P_2Y purinoceptor subtypes recruit different Mek activators in astrocytes

Abstract

1. Extracellular ATP can function as a glial trophic factor as well as a neuronal transmitter. In astrocytes, mitogenic signalling by ATP is mediated by metabotropic P_2Y receptors that are linked to the extracellular signal regulated protein kinase (Erk) cascade, but the types of P_2Y receptors expressed in astrocytes have not been defined and it is not known whether all P_2Y receptor subtypes are coupled to Erk by identical or distinct signalling pathways.

2. We found that the P_2Y receptor agonists ATP, ADP, UTP and 2-methylthioATP (2MeSATP) activated Erk and its upstream activator MAP/Erk kinase (Mek). cRaf-1, the first kinase in the Erk cascade, was activated by 2MeSATP, ADP and UTP but, surprisingly, cRaf-1 was not stimulated by ATP. Furthermore, ATP did not activate B-Raf, the major isoform of Raf in the brain, nor other Mek activators such as Mek kinase 1 (MekK1) and MekK2/3.

3. Reverse transcriptase-polymerase chain reaction (RT–PCR) studies using primer pairs for cloned rat P_2Y receptors revealed that rat cortical astrocytes express P_2Y1, a receptor subtype stimulated by ATP and ADP and their 2MeS analogues, as well as P_2Y2 and P_2Y4, subtypes in rats for which ATP and UTP are equipotent. Transcripts for P_2Y6, a pyrimidine-preferring receptor, were not detected.

4. ATP did not increase cyclic AMP levels, suggesting that P_2Y11, an ATP-preferringz receptor, is not expressed or is not linked to adenylyl cyclase in rat cortical astrocytes.

5. These signal transduction and RT–PCR experiments reveal differences in the activation of cRaf-1 by P_2Y receptor agonists that are inconsistent with properties of the P_2Y1, P_2Y2 and P_2Y4 receptors shown to be expressed in astrocytes, i.e. ATP≠UTP; ATP≠2MeSATP, ADP. This suggests that the properties of the native P_2Y receptors coupled to the Erk cascade differ from the recombinant P_2Y receptors or that astrocytes express novel purine-preferring and pyrimidine-preferring receptors coupled to the ERK cascade.

Introduction

Numerous studies have indicated that ATP is a neurotransmitter which can affect both neurones and glial cells through a great variety of purinoceptors (reviewed by Burnstock, 1997; North & Barnard, 1997). ATP mediates several effects in the CNS, such as excitatory neurotransmission (Edwards et al., 1992) and activation of voltage-gated Ca²⁺ channels in neurones (reviewed by Bean, 1992) as well as increases in intracellular Ca²⁺ in glial cells (Neary et al., 1991; Kastritsis et al., 1992; Magoski & Walz, 1992; Bruner & Murphy, 1993; Lyons et al., 1994; Salter & Hicks, 1994). ATP is released upon electrical stimulation of rat sensory motor cortex (Wu & Phillis, 1978) and hippocampal slices (Wieraszko et al., 1989). ATP can also be co-released from nerve endings with noradrenaline or acetylcholine (reviewed by Burnstock, 1990). In addition, ATP released from damaged or dying cells can stimulate hypertrophic and hyperplastic responses in astrocytes reminiscent of the reactive gliosis that results from brain injury (reviewed by Neary et al., 1996). These findings argue in favour of a role for extracellular ATP both in normal neurotransmission and in signalling following lesions of the CNS. Thus, there is considerable interest in defining the receptors and signalling pathways responsible for the physiological and pathological actions of extracellular ATP.

The biological effects of extracellular nucleotides are mediated by cell-surface, P₂ purinoceptors, which have been classified as ionotropic P_2X receptors and metabotropic P_2Y receptors (Abbracchio & Burnstock, 1994). Until now, seven P_2X and five P_2Y receptor subtypes have been cloned. The P_2Y receptor subtypes respond differently to a variety of naturally occurring agonists, including ATP, UTP and their diphosphate analogues (Ralevic & Burnstock, 1998). Some synthetic agonists, such as 2-methylthioadenosine triphosphate (2MeSATP), 2-methylthioadenosine diphosphate (2MeSADP) and adenosine 5′-O-(3-thiotriphosphate) (ATPγS) are relatively resistant to degradation by extonucleotidases and are somewhat more specific than the naturally occurring agonists (North & Barnard, 1997). The cloned receptors have been expressed in cell lines lacking endogenous P_2Y receptors and characterized pharmacologically. P_2Y1 receptors are equally responsive to 2MeSATP and ATP but unresponsive to uridine 5′-triphosphate (UTP). Recent studies have suggested that P_2Y1 receptors are ADP-preferring receptors with ATP and 2MeSATP as partial agonists or antagonists (Léon et al., 1997; Hechler et al., 1998), although the partial agonism and antagonism are more pronounced in cells with low rather than normal levels of receptors (Palmer et al., 1998). In contrast to P_2Y1 receptors, P_2Y2 and P_2Y4 receptors in rats are equally responsive to ATP and UTP but unresponsive to 2MeSATP. Another P_2Y receptor subtype, P_2Y6, is preferentially stimulated by uridine nucleotides but to a lesser extent, or not at all, by ATP; similar properties are exhibited by human P_2Y4 receptors. By contrast, a P_2Y11 receptor subtype recently identified in human placenta is activated by ATP but not by UTP (Communi et al., 1997). These differences in agonist affinities for nucleotides and related analogues have been used in attempts to identify receptor subtypes in various cells and tissues, but in some cases there are differences in agonist responses between cloned and native P_2Y receptors, thereby raising the possibility that there are P_2Y receptors not yet cloned (King et al., 1998).

