(Received for publication, November 13, 1996, and in revised form, February 21, 1997)
From the Geneva Biomedical Research Institute, Glaxo
Wellcome, CH-1228 Geneva, Switzerland, the ¶ Laboratory for
Molecular Pharmacology, Department of Protein Chemistry, Molecular
Biology Institute, University of Copenhagen DK-2100, Denmark, the
Pfizer Central Research, Groton, Connecticut 06340, and the
** Université Pierre et Marie Curie, CNRS URA 493, F-75252 Paris, France
Ligand recognition of the NK1
receptor (substance P receptor) by peptide agonist and non-peptide
antagonist has been investigated and compared by the use of fluorescent
ligands and spectrofluorometric methods. Analogues of substance P
(SP) labeled with the environment-sensitive fluorescent
group 5-dimethylaminonaphthalene-1-sulfonyl (dansyl) at either position
3, 8, or 11 or with fluorescein at the
N position were synthesized and
characterized. Peptides modified at the
-amino group or at positions
3 or 11 conserved a relatively good affinity for NK1 and agonistic
properties. Modification at position 8 resulted in an 18,000-fold
decrease in affinity. A fluorescent dansyl analogue of the non-peptide
antagonist CP96,345 was prepared and characterized. The quantum yield
of fluorescence for dansyl-CP96,345 was much higher than for any of the
dansyl-labeled peptides indicating that the micro-environment of the
binding site is more hydrophobic for the non-peptide antagonist than
for the peptide agonists. Comparison of collisional quenching of
fluorescence by the water-soluble hydroxy-Tempo compound showed that
dansyl-CP96,345 is buried and virtually inaccessible to aqueous
quenchers, whereas dansyl- or fluoresceinyl-labeled peptides were
exposed to the solvent. Anisotropy of all fluorescent ligands increased
upon binding to NK1 indicating a restricted motional freedom. However, this increase in anisotropy was more pronounced for the dansyl attached
to the non-peptide antagonist CP96,345 than for the fluorescent probes
attached to different positions of SP. In conclusion, our data indicate
that the environment surrounding non-peptide antagonist and peptide
agonists are vastly different when bound to the NK1 receptor. These
results support recent observations by mutagenesis and cross-linking
work suggesting that peptide agonists have their major interaction
points in the N-terminal extension and the loops forming the
extracellular face of the NK1 receptor. Our data also suggest that
neither the C terminus nor the N terminus of SP appears to penetrate
deeply below the extracellular surface in the transmembrane domain of
the receptor.
Many peptide hormones and neuropeptides act via known receptors belonging to the superfamily of G protein-coupled receptors characterized by a seven membrane-spanning topology. There is considerable interest in understanding ligand-receptor recognition and the mechanisms of action of both non-peptide ligands and natural peptides for peptide receptors. The tachykinin substance P (SP)1 is a peptide transmitter that plays an important role in pain perception and neurogenic inflammation (1, 2). The cellular actions of SP are mediated by the tachykinin (neurokinin) NK1 receptor, a G protein-coupled receptor. Therefore, the NK1 receptor has been the target for the development of multiple non-peptide antagonists. The prototype NK1 non-peptide antagonist is the quinuclidine compound CP96,345, which acts as a high affinity and highly selective non-peptide inhibitor of SP in both binding and functional assays (3, 4).
Traditionally, the identification of binding domains for peptide
agonists and non-peptide antagonists has been investigated by
site-directed mutagenesis of receptors and by the construction of
chimeric receptors (5). In the case of NK1, studies by different groups
have localized a binding pocket for the antagonist CP96,345 in the
outer portion of the transmembrane domain of the receptor with presumed
contacts points clustering on transmembrane domains TM-III, -IV, -V,
VI, and VII and facing the interior of the seven-helix bundle (Fig.
1) (6-13). In contrast, the binding site
for substance P appears to involve multiple domains on the
extracellular side of NK1 including the N-terminal segment, the first
extracellular loop, and the top of TM-III and -VII (Fig. 1) (10-12).
One still open and debated question is whether substance P makes
additional contacts within the transmembrane domains. So far mutational
analysis of this region has failed to give any clear answer to this
(12, 13). The main problem with mutational mapping experiments is that
they do not necessarily reveal direct contact points between amino
acids on the receptor and ligand functional group. Loss of binding
affinity can be due to either a true contact residue or to indirect
allosteric effect on the receptor folding. Covalent labeling of
NK1 with photolabile SP analogues has identified agonist peptide-binding domains of NK1 in the N-terminal region and the second
extracellular loop using ligands labeled at either position 3 or 8 of
the peptide (14, 15, 35).
