In the last video,
we saw that there are two important major types
of ligand-gated glutamate receptors.
These are the receptors that are sitting in the postsynaptic membrane,
and respond through the release of glutamate from presynaptic terminals.
We explored the function of these ligand-gated ion channels
in isolated patches of membrane, and found that there were
two different types of glutamate receptors:
the AMPA receptor, that was extremely fast,
and the much slower NMDA receptors,
that in addition were found to be voltage dependent.
In this lesson we'll explore how these receptors function
at real synapses, in situ, in the brain.
These receptors are, of course, placed in the postsynaptic locations
immediately adjacent to presynaptic release sites,
so the action potential invading an axon,
and a presynaptic nerve terminal,
will cause release of glutamate from the presynaptic specialization,
and that will act on the postsynaptically located receptors.
And so now instead of considering what happens in response
to a 1 millimolar pulse of glutamate for 1 millisecond,
we'll consider the more physiological situation
of where an action potential evokes release of glutamate,
and that acts on these receptors present in the postsynaptic membrane.
We'll also think about how the integration of those postsynaptic conductances
occurs in the postsynaptic neuron, and how those conductances
are changed into membrane potential changes
that in their turn can effect the ligand-gated ion channels themselves,
through the voltage-dependence of the NMDA receptor,
and we'll also think about what happens in terms of dendritic integration,
and how those potentials are filtered across the dendrites
before they reach the final integration site
in the soma and the axon initial segment.
So here now we have a schematic drawing
of what happens at a real synapse,
where we're considering a presynaptic neuron,
that fires an action potential,
perhaps in response to a current injection in the experimental setting,
and simultaneously, we're recording from a small, compact cell,
that's voltage-clamped and synaptically connected to the presynaptic neuron.
So the action potential invades the presynaptic nerve terminal,
it releases glutamate, and with a short delay of around 1 millisecond,
there's an activation of the postsynaptic glutamatergic currents.
And at a negative membrane potential postsynaptically,
the total synaptic current that's recorded here is brief, and fast,
and it is almost entirely dominated by the AMPA component.
AMPA receptors, just like the isolated patch measurements
that we discussed in the previous video,
give rise to short, 2 millisecond or so duration conductance changes
that here are then measured through the voltage clamp amplifier
in terms of a excitatory postsynaptic current.
That's an EPSC, an excitatory postsynaptic current.
The NMDA receptors are, however, largely blocked at negative potentials.
The magnesium ions from the extracellular solution enter the NMDA receptor,
and block it at negative potentials.
And so at typical resting membrane potentials of a neuron,
somewhere around -60 mV,
the NMDA conductance is basically shut.
The AMPA conductance is large,
and the total excitatory postsynaptic current, the EPSC,
has this fast kinetics, dominated by the AMPA receptor.
If we now change the postsynaptic membrane potential,
still in voltage clamp mode,
so now we voltage clamp our small, compact neuron
to be at +60 mV,
the AMPA receptor looks as before,
except now, of course, the current is outward, rather than inward,
we're sitting over here,
reversal potential for AMPA and NMDA receptors is close to 0 mV,
so the amplitude and kinetics of the AMPA receptor are similar,
but now the NMDA channels can contribute.
Magnesium has been expelled
from the ion channel pore, it no longer blocks,
and we get a considerable current flowing through the NMDA receptor,
and it's the summation of these two that give rise
to the total EPSC recorded at positive potentials.
A fast AMPA component, and a slower, longer-lasting NMDA component.
Now different synapses have very different ratios
of AMPA and NMDA receptors that are present.
Some synapses are almost entirely dominated by AMPA receptors,
whereas other synapses are almost entirely dominated by NMDA receptors.
But in most synapses, there's a fair mix,
where both AMPA and NMDA receptors
are present in roughly even quantities
in terms of the total synaptic currents.
And so although there are cases
where there's a large AMPA, and just a very small NMDA,
or small AMPA and very large NMDA,
most of the time there's a mix, and both AMPA and NMDA
contribute importantly at most glutamatergic synapses.