P_2Y purinoceptors are linked to a variety of signal transduction mechanisms, such as phospholipase C and adenylyl cyclase (reviewed by Ralevic & Burnstock, 1998). In addition, occupancy of P_2Y receptors leads to activation of extracellular signal regulated kinase (Erk) (Neary & Zhu, 1994), a key component of a signalling cascade that plays a crucial role in cellular proliferation and differentiation (Seger & Krebs, 1995). The activation of Erk by nucleotide receptors is widely distributed over several cell types, including astrocytes (Neary & Zhu, 1994; King et al., 1996; Neary et al., 1999), PC12 cells (Soltoff et al., 1998), brain capillary and aortic endothelial cells (Patel et al., 1996; Graham et al., 1996; Albert et al., 1997), vascular smooth muscle cells (Harper et al., 1998; Wilden et al., 1998), kidney cells (Huwiler & Pfeilschifter, 1994; Gao et al., 1999) and cardiac myocytes (Post et al., 1996), thereby indicating the importance of the link between purinoceptor signalling and Erk pathways.

The Erk pathway is one of three distinct mitogen-activated protein kinase (MAPK) cascades identified in mammals; the others have been termed stress-activated protein kinase (SAPK, also known as Jnk for c-Jun N-terminal kinase) and p38^MAPK (reviewed by Kyriakis & Avruch, 1996). The cascades are composed of a sequence of three cytosolic protein kinases which phosphorylate and activate the next kinase in the cascade. The first discovered and most studied cascade is the Erk cascade (reviewed by Seger & Krebs, 1995). This cascade involves Erk1/2, which can be activated by a MAPK kinase (MAPKK) termed MAPK/Erk kinase 1 (Mek1), which, in turn, can be activated by a MAPKK kinase (MAPKKK) identified as a member of the Raf family of serine/threonine kinases: cRaf-1 (Kyriakis et al., 1992). Among the Raf isoenzymes that specifically activate the Erk pathway, cRaf-1 is stimulated by a variety of growth factors and ligands of G protein-coupled receptors (reviewed by Avruch et al., 1994) whereas B-Raf seems to be the major Mek1 activator in nerve growth factor-treated PC12 cells (Jaiswal et al., 1994) and in the brain (Catling et al., 1994). There are also MAPKKK enzymes that can activate both Erk and SAPK/Jnk cascades. These enzymes, termed MekK, have different specificity for the Mek subtypes, leading to a different degree of activation of the Erk and SAPK/Jnk cascades.

In this report, we have focused on the types of P_2Y receptors expressed in primary cultures of rat cortical astrocytes and the nature of the Mek activators that are recruited by these P_2Y receptors to stimulate Erk signalling. Previously, we reported that ATP, UTP and 2MeSATP stimulate Erk activity in astrocytes (King et al., 1996). In other studies, we found that stimulation of Erk by extracellular ATP was not affected by activation of the cyclic AMP-dependent protein kinase (PKA) pathway (Neary & Zhu, 1994). Because PKA can inhibit the Erk cascade by interacting with cRaf-1 (Cook & McCormick, 1993; Graves et al., 1993; Sevetson et al., 1993; Wu et al., 1993), these findings suggested that P_2Y receptors may utilize a different Mek activator. To address this question directly, we have now measured the ability of P_2Y receptor agonists to activate cRaf-1, B-Raf and MekK. Surprisingly, we find that, while activation of P_2Y receptors by ATP does not recruit cRaf-1, activation of other P_2Y receptors by UTP, 2MeSATP and ADP does stimulate cRaf-1, thereby revealing for the first time that distinct P_2Y receptor subtypes recruit different Mek activators.

Methods

Reagents

All chemicals were reagent grade or better. Anti-B-Raf monoclonal antibody (sc-166-G), anti-MekK1 (sc-252), anti-MekK2 which also recognizes MekK3 (sc-1089), and protein A/G-agarose (sc-2003) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, U.S.A.). Anti-cRaf-1 was produced as described (Kyriakis et al., 1992). Mek1 and Erk1 polypeptides were expressed as glutathione S-transferase (GST) fusion proteins; after purification by glutathione-Sepharose affinity chromatography, free Erk1 was obtained by thrombin cleavage (Luo et al., 1997). A polyhistidine-tagged Erk2 (sc-4024, Santa Cruz Biotechnology) was used in some experiments with similar results. ATP, ADP, UTP, myelin basic protein (MBP), aprotinin, leupeptin and pepstatin A were obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.). [γ-³²P]-ATP was purchased from DuPont NEN (Boston, MA, U.S.A.) and Amersham Pharmacia Biotech (Piscataway, NJ, U.S.A.). 2MeSATP was obtained from RBI (Natick, MA, U.S.A.). Human recombinant fibroblast growth factor (FGF) 2 was purchased from R&D Systems (Minneapolis, MN, U.S.A.). 4-(2-Aminoethyl)benzenesulphonylfluoride (AEBSF) was obtained from Calbiochem (San Diego, CA, U.S.A.). Protein assay kits were supplied by BioRad Laboratories (Hercules, CA, U.S.A.).

