(Received for publication, October 16, 1996)
From the A region between residues 38 and 42 of the human
cholecystokinin-A (CCK-A) receptor was shown to be involved in the
binding of CCK but not in that of JMV 179 and JMV 180, two peptides
closely related to CCK (Kennedy, K., Escrieut, C., Dufresne, M., Clerc, P., Vaysse, N., and Fourmy, D. (1995) Biochem. Biophys. Res.
Commun. 213, 845-852). In the present study, we have identified the
residues of both the receptor and the ligand responsible for this
differential binding. Residues Trp-39 and Gln-40 of the receptor were
crucial for binding of the C-terminal nonapeptide of CCK as W39F and
Q40N mutants demonstrated parallel decreases in both affinity and
potency to induce accumulation of inositol phosphates (12.9- and
20.9-fold). The W39F and Q40N mutant receptors bound CCK analogues
modified at their C-terminal end, including JMV 179 and JMV 180, as
well as the C-terminal amidated heptapeptide of CCK, with identical affinities to the wild-type receptor. In contrast, W39F and Q40N mutants bound CCK octapeptide with the same decreased affinity as the
CCK nonapeptide. The modeling of the CCK-A receptor and the docking of
the peptide agonists [Thr,Nle]CCK9 and CCK-8 indicated that their N
terminus was connected to the receptor through a strong bond network
involving Trp-39 and Gln-40 thus confirming experimental data. These
first molecular data identifying the agonist binding site of the human
CCK-A receptor represent an important step toward the complete
delineation of the agonist binding site and the understanding of the
molecular mechanisms that govern differential activation of this
receptor by CCK-related peptides.
The peptide cholecystokinin (CCK)1 is
found throughout the gastrointestinal system and the central nervous
system where it functions as both a hormone and a neurotransmitter (1).
CCK was originally isolated from porcine intestine as a 33-amino acid peptide and has subsequently been shown to exist physiologically in
multiple forms derived from the cleavage of a 115-amino acid preprohormone (2, 3). The major physiological forms are CCK-58, CCK-39,
CCK-33, and CCK-8, so designated by the number of amino acids counted
backwards from the commonly shared C-terminal (1). Post-translational
processing of CCK involves sulfation of the tyrosine at position 7 from
the C-terminal and
The actions of CCK are mediated by specific high affinity
membrane receptors on target cells. These receptors have been
characterized pharmacologically and biologically and are divided into
two subtypes based on their affinities for CCK and gastrin (5, 6). The CCK-A subtype has an approximately 500-fold higher affinity for CCK
than for gastrin, and the CCKB/G subtype has the same high affinity for
both CCK and gastrin. The cloning of the cDNA coding for these
receptors has shown them to belong to the superfamily of G
protein-coupled receptors that are characterized by seven transmembrane
domains connected by intracellular and extracellular loops with an
extracellular N-terminal and intracellular C-terminal (7, 8) (Fig.
2).
The numerous physiological effects of CCK and its possible role in some
pathological disorders have stimulated research into the design of
selective peptide and non-peptide antagonists (9). Some molecules were
synthetized on the basis of the importance of the C-terminal residues
of CCK, in particular the carboxyamidated phenylalanine, for full
biological activity. Two such molecules, JMV 180 and JMV 179, which are
closely related analogues of the C-terminal heptapeptide of CCK
(Fig. 1), have been used extensively in the
characterization of CCK-A receptors. JMV 180 acts as an agonist or an
antagonist according to the biological model and concentration of
peptide used. For example JMV 180 acts as a CCK agonist to induce
amylase release by pancreatic acini at nanomolar concentrations, while
it acts as an antagonist on the CCK-induced phase of inhibition of
amylase release at micromolar concentrations (11). Evidence exists from
pancreatic acinar cells and transfected Chinese hamster ovary cells to
show that the JMV 180 induces distinct signal transduction pathways
from the natural agonist CCK through binding to the CCK-A receptor.
However, the molecular basis for such distinct mechanisms of receptor
activation by these two closely related agonists remains unknown and is
still the object of considerable debate (12-16).
JMV 179 which differs from JMV 180 by the presence of a
D-Trp in the place of the L-Trp is a full
antagonist of the CCK-A receptor (17). Characterization of the rat
pancreatic CCK-A receptor using this full peptide antagonist led to the
identification of a truncated form of this receptor, with an altered
pharmacology (18). Peptide mapping of the labeled truncated protein
showed the receptor to be lacking in its extracellular N terminus. We recently constructed N-terminally truncated human CCK-A receptors that
enabled us to identify a region of the N terminus, close to the first
putative transmembrane domain that is essential for the high affinity
interaction with CCK but was nonessential for the binding with JMV 179 and JMV 180 (19).
