| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,,
,,,
Institut Gilbert-Laustriat, UMR 7175 CNRS/Universite-Strasbourg I, Strasbourg, France;
* Departement Recepteurs et Protéines Membranaires; ESBS, Illkirch, France;
Departement de Pharmacologie et Physicochimie des Interactions Cellulaires et Moleculaires and
Departement de Pharmacochimie de la Communication Cellulaire, Faculte de Pharmacie, Illkirch, France
1The Mt. Sinai School of Medicine, Box 1065, Department of Neuroscience, 1425 Madison Ave., New York, NY 10029, USA. E-mail: emeline.maillet{at}mssm.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: allosterism • signal transduction • GPCR • neurokinin • drug discovery
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Rather, such compounds interact elsewhere on the receptor to modulate its activity: they can be positive, negative, or neutral allosteric ligands (2
3
4)
. This behavior results in a wide range of pharmacological properties arising from a combination of their intrinsic activity and their cooperative activity on orthosteric ligand binding and efficacy (for a review, see refs. 1
, 2
, 5
).
Until recently, the most widely accepted model for GPCR activation was the extended ternary complex model (often referred to as the two-state model) (refs. 6
7
8
), according to which 1) receptor molecules interconvert between two conformations endowed with distinct structural, pharmacological, and functional properties (active and resting states), and 2) receptors can spontaneously adopt an active conformation and couple to G-proteins (references listed in ref. 6
), thereby accounting for basal response levels.
Although this model aims at describing receptor properties with a minimal number of conformations, it has become increasingly clear that the two-state model no longer adequately reflects the complexity of GPCR signaling. Hence, agonist properties (from partial to full agonist) depend on both the number and the efficiency of the active receptor conformations they stabilize (9
10
11
12)
. Moreover, studies of multiple G-protein-coupled pathways for several GPCRs (13)
suggest that GPCRs can adopt multiple active conformations, leading to different biological responses (for a review, see ref. 9
). Several mechanisims for selection among possible signaling pathways have been described; examples include phosphorylation of a GPCR (14)
, interaction of the receptor with a partner protein such as RAMP or calcyon (10
, 15
, 16)
, spatial segregation induced by microdomains (17)
, or interaction with a given agonist (ref. 13
). In the case of GPCRs capable of eliciting multiple responses, we expect that allosteric modulators will be useful tools for investigating both receptor coupling to each pathway as wellas the balancing of pathways under physiological conditions (5)
.
The tachykinin receptor subtypes, neurokinin 1 and 2 (NK1 and NK2), serve as excellent examples of GPCRs that couple to more than one G-protein in heterologous expression systems. Both receptors have been shown to couple to two different signaling pathways: a G
q pathway, which activates phospholipase Cß (PLC), thereby initiating inositol phosphate formation and calcium responses; and a Gs pathway, which activates adenylate cyclase (AC), inducing cAMP formation (18
, 19)
. Both effector systems are activated by physiological nanomolar concentrations of tachykinin peptides (18
, 19)
. For the NK1 receptor subtype, two high-affinity ligand binding states have been described in which the highest of the two corresponds to Gs coupling and the lowest corresponds to Gq coupling (20)
. For the NK2 receptor (NK2R), we have shown that fluorescent neurokinin A (NKA) rapidly stabilizes a PLC-coupled active state, followed by slow stabilization of a higher affinity AC-coupled active state (21)
. These results suggest that different conformations of NK2Rs could correspond to distinct activation states that couple to different G-proteins. The cAMP response can be eliminated by point mutations in the extracellular amino-terminal domain of NK2R (22)
or by point mutations in the TM4 domain of NK1R (23
, 24)
. Moreover, a modified and truncated fluorescent NKA, [Cys7NKA4–10-TR], apparently stabilizes only one conformation of NK2R, eliciting the calcium response only, with no increase in cAMP (21)
. This effect is referred to as "agonist-specific trafficking" or "functional selectivity" (25
26
27)
, and has been described for a variety of peptidergic GPCRs and nonpeptidergic GPCRs (including, for example, bombesin receptors, µ- and 
-opioid receptors, and calcitonin receptors) (13
, 28)
.
