INTRODUCTION

Receptors for extracellular nucleotides (P2 receptors) are found on the cell surface of all higher animal tissues, where they regulate a broad range of physiological processes (1). These receptors have been categorized into two major groups (P2X and P2Y), based on their pharmacological and electrophysiological properties, as well as their molecular structure (2, 3). P2X receptors are members of the ligand-gated ion channel superfamily, while P2Y receptors have seven transmembrane domains and belong to the G-protein-coupled-receptor superfamily. To date, seven P2X (P2X_1-7) and seven P2Y (P2Y_1-7) receptors have been cloned from mammalian and avian species (3). Different P2Y receptors show preferential selectivity for purine and pyrimidine nucleotides; whereas P2Y₁ is responsive to ATP but not UTP (4), P2Y₂ is equally responsive to both ATP and UTP (5) and P2Y₄ is responsive to UTP and much less to ATP (6, 7). p2y3¹ and P2Y₆ are selective for nucleoside diphosphates (8, 9), while p2y5 and P2Y₇ bind ATP with a greater avidity than UTP (10, 11). In most cell types P2Y receptor activation results in the hydrolysis of phosphatidylinositol (4,5)-bisphosphate to the Ca²⁺-mobilizing second messenger inositol (1,4,5)-triphosphate and diacylglycerol, a process catalyzed by PLC- (12).

Cell-cell interactions are crucially important during early embryonic development, providing the impetus to establish and maintain different cell fates, regulate morphogenesis, and control cell differentiation (13). While most of the current evidence suggests that secreted polypeptide growth factors, such as the fibroblast growth factor and transforming growth factor- superfamilies, are the major intercellular signaling molecules in embryonic development (14), there is some evidence to suggest a role for nonpeptide signaling molecules such as serotonin, noradrenaline, and dopamine (15). For example, in the fruit fly Drosophila melanogaster, a serotonin receptor is expressed in even parasegments at the cellular blastoderm, an expression pattern similar to that of the pair-rule gene fushi-tarazu (16). Although the role of this receptor in Drosophila development is currently unknown, that it must be functioning is suggested by the presence of ligand at the same stage of development. In Xenopus neurulae, while -adrenergic receptor antagonists and inhibitors of dopamine -hydroxylase inhibit neuronal differentiation, exogenous noradrenaline or dopamine can increase the number of neurons differentiating in neural cultures (17). This suggests that endogenous noradrenaline, acting via -adrenergic receptors, is part of the mechanism controlling neuronal differentiation in the central nervous system. Very little is known about the roles of extracellular ATP in early embryonic development, but this molecule has been shown to cause an increase in intracellular Ca²⁺ (Ca²⁺_i)² concentration in a number of embryonic cell types, including dissociated cells from early chick embryos (18), early embryonic chick otocyst (19) and retina (20), cultured astrocytes from embryonic rat spinal cord (21), cultured neurons from embryonic rat brain (22), myotubes cultured from embryonic chick (23), and a murine myoblast cell line (24). Increases in Ca²⁺_i are known to regulate cell proliferation, migration, and differentiation (15, 25), all important processes during embryonic development. The P2 receptor(s) responsible for most of these ATP-activated developmental signals have not been identified, and none have been cloned, although the response of embryonic chick retina is thought to be mediated by a P2Y receptor responsive to UTP (20).

In this study, we have isolated a cDNA encoding a novel P2Y receptor, XlP2Y, that is expressed during neurulation in Xenopus embryos. XlP2Y is equally responsive to all five naturally occurring nucleoside triphosphates (ATP, UTP, CTP, GTP, and ITP) and, when expressed in Xenopus oocytes, exhibits an unusually long response to agonists. Our results show that expression of this receptor occurs transiently during embryonic development, being coincident with the phase of primary neurulation, and suggest that this receptor may play an important role in the early development of neural tissue in Xenopus laevis.