It is therefore very important to complement the mutational approach with biochemical and spectroscopic techniques. Recently, we developed novel fluorescence methodology to investigate ligand-receptor recognition and to map subdomains of binding sites with fluorescent ligands. In the NK2 receptor system, we demonstrated that agonist and antagonist peptides of similar molecular size have distinct binding sites (16), and we built a molecular model for ligand-receptor interactions using experimental determination of distances by fluorescence energy transfer between a fluorophore on the ligand and another fluorophore placed at specific sites in the receptor through biosynthetic incorporation by suppression of nonsense codons (17).
Here we have used site-specific fluorescent-labeled analogues of substance P at positions 1, 3, and 11 and a fluorescent CP96,345 analogue to probe the polarity and solvent accessibility of NK1 binding pockets and to measure the motional freedom of receptor-bound ligands. In this study we present spectrofluorometric evidence for the existence of different ligand binding pockets on the NK1 receptor for substance P and the antagonist CP96,345, respectively.
Synthesis of Fluorescent Ligands
NTo 2.34 mg of SP (1.73 µmol) in 360 µl of 50 mM sodium borate, pH 9.0, at 4 °C was added 42 µl of 50 mM dansyl chloride in acetone (2.07 µmol). The mixture was stirred for 2 h at 4 °C. The mono-dansylated SP derivative was recovered as the major peak by reverse-phase high performance liquid chromatography (Nucleosil 300-7 C8 column Machery-Nagel ET 250/8/4; gradient 0-75% of 0.01% trifluoroacetic acid in acetonitrile, in 0.01% trifluoroacetic acid in water over 60 min). Compounds were detected by absorbance at 214 and 350 nm. Fractions containing pure material were pooled and lyophilized. Electrospray MS: calculated, 1580.9; found, 1581.2. N-terminal Edman degradation and analysis of PTH derivatives were as follows: at cycle 1, Arg-PTH was recovered for both SP and Dns-Lys3-SP, whereas at cycle 3, Lys-PTH was recovered for SP only. The modification of Lys3 was further confirmed by compositional amino acid analysis.
NDns-Dap8-SP
was synthesized by solid phase peptide synthesis according to the
general methods previously reported (18) with the following
modifications. A base-labile resin linker preloaded with Fmoc-Met was
used. At the third cycle, Dap was coupled using N-Fmoc and side chain
t-butoxycarbonyl protections. After coupling, the
t-butoxycarbonyl group was removed with trifluoroacetic
acid, and the resin was then incubated with dansyl chloride in acetone. The peptide synthesis was continued. The peptide was cleaved from the
resin as C-terminal amide using liquid ammonia. Electrospray MS:
calculated, 1519.86; found, 1519.69.
Dns-Hcy11-SP was prepared as described by Chassaing et al. (19).
NTo 4.35 mg of SP (2.84 µmol) in 500 µl of 25 mM sodium borate, pH 9.0, at 4 °C was added
62 µl of 50 mM fluorescein isothiocyanate in
N,N-dimethylformamide. The mixture was stirred
for 3 h at 4 °C, and then 100 µl of
N,N-dimethylformamide and 500 µl of 0.2 M acetic acid were added. Monofluorescein SP was recovered
by high performance liquid chromatography (same conditions as above) as
the major peak of the mixture. Yield: 2.3 mg (47%). Electrospray-MS:
calculated, 1737.1; found, 1736.9. Edman degradation: cycle 1 yielded
Arg-PTH for SP but not for
[N-Flu-SP]; cycles 2 and 3 yielded
Pro-PTH and Lys-PTH respectively, for both SP and
N
-Flu-SP.