Now we'd like to know, in the physiological situation,
where the postsynaptic membrane is <i>not</i> voltage clamped,
but rather that the membrane potential is free to swing
under the influence of these ionic channels
that are present in the plasma membrane,
how now do these postsynaptic conductances
of the AMPA and the NMDA receptor
interplay with membrane potential?
And so we imagine that we have a membrane potential recording
from a small, compact neuron that receives a glutamatergic input.
The action potential comes down,
the axon releases glutamate,
and of course it activates the ion channels that are sitting here,
the glutamate receptors, the AMPA and the NMDA receptors,
and that of course causes then a current flow
through the AMPA and NMDA receptors.
And we've already considered the relationship
between current flow and membrane potential
in our electrical equivalence circuits,
where we found that the phospholipid bilayer
could be thought of as a capacitor,
that we draw here, in the electrical equivalence circuit,
and the individual ion channels, with their different ionic permeabilities,
can be thought of as conductors, or resistors, alternatively,
with a specific driving force relating to their reversal potential.
So we have potassium, sodium and chloride conductances,
and now, for the synaptic conductances,
we can add the AMPA and the NMDA conductances
that have reversal potentials close to 0 mV,
and that then are involved in determining the membrane potential of the cell.
We also already know the quantitative description
of how membrane potential and currents interrelate.
So the total membrane current is made out of different components.
There's the capacitative component,
that's the filling on charging of the capacitor,
that then relates to the change in membrane potential over time,
so that's the typical equation here, from Q = CV,
we take the derivative of that, and that then gives us
that the current is equal to dVm/dt times the capacitance.
We also have our various leak conductances
relating to potassium, sodium and chloride conductances,
that we can also just rewrite in a single term, here,
that the leak is equal to the membrane potential,
minus some leak potential,
times the leak conductance overall,
and the leak here
is simply the resting membrane potential of the cell
in the absence of synaptic conductances,
and typically that's sitting at around -70 mV.
Now we also add the synaptic conductances here,
and of course the current flow through each of these synaptic conductances
relates to the driving force, or the membrane potential,
the 0 mV reversal potential,
so it's simply the membrane potential,
times the conductance of the different channels.
And of course for the NMDA conductance
we also have to, in addition, think that this is actually voltage dependent,
in addition, of course, to its dependence
upon the presence of glutamate released from the presynaptic terminals.
Now, importantly, when we think about these currents,
and their relationship to voltage,
we need to think about the charging and discharging of a capacitor.
And so there's a big difference between the synaptic currents that are flowing,
and the influence on the membrane potential,
both in terms of how the leak conductances
will determine how large that potential is, postsynaptically,
how large the EPSP is in response to a given conductance,
and of course the membrane capacitance will also affect
the kinetics of how those currents are translated
into membrane potential changes.
And so if we now try to directly compare,
for the same synapse, the same conditions,
presynaptic action potential gives rise to a fast EPSC.
If we're talking about hyperpolarized conditions at -70 mV,
we only need to think about AMPA receptors,
the NMDA receptor is blocked,
so we have these brief, transient AMPA conductances,
lasting about 2 ms, and that then generates a 2 ms EPSC,
the current follows the conductance in voltage clamp mode,
and we see these rapid synaptic currents.
Now in terms of the membrane potential,
the situation is a bit more complicated.
So we imagine, again, a synapse here, the release of glutamate.
We have the synaptic currents that are,
as stated here, about 2 ms in duration,
then the conductance closes, and that is then closed,
but there's still a lot of charge that entered during that EPSC.
And the K phase of the EPSP is determined
by how that charge then discharges the capacitor,
and that is, of course, through the leak conductances.
So we have the leak conductance here,
we have our membrane capacitor here,
and essentially the charging that occurs due to the EPSC,
that causes the depolarization of the membrane,
the decay phase of the EPSP, that's determined here
through the leak conductance and the membrane capacitance.
And you'll remember from our previous studies
that there's a membrane time constant that is related here
to the membrane resistance, times the membrane capacitance,
and of course the membrane resistance here
is simply 1 over the membrane conductance,
or the leak conductance, if you like.
