Connecting the dots on connexin function

Connexin (Cx) proteins assemble into hemichannels in the plasma membrane of various cell types, mediating the transport of signaling molecules and metabolites, as well as ions, between the cytoplasm and the extracellular space. In addition, hemichannels from neighboring cells can dock together to form gap junctions that mediate transport directly between cells.

Two connexins, Cx26 and Cx30, are widely expressed in the cochlea, and mutations in the genes encoding either protein cause hearing loss. Cx26, encoded by the GJB2 gene, appears to be particularly important: mutations in GJB2 account for up to 50% of all severe-to-profound inherited deafness cases (1), and studies in mice have shown that, while Cx26 is indispensable for hearing, loss of Cx30 does not cause deafness as long as Cx26 levels are maintained (2, 3). In this issue of JGP, companion studies by Sanchez et al. (4) and Kraujaliene et al. (5) reveal that a single amino acid difference in a pore-lining region of Cx26 causes significant functional differences between the two connexins, which could help researchers begin to understand why Cx26 is more important for hearing.

Previous studies have suggested that a 10-amino acid segment in the first extracellular loop of connexins forms a parahelical structure that lines the pore and participates in voltage-dependent gating (6). The sequence of this segment is identical in Cx26 and Cx30 except for a single difference at position 49, which is an alanine in Cx26 and a glutamate in Cx30. Both connexins have an aspartate residue at position 50. Intriguingly, mutations in this region of Cx26, including at position 50, are associated with a particularly severe hearing disorder known as keratitis-ichthyosis-deafness (KID) syndrome, indicating its importance for connexin function.

Vytas Verselis and colleagues, including Helmuth Sanchez and Lina Kraujaliene, discovered that, compared with Cx26 hemichannels, Cx30 hemichannels function poorly when expressed in Xenopus oocytes (4). However, mutating either of the charged residues at positions 49 and 50 of Cx30 resulted in robust hemichannel conductance. Conversely, the activity of Cx26 hemichannels was greatly reduced when the neutral alanine residue at position 49 was mutated to a charged glutamate or lysine. This suggests that an interaction between charged residues at positions 49 and 50—as occurs naturally in Cx30—favors hemichannel closing.

To investigate the parahelical region in more detail, Verselis and colleagues substituted cysteine for alanine at position 49 (4). Cx26(A49C) and Cx30(E49C) hemichannels were susceptible to modification by thiol-modifying reagents, confirming that position 49 is exposed to the pore. Moreover, the researchers found that Cx26(A49C) and Cx30(E49C) hemichannels are inhibited because the cysteine residues form disulfide bonds and coordinate metal ions, blocking the hemichannel pore. “This suggests that the parahelical region is highly flexible and can narrow the pore, consistent with computational studies indicating the region’s importance for hemichannel gating,” Verselis explains.

Surprisingly, given its effects on hemichannel conductance, the researchers found that cysteine substitution at position 49 does not inhibit the activity of Cx26 or Cx30 gap junctions (4). Because the extracellular loop containing the parahelical region is involved in docking apposing hemichannels together, Verselis speculates that gap junction assembly may limit the region’s flexibility, preventing the cysteine from forming disulfide bonds. This restricted flexibility might also explain why the E49–D50 charge pair in Cx30 dampens hemichannel function but does not appear to reduce the activity of Cx30 gap junctions.

Cx26 and Cx30 gap junctions do appear to differ in at least one respect, however; Cx26 channels have been reported to be much more permissive to the passage of larger anionic tracers than Cx30 channels, which could reflect a greater capacity to transport biomolecules such as ATP. In their second paper, Verselis and colleagues report that this permeability profile can be reversed by altering the amino acid at position 49 (5). Cx26(A49E) gap junctions and hemichannels are much less permeable to fluorescent anionic tracers, whereas Cx30(E49A) channels permit the passage of these molecules.

Small differences in the pore-lining parahelix therefore affect the functional properties of Cx26 and Cx30 in both the gap junction and hemichannel configurations. But what might this mean for the two connexins’ physiological role in hearing? In the developing cochlea, connexins are thought to participate in the ATP-dependent generation and propagation of Ca²⁺ waves that promote hair cell differentiation.

“We thought that Cx30 might not be so good at this because of its possible low permeability to ATP and the decreased activity of Cx30 hemichannels,” Verselis says. The researchers therefore looked at the spread of Ca²⁺ waves in mouse cochlear explants labeled with a Ca²⁺-sensitive dye. “We saw that waves originating in nearby supporting cells didn’t penetrate into the border region containing the hair cells, and we found that Cx30 is more highly expressed in this border region,” Verselis adds.

Thus, by increasing hemichannel activity and, potentially, ATP permeability, the presence of an alanine at position 49 in the pore-lining parahelix might help explain why Cx26 has a more important role in hearing than Cx30. Verselis and colleagues’ findings suggest that mutations in Cx30 could alter this connexin’s activity to compensate for the loss of Cx26. In addition, the discovery that introducing a cysteine at position 49 inhibits hemichannel, but not gap junction, function provides a tool that could help researchers distinguish their different contributions to cochlear development.