In the present work we characterize the effects of LPI805, the first allosteric modulator of tachykinin receptor activity. LPI805 was chemically optimized from a hit compound identified in a drug screening assay using NK2Rs, and the ligand NKA (29)
. Using a combination of equilibrium and kinetic binding experiments, we demonstrate that LPI805 modulates the relative proportion of the two active conformations of the receptor. These results are corroborated by physiological response measurements in which LPI805 causes a slight positive modulation of the NKA-induced calcium response of NK2Rs and a concomitant strong negative modulation of the NKA-induced cAMP response.
Our results further support the existence of multiple active and interconvertible states in GPCRs. Moreover, for the first time to our knowledge, we demonstrate that modulation of an orthosteric endogenous agonist of a GPCR by an allosteric extracellular drug can affect not only quantitative aspects of the responses generated (e.g., potentiation or inhibition of all possible transduction events of the activated receptor), but also their qualitative aspects (e.g., potentiation/preservation of one particular pathway and inhibition of another pathway). This proof of feasibility opens the way for the design of a new class of potentially powerful drugs that could fine-tune GPCR signaling pathways in vivo.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
7.89–7.82 (4H, m); 7.55–7.28 (8H, m); 4.02 (2H, s); 3.89 (2H, s). A mixture of the above secondary amine (1.24 g, 5 mmol), chloroacetonitrile (380 mg, 5.05 mmol), and triethylamine (510 mg, 5.1 mmol) in dimethylformamide (0.3M) was heated at 60°C overnight. The solution was extracted with ethyl acetate (10V) and washed with NaHCO3 sat (x1), H2O (x1), and brine (x1). The solvent was dried and evaporated. The resulting oil was purified by flash chromatography on silica gel and recrystallized in 2-propan-2-ol; yield: 84%. 1H-NMR: (300 MHz, CDCl3): 7.90–7.80 (4H, m); 7.60–7.20 (8H, m); 3.92 (2H, s); 3.81 (2H, s); 3.40 (2H, s). 13CNMR: (50 MHz, CD3OD): 136, 135, 132, 131, 130, 114, 60, 59, 41. CHN calculated: C = 83.88%, H = 6.34%, n = 9.77%. CHN found C = 83.758%, H = 6.28%, n = 9.77%.
Synthesis of a fluorescent NKA analog
NKA (HKTDSFVGLM-NH2) was derivatized with iodoacetyl-Bodipy to give NKAbo, as described previously (21
, 30)
.
Chemicals
Synthetic peptides were obtained from Bachem (Bubendorf, Switzerland). Protease inhibitors were obtained from Sigma (St. Louis, MO, USA) and Calbiochem (San Diego, CA, USA). Fluorescent labels and ion chelators were from Molecular Probes (Carlsbad, CA, USA). [3H]SR48968 was purchased from Amersham (Piscataway, NJ, USA). SR48968 was kindly given by Dr. Emonds-Alt (31)
. Emerald Enhancer and CSPD were from Applied Biosystems (Foster City, CA, USA). JetPEI was from Polyplus (San Marcos, CA, USA).
NK2R cDNA
Rat NK2R cDNA in 5' frame fusion with EGFP (enhanced green fluorescent protein) cDNA was cloned in the pCEP4 expression vector (Invitrogen, Carlsbad, CA, USA) as described (30)
.
Cell culture and buffer
HEK293 cells were stably transfected with the EGFP-rNK2 vector by calcium phosphate precipitation, followed by selection with 500 µg/ml Hygromycin B. The number of NK2Rs expressed per cell was found to be 3 x 105, as calculated from BMAX values (1700 dpm) of [3H]SR48968, as described (30)
. Unless stated otherwise, cells were assayed in HEPES-BSA buffer (137.5 mM NaCl, 1.25 mM MgCl2, 1.25 mM CaCl2, 6 mM KCl, 5.6 mM glucose, 10 mM HEPES, 0.4 mM NaH2PO4, 1% bovine serum albumin (w/v), pH 7.4) supplemented with protease inhibitors (22)
.
