changes the conformation of β-arrestin to form a complex with signaling molecules Scientific REPORts | (2018) 8:677 | DOI:10.1038/s41598-017-19130-y such as MAPK/ERK kinase (MEK) and ERK1/2, which are then phosphorylated through the interaction with β-arrestin [30][31][32] . To examine whether Arrestin CRY induces phosphorylation of ERK1/2 after the light-induced interaction of Arrestin CRY with ADRB2 CIB , HEK293 opt cells were irradiated with blue light for 60 min. The phosphorylated ERK1/2 level was evaluated with its specific antibody ( Supplementary Fig. 8a). ISO stimulation increased the amount of phosphorylated ERK1/2, whereas the phosphorylated ERK1/2 level decreased after light stimulation. To clarify the reason why light stimulation failed to increase the phosphorylated ERK1/2 level, we investigated phosphorylated ADRB2 (Supplementary Fig. 8b). The phosphorylation of ADRB2 CIB was not detected after light irradiation. Considering that the phosphorylation of ADRB2 triggers a conformational change in β-arrestin that enables the recruitment of ERK1/2 and MEK to β-arrestin 30-32 , we concluded that Arrestin CRY does not recruit ERK1/2 or MEK to the Arrestin CRY -ADRB2 CIB complex due to the lack of ADRB2 phosphorylation.

Applicability of the Light Activation System to Other Membrane Receptors.
To investigate the applicability of the present system to other membrane receptors, we replaced ADRB2 with different GPCRs such as the neurotensin receptor, muscarinic acetylcholine receptor M3, corticotropin releasing-factor receptor, and vasopressin 2 receptor. All the GPCRs that were investigated in this study localized on the cell surface before  (Supplementary Fig. 9a). Furthermore, we applied the method to a non-GPCR receptor, transforming growth factor 3 receptor (TGF3R). TGF3R localized on the cell surface before light stimulation, whereas light stimulation led to the endocytosis of TGF3R, indicating that endocytosis of TGF3R is controllable by the recruitment of β-arrestin to TGF3R. These results demonstrate that the present method is widely applicable to inducing the endocytosis of membrane receptors in living cells. A previous report demonstrated that ERK1/2 phosphorylation was promoted by chemical dimerizer-induced interaction of β-arrestin with V2R 23 , implying different regulations of MAPK activation between ADRB2 and V2R. To examine the difference, we quantified phosphorylated ERK1/2 after light-induced interaction of Arretin CRY with V2R CIB . In contrast to the case of ADRB2, the light-induced interaction of β-arrestin with V2R promoted phosphorylation of ERK1/2 30 min after light irradiation (Supplementary Fig. 9b). The results suggest that β-arrestin regulates ERK1/2 phosphorylation in different manners depending on the GPCRs.

Discussion
We developed an optogenetic method for the analysis of intracellular trafficking of ADRB2 regulated by β-arrestin. We clarified the significance of the interaction of β-arrestin in the intracellular trafficking of ADRB2 using the CRY-CIB system. Endocytosis was promoted by increasing the time that ADRB2 interacts with β-arrestin. The ADRB2 on the cell surface was recovered after β-arrestin dissociated from ADRB2 in the dark, indicating that the recycling pathway was triggered by the dissociation of β-arrestin from ADRB2. Furthermore, ubiquitination and lysosome sorting were accelerated by prolonged irradiation. These results indicate that the duration of β-arrestin interaction with ADRB2 determines whether ADRB2 takes the lysosomal pathway. In addition, we monitored Mdm2 binding to and dissociating from ADRB2 CIB -Arrestin CRY complexes using light irradiation. Our method monitors the dynamics of ADRB2 and its regulatory proteins in the trafficking pathways. Light-induced manipulation of the interaction of β-arrestin with ADRB2 consequently clarified the regulation of the intracellular trafficking of ADRB2. The present system will be a robust tool for investigating the dynamics of a wide variety of GPCRs and their regulatory proteins in trafficking pathways. Furthermore, we examined other functions of β-arrestin such as desensitization and signaling. Light-induced interaction of β-arrestin with ADRB2 did not promote degradation of cAMP or activation of MAPK. The results indicate that the present method is applicable to controlling the trafficking of ADRB2 without inducing the desensitization or signaling, which is beneficial for investigation of mechanisms of the β-arrestin-mediated trafficking of ADRB2 in living cells and animals.
