The atomic model of Ce INX
The atomic model of Ce-INX6 provides an additional interpretation to the Ce-INX6ΔN mutant structure at low-resolution . Given that the N-terminal residues from 2 to 19 are deleted in the Ce-INX6ΔN mutant, the N-terminal funnel found in the pore of the wild-type Ce-INX6 should be lost in the Ce-INX6ΔN structure. The density found in the Ce-INX6ΔN pore cavity is therefore not derived from the polypeptide. In a cryo-EM study, an unassigned density was observed in the middle of a large circular configuration in a transmembrane domain of a rotor ring of an ATP synthase , which was estimated to be substoichiometric amounts of phospholipids . One possibility is that the density in the Ce-INX6ΔN pore corresponds to lipid molecules as the channels were embedded in a lipid bilayer in 2D crystals. When constructing an oligomeric form of the Ce-INX6ΔN channel, it might be possible for lipids to fortuitously remain inside the pore unless excluded by the N-terminal funnel. Interestingly, a wide space is observed between the adjacent subunits in the transmembrane region in the high-resolution cryo-EM structure of Ce-INX6 where unassigned densities like carried-over lipids or detergent micelles were inserted close to the N-terminal funnel  (Figure 2d). High-resolution structures of the Ce-INX6 gap junction channels reconstituted in the lipid bilayer are awaited for further investigation of the contribution of lipids to channel function.
The N-terminus is essential for channel function in both connexin and innexin channels The N-terminus of connexin is thought to be essential for channel function in association with the voltage-dependent gating mechanism (summarized in Ref. ). Specifically, the voltage sensor is located in the N-terminus of connexin . This suggests that the N-terminal portion should reside in the pore vestibule because this is the only position at which the difference in voltage across the adjacent LGX818 mg can be sensed  (Figure 3a). The structural organization of the N-terminal funnel observed in the Cx26 X-ray structure is consistent with these interpretations. Despite the lack of sequence similarity, the Ce-INX6 structure also has an N-terminal funnel, suggesting that this conformation could be generally shared among gap junction family proteins. The partially resolved N-terminus of LRRC8A faces the pore , which, if it is not helical, is similar to Cx26 and Ce-INX6 [,]. All of the N-terminal deletion constructs of Cx37 are no longer able to form conducting channels , consistent with findings from functional innexin studies [41,50]. These findings suggest that an intact N-terminus is necessary for normally opening gap junction channels. Recent X-ray structures of Cx26 by Bennet et al. do not visualize the N-terminal portion ; nor does the electron crystallographic structure of Cx43 . It is possible that the N-terminus of connexin is flexible and can assume variable and unexpected conformations. The movement of the N-terminus during a functional cycle remains to be further elucidated. Functional studies with site-directed mutagenesis demonstrated the possibility that voltage-dependent closure is initiated by movement of the voltage sensor, namely the N-terminus of connexin, toward the cytoplasmic side [48,51] (Figure 3a). This movement is inconsistent with the idea that the N-termini aggregate to act as a physical blockage [,], and cannot be accounted for by electrostatic distribution in the structures without the N-terminal portion . The individual model suggests that the behavior of the N-terminus does not have to be concerted, but can be independent . The discrepancy of these interpretations is basically attributed to the lack of a high-resolution structure in a closed state representing a physiological condition. The pore size, estimated from the structures above to be over 10 Å in diameter, should be completely shut when the channel is closed. One key point that has been so far neglected in considering the functional interpretation of gap junction channels is the possible presence of lipid molecules. It should be noted that all high-resolution structures of gap junction channels by X-ray and cryo-EM are in a solubilized form surrounded by detergent micelles that would have removed most of the lipids carried over from cells. The cryo-EM structure of Ce-INX6 indicates that lipid molecules could possibly be distributed between adjacent subunits in the transmembrane domains in vivo, where is close to the N-terminal funnel  (Figure 2d). Nanodisc reconstitution  may be a possible method for obtaining a high-resolution structure of a gap junction channel in the lipid bilayer that will provide insight into understanding the N-terminal arrangements of connexins and innexins under more physiological conditions.