Cell culture and treatment

Primary astrocytes were obtained from neonatal Fischer rat cerebral cortices as previously described (Neary et al., 1994). At least 99% of the cell population was astrocytes, as determined by staining with cell-specific markers (Neary et al., 1994). Experiments were conducted with 3–5 week-old cultures. Prior to treatment with nucleotides or FGF-2, cells that had been maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% horse serum were shifted to the quiescent phase by incubation in DMEM containing 0.5% horse serum for 48 h prior to MAPKKK, Mek and Erk studies. Stock solutions of nucleotides were divided into single-use aliquots and stored at −80°C.

MAPKKK assays

Activation of cRaf-1 and other MAPKKKs was measured by means of a coupled assay in which inactive GST–Mek1 and inactive bacterially expressed Erk1 were added to immunoprecipitates of cRaf-1 or other MAPKKKs, as previously described (Kyriakis et al., 1993; Luo et al., 1997) with minor modifications. After treatment, cells grown on 60 mm plates were lysed in 1 ml lysis buffer (mM) Tris-HCl 20, EDTA 2, β-glycerophosphate 50, 1% w v⁻¹ Triton X-100, 10% v v⁻¹ glycerol, 1 mM sodium orthovanadate, 0.1 mg ml⁻¹ AEBSF, 50 μg ml⁻¹ aprotinin, 2 μM leupeptin, 2 μM pepstatin A; pH 8.0). The lysates were centrifuged (10,000×g for 5 min at 4°C) and aliquots of supernatants were set aside for Erk assay. Supernatants containing equivalent amounts of protein (80–120 μg, as determined with the BioRad Coomassie dye method using bovine serum albumin as standard) were incubated with anti-cRaf-1, anti-B-Raf, anti-MekK1 or anti-MekK2 antibodies for 30 min at 4°C. Twenty microlitres of protein A/G agarose was added and incubation at 4°C was continued overnight, with agitation. The protein A/G agarose beads were washed twice with lithium wash buffer (mM) (Tris-HCl 100, LiCl 500, dithiothreitol (DTT) 1; pH 7.6) and twice with assay buffer (mM) β-glycerophosphate 50, EGTA 1.5, DTT 1, 0.03% Brij 35; pH 7.3). Beads were divided among assay tubes, residual liquid was absorbed with paper strips and 45 μl assay buffer was added. Samples were incubated with 0.2 μCi [γ-³²P]-ATP, 10 mM MgCl₂ and 0.1 mM ATP for 20 min, at 30°C, in the presence or absence of 0.2 μg recombinant inactive GST–Mek1. Following this, 0.5 μg inactive bacterially expressed Erk1 was added for 20 min, after which 13 μg MBP was added for an additional 20 min (total incubation time, 60 min). Reactions were stopped by adding sodium dodecyl sulphate (SDS) sample buffer (Laemmli, 1970) and proteins were separated by SDS polyacrylamide (11%) gel electrophoresis (PAGE). Incorporation of ³²P into MBP was quantified with an Instant Imager^TM electronic autoradiography system (Packard Instruments, Meriden, CT, U.S.A.). To correct for Mek contamination in the MAPKKK immunoprecipitates, control assays were conducted without the addition of GST-Mek1 and the results from these samples were subtracted from those containing the complete reaction mix. MAPKKK activation in nucleotide- and FGF-2-treated cells was compared to basal levels in untreated cells and expressed as fold-stimulation.

Mek assay

Lysate supernatants containing equivalent amounts of protein (80–120 μg) were incubated with 0.2 μCi [γ-³²P]-ATP, 10 mM MgCl₂, 0.1 mM ATP and 0.5 μg inactive bacterially expressed Erk1 for 20 min at 30°C. Reactions were stopped and samples were subjected to SDS gel electrophoresis as described above. Incorporation of ³²P into Erk1 was quantified with a Packard InstantImager.

Erk activity and phosphorylation

Erk activity was assayed as described elsewhere (Neary & Zhu, 1994; Neary et al., 1999). For experiments in which coupled assays were performed, Erk assays were conducted using aliquots of the same samples. Erk phosphorylation was determined by probing immunoblots with an antibody that recognizes dually phosphorylated (Thr¹⁸³, Tyr¹⁸⁵) Erk1 and Erk2 (pTEpY; Promega Corporation, Madison, WI, U.S.A.). Samples containing equal amounts of protein were subjected to SDS–PAGE (Laemmli, 1970) using 11% polyacrylamide and transferred to nitrocellulose filters with a Genie^TM electrophoretic blotter (Idea Scientific Inc., Minneapolis, MN, U.S.A.) for 1 h at 12 V in a transfer buffer containing 25 mM Tris, 192 mM glycine and 20% methanol. Filters were incubated with a blocking solution containing Tween Tris-buffered saline (mM) (TTBS: Tris 20, NaCl 137, Tween 20 v v⁻¹ 0.1%; pH 7.7) and 5% fat-free dried milk for 1 h at room temperature, rinsed in TTBS and then incubated for 1 h at room temperature with anti-phosphoErk1/2 antibodies diluted 1 : 20,000 in TTBS. Following three rinses in TTBS, filters were incubated for 1 h at room temperature with peroxidase-conjugated anti-rabbit IgG (Amersham Pharmacia Biotech) diluted in TTBS (1 : 20,000). Filters were washed three times in TTBS and proteins were detected by enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech). Identical blots were probed with anti-Erk1/2 antibodies which recognize total Erk1/Erk2 (1 : 500; Santa Cruz Biotechnology).

