1. Hemichannels formed by connexin26 (Cx26) on the horizontal cell dendrites that invaginate cone terminals in the vertebrate retina have been implicated in the feedback mechanism by which horizontal cells regulate transmitter release from cone photoreceptors. However, their membrane properties had not been studied previously, and it was unclear whether they could subserve their purported function at the membrane potentials over which horizontal cells operate. 2. We used the two-electrode voltage clamp technique to record the membrane currents and pharmacological properties of Cx26 hemichannels formed in the Xenopus oocyte expression system. 3. Oocytes expressing Cx26 exhibited large membrane conductances over a broad range of hyperpolarizing and depolarizing membrane potentials, and displayed little evidence of voltage-dependent gating, indicating that the hemichannels are constitutively open. The Cx26-mediated nonjunctional currents were relatively insensitive to quinine, a cinchona alkaloid that opens hemichannels formed by several other connexins. However, the hemichannel currents were blocked by carbenoxolone, a rise in extracellular calcium, or lowering intracellular pH. The currents could also be suppressed by reducing extracellular pH, and by the chloride channel blocker NPPB through its direct interaction with Cx26 hemichannels. 4. These findings provide a basis with which to evaluate the in situ pharmacological studies that attempt to assess the putative role of Cx26 hemichannels in the feedback pathway in the distal retina.

Previous reports have shown that gap junctions relay cell death in many cell types. However, changes in electrical coupling and their dynamics during cell death are poorly understood. We performed comprehensive studies of electrical coupling following induction of cell death by single-cell cytochrome c (cyC) injection in paired Xenopus oocytes. Cell death was rapidly induced by cyC in injected cells, and cell death was also observed in uninjected bystander cells electrically coupled to the cyC-injected oocytes. Gap junction currents either remained at pre-cyC injection levels or increased dramatically as the injected cell died. Nonjunctional currents increased in injected cells immediately following cyC injection; nonjunctional currents increased slowly in uninjected bystander cells. Bystander cell death occurred only when junctional conductance was B6 lS. Both 1,2-bis-(o-aminophenoxy)-ethane-N,N,-N 0 ,N 0 -tetraacetic acid tetraacetoxy-methyl ester and Xestospongin C inhibited bystander cell death in pairs that had reached the death conductance threshold, suggesting that Ca 2 þ and inositol 1,4,5 triphosphate are involved in the process.