INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CONSIDERABLE EVIDENCE has been provided for regulating roles of extracellular ATP on neurotransmission and homeostasis in inner ear organs (27). The nucleotide is found in perilymphatic and endolymphatic fluids of the cochlea (21, 30); it modulates the transduction processes of sound and/or motion stimuli (10) and controls the functions of secretory epithelial cells (18). Thus, in Rana pipiens, ATP increases the spontaneous electrical activity of afferent fibers from semicircular canal (2, 3); in the guinea pig, the nucleotide alters gross cochlea and auditory nerve potentials and distortion product otoacoustic emissions (22); and in the gerbil, ATP regulates the K⁺ transport toward endolymphatic space by secretory cells of vestibular dark cell and strial marginal cell epithelia (18).

In electrophysiological experiments performed on guinea pig cochlea, P_2X receptors linked to large cation channels have been localized on the apical side of outer hair cells (14, 22, 26), and both P_2X and P_2Y receptors coupled to Ca²⁺-activated nonselective cation channels on the basolateral membrane of these cells (28). Biochemically, P_2X and P_2Y receptors have been identified by autoradiography in the organ of Corti, stria vascularis, and spiral prominence of guinea pig cochlea (19); P_2Y agonists enhance cytosolic free Ca²⁺ mobilization in nonsensory cells from cochlear lateral wall (15); and P_2Y agonist-sensitive phosphoinositidase C activities have been reported in rat cochlear lateral wall (23) and in ampulla from R. ridibunda semicircular canal (6).

The aim of the present work was to characterize the kinetic and pharmacological properties of ATP receptors in ampullary epithelium of R. ridibunda semicircular canal by making direct radioligand binding experiments on a few ampullas and by screening the abilities to inhibit radioligand binding of unlabeled structural ATP analogs more specific for the P_2X ligand-gated ion channels or P_2Y receptors linked to G protein signal pathways (12, 13, 33). The binding study was based on previous findings obtained with the [³⁵S]adenosine 5'-O-[2-thiodiphosphate] ([³⁵S]ADPS), a potent marker of the P_2Y receptors from avian erythrocyte membranes and cultured bovine aortic endothelial cells that binds with a lower affinity to the P_2X receptors and barely labels the pyrimidine nucleotide-sensitive P_2(Y) receptors (7, 12, 13, 31). Our results indicate that the population of [³⁵S]ADPS-labeled binding sites is heterogeneous and contains mainly both P_2X-like and P_2Y-like receptors whose molecular subtypes remain to be defined.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Products used. [³⁵S]ADPS (1,400 Ci/mmol) was purchased from New England Nuclear (DuPont de Nemours, Les Ulis, France).

Other chemicals were provided from the following sources: adenosine, cAMP, AMP, ADP, ADPS, ATP, adenosine 5'- (3-thiotriphosphate) (ATPS), ,-methyleneadenosine 5'-triphosphate (,-Me-ATP), 2'- and 3'-O-4-(benzoylbenzoyl)-adenosine 5'-triphosphate (Bz-ATP), diadenosine tetraphosphate (A_p4A), guanosine 5'-triphosphate (GTP), inosine-5'-triphosphate (ITP), xanthosine 5'-triphosphate (XTP), cytosine 5'-triphosphate (CTP), 2-desoxythymidine 5'-triphosphate (dTTP), uridine 5'-diphosphate (UDP), uridine 5'-triphosphate (UTP), bacitracin, BSA, DIDS, reactive blue 2 (basilen blue E-3G), and urethan from Sigma (St. Louis, MO); 2-methylthioadenosine 5'-triphosphate (2-MeS-ATP) from Research Biochemicals International (Bioblock Scientific, Illkirch, France); and suramin from Miles (Naperville, IL).

Animals. Experiments were performed using a total series of 685 adult frogs of both sexes of the species R. ridibunda purchased from the Ardenay Breeding Center (France). Frogs were kept in tanks containing tap water at 8°C. After percutaneous anesthesia with urethan, they were killed by decapitation.