This observation and the fact that to date there is no information on
the amino acids that compose the ligand binding site of the CCK-A
receptor prompted us to mutate the conserved residues in the region
38-42 of the human CCK-A receptor. By doing so, we have been able to
identify for the first time two residues, Trp-39 and Gln-40, that
interact with CCK but not with JMV 179 and JMV 180. Screening of large
variety of peptide analogues of CCK for binding to the mutated CCK-A
receptors together with the three-dimensional modeling of both the
receptor and the receptor-ligand complex enabled us to demonstrate that
the two residues Trp-39 and Gln-40 interact with the N-terminal region
of CCK octa- and nonapeptides that are not present in JMV 179 and JMV
180.
The C-terminal nonapeptide of CCK,
[Thr28,Ahx31]CCK-25-33 ([Thr,Nle]-CCK-9),
the heptapeptide [Nle]CCK-7, the octapeptide CCK-8, and the
decapeptide CCK-10 were synthesized by Luis Moroder (Max Planck
Institut für Biochimie, München, Germany) (20). The other
C-terminal analogues of CCK, namely JMV 170, 179, 180, 195, 211, 262, and 295 and Boc-[Nle]CCK-7 (see Table II for complete primary
structures), were synthesized by Jean Martinez's group. (1-[2-(4-(2-Chlorophenyl)thiazol-2-yl) aminocarbonyl indoyl] acetic acid) (SR27,897) and its tritiated derivative,
[3H]SR27,897 (41 Ci/mmol), were donated by Sanofi
Research (Toulouse, France) (21). 125INa was from Amersham,
Les Ulis, France. [Thr,Nle]CCK-9 and JMV 179 were conjugated with
Bolton-Hunter reagent, purified, and radioiodinated as described
previously (22). The specific activity of radioiodinated peptides was
1600-2000 Ci/mmol. Oligonucleotide primers were from Bioprobe Systems
(Montreuil, France). All other chemicals were obtained from commercial
sources.
INSERM U151, Institut Louis Bugnard, CHU
Rangueil, Bat. L3, 31054 Toulouse Cedex, the § Laboratoire
de Chimie Théorique,
Max Planck Intitut
für Biochemie, 82143 Martinsried, Federal Republic of Germany,
and the ** Sanofi-Recherche, 195 Route d'Espagne,
31036 Toulouse Cedex, France
-amidation of the C-terminal phenylalanine
residue (1). Studies using chemically synthesized fragments have shown
that the C-terminal sulfated, amidated heptapeptide (Fig. 1) is
essential for full biological activity; however, fragments as small as
the C-terminal pentapeptide which CCK has in common with the related
peptide gastrin and the C-terminal tetrapeptide retain biological
activity (4).
Fig. 1.
Ligands of the CCK-A receptor. The
synthetic agonist, [Thr,Nle]CCK-9, is closely related to the
natural ligand CCK-8. However, the two methionines are substituted by
Thr and Nle residues in order to avoid oxidation which leads to a loss
of the biological activity of CCK. JMV 180 corresponds to the
heptapeptide of CCK in which the phenylalanylamide residue is
substituted by a phenylethyl ester. In JMV 179, the
phenylalanylamide residue and the L-tryptophan are substituted by a phenylethyl ester and a D-tryptophan,
respectively. The highly potent and selective non-peptide antagonist
(1-[2-(4-(2-chlorophenyl)thiazol-2-yl)aminocarbonyl indoyl]
acetic acid) (SR27,897) is from Sanofi Recherche.
[View Larger Version of this Image (19K GIF file)]
Fig. 2.
Serpentine drawing of the human CCK-A
receptor. The segment from Glu-38 to Ala-42 in the N-terminal
region of the protein (drawn in square bracket), close to
the first putative transmembrane domain, which has been previously
identified as being involved in the binding of CCK was subjected to
site-directed mutagenesis. In the inset is shown a sequence
alignment indicating that all five amino acids except Pro-41 are
conserved in the four species.
