Highlight text to annotate itX
Welcome back. I'm going to continue this discussion of self-organization,
but now get a little bit more into the details.
And I ... What I was trying to do in the last talk was really
give you my sense of where molecules become living systems.
But now, we're going to try and get really... look a bit more
about how do actual molecules do that?
And I'm going to use this example of the meiotic spindle here.
Again, here are some outline and concepts that I won't read out now.
The examples I'm going to use are all in the context of cell division,
and I thought it would be nice to start this with the very first description of cell division
by Flemming in 1882 -- more than 100 years ago.
And he looked at fixed and stained salamander cells
and drew these absolutely exquisite pictures.
And you can see here, these lines here are the mitotic spindle,
and Flemming realized what the job of the spindle was,
which was to split the chromosomes into two groups.
You can see them here being split at anaphase
so that the two daughter cells get identical numbers.
What Flemming, of course, didn't know
is why is it so important to split the chromosomes in two.
Now we know that's where the DNA is.
It's important that every cell in our body get an identical copy of the genome,
so the mitotic spindle has to very precisely split the chromosomes in two.
Here's a slightly more updated view of this.
This is a movie of a tissue culture cell dividing.
The chromosomes are these dark objects in the middle,
and you can see here they are lined up in the middle.
And in this movie, you'll see them be split in two,
and the forces are coming from the mitotic spindle.
This is a movie I took -- actually in 1988 --
I took myself when I got my first microscope in my first lab in San Francisco.
I'm very fond of this movie.
I think it shows the process rather nicely.
And then, after the chromosomes, this pinching in is called cytokinesis.
That's where the cell actually divides in two.
In this image, I've super imposed on the phase contrast.
This is not the same cell. This is a much later image, but it's the same cell type.
You can see here the chromosomes stained red,
and the microtubules which are exerting force on the chromosomes
stained green. This is an image by Julie Canman and Ted Salmon.
Microtubules and DNA.
And just a little bit more terminology.
I'm going to use the word spindle pole for this point,
where the minus ends -- the microtubules are polar --
all the minus ends gather together here at the spindle pole.
And in some types of spindles, you have a centrosome here --
a nucleating site.
And then at this point here, the plus ends of some of the microtubules
-- a subset, actually --
meet chromosomes at kinetochores.
This is where a lot of the force is generated to actually move the chromosomes.
I'll be using those words.
I'm not in this talk going to focus on how chromosomes are segregated...
how that's made so accurate.
That's a fascinating topic, and many people doing beautiful work on that.
I'm going to stay more on the question of how you self-organize this shape
and how it gets its life-like properties.
The system that my lab has used to do a lot of this
are eggs here. You can see a frog egg dividing,
and these frog eggs come from the Xenopus laevis,
the African clawed toad,
which is an amphibian that doesn't seem to mind growing in the laboratory.
They live for ages in the laboratory. They aren't hurt.
They lay eggs, and then we have them recover, and then they lay eggs again a few months later.
And you can see it divides beautifully.
And so this is a system I like very much for studying cell division.
And, we study one particular type of spindle in these eggs
which is actually the meiosis II spindle.
And, you can see it in this image here.
This is an unfertilized frog egg, this big sphere.
And the meiosis spindle is this tiny little point up at the top there.
And, I really like this image.
I have a huge one. Martin Wuhr, my student who took this image,
blew up one of these to be 4 feet in diameter,
and it's hanging outside my office.
It's sort of symbolic of my life.
Here's this world of the cell, and I've spent far too many years perhaps
studying this one tiny little point here, up at the top.
But, if you zoom... oh, and let me also emphasize the size here.
The frog egg -- the Xenopus eggs -- are 1.2 mm in diameter,
which really emphasizes the challenge of
actual organization of such a large egg.
Here's a zoom-in to the meiotic spindle, here.
It's about 30 microns in diameter. You can see the DNA in the middle.
And the job of this little spindle here
is to complete female meiosis.
So, when the egg is fertilized, it separates the two copies of the maternal genome --
splits them into two.
One copy is thrown away, and the other copy meets the paternal genome,
which comes in ... a *** would come in something like this.
And the *** nucleus moves into the center
and meets the one half of the female genome.