RT–PCR analysis

Total RNA was isolated by a modification of the method of Chomczynski & Sacchi (1987) using the Trizol^TM reagent (GibcoBRL Life Technologies, Rockville, MD, U.S.A.) according to the manufacturer's recommendations. Prior to RT–PCR, the RNA was pre-treated with 1 unit ribonuclease-free deoxyribonuclease (Promega) per microgram RNA for 15 min at 37°C to remove any genomic DNA contaminants. One-step RT–PCR was conducted using the Titan RT–PCR system (Boehringer Mannheim, Indianapolis, IN, U.S.A.). Complementary DNA was synthesized from 0.6 μg total RNA in the presence of (mM) nucleotide triphosphates 0.2, DTT 5, MgCl₂ 1.5, 0.4 μM upstream and downstream primers and a mixture of avian myeloblastosis virus (AMV) reverse transcriptase, Thermus aquaticus (Taq) DNA polymerase and Pyrococcus woesei (Pwo) DNA polymerase in a total volume of 25 μl, according to the manufacturer's recommendations. Oligonucleotide amplification primers were designed from the published rat P_2Y1 (Tokuyama et al., 1995), P_2Y2 (Chen et al., 1996), P_2Y4 (Bogdanov et al., 1998) and P_2Y6 (Chang et al., 1995) sequences. P_2Y1 (nt 1097–1487): sense primer 5′-CTGGGCAGGCTCAAGAAGAAGAAT-3′, antisense primer 5′-AATCCAGCCGTGCCCTCAAAT-3′; P2Y₂ (nt 735–1337): sense primer 5′-CTCCCTGCCGCTGCTGGTTTATTA-3′, antisense primer 5′-ATGTTGATGGCGTTGAGGGTGTG-3′; P2Y₄ (nt 1739-2115): sense primer 5′-CACCGATACCTGGGTATCTGCCAC-3′, antisense primer 5′-CAGACAGCAAAGACAGTCAGCACC-3′; P2Y₆ (nt 689–1091): sense primer 5′-TATAACTACGCCAGAGGGGACCAC-3′, antisense primer 5′-GGCGGGCCATGCGACAGTAG-3′. RT–PCR was conducted as follows. Samples were incubated for 30 min in a thermocycler at 50°C, after which denaturation was conducted at 94°C for 2 min. Ten cycles of denaturation at 94°C for 30 s, annealing at 58°C (P_2Y1), 61°C (P_2Y2), 65°C (P_2Y4) or 60°C (P_2Y6) for 30 s and elongation at 68°C for 45 s was then conducted and followed by 25 cycles of the same denaturing and annealing conditions but with an additional 5 s of elongation for each cycle. The final elongation was conducted for 7 min at 68°C. Reactions were also conducted without the reverse transcriptase step as a control for genomic DNA contamination in the RNA sample. Amplification products were resolved by agarose (1.2%) gel electrophoresis; bands were excised and eluted from the gel, purified (Wizard DNA purification kit; Promega) and sequenced. Three RNA samples from independent culture seedings were analysed by RT–PCR with identical results.

Cyclic AMP determination

Cyclic AMP was assayed as previously described (Tasca et al., 1995). In brief, astrocyte cultures (60 mm plates) were maintained and treated in DMEM supplemented with 10% horse serum. After treatment, cells were lysed in 200 μl 50 mM Tris-HCl (pH 7.4) containing 4 mM EDTA. Lysates were centrifuged, protein was determined and 50 μl of the supernatant (or standard solution) was used in the cyclic AMP assay. [³H]-cAMP (50 μl) and 100 μl binding protein (3 μg of protein kinase A, Sigma) were added and incubations were conducted for 4–16 h on ice. The reaction was stopped with 100 μl activated charcoal, solutions were centrifuged, and [³H]-cAMP in the supernatant was determined by scintillation counting. Data were fitted to the standard curve and corrected for protein content; results were expressed as nmol mg⁻¹ protein.

Statistical analyses

All experiments were conducted a minimum of three times, each time with cultures from different seedings. Data were analysed by Student t-tests for two groups or by ANOVA followed by Bonferroni post hoc comparisons for multiple groups using the InStat® software package (GraphPad^TM Software, San Diego, CA, U.S.A.).

Results

Erk phosphorylation and Mek activation by P_2Y receptor agonists in rat cortical astrocytes

Erk isoforms 1 and 2 (p44 and p42, respectively) are activated by phosphorylation of Thr¹⁸³ and Tyr¹⁸⁵; this dual phosphorylation reaction is catalysed by Mek. To examine the ability of P_2Y receptor agonists to stimulate phosphorylation of Erk, we conducted immunoblot studies with an antibody that recognizes phosphorylated Thr¹⁸³ and Tyr¹⁸⁵ residues of Erk1 and Erk2. We found that the P_2Y receptor agonists ATP, UTP and 2MeSATP stimulated Erk phosphorylation (Figure 1). ATP and UTP (both at 100 μM) were approximately equivalent while 2MeSATP (10 μM) was slightly less effective. The extent of Erk phosphorylation stimulated by these P_2Y agonists is similar to that obtained with FGF-2 (Figure 1). These findings confirm and extend our previous observations that ATP, UTP and 2MeSATP increase Erk activity (King et al., 1996). The concentrations of nucleotides used in these studies stimulate maximal or near maximal activation of Erk. We also investigated the ability of P_2Y agonists to stimulate Mek. ATP and UTP (both at 100 μM) activated Mek; 2MeSATP (10 μM) appeared to be slightly less effective but this difference was not statistically significant (Figure 2). FGF-2 also activated Mek (Figure 2).

Figure 1.

