Tip:
Highlight text to annotate it
X
Hello everyone. My name is Tim Mitchison.
I'm a scientist at Harvard Medical School,
although, at the moment -- this is summer 2011 --
I'm at the Marine Biological Laboratory in Woods Hole,
where I'm a Whitman Investigator.
Woods Hole is a beautiful marine lab by the sea
with a long history, and the topic I'm going to talk about today --
self-organization in biology, it's particularly appropriate
at Woods Hole, since a lot of classic experiments were done here by
Shinya Inoue. And I'll show you some movies that were taken here
a while ago, and then some recent ones my group has taken here,
working here as a Whitman Investigator.
And I'm going to talk today about self-organization in biology.
Let me go to... I'm going to start with a kind of introduction
about self-organization as a kind of organizing principle in biology.
As a way the genome gets turned into cells
And then, I'm going to talk in slightly more length about
a particular case study in self-organization, which is meiotic spindles.
And, in the first part, I'll be... already, this is a meiotic spindle here...
I'll be using this image quite a lot to illustrate my point.
I'm not going to introduce the biology of meiosis or mitosis in the first talk,
I just would like you to get a kind of feeling for what a self-organizing system is like.
And so, if I use language -- names of pieces of a spindle,
I will introduce those properly in the second talk,
where I talk more in detail about cell division.
So, here I want you to get a kind of feeling for the biology.
And what I'm hoping to transmit is a sense of
kind of wonder at living processes that I felt
as a child, and I continue to feel watching living cells under the microscope
go about the processes of life.
And studying how those processes work has been my life's work,
and that of many other people.
OK. Here's some outlines and key concepts.
This is more for people who may be taking this as part of a course,
but this is the things we're going through.
I want to -- let me just step back here for a second...
Here's a diagram that I hope is familiar to any student of biology.
Genes going to polypeptides, folding into proteins.
This is the central dogma. This is the way the genome is read out,
and I think nowadays, we would add another layer to this.
Proteins can bind to each other and to other macromolecules like nucleic acids
and assemble into very sophisticated protein machines.
And, at least to a first approximation, this genes into proteins into protein machines
is something I think we understand conceptually.
And, perhaps the most sophisticated example of this... I'll call cell self-assembly process,
where genes rather directly encode the structure and dynamic biological processes.
This is a lovely example of bacteriophage assembly.
So this is a bacterial virus down here. This is the genome of that virus. This is bacteriophage T4.
This is taken from a classic review by Bill Wood.
And you can see here in the genome where these different pieces of the bacteria are
encoded. You can see over here, they fold up into these polypeptides
and stick together like a kind of very sophisticated lego set essentially.
And you get this really remarkable machine here, a bacteriophage --
one of the more sophisticated biological machines you can find.
So this is a fantastic example of genes building a complex object.
But I think many people looking at this bacteriophage would not say that it was really alive.
It doesn't move, it doesn't change shape, it doesn't grow.
The genes specify a very precise structure that
does its job, but can't adapt to changing circumstances
and that kind of thing.
And if I go forward one, I want to contrast that to... I've put an amoeba here,
growing and moving, and I think even quite a young child
looking at one of these amoeba down a microscope...
which is something I did with my kids' classes when they were younger...
You can see an amoeba move, even with your eyes,
and everyone can tell it's kind of alive in some way.
And how these proteins and protein machines become living processes
is really what I want to talk about today,
mainly using the example of mitotic spindles,
but trying to generalize a little bit from that.
And, in thinking about this problem, I want to highlight or step back for a second here
a kind of scaling problem. How do proteins -- here --
proteins and protein machines, which have nanometer diameters
so they're quite small... How can they get together to organize cells,
which can be much larger? And the example I've chosen here is a particularly large cell,
a frog egg, which I'll talk more about frog eggs dividing
in my second talk, because it's an experimental system we use.
But, you can see here the task of building, here, a millimeter-size object --
although microns would be more typical for cells --
out of nanometer objects and organizing them in space and time.
It's kind of distinctly non-obvious how you would do that.
And the organizing principles of self-assembly that make protein machines
like bacteriophages are going to be of limited help, I think.
And sort of stepping back from this, I think one question I like to ask is
how, in general, could one organize a cell so different bits of the cell
knew what the other bits were doing and acted as a coordinated whole?
As in that frog egg dividing, over many, many microns, a scale
much larger than molecules.
And, I think there's in theory a variety of ways that this could occur.
One way that I'm going to really emphasize in my talk is making long protein polymers --
joining proteins together, end-to-end, to make long filaments that can
physically organize a cell.
Another way I'll come to in my next talk,
at least a little bit, is using chemical gradients.
So, a molecule that is created in one place and diffuses
across the cell is another way of doing it.