So, the job of this spindle is to essentially throw away one copy
of the maternal genome and keep the other to make it haploid.
And, a little bit of review here:
I've shown you a somatic spindle dividing -- that PTK2 spindle.
Those spindles have centrosomes here.
But, the spindle we're going to focus on -- these egg meiotic spindles
are organized a little differently.
They don't have centrosomes.
They're sometimes called anastral spindles,
and other organisms make spindles that way.
For example, higher plants lack centrosomes,
and they have spindles that look a lot like these egg meiotic spindles.
Now, Xenopus eggs, although they divide
beautifully, they're large and opaque, which makes it hard to visualize
what's going on with a microscope, and I'm sure it's evident to everyone by now
that the favorite thing in my lab, my favorite thing to do
in biology is to peer at things through a microscope as they happen.
And it turns out that we can get a way around this problem
of opacity of the egg by preparing an extract from the egg
and using that extract to recapitulate biological processes
either in a test tube or under a microscope slide.
And, when you make an extract, in addition to being better for microscopy,
it's great for biochemistry.
You can add and remove components and do many kinds of manipulations
that aren't possible in a living system.
So, this is an experimental system that's very dear to my heart.
The female frog lays eggs -- she's injected with hormones -- lays eggs
and when that's done, she goes back to the colony to recover.
We collect those eggs. We wash them. We pack them in a centrifuge tube
to squeeze out all the pond water that would dilute them (or the buffer we've washed them with).
Then, we crank up the centrifuge to 10,000xg.
That puts a lot of force that crushes the eggs
and also separates their internal content here.
So, the yolk is down at the bottom. There's some fat up at the top.
And, here is the kind of liquid part of the egg -- the cytoplasm --
and this is essentially undiluted. This is a kind of 1x cytoplasm.
This is a method developed by Andrew Murray when he was in Marc Kirschner's lab
and actually my very first PhD student, Ken Sawin,
implemented this system for studying meiotic spindle assembly.
This extract, as I say, it's really cytoplasm.
It contains abundant organelles, mitochondria, ER, etc.
It also contains glycogen, an energy resource, ribosomes to make proteins...
really everything in the cell, except nuclei.
It's metabolically active. It actually consumes about 1mM of ATP a minute,
which it remakes.
And the mitochondria deplete oxygen, which turns out to be very good for fluorescence imaging
because oxygen tends to bleach fluorochromes.
And finally, the cell cycle state can be precisely controlled,
which is essential for these experiments.
So, I wanted to walk you through a very typical extract experiment.
And this is a kind of experiment, hopefully I'm going to be doing later this afternoon
in Woods Hole.
So, the first thing is to prepare this extract.
And on a typical day, we would, for example, collect the eggs before lunch,
and the extract would be ready after lunch.
And you can store it on ice for up to 8 hours.
You can freeze and thaw it, but it doesn't work so well,
so we always prefer to do it fresh if we can.
If you want, you can remove a protein of interest with an antibody.
We almost always add fluorescent proteins so we can do imaging.
A lot of the experiments I'll show you add tubulin labeled with rhodamine,
so we can see microtubules growing.
We typically add a nucleus, or I'll show you some experiments with DNA on beads
to trigger spindle assembly.
You can add a drug, an inhibitory reagent, something like that.
Then, typically, you take a few microliters, put it on a slide,
drop a coverslip on top, seal it with wax,
and then image by live fluorescence microscopy.
And then a very essential part, when you get a good movie...
Later. Not the same day. You have the movie stored, and you would do quantitative image analysis
to figure out what happened and quantify what you've perturbed
compared to wild-type.
And this whole procedure, from here on down,
takes less than an hour.
One reason I love this system, is you can come to the extract
at 1 o'clock in the afternoon with some big theory,
and by 3 o'clock, you've already disproved that theory,
and you're all despondent. You try a few more things,
and then by 7 o'clock in the evening, you've got some new theory,
and you're working crazily into the night to figure out if it's right.
So, it has this very fast pace,
which I just absolutely love. Of course, you have to
make a bunch of reagents beforehand
to allow these experiments to go that fast.
And, we can recapitulate at least two kinds of microtubule assembly.
