Arrestin2-clathrin-N interactions detected by arrestin2 NMR spectra.

(A) AlphaFold2 model of arrestin21-418 (arr2) showing the N- and C-domain (gray) followed by the clathrin-binding loop (CBL, orange), β20 strand (yellow) and the disordered C-terminal tail (blue). The model was used to visualize the clathrin-binding loop and the 344-loop of the arrestin2 C-domain, which are not detected in the available crystal structures of apo arrestin2 [bovine: PDB 1G4M (Han et al., 2001), human: PDB 8AS4 (Isaikina et al., 2023)]. In the other structured regions, the model is virtually identical to the crystal structures. The C-terminal strand β20 forms a parallel β-sheet with the N-terminal strand β1 (pink) in inactive arrestin2 and is released from arrestin2 upon activation by phosphopeptides. Arrestin2 residues interacting with clathrin-N are shown as green (red for main interacting residue I377) spheres. The clathrin and AP2 binding motifs are indicated in the arrestin2 sequence on the top. (B) Part of 1H-15N TROSY spectrum of apo arrestin21-393 (black) and upon addition of an equimolar amount of clathrin-N (green). Assigned CBL arrestin21-393 residues are indicated. (C) Intensity ratios of assigned complexed vs. apo arrestin21-393 CBL resonances. The region between T374 and T381 undergoes significant intensity attenuation. (D) 1H-13C HMQC spectra of apo Ile-δ1-13CH3, 2H-labeled arrestin21-393 (black) and upon addition of an equimolar amount of clathrin-N (green). (E) Intensity attenuation (top) and chemical shift perturbation (bottom) of arrestin21-393 isoleucine 1H3-13Cδ1 resonances upon clathrin-N addition. Out of 15 isoleucine residues, significant changes are observed only for I377. (F) Left: part of 1H-13C HMQC spectrum showing resonance shifts of I377 upon clathrin-N binding. Right: detected chemical shift changes as a function of clathrin-N concentration. The solid line depicts a non-linear least-square fit to the data points with respective dissociation constant.

Arrestin2-clathrin-N interactions detected by clathrin-N NMR spectra.

(Left): intensity attenuation I/I0 of clathrin-N 1H-15N TROSY resonances upon addition of equimolar amounts of various arrestin2 constructs: (A) apo arrestin21-393, (B) CCR5pp6-activated arrestin21-393, (C) full-length arrestin21-418 and (D) arrestin2ΔCBL with removed clathrin-binding motif. Regions that undergo significant attenuation upon interaction with arrestin2 are indicated by arrows. Right: residues undergoing significant intensity attenuation (one standard deviation below the average signal attenuation) are marked on the structure of clathrin-N (PDB: 3GD1) together with a schematic representation of the respective arrestin2 construct.

Arrestin2 interaction with the C-terminal domain of the AP2β2.

(A) Structure of AP2β2701- 937 (orange) in complex with arrestin C-terminal peptide (blue) (PDB: 2IV8). Important residues from both chains stabilizing the interaction are depicted in stick representation. (B-F) SEC profiles of arrestin21-418, AP2β2701-937, various phosphopeptides and their mixtures. The annotated color coding below the SEC profile indicates the individual sample composition. In panels (B, C, E) ‘mixture’ refers to a SEC sample containing all of the individual components indicated on the left. In panels (D, F) the primary sample composition is indicated above the SEC profile, and the annotated color coding below the SEC profile indicates the added component to the primary sample composition. (G) Apparent affinities for arrestin2-AP2 complex formation in the presence of phosphopeptides derived by integrating the SEC peaks marked in Figure 4F and scaling the integrals by the respective extinction coefficients. ‘nd’ indicates ‘not detectable’. (H, I) (Left): intensity attenuation I/I0 of AP2β2701-937 1H-15N TROSY resonances upon addition of equimolar amounts of apo arrestin21-418 (H) or CCR5pp6-activated arrestin21-418 (I). Regions that undergo extensive specific attenuation upon interaction with arrestin2 are indicated by arrows. Right: residues undergoing extensive specific intensity attenuation are marked on the structure of AP2β2 (PDB: 2IV8) together with a schematic representation of the respective arrestin2 construct.

CCR5 internalization in the presence of the chemokine ligands.