A third way is using mechanical forces.
If you were pulling more at C, for example, A might detect that pulling.
These are not mutually exclusive models. In fact, the spindle example
I'm going to talk to you about, in some sense, uses all three.
And also, you might want to think about,
are these the only ways you could physically organize the cells?
I think there are probably some other principles.
For example, how a neuron moves information, which I won't talk about.
And, I'm going to focus on this way of making large objects --
protein polymerization. Here's an example of tubulin,
which is the subunit of microtubules.
And, microtubules form by an end-to-end polymerization reaction.
They break down again by a depolymerization reaction.
This is a non-covalent reaction.
The tubulin molecules stick together by the weak bonds
that characterize all or most interaction between proteins,
and they can form and dissociate quite quickly.
So, this in some way is a simple way of getting a much larger molecule out of a smaller one.
And we can see, here's an example in a rather small cell,
a Drosophila tissue culture cell. You can see the radial array of microtubules,
kind of spanning the cytoplasm
and organizing it.
A more spectacular example, I think,
is in a much larger cell... not a frog egg, in this case, but a fish egg.
I'll show you a movie taken by a student in my lab, Martin Wuhr,
of microtubules in a fish egg dividing -- a zebrafish embryo
that's been transfected with a protein that binds to microtubules
and then followed with a laser confocal microscope.
And the field here is a whole 1/2 millimeter -- 500 microns -- in diameter.
So, I can step back from this so you can see it.
This is... mitosis here is just finishing, and two radial arrays of microtubules are growing out
and stretching through the cytoplasm.
And these two sets of microtubules will specify where the cleavage furrow comes in,
which is that dark line coming down, so the cleavage furrow is cutting down to the top,
and it will know where to go because these microtubules will tell it.
But you get the sense these microtubules grow through the cytoplasm and organize it,
and we're actually spending quite a lot of time trying to understand
these large asters in my lab at the moment.
So, I talked at the beginning in rather general terms about self-assembly.
What I'm going to argue now is that microtubule assembly, and particularly mitotic spindles,
provide a great example of this kind of organization of molecules
into living systems. And Shinya Inoue, who works still here at
the MBL in Woods Hole was one of the great students of self-organization
in the cytoplasm, and you can see here in this movie of jellyfish embryos dividing.
Mitotic spindles popping up in the cytoplasm and then going away again.
The spindles... there's an example here. This is polarization microscopy.
It's sensitive to aligned protein filaments.
When the movie rewinds, we'll look at it just for a second again,
but you'll see these mitotic spindles kind of magically appear,
do their job, and then disappear again.
And we'll talk more about their job segregating chromosomes in a while.
So, there's the first mitotic spindle appearing and disappearing.
And we'll see the second mitotic spindle (sort of magically, to me)
appearing right... where is it going to come? Just right here.
Those sort of black and white quadrants there, by polarization microscopy.
And then disappearing again.
And Inoue actually inferred from this kind of microscopy
that they're made of aligned protein filaments,
which are in fact, microtubules.
And so that kind of spontaneous appearance and disappearance is one of the things
we want to explain.
Now, the simplest way of building a microtubule would be to have
a tubulin subunit bind and dissociate
as a forward and reverse of the exact same reaction with no other chemical change.
This would be a very simple example of self-assembly.
In this case, creating an endless chain of molecules,
but it's just like the bacteriophage -- it's kind of a Lego block principle.
And this simple sticking without any other chemical change,
without any ATP hydrolysis is the way most protein machines are built.
And so this is sort of how you might imagine a microtubule being built.
However, there's a problem that would occur
if, for microtubules, if the proteins polymerizing and depolymerizing
were simply forward and reverse reactions of the exact same chemical reaction.
And that is once a microtubule assembled to equilibrium,
and you looked at its length over time, it would remain very constant.
There might be small fluctuations as subunits came and went,
but, in general, the polymer would stay at roughly the same length
while the amount of tubulin in the solution was constant.
And that would be a fine way to build a stable rod, for example
those bacteriophage tail fibers, but it wouldn't be suitable for building one of these mitotic spindles,
which, as I showed you in the Inoue movie, has to form in a minute or two
and then later dissolve again in a minute or two.
So, it couldn't work like this thermodynamically.
And, in fact, tubulin has evolved a way of getting around
the constraints of equilibrium thermodynamics
by coupling polymerization -- this is polymerization here --
to hydrolysis of a high-energy nucleotide, GTP.
So, tubulin polymerizes here in the GTP form.
This is a growing microtubule.
When the microtubule starts shrinking again, it's shrinking in the GDP form,
because you have a hydrolysis event here.
And that means, for every tubulin molecule that goes one full cycle here,
you've hydrolyzed a molecule of GTP.
That's 12 kcals per mole of chemical energy... something like that.