To
a 150-ml round-bottomed flask equipped with condenser and
N2 inlet were added 3.42 g (11.71 mmol) of
()-diphenylmethyl-1-azabicyclo[3.2.2]octan-3-amine, prepared as
described in Ref. 20, 2.51 g (15.23 mmol) of 5-nitroanisaldehyde,
prepared according to Ref. 21, and 60 ml of methanol. The solution was
stirred 40 min at room temperature, and the resulting precipitate was
collected by filtration and dried. It was then taken up in 46 ml of dry
tetrahydrofuran, 28 ml (56 mmol) of a 2.0 M solution of
borane methyl sulfide in tetrahydrofuran added, and the reaction
refluxed for 4 days. The reaction was cooled, evaporated, and taken up
in 50 ml of ethanol, treated with 5 g of sodium carbonate and
4 g of cesium fluoride. The resulting mixture was refluxed 3 days,
cooled, and taken up in methylene chloride and water. The organic layer
was separated, dried over sodium sulfate, and evaporated. The residue
was chromatographed on silica gel using methanol/methylene chloride as
eluant and then triturated with isopropyl ether to afford 2.098 g (42%
overall) of a white solid, m.p. 185-189 °C, after evaporation from
methylene chloride. 1H NMR (
, CDCl3): 1.30 (m, 1H), 1.57 (m, 1H), 1.66 (m, 1H), 1.97 (m, 1H), 2.10 (m, 1H), 2.68 (m, 1H), 2.84 (m, 2H), 2.97 (m, 1H), 3.30 (m, 1H), 3.40 Abq, J = 15,
= 64, 2H), 3.52 (s,
3H), 3.68 (dd, J = 8, 12, 1H), 4.49 (d,
J = 12, 1H), 5.80 (d, J = 2, 1H), 6.5-7.5 (m, 12H). 13C NMR (
, CDCl3): 20.1, 24.7, 25.5, 42.0, 45.7, 49.3, 49.6, 54.2, 55.9, 61.7, 111.5, 114.0,
116.6, 125.9, 126.2, 127.5, 127.7, 127.8, 128.4, 128.9, 129.1, 139.5, 143.6, 145.5, 150.7. MS (%): 428 (parent + 1, 1), 291 (20), 274 (18),
260 (100), 136 (65), 106 (28). High Resolution MS: calculated for
C28H33N3O, 427.2624; found, 427.26131. [
]D =
16.6° (c = 1, CH2Cl2).
![]() |
![]() |
![]() |
To a 30-ml round-bottomed flask equipped
with condenser and N2 inlet were added 100 mg (0.234 mmol)
()-diphenylmethyl-3-((2-methoxy-5-aminophenyl)methylamino)-1-azabicyclo[3.2.2]octane, 76 mg (0.281 mmol) of 5-dimethylaminonaphth-1-ylsulfonyl
chloride, and 5 ml of dry toluene. The reaction was heated at reflux
for 18 h, cooled, and diluted with ethyl acetate. The organic
layer was washed with 0.5 M aqueous sodium hydroxide and
brine and then dried over sodium sulfate and evaporated. The residue
was purified by flash chromatography on silica gel using
methanol/methylene chloride as eluant to afford the product as a light
yellow powder from methylene chloride/2-propanol, 64 mg (48%), m.p.
198-202 °C. 1H NMR (
, CDCl3): 1.25 (m,
1H), 1.42 (m, 1H), 1.60 (m, 1H), 1.78 (m, 1H), 1.92 (m, 1H), 2.6-2.8
(m, 4H), 2.86 (s, 6H), 3.15 (m, 1H), 3.22 Abq,
J = 12,
= 60, 2H), 3.40 (s, 3H), 3.61 (dd,
J = 8, 12, 1H), 4.40 (d, J = 12, 1H),
6.26 (broad s, 1H), 6.41 (d, J = 6, 1H), 6.77 (m, 1H),
7.0-7.3 (m, 12H), 7.44 (m, 1H), 7.60 (m, 1H), 8.15 (m, 1H), 8.42 (m,
1H), 8.51 (m, 1H). 13C NMR (
, CDCl3): 19.9, 24.7, 25.3, 41.9, 45.3, 45.4, 49.1, 49.4, 53.8, 55.4, 61.7, 110.2,
115.2, 118.7, 122.7, 123.1, 124.6, 126.0, 126.2, 127.5, 127.6, 128.3, 128.5, 128.8, 129.0, 129.7, 129.8, 130.2, 130.6, 134.5, 143.2, 145.3, 152.1, 155.5. MS (%): 644 (parent + 1, 1), 493 (40), 291 (100), 125 (70), 105 (75), 96 (60), 84 (70). High resolution MS: calculated for
C40H44N4O3S, 660.31341; found, 660.31281.