[View Larger Version of this Image (43K GIF file)]
Materials
CCK-related
peptides
Receptors
Wild-type
W39F
Q40N
Ki, nM
(n)
Ki, nM
(n)
F
Ki,
nM (n)
F
Arg-Asp-Tyr(SO3H)-Thr-Gly-Trp-Nle-Asp-Phe-CONH2
[Thr,Nle]CCK-9
0.24
± 0.01 (4)
3.1
± 0.8 (5)
12.9
5.0 ± 1.1 (5)
20.8
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-DPhe-CONH2
JMV
203
47 ± 17 (3)
56 ± 22 (3)
1.2
54
± 20 (3)
1.1
Boc-Tyr(SO3H)-Nle-Gly-DTrp-Nle-Asp-O-CH2-CH2-
JMV
179
4.8 ± 0.1 (3)
4.8 ± 0.2 (3)
1.0
4.9
± 0.1 (3)
1.0
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-O-CH2-CH2-
JMV
180
6.0 ± 0.8 (3)
5.9 ± 1.2 (3)
1.0
6.2
± 1.5 (3)
1.0
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-CH2-CH2-
JMV
170
13.3 ± 1.1 (3)
12.1 ± 1.3 (3)
0.9
14.3
± 1.6 (3)
1.1
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-Phe-OOCH3
JMV 295
22.8
± 8.2 (3)
32.2 ± 11.3 (3)
1.4
44.3
± 7.2 (3)
1.9
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-
Ala-CONH2
JMV
211
768 ± 150 (3)
589 ± 68 (3)
0.8
768
± 125 (3)
1.0
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-CONH2
JMV 195
397
± 43 (3)
467 ± 61 (3)
1.2
414 ± 49 (3)
1.0
Boc-Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-Phe-CONH2
Boc-[Nle]CCK-7
0.75
± 0.25 (3)
1.15 ± 0.80 (3)
1.5
0.86
± 0.33 (3)
1.1
Tyr(SO3H)-Nle-Gly-Trp-Nle-Asp-Phe-CONH2
[Nle]CCK-7
2.8
± 0.8 (3)
2.5 ± 0.9 (3)
0.9
3.3 ± 1.5 (3)
1.2
Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-CONH2
CCK-8
0.27
± 0.02 (3)
3.2 ± 2.1 (5)
11.8
4.7
± 1.8 (5)
17.4
We have previously
identified, through the construction of receptors truncated to varying
degrees in their N terminus, a region of the human CCK-A receptor,
amino acids 38-42 inclusive, that is essential for the high affinity
interaction with the agonist [Thr,Nle]CCK-9 (19). Our initial aim was
to identify the individual amino acid(s) in this region that
interact(s) with the CCK. We decided to use site-directed mutagenesis
of evolutionarily conserved amino acids to assess their involvement in
the interaction with CCK. Alignment of the sequences of all CCK-A
receptors cloned to date showed that all five amino acids in this
region are conserved between human, guinea pig, and rabbit and that the
only difference in the rat receptor is a serine residue at position 41 compared with a proline in the human, guinea pig, and rabbit receptors (Fig. 2). Assuming that the binding site of the natural
ligand is most likely to be conserved during evolution, we mutated all the amino acids in this region with the exception of the proline which
is likely to have an important structural role rather than a direct
involvement in binding. We started with conservative mutations of the
residues: Glu-38 Asp, Trp-39
Phe, Gln-40
Asn, and Ala-42
Val. In a second step, the amino acids that were found to be
critical for CCK binding were exchanged by chemically distinct
residues:Trp-39
Ile and Gln-40
Glu.
Site-directed mutagenesis was carried out using a Chameleon 228 Double-stranded Site-directed Mutagenesis kit (Stratagene) following the manufacturer's instructions. The protocol is based on the method of mutagenesis by unique site elimination (23). Mutations were introduced into the human CCK-A receptor cDNA cloned into the pRFENeo vector (19) using mutagenic primers based on the published human CCK-A receptor cDNA sequence (7, 8). Selection primers mutated a unique SmaI restriction site in the cDNA to a unique NruI site and vice versa. Plasmids were isolated from individual colonies and screened for the incorporation of the desired mutations by restriction site analysis. The mutations were confirmed by sequencing either manually, using the Double-stranded DNA Cycle Sequencing System (Life Technologies, Inc.) or on an automated sequencer (Applied Biosystem).
Transient Transfection of COS-7 Cells and Membrane PreparationCOS-7 cells (1.2 × 106) were plated
in 10-cm culture dishes and grown in Dulbecco's modified Eagle's
medium containing 5% fetal calf serum in a 5% CO2
atmosphere at 37 °C. After an overnight incubation, cells were
transfected with 2 µg/plate of the pRFENeo vectors containing the
cDNA for the wild-type or mutated CCK-A receptors, using a modified
DEAE-dextran method as described previously (24). Approximately 65 h post-transfection the cells were washed three times with
phosphate-buffered saline, pH 6.95, scraped from the plate in 10 mM Hepes buffer, pH 7.0, containing 0.01% soybean trypsin
inhibitor, 0.1% bacitracin, 0.1 mM phenylmethylsulfonyl fluoride and frozen in liquid N2. After thawing at
37 °C, the cells were subjected to another cycle of freeze/thawing
and then centrifuged at 25,000 × g for 20 min. The
membrane pellet was resuspended in 50 mM Hepes buffer, pH
7.0, containing 115 mM NaCl, 5 mM
MgCl2, 0.01% soybean trypsin inhibitor, 0.1% bacitracin, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride
(binding buffer), aliquoted, and stored at 80 °C until use.