EXPERIMENTAL PROCEDURES

Degenerate oligonucleotide primers based on the sequence of transmembrane domains III and VII of chick P2Y₁ (4) and murine P2Y₂ (5) were used to amplify fragments of approximately 574 bp from cDNA synthesized from Xenopus neurula total RNA. The forward primer was 5-GCAGCATCCT(C/G)TTCCTCAC(C/G)TGCAT-3 (amino acid sequence SILFLTCI), and the reverse primer was 5-CCC(G/A/T)GCCAGGAAGTAGAG(G/T/C)A(C/T)(G/C)GG-3 (amino acid sequence P(M/I/V)LYFLAG). The polymerase chain reaction (PCR) amplification conditions were 94 °C for 60 s, 55 °C for 45 s, 72 °C for 60 s for 30 cycles, followed by 72 °C for 10 min. PCR fragments were subcloned into the pCRII TA cloning vector (Invitrogen) and sequenced by the dideoxy chain termination method.

The Xenopus P2Y PCR fragment was used as a probe to screen 7 × 10⁵ recombinant phage of a Xenopus stage 17 (mid-neurula) cDNA library in gt10 (26). Hybridization was performed at 65 °C in buffer containing 4 × SSC, 5 mM EDTA, 5 × Denhardt's solution, 1% SDS, 0.1 mg/ml salmon sperm DNA (Sigma). Final washing of membranes was at 65 °C in 0.2 × SSC, 0.1% SDS. The cDNA of the longest positive clone was subcloned into pBluescript II KS(+) (Stratagene) and sequenced by the dideoxy chain termination method.

Unfertilized eggs were obtained from females of X. laevis previously injected with 500 units of human chorionic gonadotrophin (Inervet). They were fertilized with a piece of macerated testis, dejellied in 2% cysteine hydrochloride (Sigma), reared in 10% Barth's solution (Barth's solution is 110 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 7.5 mM Tris-HCl, 0.33 mM Ca(NO₃)₂, 0.41 mM CaCl₂, 0.82 mM MgSO₄, 50 µg/liter gentamycin, adjusted to pH 7.45) and staged according to Nieuwkoop and Faber (27). Exogastrulae were generated by incubating embryos in full-strength Barth's solution throughout development. Embryos were UV-irradiated at the vegetal pole 35 min after fertilization. Dissections were carried out in full-strength Barth's solution, using forceps and electrolytically sharpened tungsten needles.

Total RNA was isolated from staged embryos and embryo fragments, separated on formaldehyde-agarose gels, and transferred to Hybond N⁺ (Amersham) using standard techniques (28). Blots were sequentially probed, under high stringency, with random primed probes for XlP2Y and Xenopus histone H4, the latter acting as a loading control.

Whole mount in situ hybridization was performed on albino embryos as described previously (29), with the exception that an RNase digestion step was not included, CHAPS buffer was replaced by maleic acid, and BM Purple (Boehringer Mannheim) was used as a substrate. Sense and antisense digoxygenin (Boehringer Mannheim) labeled probes for XlP2Y were made by transcription of the initial PCR clone with either T7 (antisense) or SP6 (sense) RNA polymerase according to the manufacturer's instructions.

A 2.361-kilobase fragment of XlP2Y was subcloned into the pRN3 transcription vector (30), and capped synthetic RNA was transcribed with T3 RNA polymerase using the relevant Megascript kit (Ambion). For detection of the XlP2Y translation product, XlP2Y in pRN3 was added to an aliquot of a TNT-T3-coupled reticulocyte lysate system (Promega) along with 40 µCi of Prolabel (Amersham). As controls, either water or a luciferase cDNA provided by the manufacturers was added to parallel reactions. The subsequent translation products were separated by SDS-PAGE on a mini-gel system (Bio-Rad) according to the manufacturer's instructions, and prepared for fluorography by immersion in EN³HANCE (DuPont NEN).

