Genomic structure, developmental distribution and functional properties of the chicken P2X₅ receptor

Materials

Solutions for electrophysiology were made with highly purified water (NANOpure, Barnstead, Debuque, IA, USA) using chemicals of analytical grade. ATP (potassium salt); α,β-methylene-ATP (αβmeATP; lithium salt) and 2-methylthio ATP (2meSATP; sodium salt) were purchased from Sigma (St Louis, MO, USA). Suramin was purchased from RBI (Nactick, MA, USA) and pyridoxal 5-phosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS) was obtained from Tocris Cookson (Bristol, UK). Solutions containing ATP or other nucleotides were prepared freshly each day and the pH of the solution was readjusted.

cDNA cloning and genomic screening

A partial cDNA for the cP2X₅ receptor (∼400 bp) was obtained using degenerate PCR as described (Soto et al. 1996) and used to screen a 17-day-old chicken embryo brain cDNA library (Stratagene, La Jolla, CA, USA) (Ruppelt et al. 1999). One positive phage was isolated and cloned. Protein-sequence analysis revealed a cDNA encompassing the 3′-coding sequence of cP2X₅ up to amino acid 257. To isolate the unknown 5′-coding sequence of the cP2X₅ mRNA, we used RACE techniques (Marathon cDNA amplification kit, Clontech, Palo Alto, CA, USA). Specific antisense oligonucleotide (5′-ATCCAGACGGCTAAAAGAAT-3′) targeted to cP2X₅ was used for PCR on a library of adaptor-ligated 14-day-old embryo heart cDNA following manufacturer's recommendations. A clone that contained an inframe stop codon in the 5′-untranslated sequence was joined to the previously isolated partial cDNA using an internal EcoRI site to form the full-length cP2X₅ cDNA. In order to determine if additional C-terminal domains with higher homology to rP2X₅ were present in chicken embryos, 3′-RACE was performed on adaptor-ligated cDNA libraries synthesized by using mRNA isolated from 14-day-old embryo brain and skeletal muscle and adult brain as described above. The specific sense oligonucleotides used for nested PCR were: 5′-CGAAACCTCCCACCACACTATCCT-3′ and 5′-CCGTTCCGTCCCTTTAAATTGTAGT-3′.

To screen a commercially available chicken genomic library (Clontech, Palo Alto, CA, USA) a 1042-bp probe (nt 248–1289) was used. Fourteen positive phages were isolated. One 17-kb long phage (λ71) that hybridized with probes containing N- and C-terminal sequences of cP2X₅ cDNA was used for further analysis. The isolated λDNA was digested with XhoI/Asp718I (containing exon 1–2), Asp718I/HindIII (exon 3–5), HindIII (exon 6–12), cloned in pBluescript II KS(+) and sequenced. Comparison of the genomic sequence with the cDNA sequence was used to determine the exon–intron junctions.

All cDNA clones were sequenced in both strands using a Dye Terminator Cycle Sequencing Kit and an ABI377 DNA sequencer (Applied Biosystems, Foster City, CA, USA).

RNAse protection analysis

In the RNAse protection assay, a 575-bp fragment containing the last 439 bp of the cP2X₅ coding sequence plus 23 bp of 3′-untranslated sequence (nt 715–1289) was used to generate a [³²P]UTP-labeled antisense RNA probe using T7-RNA polymerase. Twenty micrograms of total RNA obtained as previously described (Chomczynski and Sacchi 1987) were hybridized with the probe and unprotected RNA was digested with RNAse A and RNAse T1 following the manufacturer's instructions (Ribonuclease protection assay RPA III, Ambion, Austin, TX, USA). DNA ladder (100 bp, Gibco BRL, Karlsruhe, Germany) was [³²P]CTP-labeled using Terminal transferase (Boehringer Mannheim, Mannheim, Germany) following the protocol described by the manufacturer.

In situ hybridization

For whole mount in situ hybridization, 450 bp of cP2X₅ C-terminal sequence (nt 825–1274) or 506 bp of extracellular domain sequence (nt 370–875) were used to generate antisense digoxigenin-labeled RNA probes using the Boehringer Mannheim kit (Boehringer Mannheim, Mannheim, Germany). The hybridization pattern obtained with both antisense probes was essentially identical. The corresponding sense probes were used as negative control. Embryos corresponding to Hamburger and Hamilton developmental stages 11 (HH11), 15 (HH15) and 18 (HH18) (Hamburger and Hamilton 1951) were fixed overnight in 4% paraformaldehyde. Whole mount in situ hybridization was performed as described (Pera and Kessel 1997).

