Share This article

Real neurons don’t actually talk to each other with electrical spikes. For that, their synapses deploy vesicles filled with transmitters. Although there is clearly a relation between the spikes a synapse receives, and the transmitter it releases, neuroscientists have very little idea what this relation really is. Part of the problem is that they don’t even know how the vesicle release mechanism behind this process of exocytosis operates.

New work just published in Nature now gives us an atomic-scale picture of the protein conglomerate that docks vesicles to the presynaptic membrane and actuates the fusion pore through which the transmitter flows. Led by recent Nobel laureate Thomas C. Südhof, the researchers used this data to create a model of exocytosis based on the operation of two key elements — synaptotagmins and SNAREs. Anyone who has every looked to see how Botox injections work their magic might recognize these guys as the places where the botulinum toxin works its own magic. Incidentally, the tetanus toxin also directly targets these protein intermediaries.

Synaptotagmins are proteins that bind the calcium that enters a synapse after a spike comes along and triggers the opening of voltage-sensitive calcium channels. SNAREs are proteins with precision mates between long amino acid helices that preferentially insert themselves into both the vesicle membrane and the synapse membrane at the release site. ‘SNARE’ is a monster of an acronym, multiple-nested in a way that only a biologist could dream up. On its surface it stands simply for SNAP Receptor, where ‘SNAP’ is Soluble NSF Attachment Protein. Within that, the ‘NSF’ stands for N-ethylmaleimide sensitive factor.

To get their atomic-level images, the group first carefully grew crystals of regularly repeating units of a synaptotagmin-SNARE complex. These proteins were also bound to different enlightening accessory molecules like calcium or various toxins that are known to bind or cleave the proteins at precise places. They collected diffraction data using the X-ray free-electron laser at Stanford’s accelerator facilities. As you can see in the image above, some fairly sophisticated vacuum equipment is needed to control conditions and manipulate the orientation of the crystal relative to the beamline.

What the researchers now think is happening is that when the synapse is in standby mode waiting for a spike to come along, a few vesicles are already tethered in primed state at the release site. As shown above, membrane fusion and transmitter release can then be quickly triggered by a change in the conformation of the protein complex after it binds calcium. Presumably, when enough protein complexes are similarly activated in a ring around the impending fusion pore, the whole geometry changes and a synaptic wormhole of sorts is transiently created.

In Fig. C above we see the fusion machinery in the ‘primed state’ with the vesicle already attached by the large protein complex at the release site. When sufficient signal in the form of spike energy invades the synapse, fusion is triggered along with creation of the exit pore. Typically, the ‘signal’ is understood to be a change in potential followed by opening of calcium channels, and then calcium binding to synaptotagmin or other binding sites on the complex.

In theory it should be possible to build an organic computing machine that reliably releases vesicles from its synapses every time a spike comes along, and furthermore only when a spike comes along. That might be an interesting device, but it wouldn’t be a brain. If spikes alone could do the trick, as the neural network crowd now presumes, then evolution would not have provided for the elaborate synapses we see in real brains. In the larger scheme of things, spikes themselves are actually fairly cheap to send. On the other hand, a neuron’s many thousands of synapses spend not only considerable time doing their business, they expend comparatively vast amounts of energy reclaiming spent transmitter and pumping it back into vesicles.

It would certainly have been simple enough for brain cells to connect directly using ‘gap junction’ proteins to send electrical signals using using raw ions. Many cells routinely use such junctions for various low-levels communication tasks, and there are several varieties maintained in the genome to meet needs as they arise through deep evolutionary time. However, it seems gap junctions and direct electric is not the path chosen by nature when it come to brains.

There are still a lot of unknowns in how the SNARE complex actually powers the creation of the fusion pore. Although bare lipids vesicles are capable of spontaneous transitions in size and geometry — regularly fusing or splintering into tubes, sheets, and lamellae according to the local ebb in the entropic background — to do it on demand generally takes energy and auxiliary control molecules. It is now believed that some of the energy for fusion is transduced by partial disassembly of the SNARE complex, and perhaps also indirect use of energy provided by ATP hydrolysis as in muscle contraction.

With an understanding of what drives signaling in neurons will come better ways to incorporate those features into the neural networks that are now creeping into many of our electronic devices.

Main image credit: SLAC National Accelerator Laboratory