Fluorescence binding
Interaction of fluorescent peptides with chimeric EGFP-rNK2 was monitored on a Fluorolog 2 spectrofluorometer (SPEx) with excitation set at 470 nm (bandwidth=5 nm). A decrease in the EGFP emission at 510 nm (bandwidth=10 nm) indicated fluorescence resonance energy transfer (FRET) toward the acceptor group Bodipy as described (30)
. Cell (106 cells/ml) or membrane suspensions (100 µg protein/ml) in HEPES-BSA buffer were placed in a 1 ml quartz cuvette under constant stirring. Addition of fluorescent agonist to the suspension resulted in a decrease of EGFP fluorescence at 510 nm. The intensity of the fluorescence variation was proportional to the occupancy of receptor sites and is expressed as raw FRET in % (
F/F). Typically, saturation of the receptor sites resulted in a 40% decrease in initial fluorescence of the cell suspension. Measurements of dissociation were initiated by addition of an excess of unlabeled NKA (20 µM) and recorded as the recovery of EGFP fluorescence intensity as a function of time. Binding and kinetics studies were performed at 
20°C to avoid internalization of receptors during experiments.
Calcium imaging
The cells were trypsinized and reseeded onto polylysine-coated 96-well assay plates (Corning, NY, USA) at a density of 40,000 cells/plate. After 24 h, the cells were washed once, then loaded with the calcium-sensing dye Fluo-4 (75 µl at 3 µM) in HEPES-BSA buffer for 2 h. The cells were washed twice and assayed using a FlexStation II (Molecular Devices Corp., Palo Alto, CA, USA). Fluorescence changes (excitation at 488 nm, emission at 525 nm, and cutoff at 515 nm) were monitored after the addition of HEPES-BSA buffer supplemented with ligands. For each response trace, the data were acquired at 2 s intervals; samples were added at 30 s and scanning was continued for an additional 200 s. The intrinsic fluorescence of cells expressing EGFP or nontransfected cells did not contribute significantly to the strong fluorescence signal arising from Fluo-4, and was ignored. Calcium mobilization was quantified as the change of peak fluorescence (F) over the baseline level (F).
Cyclic AMP accumulation
Receptor coupling to adenylate cyclase activation was determined by radio-immunoassay for adherent cells plated in 24-well plates, as described in ref. 22
. Agonist was added at the desired concentrations at 20°C and incubated for 30 min. The reaction was stopped by addition of one volume ice-cold 0.2M HCl. Preincubation with LPI805 or other drugs was carried out for 20 min.
cAMP dependent reporter gene assay
The MercuryTM SEAP System was used to quantify PKA activation in EGFP-NK2R-expressing HEK293 cells (32)
. The SEAP assay was carried out as described previously (33)
, except that transfection was performed with JetPEI reagent on a single cell batch, from which 50,000 cells/well were seeded in 96-well collagen-coated plates the day after transfection. A range of concentrations of agonist diluted in culture medium was applied to wells of adherent cells at 37°C for 5 h. Preincubation with LPI805 or other drugs was performed for 20 min before agonist addition.
Data analysis
GraphPad PrismTM software was used to calculate EC50, KD, KI, BMAX, and EMAX and to generate bar graphs and saturation binding curves. KaleidaGraphTM software was used for adjustment of NKAbo binding traces. Biexponential dissociation was defined by the equation: [F] = f(t) = Feq + A1*exp(–V1*t) + A2*exp(–V2*t), where t is time in seconds; [F] is the fluorescence signal: V1 and V2 are off rate constants koff (s–1) of rapid and slow dissociation events, respectively; and A1 and A2 are the amplitudes of rapid and slow dissociation, respectively. Biexponential association was defined by the equation: [F] = f(t) = Ft = 0 – A1*exp(–Vapp1*t) – A2*exp(–Vapp2*t). When raw FRET values were not used, data from association and dissociation fluorescence experiments were normalized between 1 (basal fluorescence prior to addition of NKAbo and/or recovered fluorescence after dissociation of NKAbo) and 0 (fluorescence after association of NKAbo and/or after saturation of the receptor). When expressed in terms of "bound" or "amplitude," NKAbo binding was normalized between 0 (no binding) and 1 (saturation of receptor occupancy).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A) belongs to a new family of chemicals that act on tachykinin receptors (29)
. Hit compounds were identified in drug screening trials by using FRET (30)
to detect altered equilibria (34)
and/or kinetics (29)
between N-terminally EGFP-tagged NK2 receptors (EGFP-NK2R) and bodipy-labeled neurokinin A (NKAbo). These compounds were then chemically improved (C. Valant et al., unpublished results) and LPI805 was selected for extensive characterization regarding its specificity, affinity, and biological activity on NK2Rs.
