Nature | Vol 579 | 12 March 2020 | 301structure shows a large twist in the orientation of the finger loop, due
to the N-terminal segment of the phosphopeptide inducing a twist in
the flanking β-strands S5 and S6. This V2Rpp segment must disengage
before the finger loop can be inserted into M2Rpp and is not ordered
in our map. The positioning of the N-terminal V2Rpp segment in the
crystal structure could be due to lattice packing. Alternatively, these
differences could reflect a multi-step binding mechanism, in which
the initial binding of βarr1 to the full length of the phosphopeptide
is followed by release of the N terminus of the phosphopeptide, flip-
ping of the βarr1 finger loop, and burial of the hydrophobic residues
of the finger loop in the intracellular cavity of the receptor (Extended
Data Fig. 6a).
The M2Rpp ICL2 rests in a cleft between the N- and C domains of
βarr1, which comprises portions of the finger, middle, gate and C-loops.
Although this portion of the map lacks well-resolved side-chain fea-
tures, the density suggests that ICL2 adopts the same helical conforma-
tion as in M2R–Go (Fig. 3c). This ICL2 orientation places L12934.51 facing
a βarr1 hydrophobic cleft that is capable of accommodating even larger
side chains (Extended Data Fig. 6b). Although leucine is most prevalent
at this position in other class A GPCRs that couple to β-arrestins, larger
hydrophobic amino acids (such as methionine and phenylalanine) are
common (Extended Data Fig. 6c), suggesting one source of the promis-
cuity of β-arrestins. Notably, the same hydrophobic ICL2 residue often
engages in van der Waals interactions with G proteins^3 ,^23 –^25.
Membrane interaction of βarr1 C domain
The cryo-EM maps reveal that the C-edge of βarr1 interacts with the
nanodisc, with one C-edge loop contacting the membrane (loop 1,
residues 191–196, LMSDKP) and a second loop burying itself in the lipid
bilayer (loop 2, residues 330–340, SRGGLLGDLAS) (Extended Data
Fig. 7). Interaction of the C-edge of arrestin with the membrane has
been suggested by the structure of Rho–Arr1, and by site-specifically
labelling the C-edge of visual arrestin with environmentally sensitive
probes^12 ,^19. Given that the affinity of βarr1 for the 7TM core of M2Rpp is
higher in nanodiscs than in detergent (Fig. 1b, c), we proposed that there
are functional consequences to the C-edge–lipid interaction. First, we
performed multiple, independent molecular dynamics simulations^21
with M2Rpp–βarr1 initially embedded in either a continuous membrane
(enabling C-edge interaction) or a 9-nm-diameter nanodisc (too small
for C-edge interaction) (Fig. 4a, Extended Data Fig. 8a). In four of the
five simulations using the full membrane, βarr1 remained in its active
conformation with an average interdomain twist angle of around 17°—
similar to the interdomain twist observed in the crystal structure of
the βarr1–V2Rpp–Fab30 complex^7. However, in small nanodiscs, βarr1
adopted a more heterogeneous conformational ensemble dominated
by inactive conformations (with an interdomain twist angle of between
0° and 7°) (Fig. 4b); this suggests that the interaction between the
C-edge and the membrane stabilizes the active conformation of βarr1.
We site-specifically labelled C-edge loop 2 with mBr (L338mBr) to
assess whether the βarr1–lipid interaction requires the binding of βarr1
to the M2Rpp phosphorylated tail or the 7TM bundle^19. An increase in
fluorescence of βarr1 L338mBr, which reflects insertion of the C-edge
into a hydrophobic environment, requires M2Rpp phosphorylation
and nanodisc reconstitution (Fig. 4c, Extended Data Fig. 8b). This is
consistent with previous observations that the engagement of visual
arrestin with membranes requires rhodopsin phosphorylation^19. How-
ever, the change in βarr1 L338mBr fluorescence is not agonist (iperoxo)-
dependent, which indicates that the C-edge–lipid interaction does not
require interaction between βarr1 and the 7TM bundle. Because the
C-edge–membrane interaction is observed in nanodiscs but not in
detergent (Fig. 1b, c), we suggested that it might facilitate βarr1–7TM
coupling. We sought to test this by attenuating the hydrophobicity of
the C-edge with aspartic acid substitutions at L335, L338 and S340 of
βarr1 (βarr1(3×D))^19. Fluorescence of βarr1 in which the finger loop was
labelled with bimane (Fig. 4d, Extended Data Fig. 8c) confirmed that
the 3×D mutation markedly reduced the interaction of βarr1 with the
7TM bundle of HDL-M2Rpp. Thus, maximal coupling of βarr1 to the
7TM bundle of M2Rpp requires not only receptor phosphorylation but
also interaction of the C-edge of βarr1 with the phospholipid bilayer.
