Tip:
Highlight text to annotate it
X
Hello. My name is Tony Hyman.
I'm a director of Max Planck Institute in Dresden in Germany.
And I'd like to talk to you today about formation of a structure known as a P granule.
Before I start, on the right hand side,
is a picture of our institute in Germany that we built
when we went there about 10 years ago.
And, this was a new institute. It was put up in old East Germany
after the wall came down.
I was involved in designing this institute.
And, before we started building it, I climbed up next door in a house
and put a camera to time-lapse the building of this particular institute.
And that's shown in this movie. You can see that we can follow
the building of this Max Planck Institute through the different seasons.
It took about two years to build.
You can see that first, the wings of the building have gone up.
You can see the two separate wings of the building going up.
You can see them getting built from bottom to top.
And, you can see that sadly, the weather is not as good as it is here.
I'm in California. Sadly, there's a lot of snow and a lot of rain.
You can follow the changing conditions around a building.
You can see the intermediates in the process of building,
with these coverings over the structures once they're built.
And, we can then follow the process through to its completion.
And from that, we can get an understanding of how this building is put together.
I show that to remind you how important it is to get kinetic information
on how things are put together, and that's also true in any biological system.
If you simply look at the fixed, finished building,
it's very hard to work out how it's put together, and if you guessed,
you are likely to make mistakes.
And so, in biology, it's time lapse microscopy which has been absolutely key
for understanding how these different compartments are put together.
And getting kinetic information can really tell you a huge amount about
how the process is put together.
And I'd like to tell you today about how kinetic information,
and looking at the kinetics of a process,
told us so much about P granules that we couldn't get by looking at the static picture.
Coming back to our scale,
I've shown you some steps of organization of the cell
from individual tubulin molecules to microtubules to centrioles,
and there are many, many different protein complexes
that exist in this nanometer scale space.
Other protein complexes, such as ribosomes and nuclear pores, and
proteasomes exist around that scale.
But, there's a whole set of compartments
that exist at this scale, where actually we have much less idea how they're put together.
As I showed you from the first 3 segments of my talk,
we understand quite a lot about the rules by which we put together
even very complex things, such as a microtubule
or a centriole. We can understand the structure.
We can understand the mechanism by which the proteins are working.
But, at this level of scale, we understand much, much less.
And this is the problem I've brought up in my introductory segment,
which is, these are non-membranous bound compartments
which contain many different protein complexes, whose individual structures
we can understand -- the protein complexes.
But, we don't really understand the rules by which these protein complexes live together
to form these compartments.
Remember, I mentioned that they're extremely dynamic.
And so, that's essential for us in cell biology.
If we're going to understand how the cytoplasm is organized,
we need to understand how compartments at this scale
are put together.
So, a more general question that we can ask is
how do we structure large, non-membranous bound organelles?
If we think about a problem like that, one thing we can do
is we can ask what can we learn from non-biological systems?
Because, actually, non-biological systems take up very complex patterns
as well, such as water.
Very simple... water goes between many different... gas, ice crystals...
And so simple, non-biological systems actually take up
quite complex structures.
And so, we can ask the following question then,
which is, what do non-biological structures have to do with biological assembly?
Do they give us a clue for how these complex compartments
are organized, which so far have been very difficult to understand.
And the work I'd like to talk to you about today
was primarily the work of Cliff Brangwynne
in collaboration with Frank Julicher, who is a colleague of mine
in the Max Planck Institute for Physics of Complex Systems,
also in Dresden.
The Max Planck is a society which consists of many different institutes,
some in physics, some in chemistry
some in humanities, some in biology,
and so we can then often talk to our colleagues in these other Max Plancks
in other subjects, and Frank Julicher is a theoretical physicist.
Christian Eckmann, also in our institute, a group leader who works on
the formation of the germ-line, and Carsten and Agata,
were also involved in this particular project.
And the question is, how do P granules form?
So, P granules are cytoplasmic hubs of proteins and RNAs,
and they were discovered in C. elegans all the way back in 1983,
actually, just before I started my PhD.
They are essential for forming the germ-line, and they tend to be very big.
You can see, they're right in this mesoscale that I was mentioning
of organization of compartments at this scale.
P granules are also complex.
So, here, there are many, many different proteins
in a P-granule.
And, there are RNA and RNA-binding proteins in P-granules.
And they're believed to be important for the totipotent state of the germ line.
In other words, maintaining the ability of the germ line to carry on to the next generation.