Controlling the interaction of GPCRs with β-arrestin using the CRY-CIB photoactivation system has several advantages. Previous methods to control the interaction of GPCRs with β-arrestin were based on a chemical approach 23,33 . A pair of chemical dimerizers (FKBP-FRB) were used to control the interaction of β-arrestin with vasopressin receptors or chemokine receptors. However, it is difficult to eliminate rapamycin by simple washing procedures due to its high affinity for FKBP 34 , suggesting that the chemical approach is not suitable for the temporal manipulation of these interactions. The CRY-CIB system is a robust tool to temporally control the reversible interaction of β-arrestin with membrane receptors. The other advantage of the present method is that it can be applied to a wide variety of membrane receptors by connecting CIB to the C-terminus of the target membrane receptor. This approach is useful for the manipulation of orphan GPCRs, whose ligands are undetermined. Recent studies have demonstrated that the trafficking of an orphan GPCR contributes to the pathology of Alzheimer disease 35,36 . However, knowing its specific ligands is required for further analysis of the pathological process. The present method will be useful for the analysis of such orphan GPCRs, without requiring their specific ligands, because the trafficking of GPCRs is controllable by external light. The present method can potentially determine the effects of trafficking a wide variety of membrane receptors on important diseases.
We induced prolonged interactions of β-arrestin with ADRB2 using light irradiation. GPCRs are divided into two classes (class A and class B) depending on their affinity with β-arrestin 14 . ADRB2 is a class A GPCR, members of which interact with β-arrestin for a short period after stimulation by their ligands. We artificially prolonged the interaction of β-arrestin with ADRB2 using light stimulation. The sustained interaction led to greater ubiquitination and lysosome sorting of ADRB2 than stimulation with ISO, suggesting that the duration of the interaction between ADRB2 and β-arrestin determines the amount of ubiquitination and lysosome sorting. In addition, we demonstrated the light-induced endocytosis of 4 different types of GPCRs and a TGF3R. The light-endocytosed GPCRs are possibly transported in different kinetics because the trafficking is largely influenced by the amino acid residues in the C-terminus 12 . The investigation of the differences provides us valuable information about the significant roles of C-terminus of GPCRs in kinetics regulation of their trafficking without effects of other complicated factors such as ligand affinity to GPCRs and G protein signaling.
We examined the significance of the phosphorylation of ADRB2 in trafficking and signal induction by artificially manipulating the interaction of β-arrestin with ADRB2. The direct manipulation of interactions between specific proteins has been a robust approach for analyzing the significance of protein-protein interactions in complicated intracellular networks 15 . The ligands of ADRB2 activate G proteins via ADRB2 on the cell surface, and the G protein signaling leads to phosphorylation of ADRB2, which increases the affinity of ADRB2 for β-arrestin. The interaction of β-arrestin with ADRB2 on the cell surface induces the endocytosis of ADRB2. We artificially induced the interaction between ADRB2 and β-arrestin via the CRY-CIB system without phosphorylating ADRB2. Light stimulation triggered the endocytosis of ADRB2 CIB , indicating that phosphorylation of ADRB2 is not necessary to initiate the recruitment of endocytic proteins such as clathrin and Mdm2 to ADRB2-β-arrestin complexes. However, the maximum induction level of the endocytosis of ADRB2 CIB after light stimulation was lower than that induced by 0.1 and 1.0 μM ISO. The result implies that phosphorylation of ADRB2 is required for the efficient endocytosis of ADRB2 from the plasma membrane. We also evaluated the induction level of MAPK