In situ hybridization was performed on tissue slices of embryonic skeletal muscle and on heart cryostat sections using synthetic antisense and sense oligonucleotide probes. The oligonucleotides used were: antisense 5′-AACCCTTCCAGGACACCTGCCCCAT TCTCCCAGCTCTCTTTA-3′ or 5′-TCTTTCAAACAACGGCCA GTCTTAACCCCATTACCAGCCACCAC-3′; sense: 5′-GTGGTG GCTGGTAATGGGGTTAAGACTGGCCGTTGTTTGAAAGA-3′. Tissue preparation, hybridization, dipping and counterstaining were performed as previously described (Stocker and Pedarzani 2000). The following controls were performed: (i) adjacent sections were hybridized with sense oligonucleotide, or (ii) hybridized with a mixed oligonucleotide probe containing a 500-fold excess of cold oligonucleotide. The silver grain densities observed in control sections were considered as background.

Stable transfection of HEK-293 cells and electrophysiological measurements

The full-length cP2X₅ cDNA was subcloned into the mammalian expression vector pcDNA3 (Invitrogen, Groningen, the Netherlands). Human embryonic kidney (HEK-293) cells were stably transfected using a modification of the calcium phosphate precipitation method (Chen and Okoyama 1987). Independent foci were selected in the presence of 1 mg/mL and expanded in the presence of 0.5 mg/mL geneticin (Sigma, St Louis, MO, USA). The cells were grown as monolayer cultures in 1 : 1 mixture of Dulbecco's modified Eagle's medium and Ham's F12 nutrient mix (GibcoBRL, Rockville, MD, USA) supplemented with 10% fetal bovine serum and 0.5 mg/mL geneticin at 37^°C, 5% CO₂. For electrophysiological measurements, cells from subconfluent cultures were plated onto glass coverslips 24–48 h before use and maintained in culture medium containing 3% fetal bovine serum.

Membrane currents were recorded using the standard whole-cell configuration of the patch-clamp technique (Sakmann and Neher 1995) as previously described (Korngreen et al. 1998). Once the whole-cell configuration was established, the cell was continuously superfused with the desired extracellular solutions via a gravity-fed perfusion system. Experiments were initiated at least 2 min after establishing the whole-cell configuration to allow equilibration of the cytosol with the pipette solution. Membrane currents were recorded under voltage-clamp using an Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA, USA) and stored on VCR tape for subsequent analysis (VR-10B, Instrutech, Port Washington, NY, USA). The sampling frequency was set to at least two times the corner frequency of the low-pass filter. The standard extracellular solution contained (mm): 150 CsCl, 1.5 CaCl₂, 5 TES, 10 d-glucose, adjusted to pH 7.4 with CsOH. The standard pipette solution contained (mm): 155 CsCl, 5 TES, 0.1 EGTA, 0.096 CaCl₂ (0.1 µm Ca²⁺, estimated by Max Chelator 6.0), adjusted to pH 7.2 with CsOH. Deviations from these solutions are noted in the figure legends.

Sigma Plot™ 2.0 scientific software (Jandel Scientific, San Rafael, CA, USA) was used for non-linear curve fitting.

Isolation and analysis of cP2X₅ cDNA and genomic sequences

A cDNA fragment with 62% sequence identity to rat P2X₅ was amplified by degenerated PCR on 14-day-old chicken embryo cardiomyocyte mRNA as previously described (Soto et al. 1996; Ruppelt et al. 1999). The full-length cDNA was then isolated by combining the screening of a chick brain cDNA library with PCR amplification of the 5′ end by RACE techniques. The predicted amino acid sequence (402 aa) shows 67% and 69% similarity to rat (Collo et al. 1996; Garcia-Guzman et al. 1996) and human P2X₅ (Lêet al. 1997) with a significantly lower sequence identity to other P2X subunits, suggesting that at least structurally this protein represents the chicken ortologue of P2X₅. The identity between rP2X₅ and cP2X₅ subunits is well distributed along the sequence with the exception of the C-terminal domain (Fig. 1). cP2X₅ is 54 aa shorter than rP2X₅ and the similarity to the rat orthologue disappears shortly after the second transmembrane domain. We performed additional 3′-RACE reactions using mRNA isolated from embryonic skeletal muscle, heart and brain, and adult chicken brain, and while the C-terminal domain described here could always be amplified, a C-terminal domain with higher similarity to the rat P2X₅ sequence was not found by this approach.

