Synaptic plasticity

Synapses are the gaps between two communicating neurons. Most synapses, especially in the brain are chemical synapses, this means there is no contact between the two neurons and a chemical transmitter is used to traverse the gap. these transmitters are neurotransmitters, there are some that are exclusively to activate the next neuron, such as glutamate and those that exclusively inhibit the activation the next neuron, such as gamma aminobutyric acid (GABA) or glycine, then there are those that can be either depending on the receptor on the post synaptic membrane that they are binding to, for example serotonin (5-hydroxytriptamine) will activate the next neuron if it binds to the 5-HT2A receptor but can inhibit its own release if it binds to 5-HT1A receptor which is present on its own neuron and will stop further release.

It is the interaction between these neurotransmitters and receptors that govern whether or not the electrical signal is passed onto the next neuron, this is because the excitatory transmitters will depolarise the next neuron (depolarising makes the neuron more positive = activated) but inhibitory transmitters will hyperpolarise the membrane this means more excitatory neurotransmitter is needed to reach threshold for activation, If the previous signal was not strong enough to provide the extra transmitter release then the post synaptic neuron remains/becomes inactivated so the signal is stopped. This is seen when we smell something, at first the smell is very prominent but it gradually goes away, if it is a perfume it will stay in the air quite a while, yet we soon stop smelling it. The perfume is still going into the nose and therefore there is still a stimulus activating GPCRs on the nasal neurons, however we can no longer smell it. What is happening is that the first neurons are still activated by the stimulus and synapsing in the olfactory bulb but at some point in the chain of neurons (most likely more than one point in the chain) from nose to the olfactory cortex there has been GABA released onto the postsynaptic neuron to hyperpolarise the neuron. This is strong enough to push the neuron below threshold potential, therefore even though the same amount of glutamate is being released to the synapse, it cannot depolarise the next neuron enough to cause a response. This is a temporary change, so if you leave the room and come back a few minutes later, you can smell that same smell again. This however, is not synaptic plasticity, it is a temporary change that acts to stop excessive signalling or to stop signalling when it is unnecessary. Dopamine in the striatum is both excitatory and inhibitory, it acts constantly in conjunction with GABA on motor neurons here to stop excessive movement like twitching or hand tremors, this is why early signs of dopamine neuron loss in Parkinson's disease include hand tremors.

So I've basically just taken signalling that seems straight forward and took a hatchet to it and it gets more detailed. But the important bit is the understanding that neurons are governed by up to hundreds of synapses, some excitatory, some inhibitory and understand that these mechanisms are part of the normal functioning of synapses in all creatures and that the type of receptor generally governs the response of the neuron, not the neurotransmitter.

Synaptic plasticity is not as simple as the above which is just changing the level of neurotransmitter to find the correct signalling level and it is more permanent. What synaptic plasticity is, is the synapses ability to change upon repeated exposure to the stimulus, either by strengthening the response to the stimulus or weakening it. This occurs through two processes, long term potentiation or long term depression and they involve changes in phosphorylation, changes in translation and changes in transcription.

If we consider the components in the synapse that could be changed in either structure or quantity, the first thing is the number of receptors and the state of receptors (ion channels have 3 states, open, closed and inactive, changing the time spent in these states will allow the rate of depolarisation). The second thing is the level of neurotransmitter being released to the receptors. Third is the rate of protein synthesis, not just of receptors but of enzymes, co-enzymes and intracellular signalling proteins. The forth main thing that can be changed is the rate of transcription. As may seem obvious when reading these changes, they are all hugely inter-related which is important because synaptic plasticity requires various different processes working at the optimal time to cause a sustained change to the synapse. An example I will be referring to a lot is memory, transcription takes a while to take effect so if this was the first or only process, the repeated stimulus would die away before an increase in transcription could take effect therefore no synaptic plasticity, so when trying to commit something to memory, by the time you had tried, you would forget because there would be no processes to sustain that stimulus until transcription rate increased.

This may seem a little vague and a lot of bouncing around between information at the moment so I will make it more clear. However, anyone getting into neuroscience, there is a lot of bouncing around between info trying to take sections in and then ending up nowhere near the topic you started on and spending hours or days to weeks trying to link up things that you would never imagine. So I am going to make more of an overview of the system before I get into the details.

