Evidence for the Functional Expression and Pharmacological Characterization of Adenine Receptors in Native Cells and Tissues
Pharmaceutical Institute, Pharmaceutical Chemistry Poppelsdorf, (S.G., C.E.M.) and Department of Pharmacology and Toxicology (I.v.K.), University of Bonn, Bonn, Germany
Department of Chemical Sciences, University of Camerino, Camerino, Italy (S.V., R.V., G.C.)
Abstract
An orphan G protein-coupled receptor from rat has recently been discovered to be activated by the nucleobase adenine (Proc Natl Acad Sci USA 99:8573eC8578, 2002). In the present study, we show for the first time that the adenine receptor is expressed in membrane preparations of native tissues and cell lines in high density, including rat brain cortex, rat brain striatum, and the mouse neuroblastoma x rat glioma hybrid cell line NG108-15. Saturation analysis with [3H]adenine at rat brain cortical membranes exhibited a single high-affinity binding site with a KD value of 27.2 nM, and a binding capacity of 2.28 pmol/mg of protein. Kinetic studies revealed unusual binding kinetics of [3H]adenine with rapid association and slow dissociation. A series of compounds were investigated in [3H]adenine competition experiments at rat brain cortex. Only minor substitution of the adenine structure was tolerated, the most potent compounds of the present series being 2-fluoroadenine (Ki value of 620 nM), 8-thioadenine (Ki value of 2.77 e), N6-methyladenine (Ki value of 3.64 e), and 7-methyladenine (Ki value of 4.13 e), all of which were partial agonists (40eC60% intrinsic activity). Adenine dose dependently inhibited forskolin-stimulated adenylate cyclase in membrane preparations of NG108-15 cells as well as in intact cells, showing that the receptor is functional in NG108-15 cells. Reverse transcriptase-polymerase chain reaction experiments followed by sequencing indicate that the NG108-15 cells express the murine ortholog of the adenine receptor. Moreover, preliminary radioligand binding studies with [3H]adenine at membranes of human astrocytoma 1321N1 cells suggest that a human ortholog of the rat adenine receptor exists.
Adenine was recently identified as the endogenous ligand of an orphan rat G protein-coupled receptor by a "reverse pharmacology" approach. mRNA localization studies revealed the highest expression in the small neurons of the dorsal root ganglia, suggesting a role in nociception (Bender et al., 2002). Potent neurotrophic effects of adenine have recently been described; in primary cultures of rat cerebellar Purkinje cells, the number of cells was greatly increased by the addition of adenine, although relatively high concentrations (1eC2 mM) were required (Yoshimi et al., 2003). Adenine was the most potent compound in a series of purine and pyrimidine derivatives that showed protective effects on Purkinje cells and prevented cell death, whereas other adenine-based purines, including adenosine, AMP, ADP, and ATP, did not promote Purkinje cell survival (Watanabe et al., 2003).
Radioligand binding studies had previously been performed at membrane preparations of CHO cells stably transfected with the rat adenine receptor using [3H]adenine as a radioligand (Bender et al., 2002). The recombinant rat adenine receptor expressed in CHO cells was found to be coupled to inhibition of adenylate cyclase via Gi protein (Bender et al., 2002). However, the postulated adenine receptor has not been demonstrated to be present in native tissues on the protein level, and therefore it has not yet been officially accepted as a novel receptor by the International Union of Pharmacology committee on receptor nomenclature. A mouse ortholog (mMrgA10), but no human ortholog, of the rat adenine receptor could be identified by sequence analysis and comparison (Bender et al., 2002).
The present study proves the existence of the adenine receptor in brain tissues and native cell lines and its functional coupling to inhibition of adenylate cyclase in NG108-15 cells. Furthermore, we have explored the structure-activity relationships of a large series of compounds structurally related to adenine.
Materials and Methods
Radioligand Binding Studies
[3H]Adenine Saturation Experiments. Saturation binding experiments were carried out at rat brain cortical and rat brain striatal membranes using [8-3H]adenine (27 Ci/mmol; Amersham Biosciences Europe GmbH, Freiburg, Germany) as a radioligand. Frozen rat brain membranes were obtained by Pel Freez (Rogers, AR) and thawed at 4°C. Cortex and striatum were dissected, and membrane fractions were prepared as described previously (Meler et al., 2002). Protein concentrations of 100 e/ml were used in the saturation binding experiments, determined by the method of Lowry (Lowry et al., 1951) using a Sigma Chemie protein assay kit. Incubations were carried out in a total volume of 200 e of 50 mM Tris-HCl buffer, pH 7.4, supplemented with 10 mM MgCl2 and 1 mM EGTA, at room temperature (23°C) for 60 min. Nonspecific binding was determined with 10 e adenine. In some experiments, the stable GTP analog guanosine-5'-(,-imido)triphosphate (GppNHp, 100 e) was present.
Incubation was terminated by rapid filtration through GF/B glass fiber filters, presoaked in 0.5% aqueous polyethylenimine solution for 60 min using a Brandel 48-channel cell harvester. Filters were rinsed three times with 2 ml each of ice-cold 50 mM Tris-HCl buffer, pH 7.4. The wet filters were transferred to mini-vials and scintillation cocktail was added. Drying of the GF/B glass fiber filters before the addition of scintillation cocktail did not influence the results. Radioactivity on the punched-out filters was measured after 9 h of preincubation with 2.5 ml of Ultima Gold scintillation cocktail (Canberra Industries, Dreieich, Germany).
[3H]Adenine Competition Assays and Kinetic Studies. The adenine binding assays were carried out at rat brain cortical membranes using 10 nM [3H]adenine as a radioligand. Protein amounts of 100 e/tube were used in the assays. Inhibition curves were determined using seven to nine different concentrations of the test compound, spanning 3 orders of magnitude. Three separate experiments were performed each in triplicate, unless otherwise noted. Nonspecific binding was determined with 10 e adenine and amounted to 10 to 20% of total binding. Incubations were carried out in a total volume of 0.5 ml of 50 mM Tris-HCl buffer, pH 7.4, supplemented with 10 mM MgCl2 and 1 mM EGTA, at room temperature (23°C) for 60 min. Further experimental procedures were as described above for saturation studies. Kinetic studies were performed using 15 nM [3H]adenine. Dissociation was initiated by the addition of 1 e adenine after 60 min of preincubation. Three separate experiments, each in triplicate, yielded similar results. Competition binding assays at membrane preparations of NG108-15 cells and astrocytoma 1321N1 cells were performed as described above for rat brain cortical membranes using 100 e of protein (NG108-15), or 150 e of protein (astrocytoma), respectively, per tube.
Cell Culture. NG108-15 cells were provided by Dr. Bres (Pharmacological Institute, University of Bonn, Bonn, Germany) and grown as described previously (Kaulich et al., 2003). For radioligand binding studies, the culture medium was removed and the cells were kept frozen ateC80°C until the preparation of the membranes. Cells were thawed and detached in ice-cold 50 mM Tris-HCl buffer, pH 7.4, with a cell scraper. The cell suspension was homogenized on ice and spun down for 10 min at 1000g at 4°C. The supernatant was centrifuged at 20,000g for 30 min at 4°C, and the resulting pellet was resuspended in 50 mM Tris-HCl buffer, pH 7.4. The last centrifugation step was repeated once. The protein concentration was determined by the method of Lowry (Lowry et al., 1951). Membranes were kept frozen ateC80°C until use. For cAMP assays, membranes were used immediately. Astrocytoma 1321N1 cells were grown adherently and maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 100 U/ml penicillin G, 100 e/ml streptomycin, and 1% GlutaMAX at 37°C, 5% CO2. Cells were grown to confluence. For binding assays, culture medium was removed and cells were kept frozen ateC80°C until membrane preparation following the procedure described by Griessmeier and Meler (2004).
Molecular Biological Experiments. Poly A+ mRNA was isolated from NG108-15 cells cultured as described above using the Oligotex Direct mRNA mini kit (QIAGEN GmbH, Hilden, Germany). Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed using gene-specific primers for the rat adenine receptor (GenBank accession no. NM_145787 [GenBank] ; sense primer, CACCATGGGGGAAAGCTTCACCGGTACAG; antisense primer, CGGCTCCACCTTGCTGCTTGACATC) as well as the putative murine ortholog MrgA10 (GenBank accession no. XM_195647 [GenBank] ; sense, CACCATGGGGGAAAGCAGCACCGGTGCAG; antisense, TGGCTCTGCTTTGCTTTTTGACATCTCC) and the Superscript One Step RT-PCR kit (Invitrogen, Karlsruhe, Germany). Annealing temperature for the PCR reaction was 56.0°C (35 cycles). For control experiments, the enzyme reverse transcriptase was omitted (instead of the RT-PCR enzyme mix, Taq Platinum Polymerase was used; Invitrogen). PCR products were analyzed by ethidium bromide staining after agarose (0.9%) gel electrophoresis. After amplifying the product of one RT-PCR reaction using Platinum Pfx polymerase (Invitrogen), the sequence was cloned into the expression vector pcDNA3.1D/V5-His-TOPO (Invitrogen). The sequence was then identified by cycle sequencing (SequiTherm Exel II DNA sequencing kit; Epicenter Technologies, Madison WI) using a LICOR Gene READIR 4200 sequencer (MWG-Biotech, Ebersberg, Germany).
[35S]GTPS Binding Assay. The [35S]GTPS binding assay was carried out as described previously (Bender et al., 2002) with minor modifications. The incubation buffer consisted of 20 mM HEPES, pH 7.4, containing 100 mM NaCl, 10 e GDP, 10 mM MgCl2, and 0.5% bovine serum albumin. After a preincubation period of 30 min of rat cortical membranes (10 e/tube) with agonist (adenine or N6-cyclopentyladenosine) in the presence of 2 IU/ml adenosine deaminase, [35S]GTPS was added in a final concentration of 0.5 nM (final volume 200 e). After a further incubation period of 60 min, bound and free [35S]GTPS were separated by filtration through a GF/B glass fiber filter, presoaked in 50 mM Tris-HCl buffer, pH 7.4, containing 10 mM MgCl2.