And, this kind of frees the microtubule from the constraints of thermodynamic equilibrium,
and allows it to do some more exotic things.
And, how does tubulin use that energy?
One of the main ways tubulin uses that energy is to create
a very large length fluctuations in the microtubules.
So, they grow and then shrink and then grow and then shrink.
And these are much larger and more rapid length fluctuations
than would be possible at equilibrium.
This process is called dynamic instability.
And it's easier to actually appreciate this if you see a movie.
Here's a centrosome in frog egg extract, which I'll talk about in the next talk.
You can see, I hope, individual lines growing out and then shrinking back.
And each microtubule in this movie lasts on the order of 30 seconds or so.
So, the microtubules are growing out, shrinking back, growing out, shrinking back.
And to do that, they need to burn GTP
to escape the constraints of equilibrium.
And although I won't talk about this, in fact, this evolutionary strategy
for a protein polymer to couple nucleotide hydrolysis
to interesting and complicated polymerization dynamics, to
allow self-organization and work is conserved among a number of cytoskeleton polymers
in the tubulin and also the actin gene family.
The tubulin family, which has both animal and plant eukaryotic homologs
and also bacterial cousins, essentially with the same fold.
They bind and hydrolyze GTP during polymerization.
And the actin filaments -- actin in eukaryotes and ParM and other molecules in bacteria --
they bind and hydrolyze ATP.
But in both cases, the point mechanistically and biologically
is that by coupling polymerization to nucleotide hydrolysis,
you can get more interesting dynamics... more lifelike dynamics
than would be possible at equilibrium.
And, you know, it'd be fun to talk about how some of these other polymers use nucleotide,
but I'm going to focus only on microtubules.
But I definitely encourage people to read about the other ones.
So, we have these large length fluctuations.
But how is that going to help us? How is it going to help the cell
rapidly assemble a mitotic spindle
out of soluble precursors -- this rather magical self-organization process?
And, I think it's fair to say we don't really know. That it's still an open research question.
I'll share with you one an idea. It's an old idea.
This is an idea Marc Kirschner and I developed when I was a PhD student in his lab
a number of years ago. And that's that these length fluctuations,
these large changes in length that are quite rapid
promote a kind of spatial exploration. They allow the microtubules to explore the space
of the cell much more rapidly than would be possible without GTP hydrolysis.
And I'll illustrate that with the example of hooking up a centrosome
to the kinetochores of the chromosome.
So, microtubules are going to grow from the centrosome.
They have to find these kinetochores. I'll explain why that is more in the next segment.
But, here, let's just say this is a challenging spatial problem -- how can the microtubules do this?
And, one idea -- the idea here -- is that microtubules are growing out,
growing out, shrinking back, growing out, shrinking back,
and exploring space randomly.
Some fraction of them will hit the target here and become captured.
And in that way, you can effect this connection without needing any other kind of information
to be exchanged between the target and the centrosome.
That's still, I would say... this is a long time later ... still an unproven idea.
But I think the general idea of spatial exploration is likely to be correct.
What I want to talk about now...
I'm not going to talk more about details of microtubules and how they work in this talk.
I'll go into that a bit more when we talk in the next segment,
where I talk more about mitotic spindles
What I more want to talk about now are what are the kind of general characteristics --
the life-like properties, I'm going to call it --
of a self-organized system.
I already talked about spontaneous assembly from soluble precursors --
this kind of magical... growing in the cytoplasm.
I want to talk about two other characteristics.
A characteristic, but adaptable size and shape.
I'll explain what I mean by adaptable here.
And then also rapid recovery from damage.
And I would argue, although I'll show only microtubule assemblies,
that these are rather general properties of self-organized systems in cells.
And, let me start with adaptable size and shape.
So, that bacteriophage I showed you, which we can trace
the assembly pathway directly from the genes
that has a very fixed size and shape that are directly encoded in the genes
and the self-assembly pathway.
For microtubule assemblies, although they have a certain shape,
the characteristic shape, let's say, of the mitotic spindle,
the size, in fact, can adapt to the size of the cell,
unlike the bacteriophage.
And I'll show you an example here with cleavage divisions in a zebrafish embryo.
This is actually a continuation of that movie I showed earlier
at the one-cell stage.
So I'll step back here so you can see it.
That was second mitotic spindle. Now, you'll see the third mitotic spindle.
And, as they go on.
And you can see here, as cleavage makes the cells smaller,
of course the mitotic spindle has to get smaller.
Otherwise, it would be too large for the cell.
But, I hope you appreciate the way here that the self-organizing structures --
the mitotic spindles -- are able to adapt naturally
and rapidly to a change in the physical circumstance --
in this case, size.
Let me show you another ... it's kind of related in my mind --
recovering from damage.