|
Noncompetitive antagonist behavior of LPI805 on NK2Rs
As shown in Fig. 1B
, LPI805 (50 µM) induced a small shift in the NKAbo equilibrium saturation curve on EGFP-NK2Rs, from an apparent KD of 5.7 ± 0.6 nM to 12.7 ± 2.1 nM. The maximal FRET amplitude was also slightly reduced, although the difference was not significant. In competition experiments carried out at a subsaturating concentration of NKAbo (20 nM), increasing concentrations of LPI805 reversed a portion (up to 33±2%) of the fluorescent signal from bound NKAbo (IC50=19±4 µM) (Fig. 1C
). In contrast, the well-known competitive antagonist SR48968 (30)
fully inhibited NKAbo binding (IC50=2.6±0.6 nM) (Fig. 1C
).
This result suggests that LPI805 partially reduces the NKAbo occupancy of native receptor binding sites. Thus, LPI805 might act as a noncompetitive allosteric ligand on the receptor, modifying either the ratio of high- to low-affinity receptor states or/and NKA affinity for each individual receptor state (see ref. 5
). To determine the parameters affected by an allosteric effector, its effects on dissociation and association rate constants of classical orthosteric ligands are typically analyzed (5)
.
Quantitative analysis of LPI805-effects on NKAbo dissociation kinetics
To monitor the effect of LPI805 on orthosteric ligand binding and dissociation kinetics, cells expressing EGFP-NK2Rs were treated with NKAbo, and EGFP fluorescence emission was recorded as a function of time (Fig. 2
). Addition of an excess of nonfluorescent competitive displacers like the agonist NKA (20 µM) or the nonpeptidic antagonist SR48968 (data not shown) led to complete dissociation of bound NKAbo, with similar dissociation kinetics 1) during the association event (Fig. 2)
or 2) after association equilibrium had been reached (Fig. 3
, 1200 s).
|
|
NKAbo stabilizes two apparent states of NK2Rs
We previously reported that fluorescent NKA-Texas red (NKA-TR) associates with several NK2R conformations (21
, 22
, 30)
. Like NKA-TR, association and dissociation kinetics of NKAbo follow a multiphasic process that can be described using a biexponential equation that accounts for two different conformations of NKA-bound receptors (see Materials and Methods and refs. 21
, 35
):
 |
The dissociation of bound NKAbo at equilibrium was fitted with average rate constants of 0.04 s–1 and 0.008 s–1. These values were subsequently imposed to fit the whole dissociation traces from pre-equilibrium NKAbo association experiments (Fig. 2A
). During the association process, NKAbo progressively stabilized high-affinity (slow dissociation) conformations of receptor (A2L) at the expense of lower affinity (rapid dissociation) conformations (A1L). Indeed, the number of A1L conformations exhibited a bell-shaped curve at early association times, accounting for a strong interconversion of these fast-dissociating conformations toward the more stable slow-dissociating conformations A2L (Fig. 2B
). After association equilibrium was reached, the slow dissociation amplitude typically represented 60–70% of total dissociation amplitude (Fig. 2
and Fig. 3A, B
, black lines). Macroscopically, association of NKAbo with the receptor presented the appearance of a sequential binding process (Scheme 1
) in which the A2L conformation was more likely to arise from interconversion of the previously bound A1L conformation (conversion CA1L
A2L) rather than from direct binding to the free receptor (CR+LA2L). As shown in Scheme 1
, the kon constant k3 should be lower compared with k1 and k2 to permit prior binding of NKAbo on the fast dissociating (k–1>k–3) A1L state rather than the slow dissociating A2L state.