Functional Studies
Cyclic AMP Accumulation. Using membrane preparations. Culture flasks with almost confluently grown NG108-15 cells were taken, and cells were detached with a cell scraper. A membrane preparation of the cells was obtained according to the method described above. The membrane fraction was resuspended in 50 mM Tris-HCl buffer, pH 7.4, supplemented with 4 mM EDTA, 10 mM MgCl2, and 5 e GTP. Stock solutions of compounds in dimethyl sulfoxide were added to 50 mM Tris-HCl buffer, pH 7.4, containing 4 mM EDTA, 10 mM MgCl2, 5 e GTP, 10 e forskolin, and 1 mM ATP, and the mixture was incubated for 20 min at 37°C with 90 eof the membrane suspension. Incubation was terminated by rapidly heating to 98°C for 2 min. After cooling down to room temperature, the tubes were centrifuged at 2000g for 15 min at 4°C, using an Allegra 21 R centrifuge (Beckman Coulter GmbH, Krefeld, Germany). The supernatant (50 e) was assayed in a charcoal adsorption-assay using a commercially available kit (Amersham Biosciences Inc.) to determine the cAMP concentration. Each experiment was carried out in duplicates at least three times.
Using whole cells. Cyclic AMP accumulation was determined in intact NG108-15 cells plated on a 48-well culture plate. After cells were almost grown to confluence the cell culture medium was removed, cells were washed twice with Krebs-Ringer-HEPES buffer, pH 7.4, containing 130 mM NaCl, 4.7 mM KCl, 4.0 mM NaHCO3, 1.2 mM KH2PO4, 1.2 mM MgSO4, 1.8 mM CaCl2·2H2O, 11.5 mM glucose, 10 mM HEPES, and 0.1% bovine serum albumin, and were then preincubated with adenosine 1 IU/ml deaminase, and the phosphodiesterase inhibitors Ro20-1724 (500 e) and 3-isobutyl-1-methylxanthine (500 e) in Krebs-Ringer-HEPES buffer, pH 7.4, for 10 min at 37°C. Forskolin (final concentration 10 e) and a range of different concentrations of adenine were added. The incubation at 37°C was stopped after 10 min by removal of the reaction buffer. Cells were lysed with 0.1 M HCl for 30 min at room temperature. After neutralization with 1 M NaOH solution and further dilution, the amount of cAMP was determined as described above. Each experiment was carried out in duplicates at least three times. In some experiments, cells were preincubated overnight in medium in the presence of pertussis toxin (PTX, 10, 100, or 1000 ng/ml, respectively). All three concentrations gave virtually identical results.
Data Analysis. Data were analyzed using Prism 3.0 (GraphPad Software Inc., San Diego, CA). IC50 values were determined by fitting data to a sigmoidal curve with variable slope. The Cheng-Prusoff equation and the KD value of 27.2 nM for [3H]adenine was used to calculate Ki values from the obtained IC50 values.
Chemicals. 1-Methyladenine, 7-methyladenine, and 9-methyladenine were purchased from Acros Organics (Geel, Belgium). 2-Chloroadenine, 7-ethyladenine, 9-dihydroxypropyladenine, 9-ethyladenine, 8-bromoadenine, 1-deazaadenosine, 3-deazaadenosine, 2',3'-dideoxyadenosine, 2-[3-(1-phenylprop-2-yn-1-ol)-9-ethyladenine, 2-phenylethynyl-9-ethyladenine, 2-(hex-1-yn-1-yl)-9-ethyladenine, 8-phenyl-9-ethyladenine and 8-bromo-9-propyladenine were synthesized according to or in analogy to described procedures (Cristalli et al., 1987; Camaioni et al., 1998; Klotz et al., 2003). All other reagents were obtained from Sigma Chemie (Deisenhofen, Germany).
Results
Saturation binding experiments at rat brain cortical membranes performed with [3H]adenine revealed a single high-affinity binding site with a KD value of 27.2 ± 0.9 nM (n = 5, Fig. 1A). A binding capacity (Bmax) of 2.28 ± 0.01 pmol/mg of protein was determined. The presence of a stable analog of GTP, 100 e GppNHp, led to a small increase in the KD value [from 27.2 ± 0.9 (n = 5) to 32.6 ± 2.7 nM (n = 3)], which seemed to be reproducible, but did not quite reach statistical significance (P = 0.058, two-tailed t test). The Bmax value was virtually unaltered [2.28 ± 0.01 (n = 5), without GppNHp versus 2.58 ± 0.37 pmol/mg of protein (n = 3), with GppNHp, P = 0.47]. Adenine (0.01 nMeC500 e) did not have any effect on [35S]GTPS binding to rat brain cortical membranes. In membranes prepared from rat brain striatum, [3H]adenine saturation experiments exhibited a single binding site with a KD value of 72.4 ± 5.0 nM and a Bmax value of 8.23 + 0.8 pmol/mg of protein (n = 3, Fig. 1B). For the saturation experiments, a two-site model did not give an improved fit over a one-site model. Kinetic studies (Fig. 2) showed that [3H]adenine bound very rapidly to membranes from rat brain cortex. Steady state seemed to be reached as early as the first time point examined (30 s) and was sustained for at least 200 min (not shown). Binding was slowly reversed by the addition of adenine in a final concentration of 1 e, with 60% of dissociation being evident after 60 min. A kinetic KD value could not be determined. The law of mass action did not apply to the system.
A concentration of 10 nM [3H]adenine was suitable for competition binding assays at rat brain cortical membranes. At this concentration, specific binding amounted to 80 to 90% of total binding. The addition of dimethyl sulfoxide in concentrations up to 10% had no effect on [3H]adenine binding (data not shown). NaCl (100 mM) did not alter [3H]adenine binding significantly (data not shown).
Binding affinities for a series of compounds were determined in competition experiments at rat brain cortical membranes. Adenine itself showed a Ki value of 29.9 ± 3.4 nM (Fig. 3). The investigated compounds included adenine derivatives with substitution in the 1-, 2-, 3-, N6-, 7-, 8-, or 9-position (Tables 1, 2, 3; Fig. 3), xanthine derivatives (Table 2), and a selection of drugs with analgesic properties (salicylic acid, acetylsalicylic acid, acetaminophen, indomethacin, and propyphenazone) and with effects against neuropathic pain (the antidepressants amitriptyline, imipramine, and doxepin). Only adenine derivatives with small structural modifications showed appreciable affinity for the rat brain adenine receptor. All compounds investigated were considerably (at least 20-fold) less potent than adenine. The most potent derivatives were 2-fluoroadenine (3, Ki value of 620 nM), 8-thioadenine (14, Ki value of 2.77 e), N6-methyladenine (6, Ki value of 3.64 e), and 7-methyladenine (11, Ki value of 4.13 e) (Fig. 4). The following rank order of potency could be established: adenine >> 2-fluoroadenine > N6-methyladenine = 7-methyladenine > 9-methyladenine 1-methyladenine > 3-methyladenine/N6-dimethyladenine (Table 1). Adenosine (18) exhibited binding affinity in the micromolar range (Ki value of 19.4 ± 6.3 e) and its 2',3'-dideoxy analog (19) was equally potent (Ki value of 18.2 ± 3.4 e) (Table 2). 1-Deazaadenosine (20), 3-deazaadenosine (21), and the adenosine derivative NECA (22) (Table 2) showed no significant affinity for the adenine receptor. The addition of adenosine deaminase, which is used in radioligand binding studies at adenosine receptors to remove endogenous adenosine by converting it to inosine, did not have any effect on the [3H]adenine binding studies (data not shown). A selection of adenine derivatives, with combinations of substituents in the 2-, 8-, and 9-position that had previously been shown to be potent adenosine A2A receptor antagonists, were found to be inactive at the rat brain adenine receptor (Table 3).
[3H]Adenine binding studies at membrane preparations of two permanent cell lines, the mouse neuroblastoma x rat glioma cells NG108-15 and the human astrocytoma 1321N1 cell line, revealed that both contained specific binding sites for adenine (Fig. 4). Competition assays at NG108-15 cell membranes using 10 nM [3H]adenine yielded an IC50 value of 154 ± 56 nM for unlabeled adenine and similar affinities and the same rank order of potency for a selection of adenine derivatives as seen in binding studies at rat brain cortex (Table 4).
Freshly prepared membranes from NG108-15 cells and intact NG108-15 cells were used for functional studies (Fig. 5). At membranes adenine dose dependently inhibited forskolin-stimulated adenylate cyclase with an IC50 value of 21 nM (95% confidence interval 7eC67 nM). Forskolin stimulated the basal cAMP production of 6 pmol/mg of protein/min about 4-fold to 25 pmol/mg of protein/min. Maximal inhibition of 50% was observed at 3 e adenine. The addition of adenosine deaminase had no effect on cAMP accumulation (data not shown). Selected adenine derivatives were investigated in functional assays determining the inhibition of forskolin-stimulated cAMP accumulation at a single, high concentration. All of the compounds were partial agonists at the rat adenine receptor with somewhat weaker maximal inhibition as the full agonist adenine (Table 5). Functional studies were additionally performed using whole NG108-15 cells, where adenine dose dependently inhibited forskolin-stimulated adenylate cyclase activity with an IC50 value of 2.54 ± 0.99 e. Maximal inhibition of 70% was observed at 30 e adenine. Preincubation of the cells with pertussis toxin (10, 100, or 1000 ng/ml, respectively) led to a decrease of the maximal stimulation by forskolin of the cAMP production by 50% (data not shown). Pertussis toxin completely prevented adenine-induced inhibition of the forskolin-stimulated cAMP production (Fig. 6).
RT-PCR with primers specific for the coding sequence of the putative mouse ortholog but not with primers for the rat adenine receptor revealed products of the expected length from mRNA preparations prepared from cultured NG108-15 cells (Fig. 7; one of three independent runs). Control reactions without the enzyme reverse transcriptase showed no products confirming that mRNA but not genomic DNA acted as template for the PCR reaction. Cycle sequencing of one RT-PCR reaction product confirmed that it was a murine and not a rat sequence.
