The Extracellular Adenosine Deaminase Growth Factor, ADGF/CECR1, Plays a Role in Xenopus Embryogenesis via the Adenosine/P1 Receptor*

Adenosine deaminase-related growth factors (ADGF), also known as CECR1 in vertebrates, are a novel family of growth factors with sequence similarity to classical cellular adenosine deaminase. Although genes for ADGF/CECR1 have been identified in both invertebrates as well as vertebrates, their in vivo functions in vertebrates remain unknown. We isolated cDNA clones for two cerc 1s from Xenopus laevis. Both recombinant Xenopus CECR1s exhibited adenosine deaminase and growth factor activity, and the adenosine deaminase activity was found to be indispensable for growth factor activity. The Xenopus cerc 1s are expressed in the somites, pronephros, eyes, cement gland, neural tube, and neural floor plate of the embryos. Knock-down of these two genes using morpholino oligonucleotides caused a reduction in the body size and abnormalities of the body axis in the Xenopus embryos, accompanied by selective changes in the expression of developmental marker genes. Injection of adenosine, agonists for adenosine/P1 receptors, or adenosine deaminase inhibitor into late gastrula archenteron embryos resulted in developmental defects similar to those caused by morpholino oligonucleotide injection. These results show, for the first time, the involvement of CECR1s via the adenosine/P1 receptors in vertebrate embryogenesis via regulation of extracellular adenosine concentrations.

the genomes of other vertebrate species, such as chimp, baboon, pig, chicken, frog, zebrafish, and pufferfish (2), indicating that the ADGF/CECR1 genes represent a family of evolutionally conserved genes. Interestingly, no obvious CECR1 homologue has yet been identified in rodents (20).
ADA activity has been reported for some ADGF/CECRs (9 -14), including IDGF (3), where the K m , V max , and k cat values are comparable with those of mammalian cytoplasmic ADAs (21). As IDGF was originally identified as a growth factor in an embryonic cell line (1) and ADA activity was shown to be indispensable for growth factor activity (3), it is plausible that the ADGF/CECR1 family plays a role in cell proliferation during morphogenesis and organogenesis via its ADA activity. In Drosophila, the lack of Adgf-a leads to precocious metamorphic changes during metamorphosis, including the differentiation of macrophage-like cells, fat body disintegration and developmental delay (16,22). Transgenic expression of human CECR1 in mice results in abnormal heart and kidney development, suggesting a possible role in mammalian development (23). The in vivo function of the CECR1 family in vertebrates, however, remains unclear, as there is no known CECR1 homologue in mice.
In the present study, we analyzed the molecular properties and in vivo functions of Xenopus CECR1s to increase our understanding of the roles of the ADGF/CECR1 family in vertebrates. The results demonstrate a key role of Xenopus CECR1s in embryonic development via their ADA activity.

EXPERIMENTAL PROCEDURES
Embryos and Their Culture-Xenopus laevis embryos were collected and cultured as described previously (24). The embryos were staged according to the method described by Nieuwkoop and Faber (25).
cDNA Cloning for Xenopus cecr1s-To isolate Xenopus cecr1 cDNA, PCR was performed using Xenopus cDNA library (embryo, stage 30, Stratagene). The nucleotide sequences of the degenerated primers used for the PCR were 5Ј-GT(G/T)GGI(C/A)(G/A)IGA(A/G)GA(C/T)ACIGG-3Ј and 5Ј-CCA(A/G)TCNAT(T/C)TCICCIGC (G/A)TG(G/ A)AA-3Ј, which correspond to 325 VGHEDTG 331 and 355 FHAGETDW 362 , respectively. A 114-bp cDNA fragment was amplified and used as a probe for subsequent cDNA cloning. Two hybridization-positive clones were isolated from 1.3 ϫ 10 5 clones of a Xenopus embryo (stage 30) cDNA library (Stratagene) by plaque hybridization, and their sequences determined.