And typically these membrane time constants
are on the order of many tens of milliseconds,
so it's somewhere 10, 20 ms in duration
for the membrane time constants,
And that's then the conversion factor, if you like, between the EPSC,
which has this rapid time constant,
and the EPSP, that's through the filtering here,
the discharge properties of the membrane capacitor
that has a built-in time constant sitting somewhere around 10 ms,
and so there's a major difference
in the time course between EPSCs and EPSPs,
when EPSP is a much slower and longer-lasting event
than the underlying conductance that drives it.
So far we've been looking at schematic example-type drawings
of how EPSCs and EPSPs are.
Now I'd like to look at real data,
how synaptic transmission looks like in physiological conditions,
at least in brain slice-type recordings.
In this particular example, we're recording simultaneously
from three different neurons.
There's a presynaptic neuron that we stimulate to fire an action potential
at a well-controlled point in time,
and simultaneously we record from two postsynaptic neurons,
a red neuron and a blue neuron,
each of which is synaptically connected,
and so the action potential here propogates down the axon,
releases glutamate onto the red cell,
and also simultaneously releases glutamate onto the blue cell.
And it turns out that the properties of the postsynaptic potentials
are really very different in this example,
between these two cells, postsynaptically.
These potentials can be thought of as unitary potentials,
in the sense that a major and important unit of information in the brain
is one cell firing a single action potential.
And we then measure the so-called unitary postsynaptic potentials again
in single cells, at the somatic location here,
how the action potential of one neuron here
influences the postsynaptic membrane potential
of two other neurons.
And in this particular example, an action potential in the black neuron
induces a large depolarization in the red neuron,
so let's look at the scale bar here.
This is about 0.5 mV depolarization,
lasting some tens of milliseconds in the red cell,
whereas in the blue cell you'll see that the amplitude is much smaller,
there's a different scale bar,
it's about ten times smaller,
and also the duration of the PSP in this case is also much faster,
the decay time here is maybe 10 ms,
whereas here maybe it's 30 ms,
and so there are differences in the amplitude
and the kinetics of the responses,
and in addition there's also an obvious difference
in the reproducibility, or the variability of the postsynaptic potential.
Here we have ten consecutive trials,
and action potential is evoked in each one,
and in the red cell you see that there's
a relatively constant depolarization that's measured.
Here we superimpose those ten trials, and you see that it's almost
the same postsynaptic depolarization we see in the red cell,
whereas in the blue cell there seems to be enormous variability,
sometimes there's a response, sometimes there isn't,
and that corresponds to failures and successes, as we record here.
And we can envisage a sort of a simple explanation
for why we might have this type
of large-amplitude, reliable transmission,
compared to this small-amplitude, unreliable synaptic transmission,
and part of the reason for this explanation then might be
that there's many different synapses here,
that actually independently form
release sites and postsynaptic contact sites
between this single axon here.
It might make multiple contacts, say it might make 10 different synapses,
each of which can independently have released glutamate onto this red cell
when an action potential is released,
and that will then increase the amplitude of the response,
and in addition decrease the variability,
because all though these might be independently releasing glutamate or not,
in the probabilistic manner of synaptic transmission,
if we have 10 synapses, then the chance that some of them
will be releasing simultaneously, and the variance here will decrease,
whereas for the blue cell, it really might be just a single synaptic contact,
and then the action potential
either causes release of a neurotransmitter or it doesn't,
and then we can get this all-or-none type of synaptic transmission.
And of course there's also some smaller amplitude responses,
and that might then correspond to vesicles
that have less glutamate in them than other vesicles.
So typically when we think about synaptic transmission in the brain,
there are small-amplitude PSPs, and even this one,
which is amongst the larger that are seen at most synapses.
This is still a small depolarization, compared to what's needed
in order to fire an action potential.
So here the postsynaptic cell might be at -70 mV,
we get a 5 mV depolarization,
but we're still more than 20 mV away from action potential threshold,
and so unitary inputs onto most cells
don't cause postsynaptic action potential firing.