Discussion
Purinergic membrane receptors for the nucleoside adenosine (P1 receptors) and the adenine nucleotides ATP and ADP (P2 receptors) have been known to exist for more than 20 years. The aglycon adenine has been found to activate an orphan rat G protein-coupled receptor belonging to the MAS-related gene (Mrg) receptor family with high affinity (Ki value of 18 nM) (Bender et al., 2002). mRNA expression of the rat adenine receptor was detected in many tissues with highest levels found in the small neurons of the dorsal root ganglia (Bender et al., 2002). We have now for the first time investigated native adenine receptors on the protein level with radioligand binding and functional studies.
In saturation studies at rat brain cortical membranes, [3H]adenine exhibited a KD value of 27.2 nM. This is well in accordance with literature data for the recombinant rat adenine receptor expressed in CHO cells (KD value of 24 nM) obtained at virtually identical conditions (Bender et al., 2002), demonstrating that [3H]adenine actually labels the adenine receptor in rat brain cortex. At rat brain striatal membranes, the determined KD value was somewhat higher (72.4 nM) but still in the same concentration range. In both preparations, [3H]adenine labeled only a single binding site with high affinity, indicating that only a single adenine receptor seems to be present.
The adenine receptor density in rat cortical membranes was determined to be quite high, with a Bmax value of 2.28 pmol/mg of protein; it was even higher in striatal membranes (8.23 pmol/mg of protein). In comparison, the density of the highly expressed adenosine A1 adenosine receptor in the rat brain is severalfold lower (Patel et al., 1982, Lloyd and Stone, 1985). The mRNA profile determined by Bender et al. (2002) had indicated low expression in striatum and somewhat higher expression in brain cortex. Our results show that mRNA levels do not necessarily correlate with protein expression levels of adenine receptors. The detected high densities of adenine receptors in rat brain cortex and striatum point to an important physiological role of these receptors.
Guanine nucleotides have been shown to decrease the affinity of inhibitory agonists for their receptor (Yeung and Green, 1983; Stiles, 1988). This has been explained by a shift from a high-affinity to a low-affinity receptor state in the presence of GTP. Such an effect is sometimes not very pronounced or even undetectable (Kelly et al., 1998). Saturation studies in the presence of the stable GTP analog GppNHp (100 e) at rat brain cortical membranes revealed that the binding of [3H]adenine was somewhat influenced by the guanine nucleotide, which led to a moderate increase in the KD value. Although GTP and its analogs can lead to decreased binding of an agonist to a Gi-coupled receptor, the reverse allosteric effect has also been observed, namely, a stimulation of [35S]GTPS binding to the Gi protein by a receptor agonist. [35S]GTPS binding studies work well in artificial systems with high receptor expression levels. In fact, Bender et al. (2002) had shown that adenine increased [35S]GTPS binding in membranes prepared from CHO cells recombinantly expressing the rat adenine receptor. In native rat cortical membranes, we could not show adenine-induced enhancement of [35S]GTPS binding. Showing such an effect in native tissues is sometimes difficult or even impossible. N6-Cyclopentyladenosine, an agonist at the Gi-coupled adenosine A1 receptor, also did not show any measurable increase in [35S]GTPS binding at native membranes in our hands, whereas it strongly increased [35S]GTPS in recombinant CHO cells expressing the human adenosine A1 receptor in high density (data not shown).
Kinetic studies demonstrated that the binding of [3H]adenine to rat brain cortical membranes was reversible. However, after a very fast association, dissociation initiated by the addition of a high concentration of unlabeled adenine was slow in comparison (Fig. 3). A kinetic KD value could not be determined. An explanation for the observed unusual kinetics could be that the adenine receptor undergoes a conformational change after binding of [3H]adenine; an initial low-affinity conformation may be converted to a high-affinity conformation for adenine upon binding.
The binding site of the adenine receptor in rat brain cortex was further characterized by competition assays, and structure-activity relationships were determined by investigating a series of differently substituted adenine derivatives. Bender et al. (2002) had evaluated only a very limited selection of compounds at the recombinant rat adenine receptor, namely, 1-methyladenine, 6-benzyladenine, adenosine, AMP, ADP, ATP, hypoxanthine, and guanine, none of which was remarkably potent. Our results for this collection of compounds at the rat brain cortex receptor were mainly consonant with the results obtained by Bender et al. (2002) at the recombinant adenine receptors. The only major discrepancy was that adenosine seemed to exhibit measurable affinity for the adenine receptor in rat brain cortical membranes (Ki value of 19.4 e), whereas it was found inactive at the recombinantly expressed receptor in CHO cell membranes. However, the determined Ki value at rat brain cortex was almost 3 orders of magnitude higher than that for adenine. It cannot be excluded at present that adenosine is degraded by a purine nucleoside phosphorylase during the 60-min incubation with rat brain cortical membranes and the detected binding is the result of formed adenine. The same may also be true for the adenine nucleotides AMP (23), ADP (24), and ATP (25), which show very weak affinity (39eC51% inhibition of radioligand binding at 100 e): the observed inhibition may be caused by adenine formed through enzymatic degradation, although EGTA was present in the assays to inhibit nucleotidases.
In the competition studies, adenine itself exhibited a Ki value of 29.9 nM, which is in good agreement with literature data for the recombinant receptor (18 nM; Bender et al., 2002) and with the determined KD value (27.2 nM). Even only small modifications at the adenine core structure led to a major reduction or loss of affinity. In a series of mono-substituted adenine derivatives (Table 1), substitution with a fluorine atom in the 2-position was best tolerated, resulting in 2-fluoroadenine (3); however, it was 20-fold less potent than the unsubstituted adenine (620 versus 30 nM). A fluorine atom is sterically similar to a hydrogen atom but has strong electronic effects (negative inductive effect) that must be responsible for the reduced affinity of 3 in comparison with adenine. Replacement of the fluorine atom by the larger halogen atom chlorine (compound 4) further reduced affinity by 60-fold (Ki value of 37.9 e), showing that the receptor does not have any bulk tolerance in that region. Substitution of the exocyclic amino group was investigated. A methyl group was best tolerated (6, Ki value of 3.64 e), whereas the larger benzyl (7) and benzoyl (8) residues resulted in very weak binding. It is interesting that the more polar benzoyl substitution was better tolerated (Ki value of 46.6 e) than a benzyl residue (Ki value >100 e). The N6-hydroxypentenyl-substituted adenine derivative trans-zeatin, a cytokinin (phytohormone), was somewhat more potent than 8 (Ki value of 22.6 e); it seems that polar substituents are better accepted by the receptor than unpolar residues. Dimethylation of the amino group abolished affinity, indicating that a hydrogen bond donor (N-H) in that position was probably required for binding to the receptor, similarly as in adenosine receptors (Meler, 2000). Small polar substituents in the 8-position, such as thiol (14, Ki value of 2.77 e) and bromine (13, Ki value of 17.3 e) were tolerated by the receptor leading to moderately potent compounds. Methylation at the ring nitrogen atoms of adenine was explored. Methylated adenine derivatives exhibited quite different binding affinities depending on the position of the methyl group. The rank order of potency was 7-methyladenine (11, Ki value of 4.13 e) > 9-methyladenine (15, Ki value 17.5 e) 1-methyladenine (2, Ki value of 29.3 e) >> 3-methyladenine (5, Ki value >>100 e). Replacement of the methyl group in 7- or 9-methyladenine by an ethyl residue resulted in a large drop in affinity (compare 11/12 and 15/16), indicating that the receptor did not have bulk tolerance in that positions. However, a more hydrophilic substituent in the 9-position, a dihydroxypropyl residue (compound 17), was somewhat better tolerated than an ethyl residue.
A series of 9-alkyl-substituted adenine derivatives that bear an additional substituent either in the 2- or 8-position were recently described as relatively potent adenosine A2A receptor antagonists (Klotz et al., 2003). A2A antagonists acting at receptors in the striatum (caudate putamen) are currently developed as novel therapeutics for Parkinson's disease (Chen, 2003). Some of the disubstituted adenine derivatives, including compound 31, had been evaluated in animal models of Parkinson's disease. The compounds had induced contralateral turning in 6-hydroxydopamine-lesioned rats after systemic application in a dose of 5 mg/kg. In contrast, the standard A2A antagonist SCH58261had required priming by dopamine to induce the same effect (M. Morelli, oral presentation at the Satellite Symposium on Purinergic Receptors, Camerino, Italy, 2003). Because we have found that adenine receptors are highly expressed in the rat striatum, we investigated whether the adenine derivatives described as adenosine A2A receptor antagonists were selective for the adenosine receptors or whether they showed additional binding to the adenine receptor. In fact, the disubstituted adenine derivatives did not exhibit any affinity for adenine receptors in concentrations up to 100 e (Table 3).
Adenosine (18) exhibited a more than 600-fold lower affinity in comparison with the natural ligand adenine. The potent, nonselective adenosine receptor agonist NECA (22), which has been found to also bind with high affinity to nonadenosine receptor binding sites (Lorenzen et al., 1993, 1996; Yoshioka et al., 2001), showed no significant binding affinity for the adenine receptor. As pointed out above, adenosine might be partially degraded to adenine during the assay, whereas NECA is enzymatically more stable. 2',3'-Dideoxyadenosine (19), which lacks the 2'- and 3'-hydroxyl groups at the ribose moiety, exhibited a similar Ki value as adenosine. 1-Deazaadenosine (20) and 3-deazaadenosine (21) showed only low affinity for the adenine receptor, but it seems that the N-atom in position 1 is more important for binding than the N-atom in position 3. The adenine nucleotides AMP (23), ADP (24), and ATP (25) seemed to exhibit some affinity to the adenine receptor in similar, high concentrations (39eC51% inhibition of radioligand binding at 100 e), but this might be caused by degradation of the nucleotides to adenosine or adenine by enzymes present in the membrane preparations. The obtained structure-activity relationships are not consistent with those obtained for the ATP binding site of adenylate cyclases, the so-called P-site (Dessauer et al., 1999; Levy et al., 2003).