![]() |
![]() |
![]() |
Cell Culture
Fluorescence experiments were performed with either CHO cells stably transfected with the human NK1 receptor and expressing about 120,000 receptors per cell or CHO cells transiently expressing about 150,000 NK1 receptors per cell and obtained using the Semliki forest virus system as described (22). All cells were cultured as monolayers in a humidified 5% CO2 atmosphere at 37 °C, in Dulbecco's modified Eagle's medium/Ham's F12 medium supplemented with 10% fetal calf serum, 2% (w/v) Pen-Strep, and 0.5% (w/v) Nystatin. Cells were harvested at 80% confluence with PBS containing 1 mM EDTA and washed with ice-cold PBS. Membrane fractions were prepared as described previously (23).
For radioligand binding assays, the expression plasmids containing the receptor cDNAs were transiently transfected into COS-7 cells by the calcium phosphate precipitation methods as described previously (24).
Competition Binding Experiments
Mono-iodinated 125I-Bolton-Hunter-labeled substance P (125I-BH-SP) was prepared and purified as described (24). The transfected COS-7 cells were transferred to 12-well culture plates, 0.1-0.6 × 105 cells/well, 1 day after transfection and 24 h before performing the binding experiments. The number of cells per well was adjusted according to the expression efficiency of the individual plasmid aiming at 5-10% binding of the added radioligand in the competition binding experiments. Binding experiments were performed for 3 h at 4 °C with 50 pM 125I-BH-SP plus variable amounts of unlabeled peptide or non-peptide compound in 0.5 ml of a 50 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 5 mM MnCl2, and 0.1% (w/v) bovine serum albumin (Sigma) supplemented with 100 µg/ml bacitracin (Sigma). All determinations were performed in triplicate, and the nonspecific binding was determined as the binding in the presence of 1 µM SP. The specific binding constituted more than 80% total binding. The binding data were analyzed and IC50 values determined by computerized nonlinear regression analysis using InPlot (GraphPad Software, San Diego, CA).
Spectroscopy
UV-visible absorbance spectroscopy measurements were made using a Hewlett-Packard 8452A diode array spectrophotometer at ambient temperature. Steady-state fluorescence measurements were recorded using a Jasco FP-777 spectrofluorimeter with 0.2 × 1.0 cm and 0.5 × 1.0 cm quartz cuvettes with 0.2- and 0.5-ml sample volume, respectively. Unless stated otherwise, fluorescence measurements were made on cell membrane preparations at 20 °C. Solvents and buffers were degassed by bubbling nitrogen to prevent quenching of fluorescence by solubilized oxygen. The fluorescence emission signals were stable to photobleaching under measuring conditions. For dansyl-labeled ligands, the excitation wavelength was 343 nm, and the emission was recorded from 400 to 600 nm or at 530 nm in collisional quenching and anisotropy experiments. For fluorescein-labeled SP, excitation was at 475 nm and emission from 500 to 600 nm. For collisional quenching and anisotropy measurements, the emission was set at 520 nm. The excitation and emission bandwidth were 5 and 10 nm, respectively.
Fluorescence Experiments
Membrane suspensions from NK1/CHO cells were suspended in PBS,
pH 7.2, containing MgCl2 (3 mM), bovine serum
albumin (0.2 mg/ml), at a concentration of 5 nM
3H-SP binding sites and incubated with 50 nM
fluorescent ligand in the presence (nonspecific binding) or absence
(total binding) of 50 µM nonfluorescent ligand at
20 °C for 1 h. As an additional control, membranes were also
incubated in the absence of fluorescent ligand. Membranes were then
washed twice with ice-cold PBS, pH 7.2, resuspended in ice-cold PBS,
and homogenized by sonication in an ice-cold bath for 3 min immediately
prior fluorometry measurements. During fluorescence recordings,
membranes were kept in suspension by continuous stirring. Fluorescence
was corrected for light scattering and background fluorescence from
control samples. For dansyl compounds excited at 343 nm, the scattering
caused by irradiation of membranes particles was not important in the
520-540-nm emission region. For the fluorescein ligand
N-Flu-SP, the effect of scattering was
low in comparison to the high quantum yield of this fluorophore.