|
Acceleration of NKAbo dissociation rate by LPI805
When increasing concentrations of LPI805 were added to the equilibrated NKAbo/receptor complex (after 1200 s), weak dissociation of NKAbo was detected and new equilibrium conditions were reached. When NKA (20 µM) was further added, complete dissociation of NKAbo was observed, with a higher rate than in the absence of LPI805 (Fig. 3A
). The time course of NKAbo dissociation was accelerated, with t1/2 dissociation times ranging from 68 s (0 µM LPI805) to 57 s (1 µM), 46 s (5 µM), and 25 s (50 µM LPI805) (Fig. 3A
). These effects of LPI805 could result either from a change in the proportion of receptors in the A1L and A2L conformations or from a change in the affinity of NKAbo for these two conformations.
Based on the measurement of NKAbo association in the presence of LPI805, where apparent dissociation constants were unaltered (see Results below), we hypothesize that the parameter primarily altered by LPI805 is the ratio of the amplitudes of each conformation rather than intrinsic koff rates (Scheme 2
). All dissociation traces were well-fitted by imposing the same rate constants V1 = 0.04 s–1 and V2 = 0.008 s–1, identical to those provided by fits of control dissociations in the absence of LPI805 (Fig. 3A
), with the determined parameters being the amplitudes A1 and A2 of each exponential (A1+A2=total dissociation amplitude).
|
Analysis of dissociation in these conditions indicated that LPI805 increased the proportion of rapid dissociation amplitude (A1%) in a dose-dependent (EC50=7.8±1.8 µM) and saturable (from 33±3 to 83±4%) manner (Fig. 3A
). In other words, LPI805 sped up the global dissociation rate by increasing the amplitude of the rapid dissociation process (EC50-A1=2.5±0.9 µM) and by concomitantly decreasing the amplitude of the slow dissociation process (IC50-A2=8.2±1.4 µM) (Fig. 3C
). Similar effects of LPI805 were found when different concentrations of NKAbo (5, 100, 200 nM) were used instead of 20 nM (data not shown), as expected for a noncompetitive/allosteric modulator of NK2Rs.
In contrast, the competitive antagonist SR48968 (from 0 to 5 nM) progressively reduced NKAbo dissociation amplitude due to competition between SR48968 and NKAbo, without altering kinetics (Fig. 3B
). All traces had similar dissociation t1/2 values (45 s at 0 nM, 44 s at 0.5 nM, 49 s at 2 nM, and 53 s at 5 nM SR48968) and the ratio between the relative amplitudes A1 and A2 remained constant (Fig. 3B, D
).
The effects of LPI805 on NKAbo binding were reversible (Supplemental Fig. 1
). Indeed, when cells preincubated with LPI805 were washed, the binding and dissociation kinetics of NKAbo were restored close to control values. Moreover, LPI805 had no effect on the dissociation of the fluorescent ligand pirenzepine-bodipy from EGFP-M1 muscarinic receptors assessed by FRET (36)
(data not shown). LPI805 had the same effect on the dissociation kinetics of both Texas red-labeled NKA and bodipy-labeled NKA (data not shown). Acceleration of the NKAbo dissociation rate and the partial antagonism of LPI805 were also observed on membrane preparations of the EGFP-NK2 receptor (data not shown). Thus, LPI805 appears to behave as an allosteric modulator of the NK2 receptor.
LPI805 alters the equilibrium between the two active conformations of NK2 receptors
The fluorescent NKA ligands, NKA-TR (21
, 22)
and NKAbo (this work), are agonists that stabilize NK2Rs in at least two conformations: A1L (fast dissociation) and A2L (slow dissociation). These two conformations correspond to two active states that function to trigger either the PLC pathway (A1L) or the AC pathway (A2L) (Scheme 2
). By itself, NKAbo predominantly stabilizes the A2L conformation (70%); however, in the presence of LPI805, NKAbo predominantly stabilizes the A1L conformation (up to 90%) (Fig. 2B, C
, Fig. 3C
). Therefore, we predicted that LPI805 modulation of NK2Rs should function to increase A1L-associated responses and to decrease A2L-associated responses (Scheme 2
). To test these predicted functional allosteric effects, we measured changes in the calcium and cAMP responses associated with the NKA-activated NK2 receptor in the presence or absence of LPI805.