P_2Y receptor agonists stimulate phosphorylation and activation of Erk. Primary cultures of rat cortical astrocytes were treated for 10 min with FGF-2 (25 ng ml⁻¹), ATP (100 μM), UTP (100 μM), or 2-methylthioATP (2MeSATP, 10 μM). Cells were lysed, and lysates containing equivalent amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis. Immunoblots were probed with antibodies which recognize dual-phosphorylated Erk1/Erk2 (upper panel) or total Erk1/Erk2 (lower panel). All three P_2Y receptor agonists stimulated Erk phosphorylation to an extent similar to that obtained with FGF-2.

Figure 2.

Mek activation by P_2Y agonists and FGF-2. Primary astrocyte cultures were treated for 2 min with ATP (100 μM), UTP (100 μM), 2MeSATP (10 μM) or FGF-2 (25 ng ml⁻¹); Mek activity was determined as described in Methods. Data (mean±s.e.mean) were obtained from a minimum of six experiments. ATP, UTP, 2MeSATP and FGF-2 stimulated Mek activity (P<0.05).

Differences in coupling of P_2Y receptor subtypes to the Erk pathway

To further explore the coupling of native P_2Y receptors in rat cortical astrocytes to the Erk signalling pathway, we investigated whether P_2Y receptor agonists recruit cRaf-1, an activator of Mek (Kyriakis et al., 1992). For comparison, experiments were also conducted with FGF-2, a known activator of cRaf-1 in astrocytes (Kurino et al., 1996). Previously, we found that activation of the cyclic AMP/PKA pathway, which can inhibit the Erk cascade by interacting with cRaf-1 (Cook & McCormick, 1993; Wu et al., 1993; Graves et al., 1993; Sevetson et al., 1993), did not affect extracellular ATP-evoked stimulation of Erk (Neary & Zhu, 1994); this suggested that one or more P_2Y receptor subtypes may utilize a Mek activator other than cRaf-1. Consistent with this, we found that, although FGF-2 did activate cRaf-1, ATP did not, even though ATP and FGF-2 were equally effective in activating Erk (Figure 3). Time course studies were then conducted to determine whether ATP might activate cRaf-1 earlier than the 5 min point described in the preceding experiment. However, ATP did not activate cRaf-1 at 1 and 2 min (Figure 4A); subsequent experiments indicated that ATP was also ineffective at 30 s (data not shown). Surprisingly, when other P_2Y receptor agonists were examined in these studies, we found that cRaf-1 was activated by 2MeSATP and UTP at 2 and 5 min (Figure 4A). It should be noted that, even though cRaf-1 was not activated by ATP, Erk activity was stimulated by ATP, UTP and 2MeSATP with a similar time course (Figure 4B). The effect of ADP on components of the Erk cascade was also examined (Table 1). A similar, but not greater, increase in cRaf-1 activity was observed with ADP as compared to 2MeSATP and UTP. ADP also activated Mek to a similar extent to ATP, UTP and 2MeSATP. At the peak time of Erk activation (10 min), the ADP response was approximately 30% less than that of ATP (data not shown), a finding that confirms our previous observation that ADP was less effective than ATP in stimulating Erk activity (King et al., 1996).

Figure 3.

Inability of extracellular ATP to activate cRaf-1. Primary astrocyte cultures were treated with ATP (100 μM) or FGF-2 (25 ng ml⁻¹) for 5 min. cRaf-1 was immunoprecipitated from lysates, and cRaf-1 activity was determined by the coupled assay described in Methods. Erk assays were conducted using aliquots of the same samples. Data (mean±s.e.mean) were obtained from seven independent experiments. FGF-2 activated Raf-1 but ATP did not (left panel), even though both FGF-2 and ATP stimulated Erk to the same extent (right panel).

Figure 4.

Time course studies: cRaf-1 activation by UTP and 2MeSATP. (A) Primary astrocyte cultures were treated with ATP (100 μM), UTP (100 μM), 2MeSATP (10 μM) or FGF-2 (25 ng ml⁻¹) for the indicated times. cRaf-1 was immunoprecipitated from lysates, and cRaf-1 activity was determined by the coupled assay described in Methods. Data (mean±s.e.mean) were obtained from at least three independent experiments. The data for the 1 min time points obtained with P_2Y agonists and all ATP time points were not significantly different from control values (P>0.05). All other data were significantly different from control values (P<0.05). UTP and 2MeSATP, but not ATP, activated cRaf-1. (B) Aliquots of samples from A were assayed for Erk activation. All P_2Y agonists tested stimulated Erk activity from 2 to 10 min; by 60 min Erk activity in ATP- and 2MeSATP-treated cells had returned to near basal levels, whereas Erk activity in UTP- and particularly in FGF-2 (data not shown)-treated cells remained elevated.

Table 1.

Effect of ADP on components of the ERK cascade

In an effort to determine the Mek activator recruited by ATP, we measured the ability of ATP to activate B-Raf, the main form of Raf found in the brain (Catling et al., 1994), as well as MekK1 and MekK2/3. However, ATP did not stimulate any of these Mek activators (Figure 5). The effects of 2MeSATP and UTP on B-Raf, MekK1, and MekK2/3 were also examined. We found that 2MeSATP was also ineffective in stimulating these Mek activators, whereas UTP significantly, but moderately (50%), stimulated MekK1 (Figure 5). UTP also appeared to stimulate MekK2/3 but the results were not significant owing to variability in the four experiments. FGF-2 stimulated MekK1 and MekK2/3 but not B-Raf, although B-Raf was detected in immunoprecipitates by Western blot analysis (data not shown).