signaling after light irradiation. The level of phosphorylated ERK1/2 unexpectedly decreased after light irradiation. Although the reason for the phosphorylation decrease after light stimulation is not clear, the result indicates that other factors increase the amount of phosphorylated ERK1/2. For example, phosphorylated ADRB2 facilitates β-arrestin adopting a specific conformation to form an ADRB2-β-arrestin-MEK-ERK1/2 complex [30][31][32] . We speculated that the lack of increase in phosphorylated ERK1/2 after light stimulation was due to the lack of phosphorylated ADRB2. To confirm the conformation of β-arrestin, several FRET-and BRET-based indicators were reported 32,37 . However, these indicators contain bioluminescent, fluorescent proteins or chemical dyes, of which absorption and emission wavelength overlap with absorption wavelength of CRY, suggesting a possibility that the overlap hampers a precise detection of the conformation of β-arrestin. To overcome the issue, use of a FRET indicator with another absorption and emission wavelength such as mKate and iRFP will be needed. Combination of a novel indicators and the present system will provide powerful approaches for investigation of the relevance between conformation of β-arrestin and intracellular trafficking on cell surface and endosomes. In addition, several studies have demonstrated that the rapamycin-induced interaction of β-arrestin with vasopressin receptors or chemokine receptors promotes the phosphorylation of ERK1/2 23,33 . Light-induced interaction of β-arrestin with V2R also activated the phosphorylation of ERK1/2. The result using ADRB2 is completely different from the results of the reports and our result, suggesting that the activation of MAPK signaling depends on the type of GPCR. We consequently suggested that the light-induced interaction of β-arrestin with ADRB2 is not sufficient for the activation of MAPK signaling, which is a significant clue to understanding the mechanisms of activation of β-arrestin-mediated signals. We developed the photoinducible GPCR-β-arrestin interaction systems, and utilized them for investigation of the significant roles of β-arrestin on the endocytosis and intracellular trafficking of GPCRs. In addition, there are other potential applications of the present methods. Previous studies demonstrated importance of GPCR conformations for activation and inactivation of downstream signaling 28 . However, it is not well investigated whether the conformational state of GPCRs influences on their intracellular trafficking. The trafficking of GPCRs in a specific conformation may be possibly investigated using the present system together with antagonists or inverse agonists that induce specific GPCR conformations. This approach will clarify significant roles of conformational states of GPCRs on intracellular trafficking. In addition, a previous study also suggested the possibility of the endocytosis of G proteins with some GPCRs 38 , whereas another study on single molecule imaging reported that G protein-ADRB2 complexes did not colocalize with clathrin-coated pits 39 . Thus, it is not clear whether the interaction of G protein with GPCRs affects their endocytosis. The present system may be applicable to clarify the impact of G proteins on the receptor internalization by comparing the dynamics of endocytosis between ligand-activated GPCRs interacting with G proteins and light-stimulated GPCRs without G proteins. Furthermore, the present method has a potential to control the oligomeric state of the GPCRs. GPCRs form dimers and oligomers upon stimulation of their ligands, which is important for recruitment of their regulatory proteins on cell surface 40 . Based on the fact that the light-activated CRY and CIB forms oligomers 41 , ADRB2 CIB and Arrestin CRY also form oligomers upon light stimulation. Light control of the oligomeric state of ligand-stimulated GPCRs will be a useful approach to investigate the oligomerization of GPCRs affecting their regulatory proteins in living cells.