Microdissection of ampullas from frog semicircular canals. Microdissection of the three semicircular canals from each inner ear was performed under stereomicroscopic observation in a chilled modified amphibian Ringer solution, medium A (in mM: 20 HEPES-NaOH, pH 7.5; 82 NaCl; 3 KCl; 1.8 CaCl₂; 1 MgCl₂; 0.33 Na₂HPO₄; 0.44 NaH₂PO₄; 5 glucose; 3 lactic acid; 10 sodium acetate; and 0.1% BSA). The ampullas were separated from the adjacent regions of the semicircular canals and opened by sagittal incision of their dorsal level. In some experiments, the dorsal ampullary regions formed of undifferentiated epithelial cells were separated from the ventral regions containing secretory dark cells, transitional cells, sensory hair cells, and undifferentiated cells (24) by cutting the higher frontal level of ampullas. Dark cell and hair cell areas with their adjacent connective tissue were microdissected from other structures in ventral regions of ampullas. Epithelial structures were kept overnight at 4°C in medium A in the same petri dish until use.

[³⁵S]ADPS binding assays. Assays were performed using a microtechnique adapted to ampullas from frog semicircular canals derived from that described earlier for vasopressin binding to rat kidney tubules (5).

Ampullas were first washed at 4°C in a large volume of medium B devoid of divalent cations (in mM 40 HEPES-NaOH, pH 7.5; 0.25 EDTA-NaOH, pH 7.5; 82 NaCl; 3 KCl; 0.33 Na₂HPO₄; 0.44 NaH₂PO₄; 5 glucose; 10 sodium acetate; and 0.1% bacitracin). Binding assays were further performed on siliconized bacteriological slides pretreated with a 2-µl droplet of a 0.1% BSA solution put on the hollow slide and evaporated to dryness to limit the area of the droplet.

Except where otherwise indicated, for each individual determination, three ampullas presumably removed from different frogs were transferred onto a 5-µl droplet of a medium C (medium B containing 20 nM [³⁵S]ADPS and various amounts of unlabeled ATP analogs or antagonists) put on the hollow slide, and samples were tightly covered with a petroleum jelly-coated slide to obtain a watertight seal. The binding reaction was carried out for 3 h at 4°C and stopped by adding 200 µl chilled medium D (medium B devoid of bacitracin). Epithelial structures were removed and washed 5 times in 200 µl of medium D at 4°C. A 5-µl droplet containing ampullas was sucked out from the last rinse, whereas an equivalent volume of this rinse was used as blank reference. Samples were treated with 1 ml of a 3% sodium deoxycholate solution and counted for 10 min by liquid scintigraphy. Generally for each experiment, six to eight different concentrations of a given ATP analog were tested in three replicates using three ampullas per determination; therefore, each dose-dependent inhibition binding curve was drawn from results obtained with 54-72 ampullas microdissected from 9-12 frogs.

Specific binding was defined as the difference between total binding measured with [³⁵]SADPS only and nonspecific binding determined in the presence of both radioligand and 0.1 mM unlabeled ADPS added together at the beginning of the reaction.

As an example, the actual counts per minute (cpm) measured in an experiment performed using for each single measurement samples containing three ampullas incubated with 20 nM [³⁵S]ADPS (203,100 cpm) were as follows (means ± SE of 3 replicates): background, 20 ± 1 cpm; blank, 83 ± 10 cpm; total binding over blank, 56,102 ± 3,714 cpm; and nonspecific binding over blank, 2,571 ± 254 cpm.