LPI805 slightly potentiates NK2R-mediated intracellular calcium elevation
When the EGFP-NK2 (as well as NK2) receptors are stimulated with the full agonist NKA (as well as NKAbo), a Gq/11-mediated intracellular calcium elevation is induced in the cells (see ref. 30
). We followed this response by spectrofluorimetry using the calcium-sensitive intracellular dye Fluo-4. NKA-induced calcium signals were quantified using the ratio of peak fluorescence intensity over baseline (F/F), as described (37)
. LPI805 did not affect the basal level of intracellular calcium in control or EGFP-NK2R-expressing HEK cells. A subsaturating concentration of LPI805 (20 µM) did not change the EC50 of the calcium dose-response curve of NKA on EGFP-NK2R-expressing cells (EC50=6±0.8 nM; EC50(+LPI805)=6.1±1.3 nM) (Fig. 4
A). However, some subtle effects could be noted. LPI805 decreased the slope of the NKA dose-response calcium curve from 1.17 ± 0.06 to 0.83 ± 0.1 (Fig. 4A
), thus accounting for the moderate potentiation of calcium responses at low concentrations of NKA (Fig. 4A
). In addition, LPI805 (20 µM) increased the maximal NKA-induced calcium response by 20% (Fig. 4B
). LPI805 did not affect Gq-dependent muscarinic M1 receptor calcium responses in the same cells (data not shown), ruling out indirect effects of LPI805 on the calcium signaling pathway. Hence, LPI805 behaves as a weak positive modulator of NKA-induced calcium signaling of NK2Rs.
|
LPI805 inhibits NK2R-mediated cAMP signaling
Several groups have described intracellular accumulation of cAMP mediated by tachykinin receptors in recombinant expression systems (18
, 19
, 21)
. Here, we followed cAMP accumulation by determining the enzymatic activity of secreted alkaline phosphatase (SEAP) expressed under the control of a cAMP-responsive element (CRE-SEAP assay; ref. 32
). We found that NKA stimulated CRE-dependent alkaline phosphatase (AP) expression, with an EC50 value of 7 ± 1 nM (Fig. 4C
). All responses were fully abolished in the presence of the PKA inhibitor H89 (10 µM) (Fig. 4D
; see also ref. 32
). Moreover, H89 decreased the basal level of AP synthesis, accounting for the inhibition of the constitutive activity of PKA in HEK cells.
In the presence of the conventional antagonist H8565, at 200 nM (KD value
20 nM, ref. 38
) the EC50 of the NKA dose-response for cAMP was shifted rightward by 4.5-fold, with no change in the maximal amplitude, as expected for a competitive antagonist (data not shown).
In contrast, the amplitude of the cAMP response was decreased 3-fold by LPI805 at 10 µM and was abolished at 40 µM (Fig. 4C
), with no significant alteration of the EC50 for NKA (EC50(10µMLPI805)=23±15 nM). Reduction of the maximal NKA response amplitude (EMAX) by LPI805 was dose dependent, with an IC50 of 3.4 ± 1.1 µM (Fig. 4F
), which is comparable to the apparent affinity of LPI805 for dissociation rate alterations (Fig. 3C
). LPI805 added alone weakly reduced the basal cAMP reporter level in EGFP-NK2-expressing cells (Fig. 4D
) and in HEKs (Fig. 4E
). Alteration of the adenylate cyclase signaling system in HEK cells by LPI805 required the presence of NK2Rs; LPI805 had no effect on the isoproterenol-induced cAMP response of the endogenously expressed Gs-coupled ß2-adrenergic receptors in HEKs (Fig. 4E
). Similar results were obtained when cAMP concentrations were determined directly by radio-immunoassay (data not shown).
LPI805 dramatically decreased the NKA-induced cAMP responses of NK2Rs without significantly altering the EC50 of NKA for this pathway. Thus, LPI805 is a negative modulator of the NKA-induced cAMP response of NK2Rs. It most likely acts by reducing the total number of receptors capable of signaling through this pathway.