Overview: long term potentiation requires a series of stimuli onto that synapse in rapid succession (high frequency stimulus) this causes sustained depolarisation of the synapse through GPCRs but also if it is glutamate released, through NMDA receptors. When a stimulus causes glutamate release it binds to AMPA receptors allowing sodium ions into the neuron= depolarisation. repeated stimuli depolarise the membrane such that NMDA receptors open, these are calcium ion channels that allow calcium in. Calcium sustains the depolarisation as well as acting as an extremely important second messenger. Calcium will begin a signalling cascade that will lead to the phosphorylation of existing AMPA receptors and movement of vesicle bound AMPA receptors to the membrane. Both of these processes increase the strength of the signal/synapse because more channels/longer open = more depolarisation in response to even a lesser stimulus. Calcium is an activator of kinases which phosphorylate but this will also occur on initiation factor binding proteins to increase the synthesis of proteins. A lot of these proteins will be AMPA receptor subunits so will further increase the number. This is still temporary so calcium and kinases are also involved in increasing the activation/phosphorylation of transcription factors which will increase the number of mRNA encoding for AMPA subunits which is a more long lasting change that has the same effect as the previous steps. The presynaptic terminal is largely separate from these changes but also incurs its own changes.

I will now break it in to the stages and describe each process in more detail, however, I promise not to bore anyone with hundreds of specific examples that essentially have the same process, so I will use as few examples as possible.

The presynaptic terminal in this sense is mainly controlling the release of neurotransmitter to the postsynaptic membrane. Neurotransmitters are not free in the cytoplasm but are stored in vesicles. Peptide transmitters in dense core vesicles and all others in small clear vesicles. The small clear vesicles lie in the axon terminal near the active zone in readily releasable pools. These are pools of vesicles ready to be activated to fuse with the membrane and release the transmitter. In order to do so the vesicles are primed which means they are ready to fuse with the membrane in response to calcium ion entry. The calcium binds to calcium sensor synaptotagmin which allows the interaction between synaptobrevin and the membrane SNARE protein SNAP-25. As you may imagine, increasing calcium entry would not really increase the release if it only bound to synaptotagmin because only a certain number of vesicles have been primed, therefore there must be a mechanism for increasing the priming of vesicles. Priming requires a family of proteins called Munc13 proteins, these proteins can be bound by DAG and other second messengers to regulate them but there has long been proof that calcium levels increase along with priming and synaptic plasticity, therefore there must be another mechanism. It has now been proven that Munc13 also has a binding site for the messenger calmodulin, but only when calmodulin is bound to calcium ions. Therefore, upon repeated stimulation, voltage gated calcium channels are open longer so excess calcium enters which binds to calmodulin, causing its binding to Munc13 to increase its activity of interacting with syntaxin (SNARE protein) causing an increase in priming rate and fusion rate of vesicles. As you might expect, this is only a short term change because a decreases in stimulus frequency will reduce calcium channel opening and the calcium-calmodulin-Munc13 complex is transient.

However, for a short time it keeps the binding of glutamate to AMPA and kainite receptors high, which means they stay open longer and cause sustained depolarisation. So why is sustained depolarisation important?

The image above is of an NMDA receptor. It is different to AMPA and Kainite receptors as it allows both sodium and calcium ions into the neuron, it also requires the binding of both glutamate and glycine to open (as you will recall I said glycine was inhibitory and glutamate excitatory which is confusing to say the least so just do what I do and run away from that realisation). In addition, once they have bound and the channel is active, it still is not open, as you can see from the diagram there is a magnesium ion blocking the channel. This ion is positively charged, so the neuron needs to sustain depolarisation (become positively charged) long enough to repel the magnesium ion because the positive ion will not want to be in a positively charged area and so moves out, unblocking the receptor. However, if the membrane is already depolarised, the driving force of sodium and calcium is also out so they will not move in either, even though it is open (seems pointless). But after the stimulus the membrane will repolarise, usually causing the magnesium ion to re-enter, unless the next stimulus is very soon after. So because the glutamate level is high in the synapse, the next stimulus and all subsequent stimuli appear at a high frequency, this means there is no time for magnesium to re-enter and now the NMDA receptors can have calcium move into the neuron to depolarise it. This receptor and calcium influx is the cornerstone of memory and a lot of long term potentiation.