Control experiments. Control experiments showed the following. 1) Overnight storage of ampullas at 4°C did not impair total, nonspecific, and specific binding of 20 nM [³⁵]SADPS; the corresponding values (means ± SE of 3 replicates) obtained for samples assayed immediately after dissection were 13.9 ± 1.7, 3.5 ± 0.8, and 10.4 ± 1.9 fmol [³⁵S]ADPS bound/ampulla and for samples kept overnight were 14.4 ± 1.8, 2.1 ± 0.5, and 12.3 ± 1.8 fmol [³⁵]SADPS bound/ampulla, respectively. 2) The presence of 1.8 mM CaCl₂ and 1 mM MgCl₂ in the incubation medium decreased by 32% the specific binding of 20 nM [³⁵S]ADPS: binding capacities of samples assayed with divalent cations were significantly lower than those of control samples (10.3 ± 1.5 and 15.1 ± 0.4 fmol [³⁵S]ADPS bound/ampulla, respectively; means ± SE of 3 replicates, P < 0.05, Student's t-test). These data suggest that removal of Ca²⁺ and Mg²⁺ prevents the metabolic breakdown of nucleotides by frog ampullas as reported for other systems (33). 3) Raising the temperature of the binding reaction to 20°C led to a dramatic increase of nonspecific binding and a decrease of specific binding capacities (nonspecific binding was 2.8 ± 0.7 and 6.1 ± 0.9 fmol [³⁵S]ADPS bound/ampulla, and specific binding was 12.6 ± 1.7 and 5.4 ± 1.9 fmol [³⁵S]ADPS bound/ampulla for samples incubated at 4°C and at 20°C, respectively; means ± SE of 3 replicates, P < 0.05, Student's t-test); these results indicate that frog ampullas degrade the ligand used at 20°C.

Calculations. Specific binding capacities RL were computed by

(1)

where X, Y, and B values are the radioactivities measured for total binding, nonspecific binding, and blank samples, respectively; N₁ and N₂ are the numbers of ampullas used for total and nonspecific binding determinations, respectively; and SRA is the specific radioactivity of the radioligand. Binding capacities were expressed as 10¹⁵ mol [³⁵S]ADPBS bound per ampulla (fmol [³⁵S]ADPS bound/ampulla).

Results were given as means ± SE of n replicates performed using three ampullas for each single measurement.

Kinetic parameters for binding of unlabeled structural ATP analogs and of antagonist acting as competitive inhibitor were computed from results of competition experiments performed using 20 nM [³⁵S]ADPS and increasing concentrations of unlabeled analogs. The observed dose-dependent inhibitions of radioligand binding by unlabeled analogs are adequately accounted for by the following relation

(2)

in which RL is the specific binding measured, R_T is the maximal binding capacity, L and I are the concentrations, K_B and K_i are the apparent dissociation constants (i.e., nucleotide or antagonist concentrations leading to half-maximal occupancies of specific binding sites), and n and m are the Hill coefficients for binding of the radioligand and unlabeled analog, respectively. Experimental data were fitted according to the expected linear relationship

(3)

in which RL₀ and RL are binding activities observed in absence or presence of increasing concentrations of inhibitor I, respectively. The slope and x-intercept of the plot give values of m and IC₅₀ (i.e., unlabeled analog concentration leading to half-displacement of labeled binding sites), respectively. The K_i values were further calculated from the equation

(4)

using for computations K_B and n values determined from results of competition binding experiments between 20 nM [³⁵S]ADPS and increasing amounts of unlabeled ADPS. With the assumption that no difference occurs between the apparent dissociation constants and Hill coefficients of [³⁵S]ADPS and unlabeled ADPS, the K_B value for ADPS binding was calculated as follows

(5)

It was assumed that the K_B values for binding of unlabeled nucleotides and competitive inhibitor are equal to the corresponding K_i values determined experimentally (for details, see Ref. 5).

For antagonists exhibiting pure noncompetitive potencies, the dose-dependent inhibitions of radioligand binding by unlabeled inhibitors are accounted for by the relation

(6)

in which K_i and m are the apparent inhibition constant and Hill coefficient for noncompetitive inhibitor binding. Data were fitted according to the relationship

(7)

whose slope and x-intercept of the plot give values of m and K_i, respectively.

Statistical analysis. When appropriate, differences between binding capacities were analyzed using Student's t-test or the ANOVA test followed by the Newman-Keuls multiple-comparison test. Differences were considered significant for P values <0.05.