Quantitative analysis of LPI805 effect on NKAbo association kinetics
We hypothesized that LPI805 changes the equilibrium between the two stabilized conformations of NKAbo-bound NK2Rs and promotes stabilization of the A1L (fast dissociation) conformation at the expense of the A2L (slow dissociation) conformation (Scheme 2
; Fig. 3
). This action is reversible (supplemental Fig. 1) and noncompetitive (Fig. 1)
. To characterize the effect of LPI805 on the interconversion process between the A1L and A2L conformations of the NKAbo-bound receptor, we investigated the kinetics of NKAbo association in the presence or absence of LPI805 (Fig. 5
).
|
LPI805 slowed the association kinetics of NKAbo at 20°C (not shown). However, the association kinetics at this temperature were too fast to properly analyze the effect of LPI805 without a stopped-flow apparatus (our stopped-flow apparatus did not permit stirring of the medium; thus, the association event could not be properly recorded during the whole association process). Therefore, we recorded the effect of LPI805 on association kinetics at 10°C, which permitted an acceptable time-resolved recording of both the early fast association and the slow stabilization of the NKAbo receptor interaction (Fig. 5A
). Associations traces were fitted using a biexponential equation as described (21
, 35)
.
The fast apparent association rate V1 of the biexponential fit accounts for the formation of the NKAbo-bound A1L state of the receptor:
 |
).
The slow apparent association rate V2 accounts for the formation of the NKAbo-bound A2L state of the receptor:
 |
); and k–2 represents the dissociation rate of the ligand from A2L as well as the backward rate constant of the equilibrium between A2L and A1L (Fig. 5C
). The saturation of the rate of the slow exponential did not occur in this experiment because the concentrations of NKAbo were not high enough for the temperature (saturation occurs at 
500 nM at 20°C).
At subsaturating concentrations, LPI805 (50 µM) did not modify any of the kinetic components of the fast association event: V1 = 0.34 ± 0.03 [NKAbo] + 12.6 ± 3.5 ms–1 and V1–LPI805 = 0.34 ± 0.03 [NKAbo] + 11.7 ± 3.8 ms–1 (Fig. 5B
). However, the fit for the slow association rate was nearly a flat line in the presence of LPI805 while the Y-intercept was unchanged: V2 = 0.018 ± 0.0017 [NKAbo] + 1.7 ± 0.2 ms–1; and V2–LPI805 =0.005 ± 0.001 [NKAbo] + 2.0 ± 0.1 ms–1 (Fig. 5C
). Thus, although we cannot provide exact values for the intrinsic rate constants of NKAbo slow binding to A2L conformations, these data demonstrate that LPI805 dramatically slows the on rate constant (k2) while preserving the backward rate constant (k–2).
Therefore, the main effect of LPI805 on NKAbo-NK2R interaction is to strongly impede the interconversion from A1L to A2L. LPI805 does not affect the generation of the A1L state (association and dissociation rate), nor does it affect the dissociation rate of A2L when this state is stabilized. We have also shown that LPI805 might dramatically decrease the amount of the A2L conformation generated (Figs. 2
3
4)
. The effect seen on the slow association rate is probably related to heterogeneity in the population of receptors, so that the "LPI805-free" receptors can interconvert from A1L to A2L while the "LPI805-bound" receptors cannot. This would account for a global slowdown of the interconversion rate of this mixture of receptors (Scheme 3
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
As expected for a molecule that might change the relative proportions of NKA-bound receptor conformations, we found that a weak potentiation of the calcium response correlates with increased occupancy of the intermediate A1L conformation while a reduction of the cAMP response correlates with reduced occupancy of the slowly stabilizing A2L conformation. The effects of LPI805 can thus be interpreted as a modulation of the equilibrium between the two conformations of the receptor, favoring the calcium-related A1L state at the expense of the cAMP-related A2L state. Moreover, we showed that LPI805 dramatically reduces the interconversion rate between A1L and A2L during the association of NKAbo but does not block the backward interconversion from A2L to A1L (Fig. 5C
). Thus, we propose that LPI805 renders the NKAbo-N2KR complex less likely to stabilize in the A2L state by blocking the transition toward the A2L conformation (Scheme 3
). LPI805 can thus be defined as a conformation-specific noncompetitive antagonist of NK2Rs. In the presence of LPI805, the NKA-bound receptors cannot reach the stabilized cAMP-A2L conformations, and NKA then behaves similarly to [Cys7NKA4–10-TR], a partial agonist that exclusively activates calcium signaling (21)
.