Collisional Quenching Experiments
Fluorescence collisional quenching experiments were performed at
20 °C by adding increasing amounts of quencher stock solution in
water to suspended NK1/CHO cell membranes. Typically, a complete quenching experiment was performed in 6-8 min. Under these conditions, the contribution of fluorescence signal from dissociated ligand was
negligible. The quencher stock solutions were either 0.153 M KI containing 1 mM
Na2S2O3 to prevent
I3 formation or 0.1 M
Tempo or 0.1 M hydroxy-Tempo. Changes in fluorescence due
to the addition of quencher were corrected by subtracting the
fluorescence measured in parallel samples in which quencher was added
to membranes in PBS or membranes saturated with an excess (1000-fold)
of non-fluorescent CP99,994 (36). Moreover, in quenching experiments
with the nitroxide radical compounds Tempo and hydroxy-Tempo, the
fluorescence intensities were corrected for the absorption increments
caused by added quenchers at the excitation and emission wavelengths
(inner filter effect) as described (25). The quenching of the
fluorescence emission at the wavelength of maximum emission was
calculated with the Stern-Volmer equation (26):
Fo/F = 1 + KSV [C], where
Fo/F is the ratio of fluorescence intensities in the absence and presence of quencher Q. The Stern-Volmer quenching constant KSV was determined from the
slope of Fo/F as a function of the
quencher concentration [Q].
Steady-state anisotropy measurements were recorded using a Jasco FP-777 spectrofluorometer equipped with a model ADP-301 fluorescence polarization accessory. The polarizer and analyzer were placed in the thermostatic sample chamber. The emission intensity was measured by setting the excitation-side polarizer in the vertical position (V) and the emission-side polarizer either in the horizontal (H) or vertical position. The emission intensities, in respective V and H positions, were corrected by subtracting the corresponding background signals from the cells or membrane suspensions and converted to anisotropy (A), see Equation 1:
![]() |
(Eq. 1) |
![]() |
(Eq. 2) |
A series of
undecapeptides derived from substance P and labeled with either dansyl
(Dns) or fluoresceinyl (Flu) fluorophores at different positions in the
sequence were prepared (Fig. 2). The
control of selectivity for the N-terminal amino group or the -amino
group in the lysyl side chain during chemical modification of peptides
can be controlled by the pH of the reaction and the nature of the
electrophilic reagent. Thus, substance P was derivatized with an
equimolar amount of dansyl chloride at pH 9.0 to afford the
mono-labeled compound Dns-Lys3-SP. Reaction of SP with an
equimolar quantity of fluorescein isothiocyanate gave the N-terminally
labeled N
-Flu-SP. The respective
positions of the dansyl and fluoresceinyl groups was confirmed by Edman
sequencing and amino acid analysis.
The fluorescent peptide Dns-Dap8-SP, in which
Phe8 of SP is replaced by -dansyldiaminopropionic acid,
was prepared by solid phase peptide synthesis. After the third coupling
cycle with Dap, the
-amino group of Dap was selectively deprotected
and derivatized with dansyl chloride. Coupling of the remaining seven
amino acids was then resumed. The S-dansyl homocysteine
derivative Dns-Hcy11-SP in which Met11 of SP is
replaced by S-(2-dansylamino)ethyl homocysteine was prepared
as reported before (19).
The quinuclidine derivative Dns-CP96,345, a dansyl-labeled analogue of
the well known NK1 antagonist CP96,345 (3), was prepared as shown in
Fig. 3.
Fluorescence Properties of Ligands
The excitation and emission maxima of dansyl- and fluorescein-labeled ligands are shown in Table I. The quantum yield of dansyl fluorescence was sensitive to the hydrophobicity of the medium with a higher quantum yield in low polarity solvents (Fig. 4). For Dns-Hcy11-SP, the 7-fold increase in quantum yield by going from aqueous buffer to 50% (v/v) dioxane in PBS was accompanied by a blue-shift of the emission maxima from 550 to 536 nm. Similar effects were observed for the other dansyl-labeled compounds.
|
Pharmacological Characterization of Fluorescent Peptide Ligands
The fluorescent ligands shown in Figs. 2 and 3 were
assayed for NK1 binding affinity by competitive binding analysis with 125I-BH-SP using COS-7 cells (Fig. 5). The results are
summarized in Tables II and III. Labeling of SP with dansyl at position
3 (Dns-Lys3-SP) maintained a high affinity for NK1, in the
nanomolar range and comparable to the parent SP. Modification of SP
with fluorescein at the N-terminal position
(N-Flu-SP) or at position 11 as in
Dns-Hcy11-SP resulted in a 70-100-fold decrease in
affinity compared with SP. More dramatically, Dns-Dap8-SP
showed a 18,000-fold decrease in affinity compared with SP that
prevented its further use as a fluorescent probe of the binding pocket.