I'm really only going to talk today about these meiosis II spindles.
We're recapitulating that tiny little spindle that lived at the top of the egg.
I'll show you one movie... We're also increasingly working on microtubule asters,
which is a new direction for my group.
They are functional, these spindles.
I'll show you one movie of a spindle actually segregating chromosomes.
The chromosomes aren't shown here, but in red --
I hope you'll be able to see that these little red dots here are kinetochores.
And to trigger anaphase, we add some calcium.
When the *** comes into the egg, it creates a calcium pulse
that goes through the cytoplasm.
And that's what naturally triggers anaphase
of a meiosis II spindle.
And, you can see here, this movie will loop here.
There's some kinetochores, they're being pulled on, and finally they start moving.
I'll let it loop one more time.
So, this is the meiotic spindle doing
its job, and not inside a living egg now,
but between a slide and a coverslip.
And that was taken by a good friend, Paul Maddox,
who was a real *** with the spinning disk microscope.
So, the kinds of questions we've been asking, we and other people studying the system,
What signals trigger spindle assembly?
I've been fascinated in how the microtubules behave in spindles.
You can see from the introductory bit, I'm very much a student of microtubule dynamics.
And then, how are the spindles spatially organized?
And I'll rather briefly address those and give you some references
if you want to get deeper into these.
So, first, what triggers spindle assembly?
This question of what regulates spindle assembly can be broken down
into a time question -- what triggers when the spindle occurs --
and then a space question -- where it occurs.
And none of this is work from my lab.
The time question... To trigger spindle assembly, the cytoplasm must enter
a mitotic state, and this occurs by activation of Cdk1
kinase (the Cdc2.CyclinB kinase).
That's the master kinase that turns on all of mitosis.
That's all textbook stuff.
And it's by manipulating that system that we control
whether the extract is in interphase or mitosis.
The space question is really why do you get
a meiotic spindle right here in the egg
and not somewhere else?
Why precisely in one place? And this is the exact
north pole of the egg that there's spindle assembly.
What's special about that place that makes the spindle form
there and not anywhere else?
And, what's special about that place
is that's where the DNA is -- the maternal genome.
That's where the condensed chromosomes of the maternal genome are.
And, what we now think is that the spindle forms around the DNA.
And this was really the discovery and part of the life work of a good friend of mine, Eric Karsenti.
Here's a couple of classic references.
He used to, back actually when I was a graduate student, he was in Marc Kirschner's lab,
and he was injecting DNA into eggs and watching what happened.
And he noticed, if you injected DNA, spindles would form around them ectopically.
And he hypothesized all the way back in 1984 that DNA emitted some signal here,
which he imagined would be a kind of gradient -- high at the DNA
and declining elsewhere -- that would trigger spindle assembly to happen
right there at the top of the egg
and not anywhere else.
And, for many years, the nature of that signal was completely mysterious,
but now we know in molecular terms of at least two signals.
And, there was a beautiful experiment to prove this, that the DNA is the trigger,
done by Rebecca Heald when she was a post-doc in Eric Karsenti's lab.
She took a completely random DNA sequence from a bacteriophage,
coupled it to plastic beads, using magnetic beads as those are convenient to manipulate,
put those in the extract, and found that they triggered
microtubule polymerization and spindle assembly.
And there's one of their spindles again.
So it's really just a classic proof of Karsenti's concept
that the DNA is the spatial trigger.
I'll show you a movie of that, taken here at Wood's Hole by Aaron Groen and Tom Maresca
in the egg extract system. This is the scale bar here.
I hope you can see here, this clump of circles is the DNA beads.
And you can see the microtubules first form rather randomly around the beads
and gradually organize themselves -- this is hours and seconds playing out here --
into this characteristic bipolar structure.
Now, this kind of spindle can't actually separate the beads,
You know, unlike chromosomes, you can't split them in two and separate them.
But, we believe the actual assembly process uses all the same factors and is
in many ways similar.
And so our picture of self-organization of meiotic spindles is
that you start with DNA, assembled into chromosomes. We call that chromatin.
That triggers microtubule nucleation near the DNA.
