Tip:
Highlight text to annotate it
X
My name is Kai Simons. I come from the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden,
and now I am going to give my second iBioSeminar, and this is about lipid rafts as a membrane organizing principle.
Now this work started with studies on epithelia, on epithelial cells.
Now if you look at simple columnar epithelia, they form a barrier between the outside world and the inside.
Take the digestive tract. There is an epithelial cell layer that is now protecting the body from everything
that is bad outside there in our digestive tract. And that means that each cell has to be constructed
in such a way that they can take care of this task.
For instance, the plasma membrane is divided into two domains,
the apical membrane and the basolateral membrane that you see here of the plasma membrane.
So this polarized plasma membrane serves a function such that the apical membrane is robust.
It has to be able to even withstand detergents, bile salts, and enzymes digesting our food.
And that means that the lipid composition and the protein composition of these two membranes is different.
So in the apical membrane we have different proteins, but we also have different lipids.
We have sphingolipids enriched and the glycolipids enriched depending on the tissue in the apical membrane.
And instead there is more phosphatidylcholine in the outer leaflet of the basolateral plasma membrane domain.
So now comes the question how does the cell generate these two plasma membrane domains?
How do they get the proteins to the apical side and the proteins to the basolateral side?
And this happens in the trans-Golgi network
of the Golgi complex on the trans side where they are sorted into different pathways.
But also the lipids are sorted there in the trans-Golgi network.
And Gerrit van Meer years ago showed that the simple red lipid, the glycolipid, was preferentially sorted to the apical side in the Golgi.
And therefore sorting away, segregating away, from these blue lipids, the phosphatidylcholines, the glycerol lipids.
And on the basis of this we postulated that protein and lipid sorting are connected.
That you form a platform of these lipids and those proteins that should go apically, segregating away from those that should go basolaterally.
And then you can generate these carrier vesicles and the idea was then that these are apical carriers
or lipid rafts containing both sphingolipids, cholesterol, and proteins.
And this gave rise then to the raft concept,
which was generalized then to not only work in epithelial cells, but in all cells in our body.
And the idea here is that you have these red lipids, the sphingolipids, ceramide based,
which associate with each other through hydrogen bonding, intermolecular hydrogen bonding, but also with cholesterol.
Cholesterol likes to be underneath the umbrella of these head groups:
the phosphocholine of sphingomyelin and the carbohydrate head groups in the glycosylated lipids.
And then you see that this fatty acid is longer than the sphingosine, and probably penetrates into the inner leaflet,
connecting the two leaflets in the bilayer.
And then you get this liquid ordered, more tightly packed assembly, which is the raft.
And this assembly is also somewhat thicker because the unsaturated fatty acids in the glycerol lipids like phosphatidylcholine is kinked
and therefore not as long, even if the length is the same.
So rafts are therefore sub-compartmentalizing membranes through their ability also to include specific sets of proteins,
GPI-anchored proteins, glycophosphatidylinositol anchored proteins with two saturated fatty acids
in the outer leaflet, and then you also have in the inner leaflet you have two saturated fatty acids on many tyrosine kinases like this yes kinase.
Or you have transmembrane proteins which like to be there and then there are the proteins that are excluded.
So this hypothesis went through hard times.
And there was lots of criticism of the lipid raft hypothesis. So here you see one of them: "Lipid rafts: Elusive or Illusive?"
So what was the idea? Well, are there now, do these rafts exist which can segregate and co-cluster proteins and lipids
that can then function in signal transduction and be segregated from the others?
Or, is the plasma membrane a continuous bilayer, and you don't have any lipid-protein,
lipid-lipid interactions but you only have protein-protein interactions?
What was the problem? The problem was that there are difficult methods and there are easy methods.
And the easy methods were simply take detergents, Triton X-100 at four degrees,
and you solubilize the membrane and what is left is a raft.
That is of course not very sophisticated.
So this could lead to artifacts. And then there was a general problem. They are so difficult to visualize. You cannot see them.
Why? Seeing is believing.
"Quarks. Neutrinos. Mesons. All those damn particles you can't see." Look at this physicist sitting there.
And then, "that's what drove me to drink. And now I can see them."
Well, and don't forget those pesky lipid rafts.
Well, things have improved. Also in this area, and today we have a three state model of the raft concept.
The resting state, cells like fibroblasts and the plasma membranes of immune cells, there you have small nano-scale assemblies.
This is one state. Another state is then the stabilized raft. That these nano-scale assemblies come together and form a platform.
And then we have a third state. We can form micrometer raft phases, which can be seen in the microscope.
So let's start with the nano-assemblies.