The immediate impact of the calcium influx is its activation of Src tyrosine kinases that phosphorylate AMPA receptors in readily releasable pools (postsynaptic membrane) which moves them to the membrane which increases the number of AMPA receptors to strengthen the signal after the presynaptic frequency is reducing and thus the glutamate in the synapse is reducing (this is why first mechanism is short term). Some of these AMPA receptors with be homomers (GluR1-GluR1) rather than the usual GluR1-GluR2 arrangement, the homomer allows calcium in too so adds to calcium concentration.

The third mechanism is also due to calcium influx but is not as immediate. I will explain this in a moment, but first I want to clear up some myths/oversimplifications that people have been taught for years. mRNA is indeed single stranded and enters ribosomes where it is read to produce a protein. However it is not naked or linear like in (A) which is how many of the people reading this will have seen it. mRNA consists of a 5' (5 prime) untranslated cap, an encoded region then a 3' untranslated region and finally a poly adenine tail, but is still not straight as seen in (B). The best representation is this near circular structure seen in (C) and the reason this is the correct representation is because the untranslated regions on the mRNA are not wasted space but areas for proteins known as factors. At the 3' end there are poly A binding proteins and general cellular factors. Poly A binding proteins will bind to all mRNA as all have a poly A tail so are often export factors. Those that bind to the untranslated region can only bind if the correct codon is present (so matching pairs). These factors can be stabilising to make the mRNA survive longer or transport factors which take the mRNA to a specific part of the neuron i.e. the synapse that is undergoing plasticity. At the 5' end proteins called initiation factors bind which control the rate of translation, but these factors need to interact with the factors on the 3' end, this is only possible if there is a more circular structure like in (C), in (B) the two ends are too far apart.


(A)

(B) http://www.pier1.dk/dropshippers/peranan-guru-mrnangani-masalahdisiplin-di-sekolah

(C) http://humanbiologylab.pbworks.com/w/page/50107100/polyribosomes

Another thing that is often not mentioned is that many ribosomes can bind to the same mRNA and translate, it is not just one, so the process is slightly more complicated.
Now that is a bit clearer and we have established that mRNA can be transported to the synapse that needs it before being translated it might be becoming more clear as to why protein synthesis is needed for synaptic plasticity, because as we have said, there are only so many pools of vesicles and only so long phosphorylation lasts, once the previous steps stop everything would return to normal, so no lasting strength = no memory. But how do we increase the rate of protein synthesis? in simple terms we activate the initiation factors on the mRNA, but because it is neuroscience, it is not that simple.

An initiation factor interacts with the ribosome to allow translation, this is needed for all protein synthesis so how can it be increased? well rather than just interacting with the ribosome, the initiation factor also interacts with an initiation factor binding protein. When it is interacting with the binding protein it cannot be interacting with the ribosome, thus regulating translation. These interactions are at equilibrium under normal circumstances so it is 50/50 which the initiation factor is bound to. As you may now expect, there is a process which reduces the ability of the binding protein to bind to the initiation factor which means the initiation factor spends longer interacting with the ribosome thus increasing the rate of protein synthesis. This process (as with all the other processes in synaptic plasticity) is localised to the synapse that is changing rather than affecting the whole neuron and it is localised because of this excessive calcium influx at just this synapse. The calcium activates a signalling cascade to increase protein kinase levels, the kinase will phosphorylate the binding protein which makes it unable to bind to the initiation factor, thus reducing the inhibition on the ribosome allowing more protein to be produced.

Putting this into a quick but specifically related example may help. The mRNA for AMPA subunits have the initiation factor eIF4E at the 5' end. eIF4eE interacts with the ribosome and also interacts with its binding protein eIF4EBP to form the equilibrium. When calcium enters due to the NMDA receptor opening it causes the activation of mTORC1 which is a multifunctional protein including acting as a kinase to phosphorylate eIF4EBP, this prevents it binding to eIF4E to allow more AMPA receptor subunits to be translated and therefore more AMPA receptors that can be added into the membrane to keep the signal strong. This is still not long term though because the phosphorylation can only last 120 minutes (I will explain how this is known at the end). So there must be another mechanism.