The microtubules are then sorted out and bundled,
and motor proteins are important for that. I'll come to those in a little bit.
But, it's also very important, going back to the themes of my introduction,
not to think of this as just a linear pathway,
because once you have this spindle, assembly isn't finished,
this structure is continuously rebuilding itself.
So, all the time, this same process is going on, so it's building itself continuously.
That's steady state.
DNA is sending out some kind of signal. What is the nature of that signal?
And I think Eric, early on, felt that it would likely be what we could call a reaction diffusion system.
That's something where a local source, stuck to DNA or stuck to chromatin,
creates some unstable molecule that can diffuse away.
That molecule is unstable because it's broken down, here by a second enzyme.
And, if you have a reaction-diffusion system like that,
it'll give rise to a spatial gradient in concentration
of the signal. And those of you who are mathematically inclined could figure out
what the shape of this curve is -- what mathematical function it would be
for a point source.
Here, it suffices that the signal is high at the DNA and then decreases
due to a combination of production, diffusion, and break down.
And this concept is actually very similar to the concept of the morphogen
in developmental biology -- a molecule that diffuses to pattern an embryo.
The difference here is that, instead of diffusing between cells,
it's diffusing inside of a cell.
And I think the spindle morphogens, if you want to use that term,
diffusing away from DNA is one of the best examples we have
of a reaction-diffusion system operating inside of a living cell.
So, I won't go into all the history of this.
We actually now know of two signals emitted by DNA.
One is a small GTPase, Ran, in the GTP form,
and another is protein phosphorylation. The Ran signal is created by
Ran GEF and then broken down locally on the chromosome,
and then a GAP... this [the GEF] is a GTP exchange factor.
This [the GAP] is a GTPase activating protein.
The phosphorylation gradient is generated by Aurora B kinase
and then opposed globally by phosphatases.
There's a review for people who want to learn more about this.
I'm only going to talk today about the Ran gradient.
Sorry, this one. But this phosphorylation gradient is also very important.
I'm trying to keep it as simple as I can here.
Here's a cartoon of the Ran system. The GTP exchange factor
is localized on DNA. Ran-GDP -- that's the inactive form of this small GTPase --
is then switched on catalytically (GTPase exchange) by Rcc1.
This [Ran-GTP] is the active factor that triggers spindle assembly.
Then the Ran-GTP is converted back to Ran-GDP by a GTPase activating protein
And this was discovered by a bunch of people. I've put some names there.
My lab was not one of them. I think this is absolutely fabulous biology.
One other piece of information. This is the exact same system that powers nuclear import
during interphase. So, this is the system that powers NLS-dependent transport
through nuclear pores, although I won't be talking about that today.
And what Ran does -- Ran-GTP... There are important spindle assembly molecules
(here's one, called TPX2) that are sequestered.
They're sequestered by binding to a molecule called importin,
which is an alpha-beta heterodimer.
So, TPX is inactive here, because it's sequestered.
Ran-GTP binds to importin-beta and dissociates this complex into three parts.
Now, the TPX is free to do it's job, which is organizing spindles.
TPX2 is actually one of the proteins Eric Karsenti worked on and my lab works on.
We actually don't know precisely how it works.
We do know it's a key spindle assembly factor.
So, this is the biochemistry.
And, one question I think you'd very naturally like to ask is,
Ok, I'm asserting there's a spatial gradient triggering that self-organization
I showed you around the DNA beads, but can we actually see that spatial gradient?
Can we visualize it?
This is very difficult to do, because, for example, if you fixed a cell for immunofluorescence,
these are putatively soluble molecules,
they'd probably all move around during your processing.
And so this is a very difficult challenge.
And this challenge was solved by Rebecca Heald and Karsten Weiss.
Heald was the student of Karsenti's who did that DNA bead experiment.
They built an elegant FRET biosensor. This is fluorescence resonance transfer.
And this is a kind of biosensor that people are using
for a lot of applications in cell biology nowadays,
that will allow us to see changes in the chemical state of the cytoplasm with a microscope.
And the way this works is there's a Ran-binding domain, shown here in green.
This is easier to see here.
Fused on it's N or C terminus (I'm not sure which is which,
I've put the reference here if people are interested)
to CFP and YFP.