And what you need in order to get
action potentials fired in postsynaptic cells
are perhaps inputs from many cells around it, that all provide input,
perhaps at the same time, onto the postsynaptic cell,
so that you can get a larger degree of summation,
and then you can cross action potential thresholds, and fire.
Although this is the most common situation,
where we have relatively small unitary EPSPs,
on the range of a few millivolts, lasting a few tens of milliseconds,
there are other examples in the brain,
and one important one that's studied,
in much detail,
is the so-called <i>Calyx of Held</i> synapse,
where an individual axon comes and envelops the postsynaptic cell,
and has really a large number of release sites present in it,
maybe some 200 or more synapses,
and because the action potential comes in,
and it then releases glutamate onto the postsynaptic cell,
which has many receptors, this then causes automatically
an action potential to fire in the postsynaptic neuron,
and so there are cases where an action potential
is converted into an action potential in the postsynaptic cell
through a large number of synaptic connections,
like, for example, here in the Calyx of Held,
an example from the auditory brain stem.
But for most synapses in the brain
we need to think about small depolarizations,
lasting relatively short amounts of time,
and that these potentials need to be integrated, postsynapticallly,
in order to have further processing of information in the brain.
We can begin to explore this process of postsynaptic integration
artificially, at first, by thinking about what happens
as we simply depolarize the postsynaptic cell.
And so we can imagine having a recording electrode
sitting on a cell, that's receiving a synaptic input,
glutamate is being released, opening AMPA and NMJ receptors,
and we're measuring the membrane potential of the cell.
And in addition to measuring the membrane potential of the cell,
we can also use this electrode to inject current.
And so we can inject 0 pA of current,
in which case we measure the membrane potential at rest,
so this is the current being injected.
We can inject some current, and that will then cause
a filling of the membrane capacitance, a depolarization.
We can depolarize, even inject more current,
and cause further depolarization, and then we can, of course,
stimulate the presynaptic cell at different time points,
and see what happens to the PSP at different membrane potentials.
And here are examples, schematic drawings of what might happen.
At -65 mV the membrane potential is relatively hyperpolarized.
The NMDA receptors don't get activated,
and so we have an AMPA-driven EPSP,
the rapid 2 ms conductance of the AMPA receptor,
and then the discharge of the membrane,
through the membrane time course,
discharging through the leak conductances.
If we depolarize the postsynaptic cell, here by 10 mV,
and it gives the same presynaptic action potential,
same amount of glutamate is being released,
same activation of AMPA receptors,
but now, in addition, also perhaps some contribution of NMJ receptors,
because we're depolarized a bit.
And so there are now two different phenomena that are going on.
If we think about our current voltage relationships,
we've gone from having large currents,
to having somewhat smaller currents,
and so there might be a decrease in the amplitude of the EPSP.
if we're just thinking about the AMPA component,
but on the other side we have the NMJ component
that might now be making a larger contribution.
And so we can imagine that perhaps at slightly later times during the PSP
the AMPA component is supplemented by an NMDA component,
and the EPSP might have a slightly longer time course,
and a smaller overall amplitude,
because of the change in the driving force,
that is also, of course, involved in helping
the NMDA receptor get recruited.
If we depolarize even further,
there'll be a further drop in the amplitude of the response,
because we're moving towards smaller and smaller currents,
the driving force for the AMPA receptor
has decreased, but we're, of course, much closer to AP threshold,
and if we hit AP threshold, an action potential wlll be fired
in the postsynaptic neuron, and the membrane potential
will zoom up to something like +14 mV.
And so you can see here that there are some relatively complex
membrane potential dynamics that we need to think about.
As we depolarize, the driving force gets smaller,
but at the same time we're also recruiting more NMDA receptors,
and so, again, depending upon
how much NMDA receptors are present at the synapse,
you can get quite different effects of depolarization at different synapses.
The situation is, in fact, much more complicated
because these synapses that we've been talking about
are typically distributed over the dendrites of a neuron,
and so there will be glutamatergic inputs
that are distributed all over the dendrites,
and, in fact, relatively few synapses are present
on the actual soma of a neuron,
and so, typically, the input that arrives on a cell
occurs on dendrites
at electrotonically remote locations to the soma.
