The subunits that make up the P2X receptor family form both homomeric and heteromeric ATP-gated non-selective cation channels. Assemblies of P2X₃ and of P2X₂ and P2X₃ subunits are believed to dominate the P2X receptor complement of sensory neurones, and consequently have been associated with nociceptive signalling and the search for novel analgesics.

We have characterised a CHO cell-line stably expressing both the human P2X₂ and P2X₃ subunits, using whole cell voltage-clamp combined with concentration-jump techniques. Currents mediated by P2X₃ and/or P2X₂/3 receptors were activated using 10 µM α,β-me ATP, a concentration previously demonstrated not to significantly activate homomeric P2X₂ receptors. Data are expressed as mean ± S.E.M.

At -30 mV, single 1.5 s applications of 10 µM α,β-me ATP, characteristically produced a biphasic current response consisting an early component which activated and desensitised rapidly followed by a sustained plateau current. When a second identical agonist application was made 5 s later the initial desensitising current component was entirely absent, whereas the sustained current component remained largely intact. By subtracting the second response from the first it was possible to characterise the properties of the desensitising current component (I_desen), which is likely to reflect P2X₃ homomers, and compare it to the response to the second agonist challenge (I_resis), which is likely to represent only P2X₂/3 heteromers. An agonist free period of around 3-5 min was required to allow I_desen to completely recover. The ratio of I_desen to I_resis varied from 0.34 to 35.8 and did not correlate with current amplitude. The reversal potentials of both I_desen and I_resis were indistinguishable from 0 mV. I_resis, like homomeric P2X₂ currents, exhibited strong inward rectification (I_-60/I₊₆₀ = 0.14 ± 0.02, n = 8), whereas, I_desen, like homomeric P2X₃ channels was only weakly rectifying (I_-60/I₊₆₀ = 0.71 ± 0.08, n = 8). At -30 mV, the time constants of I_resis activation and deactivation did not correlate and were 565 ± 33 ms and 635 ± 33 ms, respectively. Activation was significantly slowed and deactivation significantly speeded by depolarisation, inferring agonist affinity may be reduced by membrane depolarisation. Macroscopic desensitisation of I_resis occurred with a time constant of 16.5 ± 2.8 s, but failed to reach completion during 90 s of continuous agonist application. Unlike GABA_A receptors, the deactivation kinetics of I_resis were not sensitive to prior desensitisation. Subsequent experiments with cultured rat neonatal dorsal root ganglion neurones revealed a similar dual component response to 10 µM α,β-me ATP suggesting these cells express both homomeric P2X₃ and heteromeric P2X₂/3 channels.