Raftaugamót

Mig grunar að fæstir viti um hvað námið mitt snýst. Þess vegna ætla ég að birta hér samantekt um raftaugamót sem ég var að ljúka við. Athugið að þetta var skrifað með mig og samnemendur mína í huga. Þetta eru glósur, ekki bókmenntaverk.

 

© Heida Maria Sigurdardottir and Scott Cruikshank

Gap junctions are electrical synapses (not chemical synapses); these are almost synonyms. They are a group of channels that connect the cytoplasm of two cells.

 

The difference between electrical and chemical synapses:

Chemical   
Presynaptic calcium channels are opened by an action potential and calcium flows in.

The calcium triggers the fusing of vesicles so that neurotransmitters are released into the synaptic cleft

Neurotransmitters bind to postsynaptic receptors which leads to an increase in the open probability of postsynaptic ion channels

Depending on the channel, either cations or anions (floating around in the synaptic cleft) flow through, depolarizing or hyperpolarizing the postsynaptic membrane

Electrical
Action potential sets up a voltage gradient between the presynaptic and postsynaptic membrane (presynaptic more depolarized)

This will repel positive ions from the presynaptic neuron through the gap junction AND attract negatively charged ions from the postsynaptic neuron to go through (gap junctions are fairly non-selective, almost any ion can flow through)

This will create a little depolarization (spikelet) in the postsynaptic cell that is shaped similarly to the action potential

Chemical ---- Electrical

Complicated  ----  Simple
Metabolically expensive (synthesize all the neurotransmitters, have all the proteins etc.)  ----   The gap junctions are basically just tunnels, all you need is the channel itself

Can be some delay from the arrival of the action potential to the postsynaptic event BUT in mammalian cells this is super fast anyway  ----   Relatively instantaneous (as fast as the current can flow)

Ion selective (cations or anions, not both), gives you either depolarization or hyperpolarization ----    Most ions can flow across gap junction channels (although monovalent cations flow better than calcium or anions). Some other molecules can go across as well

Subthreshold potential change in the presynaptic membrane by definition doesn’t cause transmitter release and therefore doesn’t affect the postsynaptic membrane potential   ----  Subthreshold potential changes get conducted (a large or a small spike will set up a voltage gradient across the synapse and drive ions through)

Hyperpolarizing the presynaptic membrane has no effect on the postsynaptic membrane   ----  Hyperpolarizing the presynaptic membrane hyperpolarizes the postsynaptic membrane (ions move in the opposite direction)

In general conduction is unidirectional: From presynaptic to postsynaptic  ----   Conduction is bidirectional: There are no real “presynaptic” and “postsynaptic” membranes

Postsynaptic density  ----   No PSD?

History:

There was a controversy whether neurons communicated by electrical or chemical synapses (in the 1950´s). John Eccles, a neurophysiologist, was the advocate for electrical synapses. The work of Hodkin and Huxley (and others) made it clear that chemical synapses were real, so finally Eccles completely changed his mind (“the conversion of Eccles”) after he did an experiment where he stimulated inhibitory axons in the spinal cord and recorded from motor neurons. He found hyperpolarized potentials in the postsynaptic cell even when he DEPOLARIZED the presynaptic cell. That means that the synapse was not electrical, because then the potential change of the postsynaptic membrane would mimic the presynaptic membrane potential. Work on electrical synapses “died”.

However, a few years later, people found electrical synapses in invertebrates and then also in the vertebrate CNS. The recorded from two adjacent supramedullary cells (above the medulla) that spiked in synchrony and found that if they hyperpolarized one cell the other cell hyperpolarized as well. In general that is a sign of an electrical synapse. Importantly, they found that this was bidirectional, so they could inject current into either cell and get a response in the other.

What about mammals? For some time people thought that electrical synapses were only in a few specialized areas. In 1999 there were two breakthrough papers published in Nature that showed widespread electrical coupling in the cortex. They optically (DICR optics?) identified various cell types (interneurons, excitatory neurons etc.) and chose to record from particular types and from cells that were near one another. When they injected current into cell 1 the cell depolarized until it spiked. The interesting part was that they found depolarization of cell 2 BEFORE cell 1 spiked, indicating that they were coupled by electrical synapses (chemical synapses don’t allow conduction of subthreshold potential changes). The gold standard was that hyperpolarization of cell 1 also hyperpolarized cell 2.

Fast-spiking cells are electrically coupled to other fast-spiking cells, as are low-threshold spike cells to other low-threshold spike cells (late-spiking cells, multipolar bursting, receptor 1 positive cells etc.), but there aren’t many connections across different types of interneurons. There seem to be almost no electrical connections between inhibitory neurons and excitatory neurons, or between two excitatory neurons.

The interneurons were electrically coupled to many other interneurons of the same type, so it’s a massive network of connected neurons.

SO WHY DID NO ONE SEE THIS BEFORE?

Until the 80’s people mostly recorded in vivo, not in vitro, so it was virtually impossible to record from two adjacent cells.

In the 90´s most recordings were done “blind”, so you didn’t really know where exactly you were sticking your electrodes. If you don’t see, there is a very low probability to record from two interneurons of the same type at the same time, because they are so few.



 

Structure and molecular composition of electrical synapses:

Electrical synapses are made from pore-forming intercellular channels that together form a kind of honeycomb structure. Ions flow through them directly from one cell to another. A functional channel needs two hemichannels (also called connexons), one from each cell. Each connexon is a hexamer, so it has six subunits, and each subunit is called a connexin.