In conclusion, we developed a novel method that uses the CRY-CIB system to manipulate the interaction between β-arrestin and ADRB2. Temporal manipulation of the interaction between β-arrestin and ADRB2 using this optogenetic system revealed the β-arrestin-mediated regulation of the intracellular trafficking of ADRB2. Endocytosis of ADRB2 was promoted by increased light intensity and irradiation time. The dissociation of β-arrestin from ADRB2 under dark conditions triggered the recycling of the endocytosed ADRB2 to the cell surface, whereas the prolonged interaction of β-arrestin with ADRB2 promoted the sorting of ADRB2 to lysosomes. The duration of the reversible interaction of β-arrestin with ADRB2 therefore determines the intracellular trafficking of ADRB2. In addition, the recruitment of the regulatory protein Mdm2 to ADRB2-β-arrestin complexes was driven by light induction of the interaction between β-arrestin and ADRB2. Our method will be effective for investigating the dynamics of GPCRs and their regulatory proteins in trafficking pathways. Furthermore, MAPK signaling was not activated through the light-induced interaction of β-arrestin with ADRB2; this behavior may be due to the light-induced interaction not causing a conformational change in β-arrestin that results in recruitment of MEK and ERK1/2. We consequently showed that the artificial induction of interactions between membrane receptors and their regulatory proteins via an optogenetic tool is a useful approach for investigating the intracellular trafficking of membrane receptors. Light-induced internalization of membrane receptors on cell surfaces and the manipulation of intracellular trafficking will be beneficial to unveiling significant roles of membrane receptors in living tissues and will also be useful for investigating the pathological processes of important diseases.

Methods and Materials
Materials and cDNA Construction. SNAP-Surface Alexa Fluor 647 and anti-SNAP-tag antibodies were obtained from New England Biolabs (USA). ISO and Dyngo4a were purchased from Tokyo Chemical Industry (Japan) and Abcam (USA), respectively. Anti-Rab5, anti-Rab7, anti-LAMP1, anti-ERK1/2 and anti-phosphorylated ERK1/2 antibodies were obtained from Cell Signaling Technology (USA). Anti-β2-adrenergic receptor (ADRB2) and anti-phosphorylated ADRB2 antibodies were purchased from Santa Cruz Biotechnology (USA). The SNAP-Surface Alexa Fluor 647 and Dyngo4a were solved in DMSO. The ISO was solved in H 2 O. The DNA fragments encoding SNAP-tag, ADRB2 and CIB were inserted between HindIII and XhoI sites in pcDNA3.1/ V5-His (B) (Thermo Fisher Scientific, USA). The genes encoding CRY, mCherry and β-arrestin 2 were subcloned between BamHI and XbaI sites in pcDNA4/myc-His (B) (Thermo Fisher Scientific).

Confocal Fluorescence Microscopy.
Cells were cultured on 35-mm glass-bottomed dishes. To label the SNAP-tag, 0.5 μM SNAP-Surface Alexa Fluor 647 was added to the medium, and the cells were incubated for 30 min at 37 °C. After the cells were washed three times, the medium was exchanged by 2 ml of phenol red-free D-MEM containing HEPES, pH 7.4 (Wako Pure Chemical Industry). The cells were observed under a confocal microscope (IX-81, FV-1000D, Olympus). The images were acquired every 1 min. CRY activation was performed under a confocal microscope using a 20 mW 440-nm laser at 0.5% output with a scan speed of 2.0 μs/pixel. The fluorescence intensity of Arrestin CRY was evaluated using the image analysis software FV10-ASW (Olympus). The number of ADRB2 CIB or Mdm2 fluorescence spots was counted using ImageJ. The diameter to detect the spots was set at 0.5-5 μm.
Immunostaining of Early Endosomes, Late Endosomes and Lysosomes. HEK293 opt cells were cultured on 35-mm glass-bottomed dishes. ADRB2 CIB was labeled with SNAP-Surface Alexa Fluor 647 for 30 min and then washed three times with D-MEM. The cells were cultured in D-MEM and stimulated with blue light at an intensity of 3 mW/cm 2 using an LED device (TH-211 × 200BL, Creating Customer Satisfaction, Japan) for the indicated times. The cells were fixed using 4% formaldehyde and then permeabilized with 0.2% Triton X-100. To avoid the non-specific adsorption of antibodies, the cells were treated with 0.2% gelatin from cold-water fish skin (Sigma-Aldrich) for 1 h at room temperature. Rab5, Rab7 and LAMP1 were labeled with the corresponding antibodies in PBS(+) overnight at 4 °C. After the cells were washed three times with PBS(+), they were stained with Alexa Fluor 488-conjugated goat anti-Rabbit IgG Secondary Antibody (Thermo Fisher Scientific) for 1 h at room temperature. The cells were washed three times with PBS(+) and then observed under a confocal microscope.