And this is going to be... this is a meiotic spindle here in a parchment worm egg
polarization microscopy. This is a lovely example from the work of Ted Salmon
when he was a PhD student with Shinya Inoue.
This movie, I'm pretty sure, would have been taken here at Woods Hole.
And, Salmon and Inoue are going to torture this egg,
if you will, by applying high pressure to it.
Ted Salmon developed this pressure cell
where he could put hundreds of atmospheres very quickly
and release it. And the pressure causes microtubules to depolymerize,
and then when you release it, they can polymerize again.
And so, if you watch this spindle here... It's low pressure.
The pressure is applied. You can see the spindle dissolve. It dissolves within a minute or so.
The pressure is released. It's comes back.
So, this ability to recover very quickly is very characteristic of a living system.
I think there's going to be another cycle of pressure in a second.
The pressure is going to go... Oh, well maybe the movie's looping.
Anyway, there it is dissolving
and coming back again. I just absolutely love watching.
This is very similar to the egg extract spindle that I'm going to talk about in the next talk,
just in shape and form.
It's a meiosis spindle in the egg.
Here's another example.
This is two meiotic spindles brought in contact with each other
in an egg extract. Again, the details don't matter here.
You can see the two spindles right at the beginning of the movie.
It's a little more difficult here. They're first separate and then they jackknife together
and fuse. And I think this fusion of two structures into one...
Again, a bacteriophage certainly couldn't do that.
It's very characteristic of one of these self-organized assemblies.
So, I'll talk in more detail about some of the molecules... not too many molecules
in these spindles.
Here, I want to ... I'll end the introduction by
talking in rather general terms about the sorts of properties that make these spindles so life-like.
That allow them to, for example, recover from damage or fuse together
in that last experiment.
One important aspect is spindles are continuously exchanging subunits.
Inoue first proved this and then I'll show in my next talk some more
recent experiments. But, the tubulin subunits are continuously exchanging
with the cytoplasmic pool, and one way of thinking about that
is the spindles are continuously rebuilding themselves.
Although they are at a steady state in size and shape
that can last for many hours in the case of a meiotic spindle in an unfertilized egg,
every subunit in that spindle, except actually for the DNA itself,
is cycling in and out rather rapidly.
So, I like to think of them as rebuilding themselves all the time.
So, it's not like the bacteriophage that builds itself and then its finished.
In the case of the spindle, it builds itself, and it keeps rebuilding and rebuilding.
And, I hope you can understand that makes it easier in some way for it
to change shape or recover from damage.
There are also a lot of mechanical forces in the spindle
generated by motor proteins.
And they're pushing and pulling on different parts of the spindle,
and those forces have to be balanced at steady state so that things don't move.
And, again, the way I like to think about it,
it isn't these forces act and then they're done.
They're continuously acting. They're continuously rebalancing themselves
to get everything right
on a timescale of seconds.
And again, that helps it be life-like.
So, continuous turnover of subunits and rebalancing of forces
generate a steady state -- that means constant over time --
but still dynamic in both mass and shape.
These process require substantial
energy dissipation. I told you about how microtubule polymerization uses GTP,
but the motors in the spindle I'll talk about later burn a lot of ATP as well.
So, it costs biological energy to do this continuous rebuilding and rebalancing.
But the net effect is an assembly that continuously renews itself
and can thus adapt... adapt, for example, to a changing shape of the cell,
repair damage, and other life-like properties.
And let me end the introduction here.
So here's this bacteriophage.... a really exquisite biological assembly.
So complicated, so amazing that it does its job amazingly well
through a self-assembly pathway leading to an invariant size and shape.
Here in the spindle, we're going to use this word self-organization
to indicate assembly, but now to a steady state
that continuously is dissipating biological energy
to give a much more kind of life-like assembly
that can change and adapt and repair damage.
Both beautiful biology.
I'd say the spindle uses all the self-assembly principles of the bacteriophage,
but has added to it this energy dissipation, this continuous change and rebuilding.
And, just to end here...
Although I've only talked about the microtubule cytoskeleton...
I've really only talked about mitotic spindles,
I think this principle of self-assemblies
that continuously reorganize themselves, rebuild themselves,
burning energy to do so is also true for the actin cytoskeleton,
for the membrane systems of the cell, for the plasma membrane,
things like this are going on in the nucleus.
Indeed, I think you can view the whole cell as a sort of self-organizing system
along the same principles.
So, I think this way of looking at biology
which is... I'm going to share with your sort of the way I look at biology...
through studying the spindle and microtubules
is a fairly general lens.
And, we'll leave you with that.
And in the next segment, we'll talk in more detail about
how the mitotic spindle and the microtubule cytoskeleton
in a molecular way achieve some of these properties.
Thank you.