Compounds N
-Flu-SP,
Dns-Lys3-SP, and Dns-Hcy11-SP were still
agonists at NK1 as assayed by their ability to 1) mobilize
intracellular Ca2+ in CHO cells stably transfected with NK1
(Table II), and 2) evoke Ca2+-dependent chloride currents in
Xenopus oocytes expressing NK1 (data not shown).
|
|
Dns-CP96,345 had a high, nanomolar affinity for NK1 with only an 8-fold decrease compared with the parent compound CP96,345 (Fig. 5). To compare the mode of binding of Dns-CP96,345 to that of CP96,345, binding constants were determined in a series of NK1 point mutants (Table III). These mutations included sites that were known to be important for CP96,345 binding. For most of the mutants tested, the change in affinity compared with wild-type receptor (fmut) was similar for CP96,345 and Dns-CP96,345. However, for mutant F264A near the top of TM-VI, the binding affinity for Dns-CP96,345 increased 2-fold. In contrast, the affinity of the parent CP96,345 decreased 17-fold in this mutant (Table III). A similar but less pronounced effect was observed for mutants F267A and F268A. There was only one mutant, F264Y, for which Dns-CP96,345 lost significantly more affinity than did CP96,345. Taken together, these data indicate that CP96,345 and its fluorescent analogue Dns-CP96,345 appear to have the same binding determinants on the NK1 receptor. The data also suggest that the dansyl group of Dns-CP96,345 could be located in the vicinity of residue Phe264 and, possibly, residues Phe267 and Phe268.
Interactions of Fluorescent Ligands with the NK1 ReceptorFluorescence emission of ligands bound to NK1 was
measured in a suspension of membranes from stably transfected CHO cells with an estimated total concentration of binding sites of 3-5 nM. Fig. 6 shows for each
ligand the fluorescence spectra of total observed fluorescence, the
nonspecific fluorescence due to binding of ligands into the membrane
and determined in the presence of a 1000-fold excess of unlabeled
ligand, and the fluorescence of an equivalent amount of compound (5 nM) in solution without membranes. The nonspecific binding
of fluorescent ligands in membranes was about 5% for
N-Flu-SP, 25% for
Dns-Hcy11-SP, and 45% for Dns-CP96,345. The comparison of
the specific fluorescence intensity emission for the dansyl-labeled
ligands bound to NK1 and in solution, after correction for the
fluorescence of membranes labeled in the presence of an excess of
nonfluorescent ligand, is a direct indication of the relative polarity
of the environment around the bound fluorescent group. For Dns-CP96,345 and Dns-Hcy11-SP, there was a 22- and 10-fold increase,
respectively, in fluorescence intensity accompanied by a blue-shift of
the emission maximum from 540 to 515 ± 5 nm and 525 ± 5 nm,
respectively, upon binding to NK1. This indicates an increase in
hydrophobicity around the fluorescent group. In contrast, for
Dns-Lys3-SP the lack of detection of fluorescence signal
indicated a different, more hydrophilic environment.
Flu-N
-SP which is brighter and less
sensitive to the polarity than dansyl was used to probe the N terminus
of SP.
Probing the Accessibility of Bound Ligands by Collisional Quenching
The nitroxide radical compounds hydroxy-Tempo
(water-soluble) and Tempo (lipid-soluble) were used to probe the
accessibility of bound dansyl-labeled ligands. In aqueous solution,
intermolecular quenching of singlet excited state radicals occurs via
an electron exchange (transfer) mechanism (27). The fluorescence of
ligands shown in Figs. 2 and 3 free in solution was efficiently
quenched by addition of increasing amounts of Tempo or hydroxy-Tempo to the solution. The very weak fluorescence emission of NK1-bound Dns-Lys3-SP precluded quenching experiments with this
ligand. Iodide anion and neutral hydroxy-Tempo were used as fluorescein
quenchers in experiments with
Flu-N-SP. Linear Stern-Volmer plots
were obtained indicating collisional quenching. The Stern-Volmer
constants (Table IV), which are the slopes of Fo/F as a function of the
quencher concentration, provide a relative measure of the degree of
accessibility of the fluorescent group to the quencher. The Tempo
compounds efficiently quenched the dansyl fluorescence of Dns-CP96,345
and Dns-Hcy11-SP in solution with
KSV values of 27-28 M
1 comparable to the reported value of 25 M
1 for quenching of dansyl-choline in 1-butanol (28). The
KSV value for hydroxy-Tempo quenching of
fluorescein in the Flu-N
-SP peptide in
PBS was 21 M
1, a value comparable to the
previously reported value of 20 M
1 for
fluorescein-labeled compounds (29).
|
The specific fluorescence of Dns-CP96,345 bound to NK1 was not quenched
by either water-soluble hydroxy-Tempo or lipid-soluble Tempo (Fig.
