Discussion
In this study, we intended to find out the critical residues of μ-opioid
receptor governing ligand bias by integrating computational simulations
and cell assay based bias analysis. We first tried to predict the key
residues influencing ligand bias at MOR by computer modeling. Then the
site-directed mutation on MOR was conducted and CHO cell lines stably
expressing wild-type and mutant MORs were established in order to
characterize the binding affinities and functional activities of
morphine and fentanyl derivatives with wild-type and mutant MORs. Based
on results from functional assays, the bias profiles of morphine and
fentanyl derivatives were obtained by operational model analysis. We
found that D3.32 and H6.52 were critical residues in driving morphine
and fentanyl derivatives to bind with μ-opioid receptor. W6.48 was a
pivotal residue in governing the bias signaling of MOR. The interactions
among ligand and W6.48, Y7.43 were responsible for the signaling bias.
The stabilization of Y7.43 by ligand and the stable interactions of
ligand with W6.48 and Y7.43 were propitious to the β-arrestin signal,
while the stabilization of W6.48 is propitious to biased G-protein
signal.
Recent years, unremitting efforts continually put into the research
about structures of GPCR. People were trying to illustrate the mechanism
underlying the activation of receptor through analysis of the structural
characteristics of active receptors. It is indicated that the
ligand-induced comformation of receptor is conducive to the bias
signaling(24 ,25 ). In this work, we found that
the residue W6.48 and Y7.43 were paired activation switches of ligand
bias at MOR, which is consistent with the computational result of Chen’s
work(26 ). And we proposed that the
interaction between W6.48 and Y7.43 with fentanyl derivatives lead to
two contrary states of transmembrane helix. When H6 moved towards H5,
the “close” state impeded the activation of G-protein; When H6 moved
far away from H5, the “open” state is in favour of the
G-protein-dependent signaling. This is in agreement with the previous
structural researches of MOR and other GPCRs. It is reported that,
compared with the structure of
MOR-Nb39(15 ), the TM6 of MOR-Gi is
outward from TM5(14 ). The
structures of other active GPCRs, such as β2 adrenergic
receptor and glucagon like peptide-1 receptor, revealed that activation
always involves the relative outward movements of TM5 and TM6, which
allows cavity structure for accommodating the α subunit of
G-protein(27-29 ).
Opioids were used to alleviate or cure pain for centuries, but the
opioid side-effects still limited their use in clinic. A great number of
efforts were put in to solve the fatal hypoventilation and other adverse
effects, but no completely ideal ligand was found. People have ever
tried to design µ or κ subtype-selective
ligands(30 ,31 ), as well as allosteric
modulators(32 ), to obtain a
pain-killer without those unwanted effects, but the result is not
satisfactory. Recently, the bias agonism stands a chance of separating
analgesia and adverse effects for opioid ligands, the discovery of
potential drugs TRV130 and PZM21 was a good signal. However, there is
still a long way to go. On one hand, the mechanism of biased signaling
was not clear enough. We are not sure which structural features
contribute to the specific conformation of the receptor that facilitate
G-protein-biased signaling. On the other hand, there is controversy in
the quantification and assessment of ligand bias, the method of bias
analysis was not compeletely developed. In addition, TRV130 was not
approved to use clinically for its little difference of hypoventilation
versus morphine in phase Ⅲ
study(33 ). In this study, we found
the critical sites affecting ligand bias and tried to explain the
mechanim of ligand bias at MOR, which contributed to better design and
discovery of opioid ligands with reduced adverse effects.