Screening of a chicken genomic library allowed us to determine the gene structure of cP2X₅ including the exon–intron borders by comparison with the cDNA sequence (Figs 2a and b). Chicken P2X₅ gene consists of 12 translated exons spanning a stretch of genome of approximately 10 kb (Fig. 2a). The gene structure is strongly conserved between cP2X₅ and the mammalian P2X genes described so far (Brandle et al. 1997; Souslova et al. 1997; Buell et al. 1998; Nawa et al. 1998; Lynch et al. 1999). The last coding exon of cP2X₅ (exon 12) starts at the point where the divergence between the rat and the chicken P2X₅ protein sequence is apparent (Fig. 2b) indicating that an exchange or modification of that exon during evolution gave rise to two different C-terminal domains.

Additionally, we performed RNAse protection analysis (RPA) on RNA isolated from heart and skeletal muscle of a 14-day-old chicken embryo in order to confirm the existence of chicken exon 12 and to investigate the possible presence of mRNA containing an alternative C-terminal exon. For this purpose, a ³²P-labeled 575 bp long antisense cRNA probe that overlaps the last 439 bp of cP2X₅ coding sequence plus 23 bp of 3′ untranslated region was synthesized (exon 8–12) (Fig. 3, undigested probe). The main band detected in both tissues corresponds to a protected RNA containing exon 8–12 (463 bp) (Fig. 3), strongly supporting the idea that the cP2X₅ predominantly expressed in chicken heart and skeletal muscle contains exon 12 as last coding exon. Additionally, two faint bands of approximately 375 and 325 bp were detected (Fig. 3). The 375 bp band corresponds to the length of an additional C-terminal domain found by 3′-RACE containing exon 8–11 and going into intron 11–12 where an immediate stop codon in frame and polyadenylation signal are present (not shown).

Tissue distribution of cP2X₅ subunits

Since the highest expression level of cP2X₅ mRNA was found by northern blot analysis in skeletal and heart muscle tissue of chicken embryos (not shown), we used radioactively labeled oligonucleotides to determine the transcript distribution in slices from both tissues. Figure 4(a and e) show an overview of the labeling from an X-ray film image, showing a homogeneous distribution of cP2X₅ transcripts in cardiac (heart; frontal section) and skeletal muscle (lower limb; transversal section), respectively. Figure 4(b and f) present the corresponding sense controls. Both dark-field images and bright-field images showing cellular resolution of emulsion coated sections of the ventricular myocardium and skeletal muscle are presented in Fig. 4(c, d, g and h). In heart, no preferential localization of P2X₅ transcripts to atria vs. ventricle could be detected.

cP2X₅ transcripts were also detected in 3–4-day-old (HH18–HH22) embryos by northern blot analysis (not shown). Therefore, we used digoxigenin-labeled probes to investigate the distribution of P2X₅ mRNA in whole-mount embryos (Fig. 5b; HH18). Chicken P2X₅ mRNA was found in somites, otic vesicle, heart, brachial arches and embryonic brain (Fig. 5b). Expression in somites was also present in HH15 embryos (not shown) while no transcripts at all were detected in younger animals (HH11) (Fig. 5a). Transversal (Fig. 5c) and sagittal (Fig. 5d) sections of somites in a late stage of maturation (stage XI–XX; Christ and Ordahl 1995) showed a strong staining in the myotome, the first differentiated skeletal myocytes to appear and source of skeletal muscle tissue of trunk and limbs (Ordahl et al. 2000). No labeling could be detected in early stage somites (Figs 5a and b).

Functional characterization of cP2X₅ receptors

To characterize the functional properties of cP2X₅ receptors, the full-length cP2X₅ clone was stably transfected in HEK-293 cells. The currents elicited by ATP in the transfected cells were then investigated using the whole-cell configuration of the patch-clamp technique. In extracellular solutions containing 1.5 mm CaCl₂, at a holding potential of − 60 mV, application of ATP (10 µm) induced a rapid inward current (100–700 pA) that desensitized with a time constant of 4.4 ± 1.3 s (n = 7) (Fig. 6a, lower trace). Only 30% of the initial current recovered after washing for 5 min (n = 5). The observed desensitization properties of the cP2X₅ receptor are unlikely to arise from the activation of endogenous P2Y receptors (Schachter et al. 1997), since including a non-hydrolysable analogue of GTP (GDP-βS; 200 µm) in the pipette solution did not significantly affect the rate of desensitization at either negative or positive potentials (n = 8).