Figure 5.

Effect of the P_2Y agonists and FGF-2 on different Mek activators. Primary astrocyte cultures were treated for 2 min with ATP (100 μM), UTP (100 μM), 2MeSATP (10 μM) or FGF-2 (25 ng ml⁻¹). B-Raf, MekK1 and MekK2/3 were immunoprecipitated from lysates, and B-Raf, MekK1 and MekK2/3 activities were determined by the coupled assay described in Methods. Data (mean±s.e.mean) were obtained from a minimum of three independent experiments. ATP and 2MeSATP did not activate any of the Mek activators tested (P>0.05). UTP and FGF-2 activated MekK1 and MekK2/3 (^*P<0.05).

Expression of P_2Y receptors in rat cortical astrocytes

To identify P_2Y receptor subtypes expressed in rat cortical astrocytes, we conducted RT–PCR analysis using primer pairs designed to amplify regions of the cDNA encoding the four cloned rat P_2Y receptor subtypes, P_2Y1, P_2Y2, P_2Y4, and P_2Y6. PCR analysis revealed that three receptor sequences could be amplified from rat cortical astrocytes: P_2Y1, P_2Y2 and P_2Y4 (Figure 6). All three amplification products were of the expected sizes (P_2Y1, 391 base pairs; P_2Y2, 603 base pairs; P_2Y4, 377 base pairs) and their identities were confirmed by sequencing. An amplification product for P_2Y6 was not found in rat astrocytes, although P_2Y6 expression was detected using the same primers with cDNA templates from rat heart tissue (data not shown). ATP and UTP are equipotent agonists at recombinant rat P_2Y2 and P_2Y4 in terms of stimulating increases in intracellular calcium (Chen et al., 1996) or membrane currents (I_C1,Ca) (Bogdanov et al., 1998), respectively. To examine the potency of ATP and UTP on native P_2Y receptors in rat cortical astrocytes, concentration-response studies were conducted. We found that although ATP and UTP recruit different Mek activators, these nucleotides have similar potencies (EC₅₀ values: ATP, 5.4 μM; UTP, 3.0 μM) in terms of stimulating Erk activity (Figure 7).

Figure 6.

RT–PCR analysis of P_2Y receptor subtype expression in rat neonatal cortical astrocytes. Primer pairs specific for the amplification of rat P_2Y1, P_2Y2, P_2Y4, and P_2Y6 receptor subtypes were used in these experiments. Amplification products were resolved by agarose gel electrophoresis. M, size markers as indicated; –, PCR reactions conducted without the reverse transcriptase step to control for amplification of genomic DNA in the RNA samples. Results are representative of experiments conducted with RNA extracted from at least three independent culture seedings. P_2Y1, P_2Y2 and P_2Y4, but not P_2Y6, receptor subtypes were expressed in rat neonatal cortical astrocytes.

Figure 7.

Concentration dependence of ATP and UTP stimulation of Erk activity. Primary cultures of rat cortical astrocytes were treated for 10 min with the indicated concentrations of ATP or UTP, and Erk activity was determined. Data (mean±s.e.mean) are from four independent experiments and were fitted in a sigmoid log concentration-response curve by Prism® v2.0 (GraphPad™). The concentration-response relationships revealed that ATP and UTP have similar potencies in terms of Erk activation.

Effect of extracellular ATP on cyclic AMP synthesis

Because of the differences in P_2Y agonist stimulation of cRaf-1 and Erk activities, we tested for the presence of P_2Y11, an ATP-preferring receptor recently found in human placenta (Communi et al., 1997). This P_2Y subtype has not been cloned from rat tissue but functional studies demonstrated that it is linked positively to adenylyl cyclase (Communi et al., 1997). To assess whether P_2Y11 is involved in signal transduction in cortical astrocyte cultures, cells were treated with ATP (100 μM) for 10 min and cyclic AMP assays were performed. Extracellular ATP did not increase cyclic AMP levels, although treatment with forskolin (10 μM) did stimulate cyclic AMP production (Figure 8). ATP treatment resulted in a decrease in cyclic AMP synthesis that was significant (P<0.01) when normalized to basal levels within each experiment. Although extracellular ATP did not stimulate basal cyclic AMP production, it potentiated the forskolin-induced increase in cyclic AMP levels (Figure 8), an effect reported previously in endothelial cells (Albert et al., 1997).

Figure 8.

Effect of ATP and forskolin on cyclic AMP synthesis. Primary astrocyte cultures were treated for 10 min with ATP (100 μM), forskolin (Fsk, 10 μM) or ATP+forskolin, and cyclic AMP synthesis was determined as described in Methods. Data (mean±s.e.mean) were obtained from five experiments. ATP did not stimulate cyclic AMP formation but it significantly potentiated forskolin-induced cyclic AMP synthesis (P<0.01, ANOVA).

Discussion

The role of purines and pyrimidines as transmitters and trophic factors in a greater variety of cells and organs has seen a rapid growth in the last 5 years. Seven subtypes of the ionotropic purinoceptors (P_2X) and five subtypes of the metabotropic purinoceptors (P_2Y) have been cloned from mammalian tissues (Ralevic & Burnstock, 1998). Although eleven P_2Y receptor subtypes have been classified, doubts have been raised about the real identity of some of these receptors. The receptor cloned as being a purinoceptor and named P_2Y7, actually was shown to be a leukotriene B₄ receptor (Yokomizo et al., 1997), and Janssens et al. (1997) did not find any effect of purinergic agonists on cells transfected with a plasmid containing a receptor closely related to P_2Y5. This observation is strengthened by the lack of four positively charged amino acids conserved in all other P_2Y subtypes and not present in P_2Y5 and the related P_2Y9 and P_2Y10 subtypes (Janssens et al., 1997). P_2Y8 has been described in Xenopus embryos but its presence in mammals is uncertain (Bogdanov et al., 1997).