If you think about it, it's really amazing that all over millions of years of evolution,
of any particular organism, each time, the germ-line has to be able to deliver
a perfect copy of the previous generation onto the next generation.
There will be some slow mutation, but basically, unless it can do that,
they're not going to be able to propagate.
So, your cells in your body slowly age, and they slowly die away,
but the germ line stays totipotent. It does not age,
and that's why we can propagate ourselves.
So, these germ-line granules have been known to be important
for maintaining the totipotency of the germ-line.
So, how do they actually form, and how do they function?
Let's look at a movie of P granules.
And this is a particularly beautiful movie, where we've labeled P granules
with a GFP marker. So, we GFP-tagged one of the P granule
components, and then we're looking at them moving in the embryo.
Now, watch what happens. You'll see them all sweep down the embryo
towards the end. So, they start here, and they sweep down towards the end,
and they end up in one end of the embryo,
and they're inherited only by one cell.
And that's what classically is known as an asymmetric cell division,
where one cell is different than the other
because it inherits different amounts or different types of cytoplasm.
So, let's look at this movie.
You see how the P granules are sweeping down towards the posterior.
And, let's show you that again.
The P granules sweep down towards the posterior of the embryo,
and you can see how they seem to be moving with a cytoplasmic flow.
So, that was fascinating. You see this large-range flow of cytoplasmic granules
and you also see the P granules moving with them.
So, that led to my hypothesis that perhaps P granules are segregated by cytoplasmic flow.
However, we did an experiment which suggested that couldn't be all that was going on.
We did that by 3D particle tracking.
We track individual P granules as they're moving.
And, what that showed is that there is no net flux of P granules to the posterior of the embryo.
And that's because you can measure the P granules going one way and the other way.
And, this particular bar chart over here shows the average number crossing in the middle,
going from the anterior to the posterior and the posterior to the anterior.
And you can see the same number go both ways.
So, there's no net flux to the posterior.
So, it can't be that the flows are simply moving P granules
from one end of the embryo to the other.
So, how could they be being segregated?
Well, you can look at the fate of individual P granules.
Let's do time-lapse microscopy.
As I brought up in the introduction to this talk, let's not look at a static picture of a P granule.
Let's ask what is the life history of a particular P granule.
Maybe that will tell us something.
And, indeed it did. Because, this graph here shows the lifetime
of many P granules over the division of the embryo.
And this actually looks a little bit complex, but just bear with me.
On this axis, we have the intensity of different P granules.
By intensity, I mean the amount of fluorescent molecules in any particular P granule.
On this axis, we have the time.
So, you can see from -14 to 0, which is an arbitrary time point we set during cell division.
Now, look what happens.
Right out of fertilization, there are P granules
on the anterior (that are in blue) and the posterior (that are in red).
Can you see how there are lots of them in both sides?
And that's what you saw in the movie, at the beginning they're in the whole embryo.
But look what happens first. They all depolymerize.
Or, they shrink, let's say, until there are very few P granules left in the cytoplasm.
And then, an interesting thing begins to happen.
They start to grow in the posterior, but they don't grow in the anterior.
So, P represents the posterior, A represents the anterior.
You can always think of an embryo having a polarity axis,
and anterior-posterior is the basic polarity axis of any embryo that first starts.
You remember, you end up with three axes.
You have anterior-posterior, left-right, and dorsal-ventral.
That's... when you have a human... anterior-posterior,
you have left-right, and you have dorsal-ventral.
So, we have an anterior-posterior axis here,
and P granules are forming only in the posterior.
So, we learned they're dynamic, and also we learnt that the dynamics are different
in different parts of the embryo.
They dissolve at the anterior, and then they dissolve at the posterior,
but then they form again later on in the cell cycle.
We can look at that in a bit more detail
by graphing the intensity of P granule assembly.
This is a graph here to show you there is a gradient of P granule assembly.
So, this is on the X-axis, we've got the position along the anterior-posterior.
Ok, the X is down here.
And on the other one, we've got the intensity.
And the dotted line is the line above which the net growth tends to be more than 0.
So, they don't ... they're tending to grow.
So, each point is an average of many different P granules
that either tend to grow or tend to shrink, but the net...
what we call the net assembly goes over this point, about 0.6 of the cell axis.
We can actually do a nice little dynamic graph,
where we took graphs during the different points of the cell cycle
and look at the growth, and you see that the growth and shrinkage changes
across the embryo, through time.
You see the way the graph... they're all shrinking at the beginning
And now, as you move through the cell cycle, some at the posterior are now tending to grow.
So, P granule assembly/disassembly changes with time.