How small are they? How small? So work 10 years ago at the EMBL, we were working with Heinrich Horber and Arnd Pralle.
And we were using a new microscope-an atomic force microscope-
and the idea was here was to look at... we had a trap and in the trap there was a bead.
And this bead contained antibodies against proteins on the cell surface of the cells we were studying.
This trap was then lowered so that the bead touched the membrane,
and then after some while, the antibody could bind to an antigen on the surface.
And the microscope could measure the thermal fluctuations in x, y and z to get a measure for the viscous drag on the bead.
Now for that we used then proteins which were considered to be raft proteins, like influenza virus hemagglutinin,
or this GPI-anchored proteins, two types, and we compared with these green proteins,
the transferrin receptor or the low density lipoprotein receptor which were non raft proteins.
And now we want to see how they compare.
And we want to see also how these measurements will be influenced by taking away the cholesterol from the cell.
So the bead was very large. It's 100 or 50 nanometers in size,
so you can see here now this is an antibody binding to an antigen, and we even used the soluble part of this protein.
We took away the ectodomain and added it so that we would get just binding of one antibody to one antigen.
And here are the results when you now look at the drag on the bead.
Here is now the free bead, somewhat higher above the plasma membrane on the living cell.
Now we lower it, and then we see how it binds, how the drag on the bead increases but no binding yet.
What you see then is that it takes about 30 seconds, and now you see a change.
And this is when the antibody binds to the antigen.
Now this can be translated into local viscous drag measurements, and when we look at the raft proteins,
we see all of them are gray bars. They are... the drag on these is much higher than the gray bars for the non-raft proteins.
But when you take away the cholesterol, look.
The GPI anchored proteins then have much less drag, HA looks the same now as the non-raft proteins.
So this means that we can now measure, get an estimate, of the radius of this assembly in the membrane through these measurements.
And these measurements give an estimate that the raft proteins,
GPI anchored proteins, and the transmembrane HA protein, diffuse as an assembly of 50 nanometer diameter assemblies.
So it means that they have proteins in them and they have lipids in them.
And these lipids disappeared when the cholesterol was taken away.
So it gave us a clear idea that raft proteins indeed have a small cloud, a nano-scale cloud, of lipids around them.
So now let's move into the 21st century and microscopes have changed.
In cell microscopy in the '90s the resolution was not as high as today.
Today we can look at cells with super resolution.
And Stefan Hell has developed a new microscope called the STED microscope-Stimulated Emission Depletion-
and with that you get much higher resolution than with confocal.
Confocal is about 250 nm, and with the STED you can see below 50 nm.
So now we can start to look at these assemblies with these microscopes.
Now what he did, Christian Eggeling in his lab, took labeled sphingolipids
and labeled glycerol lipids and looked at how they moved in the living membrane.
So here is now the confocal microscope and this is done in the fluorescence scanning mode
so if a peak goes like this it means that the lipid is moving through this confocal 250 nm.
It peaks and moves out.
So the glycerolipid, phosphoethanolamine, moves, you see here. And if you compare it with the sphingomyelin, there is no difference.
You can't see any differences here.
But now, look at STED. Then you see that the glycerolipids...it is a much smaller area that you are looking at.
So if we just see a peak like this-look at these graphs-but now the sphingolipids,
some of the sphingolipids moving through this 50 nm area is staying longer. 10-30 milliseconds.
Meaning that they are trapped there. And this way you see a different behavior
between raft lipids and non-raft lipids, giving support for this raft concept.
And they can estimate that these nano-scale assemblies are smaller than 20 nm.
So smaller than we measure with our system, but that may be due to these big beads that somehow also influence the measurements.
So the take away, also with other methods, is now that in this resting state in the fibroblast plasma membrane
you see nano-scale dynamic assemblies which associate and dissociate on a very rapid, sub-second timescale.
So now what is the physical basis of these nano-scale assemblies?
How do they come together? So there is a whole area out there of fluid physics which can be applied now to phenomena in complex cells.
And there have been three ideas: critical fluctuations, micro-emulsions, cytoskeletal esters.
So let's start with critical fluctuations. Sarah Veetch, she showed that if you take plasma membrane blebs that have been released from the cell,
and you look at them then when you decrease the temperature, you get the phase transition as we will discuss later.
And just before you get to this phase transition you see fluctuations.
And in these model systems there is a phenomenon called critical fluctuations
meaning that when you go from a single phase here and you now go to the phase boundary,
and if you are close to the critical point then as you go towards the critical point
then you see fluctuations because the energy between the two phases is so small
before they phase separate and get two phases which are microscopically separated.