For each analog tested, the 95% confidence interval variation range of pK_B (or pK_i) value (pK_B = log K_B and pK_i = log K_i) was calculated by computerized analysis of the corresponding dose-dependent inhibition binding curves (GraphPad Software; sigmoidal dose-response fit with variable slope). It has been assumed at the 95% probability level that two analogs bind to ampullas with differing affinities only when the lower limit of pK_B (or pK_i) variation range of the best analog is higher than the upper limit of pK_B (or pK_i) variation range of the worst analog.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Figure 1 shows that total and nonspecific binding of 20 nM [³⁵S]ADPS increased linearly with the number of structures used. On these grounds, binding capacities were further expressed as 10¹⁵ mol radioligand bound per ampulla (fmol [³⁵S]ADPS bound/ampulla).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Effects of the number of ampullas from frog semicircular canals on total and nonspecific binding of [35S]adenosine 5'-(2-thiodiphosphate) ([35S]ADPS). Indicated numbers of ampullas were incubated for 3 h at 4°C with 20 nM [35S]ADPS in absence (total binding, ) or presence of 0.1 mM unlabeled ADPS (nonspecific binding, ). Results are means ± SE of 3 or 4 replicates. Equations of linear regression lines are total binding, y = 9.74x + 1.32, r = 0.99; nonspecific binding, y = 0.27x + 0.10, r = 0.97.

Specific binding of 20 nM [³⁵S]ADPS increased exponentially with incubation time at 4°C up to a fairly steady-state level, and the half-time for binding was ~40 min (Fig. 2A). Binding reversibility was checked at 4°C after elimination of free radioligand induced by large dilutions of concentration and of specific radioactivity (Fig. 2B). Clearly, the dissociation kinetics were biphasic and two components could be distinguished: an initial component representing ~85% of specific binding that dissociated rapidly (t_1/2 ~5 min) and a second component that was slowly reversible (t_1/2 ~85 min).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Kinetics of association and dissociation of [35S]ADPS binding to frog ampullas. Association and dissociation time courses were studied in separate experiments. Ampullas were incubated at 4°C with 20 nM [35S]ADPS, and binding was measured as a function of incubation time (A). Binding reversibility was induced by adding 200 µl chilled medium D (see MATERIALS AND METHODS) containing 0.1 mM unlabeled ADPS to ampullas preincubated for 3 h with 20 nM radioligand at 4°C (0 min), and the remaining binding was measured as a function of dissociation time at 4°C (B). Data (means ± SE of 3 replicates performed using 3 ampullas per single measurement) were corrected for nonspecific binding determined in presence of 0.1 mM unlabeled ADPS. Arrow indicates half-time for association binding reaction.

Results depicted in Fig. 3 summarized the main characteristics of ADPS binding to ampullas measured under equilibrium conditions. Specific binding of [³⁵S]ADPS was competitively inhibited by the corresponding unlabeled nucleotide through a concentration range greater than two orders of magnitude; unlabeled ADPS interacted with the population of labeled binding sites with the following kinetic parameters (means ± SE of data from 4 experiments): K_B = 0.48 ± 0.09 µM, Hill coefficient = 0.70 ± 0.06, and R_T = 488 ± 85 fmol [³⁵S]ADPS bound/ampulla. Fitting the dose-dependent inhibition binding curve in Scatchard coordinates generated a curvilinear plot that, after analysis, yielded two straight lines, suggesting the presence in ampullas of two classes of receptors exhibiting differing affinities for [³⁵S]ADPS: a minor class of high-affinity binding sites (maximal binding capacity R_T1 = 52 ± 11 fmol [³⁵S]ADPS bound/ampulla and dissociation constant K_d1 = 0.15 ± 0.04 µM) and a major class of low-affinity receptors (R_T2 = 436 ± 79 fmol [³⁵S]ADPS bound/ampulla and K_d2 = 2.0 ± 0.8 µM).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Dose-dependent inhibition of [35S]ADPS binding to frog ampullas by unlabeled ADPS. Ampullas were incubated for 3 h at 4°C with 20 nM [35S]ADPS and increasing amounts of unlabeled ADPS. A: binding activities (means ± SE of 3 replicates performed in the same experiment using 3 ampullas per single measurement) were corrected for nonspecific binding determined in presence of 0.1 mM unlabeled ADPS. Arrow indicates ADPS concentration leading to half-maximal occupancy of specific binding sites (KB). B: experimental data were fitted as log [(RL0/RL)  1] vs. log (I), in which RL0 and RL are binding activities observed in absence or presence of unlabeled ADPS (I), respectively (see Eq. 3). Equation of linear regression line is y = 0.73x + 4.44, r = 0.99. Kinetic parameters for unlabeled ADPS binding were computed from Eqs. 3 and 5. C: Scatchard plot of the dose-dependent inhibition binding curve. Binding capacities were corrected for dilution of specific radioactivity and ratios RL/L (1/ampulla) between concentrations of bound ligand RL (fmol ADPS bound · ampulla1 · µl1) and free nucleotide L (fmol ADPS/µl) were plotted as a function of RL. Equations of linear regression lines are y = 6.9 × 103 x + 1.01, r = 0.99 (solid line) and y = 0.9 × 103 x + 0.59, r = 0.99 (dotted line). For each class of binding sites, the dissociation constant for ADPS binding (Kd) was computed from slope of the corresponding plot and the maximal binding capacity (RT) was calculated from y-intercept of the solid line (RT1/Kd1 + RT2/Kd2) and from x-intercept of the dotted line (RT1 + RT2). Mean data from 4 different experiments are summarized in text and Table 1.

