Short term plasticity Flashcards Preview

N2 > Short term plasticity > Flashcards

Flashcards in Short term plasticity Deck (30):

What is short term plasticity?

Plasticity can occur from one firing to another - synaptic strength is highly dynamic even before LTP/LTD, and will depend on frequency and duration of presynaptic activity.

Any one or more of depression, augmentation, potentiation, facilitation (can combine for non-linear effects)

Lasts anywhere between 10s of msec and 30 mins

The change in strength is only by a few percent of control



Different frequencies of stimulation will cause depression vs facilitation, leading to a tuning curve of frequency vs plastic effect
This tuning curve is different for different types of synapse, and can't be assumed. The tuning curve conveys frequency-filtering effects (i.e. a facilitating synapse is a high-pass filter, depressing is a low-pass)
Also is terminal specific - so the same cell, stimulated in the same way, can generate a different type of plasticity at different synapses it has onto other cells. Divergence!


Quantal hypothesis - what is it? Discovery

Bernard Katz and colleagues, 1950s - looked at frog NMJ in low calcium Ringer. Found mEPPs followed binomial distribution, each was a multiple of a quantum.
synaptic efficacy = npq
n = number of vesicles (though now we think it corresponds more closely to number of release sites, or active zones containing clusters of vesicles)
p = proportion of vesicles released at each stimulation. Related to Ca2+.
q = quantal amplitude (postsynaptic response to one vesicle)
m = quantal content (np)
Palade and Palay 1954 - used EM to show vesicles at NMJ. People accepted vesicles = n.
Heuser and Reese 1979 - use freeze-fracture EM after stimulating in presence of calcium channel blocker 4-aminopyridine. Were able to count vesicles, and found they matched predictions from Katz's model.


Quantal hypothesis - to bear in mind

Each factor assumed independent for maths, but in reality interactions are likely - high p means n depletes quicker and depression is likely, low p will increase more over successive presynaptic spikes so cause facilitation.
So while a single synaptic event will favour some values, they may well 'balance out' over multiple spikes.
This means stimulating once and recording is not physiologically relevant.

BUT normally, assume increase p will increase m, because normally there are enough vesicles left over.


Against vesicle theory

Deuter et al -
Vesicles shown in EM might contain synaptic proteins, not NT. Supported by experiments using styryl dyes at frog NMJ. Destaining shows exocytosis, and coincides with postsynaptic response, BUT when you apply a PKC antagonist the destaining stops but there’s still a postsynaptic response!
Tauc 1987 -
Vesigate theory. Applying cytoplasmic AChE reduces postsynaptic response (which it shouldn’t do if ACh were in vesicles). Also from Tauc, AMECh was released alongside ACh immediately, way too fast for it to have been loaded into vesices (which takes several minutes). Also vesicles in the torpedo electric organ are way too big to be quanta.


n-p interactions

In the spinal cord, substance P converts glutamatergic interneuron synapses from depressing to facilitating. THis looks like a decrease in p. However, the first presynaptic input after SP application is not smaller (like you'd expect from a decrease in p), so maybe there's an increase in n to compensate.
This is a 'metaplastic change', because SP is increasing plasticity. This change takes about 10 minutes.


Sites of synaptic regulation

AP waveform (dependent upon axon noise - in the thinnest axons, AP form varies randomly not only between APs, but also as the same AP propagates along the axon
Ca channels
Ca buffering
Ca-ATPase (augmentation)
mitochondria and leftover Ca (PTP)
presynaptic receptors
postsynaptic receptors (phosphorylation can rapidly alter q)
readily releasable vesicles
reserve vesicle pool


Paired pulse facilitation

-When two presynaptic spikes come quickly after one another, the second causes a larger response.
-The longer the interval, the smaller this effect
-The interval-facilitation curve is a double exponential at some synapses, with two different time constants, suggesting two phases of facilitation.
-Perhaps this reflects two sites of Ca binding, or perhaps [Ca] decays non-exponentially, due to diffusion away from the active zone. Perhaps the slow one is augmentation occurring as well.
-At others it's one exponential. What makes the difference?


Spike shape

Serotonin is a neuromodulator in the Aplysia gill withdrawal system, blocking K channels to increase duration of AP, which increases the amplitude of the postsynaptic response. Increasing spike width presumably increases calcium signal at the terminal, which increases quantal content via increasing p.
Different synapses have different spike shapes
Spike shape can change over the course of repetitive activity, due to activity-dependent changes in certain ion channels

**BUT** there are exceptions - in the jellyfish, broadening AP decreases transmitter release, because most calcium entry happens on the repolarising phase of the AP, and broadening the spike inactivates these channels before much calcium has entered.