And when this molecule is on its own,
CFP and YFP are close enough together to get efficient FRET --
fluorescence resonance energy transfer -- from the CFP to the YFP.
So, the FRET signal is high.
That's if Ran is in the GDP form.
If Ran is in the GTP form, it binds to the Ran-binding domain,
kind of separates the two halves of this molecule in space, and the FRET signal goes down.
And this is one of two different probes that Heald and Weiss developed.
And, here is their sensor doing its work.
So, it's a multicolor live imaging experiment.
Here, we're seeing the microtubules with fluorescence tubulin.
Here, we're seeing the YFP signal from the probe.
Here, the CFP signal, and here the FRET signal.
And this color here, indicates this is a low value of FRET.
If we go back, the low value... This is the high value, this is the low value.
That's where the Ran-GTP is, where the FRET is turned off.
And so, you can see here, this low value in the gradient
indicates that there's a high concentration of Ran-GTP here.
So, here we're actually visualizing, live in the extract,
this reaction-diffusion gradient of Ran-GTP,
which I just think is a really remarkable thing to see.
As I mentioned earlier, I think this is one of the best examples
of a reaction-diffusion gradient in cells.
There are a lot of questions about this.
One thing you'll notice is that the shape of the space in the extract which has high Ran-GTP
concentration is not the same shape as the microtubules.
So even, although we believe this high Ran-GTP area
here is triggering assembly of microtubules,
we think it can't be dictating the exact shape of the microtubules
or, for example, the exact length of the spindle.
And how those are organized, I think, is going to come more out of the
microtubules themselves and the motor proteins that I'll talk about in a bit.
How do microtubules behave in the spindle?
So, the second topic I want to take up.
So, I talked in the introduction about how the spindle is continuously rebuilding itself --
exchanging subunits with the cytoplasm.
We've known that for a long time.
But, what is the nature of that continuous rebuilding?
Just to give you a sense of the microtubules here,
this is an electron micrograph of the spindle.
And here's a cartoon representation.
One of the things I want to point out is that in these egg extract spindles,
the vast majority of microtubules are of this kind here that don't
attach to kinetochores, and only about 5% of the molecules attach to chromosomes.
When I'm talking about microtubule behavior,
using the kind of bulk measures I'm going to show,
I'm talking about these so-called non-kinetochore microtubules.
Those are the microtubules that give the spindle its size and shape.
They're not the microtubules that actually segregate the chromosomes,
but they're the most relevant microtubules for talking about the self-organization problem.
So, one thing we know is that microtubules turn over very rapidly, as I implied
in the cartoon.
A classic experiment that was done when I was still a PhD student
by Ted Salmon, *** McIntosh, and coworkers
is a fluorescence photobleaching recovery experiment.
This is 1984, so this was done with a videocamera, not a modern digital camera,
so it looks a little fuzzy.
But, the idea in this experiment is you have a spindle containing fluorescent tubulin.
This is microinjected fluorescein-labeled tubulin.
A laser was used to bleach one half of the spindle.
And key to this experiment is you're not actually damaging the spindle,
you're just turning off fluorescence.
And then, a few seconds later, you can see recovery here.
And you can plot a curve of recovery.
And in fact, the half-time of recovery is about 19 seconds.
So, Salmon and McIntosh inferred that subunits are exchanging sufficiently
rapidly that in 19 seconds, half the microtubules in the spindle have been turned over
and refreshed. So even though the spindle can last hours,
the subunits within it only persist for
tens of seconds inside the spindle.
And that really is very characteristic of the spindle
and important for those properties that I talked about in the introductory material.
I'll show you now... there's lots... I've spent a long time --
me and my coworkers -- studying microtubule behaviors and spindles.
I'll show you one recent experiment done by Dan Needleman and Aaron Groen,
which is a single-molecule imaging.
So, imaging has come a long way since that 1984 paper.
Now, here, using a spinning disk confocal
and a very sensitive camera, we can see these dots here.
These turn out to be single tubulin molecules.
And what we've done here is we've spiked a very low concentration
of fluorescent tubulin into the spindle, such that only 1 in every 200,000
approximately is fluorescent.