Xenopus oocytes (stages V and VI) were plucked off the inner ovarian epithelial lining with fine forceps and stored (at 4 °C) in Barth's solution. The enveloping follicle cells were removed from oocytes, since these epithelial cells possess P1 and P2 receptors (31, 32) while defolliculated oocytes are devoid of these receptors. Oocytes were defolliculated by a two-step process of collagenase treatment (Sigma Type 1A, 2 mg/ml, in a Ca²⁺-free Ringer solution) followed by mechanical stripping. Defolliculated oocytes were cytosolically injected with XlP2Y RNA (40 nl, 0.5 µg/µl) and incubated at 18 °C for 49 h in Barth's solution to allow full expression of this receptor. Control oocytes were injected with either sterile water (40 nl) or an antisense RNA for chick P2Y₁ (40 nl, 1 mg/ml) and incubated under the same conditions.

Nucleotide-activated currents (I_Cl,Ca) were recorded from injected oocytes held under voltage-clamp (V_h = 40 mV) using a twin-electrode amplifier (Axoclamp 2A). I_Cl,Ca is routinely used as a reporter current for the activation of phospholipase C, IP₃ production, and release of Ca²⁺_i by G-protein-coupled receptors, including P2Y subtypes (33). The voltage-recording and current-recording electrodes (1-2 megaohms tip resistance) were filled with 0.6 M K₂SO₄ and 3 M KCl, respectively. Oocytes were superfused (5 ml/min) with Ringer's salt solution (at 18 °C) containing (mM): NaCl, 110; KCl, 2.5; HEPES, 5, CaCl₂, 1.8; adjusted to pH 7.45. Nucleotides (Sigma and Research Biochemicals International) were added to the superfusate at the concentrations given in the text for periods of 60-180 s, followed by a period of washout of 60 min. Evoked responses were recorded on magnetic tape using a DAT recorder (Sony 1000ES) and displayed using a pen recorder (Gould). For the concentration-response curve for ATP, data were normalized to the amplitude of responses obtained using ATP (1 µM), which gave submaximal responses. A submaximal standard was chosen because it was only possible to test 3-4 concentrations of ATP or other nucleotides on one oocyte, given the long duration (approximately 40-60 min) of each response. Pharmacological data are expressed as the mean of three observations from separate oocytes.

RESULTS

Using degenerate oligonucleotide primers based on the P2Y₁ and P2Y₂ sequences encoding the highly conserved transmembrane domains III and VII, we used reverse transcriptase-PCR to amplify P2Y sequences from cDNA made from an RNA pool extracted from X. laevis neurulae. A 560-bp fragment was identified among the resultant products, and sequencing suggested that it encoded a novel P2Y receptor. This PCR fragment was used to screen a X. laevis neurula (stage 17) cDNA library, and a number of positive clones were identified, the largest insert size being 2.361 kilobases. Sequence analysis of this insert revealed an open reading frame of 1611 bp, but the closest fit to the Kozak translation initiation consensus sequence (34) is met by the sequence surrounding a second ATG 15 bp downstream (Fig. 1). Translation from this second ATG would produce a protein of 532 amino acids, which is somewhat larger than the 308-377 amino acids described for P2Y_1-7. This is the consequence of a relatively long carboxyl-terminal tail of 216 amino acids (Fig. 1), compared with the 16-67 amino acids for P2Y_1-7. The long carboxyl-terminal tail includes a number of potential phosphorylation sites for protein kinase C (× 1), protein kinase A (× 5), calmodulin dependent kinase (× 5), GSK3 (× 4), and tyrosine kinase (× 2). There is also a single site for phosphorylation by protein kinase C in the third intracellular loop. To confirm that XlP2Y does indeed encode a protein with a long carboxyl-terminal tail, we have produced the translation product in a combined in vitro transcription-translation system and analyzed it by SDS-PAGE (Fig. 2). XlP2Y has an M_r of approximately 56-57 × 10³ compared with the predicted M_r of 61 × 10³. Although XlP2Y migrates faster than predicted from its amino acid sequence, which is not uncommon for proteins separated by SDS-PAGE, it is still significantly slower than is observed with other P2Y receptors.