All other compounds that we screened and that were not derived from adenine, including the xanthine derivatives caffeine (26), theophylline (27), 8-cyclopentyltheophylline (28), and 1,3-dipropyl-8-cyclopentylxanthine (29), all of which are adenosine receptor antagonists, exhibited no significant binding affinity for the adenine receptor at a high concentration of 100 e. A set of analgesic and antidepressant compounds were investigated because of the finding that the adenine receptor is highly expressed in the small neurons of the dorsal root ganglia (Bender et al., 2002), suggesting a role in nociception. Neither the analgesics nor the tricyclic antidepressants, which have been shown to be effective in neuropathic pain (Mattia and Coluzzi, 2003), exhibited significant binding affinity for the adenine receptor.
Specific binding sites for [3H]adenine with similar properties as in brain cortex were also detected in the permanent mouse neuroblastoma x rat glioma cell line NG108-15. This cell line can therefore be used as a model system for studying natively expressed adenine receptors. RT-PCR experiments indicated that the cells express the murine ortholog of the adenine receptor. We found that in NG108-15 cells the adenine receptor was coupled to inhibition of adenylate cyclase, as it had previously been shown for CHO cells recombinantly expressing the rat adenine receptor (Bender et al., 2002). The determined IC50 value (21 nM) for adenine in inhibiting forskolin-stimulated adenylate cyclase in membrane preparations of NG108-15 cells was roughly in the same concentration range as the previously determined IC50 and Ki values obtained in radioligand binding studies. An about 100-fold higher IC50 value for adenine of 2.5 e was determined in adenylate cyclase assays when intact NG108-15 cells were used. The reason for this discrepancy may be the fact that the intact cells have machinery for quickly and efficiently removing physiological receptor agonists, by degradation or cellular uptake, to limit their actions. However, little is known so far about the extracellular metabolism of adenine. Transporter proteins for adenine have been described, which may contribute to the removal of adenine from the extracellular compartment (De Koning and Diallinas, 2000; Redzic et al., 2001). PTX is a modifier of Gi/Go proteins, which ADP-ribosylates a cysteine residue of the subunit of Gi/Go proteins (Yamane and Fung, 1993). This results in the irreversible inactivation of the G protein and in a loss of its ability to transduce signals from activated receptors to adenylate cyclase (Kurose et al., 1986). Preincubation of the cells with PTX resulted in a decrease in the maximal stimulation of cAMP production induced by forskolin. PTX effects on control cAMP levels have previously been described, e.g., Hughes et al. (1984) observed an enhancement of the cAMP production by PTX in NG108-15 cells stimulated with prostaglandin E1. We could show that preincubation of the cells with pertussis toxin at concentrations as low as 10 ng/ml completely prevented the inhibitory effects of adenine on forskolin-stimulated cAMP production indicating that adenine exhibited its effect via a Gi protein. A direct effect of adenine on adenylate cyclase therefore seems to be very unlikely. In addition, adenine does not fulfill the structural requirements for high-affinity binding to the P-site of adenylate cyclase (Dessauer et al., 1999).
Selected adenine derivatives were all found to act as partial agonists at the adenine receptor (Table 5). Bender et al. (2002) had tried to identify a potential human ortholog of the rat adenine receptor by sequence comparison, but they were unsuccessful and concluded that adenine may not be a signaling molecule in humans. In radioligand binding studies, we have detected high specific binding of 10 nM [3H]adenine to membrane preparations of a human cell line, astrocytoma 1321N1 cells (Fig. 5). Therefore, we conclude that a human ortholog of the rat adenine receptor is very likely to exist.
In conclusion, we have shown for the first time that adenine receptors are expressed in high density in the rat brain. Functional coupling to inhibition of adenylate cyclase has been found in a native cell line (NG108-15) that most likely expresses the murine ortholog of the receptor. Structureactivity relationships of adenine derivatives have been investigated. Based on radioligand binding studies, we postulate that a human ortholog of the rodent adenine receptor exists. Thus, the adenine receptor adds a third family (P3) besides the P1 (adenosine) and the P2 (nucleotide) receptors to the class of purinergic receptors. Like P1 and P2 receptors, adenine receptors may be of interest as novel drug targets.
S.G. and C.E.M. were supported by the Deutsche Forschungsgemeinschaft (GRK-804/1).
doi:10.1124/mol.104.006601.
References
Bender E, Buist A, Jurzak M, Langlois X, Baggerman G, Verhasselt P, Ercken M, Guo H.-Q, Wintmolders C, Van den Wyngaert I, et al. (2002) Characterization of an orphan G protein-coupled receptor localized in the dorsal root ganglia reveals adenine as a signaling molecule. Proc Natl Acad Sci USA 99: 8573eC8578.
Camaioni E, Costanzi S, Vittori S, Volpini R, Klotz KN, and Cristalli G (1998) New substituted 9-alkylpurines as adenosine receptor ligands. Bioorg Med Chem 6: 523eC533.
Chen JF (2003) The adenosine A2A receptor as an attractive target for Parkinson's disease treatment. Drug News Perspect 16: 597eC604.
Cristalli G, Franchetti P, Grifantini M, Vittori S, Bordoni T, and Geroni C (1987) Improved synthesis and antitumor activity of 1-deazaadenosine. J Med Chem 30: 1686eC1688.
De Koning H and Diallinas G (2000) Nucleobase transporters. Mol Membr Biol 17: 75eC94.
Dessauer CW, Tesmer JJG, Sprang SR, and Gilman AG (1999) The interactions of adenylate cyclases with P-site inhibitors. Trends Pharmacol Sci 20: 205eC210.
Griessmeier KJ and Meler CE (2004) [3H]BQ-123 binding to native endothelin ETA receptors in human astrocytoma 1321N1 cells and screening of potential ligands. Pharmacology, in press.
Hughes AR, Martin MW, and Harden TK (1984) Pertussis toxin differentiates between two mechanisms of attenuation of cyclic AMP accumulation by muscarinic cholinergic receptors. Proc Natl Acad Sci USA 81: 5680eC5684.
Kaulich M, Qurishi R, and Meler CE (2003) Extracellular metabolism of nucleotides in neuroblastoma x glioma NG108-15 cells determined by capillary electrophoresis. Cell Mol Neurobiol 23: 349eC364.
Kelly MDW, Hill RG, Borsodi A, Toth G, and Kitchen I (1998) Weaning-induced development of D-opioid receptors in rat brain: differential effects of guanine nucleotides and sodium upon ligand-receptor recognition. Br J Pharmacol 125: 979eC986.
Klotz KN, Kachler S, Lambertucci C, Vittori S, Volpini R, and Cristalli G (2003) Ethyladenine derivatives as adenosine receptor antagonists: 2- and 8-substitution results in distinct selectivities. Naunyn-Schmiedeberg's Arch Pharmacol 367: 629eC634.
Kurose H, Katada T, Haga T, Haga K, Ichiyama A, and Ui M (1986) Functional interactions of purified muscarinic receptors with purified inhibitory guanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J Biol Chem 261: 6423eC6428.
Levy DE, Bao M, Cherbavaz DB, Tomlinson JE, Sedlock DM, Homcy CJ, and Scarborough RM (2003) Metal coordination-based inhibitors of adenylyl cyclase: novel potent P-site antagonists. J Med Chem 46: 2177eC2186.
Lloyd HG and Stone TW (1985) Cyclohexyladenosine binding in rat striatum. Brain Res 334: 385eC388.
Lorenzen A, Grossekatthofer B, Kerst B, Vogt H, Fein T, and Schwabe U (1996) Characterization of a novel adenosine binding protein sensitive to cyclo-AMP in rat brain cytosolic and particulate fractions. Biochem Pharmacol 52: 1375eC1385.
Lorenzen A, Nitsch-Kirsch M, Vogt H, and Schwabe U (1993) Characterization of membrane-bound and solubilized high-affinity binding sites for 5'-N-ethylcarboxamido[3H]adenosine from bovine cerebral cortex. J Neurochem 60: 745eC751.
Lowry OH, Rosebrough NJ, Farr AL, and Randall RJ (1951) Protein measurements with the Folin phenol reagent. J Biol Chem 193: 265eC275.
Mattia C and Coluzzi F (2003) Antidepressants in chronic neuropathic pain. Mini Rev Med Chem 3: 773eC784.
Meler CE (2000) Adenosine receptor ligands - recent developments. Part. 1. Agonists. Curr Med Chem 7: 1269eC1288.
Meler CE, Schumacher B, Brattstrm A, Abourashed EA, and Koetter U (2002) Interactions of valerian extracts and a fixed valerian-hop extract combination with adenosine receptors. Life Sci 71: 1939eC1949.
Patel J, Marangos PJ, Stivers J, and Goodwin FK (1982) Characterization of adenosine receptors in brain using N6-cyclohexyladenosine. Brain Res 237: 203eC214.
Redzic ZB, Segal MB, Gasic JM, Markovic ID, Vojvodic VP, Isakovic A, Thomas SA, and Rakic LM (2001) The characteristics of nucleobase transport and metabolism by the perfused sheep choroids plexus. Brain Res 888: 66eC74.
Stiles GL (1988) A1 adenosine receptor-G protein coupling in bovine brain membranes: effects of guanine nucleotides, salt and solubilization. J Neurochem 51: 1592eC1598.
Volpini R, Costanzi S, Vittori S, Cristalli G, and Klotz KN (2003) Medicinal chemistry and pharmacology of A2B adenosine receptors. Curr Top Med Chem 3: 427eC443.
Watanabe S, Yoshimi Y, and Ikekita M (2003) Neuroprotective effect of adenine on Purkinje cell survival in rat cerebellar primary cultures. J Neurosci Res 74: 754eC759.
Yamane HK and Fung BKK (1993) Covalent modifications of G-proteins. Annu Rev Pharmacol Toxicol 32: 201eC241.
Yeung SH and Green RD (1983) Agonist and antagonist affinities for inhibitory adenosine receptors are reciprocally affected by 5'-guanylylimidodiphosphate or N-ethylmaleimide. J Biol Chem 258: 2234eC2239.
Yoshimi Y, Watanabe S, Shinomiya T, Makino A, Toyoda M, and Ikekita M (2003) Nucleobase adenine as a trophic factor acting on Purkinje cells. Brain Res 991: 113eC122.