Membrane protein concentrations were determined by the Bradford method
(25) using the Bio-Rad protein assay kit.
Isolated plasma membranes from COS-7 cells (3-10 µg protein) were incubated in binding buffer containing 0.5 mg/ml bovine serum albumin for 90 min at 25 °C (steady state conditions) with 60-100 pM 125I-BH-[Thr,Nle]CCK-9 or 240 pM [3H]SR27,897 in the presence or absence of competing agonist or antagonist. The binding reaction was stopped by the addition of 500 µl of ice-cold binding buffer, and bound radioligand was separated from free by centrifugation at 10,000 × g for 10 min at 4 °C. Pellets were washed and centrifuged once more, and the radioactivity was directly counted for the radioiodinated ligands or the pellet was resuspended in 0.1 M NaOH and added to scintillant and counted for the tritiated radioligand. Nonspecific binding was determined in the presence of 10 µM unlabeled ligand and was less than 10% of total binding. Binding data from at least three separate experiments from different batches of transfected cells were analyzed using the EBDA LIGAND program (26). Data are expressed as mean ± S.E.
Assay of Inositol PhosphatesTo measure the accumulation of inositol phosphates, COS-7 cells were transfected and cultured in Dulbecco's modified Eagle's medium (3 ml) containing serum in 25-cm2 plates in the presence of 2 Ci/ml myo-2-[3H]inositol (Amersham Corp.) for 48-72 h. The day of the experiments, the medium was replaced by serum- and inositol-free medium, and cells were incubated for 30 min at 37 °C. Cells were washed with serum- and inositol-free medium and were again incubated for 10 min at 37 °C in the presence of 10 mM LiCl. Finally, cells were stimulated for 20 min with or without [Thr,Nle]CCK-9, JMV 180, JMV 179, or SR27,897 at various concentrations. The reaction was stopped by the addition of 10% HClO4, and inositol phosphates were extracted on AmprepTM mini-columns (Amersham Corp.) following the manufacturer's instructions.
Computer Modeling of the ReceptorThe molecular model of
the human CCK-A receptor was constructed using the transmembrane
helical positions found in the bacteriorhodopsin crystal structure as a
starting point for the optimization procedures and using successively
the human rhodopsin and the 2-adrenergic receptor for
sequence alignment procedures. Extracellular and intracellular loops
connecting the transmembrane helices were then added to the preliminary
7-helix bundle and modeled with the use of a simulated annealing
procedure. The entire system was relaxed, with the possible translation
and rotation movements of individual transmembrane helices taken into
account. Details of the building of such models have been presented
elsewhere, for example for the AT1A receptor (27). The
homology, energy minimization, and dynamics steps were produced with
the use of the Biosym molecular modeling software. Molecular modeling
was performed independently from the experimental work to avoid any bias in subsequent analysis of the results.
The human wild-type and mutant CCK-A receptors were transiently expressed in COS-7 cells (Table I). We first determined the binding properties of the wild-type receptor using both the agonist radioligand 125I-BH-[Thr,Nle]CCK-9 and the non-peptide antagonist radioligand [3H]SR27,897. Scatchard analysis of the binding results obtained with 125I-BH-[Thr,Nle]-CCK-9 demonstrated two classes of binding sites, a high affinity site with a Kd of 0.34 ± 0.01 nM and a maximal binding capacity of 0.43 ± 0.06 pmol/mg proteins, and a low affinity site with a Kd of 109 ± 56 nM and a maximal binding capacity of 5.6 ± 0.4 pmol/mg proteins. In contrast, binding of the selective non-peptide antagonist [3H]SR27,897 revealed a single class of binding sites with a Kd of 1.55 ± 0.15 nM and a binding capacity of 15.5 ± 1.5 pmol/mg proteins (not illustrated).
|
The effects of mutations on the pharmacological properties of the
mutant receptors were studied by competition binding experiments. The
two mutations E38D and A42V had no effect on the binding of [Thr,Nle]CCK-9 nor on the binding of the antagonist SR27,897 (Table I). However, the mutants W39F and Q40N demonstrated significantly decreased binding for [Thr,Nle]CCK-9 (Fig. 3 and Table
I). The exchange of Trp-39 for a Phe caused a 11.9-fold decrease in the affinity for [Thr,Nle]CCK-9 without affecting the affinity of the
non-peptide antagonist, and the exchange of Gln-40 for an Asn produced
a 19.2-fold decrease in the affinity for [Thr,Nle]CCK-9 while the
affinity for SR27,897 remained unchanged (Table I and Fig. 3). For both
mutants W39F and Q40N, Scatchard analysis of agonist binding revealed
two classes of binding sites having the following dissociation
constants: W39F mutant, Kd1 = 3.8 ± 0.9 nM, Kd2 = 1330 ± 550 nM; Q40N mutant, Kd1 = 6.3 ± 1.4 nM, Kd2 = 1283 ± 480 nM (not illustrated). These dissociation
constants indicate that the W39F and Q40N mutations decreased the
affinity of [Thr,Nle]CCK-9 for both the high affinity and low
affinity sites to the same extent.