Hydropathy analysis of the predicted XlP2Y receptor protein revealed the presence of seven putative transmembrane domains, a feature characteristic of G-protein-coupled receptors. The NH₂-terminal domain contains two potential asparagine-linked glycosylation sites and a cysteine that is conserved in all the known members of the P2Y family. A single conserved cysteine is also found in each of the three putative extracellular loops, these four cysteines probably forming two disulfide bonds. A cysteine in the carboxyl-terminal tail, which is conserved in many G-protein-coupled receptors, may be a membrane-anchoring palmitoylation site. The amino acid sequence of XlP2Y was compared with the sequences of the seven previously reported P2Y receptors, and a total of 26 amino acids were found to be absolutely conserved (Fig. 3). We note a high degree of homology between XlP2Y and P2Y_1-6 in TM III and between XlP2Y and both P2Y₂ and P2Y₄ in TM VII. XlP2Y is most closely related to the UTP receptors P2Y₄ (62% identical) and P2Y₂ (56% identical) and least related to P2Y₇ (26% identical). However, neither P2Y₂ nor P2Y₄ possess a carboxyl terminus of a similar length as XlP2Y, suggesting that XlP2Y is not the Xenopus homologue of these mammalian receptors. Four positively charged amino acids (His²⁷³, Arg²⁷⁶, Lys³⁰⁰, and Arg³⁰³) reported (35) to play a role in P2Y₂ receptor activation by ATP and UTP are conserved in XlP2Y (Fig. 3).

Embryos of X. laevis were collected at different stages over a period of 3 days post-fertilization. Total RNA was prepared from these embryos, and the temporal pattern of XlP2Y transcription was determined by Northern blotting (Fig. 4A). XlP2Y was undetectable during the earliest stages of development, from fertilization (stage 1) to onset of gastrulation (stage 10), but a single transcript of approximately 3.5 kilobases was detected at the beginning of neurulation (stage 13). Transcript levels remained high throughout neurulation (stages 13-20) before dropping to a lower level at tadpole stages. This lower level was maintained until stage 40, the last embryonic stage that we have examined. Since it took these embryos approximately 9 h to progress from stage 10 to stage 13, we have determined the time point at which XlP2Y expression is initiated by collecting embryos at 1-h intervals following the initiation of gastrulation at stage 10. A Northern blot of total RNA prepared from these embryos shows that XlP2Y transcripts can first be detected 7 h after the onset of gastrulation, and reaches maximal levels by 8 h (Fig. 4B). XlP2Y expression is therefore initiated during the later phases of gastrulation.

The regional distribution of XlP2Y was initially determined by dissecting Xenopus neurulae (stage 17) into dorsal anterior, dorsal posterior, and ventral regions. Northern blot analysis of total RNA isolated from these regions showed that XlP2Y transcripts are most abundantly expressed in dorsal regions of the embryo, which differentiates into neural tissues, notochord, and somites, although transcripts can also be detected ventrally (Fig. 4C). Consistent with dorsal localization XlP2Y expression was reduced in embryos irradiated with UV light during the first cell cycle (Fig. 4C), a treatment that reduces the development of dorsal tissues (36). XlP2Y expression was also reduced in embryos cultured in a high salt solution during gastrulation (Fig. 4C). Under these conditions the mesoderm does not involute under the ectoderm, and the expression of several neural tissue-specific, but not mesoderm-specific, genes are reduced (37). These results suggest that XlP2Y is most abundantly expressed in the developing neural tissue during the process of neurulation, when the open neural plate folds to form the neural tube. To confirm this suggestion, we analyzed embryos by whole mount in situ hybridization using a digoxygenin-labeled probe for XlP2Y. Transcripts were detected first at stage 13-14 in an arc corresponding to the anterior ridge of the neural plate (Fig. 5A), subsequently spreading throughout the neural plate (Fig. 5B). After neural tube closure XlP2Y was no longer detectable in caudal regions of the neural tube but was detectable in neural tissue emerging from the tailbud (Fig. 5C).