CCR5 internalization monitored in HeLa cells co-transfected with plasmids containing arrestin2-YFP (green) and CCR5 genes (magenta) stimulated with (A) [5P12]CCL5 (antagonist), (B) CCL5 (natural agonist), (C) [6P4]CCL5 (super-agonist). Cells were stimulated for 0, 10, and 30 min with chemokine ligands before fixation and preparation for fluorescence microscopy. Each image is accompanied by a zoomed region of interest (white squares) showing deconvolved CCR5 and arrestin2-YFP signals. (D) Mander’s colocalization coefficients of the CCR5 and arrestin2-YFP signals in the absence and presence of chemokine ligands 30 min after stimulation. Individual values, as well as mean and standard deviation are shown for N=3 biological replicates and n=45 ROIs from 20 cells; ****: P < 0.0001; ns: not significant (P > 0.9999). (E) Density of CCR5-positive (CCR5+) and CCR5/arrestin2-positive (CCR5+/arrestin2+) puncta after 30 min [6P4]CCL5 ligand stimulation. Individual values, as well as mean and standard deviation are shown for N=3 biological replicates and n=90 ROIs from 30 cells; ns: not significant (P > 0.9999). (F) CCR5 and arrestin2 fluorescence signals along the trajectory from the plasma membrane to the nucleus (blue line in panel c) 30 min after [6P4]CCL5 ligand stimulation. (G) Lysosomal trafficking of the CCR5 (magenta) and arrestin2 (cyan) complex in presence of [6P4]CCL5 monitored in the HeLa cells using LAMP1 antibodies (yellow). No recruitment to the lysosome is observed.

Dependence of CCR5 internalization on arrestin2 interactions with clathrin or AP2 monitored in HeLa-arr2/3 cells.

(A) CCR5 internalization induced by 60 min [6P4]CCL5 ligand stimulation in HeLa cells co-transfected with plasmids containing CCR5 and arrestin2-YFP, arrestin2-YFP(LIELD, arrestin2-YFPR395A, arrestin2-YFP(AP2 or empty pcDNA3.1. (B, C) Internalization was quantified by counting the number of puncta (B) positive for CCR5 and arrestin2 (CCR5+/arrestin2+) or (C) only CCR5+ 60 min after ligand stimulation in the HeLa-arr2/3 cells. No significant difference in CCR5 internalization is detected for the arrestin2-YFP(LIELD and arrestin2-YFPR395A constructs (ns, not significant: P > 0.9999) whereas the absence of the AP2 binding motif in the arrestin2(AP2-YFP construct causes a significant (****: P < 0.0001) reduction of CCR5 internalization 60 min after ligand incubation. Mean and standard deviation are shown for N=3 biological replicates and n=90 ROIs from 30 cells.

Overall scheme of arr-class A and B GPCR internalization.

Schematic difference of arrestin2-mediated internalization of arr-class B (left) vs. arr-class A (right) GPCRs. Arr-class B GPCRs bind stably to arrestin due to their high levels of phosphorylation. This results in a robust release of the arrestin C-terminus, a stable interaction with AP2 and formation of a long-lived GPCR•arrestin complex. Arr-class A GPCRs bind weakly to arrestin due to their poor phosphorylation. They require stabilization of the arrestin complex by membrane-bound PIP2 molecules. The arrestin C-terminus is not fully released and consequently the interaction with AP2 is unstable.

CCR5 phosphopeptide sequences and their affinities towards arrestin2.

Underlined serine or threonine residues are phosphorylated.

Arrestin sequence alignment and analysis of the arrestin2-clathrin-N complex structure.

(A) Sequence alignment of four human arrestin types. The clathrin-binding motif (orange) and AP2 binding motif (purple) are indicated within the sequences. The numbering on the right corresponds to the last amino acid in the alignment. (B) X-ray structure of the arrestin2-clathrin complex (PDB:3GD145). The structure shows two arrestin chains (green and blue) interacting with one clathrin-N molecule (red). Interaction sites 1 and 2 are indicated by dashed rectangles with enlarged views on the right. The arrestin2 residues involved in the interaction are shown as sticks. Site I: between CBL of arrestin2 (green) and the edge of β-sheet blade 1 and 2 of clathrin-N. CBL residues are shown as sticks. The numbering follows the PDB deposition. Site II: between arrestin2 splice loop (blue) and β-sheet blade 4 and 5 of clathrin-N.

Mapping of clathrin-N interaction on arrestin2.

(A) 1H-15N TROSY spectrum of arrestin21-393 in apo (black) or upon equimolar addition of clathrin-N (green) recorded on a Bruker AVANCE 14.1 T (600 MHz) spectrometer equipped with a TCI cryoprobe at 303 K. Assigned resonances are marked by blue amino acid numbers. Unassigned resonances are marked by black arbitrary numbers. (B) Intensity reduction I/I0 of arrestin21-393 1H-15N TROSY resonances upon clathrin-N interaction. The dashed line is drawn at one standard deviation below the average I/I0 value. The resonance numbering follows panel (A).

Mapping of arrestin2 interaction on clathrin-N.

1H-15N TROSY spectra of clathrin-N in the presence of one molar equivalent of (A) arrestin21-393 (cyan); (B) arrestin21-393•CCR5pp6; (C) arrestin21-418 or (D), arrestin2 ΛCBL superimposed on 1H-15N TROSY spectrum of apo clathrin-N (black). The spectra were recorded on a Bruker AVANCE 21.1 T (900 MHz) spectrometer equipped with a TCI cryoprobe at 303 K. Assigned residues undergoing significant line broadening are indicated on the spectra.

Arrestin2 interaction with the C-terminal domain of AP2β2 detected by NMR.