Preparation of Recombinant Xl_CECR1s-The recombinant Xl_CECR1-1 and Xl_CECR1-2 proteins were produced in Escherichia coli BL21 star (DE3) using a pET100 directional TOPO expression kit (Invitrogen) as well as in Xenopus A6 cells (American Type Culture Collection). In the first method, the coding regions of Xl_cecr1-1 and -2 were amplified by PCR and subcloned into a pET100/D-TOPO vector. His-tagged fusion proteins were then produced in E. coli. Purification was performed using a HiTrap Chelating HP column (GE Healthcare BioSciences). In the second method, the amplified coding regions of Xl_cecr1-1 and -2 were subcloned and transfected to Xenopus A6 cells with FuGENE 6 transfection reagent (Roche Diagnostics). The recombinant proteins were used for assaying ADA and growth factor activities.
Preparation of Antibodies against Xl_CECR1s-To prepare antibodies against Xl_CECR1-1 and Xl_CECR1-2, the synthetic peptides CEDLAMNPEASEREL and CKFIKDLAMN-WKKEL, which are specific to Xl_CECR1-1 and Xl_CECR1-2, respectively, were conjugated to keyhole limpet hemocyanin and injected separately into an albino rabbit with Complete Freund adjuvant. After three booster injections with Incomplete Freund adjuvant, each antiserum was affinity-purified using the recombinant Xl_CECR1-1 or -2, produced in E. coli.
Assay of the ADA Activity of Xl_CECR1s-The ADA activity of Xl_CECR1s was assayed as described by Iwaki-Egawa et al. (29). Enzymatic activity was determined by quantifying inosine liberated from adenosine. Inosine is converted by purine nucleoside phosphorylase into hypoxanthine; the latter is then converted into uric acid and diformazan by xanthine oxidase and nitro blue tetrazolium. Thus, the optical density at 540 nm of diformazan was measured. In brief, the recombinant Xl_CECR1s were preincubated for 5 min at 37°C in a reaction mixture containing 20 mM sodium phosphate buffer (pH 6.8), 1 unit/ml xanthine oxidase, 0.24 unit/ml purine nucleoside phosphorylase, and 1 mM nitro blue tetrazolium. Following incubation, the ADA assay was initiated by adding adenosine, and the mixture was further incubated at 37°C for 15 min. The reaction was stopped by adding 1/4 volume of 50% acetic acid. The absorption at 540 nM diformazan was then measured. One unit of ADA activity was defined as the amount of enzyme that produced 1 mol of inosine/min.
Assay of Growth Factor Activity-The growth factor activity assay was performed as described previously by assessing cell proliferation (7). NIH-Sape-4 cells were cultured in M-M medium (30) at 25°C, as described previously (31). Briefly, NIH-Sape-4 cells were suspended at a density of 2.5 ϫ 10 5 cells/ml in M-M medium, and then 100 l of the cell suspension was poured into each well of a 96-well microtiter plate. Cell proliferation was assessed by measuring the enhancement of cellular metabolism using the Alamar Blue assay, after the addition of 1 ng of recombinant Xl_CECR1-1 and 10 ng of Xl_CECR1-2 (32,33). The cells were then incubated for 2 days at 25°C in the presence or absence of the test sample. Following this, 15 l of Alamar Blue (Alamar Biosciences, Inc.) was added, and changes in the color of the culture medium were determined by measuring A 570 and A 610 . One unit of growth factor activity was defined as the amount giving half-maximal stimulation of cell proliferation.