The stereospecificity of ampulla receptors for recognition of a series of unlabeled structural ATP analogs and drugs acting as antagonists in other systems (4, 12, 17, 33) was investigated in competition experiments similar to those illustrated in Figs. 4 and 5, and all results are summarized in Table 1.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Dose-dependent inhibitions of [35S]ADPS binding to frog ampullas by unlabeled structural ATP analogs. Ampullas were incubated for 3 h at 4°C with 20 nM [35S]ADPS in absence () or presence of increasing amounts of either ,-methyleneadenosine 5'-triphosphate (,-Me-ATP, ), diadenosine tetraphosphate (Ap4A, ), 2-methylthioadenosine 5'-triphosphate (2-MeS-ATP, ), guanosine 5' triphosphate (GTP, ), or uridine 5' triphosphate (UTP, ). A: binding activities (means ± SE of 3 replicates performed in the same experiment using 3 ampullas per single measurement) were corrected for nonspecific binding determined in presence of 0.1 mM unlabeled ADPS and expressed as percent of values measured in absence of inhibitors (: 9.9 ± 0.5 fmol [35S]ADPS bound/ampulla). B: experimental data were fitted as log [(RL0/RL)  1] vs. log (I), where RL0 and RL are binding activities observed in absence or presence of unlabeled inhibitors I, respectively (see Eq. 3). Equations of linear regression lines are y = 0.70x + 3.65, r = 0.99 [,-Me-ATP ()]; y = 0.73x + 3.05, r = 0.99 (Ap4A, ); y = 1.41x + 4.71, r = 0.98 (2-MeS-ATP, ); y = 1.65x + 4.37, r = 0.99 (GTP, ); and y = 1.70x + 3.97, r = 0.99 (UTP, ). Kinetic parameters for binding of unlabeled analogs were computed from Eqs. 3 and 4. Mean data from 2 different experiments for ,-Me-ATP, Ap4A, GTP, and UTP and from 3 different experiments for 2-MeS-ATP are summarized in Table 1.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Inhibitions induced by unlabeled antagonists of [35S]ADPS and unlabeled ADPS binding to frog ampullas. Ampullas were incubated for 3 h at 4°C in presence of 20 nM [35S]ADPS and either with increasing concentrations of suramin () or reactive blue 2 () or with increasing amounts of unlabeled ADPS in presence of 0.1 mM suramin () or 0.5 mM reactive blue 2 (). A: dose-dependent inhibitions of [35S]ADPS binding by unlabeled drugs. Binding capacities (means ± SE of 3 replicates performed in the same experiment using 3 ampullas per single measurement) were corrected for nonspecific binding determined in presence of 0.1 mM unlabeled ADPS and expressed as percent of values measured in absence of inhibitors (: 8.6 ± 0.8 fmol [35S]ADPS bound/ampulla). B: experimental data were fitted as log [(RL0/RL)  1] vs. log (I), in which RL0 are binding capacities measured either in absence of antagonists (closed symbols) or presence of a constant amount of antagonist (open symbols) and RL are the corresponding binding activities observed in presence of increasing concentrations of inhibitors I (see Eqs. 3 and 7). Equations of linear regression lines are y = 0.79x + 3.14, r = 0.99 (suramin, ); y = 1.03x + 5.93, r = 0.96 (ADPS + 0.1 mM suramin, ); y = 1.01x + 3.32, r = 0.97 (reactive blue 2, ); and y = 0.59x + 3.68, r = 0.99 (ADPS + 0.5 mM reactive blue 2, ). Kinetic parameters for suramin binding and for ADPS binding in presence of antagonists were computed from Eqs. 3 and 4 and for reactive blue 2 binding from Eq. 7. Mean data from 2 different experiments are summarized in Table 1.