Since interactions of CCK with Trp-39 are likely to involve hydrophobic
interactions with the indole moiety, we verified that such an
interaction was completely eliminated by the exchange of a tryptophan
to a phenylalanine by performing the mutation Trp-39 Ile. This new
W39I mutant bound the agonist [Thr,Nle]CCK-9 with an affinity close
to the W39F mutant (Kd, 2.1 ± 0.60 nM versus 3.10 ± 0.30 nM)
suggesting that any contribution to the affinity of the agonist ligand
was already eliminated by the substitution of the indole ring of Trp-39
by the aromatic ring of Phe-39. This suggested that the integrity of
the indole ring at site Trp-39 in the human CCK-A receptor was in fact
required for the full interaction of [Thr,Nle]CCK-9.
Interactions of the CCK with Gln-40 are likely to involve hydrogen bonding, and the mutation Q40N consisted only in a reduction of the length of the side chain without changing the functionality. We therefore tested the possibility of restoring the affinity of the mutant receptor by exchanging the Asn for a Glu, thus restoring the length of the side chain and maintaining the possibility of hydrogen-bonded interactions. The Q40E mutant bound [Thr,Nle]CCK-9 with an affinity 7.6-fold higher than the Q40N mutant and only 2.7-fold lower than the wild-type CCK-A receptor. Affinity of the non-peptide antagonist SR27,897 for W39I and Q40E mutants remained unchanged compared with the wild-type CCK-A receptor (Table I).
Functionality of the Mutant ReceptorsTo analyze the effects
of the mutations of Trp-39 and Gln-40 on the functional coupling of the
human CCK-A receptor, we first studied the dependence of agonist
binding on activation of G proteins by GTP[S] a non-hydrolyzable
analogue of GTP. Binding of the agonist
125I-BH-[Thr,Nle]CCK-9 to the high affinity sites of the
wild-type, W39F, and Q40N receptors was inhibited in the same manner by
GTP[
S] (IC50, 2.1 nM, not shown), whereas
binding of the non-peptide antagonist [3H]SR27,897 was
unaffected. We then measured [Thr,Nle]CCK-9-induced accumulation of
inositol phosphates. As illustrated in Fig. 4, [Thr,Nle]CCK-9 induced a dose-dependent stimulation
of the production of inositol phosphates to a maximal increase of
8-10-fold in COS-7 cells transfected with the wild-type and mutant
receptors. However, the concentrations giving half-maximal responses
(kact) were significantly higher for the W39F
and Q40N mutants (15.0 ± 1.8 and 22.6 ± 2.1 nM,
n = 3) than for the wild-type CCK-A receptor (1.0 ± 0.5 nM, n = 4). Comparison of these
values with the Kd values for binding of
[Thr,Nle]CCK-9 to the corresponding receptors indicated that the
increase in kact values obtained for both the
W39F and Q40N mutants paralleled the loss in [Thr,Nle]CCK-9 binding
affinities. Indeed, Ki/kact
ratios were 0.23, 0.21, and 0.22 for the wild-type, W39F, and Q40N
receptors, respectively, indicating that the two mutant receptors were
as effectively coupled to phospholipase C as is the wild-type CCK-A
receptor.
The two CCK-related analogues JMV 179 and JMV 180 as well as the non-peptide compound SR27,897 tested at concentrations of up to 1 µM did not stimulate the production of inositol phosphates in COS-7 cells expressing the wild-type or the mutated receptors (not illustrated).
Identification of the Residues of CCK Interacting with Residues Trp-39 and Gln-40 of the CCK-A ReceptorIn a previous study, we showed that deletion of the first 42 extracellular amino acids of the CCK-A receptor did not affect the binding affinities for JMV 180 and JMV 179 which are structurally related CCK analogues (Fig. 1). This observation and the data presented here on the effects of mutating Trp-39 and Gln-40 on the binding of [Thr,Nle]CCK-9 led us to attempt to determine which amino acid(s) and chemical function(s) present in [Thr,Nle]CCK-9, but absent in JMV 179 and 180, were likely to interact with Trp-39 and Gln-40 in the CCK-A receptor. Since the primary structure of JMV 180 differs from that of [Thr,Nle]CCK-9 by the substitution of the C-terminal amidated phenylalanyl residue by a phenylethyl ester, and by the presence of the blocking group butyloxycarbonyl instead of amino acids Arg and Asp, we screened a series of CCK analogues modified at the C terminus or varying in their N terminus by the length of the peptide chain, for binding at the wild-type, W39F, and Q40N receptors. We postulated that only peptides containing residues capable of interacting with the amino acids Trp-39 and Gln-40 of the receptor would bind to the mutated receptors W39F and Q40N with decreased affinities compared with the wild-type receptor. The results of these binding experiments are shown in Table II.
