The molecular mechanism by which receptors are activated by a ligand, distinguish between ligands (where multiple ligands exist eg, Delta and Serrate both bind to Notch) and alter transcription appropriate to a received stimulus is complex and is an area receiving detailed investigation. Gap junctions are even less-well understood. Currently, no mechanism has been proposed to explain how disruption of gap junctions can alter ligand/receptor pathways leading to aberrant development. But there are a few hypotheses. But firstly, what observations suggest that defective gap junction comunication, or gap junction subunits themselves, lead to aberrant development by disrupting ligand/receptor pathways?
- Connexin mutants display defects in cell proliferation (Temme et al. 1997), cell fate determination (Hirschi et al. 2003) and pattern formation (Levin and Mercola 1999) - all processes that have previously been associated with mutation of ligand/receptor pathways (Proliferation - Giraldez and Cohen, 2003, Differentiation - Artavanis-Tsakonas et al. 1999, Patterning - De Celis, 2003).
- Genetic interaction phenotypes observed during a screen to identify loci that interact with dominant Inx2 mutants in Drosophila resemble ligand/receptor pathway mutants (for example, wing phenotypes similar to those of Notch mutants). Phenotypes generated by UAS-innexin overexpression also feature similarities to phenotypes observed in ligand/receptor mutants. Expression of UAS-ogre in leg discs results in tarsal segment fusion analogous to that seen in Serrate mutants, UAS-inx2 expression at the wing margin generates serrated wings similar to dominant alleles of Serrate mutants (eg. Ser1), UAS-inx7 induces excess vein material in wings resembling Delta mutants, UAS-inx7 expressed in the thorax disrupts bristle patterning like phenotypes seen in Notch mutants that perturb lateral inhibition (eg. Notch pathway).
- Gap junction- and ligand/receptor- signalling pathways regulate each other at the transcriptional level in vertebrates and invertebrates (Ai et al. 2000, Lechner et al. 2007). Consequently, for example, Cx43 null mutants exhibit altered transcription of ligand/receptor pathway molecules such as the Notch receptor (Iacobas et al. 2004). In flies, expression of UAS-ogre can reduce E(spl)mβ reporter expression in leg discs and UAS-inx7 increases rhomboid transcription in the wing. Both E(spl)mβ and rhomboid are downstream targets of Notch pathway signalling which suggests that manipulating innexin levels somehow effects the Notch ligand/receptor pathway.
Outside of the transcriptional interdependence of gap junction and ligand/receptor genes, what other mechanisms could allow these pathways to interact under normal conditions and in mutants? The illustration at the top of the page labelled three areas of the cell where interactions between these two separate signalling pathways could hypothetically take place:
- - For ligand/receptor signalling to function optimally, are there molecules that must pass from cell to cell via gap junctions? Similar developmental defects might then arise when either one of these pathways, ligand/receptor or gap junction components, are significantly disrupted. No candidate signalling molecule is known, but ions (Goldberg et al. 2004), second messengers (Kanaporis et al. 2008), short interfering RNAs (Valiunas et al. 2005) and possibly even proteins (Brooks and Woodruff, 2004) might have to pass through gap junction channels as part of normal ligand/receptor signalling.
- - If there's a mutual inter-dependence between the signalling pathways, as suggested above, then why don't all innexin mutants exhibit phenotypes of similar severity to those observed in mutants of a hypothetically associated ligand/receptor pathway? For example, escapers of putative lethal ogre alleles have no obvious defects in external morphology (Lipshitz and Kankel, 1985) despite the expression of Ogre in both leg and wing discs (Stebbings et al, 2002) and the phenotypes observed in inx2null alleles appear to be restricted to certain tissues (such as the proventriculus, Lechner et al, 2007) despite the extensive expression pattern of Inx2 in the embryo. Gap junction channels are multimeric structures, so, even if one family member is absent, functional channels could possibly still form in some cases. These non-wild type compensatory channels may be similar enough, regarding their permeability and selectivity properties, to the original channels to permit messengers essential for ligand/receptor signalling to function almost normally. Functional redundancy has been proposed for connexins (White, 2002 and Simon et al. 2004) and may also apply to innexins, for example, Inx7 can replace Ogre in neural circuits involved in vision and can thus rescue ogre mutant electrophysiological phenotypes (Curtin et al, 2002).
- - Gap junction plaques are relatively large structures that physically connect cells together and possibly have a role in cell adhesion (Prochnow and Dermietzel, 2008) and potentially a role in restricting the lateral dispersal of plasma membrane molecules eg, ligands and receptors. Gap junctions are also somehow involved in regulating paracellular flow (Morita et al. 2004). All of these properties are similar to those of tight (..or septate, in flies...) junctions. Indeed functional links between vertebrate gap junctions and tight junctions are reported in the literature (Kojima et al. 2002, Go et al. 2006). If gap junctions are involved in the formation/maintenance of septate/tight junctions then altered gap junctions could disrupt the micro-environment in which ligands and receptors interact (Woods and Bryant, 1993). By failing to hold apposing cell membranes close together enough for ligands and receptors to interact, or by disrupting local ion concentrations that ligands/receptors are sensitive to, abberant gap junctions could lead to sub-optimal ligand/receptor-mediated signalling.
- - Only some mutations actually delete gap junction channel subunits. Most point mutations could produce non-functional subunits or subunits with a modified activity. One potential effect of substitution mutations is to generate leaky hemichannels.
- - Leaky connexin-based channels have been documented at the plasma membrane (Stong et al. 2006, Liang et al. 2005). Such channels could disrupt cellular homeostasis, impeding ligand/receptor pathways from transducing appropriate developmental input →output signals. Disruption of the solute composition of the extracellular milieu could prevent ligands and receptors from interacting optimally (for example, the Notch receptor is sensitive to Ca2+ concentration (Rand et al, 2000, Raya et al, 2004)).
- - There are reports of non-gap junction-channel mutants synergistically interacting with ligand/receptor pathways (Doherty et al. 1997). Perhaps mutation of gap junctions can result in similar interactions. Many plasma membrane channel mutants exist in Drosophila, making it an ideal candidate to use for identifying potential links between ligand/receptor pathway phenotypes and cellular homeostasis defects arising from mutant plasma membrane channels.
- - Not all develomental defects need arise due to channel activity at the plasma membrane. Connexin fragments that are not incorporated into the membrane can induce developmental phenotypes, possibly through some property inherent in the carboxy terminal region (Dang et al. 2003). Transgenically induced phenotypes (in Drosophila, at least) may arise due to gap junction channel subunits accumulating within the cytoplasm (as observed when UAS-Ogre-myc is overexpressed in salivary gland cells) - although, what processes that might disrupt is anyones guess. Overexpressed innexins could potentially also enter the nucleus and mediate changes in expression there...but so far antibody staining has failed to reveal overexpressed innexin at this location and until very recently there was no indication at all, or reason to believe, that innexins might have a role in the nucleus. That changed when immunocytochemical analysis revealed Inx7 in the nucleus of certain cell types at specific developmental stages (Ostrowski et al, 2008).