Whole Mount in Situ Hybridization-Whole mount in situ hybridization of Xenopus embryos was performed as described previously (34). The Xl_cecr1 probe consisted of digoxigeninlabeled RNA corresponding to a 605-bp fragment (ϩ904 to ϩ1508) of Xl_cecr1-2 cDNA. The Xmyod probe consisted of digoxigenin-labeled RNA corresponding to a 350-bp fragment (ϩ157 to ϩ506) of Xmyod cDNA (35). The Xa2a probe consisted of digoxigenin-labeled RNA corresponding to a 457-bp fragment (ϩ479 to ϩ935) of Xa2a cDNA clone (MGC:85585) (36). The Xmyf-5 probe consisted of digoxigenin-labeled RNA corresponding to a 405-bp fragment (ϩ29 to ϩ433) of Xmyf-5 cDNA clone (GenBank TM accession number X56738) (35). Semiquantitative RT-PCR Using Primers for Various Developmental Marker Genes-Total cellular RNA was isolated from each embryo using TRIzol reagent (Invitrogen). Total RNA (1.4 g) was treated with RNase-free DNase I (Invitrogen) and reverse-transcribed using SuperScript III (Invitrogen) with oligo(dT) primers. The ornithine decarboxylase transcript was amplified to normalize the cDNA content of the samples. Equal amounts of cDNA were subsequently used for the amplification of marker genes with gene-specific primers. The linear range of amplification was determined for each primer set. For semiquantitative determination of relative gene expression levels, 10-l samples of the amplificates were collected during PCR after the completion of three different cycle numbers in the linear range. The amplicons were analyzed on a 1% agarose gel stained with ethidium bromide. The primers (5Ј to 3Ј) and cycle numbers used were: Xmyf5 (35) Other Methods-SDS-PAGE was performed using the method of Laemmli (49). Immunoblotting was performed as described by Luttrell et al. (50).

Identification and Characterization of X. laevis cecr1 cDNAs-
To investigate the role of the ADGF/CECR1 family in vertebrates, we attempted to isolate cDNA (s) for the X. laevis homologue of CECR1. For this, we first amplified a cDNA fragment of the Xenopus cecr1 homologue using degenerate primers designed according to the amino acid sequences conserved among human, pig, and zebrafish CECR1s; this was due to the lack of information regarding the nucleotide sequence of the Xenopus homologue of CECR1 on commencement of this study. We identified two cDNAs that encode relevant but distinct proteins, named Xl_CECR1-1 and Xl_CECR1-2. Comparison of the deduced amino acid sequence of Xl_CECR1-1 and -2 revealed that they shared ϳ40% sequence identities with Sarcophaga IDGF and human CECR1, respectively (Fig. 1A). Based on their high sequence identities (91%), Xl_CECR1-1 and -2 were considered as isoforms. The Xl_CECR1-2 nucleotide sequence was almost identical to that of a clone (accession number AY902778) deposited in the NCBI data base. Both Xl_CECR1-1 and -2 contained putative signal sequences in their N termini, suggesting that they are secretory ADGFs (51). Phylogenetic tree analysis revealed that Xl_CECR1-1 and -2 are members of the CECR1 family and that their closest relative is the chicken Gg_CECR1 (Fig. 1B). Consensus amino acids in the active site of cellular classical ADA were conserved in many CECR1s including Xl_CECR1-1 and -2 except for Val 431 (Fig. 1C). We previously reported that Asp 462 and His 405 in Sarcophaga IDGF were essential for ADA and growth factor activity (3). These two residues are also conserved in Xl_CECR1-1 and -2 (Fig. 1A). We performed BLAST search using the Xenopus tropicalis genome data base and confirmed that the genome (scaffold 341) also encodes the orthologues for both Xl_cecr1-1 and -2, and that there is no other gene more closely related to the human CECR1 gene than the X. tropicalis gene. Thus, we concluded that the cloned cDNAs are X. laevis homologues of human CECR1.