Defolliculated oocytes injected with XlP2Y RNA responded to low concentrations of ATP (10-300 nM) with a prolonged period of oscillatory currents and to higher concentrations of ATP (1 µM and greater) with a biphasic current, where the second phase persisted for 40-60 min after brief superfusion (60-180 s) of agonists (Fig. 6, A and B). Because of these prolonged responses, it was necessary to leave a period of 60 min between ATP applications to evoke responses of similar amplitude and without sign of receptor desensitization. This basic feature of XlP2Y receptor activation is markedly different from the pattern of agonist activation of other recombinant P2Y receptors (e.g. P2Y_1-3) that have been expressed in oocytes, where responses are shorter (1-3 min) and reproducible within 20 min of the first agonist application (4, 5, 8). ATP responses at XlP2Y were gradually reduced, then abolished, in Ca²⁺-free conditions but returned when extracellular Ca²⁺ was restored, indicating XlP2Y mobilized Ca²⁺_i, which was replenished from an extracellular pool. ATP responses were reduced by thapsigargin (100 µM), which activates and then desensitizes IP₃ receptors and inhibited by the chloride channel (I_Cl,Ca) blocker 5-nitro-2-(3-phenylpropylamino)benzoic acid (60 µM), indicating XlP2Y activation elevated PLC- production of IP₃ to release Ca²⁺_i and open Ca²⁺-activated chloride channels that carried a biphasic current.

Half-maximal activation (in terms of the EC₅₀) by ATP required 103 nM for the first phase of current and 80 nM for the second slower phase (Fig. 6B). The Hill co-efficient (n_H) was approximately 1, indicating one molecule of ATP is necessary to activate the receptor. XlP2Y was also activated by all other naturally occurring nucleoside triphosphates (CTP, GTP, ITP, and UTP) but not by inorganic triphosphates and diphosphates (trisodium trimetaphosphate, pentasodium tripolyphosphate, and sodium pyrophosphate; all 100 µM), confirming the requirement for a nucleotide and not just a phosphate chain. The rank order of potency for nucleoside polyphosphates was (at 100 µM): ADPS > ATP = CTP = GTP = ITP = UTP > ATPS > ADP = ,-methylene ATP > AMP = 2-methylthio-ATP > ,-methylene ATP = 2,3-O-(4-benzoylbenzoyl)-ATP. XlP2Y was weakly stimulated by diadenosine polyphosphates and adenosine, where (at 100 µM): ATP > ADP > Ap₃A = Ap₄A > Ap₂A = Ap₅A > adenosine. On the basis of structure/activity relationship, XlP2Y required a nucleotide and either triphosphate chain or a diphosphate chain with a phosphorothioate or methylenephosphonate extension, but did not tolerate substitution at the C-2 position on the adenine base.

Oocytes expressing XlP2Y showed a persistent, strongly rectifying inward current (I_X), which reversed to outward current at 35 mV, then reverted to an inward current at 10 mV (Fig. 6, C and D). This persistent current was not found in control oocytes. The amplitude of I_X was inhibited by the P2 receptor antagonist suramin (1-100 µM) with an IC₅₀ value of 27 ± 4 µM (Fig. 6C), indicating a low level activation by endogenous ATP and suggesting I_X was, in part, I_Cl,Ca. Measurement of ATP release by the firefly assay showed a rate of release of 2 ± 0.25 nmol/h/oocyte. Since there is little enzymatic breakdown of extracellular ATP by oocytes (38), the basal efflux of ATP may continuously stimulate XlP2Y and elevate Ca²⁺_i to persistently activate I_Cl,Ca.

This persistently activated I_Cl,Ca had a significant impact on the resting membrane potential and input resistance of oocytes expressing XlP2Y receptors. In comparison to uninjected (control) oocytes, XlP2Y oocytes were depolarized by more than 30 mV and their input resistance lowered by as much as 4-fold (see Table I). The resting membrane potential (E_m) of XlP2Yoocytes lay close to the reversal potential for chloride ions (oocyte E_Cl = 24 mV) (39), indicating that an ATP-activated chloride conductance was a major factor in determining E_m. Suramin (100 µM) significantly increased E_m and input resistance of XlP2Y oocytes, blocking XlP2Y receptors and preventing their activation by a persistent ATP efflux. The electrical properties of XlP2Y oocytes in the presence of suramin closely matched E_m and input resistance of oocytes expressing other P2Y subtypes, including chick P2Y₁ and a P2Y₂-like subtype found in rat cortical astrocytes (40). Values for E_m and input resistance for oocytes expressing either rat P2X₃ or rat P2X₄ were significantly higher than XlP2Y oocytes, even in the presence of suramin (Table I). These P2X receptors show a higher concentration threshold for ATP activation and also desensitize rapidly; accordingly, the impact of ATP efflux appears to be negligible. Differences in the input resistance of oocytes expressing P2X subtypes and uninjected oocytes may reflect the damage caused to the membrane by the intracellular injection of P2 receptor transcripts.