Facilitation, possible mechanisms

Occurs over short interspike intervals, decays rapidly (lasts msec). Believed to be due to leftover calcium allowing more vesicles to be primed.
-linear relationship between enhancement and presynaptic calcium
-blocked by calcium chelator EGTA
-increased by deletion of calcium BP parvalbumin at cerebellar synapses
-could also be due to changes in calcium channels (reducing calcium entry decreases facilitation, but hard to interpret as also increases depression)
-could be saturation of calcium buffers
-could be activity-dependent removal of polyamine AMPA receptors block postsynaptically (as block is removed, activation of receptors is increased)



Results from brief stimulation, longer-lasting effect (seconds)
-linear relationship between enhancement and presynaptic calcium
-Prolonged stimulation loads the terminal with Na and Ca
-The consequent reduction in Na gradient reduces Na/Ca exchanger rates
-Buffers become saturated, in equilibrium with cytoplasmic Ca
-Residual calcium decays with fast and slow components - the slow is due to leakage out of mitochondria, which is faster due to the increased cytoplasmic sodium (which is given buffering priority)



Results from sustained stimulation, longer-lasting effect (minutes)
Due to increased migration of vesicles from the reserve to the readily releasable pool
-Some think there's a contribution from reversal of Na/Ca exchanger, like augmentation but after longer tetani



Presynaptic - vesicle depletion (takes 40s for a vesicle to recycle. After depletion, recovery takes seconds to minutes)
Presynaptic - spike narrowing (although less scope for reduction than broadening, cell specific duration)
Postsynaptic - receptor desensitisation (blocking desensitisation with cyclothiazide reverses this depression)
Postsynaptic - NTs still bound to receptors, so they can't respond to a second stimulus.


Depression mechanisms at the calyx of Held

Shows a pronounced depression, to help with sound localisation
depression correlates with reduced calcium currents presynaptically.
But depression here is blocked by cyclothiazide, which blocks AMPA desensitisation.
Desensitisation happens at high frequencies >100Hz
So maybe both mechanisms are involved, at different frequencies.


Synaptic vesicles cycle

1) vesicles move from reserve pool
2) vesicles dock using Rab3/27 and active zone protein RIM
3) vesicles primed for fusion
4) Ca2+ opens fusion-pore
5a) vesicles recycle locally immediately after pore opens (kiss-and-run)
5b) vesicles endocytosed via rapid clathrin-independent pathway (40-60s)
5c) vesicles endocytosed via clathrin-dependent pathway and filled


Vesicle pools - what are they?

One or more vesicles docked and primed at the plasma membrane = immediately releasable pool
Vesicles docked but not primed = readily releasable pool
Vesicles tethered away from active zone = reserve pool (regulated by synapsin)


Ca-dependent replenishment of RRP - lamprey e.g.

Excitatory synapses are depressed by 5Hz, but plateau at 20Hz. This effect is Ca-dependent (blocked by EGTA, synapse fails in EGTA because there's no refill so vesicles deplete). Perhaps an equilibrium is reached here between vesicles exocytosis and refilling from reserve pool.

This is necessary because the lamprey swims constantly


Ca-dependent replenishment of RRP - mouse Calyx of Held e.g.

Stronger stimulation evokes greater depression, but also faster recovery. Perhaps more calcium enters when more strongly stimulated.
Broaden the AP using TEA (so more vesicles are released), and depression and recovery are both faster (total recovery in 200msec vs a second or two in control)
block calcium channels with cadmium and this super fast recovery is blocked


Pool sizes and dynamics - overview

RRP is always a small proportion of total vesicles
90% of vesicles are in the reserve pool
Absolute number varies - 200 in mouse hippocampus, 4000 in frog NMJ
-NMJ is one nerve onto a muscle fibre that /must/ contract, so more vesicles are released to ensure this
-hippocampus is up to 30,000 synapses onto one purkinje cell, to be activated intermittently
Vesicle movements required for sustained release
Must affect short term plasticity


Methods that have been used to determine pool size

For small pools, you can use serial sections under EM and count.
For bigger pools, you can outline the whole area, work out volume, count a sample section and multiply up
Readily releasable pool - people have tried getting all these to release before the pool can be refilled, typically using osmotic shock (sucrose solution). But this is a crude estimate.


Synapsin - original hypothesis

Synapsin, only found on vesicle membranes away from the active zone, is the 'glue' tethering them to the actin cytoskeleton in the reserve pool. CamKII releases them. Synapsin KO gets rid of reserve pool but doesn't affect RRP. Synapsin KO synapses depress faster and further (to 30% rather than 60% of control function) but do not fail.