Calculation of Manders' Colocalization Coefficient.
The extent of the colocalization of ADRB2 CIB with lysosomes was estimated using Manders' colocalization coefficient.
where FI i is the fluorescence intensity of Alexa Fluor 647 molecules attached to ADRB2 CIB in the pixel i, and FI jcoloc = 0 if the fluorescence intensity of Alexa 488, which is attached to LAMP1, in the pixel j is lower than a particular threshold, and FI jcoloc = FI j if the fluorescence intensity of Alexa 488 in pixel j is equal to or larger than the threshold. The coefficient was calculated using the JACoP plug-in for ImageJ after applying a threshold of fluorescent intensity to each time-lapse image 42 . The thresholds were automatically determined based on the JACoP program.
Quantification of ADRB2 CIB on the Cell Surface Using the ELISA Assay. HEK293 opt cells in 4-well plates were treated with 1.0 μM ISO or irradiated with blue light at 3 mW/cm 2 for the indicated time. After the cells were fixed with 4% formaldehyde, they were treated with gelatin from cold-water fish skin for blocking. SNAP-tag moieties on the cell surface were labeled with an anti-SNAP-tag antibody for 60 min at room temperature. After the cells were washed twice with PBS(+), ECL anti-rabbit antibody linked to HRP was added to each well. The cells were incubated for 60 min at room temperature and then washed three times with PBS(+). After the addition of 350 μl of TMB solution (Thermo Fisher Scientific) and 30 min of incubation, 350 μl of 2 M sulfuric acid was added to each well. The OD 450 value of 100 μl of solution was measured using a microplate reader (Bio-Rad, USA). The amount of ADRB2 CIB on the cell surface was defined by (OD − OD mock )/(OD basal − OD mock ) × 100, where OD mock and OD basal correspond to the OD of non-stimulated HEK293 cells and the OD from non-stimulated HEK293 opt , respectively. To promote recycling of ADRB2 CIB in ISO-treated cells, cells were washed three times with 500 μl of acidic buffer (150 mM NaCl and 5 mM acetic acid) and then twice with 500 μl of PBS(+) to eliminate ISO.
Detection of Ubiquitinated ADRB2 CIB Using Immunoprecipitation. HEK293 opt cells that had been cultured on a 4-well plate were stimulated with 3 mM/cm 2 blue light using the LED device or 1.0 μM ISO for the indicated time. The cells were harvested using 200 μl of lysis buffer (50 mM HEPES (pH 7.5), 0.5% NP-40, 250 mM NaCl, 2 mM EDTA, 10% glycerol, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml leupeptin and MG-132 proteasome inhibitor). Anti-V5 antibody (Life Technologies) was added to each lysate, which was then gently shaken at 4 °C for 1 h. After protein G Sepharose (GE Healthcare, USA) was added, the lysate was shaken at 4 °C for 1 h. The precipitate was washed three times with 1 ml of lysis buffer and dissolved in 100 μl of sample buffer (125 mM Tris, pH 6.8, 10% glycerol, 4% SDS, 0.006% bromophenol blue and 10% mercaptoethanol). The samples were heated at 65 °C for 10 min and then subjected to SDS-PAGE using 10% acrylamide gels. The proteins were transferred onto nitrocellulose membranes (Bio-Rad, USA). To block the non-specific binding of antibodies, the membrane was treated with 1% skim milk in Tris-buffered saline for 1 h. ADRB2 and ubiquitin were blotted using their corresponding antibodies. The blotted bands were detected using an image analyzer (ImageQuant LAS-4000, GE Healthcare), and the intensity of each band was measured with an ImageQuant TL software. Statistical Analysis. Statistical significance was determined using unpaired Student's t-tests (two-tailed) and Bonferroni post-hoc test. Differences with P values < 0.05 were considered statistically significant.