It changes spatially, and that's why P granules form at the posterior.
So, we can then conclude from what we've done so far
that P granules separate by a gradient of assembly/disassembly,
where dissolution is favored at the anterior and formation is favored at the posterior.
So, then you can ask a question, right? As biologists, it's always improtant to think of a question.
And the question that came out of that study was
Why do P granules form at the posterior?
Because that's why and how they segregate.
If we understand that, we can understand how these granules
end up in the posterior of the embryo.
The breakthrough in this project came in Woods Hole.
So, the Woods Hole is a very old marine station
that every year runs summer courses, and I can highly recommend
these summer courses to anybody that's interested
in trying to understand problems in cell physiology,
and there are also other courses in embryology.
And every year, I taught that course for five years.
Every year, one would go for the summer for a couple of weeks.
You'd get 10 very motivated students who want to do something interesting,
and you have to think up different projects for them to have a go at.
And one year, Cliff, who had actually been a student in the course,
decided to come and teach the course and look at in more detail
the details of P granule assembly.
What we discovered is that P granules have characteristics of liquid-like drops.
Let me show you some movies which illustrate what I mean by that.
So, look at this little P granule here, and watch it approaching this big one
And watch it getting sucked up and fusing.
Like two drops coming together.
Here's another P granule. Look at the two of them coming together and fusing
like two drops.
See them just squeezing together and making one droplet, just like you had two droplets of liquid.
So, from that, we thought, yeah they seem to have liquid-like properties.
And so, we want to do other experiments, like dissect them out of their
endogenous context, put a sort of flow of buffer, and now what you can see,
they seem to be dripping, just like the honey dripping off your spoon in your morning breakfast.
So, look at them dripping. Just imagine that you had your honey on a spoon
dripping off. They're dripping off the sides of the nuclei, just like
a very viscous liquid.
And a number of a different things we did suggested that indeed
they behave like liquid drops. First of all, they're spherical.
It turns out, that it's quite hard to be a sphere unless you have something called surface tension,
which is a property of liquids.
You can look it up in Wikipedia and find out more about surface tension there,
but you can see it's a key property of liquids.
The other thing they do is they tend to wet when attached to the nucleus.
So, if you think about a drop, and you take a drop of any liquid,
like a viscous liquid, and you put it on a surface,
it will tend to wet onto the surface.
Unless, you take steps to stop it wetting, like putting some kind of a surface on it.
And the other thing, they can form and fuse into larger spheres.
And all of these were characteristics that, the state of P granules
that is in, is a liquid-like state.
Now, what does that mean to be a liquid?
It's very confusing to talk about that, first of all. And what does it mean to be liquid?
Well, what does it mean to be a solid? Let's first think about a solid.
If you're a solid, you can come back over and over again,
and the amino acids or the atoms will always be in the same place that they were before.
You can do a crystal, of course, but you can take a microtubule
or a centriole, and you do any kind of structural biology, the components
will always be in the same place time and time again.
A liquid, that's not true.
The internal contents of a liquid are moving around very quickly.
So, over the time scales we're interested in --
minutes -- the contents of these liquids continuously rearrange,
whereas the solids would not.
And, the best analogy that I know of to think about what the difference is
let's think about a school. When the kids are in the classroom, they're all
sitting at their chairs, and so you know, time and time again,
that Johnny will be in this chair and Jackie will be in this other chair.
So, that is something you can then predict where they're going to be.
The liquid-like state is more like recess,
where the kids are running around. You know they're in the playground somewhere,
but you can never predict exactly where they're going to be.
And that's what we call liquid-like state.
That, of course, also defines the issue that time scale is very important,
because, of course, if we looked at the playground over milliseconds,
the kids would also always be in the same place.
So, we're talking about the time scales in many, many minutes,
things are no longer in the same place.
So, that's a problem for us, of course,
as biologists, because all of the techniques that have been built up
over the last 50 years rely on molecules being in the same place over and over again.
They rely on series of interactions, and they rely on the ability to do averaging
to understand the molecular arrangements.
So, that's one issue is we can't use standard techniques to think about how liquids form.
Another interesting thing about liquids
is they undergo phase transitions.
Now, what is a phase transition?
Well, there's some very simple examples of phase transitions out in nature.
For instance, let's think about water vapor.
Let's say you did the following experiment,
and you can do this in your own home.
You take water vapor, and you have it in a hot chamber.
If you cool it down, the water vapor will condense into little droplets.
And that's because the saturation point of water changes at the temperature.