Yoshioka K, Matsuda A, and Nakata H (2001) Pharmacology of a unique adenosine binding site in rat brain using a selective ligand. Clin Exp Pharmacol Physiol 28: 278eC284., 百拇医药(Simone Gorzalka, Sauro Vi)
Department of Chemical Sciences, University of Camerino, Camerino, Italy (S.V., R.V., G.C.)
Abstract
An orphan G protein-coupled receptor from rat has recently been discovered to be activated by the nucleobase adenine (Proc Natl Acad Sci USA 99:8573eC8578, 2002). In the present study, we show for the first time that the adenine receptor is expressed in membrane preparations of native tissues and cell lines in high density, including rat brain cortex, rat brain striatum, and the mouse neuroblastoma x rat glioma hybrid cell line NG108-15. Saturation analysis with [3H]adenine at rat brain cortical membranes exhibited a single high-affinity binding site with a KD value of 27.2 nM, and a binding capacity of 2.28 pmol/mg of protein. Kinetic studies revealed unusual binding kinetics of [3H]adenine with rapid association and slow dissociation. A series of compounds were investigated in [3H]adenine competition experiments at rat brain cortex. Only minor substitution of the adenine structure was tolerated, the most potent compounds of the present series being 2-fluoroadenine (Ki value of 620 nM), 8-thioadenine (Ki value of 2.77 e), N6-methyladenine (Ki value of 3.64 e), and 7-methyladenine (Ki value of 4.13 e), all of which were partial agonists (40eC60% intrinsic activity). Adenine dose dependently inhibited forskolin-stimulated adenylate cyclase in membrane preparations of NG108-15 cells as well as in intact cells, showing that the receptor is functional in NG108-15 cells. Reverse transcriptase-polymerase chain reaction experiments followed by sequencing indicate that the NG108-15 cells express the murine ortholog of the adenine receptor. Moreover, preliminary radioligand binding studies with [3H]adenine at membranes of human astrocytoma 1321N1 cells suggest that a human ortholog of the rat adenine receptor exists.
Adenine was recently identified as the endogenous ligand of an orphan rat G protein-coupled receptor by a "reverse pharmacology" approach. mRNA localization studies revealed the highest expression in the small neurons of the dorsal root ganglia, suggesting a role in nociception (Bender et al., 2002). Potent neurotrophic effects of adenine have recently been described; in primary cultures of rat cerebellar Purkinje cells, the number of cells was greatly increased by the addition of adenine, although relatively high concentrations (1eC2 mM) were required (Yoshimi et al., 2003). Adenine was the most potent compound in a series of purine and pyrimidine derivatives that showed protective effects on Purkinje cells and prevented cell death, whereas other adenine-based purines, including adenosine, AMP, ADP, and ATP, did not promote Purkinje cell survival (Watanabe et al., 2003).
Radioligand binding studies had previously been performed at membrane preparations of CHO cells stably transfected with the rat adenine receptor using [3H]adenine as a radioligand (Bender et al., 2002). The recombinant rat adenine receptor expressed in CHO cells was found to be coupled to inhibition of adenylate cyclase via Gi protein (Bender et al., 2002). However, the postulated adenine receptor has not been demonstrated to be present in native tissues on the protein level, and therefore it has not yet been officially accepted as a novel receptor by the International Union of Pharmacology committee on receptor nomenclature. A mouse ortholog (mMrgA10), but no human ortholog, of the rat adenine receptor could be identified by sequence analysis and comparison (Bender et al., 2002).
The present study proves the existence of the adenine receptor in brain tissues and native cell lines and its functional coupling to inhibition of adenylate cyclase in NG108-15 cells. Furthermore, we have explored the structure-activity relationships of a large series of compounds structurally related to adenine.
Materials and Methods
Radioligand Binding Studies
[3H]Adenine Saturation Experiments. Saturation binding experiments were carried out at rat brain cortical and rat brain striatal membranes using [8-3H]adenine (27 Ci/mmol; Amersham Biosciences Europe GmbH, Freiburg, Germany) as a radioligand. Frozen rat brain membranes were obtained by Pel Freez (Rogers, AR) and thawed at 4°C. Cortex and striatum were dissected, and membrane fractions were prepared as described previously (Meler et al., 2002). Protein concentrations of 100 e/ml were used in the saturation binding experiments, determined by the method of Lowry (Lowry et al., 1951) using a Sigma Chemie protein assay kit. Incubations were carried out in a total volume of 200 e of 50 mM Tris-HCl buffer, pH 7.4, supplemented with 10 mM MgCl2 and 1 mM EGTA, at room temperature (23°C) for 60 min. Nonspecific binding was determined with 10 e adenine. In some experiments, the stable GTP analog guanosine-5'-(,-imido)triphosphate (GppNHp, 100 e) was present.
Incubation was terminated by rapid filtration through GF/B glass fiber filters, presoaked in 0.5% aqueous polyethylenimine solution for 60 min using a Brandel 48-channel cell harvester. Filters were rinsed three times with 2 ml each of ice-cold 50 mM Tris-HCl buffer, pH 7.4. The wet filters were transferred to mini-vials and scintillation cocktail was added. Drying of the GF/B glass fiber filters before the addition of scintillation cocktail did not influence the results. Radioactivity on the punched-out filters was measured after 9 h of preincubation with 2.5 ml of Ultima Gold scintillation cocktail (Canberra Industries, Dreieich, Germany).
[3H]Adenine Competition Assays and Kinetic Studies. The adenine binding assays were carried out at rat brain cortical membranes using 10 nM [3H]adenine as a radioligand. Protein amounts of 100 e/tube were used in the assays. Inhibition curves were determined using seven to nine different concentrations of the test compound, spanning 3 orders of magnitude. Three separate experiments were performed each in triplicate, unless otherwise noted. Nonspecific binding was determined with 10 e adenine and amounted to 10 to 20% of total binding. Incubations were carried out in a total volume of 0.5 ml of 50 mM Tris-HCl buffer, pH 7.4, supplemented with 10 mM MgCl2 and 1 mM EGTA, at room temperature (23°C) for 60 min. Further experimental procedures were as described above for saturation studies. Kinetic studies were performed using 15 nM [3H]adenine. Dissociation was initiated by the addition of 1 e adenine after 60 min of preincubation. Three separate experiments, each in triplicate, yielded similar results. Competition binding assays at membrane preparations of NG108-15 cells and astrocytoma 1321N1 cells were performed as described above for rat brain cortical membranes using 100 e of protein (NG108-15), or 150 e of protein (astrocytoma), respectively, per tube.
Cell Culture. NG108-15 cells were provided by Dr. Bres (Pharmacological Institute, University of Bonn, Bonn, Germany) and grown as described previously (Kaulich et al., 2003). For radioligand binding studies, the culture medium was removed and the cells were kept frozen ateC80°C until the preparation of the membranes. Cells were thawed and detached in ice-cold 50 mM Tris-HCl buffer, pH 7.4, with a cell scraper. The cell suspension was homogenized on ice and spun down for 10 min at 1000g at 4°C. The supernatant was centrifuged at 20,000g for 30 min at 4°C, and the resulting pellet was resuspended in 50 mM Tris-HCl buffer, pH 7.4. The last centrifugation step was repeated once. The protein concentration was determined by the method of Lowry (Lowry et al., 1951). Membranes were kept frozen ateC80°C until use. For cAMP assays, membranes were used immediately. Astrocytoma 1321N1 cells were grown adherently and maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 100 U/ml penicillin G, 100 e/ml streptomycin, and 1% GlutaMAX at 37°C, 5% CO2. Cells were grown to confluence. For binding assays, culture medium was removed and cells were kept frozen ateC80°C until membrane preparation following the procedure described by Griessmeier and Meler (2004).
Molecular Biological Experiments. Poly A+ mRNA was isolated from NG108-15 cells cultured as described above using the Oligotex Direct mRNA mini kit (QIAGEN GmbH, Hilden, Germany). Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed using gene-specific primers for the rat adenine receptor (GenBank accession no. NM_145787 [GenBank] ; sense primer, CACCATGGGGGAAAGCTTCACCGGTACAG; antisense primer, CGGCTCCACCTTGCTGCTTGACATC) as well as the putative murine ortholog MrgA10 (GenBank accession no. XM_195647 [GenBank] ; sense, CACCATGGGGGAAAGCAGCACCGGTGCAG; antisense, TGGCTCTGCTTTGCTTTTTGACATCTCC) and the Superscript One Step RT-PCR kit (Invitrogen, Karlsruhe, Germany). Annealing temperature for the PCR reaction was 56.0°C (35 cycles). For control experiments, the enzyme reverse transcriptase was omitted (instead of the RT-PCR enzyme mix, Taq Platinum Polymerase was used; Invitrogen). PCR products were analyzed by ethidium bromide staining after agarose (0.9%) gel electrophoresis. After amplifying the product of one RT-PCR reaction using Platinum Pfx polymerase (Invitrogen), the sequence was cloned into the expression vector pcDNA3.1D/V5-His-TOPO (Invitrogen). The sequence was then identified by cycle sequencing (SequiTherm Exel II DNA sequencing kit; Epicenter Technologies, Madison WI) using a LICOR Gene READIR 4200 sequencer (MWG-Biotech, Ebersberg, Germany).
[35S]GTPS Binding Assay. The [35S]GTPS binding assay was carried out as described previously (Bender et al., 2002) with minor modifications. The incubation buffer consisted of 20 mM HEPES, pH 7.4, containing 100 mM NaCl, 10 e GDP, 10 mM MgCl2, and 0.5% bovine serum albumin. After a preincubation period of 30 min of rat cortical membranes (10 e/tube) with agonist (adenine or N6-cyclopentyladenosine) in the presence of 2 IU/ml adenosine deaminase, [35S]GTPS was added in a final concentration of 0.5 nM (final volume 200 e). After a further incubation period of 60 min, bound and free [35S]GTPS were separated by filtration through a GF/B glass fiber filter, presoaked in 50 mM Tris-HCl buffer, pH 7.4, containing 10 mM MgCl2.