When the membrane potential was held at + 40 mV, the rate of desensitization of the ATP-activated current was dramatically reduced (Fig. 6a, upper trace). The rate of desensitization at −60 mV was also reduced by lowering the concentration of extracellular Ca²⁺(Fig. 6b). The effect of extracellular Ca²⁺ on desensitization was mimicked by both extracellular Ba²⁺ and Mg²⁺ (1.5 mm) (data not shown). Thus, divalent cations appear to facilitate the rate of desensitization of cP2X₅ receptors by interacting with a site within the voltage field.

Due to the desensitization and lack of recovery from desensitization that cP2X₅ receptors exhibited at negative holding potentials with 1.5 mm Ca²⁺ in the extracellular solution, most experiments were carried out at + 20 mV with 0.5 mm CaCl₂ in the extracellular solution. Under these experimental conditions, the decline in the ATP response was substantially slower, and thus reliable responses to multiple applications of agonists could be obtained (Fig. 7a). Dose–response relationships for ATP and 2MeSATP were constructed by bracketing each tested concentration of agonist by a measurement with 3 µm of the agonist, which served as a reference. The currents were then corrected for rundown by normalizing the current activated by a given concentration of agonist to the preceding response to 3 µm. The resulting dose–response relationships for added ATP and 2meSATP, shown in Fig. 7(b) (n = 5–15), gave an apparent concentration for half-maximal response (EC₅₀) to ATP and 2MeSATP of 2.0 µm and 1.8 µm, respectively. Application of 100 µmαβmeATP elicited currents that were 80 ± 3% (n = 3) of the maximal current evoked by 10 µm ATP (Fig. 7a). Figure 7(c) shows that the currents elicited through cP2X₅ receptors by 10 µm ATP were fully inhibited either by 1 µm suramin or by 1 µm PPADS. Dose-inhibition curves were not constructed since it was not possible to reliably determine the contribution of current rundown to the observed reduction in current amplitude due to partial irreversibility of the antagonist effect.

P2X receptors in chicken skeletal muscle have been reported to be permeable to Cl⁻ (Thomas and Hume 1990b). Hence, we estimated the relative permeability of cP2X₅ receptors channels to Cs⁺ and Cl⁻ with the Goldman–Hodkin–Katz (GHK) equation by measuring the zero-current (reversal) potential of the net current activated by ATP. The reversal potential was measured using an intracellular solution, which contained 155 mm CsCl, and an extracellular solution, which contained either 150 mm CsCl or 30 mm CsCl. The osmolarity of the solution was maintained constant with sucrose. The net current activated by ATP was assessed by subtracting from the current activated by 10 µm ATP either the current recorded in the absence of ATP (ATP minus control) or the current remaining after inhibition by 10 µm suramin (ATP minus-ATP + suramin). Examples of net currents activated by ATP in either extracellular solution in response to voltage ramps are presented in Fig. 8(a). In 150 mm CsCl, the net current activated by ATP reversed at + 0.5 mV, and exhibited little rectification between −40 and + 40 mV (Fig. 8a) as described for the rat orthologue (Garcia-Guzman et al. 1996). The reversal potential shifted to −8.5 mV when the extracellular solution was changed to 30 mm CsCl. The average reversal potentials obtained using the two different estimates of the net current activated by ATP are presented in Fig. 8(b). The average reversal potential in 30 mm CsCl was −10.7 ± 1.3 mV (n = 5) and −9.5 ± 1.0 mV (n = 12) and in 150 mm CsCl was + 0.8 ± 0.8 mV and 0.0 ± 0.6 mV in the suramin and control subtracted currents, respectively. Assuming activity coefficients of 0.847, and 0.738 for 30 and 155 mm CsCl, respectively, a relative permeability ratio (P_Cs+/P_Cl-) of 1.98 and 1.82 was obtained using the GHK equation for control and suramin experiments, respectively. These results show that cP2X₅ receptors are highly permeable to Cl⁻, as has been reported for native P2X receptors in chicken skeletal muscle (Thomas and Hume 1990b).