At saturating concentrations, LPI805 doubles the relative number of calcium-related A1L conformations that are stabilized by NKA after equilibrium has been reached (Fig. 3C
). LPI805 is a weak positive modulator of NKA-induced calcium responses. Since calcium responses occur during the first seconds of agonist/receptor interaction, we can argue that the magnitude of the calcium response is highly related to receptors in the A1L state within the first seconds of NKA binding rather than to A1L state receptors at equilibrium (1200 s), when most are presumably desensitized. As shown (Fig. 2
; ref. 21
), NKAbo binding follows an apparently sequential process: a fast binding on a A1L state (fast dissociation) followed by a slower stabilization into a higher affinity A2L state (slow dissociation). LPI805 does not accelerate the kinetics of binding to A1L, but rather impedes the interconversion step leading to A2L (Figs. 2
, 5)
, thereby promoting stabilization and accumulation of NKAbo-bound receptors in A1L states (Figs. 2
, 3)
. Thus, the slight potentiation of the calcium response by LPI805 would be related to NKAbo-bound receptors that spend "more time" in the intermediate A1L state instead of interconverting to A2L, as would occur in the absence of LPI805 (Scheme 3
). Recently, Bartl and colleagues (39)
demonstrated that cis-acyclic retinals, lacking four carbon atoms of the ring, can only partially activate rhodopsin. With this agonist ligand, receptor Meta II state is formed with almost normal kinetics. However, this state is unstable and results into a low amount of Meta II occurrence and a fast decay of activity. On the contrary, the action of LPI805 might dramatically increase the energy level for the A2L state without affecting the intrinsic kinetics accounting for the energy barriers of A1L, thereby impeding the formation of A2L states to the benefit and stabilization of A1L intermediate states. This phenomenon could also underlie the notion of "conformational desensitization" of GPCRs in which termination of the response occurs not only because of postactivation modifications of the receptor (such as phosphorylations or interactions with partner proteins), but also because the agonist by itself triggers the receptor to undergo additional conformational changes that are differentially stabilized and/or coupled to different G-proteins (6
, 9
, 13
, 39)
.
By using the tachykinin type 2 receptor as a model of multiple G-protein-coupled receptors, we demonstrate that it is possible to differentially modulate the activity of the two signaling pathways of this receptor in three ways: 1) by using different agonists as described in ref. 21
; 2) by mutations of the receptor as described in refs. 22
23
24
; 3) by allosteric modulation of the receptor as described in this work. This latest demonstration opens the way to select one from among several possible orthosteric agonist-generated signaling cascades in order to produce a desired physiological response in integrated systems at the cellular, organ, or even whole animal level. Finally, we provide proof of the feasibility of a theoretically predicted (27)
, but until now undescribed, feature of allosteric ligands: the design of conformation-specific inhibitors or enhancers of GPCRs might lead to powerful new approaches in therapeutics.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 3, 2006. Accepted for publication February 1, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Quiniou, P. Sapieha, I. Lahaie, X. Hou, S. Brault, M. Beauchamp, M. Leduc, L. Rihakova, J.-S. Joyal, S. Nadeau, et al. Development of a Novel Noncompetitive Antagonist of IL-1 Receptor J. Immunol., May 15, 2008; 180(10): 6977 - 6987. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kenakin Functional Selectivity through Protean and Biased Agonism: Who Steers the Ship? Mol. Pharmacol., December 1, 2007; 72(6): 1393 - 1401. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. Michel and A. E. Alewijnse Ligand-Directed Signaling: 50 Ways to Find a Lover Mol. Pharmacol., November 1, 2007; 72(5): 1097 - 1099. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||