Functional Studies
Cyclic AMP Accumulation. Using membrane preparations. Culture flasks with almost confluently grown NG108-15 cells were taken, and cells were detached with a cell scraper. A membrane preparation of the cells was obtained according to the method described above. The membrane fraction was resuspended in 50 mM Tris-HCl buffer, pH 7.4, supplemented with 4 mM EDTA, 10 mM MgCl2, and 5 e GTP. Stock solutions of compounds in dimethyl sulfoxide were added to 50 mM Tris-HCl buffer, pH 7.4, containing 4 mM EDTA, 10 mM MgCl2, 5 e GTP, 10 e forskolin, and 1 mM ATP, and the mixture was incubated for 20 min at 37°C with 90 eof the membrane suspension. Incubation was terminated by rapidly heating to 98°C for 2 min. After cooling down to room temperature, the tubes were centrifuged at 2000g for 15 min at 4°C, using an Allegra 21 R centrifuge (Beckman Coulter GmbH, Krefeld, Germany). The supernatant (50 e) was assayed in a charcoal adsorption-assay using a commercially available kit (Amersham Biosciences Inc.) to determine the cAMP concentration. Each experiment was carried out in duplicates at least three times.
Using whole cells. Cyclic AMP accumulation was determined in intact NG108-15 cells plated on a 48-well culture plate. After cells were almost grown to confluence the cell culture medium was removed, cells were washed twice with Krebs-Ringer-HEPES buffer, pH 7.4, containing 130 mM NaCl, 4.7 mM KCl, 4.0 mM NaHCO3, 1.2 mM KH2PO4, 1.2 mM MgSO4, 1.8 mM CaCl2·2H2O, 11.5 mM glucose, 10 mM HEPES, and 0.1% bovine serum albumin, and were then preincubated with adenosine 1 IU/ml deaminase, and the phosphodiesterase inhibitors Ro20-1724 (500 e) and 3-isobutyl-1-methylxanthine (500 e) in Krebs-Ringer-HEPES buffer, pH 7.4, for 10 min at 37°C. Forskolin (final concentration 10 e) and a range of different concentrations of adenine were added. The incubation at 37°C was stopped after 10 min by removal of the reaction buffer. Cells were lysed with 0.1 M HCl for 30 min at room temperature. After neutralization with 1 M NaOH solution and further dilution, the amount of cAMP was determined as described above. Each experiment was carried out in duplicates at least three times. In some experiments, cells were preincubated overnight in medium in the presence of pertussis toxin (PTX, 10, 100, or 1000 ng/ml, respectively). All three concentrations gave virtually identical results.
Data Analysis. Data were analyzed using Prism 3.0 (GraphPad Software Inc., San Diego, CA). IC50 values were determined by fitting data to a sigmoidal curve with variable slope. The Cheng-Prusoff equation and the KD value of 27.2 nM for [3H]adenine was used to calculate Ki values from the obtained IC50 values.
Chemicals. 1-Methyladenine, 7-methyladenine, and 9-methyladenine were purchased from Acros Organics (Geel, Belgium). 2-Chloroadenine, 7-ethyladenine, 9-dihydroxypropyladenine, 9-ethyladenine, 8-bromoadenine, 1-deazaadenosine, 3-deazaadenosine, 2',3'-dideoxyadenosine, 2-[3-(1-phenylprop-2-yn-1-ol)-9-ethyladenine, 2-phenylethynyl-9-ethyladenine, 2-(hex-1-yn-1-yl)-9-ethyladenine, 8-phenyl-9-ethyladenine and 8-bromo-9-propyladenine were synthesized according to or in analogy to described procedures (Cristalli et al., 1987; Camaioni et al., 1998; Klotz et al., 2003). All other reagents were obtained from Sigma Chemie (Deisenhofen, Germany).
Results
Saturation binding experiments at rat brain cortical membranes performed with [3H]adenine revealed a single high-affinity binding site with a KD value of 27.2 ± 0.9 nM (n = 5, Fig. 1A). A binding capacity (Bmax) of 2.28 ± 0.01 pmol/mg of protein was determined. The presence of a stable analog of GTP, 100 e GppNHp, led to a small increase in the KD value [from 27.2 ± 0.9 (n = 5) to 32.6 ± 2.7 nM (n = 3)], which seemed to be reproducible, but did not quite reach statistical significance (P = 0.058, two-tailed t test). The Bmax value was virtually unaltered [2.28 ± 0.01 (n = 5), without GppNHp versus 2.58 ± 0.37 pmol/mg of protein (n = 3), with GppNHp, P = 0.47]. Adenine (0.01 nMeC500 e) did not have any effect on [35S]GTPS binding to rat brain cortical membranes. In membranes prepared from rat brain striatum, [3H]adenine saturation experiments exhibited a single binding site with a KD value of 72.4 ± 5.0 nM and a Bmax value of 8.23 + 0.8 pmol/mg of protein (n = 3, Fig. 1B). For the saturation experiments, a two-site model did not give an improved fit over a one-site model. Kinetic studies (Fig. 2) showed that [3H]adenine bound very rapidly to membranes from rat brain cortex. Steady state seemed to be reached as early as the first time point examined (30 s) and was sustained for at least 200 min (not shown). Binding was slowly reversed by the addition of adenine in a final concentration of 1 e, with 60% of dissociation being evident after 60 min. A kinetic KD value could not be determined. The law of mass action did not apply to the system.
A concentration of 10 nM [3H]adenine was suitable for competition binding assays at rat brain cortical membranes. At this concentration, specific binding amounted to 80 to 90% of total binding. The addition of dimethyl sulfoxide in concentrations up to 10% had no effect on [3H]adenine binding (data not shown). NaCl (100 mM) did not alter [3H]adenine binding significantly (data not shown).
Binding affinities for a series of compounds were determined in competition experiments at rat brain cortical membranes. Adenine itself showed a Ki value of 29.9 ± 3.4 nM (Fig. 3). The investigated compounds included adenine derivatives with substitution in the 1-, 2-, 3-, N6-, 7-, 8-, or 9-position (Tables 1, 2, 3; Fig. 3), xanthine derivatives (Table 2), and a selection of drugs with analgesic properties (salicylic acid, acetylsalicylic acid, acetaminophen, indomethacin, and propyphenazone) and with effects against neuropathic pain (the antidepressants amitriptyline, imipramine, and doxepin). Only adenine derivatives with small structural modifications showed appreciable affinity for the rat brain adenine receptor. All compounds investigated were considerably (at least 20-fold) less potent than adenine. The most potent derivatives were 2-fluoroadenine (3, Ki value of 620 nM), 8-thioadenine (14, Ki value of 2.77 e), N6-methyladenine (6, Ki value of 3.64 e), and 7-methyladenine (11, Ki value of 4.13 e) (Fig. 4). The following rank order of potency could be established: adenine >> 2-fluoroadenine > N6-methyladenine = 7-methyladenine > 9-methyladenine 1-methyladenine > 3-methyladenine/N6-dimethyladenine (Table 1). Adenosine (18) exhibited binding affinity in the micromolar range (Ki value of 19.4 ± 6.3 e) and its 2',3'-dideoxy analog (19) was equally potent (Ki value of 18.2 ± 3.4 e) (Table 2). 1-Deazaadenosine (20), 3-deazaadenosine (21), and the adenosine derivative NECA (22) (Table 2) showed no significant affinity for the adenine receptor. The addition of adenosine deaminase, which is used in radioligand binding studies at adenosine receptors to remove endogenous adenosine by converting it to inosine, did not have any effect on the [3H]adenine binding studies (data not shown). A selection of adenine derivatives, with combinations of substituents in the 2-, 8-, and 9-position that had previously been shown to be potent adenosine A2A receptor antagonists, were found to be inactive at the rat brain adenine receptor (Table 3).
[3H]Adenine binding studies at membrane preparations of two permanent cell lines, the mouse neuroblastoma x rat glioma cells NG108-15 and the human astrocytoma 1321N1 cell line, revealed that both contained specific binding sites for adenine (Fig. 4). Competition assays at NG108-15 cell membranes using 10 nM [3H]adenine yielded an IC50 value of 154 ± 56 nM for unlabeled adenine and similar affinities and the same rank order of potency for a selection of adenine derivatives as seen in binding studies at rat brain cortex (Table 4).
Freshly prepared membranes from NG108-15 cells and intact NG108-15 cells were used for functional studies (Fig. 5). At membranes adenine dose dependently inhibited forskolin-stimulated adenylate cyclase with an IC50 value of 21 nM (95% confidence interval 7eC67 nM). Forskolin stimulated the basal cAMP production of 6 pmol/mg of protein/min about 4-fold to 25 pmol/mg of protein/min. Maximal inhibition of 50% was observed at 3 e adenine. The addition of adenosine deaminase had no effect on cAMP accumulation (data not shown). Selected adenine derivatives were investigated in functional assays determining the inhibition of forskolin-stimulated cAMP accumulation at a single, high concentration. All of the compounds were partial agonists at the rat adenine receptor with somewhat weaker maximal inhibition as the full agonist adenine (Table 5). Functional studies were additionally performed using whole NG108-15 cells, where adenine dose dependently inhibited forskolin-stimulated adenylate cyclase activity with an IC50 value of 2.54 ± 0.99 e. Maximal inhibition of 70% was observed at 30 e adenine. Preincubation of the cells with pertussis toxin (10, 100, or 1000 ng/ml, respectively) led to a decrease of the maximal stimulation by forskolin of the cAMP production by 50% (data not shown). Pertussis toxin completely prevented adenine-induced inhibition of the forskolin-stimulated cAMP production (Fig. 6).
RT-PCR with primers specific for the coding sequence of the putative mouse ortholog but not with primers for the rat adenine receptor revealed products of the expected length from mRNA preparations prepared from cultured NG108-15 cells (Fig. 7; one of three independent runs). Control reactions without the enzyme reverse transcriptase showed no products confirming that mRNA but not genomic DNA acted as template for the PCR reaction. Cycle sequencing of one RT-PCR reaction product confirmed that it was a murine and not a rat sequence.