And so if glutamate is coming in
at synapses close to our recording site here,
because there's relatively little membrane here, so these are small,
the dendrites are something like 1 micrometer or so in diameter.
And the opening of the AMPA and NMDA conductances here,
then the charge that comes in here,
will then cause a considerable change in the voltage,
because there's relatively little membrane here that needs to get charged.
And so you get large membrane potential
changes here in dendritic areas,
because of the small surface area,
and these are also very fast in their time course,
because the amount of membrane, again, here, is relatively low,
and there's a rapid flow of current away from this,
down the axial direction of these cables
that are, ultimately, the dendrites that we're thinking about.
And so the membrane potential has these rapid,
and large amplitude excursions,
when synaptic conductances are opened,
and those potentials, of course, spread passively down the dendrites,
and are filtered, in time, because of the cable properties of the dendrites,
and so what's ultimately seen at the soma
is a much slower, and smaller amplitude version
of this synaptic potential that's been recorded here in the dendrites.
So this gives rise to even further complexity
as to how synaptic integration occurs in real neurons.
So if we have our cell body here, and there are synapses
that are distributed across the dendritic arborization of the cell
there will be a barrage of membrane potential fluctuations that occur,
some of them at the same time, and others at different times,
across the dendritic arborization.
And if there are conductances that are synaptic inputs
that come at slightly different times
then they may not give rise to summation,
but if, on the other hand, these synapses occur at the same time
then of course you can get much larger depolarizations here,
and so there's a summation process,
and that summation process can take place also locally on dendrites,
and that can even be very helpful, in terms of activating NMDA receptors.
You'll remember that the NMDA receptor is blocked at hyperpolarized potentials,
but, especially in these dendrites,
where the individual synaptic events have large amplitudes,
if you just have a few of them occurring more simultaneously
we can get very large depolarizations in the dendrites
that are likely to activate NMDA receptors extensively.
And so these distal areas here
might be involved in communicating with the soma
largely through activation of the voltage-gated NMDA receptors
that might further drive voltage-gated sodium conductances.
And so the process of dendritic integration,
in both terms of time and space,
is very complex, and there can be highly nonlinear phenomena
going on in the dendrites, as well as, of course,
there can be proximal inputs close to the soma
that in themselves give rise
to large potentials in the soma.
And so what we typically see at the membrane potential level, at the soma
is then the summated input of the synapses across the dendrites,
and it's when the membrane potential at the soma,
or more precisely at the axon initial segment,
crosses the threshold, that an action potential is fired,
giving rise to the output of that cell.
So that's, in a sense, the essence of the neuronal computation,
how you take these postsynaptic potentials,
integrate them at the axon initial segment,
and decide whether to fire an action potential, or not.
And so we've now begun to consider
what happens at glutamatergic synapses in the brain.
These glutamatergic synapses are very important to think about
because they make up about 80%
of all the synapses that are present in the brain,
and so we absolutely need to understand how these work
if we want to understand how the brain functions.
The postsynaptic responses are composed of rapid AMPA conductances,
and much slower, longer-lasting NMDA conductances
that in addition have a voltage dependence to them.
And it's that voltage dependence of the NMDA receptor
that gives rise to complicated membrane potential dynamics.
Individual, unitary PSPs that come into the soma
are typically small in amplitude,
they may be one or two millivolts in amplitude if you're lucky,
and they last some tens of milliseconds.
But typically they're not sufficient
to drive postsynaptic action potential firing.
For that we probably need the convergence
of many tens or hundreds of presynaptic neurons,
all to find action potentials with some degree of synchrony
so that we can get this postsynaptic temporal integration
on the time scale of the PSP,
which is somewhere around 10 to 20 to 50 ms,
depending again upon how much depolarization there is postsynaptically,
and to what extent the slower NMDA receptors are being activated.
And so the total conversion of activation of postsynaptic conductances,
downstream of an action potential in the presynaptic terminal,
is a complicated phenomenon,
and it's still something that's actively being investigated in research,
in terms of how these dendritic inputs are integrated in real neurons
during real behavior in the mammalian brain.