Characterization of ADA and Growth Factor Activities of Xl_ CECR1-1 and -2-We next analyzed the enzymatic properties of Xl_CECR1s and the relationship between their ADA and growth factor activities. For this, recombinant Xl_CECR1-1 and -2 proteins were expressed in Xenopus A6 cells, respectively. The conditioned medium containing Xl_CECR1-1 or -2 as the major protein was used for the experiments. Both Xl_CECR1-1 and -2 showed ADA activities with K m values of 260 and 230 M, respectively, for adenosine as a substrate. The K m values of the Xl_CECR1s were in the intermediate range between those of invertebrate (ϳ50 M) (3) and human CECR1/ADA2 (ϳ2 mM) (14). The enzymatic activities of Xl_CECR1-1 and -2 were inhibited with DCF, a potent inhibitor for cellular and extracellular ADAs, but not with erythro-9-(2-hydroxy-3-nonyl)-adenine hydrochloride (EHNA), a selective cellular ADA inhibitor (Fig. 2). This suggests that Xl_CECR1 has the enzymatic profile of an extracellular ADA. 2Ј-Deoxyadenosine was a less effective substrate than adenosine, and 5Ј-AMP was not a substrate for either Xl_CECR1-1 and -2, unlike in classical ADAs (Table 1) (21). These results indicate that the enzymatic activities of Xl_CECR1s share similar substrate and inhibitor specificities with human CECR1/ADA2 (52), although the K m is lower than that for human CECR1/ADA2. Next, we examined the growth factor activities of Xl_ CECR1-1 and Xl_CECR1-2 in NIH-Sape-4 cells, as these cells are sensitive to various invertebrate ADGFs (1,3,10,11,15). Although both Xl_CECR1-1 and Xl_CECR1-2 showed growth factor activity, Xl_CECR1-1 displayed activity at lower concentrations than Xl_CECR1-2 ( Fig. 3), consistent with the 10 times higher specific ADA activity of Xl_CECR1-1 than Xl_CECR1-2 (data not shown). To determine the relationship between ADA and growth factor activity, we examined the effects of DCF and EHNA on this activity. The growth factor activities of both Xl_CECR1-1 and -2 were inhibited by DCF, but not by EHNA (Fig. 3). These results demonstrate that the ADA activity of Xenopus CECR1s is indispensable for growth factor activity.
Analysis of the Xl_cecr1s Gene Expression during Embryogenesis-To examine the expression of the Xl_cecr1s during embryogenesis, RT-PCR was performed using RNAs from various developmental stages with primers designed to detect both Xl_cecr1-1 and -2 mRNAs. Expression of Xl_cecr1 started at stage 20 (neurula stage), gradually increased during the neurula stage, and was sustained to stage 44 (Fig. 4). There were no differences in the expression patterns between Xl_cecr1-1 and Xl_cecr1-2 in RT-PCR using gene-specific primers (data not shown).
To determine the spatial expression pattern of the Xl_cecr1 genes, we performed whole mount in situ hybridization. The probe was a 605-bp fragment of Xl_cecr1-2 cDNA with 94% homology with the corresponding sequence of Xl_cecr1-1, which suggests cross-hybridization with the Xl_cecr1-1 transcript. The Xl_cecr1 were expressed in the somites at stage 20 (neurula stage, Fig. 5A). In addition, the cement gland was also stained at stage 22 (neurula stage, Fig. 5B). At stage 30 (tail-bud    stage), the Xl_cecr1s were clearly expressed in the somites, pronephros, eye, cement gland, neural tube, and neural floor plate (Fig. 5C). At stage 36, the expression patterns were similar to those of the tail-bud stage (stage 30) (Fig. 5D). Expression in the pronephros was maintained at stage 39, but expression in the other tissues diminished (data not shown). As ADA activity is indispensable for the growth factor activity of IDGF (3, 15), we assumed that Xl_CECR1s played a role in embryogenesis through adenosine/P1 receptor(s) via regulation of extracellular adenosine levels. We therefore examined the expression profiles of Xenopus homologues of an isoform of the adenosine/P1 receptor (adenosine A2A receptor, XA2A) gene. We found that Xa2a was expressed in somitic stripes and eyes in the tail-bud stage, similar to the Xl_cecr1, suggesting that Xa2a is involved in the regulation of Xl_cecr1s (Fig. 5, F and G). Analysis of the in Vivo Functions of the Xl_CECR1s during Embryogenesis-To analyze the in vivo functions of Xl_cecr1 in embryogenesis, we used MOs to knock down Xl_cecr1. We designed two MOs specifically to target sequences surrounding the initiation codon of each of the Xl_cecr-1 and -2, and microinjected both of them at the two-cell stage to knock down both Xl_CECR1-1 and -2. Antisense MO injection significantly reduced the amounts of both Xl_CECR1-1 and -2 as shown by immunoblotting (Fig. 6A). In contrast, there was no significant change with control MOs containing four mismatched pairs, suggesting a specific effect of these MOs. Up to the neurula stage, there were no apparent morphological differences