Discussion
Purinergic membrane receptors for the nucleoside adenosine (P1 receptors) and the adenine nucleotides ATP and ADP (P2 receptors) have been known to exist for more than 20 years. The aglycon adenine has been found to activate an orphan rat G protein-coupled receptor belonging to the MAS-related gene (Mrg) receptor family with high affinity (Ki value of 18 nM) (Bender et al., 2002). mRNA expression of the rat adenine receptor was detected in many tissues with highest levels found in the small neurons of the dorsal root ganglia (Bender et al., 2002). We have now for the first time investigated native adenine receptors on the protein level with radioligand binding and functional studies.
In saturation studies at rat brain cortical membranes, [3H]adenine exhibited a KD value of 27.2 nM. This is well in accordance with literature data for the recombinant rat adenine receptor expressed in CHO cells (KD value of 24 nM) obtained at virtually identical conditions (Bender et al., 2002), demonstrating that [3H]adenine actually labels the adenine receptor in rat brain cortex. At rat brain striatal membranes, the determined KD value was somewhat higher (72.4 nM) but still in the same concentration range. In both preparations, [3H]adenine labeled only a single binding site with high affinity, indicating that only a single adenine receptor seems to be present.
The adenine receptor density in rat cortical membranes was determined to be quite high, with a Bmax value of 2.28 pmol/mg of protein; it was even higher in striatal membranes (8.23 pmol/mg of protein). In comparison, the density of the highly expressed adenosine A1 adenosine receptor in the rat brain is severalfold lower (Patel et al., 1982, Lloyd and Stone, 1985). The mRNA profile determined by Bender et al. (2002) had indicated low expression in striatum and somewhat higher expression in brain cortex. Our results show that mRNA levels do not necessarily correlate with protein expression levels of adenine receptors. The detected high densities of adenine receptors in rat brain cortex and striatum point to an important physiological role of these receptors.
Guanine nucleotides have been shown to decrease the affinity of inhibitory agonists for their receptor (Yeung and Green, 1983; Stiles, 1988). This has been explained by a shift from a high-affinity to a low-affinity receptor state in the presence of GTP. Such an effect is sometimes not very pronounced or even undetectable (Kelly et al., 1998). Saturation studies in the presence of the stable GTP analog GppNHp (100 e) at rat brain cortical membranes revealed that the binding of [3H]adenine was somewhat influenced by the guanine nucleotide, which led to a moderate increase in the KD value. Although GTP and its analogs can lead to decreased binding of an agonist to a Gi-coupled receptor, the reverse allosteric effect has also been observed, namely, a stimulation of [35S]GTPS binding to the Gi protein by a receptor agonist. [35S]GTPS binding studies work well in artificial systems with high receptor expression levels. In fact, Bender et al. (2002) had shown that adenine increased [35S]GTPS binding in membranes prepared from CHO cells recombinantly expressing the rat adenine receptor. In native rat cortical membranes, we could not show adenine-induced enhancement of [35S]GTPS binding. Showing such an effect in native tissues is sometimes difficult or even impossible. N6-Cyclopentyladenosine, an agonist at the Gi-coupled adenosine A1 receptor, also did not show any measurable increase in [35S]GTPS binding at native membranes in our hands, whereas it strongly increased [35S]GTPS in recombinant CHO cells expressing the human adenosine A1 receptor in high density (data not shown).
Kinetic studies demonstrated that the binding of [3H]adenine to rat brain cortical membranes was reversible. However, after a very fast association, dissociation initiated by the addition of a high concentration of unlabeled adenine was slow in comparison (Fig. 3). A kinetic KD value could not be determined. An explanation for the observed unusual kinetics could be that the adenine receptor undergoes a conformational change after binding of [3H]adenine; an initial low-affinity conformation may be converted to a high-affinity conformation for adenine upon binding.
The binding site of the adenine receptor in rat brain cortex was further characterized by competition assays, and structure-activity relationships were determined by investigating a series of differently substituted adenine derivatives. Bender et al. (2002) had evaluated only a very limited selection of compounds at the recombinant rat adenine receptor, namely, 1-methyladenine, 6-benzyladenine, adenosine, AMP, ADP, ATP, hypoxanthine, and guanine, none of which was remarkably potent. Our results for this collection of compounds at the rat brain cortex receptor were mainly consonant with the results obtained by Bender et al. (2002) at the recombinant adenine receptors. The only major discrepancy was that adenosine seemed to exhibit measurable affinity for the adenine receptor in rat brain cortical membranes (Ki value of 19.4 e), whereas it was found inactive at the recombinantly expressed receptor in CHO cell membranes. However, the determined Ki value at rat brain cortex was almost 3 orders of magnitude higher than that for adenine. It cannot be excluded at present that adenosine is degraded by a purine nucleoside phosphorylase during the 60-min incubation with rat brain cortical membranes and the detected binding is the result of formed adenine. The same may also be true for the adenine nucleotides AMP (23), ADP (24), and ATP (25), which show very weak affinity (39eC51% inhibition of radioligand binding at 100 e): the observed inhibition may be caused by adenine formed through enzymatic degradation, although EGTA was present in the assays to inhibit nucleotidases.
In the competition studies, adenine itself exhibited a Ki value of 29.9 nM, which is in good agreement with literature data for the recombinant receptor (18 nM; Bender et al., 2002) and with the determined KD value (27.2 nM). Even only small modifications at the adenine core structure led to a major reduction or loss of affinity. In a series of mono-substituted adenine derivatives (Table 1), substitution with a fluorine atom in the 2-position was best tolerated, resulting in 2-fluoroadenine (3); however, it was 20-fold less potent than the unsubstituted adenine (620 versus 30 nM). A fluorine atom is sterically similar to a hydrogen atom but has strong electronic effects (negative inductive effect) that must be responsible for the reduced affinity of 3 in comparison with adenine. Replacement of the fluorine atom by the larger halogen atom chlorine (compound 4) further reduced affinity by 60-fold (Ki value of 37.9 e), showing that the receptor does not have any bulk tolerance in that region. Substitution of the exocyclic amino group was investigated. A methyl group was best tolerated (6, Ki value of 3.64 e), whereas the larger benzyl (7) and benzoyl (8) residues resulted in very weak binding. It is interesting that the more polar benzoyl substitution was better tolerated (Ki value of 46.6 e) than a benzyl residue (Ki value >100 e). The N6-hydroxypentenyl-substituted adenine derivative trans-zeatin, a cytokinin (phytohormone), was somewhat more potent than 8 (Ki value of 22.6 e); it seems that polar substituents are better accepted by the receptor than unpolar residues. Dimethylation of the amino group abolished affinity, indicating that a hydrogen bond donor (N-H) in that position was probably required for binding to the receptor, similarly as in adenosine receptors (Meler, 2000). Small polar substituents in the 8-position, such as thiol (14, Ki value of 2.77 e) and bromine (13, Ki value of 17.3 e) were tolerated by the receptor leading to moderately potent compounds. Methylation at the ring nitrogen atoms of adenine was explored. Methylated adenine derivatives exhibited quite different binding affinities depending on the position of the methyl group. The rank order of potency was 7-methyladenine (11, Ki value of 4.13 e) > 9-methyladenine (15, Ki value 17.5 e) 1-methyladenine (2, Ki value of 29.3 e) >> 3-methyladenine (5, Ki value >>100 e). Replacement of the methyl group in 7- or 9-methyladenine by an ethyl residue resulted in a large drop in affinity (compare 11/12 and 15/16), indicating that the receptor did not have bulk tolerance in that positions. However, a more hydrophilic substituent in the 9-position, a dihydroxypropyl residue (compound 17), was somewhat better tolerated than an ethyl residue.
A series of 9-alkyl-substituted adenine derivatives that bear an additional substituent either in the 2- or 8-position were recently described as relatively potent adenosine A2A receptor antagonists (Klotz et al., 2003). A2A antagonists acting at receptors in the striatum (caudate putamen) are currently developed as novel therapeutics for Parkinson's disease (Chen, 2003). Some of the disubstituted adenine derivatives, including compound 31, had been evaluated in animal models of Parkinson's disease. The compounds had induced contralateral turning in 6-hydroxydopamine-lesioned rats after systemic application in a dose of 5 mg/kg. In contrast, the standard A2A antagonist SCH58261had required priming by dopamine to induce the same effect (M. Morelli, oral presentation at the Satellite Symposium on Purinergic Receptors, Camerino, Italy, 2003). Because we have found that adenine receptors are highly expressed in the rat striatum, we investigated whether the adenine derivatives described as adenosine A2A receptor antagonists were selective for the adenosine receptors or whether they showed additional binding to the adenine receptor. In fact, the disubstituted adenine derivatives did not exhibit any affinity for adenine receptors in concentrations up to 100 e (Table 3).
Adenosine (18) exhibited a more than 600-fold lower affinity in comparison with the natural ligand adenine. The potent, nonselective adenosine receptor agonist NECA (22), which has been found to also bind with high affinity to nonadenosine receptor binding sites (Lorenzen et al., 1993, 1996; Yoshioka et al., 2001), showed no significant binding affinity for the adenine receptor. As pointed out above, adenosine might be partially degraded to adenine during the assay, whereas NECA is enzymatically more stable. 2',3'-Dideoxyadenosine (19), which lacks the 2'- and 3'-hydroxyl groups at the ribose moiety, exhibited a similar Ki value as adenosine. 1-Deazaadenosine (20) and 3-deazaadenosine (21) showed only low affinity for the adenine receptor, but it seems that the N-atom in position 1 is more important for binding than the N-atom in position 3. The adenine nucleotides AMP (23), ADP (24), and ATP (25) seemed to exhibit some affinity to the adenine receptor in similar, high concentrations (39eC51% inhibition of radioligand binding at 100 e), but this might be caused by degradation of the nucleotides to adenosine or adenine by enzymes present in the membrane preparations. The obtained structure-activity relationships are not consistent with those obtained for the ATP binding site of adenylate cyclases, the so-called P-site (Dessauer et al., 1999; Levy et al., 2003).
All other compounds that we screened and that were not derived from adenine, including the xanthine derivatives caffeine (26), theophylline (27), 8-cyclopentyltheophylline (28), and 1,3-dipropyl-8-cyclopentylxanthine (29), all of which are adenosine receptor antagonists, exhibited no significant binding affinity for the adenine receptor at a high concentration of 100 e. A set of analgesic and antidepressant compounds were investigated because of the finding that the adenine receptor is highly expressed in the small neurons of the dorsal root ganglia (Bender et al., 2002), suggesting a role in nociception. Neither the analgesics nor the tricyclic antidepressants, which have been shown to be effective in neuropathic pain (Mattia and Coluzzi, 2003), exhibited significant binding affinity for the adenine receptor.