Most of the nucleotides tested inhibited [³⁵S]ADS binding to the same extent as unlabeled ADPS did, indicating that these analogs interacted with the entire population of labeled binding sites but with differing potencies as regards 1) their apparent dissociation constants [ADPS  ,-Me-ATP = ADP = ATP-S < ATP = A_p4A = AMP < Bz-ATP  2-MeS-ATP < dTTP = GTP = ITP = XTP = CTP = UTP = UDP, whereas 10 mM cAMP decreased by ~25% the number of labeled binding sites and 10 mM adenosine did not inhibit [³⁵S]ADPS binding] and 2) their Hill coefficients [ADPS, ,-Me-ATP and A_p4A bound to ampullas with negative cooperativity phenomena; ADP, ATPS, ATP, AMP, 2-MeS-ATP, XTP, dTTP, and CTP attached according to Michaelian kinetics, whereas Bz-ATP, GTP, ITP, UTP, and UDP interacted with slight positive cooperativity phenomena (Fig. 4 and Table 1)].

As expected, the unrelated nucleotide chemicals revealing antagonistic properties in other systems (4, 12, 17, 33) were able to reduce [³⁵S]ADPS binding to ampullary epithelium (Fig. 5 and Table 1). On the one hand, suramin inhibited competitively [³⁵S]ADPS binding through a concentration range greater than two orders of magnitude, and the presence of 0.1 mM suramin in the incubate increased by about three times the unlabeled ADPS concentration, leading to half-displacement of the remaining labeled binding sites, and enhanced to unity the Hill coefficient value for ADPS binding. Indeed, fitting in Scatchard coordinates the corresponding ADPS-induced inhibition binding curve generated a linear plot (r = 0.94), and computation gave an ADPS concentration, ensuring half-occupancy of residual binding sites K'_B = 1.3 µM. On the other hand, reactive blue 2 and DIDS exhibited pure noncompetitive inhibitor potencies; they decreased [³⁵S]ADPS binding according to Michaelian kinetics but did not impair the K_B and Hill coefficient values for ADPS binding to ampullas.

Finally, the distribution of [³⁵S]ADPS-labeled binding sites in the different regions of the ampulla from frog semicircular canal is illustrated in Table 2. Specific [³⁵S]ADPS binding activities were found in all structures studied (whole ampulla, dorsal region, ventral region, dark cell areas, and hair cell area). Data also show that binding capacities of whole ampulla, ventral region, and dark cell areas did not differ significantly but were higher than those of dorsal region and hair cell area (P < 0.01; ANOVA test followed by the Newman-Keuls multiple-comparison test). Results suggest that the [³⁵S]ADPS-labeled binding sites might be borne mainly by the dark cells. The lack of close additivities between binding capacities of the different ampullary regions might reflect discrepancies in radioligand accessibility to target cells in these isolated structures.

View this table: [in this window] [in a new window]   Table 2.   Distribution of [35S]ADPS binding capacities in different ampullary structures of semicircular canal from R. ridibunda inner ear

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The experiments described above provide new evidence for the presence of ATP receptors in ampulla of semicircular canal from R. ridibunda inner ear, and data also show that [³⁵S]ADPS-labeled binding sites are expressed in all structures of ampullary epithelium.