Table I. Electrical properties of defolliculated Xenopus oocytes expressing XlP2Y and other P2 receptors Membrane potential and input resistance of defolliculated Xenopus oocytes injected with mRNA for XlP2Y, chicken P2Y1 (cP2Y1; X73268), poly(A+) RNA from rat astrocytes (rP2Y; see Ref. 40), rat P2X3 (rP2X3; X90651) and rat P2X4 (rP2X4; X91200). The electrical properties of oocytes expressing XlP2Y were determined in the absence and presence of suramin (100 µM). The properties of uninjected oocytes are also shown. (Data are given as mean ± S.E.; number of observations (n) is in parentheses.) Membrane potential and input resistance of defolliculated Xenopus oocytes injected with mRNA for XlP2Y, chicken P2Y1 (cP2Y1; X73268), poly(A+) RNA from rat astrocytes (rP2Y; see Ref. 40), rat P2X3 (rP2X3; X90651) and rat P2X4 (rP2X4; X91200). The electrical properties of oocytes expressing XlP2Y were determined in the absence and presence of suramin (100 µM). The properties of uninjected oocytes are also shown. (Data are given as mean ± S.E.; number of observations (n) is in parentheses.) Expressed receptor Membrane potential Input resistance mV megaohms Uninjected  55  ± 3  (24) 2.1  ± 0.32  (24) XlP2Y  25  ± 2  (21) 0.55  ± 0.07  (21) XlP2Y (+ suramin)  37  ± 1  (3) 1.0  ± 0.21  (3) cP2Y1  37  ± 2  (20) 0.93  ± 0.07  (20) rP2Y  36  ± 2  (20) 0.86  ± 0.11  (20) rP2X3  56  ± 3  (23) 1.2  ± 0.18  (23) rP2X4  47  ± 5  (13) 1.2  ± 0.07  (13)

DISCUSSION

We have isolated a cDNA for a G-protein-coupled receptor for extracellular nucleotides (P2Y receptor) that is expressed during early embryonic development in X. laevis. To our knowledge this is the first receptor of this class to be cloned in amphibians, and the first vertebrate P2Y shown to be expressed during early embryonic development.

From expression studies in defolliculated Xenopus oocytes, XlP2Y possessed several unique pharmacological features when compared with previously described recombinant P2Y_1-7 subtypes. The first major feature involved the duration of biphasic responses to agonists, some 40-60 min for the second phase with any of the naturally occurring nucleoside triphosphates. These biphasic responses were considerably longer than the 1-3 min observed following expression of P2Y₁, P2Y₂, and p2y3 in Xenopus oocytes (4, 8, 41, 42). Prolonged membrane currents evoked by XlP2Y activation were carried mainly by I_Cl,Ca, based on their sensitivity to 5-nitro-2-(3-phenylpropylamino)benzoic acid (43) and thapsigargin (44), and long term dependence on extracellular Ca²⁺ to help replenish Ca²⁺_i stores. The long duration of XlP2Y responses may not necessarily reflect the situation in the neural plate for a number of reasons, including receptor density, agonist concentration, and receptor/signaling cross-talk. We are currently investigating this issue in neural plate-derived cells.