7). However, nonspecifically bound
Dns-CP96,345 was quenched by collision with Tempo in the membrane
(Table IV). From this we conclude that the antagonist Dns-CP96,345
bound to NK1 is buried in a hydrophobic pocket that is inaccessible to
the solvent and not in direct contact with the membrane lipids, most likely in the middle of the bundle formed by the seven
membrane-spanning domains of the NK1 receptor. Fig.
8 shows the Stern-Volmer plots obtained
for the agonists Dns-Hcy11-SP and
Flu-N-SP. The Stern-Volmer constants
for each Flu-N
-SP and
Dns-Hcy11-SP were comparable to those of ligands free in
solution, indicating that both the fluorescein group at the N-terminal
position and the dansyl group at the C-terminal position, respectively,
were fully accessible and exposed to the solvent.
Probing the Molecular Mobility of Fluorescent Ligands Bound to NK1
Fluorescence anisotropy for Dns-CP96,345,
Dns-Hcy11-SP, and Flu-N-SP
bound to NK1 or free in solution was measured in the 2-37 °C temperature range and is shown in the Perrin plots on Fig.
9. Comparable plots were obtained at
constant temperature (20 °C) by varying the viscosity (data not
shown). Limiting anisotropy, defined as the anisotropy in the absence
of all rotational freedom, was calculated from the intercepts of Perrin
plots on the y axis. Limiting values of anisotropy
(A0), ranging from 0.35 for Dns-CP96,345 and for
Flu-N
-SP to 0.24 for
Dns-Hcy11-SP, were calculated for the ligands bound to NK1
(Table V). A0
values for each ligand free in solution or bound to NK1 were similar.
The binding of ligands to NK1 receptors considerably increased the
dansyl or fluoresceinyl anisotropies compared with those of the ligands
in solution (Table V and Fig. 9). For Dns-CP96,345, the slope of the
Perrin plot for the total bound was lower than the slope for the
nonspecifically bound, indicating indirectly that the motional freedom
of the dansyl group in the ligand bound specifically to NK1 was more
restricted than that of the ligand bound nonspecifically to
membranes.
|
In this paper, we have utilized a fluorescent group in ligands to compare the binding sites of SP and the non-peptide antagonist CP96,345 in the NK1 receptor. We and others (16, 30, 37, 38) had shown that fluorescent ligands may serve as probes to investigate ligand-receptor interactions and to obtain insight into the microenvironment of binding sites. Introducing fluorescent group in ligands while maintaining high binding affinity is a prerequisite to this approach. This is generally easier to achieve for medium to large size peptides than for small molecular weight non-peptide ligands. SP was labeled at four different sites with the fluorophore dansyl to probe all regions of the peptide, both the N and C terminus and two internal positions at residue numbers 3 and 8. Modification at position 8 with the bulky dansyl group decreased drastically NK1 affinity and was not used further. It is worth noting, however, that other SP derivatives in which Phe at position 8 was replaced by larger aromatic side chains are still binding to the NK1 receptor with high affinity (14, 15). Dansyl-labeled SP analogues at positions 1, 3, and 11 conserved reasonable afffinity for SP. Dansyl was introduced into CP96,345 at a position known to tolerate different functional groups; see, for example, the use of an azide function at this position as a potent photoaffinity labeled version of CP96,345 (31). Functionally, the fluorescently labeled SP analogues maintained agonist activity, and Dns-CP96,345 was still capable of antagonizing SP at the NK1 receptor.
The putative binding site for CP96,345 has been mapped through an
extensive mutational analysis of the transmembrane regions and
extracellular loops of the NK1 receptor (for a review, see Ref. 32).
The results from several groups suggest that CP96,345 fits into a
cavity delineated by TM-III, -IV, -V, -VI, and -VII and make specific
contacts by polar interactions with receptor residues as shown in Fig.
1. The mutational binding analysis of the present study shows that
Dns-CP96,345, despite the presence of the bulky dansyl group in the
molecule, has a binding profile similar to that of its parent compound
CP96,345, as determined by the ability to displace
125I-BH-SP binding to NK1. The only significant differences
were found at positions 264, 267, and 268 which all have Phe in
wild-type NK1. These positions are located on the face of TM-VI which
is presumed to be turned inward and toward TM-VII (Fig.