between antisense and control MO-injected embryos, supporting our data showing that the expression of Xl_cecr1 starts at the neurula stage (Fig. 4). Developmental abnormalities were observed in MO-injected embryos at tail-bud stages. There was an intermediate phenotype with vertical or lateral bending (25%, Fig. 6, B and C), and an abnormal severe phenotype accompanied by both reduction in size and body axis bending (13%, n ϭ 24, Fig. 6D). Conversely, embryos injected with control MO displayed no significant abnormalities (95%, n ϭ 22, Fig. 6E). In situ hybridization experiments revealed that the expression of a gene for a basic helix loop helix transcription factor, Xmyod, was missing in some somitic stripes in MO-injected tail-bud embryos (Fig. 6, F-I), which may have contributed to the reductions in size and the abnormal body axis bending. We tested a total of 24 MO-injected embryos. Among them, 8 embryos showed a marked reduction in Xmyod expression in somites at stage 20. The phenotypes ranged from a severe one showing marked reduction in Xmyod expression in most of the somites to a mild one showing moderate reduction in Xmyod expression in a few somites. There was no tendency for the abnormal phenotype to appear in a specific region of the somite. The abnormal phenotype showing reduction in Xmyod expression was not detected in MO-injected embryos of stage 15 (data not shown), in which cecr1 mRNA had not yet been expressed, and was consistent with our notion that the phenotype is due to the knocking down of CECR1. To further confirm that the bending and truncated phenotypes observed in embryos injected with the first pair of MOs are due to the knocking down of the CECR1. As a result, similar phenotypes were observed as in the case of the first pair of MOs, and all of the MO-injected embryos showed bending and/or truncated phenotypes (n ϭ 14, data not shown).
Analysis of Developmental Marker Gene Expression in MOinjected Embryos-We next examined whether the expression of typical developmental marker genes were affected in the MO-injected embryos (Fig. 7). Of the somite marker genes (Xmyf-5, Xmyod, cardiac actin, and six1), the gene expression of Xmyf-5 (a marker of the most dorsal and ventral tips of the somites) was significantly increased (over 16-fold compared with controls), suggesting that Xl_cecr1s are involved in sup-  pressing Xmyf-5 expression. Although Xmyod gene expression in somitic stripes was ablated partially by MOs (Fig. 6, F and H), total expression levels were not affected. We also examined markers for pronephron, eye, cement gland, and hatching gland; dramatic changes were observed in the gene expression of xsmp-30, which is a Xenopus homologue of SMP-30 (senes-cence marker protein-30) expressed in pronephric proximal tubules. Its expression decreased 32-fold in MO-injected embryos compared with controls, whereas there were no significant changes in the other pronephric marker genes (XNKCC2 (distal tubules), xwt1 (glomus), and Xlim1 (pronephros, central nervous system, and tail bud)). Of the eye marker genes, ␥-crystallin, Xpitx3, and darmin-r, the gene expression of ␥-crystallin decreased 16-fold in the MO-injected embryos. Expression of the marker genes for the cement gland and the hatching gland (xa and xag, respectively) was not affected significantly, consistent with the fact that the cement gland was visibly formed even in severely affected embryos (Fig. 6, B-D). We examined whether the expression of Xa2a was affected in MO-injected embryos. The amounts of Xa2a transcript increased more than 16-fold compared with controls, suggesting a role of the adenosine/P1 receptor in Xl_CECR1 functioning in Xenopus embryogenesis. We further examined the effect of MOs on the expression pattern of the somite marker Xmyf-5 gene at stage 35/36. In the wild-type tail-bud embryo, the Xmyf-5 was expressed in the most ventral and dorsal tips of posterior somites and in the tail portion (Fig. 8, B and D) (53). On the contrary, in MO-injected embryos, the Xmyf-5 expres-  Relative gene expression levels were determined by RT-PCR. Amplification of ornithine decarboxylase was used to normalize the cDNA content of the samples. Xmyf-5, Xmyod, cardiac actin, and six1 are marker genes of somites. Xsmp-30, XNKCC2, xwt1, and Xlim-1 are marker genes of pronephros. ␥-Crystallin, Xpitx3, and darmin-r are marker genes for the eye. xa and xag are marker genes for the cement gland. Xa2a is a gene of the Xenopus homologue of the adenosine A2A receptor.
sion diminished the most in the tips of the somites (Fig. 8C). Instead, the Xmyf-5 expression was markedly up-regulated in the tail part (Fig. 8C), which may have contributed to the enhancement of the overall amount of Xmyf-5 expression detected in RT-PCR analysis (Fig. 7).