That means a single microtubule on average is only going to have 1 fluorescent spot in it.
And, if I step back and play the movie,
I hope you can see here these dots.
Usually our eyes are most attracted to the fact that these spots are moving.
They're all sort of sliding, if you look carefully. The ones in the southern half are sliding south,
and the ones in the northern half are sliding north.
But, if you look carefully at each dot and try to follow it,
you'll notice it will disappear or a new one will come up.
And, our interpretation of that movie is shown here.
So, when a tubulin molecule is just diffusing around in solution...
In cells, approximately half the tubulin typically in a somatic cell
is soluble, and about half is in polymer.
Here in the extract, the vast majority is soluble.
A soluble tubulin molecule is diffusing around very quickly.
It takes about 500 milliseconds to acquire an image on the camera.
In that time, the soluble molecule will just diffuse around and blur out.
If that molecule is captured by a growing microtubule,
it now moves much more slowly and we can now image it.
We can image it for as long as its in the microtubule.
When the microtubule depolymerizes again, it goes back to being soluble.
So, we can... I've just made that point here.
We can actually measure the lifetime in polymer of a single tubulin molecule
by seeing how long we can track it for.
And then, if you watch the spot moving --
I've shown here moving toward me --
we can also track the whole microtubule sliding. It's a kind of fiducial mark, if you will
for the microtubule sliding.
It's easier to see the turnover -- the single spots coming and going --
if you block the sliding.
In this experiment, the sliding was blocked with a drug that freezes kinesin-5,
a motor in the spindle I'll talk about.
Now, the spots aren't sliding, you can see them almost twinkling.
What that twinkling is, is each spot comes, persists for a few seconds,
and disappears again. And this experiment also shows that you can uncouple sliding
from polymerization dynamics.
And those spots are sufficiently bright and well separated
that they can be tracked by automatic software.
This is Dan Needleman's work, using software taken from physics labs
who study particles moving.
You can see here, trajectories -- these are vertical lines.
This is a control spindle, this is one of those drug-frozen spindles.
So, everything's frozen there.
And, using that kind of software, you can quantify
both the sliding and also how long each spot persists,
which is the lifetime of that spot in polymer.
And, here's a lifetime plot.
So, here's how long the spot can be imaged for,
versus the fraction of spots, and there's about 16,000 individual spots
that went into this graph here.
And, the blue spots are data here. The red line here is a fit here to a theoretical curve.
t to the -3/2 [times] e to the -t over tau.
There's 1 adjustable parameter - this tau parameter.
And then the average lifetime here is 16 seconds, so the midpoint of this distribution.
A couple of points here: this 16 seconds here is, I hope you'll recognize,
remarkably similar to that classic Salmon and McIntosh number in sea urchins.
So, this rate of turnover with this more sophisticated modern method
really confirms roughly the old value.
This equation here... I have to say this is the most complicated equation
I've ever run into in my own work.
It took me a long time to understand this.
What this equation actually is, if I go back,
to this slide here... This is the first passage time for a 1-dimensional random walk.
It's the expected time if the tip of the microtubule is doing a random walk
in position. 1st passage time is coming from the red spot back to the red spot.
So, this is exactly the curve here that you would expect if the + ends
of the microtubules are doing a random walk.
And this dynamic instability process -- these large length fluctuations --
we believe is a form of random walk.
So, this is a satisfying match to theory for what, to me, is a fairly complicated equation.
I would say it's not quite that simple.
And if people are more interested in the quantitation, we had a lot of arguments with the reviewers
of the paper about how to precisely interpret that,
and I would refer people to Dan Needleman's paper
in MBC in 2010, if they're interested in that.
So, summarizing a lot of work of that kind,
the picture of the spindle that emerges is here, where the yellow
bars are some kind of nucleating structure, the green is the microtubule,
and the white arrows are the sliding movement.
So, microtubules are nucleated throughout the spindle,
their plus ends grow and shrink by dynamic instability, which has a random walk-like character.
On average, they live 16 seconds,
and grow to about 5 microns in length.
And then they slide polewards, mostly, at variable rates.
So, there's this sliding phenomenon.