Specific binding sites for [3H]adenine with similar properties as in brain cortex were also detected in the permanent mouse neuroblastoma x rat glioma cell line NG108-15. This cell line can therefore be used as a model system for studying natively expressed adenine receptors. RT-PCR experiments indicated that the cells express the murine ortholog of the adenine receptor. We found that in NG108-15 cells the adenine receptor was coupled to inhibition of adenylate cyclase, as it had previously been shown for CHO cells recombinantly expressing the rat adenine receptor (Bender et al., 2002). The determined IC50 value (21 nM) for adenine in inhibiting forskolin-stimulated adenylate cyclase in membrane preparations of NG108-15 cells was roughly in the same concentration range as the previously determined IC50 and Ki values obtained in radioligand binding studies. An about 100-fold higher IC50 value for adenine of 2.5 e was determined in adenylate cyclase assays when intact NG108-15 cells were used. The reason for this discrepancy may be the fact that the intact cells have machinery for quickly and efficiently removing physiological receptor agonists, by degradation or cellular uptake, to limit their actions. However, little is known so far about the extracellular metabolism of adenine. Transporter proteins for adenine have been described, which may contribute to the removal of adenine from the extracellular compartment (De Koning and Diallinas, 2000; Redzic et al., 2001). PTX is a modifier of Gi/Go proteins, which ADP-ribosylates a cysteine residue of the subunit of Gi/Go proteins (Yamane and Fung, 1993). This results in the irreversible inactivation of the G protein and in a loss of its ability to transduce signals from activated receptors to adenylate cyclase (Kurose et al., 1986). Preincubation of the cells with PTX resulted in a decrease in the maximal stimulation of cAMP production induced by forskolin. PTX effects on control cAMP levels have previously been described, e.g., Hughes et al. (1984) observed an enhancement of the cAMP production by PTX in NG108-15 cells stimulated with prostaglandin E1. We could show that preincubation of the cells with pertussis toxin at concentrations as low as 10 ng/ml completely prevented the inhibitory effects of adenine on forskolin-stimulated cAMP production indicating that adenine exhibited its effect via a Gi protein. A direct effect of adenine on adenylate cyclase therefore seems to be very unlikely. In addition, adenine does not fulfill the structural requirements for high-affinity binding to the P-site of adenylate cyclase (Dessauer et al., 1999).
Selected adenine derivatives were all found to act as partial agonists at the adenine receptor (Table 5). Bender et al. (2002) had tried to identify a potential human ortholog of the rat adenine receptor by sequence comparison, but they were unsuccessful and concluded that adenine may not be a signaling molecule in humans. In radioligand binding studies, we have detected high specific binding of 10 nM [3H]adenine to membrane preparations of a human cell line, astrocytoma 1321N1 cells (Fig. 5). Therefore, we conclude that a human ortholog of the rat adenine receptor is very likely to exist.
In conclusion, we have shown for the first time that adenine receptors are expressed in high density in the rat brain. Functional coupling to inhibition of adenylate cyclase has been found in a native cell line (NG108-15) that most likely expresses the murine ortholog of the receptor. Structureactivity relationships of adenine derivatives have been investigated. Based on radioligand binding studies, we postulate that a human ortholog of the rodent adenine receptor exists. Thus, the adenine receptor adds a third family (P3) besides the P1 (adenosine) and the P2 (nucleotide) receptors to the class of purinergic receptors. Like P1 and P2 receptors, adenine receptors may be of interest as novel drug targets.
S.G. and C.E.M. were supported by the Deutsche Forschungsgemeinschaft (GRK-804/1).
doi:10.1124/mol.104.006601.
References
Bender E, Buist A, Jurzak M, Langlois X, Baggerman G, Verhasselt P, Ercken M, Guo H.-Q, Wintmolders C, Van den Wyngaert I, et al. (2002) Characterization of an orphan G protein-coupled receptor localized in the dorsal root ganglia reveals adenine as a signaling molecule. Proc Natl Acad Sci USA 99: 8573eC8578.
Camaioni E, Costanzi S, Vittori S, Volpini R, Klotz KN, and Cristalli G (1998) New substituted 9-alkylpurines as adenosine receptor ligands. Bioorg Med Chem 6: 523eC533.
Chen JF (2003) The adenosine A2A receptor as an attractive target for Parkinson's disease treatment. Drug News Perspect 16: 597eC604.
Cristalli G, Franchetti P, Grifantini M, Vittori S, Bordoni T, and Geroni C (1987) Improved synthesis and antitumor activity of 1-deazaadenosine. J Med Chem 30: 1686eC1688.
De Koning H and Diallinas G (2000) Nucleobase transporters. Mol Membr Biol 17: 75eC94.
Dessauer CW, Tesmer JJG, Sprang SR, and Gilman AG (1999) The interactions of adenylate cyclases with P-site inhibitors. Trends Pharmacol Sci 20: 205eC210.
Griessmeier KJ and Meler CE (2004) [3H]BQ-123 binding to native endothelin ETA receptors in human astrocytoma 1321N1 cells and screening of potential ligands. Pharmacology, in press.
Hughes AR, Martin MW, and Harden TK (1984) Pertussis toxin differentiates between two mechanisms of attenuation of cyclic AMP accumulation by muscarinic cholinergic receptors. Proc Natl Acad Sci USA 81: 5680eC5684.
Kaulich M, Qurishi R, and Meler CE (2003) Extracellular metabolism of nucleotides in neuroblastoma x glioma NG108-15 cells determined by capillary electrophoresis. Cell Mol Neurobiol 23: 349eC364.
Kelly MDW, Hill RG, Borsodi A, Toth G, and Kitchen I (1998) Weaning-induced development of D-opioid receptors in rat brain: differential effects of guanine nucleotides and sodium upon ligand-receptor recognition. Br J Pharmacol 125: 979eC986.
Klotz KN, Kachler S, Lambertucci C, Vittori S, Volpini R, and Cristalli G (2003) Ethyladenine derivatives as adenosine receptor antagonists: 2- and 8-substitution results in distinct selectivities. Naunyn-Schmiedeberg's Arch Pharmacol 367: 629eC634.
Kurose H, Katada T, Haga T, Haga K, Ichiyama A, and Ui M (1986) Functional interactions of purified muscarinic receptors with purified inhibitory guanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J Biol Chem 261: 6423eC6428.
Levy DE, Bao M, Cherbavaz DB, Tomlinson JE, Sedlock DM, Homcy CJ, and Scarborough RM (2003) Metal coordination-based inhibitors of adenylyl cyclase: novel potent P-site antagonists. J Med Chem 46: 2177eC2186.
Lloyd HG and Stone TW (1985) Cyclohexyladenosine binding in rat striatum. Brain Res 334: 385eC388.
Lorenzen A, Grossekatthofer B, Kerst B, Vogt H, Fein T, and Schwabe U (1996) Characterization of a novel adenosine binding protein sensitive to cyclo-AMP in rat brain cytosolic and particulate fractions. Biochem Pharmacol 52: 1375eC1385.
Lorenzen A, Nitsch-Kirsch M, Vogt H, and Schwabe U (1993) Characterization of membrane-bound and solubilized high-affinity binding sites for 5'-N-ethylcarboxamido[3H]adenosine from bovine cerebral cortex. J Neurochem 60: 745eC751.
Lowry OH, Rosebrough NJ, Farr AL, and Randall RJ (1951) Protein measurements with the Folin phenol reagent. J Biol Chem 193: 265eC275.
Mattia C and Coluzzi F (2003) Antidepressants in chronic neuropathic pain. Mini Rev Med Chem 3: 773eC784.
Meler CE (2000) Adenosine receptor ligands - recent developments. Part. 1. Agonists. Curr Med Chem 7: 1269eC1288.
Meler CE, Schumacher B, Brattstrm A, Abourashed EA, and Koetter U (2002) Interactions of valerian extracts and a fixed valerian-hop extract combination with adenosine receptors. Life Sci 71: 1939eC1949.
Patel J, Marangos PJ, Stivers J, and Goodwin FK (1982) Characterization of adenosine receptors in brain using N6-cyclohexyladenosine. Brain Res 237: 203eC214.
Redzic ZB, Segal MB, Gasic JM, Markovic ID, Vojvodic VP, Isakovic A, Thomas SA, and Rakic LM (2001) The characteristics of nucleobase transport and metabolism by the perfused sheep choroids plexus. Brain Res 888: 66eC74.
Stiles GL (1988) A1 adenosine receptor-G protein coupling in bovine brain membranes: effects of guanine nucleotides, salt and solubilization. J Neurochem 51: 1592eC1598.
Volpini R, Costanzi S, Vittori S, Cristalli G, and Klotz KN (2003) Medicinal chemistry and pharmacology of A2B adenosine receptors. Curr Top Med Chem 3: 427eC443.
Watanabe S, Yoshimi Y, and Ikekita M (2003) Neuroprotective effect of adenine on Purkinje cell survival in rat cerebellar primary cultures. J Neurosci Res 74: 754eC759.
Yamane HK and Fung BKK (1993) Covalent modifications of G-proteins. Annu Rev Pharmacol Toxicol 32: 201eC241.
Yeung SH and Green RD (1983) Agonist and antagonist affinities for inhibitory adenosine receptors are reciprocally affected by 5'-guanylylimidodiphosphate or N-ethylmaleimide. J Biol Chem 258: 2234eC2239.
Yoshimi Y, Watanabe S, Shinomiya T, Makino A, Toyoda M, and Ikekita M (2003) Nucleobase adenine as a trophic factor acting on Purkinje cells. Brain Res 991: 113eC122.
Yoshioka K, Matsuda A, and Nakata H (2001) Pharmacology of a unique adenosine binding site in rat brain using a selective ligand. Clin Exp Pharmacol Physiol 28: 278eC284., 百拇医药(Simone Gorzalka, Sauro Vi)