So, at the cold temperature, it becomes saturated,
and it condenses down into water droplets.
And that's exactly like you would see, for instance, on a cold day.
When you breathe out, your air... your warm air condenses onto the
cold air that's outside.
Phase transitions are interesting because they're a simple way of concentrating complex mixtures
of reactants. And, one way, for example of that, for instance,
is if you do an alcohol / chloroform separation.
The alcohol goes in the water layer because it has hydroxyl groups ... simple alcohols.
And so, the interesting thing is to find out what rules can be used in biological systems.
Because the simple rules can lead to large scale organizations.
A great way of concentrating, say, 50 different components
in a complex thing is the P granule.
And that will be an essential thing for the future will be to describe these rules at a molecular level.
We can ask then, why P granules actually form at the posterior.
If we think this is a phase transition, why is it they actually undergo this phase transition
at the posterior of the embryo?
And it turns out, there's a set of underlying biochemical asymmetries
which are really beautiful
and beautiful problems to study, also from a physics point of view.
You have the cortex set up into two domains.
Very interesting problem of establishing cortical asymmetry.
This cortical asymmetry then contributes to form
soluble downstream gradients of this protein called MEX-5.
And MEX-5 is thought to inhibit the formation of P granules at the anterior.
So, how does that work? Well, let's go back and look at our dynamic system, now,
and ask how MEX-5 inhibits the formation of P granules.
So here's a blow up of MEX-5. You can see a bigger picture, and you can see
the gradient of MEX-5. It's high in the anterior and low in the posterior.
And the P granules will form where MEX-5 is low.
Now, let's look at the intensity of fluorescence. And what you see from this graph
is very interesting, which is you can see in blue
the intensity of P granules -- the gradient that I showed you earlier in the talk.
But, look at MEX-5 -- it's opposite to the P granule gradient.
So, that's very interesting. Because they have opposite gradients, it might suggest
to us that indeed, P granules grow more slowly in the anterior
because the concentration of MEX-5 is high.
If we make a mutant experiment, where we RNAi MEX-5 or remove MEX-5 from the cell,
you can see something very interesting happen, which is
when you do an experiment where you remove the function of MEX-5
from the cell. You see the wild-type that I've showed you before,
where you have a gradient of assembly across the embryo.
Now, look at the red line, which is a MEX-5 mutant.
They're uniformly dynamic across the whole embryo
which is they're growing... the net tendency to grow is too slow for them to actually grow
anywhere in the embryo. They're growing a bit slower in the posterior than the wild-type,
so there's no net increase in amount of P granules, but look at the
anterior. Look at the anterior over here.
They're actually tending to grow more than the wild-type.
So, that suggests that somehow, MEX-5 suppresses growth of P granules here,
in wild-type. You take it away, and now it allows it to grow more fast.
But that somehow prevents the growth of P granules in the posterior around here.
How could that work?
We don't really understand that, but let's come back and give you some ideas
based on analogy of phase transitions in liquids.
Let's go back to our original experiment, and this is an experiment you could again do
in any classroom, which is to let segregate water
by phase transition.
So, take a water vapor and cool it down.
And, it condenses all over the slide. But, we can segregate water vapor
to one end of that chamber by putting a gradient of temperature.
So, instead of doing this experiment, we do a different experiment,
which is we make a chamber with hot at one end and cold at the other end.
And when we do that, what happens is that all the water vapor ends up at one end of the chamber
down here. And that's because water tends to condense on the cold end,
and then, you reduce the amount of soluble water vapor... there's water vapor
around the condensed out water.
Now, there's a process known as diffusive flux,
which is a basic property of diffusion, which always tends to equalize the concentration
of fast-diffusing components.
And so then, of course, the concentration of water vapor is equalized.
Anything on the cold side also goes more into water droplets,
there's diffusive flux, so slowly all of the water vapor disappears
from the chamber and ends up in the condensed water vapor.
So, there we are, we've used phase transitions and a gradient of temperature
to segregate a water vapor.
Now, let's come back and think about it for P granules.
What we think is that P granules condense at the posterior for similar reasons.
Not because there's a temperature gradient, of course,
because there is no temperature change -- they're growing at uniform temperature.
But, rather, because there's a saturation point that's established
by gradient polarity proteins, such as MEX-5.
So, you change the tendency of the components to saturate out at the posterior,
but they are still happy to stay as individual components at the anterior.
And we have evidence for that when we look at the embryo
because we actually look at the amount of soluble components.
On the left hand side, I've got green, which shows the concentration
of individual P granule components that are not in P granules.