The results presented here demonstrate that native P_2Y receptors in astrocytes recruit different Mek activators to stimulate Erk. UTP, 2MeSATP and ADP utilize cRaf-1, whereas ATP-preferring receptors recruit a different Mek activator. ATP was reported to activate Raf-1 in vascular smooth muscle cells (Yu et al., 1996) but this paper has been retracted (Wu et al., 1998). The identity of the Mek activator recruited by ATP-preferring P_2Y receptors in astrocytes remains unknown. Two major families of Mek activators have been described. The Raf family includes cRaf-1 and B-Raf, the predominant form in the brain; these serine/threonine kinases activate Mek1. A-Raf, another member of the Raf family, is not expressed in rat brain (Morice et al., 1999). The second Mek activator family comprises MekK1-5. Mek1 and Sek1 are targets of MekK1, 2 and 3, while Sek1 is stimulated by MekK4 and 5. Besides these two families, there are other proteins which function as Mek activators such as Mos (Nebreda et al., 1993), which activates specifically Mek1, and Tpl2, which activates both Mek1 and Sek1 (Salmerón et al., 1996). The mixed lineage kinases SPRK and ASK1 do not activate Mek1 but activate the other MAPK cascades (Rana et al., 1996; Ichijo et al., 1997). Besides these MAPKK activators, an atypical, non-phorbol ester-stimulated PKC (PKCζ), which is structurally related to Raf, can activate Mek1 (Liao et al., 1997; van Dijk et al., 1997; Schönwasser et al., 1998) but it does so in an indirect manner, acting through an as-yet unidentified factor. The Erk pathway stimulated by ATP in astrocytes is dependent on PKC but this pathway involves PKCδ, a calcium-independent, phorbol ester-stimulated PKC isoform, rather than an atypical PKC (Neary et al., 1999). It is not known whether calcium-independent, phorbol ester-stimulated PKC isoforms can bypass Raf but overexpression of a constitutively active mutant of PKCδ is sufficient to activate Mek and Erk (Ueda et al., 1996). In the studies presented here, ATP activated Erk without using cRaf-1, B-Raf or the known activators of Mek1, namely, MekK1, 2 or 3. Although the nature of the Mek activator recruited by extracellular ATP remains to be identified, we have found that UTP, 2MeSATP and ADP recruit cRaf-1 whereas ATP does not, thereby indicating that distinct P_2Y receptors are coupled to different Mek activators in rat cortical astrocytes.

Comparison of the pharmacological profile for the activation of cRaf-1 by P_2Y agonists with the subtypes of P_2Y receptors expressed in rat cortical astrocytes makes it difficult to identify the subtypes of P_2Y receptors coupled to the Erk cascade because the profile does not fit the properties of recombinant P_2Y receptors. Studies on the activation of Erk by ATP in rat cortical astrocytes had demonstrated previously that signalling occurs via P_2Y receptors rather than adenosine/P₁ receptors, the latter of which could be stimulated by breakdown of ATP to adenosine catalyzed by ectonucleotidases (Neary & Zhu, 1994; King et al., 1996). These studies also showed that 2MeSATP, ATP and UTP activated Erk to a similar extent, whereas ADP was less effective than ATP. This suggested the presence of endogenous ATP-, ATP/UTP- and perhaps UTP-preferring receptors. Previous RT–PCR studies demonstrated the presence of P_2Y1 receptors (Webb et al., 1996) and this has been confirmed in the results presented here. P_2Y1 receptors respond preferentially to ATP and 2MeSATP, without any response to UTP (Simon et al., 1995; Tokuyama et al., 1995; Henderson et al., 1995), and 2MeSADP and ADP have a high selectivity for this receptor subtype (Léon et al., 1997; Hechler et al., 1998; Palmer et al., 1998). However, we found that 2MeSATP and ADP recruit a different Mek activator than ATP in rat cortical astrocytes, a finding that is inconsistent with the P_2Y1 pharmacological profile. It has been suggested that 2MeSATP and ATP are partial agonists or antagonists of P_2Y1 receptors (Léon et al., 1997; Hechler et al., 1998), and thus the lack of effect of ATP on cRaf-1 activation could be interpreted to suggest that ATP is acting as partial agonist or antagonist in astrocytes. However, recent studies have shown that maximal effects of ATP and 2MeSATP on P_2Y1 receptors are very similar to ADP and 2MeSADP under conditions of normal receptor reserve; the partial agonism or antagonism of ATP and 2MeSATP may be the result of low receptor reserve (Palmer et al., 1998). Because both ATP and 2MeSATP stimulated similar responses in term of Erk activation while ATP did not recruit cRaf-1, low receptor reserve cannot explain the difference in activation of cRaf-1 by these nucleotides; thus it is unlikely that ATP and 2MeSATP would be partial agonists or antagonists of P_2Y1 receptors under the conditions used in our studies. Moreover, our findings that (1) a 10 fold higher concentration of ADP did not cause a greater activation of cRaf-1 than 2MeSATP and (2) the maximal effect of ATP on Erk activity was greater than that of ADP are inconsistent with an effect on P_2Y1 receptors because ADP is more active than 2MeSATP on P_2Y1 receptors (Léon et al., 1997; Hechler et al., 1998; Palmer et al., 1998). Furthermore, adenosine 3′-phosphate 5′-phosphosulphate, a selective competitive antagonist of P_2Y1 receptors (Boyer et al., 1996), was not effective in blocking activation of Erk by ATP or 2MeSATP (Y. Kang and J.T. Neary, unpublished observations). Collectively, these observations indicate that, if P_2Y1 receptors are coupled to Erk in rat cortical astrocytes, their pharmacological properties differ from those of the cloned receptors.