(A, B) 1H-15N TROSY spectra of AP2β2701-937 in the absence or presence of one molar equivalent of either arrestin21-418 (A) or arrestin21-418•CCR5pp6 (B). Assigned residues undergoing significant intensity reduction are indicated on the spectra. (C, D) Test of interaction of arrestin21-418 with AP2β2701-937. (C) 1H-13C HMQC (Ile-δ1-13CH3, 2H-labeled arrestin21-418) and (D) 1H-15N TROSY (15N, 2H-labeled arrestin21-418) spectra of arrestin21-418 in the absence or presence of one molar equivalent of AP2β2701-937. (E, F) Test of interaction of arrestin21-418 with CCR5pp6. (E) 1H-13C HMQC (Ile-δ1-13CH3, 2H-labeled arrestin21-418) and (F) 1H-15N TROSY (15N, 2H-labeled arrestin21-418) spectra of arrestin21-418 in the absence or presence of a ten-molar equivalent of CCR5pp6.

Arrestin2 interaction with clathrin-N and AP2β2 monitored by SEC.

(A) SEC profiles of arrestin21-418, clathrin-N and their mixture. (B) as in (A) but in the presence of AP2β2701-937 and CCR5pp6. (C) SDS-PAGE of sample from the SEC complex fraction (taken at 2.6ml) monitoring the formation of the arrestin2/AP2β2 complex for various conditions indicated on the top. The empty square indicates the expected position of clathrin-N.

Arrestin21-418 trypsin proteolysis assay.

Trypsin proteolysis of arrestin21-418 in apo form and in complexes with (A) CCR5pp6, (B) CCR5pp3 and (C) CCR5pp4 and AP2β2701-937 detected by SDS-PAGE. The incubation time is indicated on the top. ‘M’ indicates the protein size marker. The table on the right indicates sample composition and the relative abundance of arrestin21-418 after 5 min. incubation (see Materials and Methods). The activation by the CCR5 phosphopeptides accelerates arrestin21-418 proteolysis. The addition of AP2β2701-937 reverses this effect.

Uncropped SDS-PAGE gels of the arrestin21-418 trypsin proteolysis.

SDS-PAGE gels for the data in Figure 3–figure supplement 3. The composition of the sample and the reaction incubation time are indicated at the top. Protein marker bands are indicated on the left. Arrestin21-418 runs at ∼ 52 kDa, AP2β2701-937 at ∼ 30 kDa.

Arrestin2 internalization (control) and re-localization in live HeLa cells upon chemokine stimulation in the presence of CCR5.

(A) HeLa cells transfected only with arrestin2-YFP and stimulated with 6P4[CCL5]. In the absence of the receptor, arrestin2 stays distributed in the cell cytoplasm and nucleus. (B,C) Arrestin2-YFP re-localization in HeLa cells transfected with wild-type CCR5 and arrestin2-YFP genes. Arrestin2-YFP and LysoTrackerTM signals were used to monitor the protein trafficking for up to 30 min after CCL5 (B) or 6P4[CCL5] (C) incubation. No overlap between the LysoTrackerTM and arrestin2 signal has been detected. White squares indicate zoomed regions of interest.

Dependence of CCR5 internalization on the interactions of arrestin2 with clathrin or AP2 monitored in HeLa (CCL2) cells by immunofluorescence microscopy.

(A) CCR5 internalization induced by [6P4]CCL5 in HeLa cells co-transfected with plasmids containing either arrestin2-YFP or arrestin2ΛLIELD-YFP, which misses the clathrin-binding motif. No significant difference is detected as quantified by the Mander’s colocalization coefficient (ns, not significant: P > 0.9999) or by counting CCR5+/Arrestin2+ and only CCR5+ puncta 30 min after ligand incubation. Mean and standard deviation are shown for N=3 biological replicates and n=45 ROI from 20 cells (Mander’s) or n=90 ROI from 30 cells (puncta counting). (B) CCR5 internalization induced by 6P4[CCL5] in HeLa cells co-transfected with plasmids containing either arrestin2-YFP or arrestin2R395A-YFP, which bears a single mutation in the adaptin-binding motif. The mutation does not affect CCR5 internalization up to 30 min after 6P4[CCL5] stimulation. (C) Recruitment of arrestin2 ΔAP2-YFP to CCR5 in the HeLa cell plasma membrane upon [6P4]CCL5 ligand stimulation. (D) CCR5 internalization induced by [6P4]CCL5 in HeLa cells co-transfected with plasmids containing either arrestin2-YFP or arrestin2ΔAP2-YFP. Mean and standard deviation of the Mander’s coefficient and the number of puncta were determined as in (A). The absence of the AP2 binding motif in the arrestin2ΔAP2-YFP construct causes a significant (****: P < 0.0001) reduction of CCR5 internalization 30 min after ligand incubation. (E) SEC profiles of arrestin2ΔAP2, AP2β2701-937, CCR5pp6 and their mixture. No complex formation is observed.