And look at how, as the P granules form,
they're being depleted from the cytoplasm, so they all end up at the P granules.
And, presumably, as P granule components form at the posterior,
diffusive flux equalizes the concentration.
If you do photobleaching studies, these P granules turn over very, very fast.
They're diffusing fast enough, and then slowly they all end up in the posterior.
And so, if you come back and ask the final question,
Why do P granules form at the posterior?
What we would say from that is that P granules segregate
because the diffusion rate of the liquid phase
is slower than the diffusion rate of the dissolved phase.
The individual components, which are not in P granules diffuse very fast,
and they can diffuse over the embryo by diffusive flux.
The P granules diffuse very slowly,
so once they form, they tend to stay where they are
and therefore you end up with most of the P granules components in the P granules
at the posterior, where the embryo wants them.
And I bring this up to illustrate how we can use thinking about physical chemistry...
all that physical chemistry you've done at university
can be used to think about these complex problems of biological assembly.
The work I've discussed you today is just the beginning,
and we're obviously short on molecular mechanism,
but it shows how powerful these ideas are of physical chemistry
for organizing the cytoplasm of the cell
at this 100 nanometer, 1 micron to 2 micron scale,
which has been so difficult to understand.
And, I hope in the future, that these sorts of ways of thinking about it will be very useful
for thinking about other organization of other compartments in cells.
It also shows why it's worth concentrating on chemistry in university.
And, what I'd actually like to say as well
is that this idea of using physical chemistry to think about cells
is not new at all. If we go back to E. B. Wilson,
E. B. Wilson wrote a very famous book called The Cell, Development, and Heredity.
The final volume is in the 20s.
It summarized all the knowledge of the great cell biologists of the turn of the century.
This was the really hot field in those days.
And, he summarized, in this classic textbook, basically everything we knew,
which was the Bible for those of us who have started to work on these problems 80 years later.
Much of it, for instance, was in German, and inaccessible to many people.
But, of course, he was also an active scientist
and here's a paper that he published in Science in 1899,
entitled "Cytoplasm is a colloidal liquid emulsion."
So, these ideas of physical chemistry existed 100 years ago.
The ideas of colloids and physical chemistry were very topical then,
and people thought they might be useful for understanding how the cytoplasm works.
Now, the problem was, there were no molecules then, and so the field didn't really advance,
as often happens. Fields get to a certain stage and then they stop, and no advance can be made.
And if we look at a sort of short history of 20th century biology,
you can think about it, it's in the first half of the 20th century,
and the end of the 19th century, people thought about the physical chemistry of the cytoplasm,
with no knowledge of the molecules.
Then, there came the idea that we have to understand these problems at a molecular basis.
And this great challenge to catalog and understand the molecules of the cell started,
which had the DNA sequencing, as I mentioned in my introduction,
then RNA interference has been a way to really do high-throughput analysis
of molecular function in complex eukaryotes.
There are other high-throughput genetic techniques in more simple systems.
Proteomics defined the way that they're organized into complexes.
That triumph of humanity will surely be seen, looking back in a couple of years,
what a triumph that was in 30 or 40 years to really catalog the molecules in the cell.
Of course, it's not told us anything...
Well, it's told us something, but it hasn't told us a lot
about how the cell is actually organized.
And, I would contend that it makes sense now to go back with our knowledge of molecular biology
and the catalog and use the physical chemistry ideas
and join these two ideas together to try and understand how the cell is organized
in order to perform its myriad and complex functions.
Everything stems from chemistry. Originally, we were chemistry.
The initial formation of life came out, almost certainly, of
formation of complex chemical building blocks,
so it makes sense to use chemistry to think about how the cell is organized
because that would have been its original basis.
And I'll finish by summarizing the three topics I talked to you about today.
And just to show, as I mentioned at the beginning of the talk,
that you've got to use different ways to think about things
at different length scales and different organizations.
So, when we think about microtubules, we think about them as polymers.
They grow, and they shrink.
When we think about centrioles, we think about more this virus-like mechanism,
where you step in different stages of assembly.
Because... think about them as molecular complexes, but they have the intermediate states.
And then finally, this other level we talked about is liquid-like states,
where we have to think different kinds of theories.
You have to think about chemistry of liquids to understand how they work.
And, that illustrates, I think, what's so wonderful about being a biologist
in the modern era, that we can take all these different techniques,
and all these different theories.
We can bring physics, we can bring chemistry,
and try and use these to understand these myriad, complex problems
that we're confronted with every day we look at the cell.