The RT–PCR data presented here indicate that rat cortical astrocytes also express P_2Y2 and P_2Y4 receptors. However, the pharmacological profile for recombinant P_2Y2 and P_2Y4 receptors is also inconsistent with our findings on the activation of cRaf-1 by P_2Y agonists. For example, ATP and UTP would be expected to activate rat P_2Y2 and P_2Y4 receptors but our results show that UTP recruits cRaf-1 while ATP does not, thereby excluding receptors activated by both adenine and uridine nucleotides. If P_2Y2 and P_2Y4 receptors in astrocytes are not coupled to Erk, then what is the nature of the endogenous P_2Y receptor which is responsible for the activation of Erk by UTP? Among the cloned P_2Y subtypes from rat tissues, P_2Y6 receptors preferentially recognize uridine nucleotides when compared to adenosine nucleotides (Chang et al., 1995). However, this subtype can be excluded because RT–PCR analysis failed to reveal an amplification product for this receptor in rat cortical astrocytes. Recently, we reported that P_2Y receptors in rat cortical astrocytes are coupled independently to Erk and phosphatidylinositol-specific phospholipase C/calcium pathways (Neary et al., 1999). Astrocyte P_2Y1, P_2Y2 and P_2Y4 receptors may be coupled to the phospholipase C/calcium pathway but the Raf and Erk data presented here suggest that either the properties of the native P_2Y receptors differ from those of the recombinant P_2Y1, P_2Y2 and P_2Y4 receptors, or rat cortical astrocytes express novel purine-preferring and pyrimidine-preferring receptors coupled to the Erk signalling pathway.

It has been reported that P_2Y2 and P_2Y4 receptors are coupled to Erk in endothelial (Graham et al., 1996; Albert et al., 1997), smooth muscle (Harper et al., 1998; Wilden et al., 1998) and PC12 cells (Soltoff et al., 1998). In another study, stimulation of a P_2Y2-like receptor in human embryonic kidney-293 cells activated Erk in a Ras-dependent manner (Gao et al., 1999). Collectively, these observations suggest that the coupling of ATP/UTP-preferring receptors to the Erk cascade in these cells differs from rat cortical astrocytes, although cRaf-1 recruitment by different P_2Y agonists was not examined in the previous studies.

Another subtype, P_2Y11, has been recently cloned from human placenta. This receptor is positively coupled to phosphoinositide and cyclic AMP pathways (Communi et al., 1997). Because the rat homologue of this receptor has not been reported, we have not been able to investigate its expression in rat astrocytes by RT–PCR analysis. However, we have conducted experiments to measure cyclic AMP levels in response to extracellular ATP. In these studies, extracellular ATP did not evoke an increase in cyclic AMP synthesis, suggesting that P_2Y11 receptors are not present in rat cortical astrocytes or are not coupled to adenylyl cyclase in these cells. Interestingly, forskolin-induced cyclic AMP production was potentiated by ATP; this phenomenon has also been observed in endothelial cells (Albert et al., 1997). Similarly, forskolin-induced cyclic AMP formation is also potentiated by metabotropic glutamate receptor agonists in cortical astrocytes (Balazs et al., 1998).

In summary, our data suggest that purine-preferring P_2Y receptors and pyrimidine-preferring P_2Y receptors are functionally linked to the activation of the Erk/MAPK cascade in primary cultures of rat cortical astrocytes. Rat P_2Y receptors activated by both adenine and uridine nucleotides, such as P_2Y2 and P_2Y4, are expressed in these cells but they do not appear to be coupled to the Erk cascade because ATP and UTP do not recruit the same MEK activator. The ability of native P_2Y receptors to signal to Erk via different pathways provides opportunities to regulate the trophic actions of extracellular nucleotides by enhancing or interrupting signalling from specific P_2Y receptor subtypes. In addition, P_2Y receptors are coupled to phospholipase C/calcium and adenylyl cyclase/cyclic AMP signalling pathways. This suggests that a set of receptor subtypes linked to different signal transduction mechanisms will be able to produce very different responses according to the set of signalling pathways activated.

Abbreviations

2MeSADP

2-methylthioadenosine diphosphate

2MeSATP

2-methylthioadenosine triphosphate

AEBSF

4-(2-aminoethyl)benzenesulphonylfluoride

DTT

dithiothreitol

Erk

extracellular signal regulated protein kinase

FGF

fibroblast growth factor

GST

glutathione S-transferase

MAPK

mitogen activated protein kinase

MAPKK

MAPK kinase

MAPKKK

MAPKK kinase

MBP

myelin basis protein

Mek

MAPK/Erk kinase

MekK

Mek kinase

PKA

cyclic AMP-dependent protein kinase

PKC

protein kinase C

RT–PCR

reverse transcriptase-polymerase chain reaction

SAPK

stress-activated protein kinase

SDS

sodium dodecyl sulphate

Sek

SAPK/Erk kinase

UTP

uridine 5′-triphosphate