Despite the minute amounts of tissue used [in the range of 3-5 µg total proteins for 3 ampullas (11)], the validity of the microtechnique employed is verified by the following experiments: 1) total and nonspecific binding of [³⁵S]ADPS increases linearly with the number of ampullas (Fig. 1); 2) specific binding is time-dependent, saturable, and reversible after elimination of free radioligand (Fig. 2); 3) specific [³⁵S]ADPS binding is competitively inhibited by unlabeled ADPS and a series of structural ATP analogs (Figs. 3 and 4); and 4) unrelated nucleotide chemicals acting as antagonists in other systems (4, 12, 17, 33) reduce [³⁵S]ADPS binding (Fig. 5).

Data of pharmacological investigations reveal the following rank order of stereospecificity for recognition of a series of structural nucleotide analogs: ADPS  ,-Me-ATP = ADP = ATP--S > ATP = A_p4A = AMP > Bz-ATP  2-MeS-ATP > dTTP = GTP = ITP = XTP = CTP = UTP = UDP, whereas cAMP and adenosine are almost devoid of activity (Table 1). For antagonists, suramin acts as a competitive inhibitor of the binding reaction because it decreases the apparent affinity of ADPS and does not impair the maximal binding capacity, whereas reactive blue 2 and DIDS exhibit pure noncompetitive inhibitor potencies because they decrease the total number of labeled [³⁵S]ADPS binding sites and do not modify the K_B value for ADPS binding (Fig. 5 and Table 1).

It seems unlikely that the differences observed between the apparent dissociation constants for binding of nucleotides to ampullas result from ecto-ATPase and ectonucleotidase activities (8, 9, 29) because assays were performed at 4°C and in the absence of divalent cations, i.e., under experimental conditions limiting enzymatic breakdown of nucleotides (33). Indeed, 1.8 mM CaCl₂ and 1 mM MgCl₂ decrease by 32% the number of specific binding sites labeled with 20 nM [³⁵S]ADPS (see MATERIALS AND METHODS).

It is worth noting that the K_B value for ATP binding to frog ampullas is far higher than 1) the very low concentrations of ATP introduced into the perilymphatic space of R. pipiens semicircular canal, which increase the spontaneous electrical activity of afferent fibers (2, 3), and 2) the real concentrations of nucleotide assayed in endolymphatic and perilymphatic compartments of the guinea pig cochlea (21). These observations suggest that ATP would mediate its biological effects at low fractional receptor occupancy, if the so-called spare receptor phenomenon is operative in frog semicircular canal as reported for many other hormonal systems (for instance, see Ref. 16). In addition, it should be stressed that the apparent affinities for ADPS and ,-Me-ATP binding to frog ampullas are lower than the corresponding values for ADPS attachment to P_2Y receptors from avian erythrocyte membranes and cultured aortic endothelial cells (7, 31) and for ,-Me-ATP binding to P_2X receptors from rat urinary bladder membranes (12). These discrepancies might reflect some zoological stereospecificities in nucleotide sensitivity of target cells, as observed earlier for neurohypophysial hormone receptors of kidneys from various amphibian and mammalian species (1, 5, 16).

Results of pharmacological experiments argue for the absence of P₁ receptors among the population of ampullary [³⁵S]ADPS labeled binding sites because adenosine is quite inactive for inhibition of [³⁵S]ADPS binding, data in agreement with the lack of P₁ receptors coupled to phosphoinositidase C activation found in the same structure (6). On the other hand, it must be pointed out that the pyrimidine nucleotides (dTTP, CTP, UTP, and UDP) bind to ampullas with very low affinities. Moreover, the K_B value for UTP binding to ampullary epithelium is far higher than its corresponding activation constant for phosphoinositidase C stimulation in the same target structure (6). These results compare well with those obtained in cultured bovine aortic endothelial cells (25, 31, 32) and indicate that [³⁵S]ADPS labels barely the pyrimidine nucleotide-sensitive P_2(Y) receptors (13) triggering phosphoinositidase C activation in frog semicircular canal (6).