So these are critical fluctuations.
Are plasma membranes tuned to a critical point?
An issue. Another group looking at this from Brewster and Safran, they look at the membrane as a micro-emulsion.
A micro-emulsion usually means three dimensions, but this is a micro-emulsion in two dimensions.
And in micro-emulsions when you have this phase separating, then surfactants can work at the phase boundary,
decreasing the surface tension such that they become metastable or stable.
And in the membrane they would then be linactants, hybrid lipids that could sort of cover this perimeter.
So here is now a raft let's say, and here is now at the perimeter
a hybrid lipid which likes to be both in the non-raft phase and in the raft phase.
What does that mean? Well, look here. Here is a glycerophospholipid.
In the Sn1 position there is a saturated fatty acid that would like to be in the raft membrane,
and then in the 2 position there is a kinked one, which likes to be in the non-raft membrane.
So that is the idea, that such molecules would line the phase boundaries and in this way stabilize them in a metastable way.
Or you could also have proteins that could sit there. So both explanations demand an active fine-tuning
of the membrane composition to achieve the desired dynamically associating lipid and protein mixture in the bilayer.
Now there is also another way to look at it. This comes from Satyajit Mayor and Madan Rao in Bangalore, and they look at, they also want to include actin.
Actin sitting underneath the plasma membrane. So the idea would be that these nano-scale clusters of GPI anchored proteins
which you see here there are two, or three, or four of them would be together,
but that they would interact on the cytosolic side with dynamic actin
bound to the membrane and these form these asters which then are organized in these nano-scale clusters.
So presently we don't know.
It's not clear. This is an area where we have to work on, but all agree that there is a small nano-scale, dynamic assembly.
So now let's go to the next phase. These nano-scale assemblies can easily be induced to coalesce to larger and more stable platforms.
The stabilized raft. And let's look here at an early experiment.
So you take now a living cell, and there are raft proteins and non-raft proteins.
And now you add antibodies which bind to these, and you add one antibody against one raft protein, one antibody against a non-raft protein
or two antibodies recognizing raft proteins, and you see what happens.
They cluster these now in the plane of the membrane.
And what you see, this was a surprise then. So here is now transferrin receptor, as you see a non-raft protein,
and PLAP, a GPI-anchored protein, a raft protein,
and when you put antibodies to the cells that see these two proteins, they segregate into red and into green.
So they don't like each other. They go away. But now you take a raft protein, PLAP, the GPI-anchored protein, and the transmembrane HA.
And now you add antibodies against both simultaneously,
and you see strong co-clustering. They come together. You force them together by the antibodies.
So clustering of raft proteins or lipids to form raft platforms.
These are the cellularized form. And this means now that when you look at the plasma membrane,
and you look for co-localization without such tricks as antibodies co-clustering,
we would not see a specific GPI-anchored protein and its transmembrane protein,
or even a raft lipid like the ganglioside GM1 together because they are all the time associating and dissociating on a nano-scale scale.
It is only when you put them together when you see them coming together.
So this raft coalescence drives various physiological processes.
For instance in membrane trafficking from the Golgi, as we were discussing in the first lecture.
Or in signaling, all immune signaling, T cell signaling, B cell signaling, and also IgE signaling and allergens
as you will see has to do with these raft platforms coming together and creating a specific subcompartment which can then signal to the cell.
And that means that you can have parallel processing. You can have a number of these forming in parallel fashion in the membrane.
Even viruses when they bud out, ***, AIDS virus, buds out by forming a raft membrane around it.
And also influenza does the same. Otherwise they just take non-raft lipids.
So here you see now the allergens that you see here.
You have an allergen, and this allergen will bind to IgE antibodies.
And this IgE allergen will then bind to mast cells or basothelial leukocytes
through the Fc receptor which recognizes IgE. And this process will lead to clustering, a raft clustering, and this will then signal to the cell.
And the cell then releases histamine and you get all the symptoms of asthma or allergies.
Let's have a look at this process.
So here are the allergens, you see these big green balls.
And then there are these small balls. These are the IgE antibodies.
And now they bind to FcE, IgE receptors on the cell surface, and you can see now a clustering process taking place here.
And this is now a stabilized raft and then comes a tyrosine kinase in there,
and then many other things happen which are simplified here and a signal goes into the cell which will lead to histamine release.
This is now the coalescence that I have been talking about.
Okay, so let's look at other stabilized rafts.
The most famous stabilized rafts are the caveolae.
Here you see pictures from the inside of the plasma membrane by John Heuser,
and these small 60 nanometer balls here are the caveolae.
