Abstract |
The type 1 receptor (CRF1R) for the corticotropin-releasing factor (CRF) is a family B G
protein coupled receptor (GPCR) (1), which plays a key role in the maintenance of homeostasis
by regulating neural and endocrine functions (2,3). Malfunction of CRF/CRF1R systems is related
with several diseases such as depression and anxiety. The clinical importance of CRF1R is further
supported by that CRF1R-selective non-peptide small-molecule antagonists have been shown to
display anxiolytic and antidepressant properties in animal models (4).
As all GPCRs, the CRF1R consists of an extracellular amino-terminal domain (N-domain)
and seven α-helical transmembrane domains (TM1-TM7), which are connected extracellularly
with three loops (EL1-EL3). The CRF and its related peptides, such as sauvagine, bind to the
extracellular regions of CRF1R and activate the receptor (5,6). In contrast, small-molecule nonpeptide
antagonists have been proposed to interact with the TMs of CRF1R and allosterically
antagonize peptide-agonist binding and receptor activation (7). Recently our laboratory has
shown that, similar to other GPCRs, the TMs of CRF1R fold such as to form a water-accessible
crevice (binding-site crevice) within the plasma membrane (8). Amino acids of TMs that contact
non-peptide CRF analogues should be on the surface of the binding-site crevice of CRF1R.
Moreover, TM residues in family A GPCRs have been shown to participate in networks of
interactions that play important role in receptor activation (9). However the molecular
mechanisms underlying CRF1R activation and its antagonism by non-peptide molecules are still
elusive.
The present study aims to 1) obtain structural information for the TMs of CRF1R,
starting from TM3, 2) determine structural changes associated with receptor activation, 3)
determine TM residues that interact with non-peptide CRF antagonists and 4) elucidate the
molecular mechanisms underlying antagonism of receptor activation by non-peptide CRF
antagonists. The TM3 has been selected because its tilted orientation in other GPCRs, relative
to membrane, allows its residues to establish key interactions with ligands, other TMs and with
G-proteins (10).
To obtain structural information for the TM3 of CRF1R, we identified its residues that
are located on the surface of the binding-site crevice of an inactive ligand-free state (apo-state)
of the full-length receptor. To achieve this we mutated, one at a time, the TM3 residues to Cys
IX
(engineered Cys) and determined their accessibilities by applying the Substituted Cysteine
Accessibility Method (SCAM) (11,12). Among the twenty-two TM3 residues those at positions
(upper half of TM3) 1893.26, 1923.29, 1933.30, 1953.32, 1963.33, 1993.36 and 2033.40 were found to be
located on the surface of the binding site crevice. These results suggest that the TM3 of the apo
inactive state of CRF1R is positioned such that the upper half of this helix participates in the
formation of a large binding site crevice, whereas the other half is tightly packed with the other
TMs. These results are in full agreement with those obtained from a crystallization study of a Nterminally
truncated inactive non-peptide ligand-bound state of CRF1R which has been
published very recently.
Among TM3 residues, those at positions 2033.40 and 2103.47 play crucial role in receptor
activation. Mutation of G2103.47 to Cys (G2103.47C) largely decreased the high binding affinity of
sauvagine, as well as, its potency to stimulate the accumulation of cAMP in cells expressing the
receptor. According to our molecular model, which was constructed based on the crystal
structures of the inactive glucagon receptor (GCGR) and CRF1R, the G2103.47 is located one
helical turn below M2063.43. The M2063.43 forms a hydrogen bond with N2835.54 of TM5 (13,14).
Mutation of G2103.47 to Cys adds a side chain at position 2103.47 which could form an additional
hydrogen bond with the side chain of N2835.54 that further strengthens the TM3-TM5 interface,
thus stabilizing the inactive conformation of receptor. These results suggest that movements of
TM3 and TM5 that take place during activation of family A GPCRs (15), might also occur in class
B receptors and are hampered by strengthening the TM3-TM5 interactions. This suggestion is
further supported by the fact that mutation of G2103.47 to Ala (G2103.47A) did not largely
decrease the binding affinity and potency of sauvagine. In contrast to Cys, Ala at 2103.47
position of CRF1R did not form a hydrogen bond with N2835.54, and thus having a much smaller
impact on receptor activation.
As G2103.47, F2033.40 in TM3 plays an important role in receptor activation. Mutation of
F2033.40 to Cys (F2033.40C) largely decreased the ability of receptor to adopt its active state and
to bind sauvagine with high affinity. Interestingly, addition of a Lys-like chain to F2033.40C after
its reaction with the positively charged reagent, MTSEA, restored sauvagine binding to normal
(high affinity) levels. Similarly, addition of a Lys side chain at position 2033.40 by mutating
F2033.40 to Lys (F2033.40K) did not significantly alter the high binding affinity of sauvagine. As
F2033.40K substitution, mutation of F2033.40 to Trp (F2033.40 W) or Ile (F2033.40 I) did not decrease
the high affinity binding and potency of sauvagine. In marked contrast, removing the
X
hydrophobic side chain at position 2033.40 by mutating F2033.40 to Ala (F2033.40 A) reduced the
binding affinity and potency of sauvagine. Based on these results and comparing the inactive
crystal structure of CRF1R with the inactive and active crystal structures of the family A GPCR, β2
adrenergic receptor (β2AR), we propose that F2033.40 in TM3 and L3236.44 in TM6 of CRF1R, form
an aromatic-aliphatic interaction in the inactive state of receptor. As in β2AR, activation of
CRF1R is associated with a repositioning of F2033.40 relative to L3236.44 and towards a nearby
aromatic amino acid, which possibly interacts with F2033.40, Trp (F2033.40 W) or Ile (F2033.40 I) but
not with Ala (F2033.40 A). Such an aromatic-aromatic interaction most likely stabilizes the high
affinity binding active state of CRF1R and it is supported by that F2033.40K mutation did not alter
sauvagine binding. Lys (such as K2033.40) has been shown to participate in cation-pi interactions
with aromatic residues (16,17).
Despite the fact that the hydrophilic substitution F2033.40K did not alter the ability of
CRF1R to adopt its active state, this modification abolished the binding of the hydrophobic nonpeptide
antagonist, antalarmin to receptor. Similarly, F2033.40A mutation abolished antalarmin
binding to CRF1R. In marked contrast, the hydrophobic substitutions F2033.40W and F2033.40I did
not decrease the binding and antagonistic properties of antalarmin. These results suggest that
F2033.40 interacts with non-peptide CRF antagonists, as also observed in the crystal structure of
CRF1R in complex with the non-peptide antagonist CP-376395 (13). Interaction of non-peptide
CRF antagonists with TMs of CRF1R and intercalation between them most likely hampers
receptor activation-associated movements of these regions.
Conclusively the present study provided structural information for the CRF1R and
elucidated the molecular mechanisms underlying receptor activation and its antagonism by
small non-peptide molecules. In parallel this study designed, synthesized and pharmacologically
evaluated several non-peptide CRF analogues. Among the 50 compounds tested, four showed
promising binding affinities for CRF1R and were selected as lead compounds for the design of
novel molecules. All the results of the present study will put the basis for the design of novel
non-peptide CRF1R-selective antagonists with improved pharmacodynamic and
pharmacokinetic properties that will enrich the pharmaceutical armoire against CRF1R-related
disorders such as depression and anxiety.
|