Taken together, data of kinetic and pharmacological studies suggest strongly that the population of ampullary [³⁵S]ADPS-labeled receptors is heterogeneous because 1) the dissociation kinetics of [³⁵S]ADPS binding sites are biphasic (Fig. 2), 2) Scatchard plot of ADPS binding may be analyzed as a combination of a minor class of high-affinity receptors and a major class of low-affinity binding sites (Fig. 3), and 3) ADPS binds to frog ampullas with an apparent affinity that is ~40 times higher than its own affinity for stimulation of phosphoinositidase C in the same system (6). In addition, no correlation occurs between the pK_a values for phosphoinositidase C activation by potent purine or pyrimidine nucleotides and/or the pK_i values for inhibition by antagonists of ATP-induced enzyme stimulation (6) and their corresponding pK_B or pK_i values for attachment to frog ampullas (r = 0.02). Thus it seems likely that the majority of the [³⁵S]ADPS-labeled binding sites do not represent the biological receptors involved in phosphoinositidase C activation. Among all chemicals tested, only the natural principle ATP, the long-acting analog ATPS, and 2-MeS-ATP bind to frog ampullas and stimulate phosphoinositidase C with similar apparent affinities (6).

Moreover, the recognition pattern of structural ATP analogs by ampullary receptors strengthens the hypothesis that both P_2X-like and P_2Y-like receptors are coexpressed in ampulla of R. ridibunda semicircular canal. Thus 1) ,-Me-ATP, which recognizes P_2Y receptors only poorly (12, 13, 33), binds to ampullas with an apparent affinity far higher than its own affinity for phosphoinositidase C activation (6) and than that for binding of 2-MeS-ATP; 2) 2-MeS-ATP interacts with ampullary receptors with an apparent dissociation constant close to its corresponding activation constant for phosphoinositidase C stimulation (6); 3) the natural purine nucleotides GTP, ITP, and XTP bind to the population of labeled receptors with very low affinities that are about one order of magnitude smaller than their corresponding affinities for phosphoinositidase C activation (6); 4) the antagonist suramin (12, 17, 33) exhibits competitive inhibitor properties and attaches to ampullas with a K_B value that is 100 times lower than its K_i value for inhibition of ATP-stimulated phosphoinositidase C activity (6) and 10 times higher than its dissociation constant for binding to P_2X receptors of rabbit isolated ear artery (17); and 5) the P_2X antagonist DIDS (4) acts as a pure noncompetitive inhibitor for interaction with labeled [³⁵S]ADPS binding sites, whereas it inhibits competitively ATP-induced phosphoinositidase C activation (6).

On these grounds, it seems reasonable to postulate that, in ampulla from R. ridibunda semicircular canal, the major class of low-affinity binding sites for [³⁵S]ADPS might contain mainly the P_2X-like receptors and that the minor class of high-affinity binding sites might represent the P_2Y-like receptors triggering phosphoinositidase C stimulation (6).

Furthermore, it is worth noting that the pharmacological properties of the P_2X-like and P_2Y-like receptors found in frog ampulla differ from those of the ,Me-ATP-sensitive P_2X receptors and of the P_2Y1 receptors cloned from mammalian and avian species. Indeed, the P_2X1 and P_2X3 receptor channels exhibit the following recognition pattern of agonists: 2-MeS-ATP > ATP > ,-Me-ATP > ADP, and the P_2Y1 receptors trigger phosphoinositidase C activation with the following rank order of stereospecificity: 2-MeS-ATP > ATP > ADP  UTP (13). These observations suggest that the P_2X-like and P_2Y-like receptors from frog semicircular canal might represent novel P_2X and P_2Y receptor subtypes. So, further molecular cloning studies will be needful to characterize these P_2X-like and P_2Y-like receptors. On the other hand, our results do not exclude the eventual presence of P_2YAp4A receptors and of P_2X7 receptors (13) among the population of [³⁵S]ADPS-labeled binding sites because 1) the selective P_2YAp4A agonist A_p4A (12, 13, 33) binds to frog ampullas with negative cooperativity phenomena and with an apparent affinity that is 20 times lower than that of ,Me-ATP and 10 times higher than that of 2-MeS-ATP and 2) the potent marker of P_2X7 receptors, Bz-ATP (13, 33), binds to ampullas with a K_B value close to that of 2-MeS-ATP. But these observations call for additional investigations.

Perspectives