Analysis of the Mechanism of Action of Xl_CECR1s and XA2A in Xenopus Embryogenesis-Next, to determine whether the adenosine/P1 receptor influences Xl_CECR1 activity, we examined the effects of adenosine, an ADA inhibitor (DCF) and adenosine/P1 receptor agonists (CGS21680, 5Ј-N-ethylcarboxamido adenosine, or cyclopentyl adenosine) on embryogenesis by injecting them into late gastrula stage 12 embryo archenteron. Injected embryos were incubated up to stage 35/36 and their morphologies were examined. DCF caused an abnormal phenotype accompanied by reductions in both size and body axis bending, suggesting that the ADA activity of Xl_CECR1s modulates their activity (Fig. 9A). Injection of adenosine, or adenosine agonists also caused abnormal phenotypes, similar to DCF, characterized by reduction in size and abnormal body axis bending (Fig. 9, B-E). We also examined the effect of the adenosine/P1 receptor agonists on the expression of the somite marker gene, Xmyod, at stage 20. In situ hybridization revealed that the Xmyod expression in somitic stripes decreased in embryos injected with either the adenosine receptor agonist, CGS-21680 hydrochloride (Fig. 9G) or cyclopentyl adenosine (data not shown). These phenotypes were similar to those obtained with the injection of cecr1 MOs, which is characterized by the ablated Xmyod expression in somitic stripes (Fig. 6, F and H). These results are consistent with our notion that Xl_cecr1s function in Xenopus embryogenesis via the adenosine/P1 receptor by regulating extracellular adenosine levels.

DISCUSSION
In the current study, we first identified two Xenopus CECR1 homologues and showed that they catalyze adenosine deamination with K m values (260 M and 230 M) 10 times lower than those of human CECR1/ADA2 (K m ϭ 2.53 mM). Thus there might be altered levels of enzymatic activity, possibly associated with the molecular evolution from amphibian to mammal. It is noteworthy that the ADA activity of Xl_CECR1s was also linked to growth factor activity, as shown for invertebrate ADGFs (3, 9 -13). The differences in the specific growth factor activities between Xl_CECR1-1 and -2 may reflect differences in their specific ADA activities. Thus, it is plausible that Xl_CECR1s play some role in growth regulation during development via extracellular adenosine deamination.
The two Xenopus cecr1 isoforms showed similar expression profiles during embryogenesis, suggesting that they have cooperative roles. The most frequent phenotypes in the knockdown experiments were a reduction in size and abnormal bending of the body axis, which might be explained by selective changes of cell populations caused by the lack of growth factors. Given the early expression in somites of these growth factors, these results suggest that Xenopus cecr1s are involved in somite formation and differentiation during embryogenesis. This was supported by the significant increase in the amount of Xmyf-5 mRNA and the ablated expression of Xmyod mRNA in the MO-injected embryos.
The initial activation of muscle-specific genes is observed during gastrulation and concerns the two myogenetic regulatory factors, XMyf-5 and XMyoD, whose transcripts are detected at stages 9.5 and 10.5, respectively. This is quickly followed by the accumulation of transcripts of cardiac actin, the first muscle structural gene expressed in vertebrates (54). Expression of the Xl_cecr1 started at stage 20 (Fig. 4), and local elevation of adenosine in MO-injected embryos might have disturbed the function of Xmyf-5 and Xmyod present in somites thorough adenosine/P1 receptor signaling, resulting in developmental defects in the somites. As the gene expression of the Xl_cecr1s and the adenosine/P1 receptor (Xa2a) overlaps in some cell populations in somitic strips, it might be possible that the CECR1 signaling and the adenosine/P1 receptor compose an autocrine feedback loop in these cell populations to respond to a local change in the concentration of the ligand (adenosine) metabolized by the extracellular enzyme (CECR1).