10). Single mutations to alanine at
each of these three sites affected CP96,345 more than Dns-CP96,345.
Conceivably, removal of bulky phenyl side chain disturbs the shape
complementarity between the non-peptide ligand and the binding cavity.
A possible explanation for why this effect was less pronounced for
Dns-CP96,345 could be that the dansyl group could occupy the empty
space created by the Phe to Ala mutation. In contrast, adding bulk in
the cavity by introduction of an hydroxyl group in the F264Y mutation
decreases Dns-CP96,345 affinity more than that of CP96,345, presumably
by creating an unfavorable interaction between the tyrosyl chain and
the dansyl group. Overall we conclude that Dns-CP96,345 binds to NK1 in
the same mode as CP96,345 thus validating its use as a reporter for the
binding site.
Fluorescent reporter groups on ligands may provide important information on the binding pocket in the receptor. First, the polarity of the binding pocket can be estimated by using environment-sensitive fluorophores. Dansyl is very sensitive to the polarity of the medium with high fluorescence in low polarity environment. Second, we can access information on the mobility of the ligand when bound to the receptor by measuring the fluorescence anisotropy. Third, we can obtain information on the accessibility of the bound ligand by using the technique of collisional quenching.
The first spectrofluorometric evidence for the existence of different binding pockets for SP and CP96,345 was given by the differences in quantum yield of fluorescence for ligands bound to NK1 (Fig. 6). Dns-CP96345 must occupy a very hydrophobic pocket in the NK1 receptor. In sharp contrast, all labeled sites on SP were found to be located in a more hydrophilic environment. The N-terminal moiety of SP (positions 1 and 3) was in a more hydrophilic environment than the C-terminal moiety. The rotational mobility of the fluorescence-labeled ligands bound to NK1 was assessed by fluorescence anisotropy measurements. Anisotropy of fluorescent ligands increased upon binding to NK1 on CHO cells. The mostly temperature-insensitive anisotropy values of the different ligands bound to NK1 demonstrate that ligand motions are restricted at the receptor binding site(s) in the nanosecond time range. This immobilization was more important for Dns-CP96,345 than for the SP analogues, thus suggesting a more densely packed environment surrounding the receptor binding site for Dns-CP96,345.
Collisional quenching of fluorescence can be used to probe solvent accessibility of receptor-bound ligand (16). Our quenching data clearly show important differences for receptor recognition between Dns-CP96,345 and the SP analogues. Dns-CP96,345 binds in a buried pocket that is shielded from the solvent and is not in contact with the membrane lipids surrounding the receptor. Therefore we conclude that Dns-CP96,345 docks inside the pore formed by the transmembrane helices, below the extracellular water-membrane interface which is in agreement with the mutational data. In contrast, both N- and C-terminal positions of SP were fully accessible to the solvent, an observation consistent with the binding of SP in the extracellular part of the NK1 receptor. This supports the identification of contact sites between photoactivable SP analogues and the extracellular domains of NK1 (14, 15, 35). However, one important implication of our findings is that it is not necessary for SP to reach deep into the bundle of transmembrane segments to activate NK1. Interactions with the extracellular loops are sufficient to stabilize the active conformation of the receptor. In support of this hypothesis, it has been recently shown that the bradykinin B2 receptor can be activated by antibodies raised against extracellular loop peptides (33). Also, mutations in the extracellular loops of the thrombin receptor (34) can produce active receptor conformations.
In summary, our results demonstrate the existence of distinct, although possibly overlapping, binding sites for SP and the non-peptide antagonist CP96,345 in the NK1 receptor. CP96,345 binds into a hydrophobic pocket buried into the receptor and shielded from the solvent. In contrast, SP binds in the hydrophilic environment of the extracellular parts of NK1, and apparently all parts of the undecapeptide SP remain accessible to the solvent. Our data imply that receptor activation by SP does not require the peptide to make contacts deep in the transmembrane regions although we cannot totally exclude that the central portion of SP could be partly buried in the transmembrane domains. In the future, fluorescence energy transfer experiments may help define more accurately the position of the bound ligands (17). The work presented here demonstrates the complementarity of fluorescence spectroscopy and site-directed mutagenesis in investigating ligand-receptor recognition in the absence of high resolution structural data.
We thank Karin Nemeth for help with the cell culture, Charles Bradshaw for peptide synthesis, and Dr. Jonathan Knowles for enthusiastic support during these studies.