A second unique feature of XlP2Y is the broad agonist selectivity, where all of the naturally occurring nucleoside triphosphates (ATP, CTP, GTP, ITP, and UTP) proved equally effective. None of the previously described (4-11) recombinant P2Y receptors (P2Y_1-7) are stimulated by all five nucleotides, while P2Y₂ is the only other P2 receptor to be stimulated by ATP and UTP equally (5). Since defolliculated oocytes are devoid of native P2Y or P2X receptors (31, 32), it is unlikely that stimulation of an endogenous receptor contributed to this broad selectivity. A third distinguishing feature of XlP2Y is a low level of continuous activation, probably by the basal efflux of ATP from oocytes, although a constitutive activation of XlP2Y without the need of an agonist cannot be ruled out. The rate of spontaneous release of ATP was 2 nmol/h, although the local concentration of ATP at the surface of the oocyte may be higher. We found the threshold for activation of a macroscopic whole-cell current was in the region of 10 nM, but small differences in the resting conductance of the oocyte membrane may occur at ATP concentrations lower than this level. Webb and colleagues (8) suggested that chick p2y3 expressed in Jurkat cells may be activated by ATP efflux from the host cell, since this P2Y receptor remains desensitized until an ecto-ATPase, apyrase, is added to the bathing medium. A similar desensitization was observed for bovine P2Y₁ expressed in Jurkat cells, and relaxed by the addition of apyrase to the bathing medium (45). In a similar vein, Nakamura and Strittmatter (46) found that human P2Y₁ expressed in oocytes is activated transiently by ATP efflux, which is augmented following the stimulation of stretch-activated mechanosensory ion channels. Thus, it may be the case that all recombinant P2 receptors are partially activated/desensitized in most expression systems, but the prolonged responses of XlP2Y make this feature more noticeable. The persistent activation of XlP2Y expressed in oocytes considerably depolarized these cells, and we are currently looking for a similar effect in neural plate-derived cells.

Increases in Ca²⁺_i are a common response to ATP stimulation in many cell types (12), and increases in Ca²⁺_i are thought to play important roles in regulating cell proliferation, migration and differentiation (15, 25). In the Urodele amphibian Pleurodeles waltl, reagents that cause an increase in Ca²⁺_i promote neural development in ectoderm isolated from early gastrulae. Treating gastrula ectoderm with either caffeine or ryanodine has been reported to cause a transitory (10-20 min) release of Ca²⁺ from intracellular stores, the ectoderm subsequently differentiating neurons and glia (47). Similarly, reagents such as the lectin concanavalin A and phorbol esters, which can induce neural development in amphibian ectoderm (48, 49), also increase Ca²⁺_i levels (47). In contrast, preloading gastrula ectoderm with the Ca²⁺ chelator, BAPTA, suppresses neural differentiation in response to dorsal mesoderm, the source of endogenous neuralizing signals (47). These results demonstrate the potential importance of signaling pathways that control Ca²⁺_i levels in neural development. It is therefore of great interest that XlP2Y is expressed in the developing neural plate, and that it can induce prolonged (40-60 min) cellular responses, including increases in Ca²⁺_i.

In amphibians, the neural plate forms in the dorsal ectoderm as a result of inductive signals released by the underlying dorsal mesoderm, a process initiated during gastrulation (50). The onset of XlP2Y transcription during the later stages of gastrulation suggests that it is highly unlikely that this receptor mediates a primary neuralizing signal. Instead, it suggests that expression of XlP2Y is an early response to these signals, and that it might participate in secondary neuralizing signals responsible for establishing different cell identities within the neural plate, the birth of primary neurons, the maintenance of neuronal fates, and the morphogenetic movements whereby the neural plate folds about the midline to form the neural tube. In Xenopus embryos the first neurons do not differentiate until after the neural tube is formed (51, 52), and it may be significant that this coincides with a significant reduction in XlP2Y transcripts. ATP may not be the only neurotransmitter involved in neural development during neural plate stages in Xenopus. Inhibition of dopamine -hydroxylase, the enzyme catalyzing the conversion of dopamine to noradrenaline, during neural plate stages results in a substantial reduction in the number of neurons that differentiate in culture, as do antagonists of the -adrenergic receptor (17). In contrast, addition of noradrenaline or dopamine increases the number of neurons that differentiated in similar cultures. This suggests that endogenous noradrenaline is part of the mechanism controlling neuronal differentiation in the central nervous system. It will be interesting to see what effects, if any, extracellular ATP has on the differentiation of neurons.