So, the microtubules are quite a bit shorter on average
than the spindle itself, and they're all sliding.
So, this, I think matches now this picture that I gave in the introduction
of the spindle continuously renewing itself,
and it shows us how the microtubules are behaving.
It leads of course to a lot of other questions.
One question we've been very interested in is how are these short microtubules organized?
Why are they all sliding? And I'll give you some answer to that.
And then how are they nucleated?
And what are these yellow boxes here that are initiating the microtubules?
That's an unsolved problem that we and others are working on at the moment.
For the organization of sliding, we know there's a very important role
for motor proteins -- ATPase motor proteins, which I'll introduce
a bit more in a second.
We know that mainly from inhibition or depletion experiments.
Here's a series of inhibition experiments that were done here at Woods Hole in 2000.
Here's some control spindles by polarization microscopy.
Here we've inhibited the minus-end-directed motor, cytoplasmic dynein.
And I hope you can see that the poles, instead of being focused as in the controls,
are all kind of splayed out.
It's very clear, I hope, in that one.
Here, we've inhibited a plus-end-directed kinesin that I'll talk about in a second,
called Eg5 or kinesin-5.
Now, the poles are collapsed together.
This shape here is a radial array of microtubules or aster.
Fascinatingly, and we don't actually fully understand this at all,
if you inhibit both motors, the structure sort of fixes itself.
It looks a bit like control, though these are much more fragile.
They fall apart more easily, especially if you poke them.
And that kind of experiment gave us a picture of the spindle like this,
where the minus-end-directed motors help cluster the poles together.
So, if you get rid of them the pole splays out.
And the plus-end-directed motors are helping keep the spindle bipolar.
We spent a lot of time on the plus-end-directed motor in this problem, kinesin-5.
It's a plus-end-directed tetrameric kinesin,
and homologs of this motor are required for spindle assembly in most eukaryotes,
going down all the way to yeast and up to plants.
I'm cartooning it like this.
These are the four motor domains, and there's two coiled coils
that interact with each other in an anti-parallel fashion.
Here's a cartoon of it.
Here's a movie of kinesin walking along microtubules from my colleague, Ron Vale and his lab,
based on crystal and electron microscopy structures.
This is drawn for conventional kinesin,
but if you imagine this being one end of Eg5
and then the other end being up here,
and the two ends are the same, so of course, it's appealing to hypothesize
that they interact with two different microtubules
and cross-bridge. And that idea was first proposed by Jon Scholey,
quite a while ago --
a colleague in the field.
And, one reason we've been able to work on kinesin-5 in our spindle system fairly easily
is we have some nice small molecule inhibitors -- drugs.
And these drugs bind... this is the ATP-binding pocket
in the head of kinesin-5.
This is a nice drug binding pocket.
The first drug that we actually found that fits into that pocket
is a drug called monastrol,
which we found by chemical screening back in 1999.
Since then, much more potent and specific inhibitors have been developed
by industry groups in the hope of their being useful for treatment of cancer patients.
That actually is a whole story.
Our kinesin-5 inhibitors have not been successful yet in treating cancer,
and why that is, I think, is an interesting story that I'm not going to talk about today.
These class of drugs trap the motor in a state with weak affinity for the microtubule
so it's like removing the motor.
And we can do the kind of experiments shown here.
Here's a control spindle where this kind of speckles here
are ... you can see the sliding of the microtubule.
Here's the same kind of spindle treated with a kinesin-5 inhibitor.
And if I step back, I hope you can see the microtubules are sliding apart continuously in this one,
whereas here, they're static.
And, from that kind of movie, the image we get is this one,
very much similar to Jon Scholey's original hypothesis.
The kinesin-5 is bridging two anti-parallel microtubules here.
This end is walking towards the plus ends,
that end is walking towards the plus ends.
That's creating a force on the microtubules shown in green
that slides them apart
and kind of pushes the two poles of the spindle apart.
And that's a current view of how kinesin-5 works in the field.
And so, here's this kind of motor-driven sliding field,
and I want to relate that to the image I showed in the introduction,
where I talked about the poles being pushed and pulled
and now bring in specific molecules.
So, the dynein is kind of focusing the poles together,
and pulling them inwards, the kinesin-5 is pushing them out.