Each subunit (connexin) has four membrane-spanning regions, two extracellular loops, an intracellular (cytoplasmic) loop and an N-terminus and a longer C-terminus, both on the intracellular side (see fig. 3). The extracellular loops are thought to be important for docking of the two hemichannels. The C-terminus is thought to contain a phosphorylation site that affects the conductance of the channel (kinases can phosphorylate the channel). Calcium affects the channels so they are more often open when calcium is not around.

At least 20 different kinds of connexin proteins (subunits) have been identified in the mouse. They are named after their molecular weights (in kilodaltons, kDa). They are NOT only found in neurons. They can for example be found in the heart, where they are probably responsible for the synchronization of the rhythmic contraction of the heart muscle cells. About 10 types have been found in the CNS, mostly on glial cells. Neurons express connexin type 36 (Cx36) and most likely express other types such as Cx45, although less is known about that. Cx36 is exclusively neuronal and is also the main connexin in neurons. It can be found all over the brain. Knocking out Cx36 eliminates neuronal electrical coupling in various brain areas (but not all), but surprisingly the knockout animals have fairly normal phenotypes (but they do have retinal deficits so that they are nights blind, they have reduced power in the gamma band of the EEG, they have some motor control deficits, there are some deficits in circadian rhythms). Cx36 has a smaller pore (?) than other connexins so they are less permeable than other connexins to non-ion molecules.

A hemichannel can be homomeric (all subunits/connexins are the same) or hetermomeric (subunits/connexins not all the same). It is unclear whether Cx36 can connect with other types of connexins. A full channel can be homotypic (the two hemichannels/connexons are the same) or heterotypic (two different types of hemichannels/connexons), but the heterotypic seem to be uncommon.

Pannexins are another protein (super?)family that may be used in electrical synapses. Some pannexins are expressed in the brain. Vertebrates have both connexins and pannexins, whereas invertebrates only have pannexins and the closely related inexins. Connexins and pannexins have a similar structure and probably function as well, but they are completely unrelated in terms of amino acid sequence (parallel evolution).

Gap junctions and the role of development:
It was generally acknowledged that neurons were electrically coupled early in development. When people found evidence for them in mature animals people thought that it was just residual coupling from an earlier developmental stage, because most of the recordings were done in juveniles (it is easier to record from them because they aren’t as heavily myelinated to it is easier to see the neurons). Then people recorded from adult animals and still found as many electrically coupled cells as in juveniles (although the coupling wasn’t as strong).

Functional properties of electrical synapses:
Synaptic transmission through junctional channels is extremely fast, although chemical synapses can also be extraordinarily fast. No one knows for sure if this time difference is functionally important. Cx36 channels are permeable not just to monovalent ions but also to other small molecules such as calcium and IP3 (which can both function as second messengers). They can therefore have a role in things other than just synaptic transmission. Different subtypes of junctional channels have different single channel conductance, which could indicate that they mediate different kinds of activity.

Junctional channels are not all open at any given time, which leaves room for some kind of regulation of open probability. This is an area of active research. Electrical synapses are similar in size to chemical synapses, but the former are more stable (e.g. they lack short-term dynamics, such as paired-pulse facilitation and depression). Sometimes the channels are in a so-called subconductance state, where they are neither fully closed nor fully open. This happens if you apply a long-lasting voltage across the channels (some sort of activity-dependent modulation or regulation?). The IV-curve (relationship between current and voltage), however, is linear and Ohmic.

Junctional synapses behave as low-pass filters. Fast voltage changes (such as action potentials) are attenuated more than slow voltage changes (such as subthreshold passively conducted currents). This could be functionally important, e.g. calcium spikes are conducted better than sodium spikes, because calcium spikes are slower. This low-pass filtering is not caused by the channels themselves, but because of the membrane properties of the cells involved. The response of the postsynaptic cell will depend on the product of the membrane capacitance and resistance. The current needs to charge the capacitor before it can fully reach through the resistor. Therefore the postsynaptic cell can’t “keep up” with fast voltage changes. The presynaptic cell also has capacitance and resistance, so that membrane also needs to get charged and therefore it contributes to the slowing of the response.

Junctional channels are differentially voltage dependent. Conductance is maximal when there is no voltage difference between the two cells. Of course, no current will flow if there is no driving force. If you just change the voltage a little bit you will get a near maximal current flow. As the voltage difference increases the conductance decreases. Cox36 is very weakly voltage dependent, however, whereas Cx45, another connexin found in neurons, is fairly strongly voltage dependent. The gating is slow, so rapid events such as action potentials shouldn’t have a huge effect on the opening and closing of the channels. Maybe this voltage dependence is therefore not functionally important unless under abnormal circumstances (a sort of “defense mechanism” to close channels of cells that have gone haywire).

What are electrical synapses good for?

SYNCHRONY!

Cx36 knockout animals still have interneurons that oscillate, but each oscillates at its own speed (their activity is not synchronized). Gap junctions are normalizers or equalizers of voltage across cells (just a thought, could this play some role in homeostatic plasticity?). This can affect both the interneurons themselves (that are connected by gap junctions) and excitatory cells that they output to, because if two cells, A and B, get output from two synchronized cells, 1 and 2, A and be will be synchronized as well.

Interneurons are involved in coincidence detection (to find out if events go together in time) between coupled cells. When inputs reach the two cells at the same time no current flows between them because there is no voltage difference and therefore no driving force. When inputs come in non-synchronously current flows down the gap junctions between them and lowers the response to the synaptic input (cell 1 depolarizes and therefore loses some of its positive current to cell 2. When cell 2 gets its input it will depolarize more? Not sure what Scott was talking about here…).

Summary:

Gap junctions are everywhere. They change during development. Cx36 is the main connexin in neurons. Gap junctions conduct both subthreshold and suprathreshold potential changes. They behave like low-pass filters because of the properties of the cell membrane. They lead to neuronal synchrony.