JMV 180 as well as JMV 179 bound with identical affinities to the wild-type and W39F and Q40N mutant receptors, a result that is in agreement with our previous findings (19). Peptide analogues which had an N terminus identical to JMV 180 and JMV 179 and had their C-terminal amidated phenylalanine modified (JMV 170, 211, 203, 295) or absent (JMV 195) bound with identical affinities to the wild-type and mutant receptors. Furthermore, the agonist Boc-[Nle]CCK-7 and [Nle]CCK-7 which have the native amidated phenylalanine at the C terminus as does [Thr,Nle]CCK-9 but have a sulfated tyrosine at the N terminus also bound with identical affinities to the wild-type and mutant receptors. These data suggested that neither the lateral phenyl ring of the phenylalanine nor the C-terminal amide were interacting with Trp-39 and Gln-40 in the CCK-A receptor.
In contrast, CCK peptides that had a peptide chain extended backwards from the sulfated tyrosine, which is peptides longer than the heptapeptide of CCK, and their C-terminal end intact bound with different affinities to the wild-type and mutated CCK-A receptors. Indeed, compared with the wild-type receptor, the affinity of CCK-8 was decreased 11.8-fold for the W39F mutant and 17.4-fold for the Q40N mutant, and the affinity of [Thr,Nle]CCK-9 was decreased 12.9-fold for the W39F mutant and 20.9-fold for the Q40N mutant suggesting that residues located at the N terminus of these peptides were likely to be interacting with residues Trp-39 and Gln-40 of the binding site of the CCK-A receptor.
Comparison of the affinity of the wild-type receptor for CCK analogues indicated that the carboxyamidated phenylalanine contributes 1660-fold to the affinity of [Thr,Nle]CCK-9 and -CCK-8, whereas the Asp residue in CCK-8 and the fragment Arg-Asp in [Thr,Nle]CCK-9 contribute identically 10-12-fold to the affinity of [Thr,Nle]CCK-9 and -CCK-8. This observation agrees with previous data that the CCK-7 is 10 times less potent for stimulation of amylase secretion at CCK-A receptors (4).
Molecular Model for the Interaction Between the N-terminal Moiety of CCK and Residues Trp-39 and Gln-40 of the Human CCK-A ReceptorThe positioning of the CCK peptide in the modeled CCK-A
receptor was achieved using a docking model already proposed for the CCK-B/gastrin receptor (28, 29). In this model, the C-terminal moiety
of the CCK-related ligands was placed in the middle of the receptor
transmembrane region, whereas the N-terminal part was positioned in its
extracellular region near the entrance of the receptor pocket.
According to the model illustrated here for the docking of
[Thr,Nle]CCK-9 in the CCK-A receptor, it appears that the N-terminal
residues of [Thr,Nle]CCK-9 are connected to the receptor through a
strong hydrogen bond and salt bridge network (Fig. 5).
As previously demonstrated by circular dichroism (30), intramolecular
salt bridges are found inside the ligand itself connecting together on
one hand the guanidinium and carboxylate groups of the side chains of
Arg-1 and Asp-2 and on the other hand the Tyr-3 sulfate and the
N-terminal ammonium. These groups are also involved in intermolecular
interactions with receptor residues. Asp-106 forms a salt bridge
between its side chain and Arg-1 side chain of [Thr,Nle]CCK-9. Gln-40
forms H bonds between its NH amide side chain and the Asp-2 side chain
of [Thr,Nle]CCK-9 and between its carbonyl side chain and the
N-terminal ammonium of the ligand. Trp-39 forms H bonds between its
backbone carbonyl and the ligand N-terminal ammonium and between its
backbone NH and the Tyr-3 sulfate oxygen of [Thr,Nle]CCK-9
(Fig. 6).
Moreover, the correct positioning of the ligand Tyr-3 side chain and, as a consequence, of the whole of the N-terminal ligand moiety in the receptor is ensured by the position of the Trp-39 side chain; this side chain is maintained in the correct position in the receptor by an intramolecular H bond between the Trp-39 indole and the Glu-38 side chain and by an aromatic T-shaped interaction between the side chains of Trp-39 and Phe-198 residues. Owing to the restrained flexibility of its side chain, Trp-39 is involved in another T-shaped aromatic interaction with the Tyr-3 aromatic ring of [Thr,Nle]CCK-9. Such aromatic-aromatic interactions have proven to be very important for the stabilization of proteins or in ligand-protein interactions (31).