You can also see the clathrin coated pits here. This cell has a lot of them.
And these are a subset of stabilized rafts where caveolin molecules together with cavin bend the membrane and form these stabilized clusters.
So we therefore have these now at resting state. We can activate them to activate the cluster rafts
and we can even make these big large raft clusters with antibodies, as you saw.
But you can also when you use detergents, Triton X-100 at 4 degrees, what happens is that you drive raft coalescence,
and then you solubilize the rest of the membrane and the coalesced rafts are left there.
So now let's go to the third state. The two states, the nano-scale assemblies and the stabilized raft,
they all are there, seeable, observable in living cells.
This coalesced raft state you can only see when the cell is dead.
So if you look at these three lipid classes, sphingomyelins, cholesterol, and unsaturated glycerophospholipids,
dioleyl PC that you see for instance here. And you mix them in simple systems you can generate two liquid phases.
So the sphingolipids and cholesterol, they form liquid ordered, so called LO phases, so they are raft like.
Whereas these glycerophospholipids, they form liquid disordered phases which are more disordered due to these kinked fatty acids.
And these can exist together, in the same membrane.
And you can see here these giant unilamellar vesicles, where you have only dioleyl PC here,
and then so you look here, now you mix in the sphingomyelin, and you add cholesterol.
Then if you have a dye, a red dye that only sees the liquid disordered phase,
you see black holes here. These are the liquid ordered phases.
And if you add the ganglioside GM1 and cholera toxin, you can see the holes are filled with the glycolipids.
And depending on the mixtures you see a different size of phase separation.
Now what about that? These were three component systems. Three lipids only.
Now biological membranes are complex. They contain up to thousands of lipid species.
And they also contain proteins, because in the membrane 20% of the surface area in the middle of the bilayer is covered by proteins.
So it is a very tightly packed protein lipid mixture.
Can they phase separate? Nobody thought that. This was a big surprise.
They can phase separate, in the same way as simple model membranes into micrometer domains.
So here is work by Daniel Lingwood. He put the cells which have been labeled by a dye, a lipid dye. And then
three hours later he looked at cells and they had blown up into a balloon.
We don't exactly know how this works but they gained our attention.
This cell line is full of the raft ganglioside GM1, and when he then
took these balloons and he expressed the raft protein VIP17, which is a protein involved in apical traffic,
and he added the cholera toxin, he saw that all of the VIP17 went into one domain,
and when he looked at a number of other proteins, LAT transmembrane protein,
the GPI anchored proteins, or the cytosolic proteins that are palmitoylated,
then you see if you look here is before you only see the protein,
and then down here you see the protein, and now it is separated into domains which will co-localize with the GM1-cholera toxin label.
So you see phase separation where raft components are co-localizing in these big micrometer domains.
And this is selective because if you take non-raft proteins, the transferrin receptor,
then you see here is the transferrin receptor in red and green is the cholera toxin, the raft phase.
They are separated. So it is selective.
And it is also dependent on cholesterol, because if you take away cholesterol, from these blebs,
from these balloons, then you don't see this coalescence. You see small dots only.
And diffusion measurements measuring diffusion here and in the non-raft phase show that in the dark phase the diffusion is ten times faster.
So in this raft phase, it is ten times slower. So think about the disco.
Now it is early evening, you can just walk through the disco room without any problems.
It goes fast. But now it is two o'clock in the morning and then it is full of people, and now you try to get through the disco room.
And you see that it takes much longer.
That is the same thing here. The raft is more packed, therefore the diffusion is slower.
So the take home message here is that the plasma membranes freed from membrane trafficking now-
This is a systematic equilibrium now. No endocytosis. No exocytosis.
No cytoskeleton. Actin has fallen off.-
forms selective sterol dependent microscopic domains at 37 degrees.
And this is induced by oligomeric clustering, pentavalent cholera toxin, GM1, plus as you see, this phase separation being induced.
The implication: the plasma membrane bilayer is poised for phase separation.
And this capability reveals an underlying connectivity. So only by merger of small nano-scale rafts that you cannot see,
do you see now this micrometric phase separation.
So let's look at another way of making GUVs: Blebs from the plasma membrane.
So this is the work of Tobias Baumgart. He calls them giant plasma membrane vesicles.
These are induced by the calcium containing buffers and you also have to include sulphydryl blocking agents.
And then you get this phase separation in round, diffusive domains.
And this is induced by temperature. Usually the temperature has to be reduced
to about room temperature and below, and then you see one phase going into two phases.
