Background

Gap junctions are clusters of intercellular channels that directly connect the cytoplasm of adjacent cells, providing a conduit for the exchange of metabolites and ions. In chordates, gap junctions are formed by connexins, a diverse family of membrane proteins that have the ability to oligomerize into dodecameric channels. Mutations in connexin genes are the cause of a variety of inherited diseases. In many cases, these pathologies have been linked to trafficking defects that compromise the ability of the channel to form functional gap junctions [1,2]. Among the 21 connexins that have been identified in humans, only a few variants have been studied in terms of trafficking through intracellular compartments [2–4]. To study basic transport mechanisms of gap junction proteins, researchers take advantage of expression systems, such as HeLa or HEK293T cells, and transfect these cells with recombinant expression vectors. In this article, we describe a detailed protocol that can be used to determine endoplasmic reticulum (ER) retention defects for connexin 36 (Cx36). In a previous study, we described a transport defect that prevented the functional ER export of Cx36, causing the connexin to accumulate in the ER [5]. This retention mechanism promoted the formation of gap junction-like aggregates that reshaped the ER into concentric multi-membrane vesicles (Figures 1B and D), which we termed connexin whorls. These structures were characterized by several distinct features: 1) Each sheet within the whorl exhibited ultrastructural features that were indistinguishable from an actual gap junction; 2) whorls are hollow inside and their diameter varied, ranging from 0.3 to 3 µm; 3) whorls colocalized with ER-phagy receptors Tex264 (Testis-expressed protein 264) (Figure 1C) and p62; 4) whorl formation requires docking interactions of extracellular loops in Cx36 facing the ER lumen. Substituting the extracellular loop cysteines (C55 or C62) via site-directed mutagenesis prevents whorl formation. Similar whorl-like structures were reported for the lens connexin Cx50 in a previous study by Lichtenstein et al., [6], which suggests that ER-derived whorls are formed by many connexins that oligomerize in ER. Therefore, the protocol we describe here is not only applicable for Cx36 but might be used for other connexin isoforms. However, two important aspects have to be considered: 1) Some connexins, for instance Cx43, oligomerize in the Golgi [7], which would make the docking of Cx43 containing connexons in the ER impossible; and 2) whorls have to be tested for the presence of ER proteins, for instance Tex264, to determine the compartment they originated in (Figure 1C).

Figure 1. Membrane topology of Cx36. Protein structure of Cx36 is illustrated. The two extracellular loops of Cx36 (indicated by upper circle and red arrow) contain three cysteines. The C-terminal tail contains the PDZ binding motif consisting of the following amino acid sequence: SAYV.

Rationale for experimental design

The rationale for the design of connexin mutants we have described in Tetenborg et al., (2023) was based on the observation that truncation of the C-terminal tip in Cx36 prevents functional ER export leading to the intracellular formation of connexin whorls. To test if these multimembrane vesicles are formed by a gap junction-like docking mechanism of opposing hemichannels in the ER, we substituted cysteines in the extracellular loop of Cx36 with serines via site-directed mutagenesis. These mutations were previously shown to block gap junction formation [8] and served as a control experiment in our study to confirm that ER retention causes opposing Cx36 channels to dock intralumenally via the extracellular loops.