We showed somitic expression of Xa2a (a Xenopus homologue of adenosine A2A receptor) and the augmentation of gene expression in MO-injected embryos (Figs. 5, F and G, and  7). The mode of action of Xl_CECR1s during development might be explained by adenosine/P1 receptor-mediated signaling. The results of the experiments using adenosine/P1 receptor agonists suggest that adenosine/P1 receptors mediate the signal from CECR1, probably by changing the sensitivity to the ligand, adenosine, in response to a local decrease in adenosine concentration catalyzed by CECR1 (summarized in Fig. 10). This notion is consistent with the fact that human CECR1 has a binding site on the cell surface near the adenosine/P1 receptor. Human CECR1/ADA2 interacts with heparan sulfate proteoglycans and/or receptors on the cell surface to stimulate the proliferation of immune cells and tumor cells (14). The local decrease of extracellular adenosine concentration may modulate the selectivity of the receptor subtypes.
The adenosine/P1 receptor family comprises the A1, A2A, A2B, and A3 adenosine receptors, which have been cloned from various species (55). A1, A2A, A2B, and A3 adenosine receptors have distinct but frequently overlapping tissue distributions. The concentration of extracellular adenosine regulated by CECR1 might be a crucial determinant of the differential activation of coexisting adenosine/P1 receptors under different physical conditions. For example, it has been shown that low concentrations of adenosine activate the A1 adenosine receptor expressed on neutrophils, whereas high concentrations of adenosine, acting on A2 adenosine receptors, inhibit the effects produced by A1 occupancy (55). The increased expression of the Xa2a in the MO-injected embryos may reflect such a switch caused by excess amounts of extracellular adenosine. Interestingly, microinjection of adenosine receptor agonists for A1, A2B, and A3 resulted in developmental defects similar to those of A2A, suggesting the involvement in embryogenesis of other subtypes apart from the adenosine A2A receptor. The pheno-types resulting from Adgf-a (a CECR1 family in invertebrates) deficiency in Drosophila are partially explained by adenosine/P1 receptor signaling and elevations of extracellular adenosine levels (22). Similarly, adenosine concentrations could also be up-regulated in MO-injected embryos as in the Drosophila mutants (22).
We found a dramatic decrease in the number of xsmp-30 transcripts in the MO-injected embryos. Other marker genes, including that for pronephric distal tubules, XNKCC2, were unaffected in the same embryo, indicating that pronephric proximal tubules specifically failed to fully develop in the Xenopus CECR1s-deprived embryos. It was recently reported that SMP-30 acts as a survival factor in hepatocytes by regulating Akt activity (56). It might also be possible for SMP-30 to function as a survival factor during pronephros development via Akt, with expression regulated by X1_CECR1s.
Overexpression of human CECR1 is suggested to be responsible for some of the features of cat eye syndrome, such as congenital ocular coloboma and kidney defects (17,20). In contrast, as shown in our experiments, knockdown of Xenopus CECR1s resulted in a reduction in size and abnormal bending of the body axis, suggesting an important role of Xenopus CECR1 in normal development. Thus, expression of the CECR1 should be tightly regulated, possibly to maintain the appropriate adenosine levels required for normal development.
What is the physiological function of extracellular adenosine in Xenopus embryogenesis? Our results suggest that an optimum level of extracellular adenosine is needed for normal development. ATP and the reaction products (ADP/AMP) are known to be enzymatically hydrolyzed by ectonucleotidases in the extracellular space, and the physiological significance of this has been previously discussed (55,57). Interestingly, circadian-regulated 5Ј-AMP seems to be the key signal that mediates murine procolipase expression in peripheral organs and induces torpor, a hibernation-like state (58). In addition, a certain level of extracellular adenosine functions as a synaptic, angiogenic, and vasculogenic modulator (55, 59 -61). Thus, extracellular levels of adenosine or purine nucleotides are regulated both seasonally and developmentally. Our findings stress the important role of a novel group of extracellular ADAs linked with growth factor activity in the local metabolism of extracellular adenosine, which act via adenosine/P1 receptormediated intracellular signaling during embryogenesis.