Now if you look at proteins partitioning in these then like in the plasma membrane balloons,
the GPI anchored proteins will segregate away from the red lipid dye, which looks at the non-raft domains.
The transferrin receptor and the low density lipoprotein receptor, they will also be together with the dye in the non-raft domains.
And if you take two now transmembrane raft proteins, they will go segregate away from the non-raft dye.
So this is done now with non-cross-linking, non-reducing agents and they maintain the proper transmembrane protein partitioning.
And this you can quantitate just by imaging. You can see how much is in the non-raft domain, how much is in the raft domain.
And take now one of these proteins, the LAT protein involved in T cell and B cell signaling, and a raft protein. It has cysteines.
So here you see the transmembrane domain,
and there are two cysteines-this is C26-in the transmembrane hydrophobic transmembrane domain and one, C29, a little outside.
And now these have been deleted and taken out and then you see how this effects the partitioning.
Well in the wildtype, most goes into the raft domain. If you take away the cysteine at 29, then you see a little less,
and if you take away the intramembrane, in the transmembrane domain,
then you see almost no... its exclusion from the non-raft domain, from the raft domain, sorry.
It becomes non-raft. And these cysteines are palmitoylated.
They have a fatty acid added to them. So it means that cysteine palmitoylation is required for this partitioning to occur.
So you have a regulatable way of seeing how a protein can move in and out of a raft.
Now Ilya Levental did this work. He also looked at the whole slew of surface proteins in the following way.
So he biotinylated all of the cell surface proteins with the membrane impermeable Sulfo-NHS, and then he produced these blebs.
And then he removed the palmitate, and that you can do by just adding dithiothreitol.
Or he removed the GPI anchors with an enzyme, phospholipase C, and then after that,
he phase separated and labeled the proteins with fluorescent anti-biotin and quantified.
And what did he see? Well, he saw that in the control you have about 35 proteins in the raft phase,
Take away the anchors, you have less.
And take away palmitoylation, and you have even less.
So this is all shown in this pie diagram. So look at it.
Non-raft proteins going through this palmitoylation procedure, 65% outside rafts, outside this raft domain segregated in the blebs.
And then you have palmitoyl anchored proteins, 12%.
And then you have GPI-anchored, about 11%.
And then you have a third group, about 11% which are proteins like VIP17, which are not palmitoylated, but have raft-philic properties.
So 35% of the proteins are in the raft domain.
Now I should point out that the palmitoylation is important, but it is not sufficient.
Because the transferrin receptor, which we use as a non-raft protein, is also palmitoylated.
So it is not enough. There is something else, and something else means transmembrane domain length, longer in rafts.
Hydrophobicity and amino acid sequence are likely to be important factors
so it means that the transmembrane domain is the hard wired code. The palmitoylation is the regulatable event.
But what is important now is that we have a tool that we can study quantitatively how different factors influence raft partitioning.
This was not available before.
So now we know that we have a number of things that we can study,
and in this way we can get a handle on the question what makes proteins raft-philic.
How do they become associated?
What do they need to become associated with rafts?
Now one issue that is important is the protein/lipid interaction.
It is known from work by Sen-itiroh Hakomori and others, that gangliosides somehow affect, interact with protein receptors.
And we have a hypothesis that transmembrane raft proteins can interact with raft lipids for instance with gangliosides.
So there would be two conformations, one outside rafts for instance, and one in the raft, which is
a different conformation which is now binding the head group of a sphingolipid and a carbohydrate, and in this way making it raft-philic.
And we call this phenomenon lubrication, so we have to lubricate the raft proteins to get them into the raft.
And this may be one reason why we have so many head groups of glycolipids because we don't know why.
If we go into the literature, we would not know why the cell goes to the trouble of making different types of head groups.
Maybe it is due to the fact that they have to interact with the proteins.
So therefore, raft phase inclusion involves GPI anchoring, these GPI anchors,
or palmitoylation, also of proteins which are peripheral have two saturated fatty acid anchors.
And then this raft transmembrane protein/raft lipid interaction.
And remember then that there is also this palmitoylation of proteins
which have a raft transmembrane domain can then regulate the in and out of the raft.
So I think, coming to the end now, that lipid rafts have proven their worth as a membrane organizing principle.
capable of partitioning membranes at various lengths and time scales, I mean the size of rafts as you can see is not an important question.
We can go from nano-scale to micrometers depending on the way that we are studying them.
And this is based on the preferential association between sterols, sphingolipids, and raft proteins.
And I will stop here, and the next lecture will be on making these glycolipid rich apical membranes to serve as a protective barrier for epithelia.
And finally acknowledgements of people who are doing the work I was describing. Thank you.