Furthermore, we investigated whether the interaction between the
C-edge of βarr1 and the membrane modulates βarr1-mediated receptor
desensitization and internalization. We measured the ability of purified
wild-type βarr1 and βarr1(3×D) to inhibit (desensitize) HDL-M2Rpp-
mediated G-protein activation in vitro. Iperoxo-induced activation of
HDL-M2Rpp increases the GTPase activity of purified heterotrimeric Gi.
Whereas wild-type βarr1 blocks 50% of Gi activity, βarr1(3×D) has almost
no effect (Fig. 5a). We then compared the ability of GFP-tagged wild-type
βarr1 (GFP– βarr1(WT)) and GFP–βarr1(3×D) to induce the internaliza-
tion of Flag-tagged M2R (Flag–M2R) in βarr1/βarr2-null HEK293 cells.
After 5 minutes of iperoxo stimulation, GFP–βarr1(WT) translocates
from the cytoplasm to the plasma membrane. After 30 minutes, both
GFP–βarr1(WT) and Flag–M2R (more than 40%) are largely internalized
(Fig. 5b, c, Extended Data Fig. 9a, b). By contrast, GFP–βarr1(3×D)—
which is expressed at similar levels to the wild-type (Extended Data
Fig. 9c, d)—remains in the cytoplasm even after 30 minutes, with little
Flag–M2R internalization (10%). Notably, the impaired recruitment
of GFP–βarr1(3×D) to wild-type M2R and V2R (Extended data Fig. 9e)
shows that membrane anchoring contributes to the recruitment of βarr1
to natively phosphorylated GPCRs, not only the engineered M2Rpp
construct. Thus, the interaction of βarr1 with the membrane is critical
for desensitization and internalization of the receptor.Three-site interaction model
The binding of arrestin to GPCRs has been viewed as a two-part process:
binding to the phosphorylated ICLs or the C terminus of the receptorZ = 0(–) Membrane(+) Membraneβarr1 C-edge4254504755005250.000.250.500.751.001.251.50Wavelength (nm)Relative uorescence intensityIpxIpxIpxAtrpAtrpAtrpControlM2RM2Rpp4254504755005250.00.51.01.52.02.5Wavelength (nm)Relative uorescence intensity3 ×D3 ×D3 ×DWTWTWTControlM2RM2Rppβarr*Lc d βarr1 nger loopabβarr1 interdomain twist (o)Frequencyβarr1 activationβarr*LM2RN domainC domain 0.000–10–5 05101520250.0250.0500.0750.100Fig. 4 | Anchoring of the C domain of βarr1 to the lipid membrane.
a, b, Atomistic simulations of βarr1 in complex with M2Rpp (in the absence of
Fab30) to investigate the position of the C-edge of βarr1 (a) and the βarr1
interdomain twist angle in the absence (dark blue) or presence (cyan) of a lipid
membrane (dashed black line in a) (b). c, HDL-M2Rpp, but not HDL-M2R,
enhances the f luorescence of bimane at the C-edge of βarr1 (inset, red star)
independent of the presence of antagonist (atropine) or agonist (iperoxo).
d, The ability of iperoxo-activated HDL-M2Rpp to increase the f luorescence of
bimane in the finger loop of βarr1 (inset, red star) requires receptor
phosphorylation (compare with HDL-M2R) and is markedly reduced by the
C-edge 3×D mutations (L335D/L338D/S340D). Data are representative of three
independent experiments.