Synapsin - in more detail

Synapsin I - KO has little effect. Ca increases ATP binding.
Synapsin II - prevents PTP. Reduces rate of release to 50% of control. Ca has no effect on ATP binding.
KO of I and II has a synergistic effect
Synapsin III - *reduces* depression of the synapse! So v different role, hasn't been studied as much (unlike I and II, it is primarily localised to extrasynaptic regions. Thought to have a more important role in neural development. Lower adult expression than I and II, different developmental regulation). Ca decreases ATP binding

Synapsin KO terminals are smaller, so maybe just can't hold as many vesicles.
Synapsin KO causes reduced axonal growth
Ribbon synapses have no synapsin, despite being some of the most active synapses in the body. But they're constantly recycling, so maybe no need for a reserve pool?


Proteins involved in synaptic vesicle fusion

'closed' Syntaxin-1 binds Munc18-1 to 'open'.
SNARE complex assembles
Fusion core complex (synaptobrevin, syntaxin, SNAP-25) is a high affinity binding site for alphaSNAP, which binds NSF
Ca and Munc18-1 necessary for fusion pore
Endocytosis and SNARE complex disassembles.



Munc18 interacts with SNAREs and syntaxins at vesicle fusion
Munc18-1 primes vesicles for fusion
Munc18 KO stops all synaptic activity, causes 'silent mouse'



4 isoforms. 1 and 2 = forebrain, 3 = cerebellum, 4 = peripheral
KO 1, reduce glutamatergic synaptic activity
KO 1 AND 2, reduce GABA-ergic synaptic activity
Stops osmotic shock (calcium-dependent) release
No effect on latrotoxin (calcium-independent) release
Overexpression of Munc13-1 = depression in hippocampal excitatory neurons
Overexpression of Munc13-2 = facilitation

Perhaps interaction with Munc18, to aid binding to syntaxin and increase speed of pore formation?
Maybe a correlate of p, since different isoforms have different effects on STP? But also changes n...



RIM1/2 are components of the active zone, interact with most AZ and vesicle proteins
RIM1 binds Munc13 for priming and Rab3 for vesicle fusion
Loss of RIM-Munc interaction reduces RRP
RIM (Or Rab3) KO at excitatory synapses --> PPF
RIM KO at inhibitory synapses --> PPD
Both also reduce Munc13 levels
Elevated Ca could facilitate RIM-synaptotagmin interactions to give PPF



A reduced affinity mutant converts synapse from depression to facilitation. An increased affinity mutant does the opposite. Proposed to be a correlate of Katz's p
Synaptotagmin 7 is needed for facilitation in diverse synapses, but does not affect calcium influx, buffering, or initial p
Reexpressing WT Syt7 in KO synapses reinstates facilitation, but reexpressing a version with a mutated Ca binding domain (CA2) does not.
Suggested as a calcium sensor for low levels, that will transiently increase p with repetitive activity


Functional importance of STP

Experimental manipulations are difficult and often non-specific, so computational models are used a lot
Also, inhibitory synapses function completely differently, with different plasticity effects.

-frequency filter - level of depression in cells in chicken auditory processing are tonotopically arranged, as inputs from cells with higher characteristic frequencies have less depression, producing a more robust response

-Adaptation and sensitisation - Nicolaev et al 2013 - Retinal ganglion cells in zebrafish either facilitate, and are sensitised to contrast, or depress, and adapt to contrast. Facilitation in bipolar cells is mediated by depression of inhibitory feedback of the amacrine cells synapsing onto them, an example of STP having a circuit-level effect

- Gain control - Nucleus laminaris cells function as coincidence detectors in sound localisation, independent of sound intensity, because of depression at the synapse.

-Directional selectivity in electric fish - movement in the preferred direction activates slowly depressing inputs first, and subsequent inputs summate with this barely depressed activity to produce a strong activation. Movement in the opposite direction activates rapidly depressing inputs first.

-Maintaining phase constancy in CPGs (as opposed to constant latency), so that latency scales with period of oscillation.

-Speed and robustness - in the lamprey locomotor network, depression at excitatory synapses increases pattern frequency (by terminating activity quicker), and facilitation at inhibitory synapses prevents this from advancing out of control.


Multiple site hypothesis

It was originally thought that Ca acts at a single site to cause vesicle release. The only mechanism behind all enhancement (facilitation, augmentation, PTP) is residual calcium (Ca + Ca + Ca


Presynaptic vs postsynaptic mechanisms

Postsynaptic - homologous/heterologous desensitisation; activity dependent removal of AMPA block
Presynaptic - residual calcium; reversal of transporters; saturation of buffers; movement from reserve pool; AP shape alterations; altered calcium influx

Costa et al 2015 - made a computational model incorporating pre-and postsynaptic changes, modelled after horizontal connections in visual cortex. In a learning and relearning task, presynaptic changes were erased way faster. Relearning was faster than learning as long as postsynaptic changes had not yet fully decayed.