It's asymmetric: the molecular structure of gap junctions

Science Spotlight

It's asymmetric: the molecular structure of gap junctions

From the Moens Laboratory, Basic Sciences Division

July 17, 2017

Billions of neurons must communicate with one another so that animals can sense, process, and move in their environments. Information is transferred between neurons across specialized junctions called synapses using two general mechanisms: chemical and electrical transmission. Chemical synapses transfer information when one neuron releases a neurotransmitter and another uses a neurotransmitter receptor to ‘listen’ to the signal. Electrical synapses utilize direct communication between neurons by channels called gap junctions, which are composed of two hemi-channels, each one contributed by one of the neurons. The channels allow molecules and ions less than ~1 kilodalton to travel directly between the two cells and enable rapid bi-directional molecular exchanges. While most studies focus on chemical synapses, a recent resurgence in interest in electrical synapses has shown that they function broadly within nervous systems during development and adulthood. However, their molecular makeup is not well understood. Neuronal gap junctions are generally viewed as symmetric, made up of identical components contributed from each cell. More recently, scientists have discovered that invertebrate gap junctions can be asymmetric, formed of different proteins from each neuron. There is also evidence from microscopy and electrophysiology that electrical synapses can be asymmetric in vertebrates. However, the exact molecular components that make the vertebrate channels asymmetric have not been identified. In their recent publication in eLife, scientists in the Moens Lab (Basic Sciences Division) determined that unique types of channel-forming proteins called Connexins are found on each side of electrical synapses in a circuit of neurons in the model vertebrate fish Danio rerio.

Scientists in the Moens Lab study neuronal development and function using the zebrafish Danio rerio. These fish are tractable to study because during early stages of development they are optically clear so their tissues and cells can be visualized with minimal disruptions. To study the molecular composition of gap junctions in zebrafish neurons, the scientists chose to focus on a neural circuit responsible for the 'escape response' of the fish to sudden movements. They knew that an antibody to the human gap junction protein Connexin36 (Cx36) recognized the electrical synapses of this circuit so they designed a forward genetic screen to identify genes responsible for the formation of the synapses in the circuit. They created fish with random genomic mutations using a chemical and then monitored whether the pattern of Cx36 antibody staining of the neural circuit changed in each animal. They identified a mutant, which they called disconnect3, that caused a complete loss of Cx36 antibody staining but did not change the number or morphology of the neurons in the circuit. 

To identify which genes were disrupted in this mutant, they sequenced the RNA of a pool of mutant or wild-type animals and searched for regions of homozygosity, or shared sequence. They identified a ~1.8 megabase pair region on chromosome 5 that was similar among mutant animals but not among wild-type and within that region, missense mutations in a Connexin-encoding gene, called gap junction delta 1a (gjd1a). The Connexin proteins are named for their molecular weight and the gjd1a gene encodes a connexin with a predicted molecular weight of 34.1 kiloDaltons (Cx34.1). This gene and the human Cx36 gene are very similar.

By searching through the zebrafish genome, the scientists found three other related Connexin-encoding genes. They mutated each of these other zebrafish connexin genes, analyzed Cx36 antibody staining, and found that Connexin34.1 (gjd1a) and Connexin35.5 (gjd2a) are required to form gap junctions in the electrical synapses studied. They also found that the function of the synapses was disrupted when these genes were mutated. In analyzing the effects of cx34.1 and cx35.5 mutations on the escape response of the fish, the scientists found that fish with mutations to cx34.1 or cx35.5 were slower in their escape response than wild-type fish and that mutants had defects in their posture during the escape.

The scientists wondered why the synapses used two Connexins. In mammals it is often thought, though the details are not well understood, that Connexin36 alone makes most neuronal electrical synapses. To locate where the two Connexins were located in the neurons, authors raised specific antibodies to Cx34.1 and Cx35.5. They found that both localized at the gap junctions and each Cx was required for the localization of the other. The scientists wondered whether both proteins were mixed in all complexes or whether each was found exclusively in the pre-synaptic or post-synaptic hemichannel. To determine this they created chimeric animals, which have cells from two differently labeled animals, so they could identify whether mutations to each connexin gene were required in a given pre-synaptic or post-synaptic neuron in the circuit. They discovered that Cx35.5 is required exclusively in the pre-synaptic neurons and Cx34.1 is required in post-synaptic neurons.  

Cx34.1 is required in the post-synaptic neuron and Cx35.5 is required in the pre-synaptic neuron of the Mauthner/CoLo synapse of zebrafish. Arrows indicate the transplanted neuron that has either Cx34.1 or Cx35.5 mutated (top, bottom row respectively).

Image adapted from Figure 6 of the publication.

Click for high-res version

Quantification of Connexin36 staining on pres-synaptic and post-synaptic neurons in the M/CoLo circuit of wild-type and chimeric animals.

Image Figure 6Q from the publication.

Overall, their data strongly suggests that these Connexin proteins are important for forming functional electrical synapses and identifies that the channels are asymmetric at the molecular level. Whether this asymmetry biases flow through the channel or influences protein-protein interactions on either side remains to be determined. Moreover, the authors suggest that all brains, including our own, are likely to have molecularly asymmetric electrical synapses. Future studies are likely to reveal an increasing complexity in these integral, yet understudied, synapses. 

Miller AC, Whitebirch AC, Shah AN, Marsden KC, Granato M, O'Brien J, Moens CB. 2017. "A genetic basis for molecular asymmetry at vertebrate electrical synapses." eLife. 

Funding for this research was provided by the National Institute of Neurological Disorders and Stroke, the National Institute of Mental Health, the National Eye Institute, and the Eunice Kennedy Shriver National Institute of Child Health and Human Development.