This final mechanism is the changes to transcription, this happens in a lot of different ways, from directly interacting with promoters and polymerases to changing histone structure, most of which occur in synaptic plasticity via the action of dopamine activating cAMP. I am going to look at one example which has been shown to be of huge importance in memory. This example is the transcription factor CREB, this stands for cAMP response element binding-protein which kind of hints that CREB binds to cAMP response element, this is a small section of DNA within the promoter region of some genes.

A promoter region has sites for many transcription factors, what will happen is that transcription factor D will bind because it has a TATA binding protein that binds TATA on the DNA promoter region and causes the DNA to bend here, this allows TFB to bind which allows the RNA polymerase to attach and start making mRNA, so the faster the TFs can be recruited the faster mRNA is made. As you can see I have not mentioned CREB yet, well it binds in the same region but instead of directly increasing transcription, when CREB binds it needs to be phosphorylated (done by a kinase that calcium activates), once phosphorylated CREB binding protein will bind to CREB this then indirectly increases transcription. It does this because the CREB-Pi-CBP complex will modify the chromatin by acetylating (adding acetyl group) the histones that make up the chromatin in the AMPA subunit gene. DNA is wrapped around the histones to compact it which means it cannot be accessed by polymerases, CREB acetylating the histones in the area causes them to uncoil enough for the other transcription factors (D and B have been mentioned) and the polymerase to bind the promoter region faster and because acetylation is long lasting the promoter region can be activated for longer. The net effect increasing the number of mRNA for AMPA subunits being produced. This process takes the longest but once the mRNA is increased, protein synthesis can return to normal as it is having the same effect and the first two short term mechanisms can return to normal because there is more mRNA thus more AMPA receptors at the synapse. The synapse has now been strengthened so that when that synapse receives a stimulus, even at much lower levels it is able to depolarise and therefore contribute to the memory of a specific piece of information. It also means that when receiving stimuli calcium can enter and will cause CREB to acetylate the histones to keep that synapse strong (this is important because there are groups of proteins that deacetylate histones so could cause memory loss (more information will come when looking at neurodegenerative disorders)).

Just a quick word on some of the evidence for what I have said.

Phosphorylation of IF binding proteins only lasts 120 minutes: A drug called actinomycin D inhibits transcription, when a synapse is potentiated, the signal from it increases to above baseline transmission. when giving a synapse high frequency stimulation to cause potentiation and then treating the neuron with actinomycin D transmission increases above baseline for 120 minutes (potentiated) but after 120 minutes it returns to baseline. This is because there was no transcription so after 120 minutes translation returned to normal because the binding protein was no longer phosphorylated. This also shows the transcription is needed for synaptic plasticity.

Without protein synthesis potentiation occurs but is short: doing the same as above but treating with protein synthesis inhibitor anisomycin, transmission increases above baseline for a few minutes (due to phosphorylation of AMPA vesicles and open NMDA receptors) but immediately decays to baseline.

NMDA receptor involvement: Teaching mice to freeze in response to a noise = learning and memory. Then give the mice an NMDA receptor antagonist, the mice will then stop freezing in response to the noise showing that NMDA receptors are needed for memory. If these mice are made transgenic to express the gene LacZ whenever CREB binds to the promoter it can show the link between NMDA receptors and CREB. This is because LacZ produces beta galactosidase which is not normally found in mice and is easily detectable. So when the mice freeze, LacZ is expressed and beta galactosidase is produced meaning that CREB is binding to the promoter, when the NMDAR antagonist is introduced, the expression of LacZ stops meaning CREB is not binding, so NMDA receptors must be linked to CREB activation in a signalling cascade.

CREB involvement can also be identified by knocking out CREB in the hippocampus (where memories are formed) of mice and comparing their learning and memory of freezing to normal mice, the mice could learn to freeze (all processes up until transcription are unaffected so short term potentiation including protein synthesis will occur) but when left for two hours and then repeating the noise the CREB knockouts do not freeze as acetylation has not occurred to increase transcription and therefore they have no memory of learning the task.