From the above model, residues Gln-40 and Trp-39 appear to be key residues for the binding of the agonists [Thr,Nle]CCK-9 and CCK-8. Data with W39F and Q40N which demonstrated a parallel loss in both the binding affinity of these receptor mutants for [Thr,Nle]CCK-9 and their potency to induce inositol phosphate production are in perfect agreement with the data from the molecular modeling. Indeed, in the model, the exchange of the Gln-40 residue by an Asn one which shortens the side chain of the residue strongly reduces the possibility of interactions with the carboxylate side chains of Asp-2. Exchange of Gln-40 to Glu also yields decreased possibilities of docking due to electrostatic repulsive interactions between the two negative charges born by the carboxylate side chains of Asp-2 and Glu-40. The same situation is observed with the mutations of Trp-39 to Phe and Ile; when the Trp-39 side chain is replaced by that of the Phe residue, the Phe side chain moves from a T-shaped to a stacked interaction with the aromatic ring of Phe-198, making an unfavorable positioning of the ligand Tyr-3 side chain owing to unfavorable steric interactions between the sulfate oxygens and the Phe-39 aromatic ring. When the Trp-39 planar side chain is replaced by a more bulky hydrophobic side chain such as the Ile, our docking model is modified so that the salt bridge and hydrogen bond network as described above are less favorable.
The model for the docking of the agonist peptides in the CCK-A receptor which is illustrated for [Thr,Nle]CCK-9 is in good agreement with binding data for CCK-8 and CCK-7. Binding experiments demonstrated that mutations of Trp-39 and Gln-40 induced a similar decrease in the affinity for [Thr,Nle]CCK-9 and CCK-8, whereas no decrease was observed for CCK-7. It therefore appears that the overall energy involved in the interactions between the N-terminal ammonium and the two adjacent amino acids Arg-Asp of [Thr,Nle]CCK-9 and amino acids Trp-39 and Gln-40 of the receptor is nearly the same as that between the N-terminal ammonium and the Asp residue in CCK-8. Elimination of the Asp residue as in CCK-7 completely abolishes the possibility of interaction of the ligand with Trp-39 and Gln-40 (Fig. 6).
The human CCK-A receptor mediates important biological functions of CCK such as the regulation of appetite, the stimulation of exocrine pancreatic secretion, and gallbladder motility (7, 8). The cDNA for the receptor was cloned 3 years ago; however, to date, there exist no molecular data on the amino acids of the receptor that are involved in ligand binding. We have recently shown that a region of the receptor between amino acids 38 and 42 was essential for the high affinity interaction with the full agonist CCK; this observation was the starting point for studies to identify the individual amino acids involved in agonist binding (19).
Using site-directed mutagenesis of evolutionarily conserved amino acids, we have been able to identify two amino acids of the human CCK-A receptor, Trp-39 and Gln-40 which, when mutated to other amino acids of the same chemical function, lead to a loss in the affinity of the receptor for the full agonists [Thr,Nle]CCK-9 and CCK-8. The loss in affinity for [Thr,Nle]CCK-9 of the mutant receptors did not result in a loss in the biological efficacy of [Thr,Nle]CCK-9 at the receptors as shown by the identical maximum production of inositol phosphates; however, the decrease in affinity was accompanied by a similar loss in the biological potency as shown by parallel shifts in the dose-response curves for inositol phosphate production. This observation, and the fact that the binding of other peptide and non-peptide ligands was unaffected by the mutations, suggests that mutation of these two amino acids directly affects a region of the binding site for CCK on the receptor without affecting the overall integrity of the receptor. Truncation of the CCK-A receptor while affecting the binding of [Thr,Nle]CCK-9 did not affect the binding of the CCK analogues JMV 179 and JMV 180 suggesting that these last molecules do not interact with this region of the receptor (19). This was confirmed by the mutagenesis studies presented here; JMV 180 and JMV 179 are very similar structurally to [Thr,Nle]CCK-9 but do not interact with Trp-39 and Gln-40. This observation opened up the possibility of using a combination of the screening of a series of other CCK analogues and carrying out further mutations to identify the residues and chemical functions involved in the interaction between [Thr,Nle]CCK-9 and Trp-39 and Gln-40. Screening of a series of CCK analogues structurally modified at their N and C termini led to the observation that only peptides that were N-terminally extended further than the sulfated tyrosine residue were sensitive to mutations at residues 39 and 40 of the receptor, suggesting that it is in fact the N-terminal ammonium and Arg-1-Asp-2 residues of [Thr,Nle]CCK-9 and the N-terminal ammonium and Asp-1 of CCK-8 that interact with Trp-39 and Gln-40. Molecular modeling of the CCK-A receptor and the docking of the agonists [Thr,Nle]CCK-9 and CCK-8 in the receptor gave a model which suggests that Trp-39 and Gln-40 of the CCK-A receptor are connected to the N-terminal region of the agonist ligand via a network of hydrogen bonds and salt bridges. The results obtained from the molecular modeling confirm those obtained experimentally.