And those motors are coming and going from the spindles.
They're just like tubulin -- they turn over very quickly.
And so the spindle is continuously repositioning the motors
and rebalancing the forces.
And again, this is part of this dynamic steady state
that gives the spindle its life-like properties.
So, a lot of details... I hope that was interesting.
Let me quickly summarize them.
Egg meiotic spindles self-organize
in the vicinity of DNA, triggered by a reaction-diffusion
of Ran-GTP and also a kinase gradient.
They're built of relatively short microtubules
whose plus ends undergo continuous dynamic instability.
They're organized by these motor proteins,
plus- and minus-directed motors that promote sliding and clustering.
And, what I would argue, although we don't fully understand it,
is together these reactions generate this very dynamic steady state
in size and shape that's continuously rebuilding itself,
continuously adjusting the forces so they balance,
and that's what gives these very life-like forces.
So, if we go back to this movie I showed in the introduction of
two spindles fusing,
which to me, more than anything, shows the strange properties of this structure,
compared, for example, to the bacteriophage,
which certainly wouldn't do this,
I hope that that more molecular kind of picture
can help you begin to frame how this works,
although I do not want to pretend that this is a solved problem by any means.
There are many future questions.
Particularly, how are microtubules nucleated?
The size and shape of the spindle are very much governed by
where microtubules are nucleated,
and that's an unsolved problem biochemically.
We know it's downstream of that Ran gradient,
but what it is in molecular terms, we don't know.
We would like, and we're going to certainly need some help from physicists
and mathematicians here, to develop a quantitative understanding,
so we can actually explain size and shape...
these and also time scales in quantitative terms.
And, I'll leave you with one last point here.
It's interphase microtubules.
So if you change the cell cycle, everything changes.
And, for many years, we just had only focused on mitosis.
Now, we've gotten very interested in interphase.
And, that's a whole seminar, but let me just
give you one snippet here.
This is the work of Ani Nguyen and Christine Field.
This is similar kind of microscopy, lower magnification.
This is very low mag -- this whole field is 600x800 microns.
Almost a millimeter.
In blue here are *** nuclei.
In green are microtubules.
And what we're doing here is we're mimicking fertilization.
We're putting in *** and calcium at the same time
to make the extract go into interphase.
So, the *** are going to build this big aster that they use
for organizing the egg that's similar to that zebrafish aster
that I showed you very early on in the introduction.
And what you'll see when this movie runs
is these asters can move, and they can solve a kind of geometric problem
of spacing themselves out.
So, I'll stand back a little so you can see
the microtubules growing out,
interacting here, plus-end to plus-end.
And I hope you can see that they're actually
solving a fascinating geometric problem
of spacing themselves out evenly in the field.
How that works and the role in biology
of fertilization and early development is a whole other topic.
And if I want to leave you something there,
it's like there is really still many cool self-organization problems left to study.
So, let me come back to a slide
that I used in my introduction here,
where a complex and beautiful bacteriophage
that has an invariant size and shape that's kind of directly encoded by genes.
The meiotic spindle, I spent a lot of time talking about
now, where a self-organization pathway leads to a much more dynamic assembly
that's at steady state, continuously rebuilding itself,
continuously rebalancing forces,
and that gives it the ability to adapt in size,
to repair damage, and these other kind of life-like properties.
Both beautiful, but somewhat different in principle.
And to me, you know, this one is really alive.
I hope I can use that word.
And, let me finally end by thanking
the many wonderful students and post-docs
I've worked with. I didn't make a list.
I hope I mentioned many of them by name in the talk.
Particularly, my good friend Ted Salmon, who I've worked with a MBL for many years
as the MBL Cell Division Group. I showed a number of movies taken here at the MBL.
Also, inspiring colleagues, particularly Marc Kirschner, my PhD advisor,
Eric Karsenti, who discovered the idea of DNA organizing the spindle,
Shinya Inoue, who ... really one of the earliest and most thoughtful students of self-organization.
NIH-GM who funded most of this work.
They are the main supporters of basic biological research in the US.
Go, GM! They turned 50, I think, last year.
And finally, of course, the frogs, who made all this possible.