Further support for our findings lies in the good agreement between the effect of elimination, by site-directed mutagenesis, of a specific point of interaction in the receptor binding site, namely residues Trp-39 and Gln-40, and the effect of the elimination in the ligand of the residues, in this case Arg-Asp or Asp, involved in the interaction with the wild-type receptor.
The fact that JMV 179 and JMV 180 do not contain N-terminal residues capable of interacting with Trp-39 and Gln-40 may explain in part that they bind with lower affinities to the receptor than [Thr,Nle]CCK-9 and CCK-8. It is likely that JMV 179 and JMV 180 still share a minimum part of their binding site with the corresponding peptide agonists. The results from this study do not allow us to draw such a conclusion, but it is worth noting that the binding of the highly selective non-peptide CCK-A receptor antagonist SR27,897 was also unaffected by mutations of residues 39 and 40, indicating that these amino acids are not involved in the recognition and binding of this ligand. Such a minimal overlapping between the binding site for agonist and non-peptide antagonist is now well established for many G protein-coupled receptors (32-33).
Furthermore, our data lead to the conclusion that the differences in N-terminal structure between [Thr,Nle]CCK-9 and JMV 179 and 180 and therefore lack of interactions at Trp-39 and Gln-40 are not responsible for the absence of inositol phosphate production by these two peptides as observed in this study. Indeed, if such structural differences at the N terminus of the peptides were the cause of absence of inositol production by the wild-type receptor under stimulation by JMV 180 and JMV 179, we would expect that the two receptors mutated at positions Trp-39 and Gln-40 would not respond to a stimulation by [Thr,Nle]CCK-9. This was not the case. Therefore, it is highly probable that the differences in intracellular signaling previously observed between JMV 180 and the full agonist CCK are not due to stuctural differences at the N terminus of the peptides (10, 12-16). Further studies concentrating on the identification of the residue(s) of the CCK-A receptor that interact(s) with the C-terminal amidated phenylalanine residue of CCK shown to be vital for agonist activity are necessary.
This study shows the importance of interactions with residues located in the N-terminal region of the CCK-A receptor in the binding of the natural agonist CCK. Interestingly, the location of at least a part of the binding site for CCK on the extracellular surface of the CCK-A receptor which is suggested by the data from both the site-directed mutagenesis and the molecular modeling shows similarity to the binding sites of receptors for other peptides such as the AT1 receptor (34), the endothelin-B receptor (35), the neurokinin-1 receptor (36), and the neurotensin receptor (37) in contrast to those for the biogenic amines that have been demonstrated to be exclusively within the transmembrane domains (33, 38, 39).
In the absence of any other data on the CCK-A receptor, it is interesting to compare our mutagenesis results with data on the CCK-B/gastrin receptor, as the two receptor subtypes have the same high affinity for the agonists CCK-8 and [Thr,Nle]CCK-9 (5, 6). No mutagenesis experiments involving the N-terminal region of the CCK-B/gastrin receptor have been reported; however, the importance of this region was demonstrated through the discovery of a splice variant that was deleted of the N-terminal extracellular region and the upper part of the first transmembrane domain and had an altered pharmacology (40). In addition, a segment of five amino acids in the second intracellular loop of the CCK-B/gastrin receptor was shown to be essential for the high affinity of this receptor for the natural peptide agonist gastrin, suggesting that determinants of the binding site of the CCK-B/gastrin receptor are situated within the extracellular domains (41). In contrast, mutations along transmembrane domains of the CCK-B/gastrin receptor that were shown to affect binding affinity for non-peptide antagonists did not significantly affect the affinity for agonist ligands (28, 29, 42).
In conclusion, using site-directed mutagenesis of the CCK-A receptor, analysis of binding affinity, and biological potency of C- and N-terminally modified CCK analogues and molecular modeling, we have identified two amino acids Trp-39 and Gln-40 of the agonist binding site of the receptor. We have also demonstrated that these two amino acids interact with the N-terminal moiety of CCK. These findings represent an important step toward the complete delineation of the agonist binding site of the CCK-A receptor and in the identification of the interactions involved, and toward the subsequent understanding of the molecular mechanisms that govern differential activation of this receptor by CCK-related peptides.