Molecular Motor Proteins

Transcript of Part 2: The Mechanism of Dynein Motility

00:00:15.09	Hello.
00:00:16.17	I'm Ron Vale from UCSF.
00:00:18.24	In part one of my iBiology talk,
00:00:20.20	I introduced biological motility
00:00:23.14	and I focused on the mechanisms
00:00:25.05	of kinesin and myosin,
00:00:26.13	and in this talk
00:00:27.25	I'd like to discuss
00:00:29.08	our more recent research
00:00:30.26	on the mechanism of dynein motility.
00:00:33.29	Now, this year, 2015,
00:00:37.18	is the first anniversary
00:00:39.12	of the discovery of dynein,
00:00:40.29	which was made by Ian Gibbons,
00:00:43.12	and Ian described
00:00:45.11	a new type of ATPase
00:00:47.13	from cilia
00:00:49.07	that was involved in powering its motility.
00:00:53.13	Now, even though dynein was discovered
00:00:55.21	twenty years before kinesin,
00:00:57.25	we know a lot more about kinesin motility
00:01:00.22	than we do about how dynein works.
00:01:03.16	And the reason for that is that
00:01:06.13	dynein is an incredibly complicated machine.
00:01:08.23	First of all,
00:01:10.15	it's a massive protein complex,
00:01:11.29	one of the largest in the cell.
00:01:14.11	It's about one and a half megaDaltons
00:01:16.22	in size.
00:01:17.28	Even the gene that encodes
00:01:20.18	the motor polypeptide
00:01:22.20	is very large.
00:01:24.25	The motor polypeptide
00:01:26.01	is about 500 kiloDaltons in size,
00:01:29.02	so it's one of the largest polypeptides
00:01:30.18	in the genome.
00:01:32.03	Even the motor domain itself
00:01:34.12	is massive, shown here.
00:01:36.16	It's about eight times larger
00:01:38.13	than the kinesin motor domain.
00:01:40.20	So simply because this motor
00:01:42.08	has been so big and complicated,
00:01:44.14	it has made it difficult to study.
00:01:46.15	It's been difficult to express it
00:01:48.29	and obtain pure protein,
00:01:51.07	difficult to manipulate it
00:01:53.10	by recombinant protein techniques,
00:01:56.01	and also non-trivial to get a structure
00:01:59.05	by X-ray crystallography.
00:02:01.01	So all of these have posed challenges.
00:02:04.18	And in 2007
00:02:06.17	I actually gave an iBiology talk,
00:02:08.25	which you can find in the archives,
00:02:11.02	and at that moment in time
00:02:12.26	we were primarily studying
00:02:15.02	the single molecule motility of yeast dynein,
00:02:18.25	but we knew very little
00:02:20.25	about the structure of dynein
00:02:22.18	in any great detail.
00:02:24.28	Well, a lot has changed
00:02:26.19	in those intervening years
00:02:29.06	and I'd like to share to share with you today
00:02:31.21	recent progress that's been made
00:02:33.28	on the dynein structure,
00:02:36.19	and now we do have atomic structures for dynein
00:02:38.25	and we're beginning to get
00:02:40.15	some insight into how this motor works.
00:02:44.09	So first of all,
00:02:45.14	let me just tell you about the different kinds of dyneins.
00:02:48.17	A major class of dynein
00:02:50.12	are the axonemal dyneins.
00:02:52.07	These are dyneins that power
00:02:54.04	the movement of cilia,
00:02:56.04	they also power the movement
00:02:58.21	of flagella from a sperm, for example,
00:03:01.16	and this just shows what the architecture
00:03:03.17	of the axoneme looks like.
00:03:05.27	It's composed of
00:03:09.14	nine unusual types of microtubules
00:03:11.13	that are called outer doublet microtubules,
00:03:15.04	and they have this particular structure here.
00:03:18.16	And on these outer doublet microtubules
00:03:21.20	sits the dynein molecule.
00:03:23.03	In fact there are two different kinds of dyneins
00:03:25.03	-- there's an outer arm dynein
00:03:26.08	and an inner arm dynein --
00:03:28.07	and they sit statically
00:03:31.02	on one of the outer doublets,
00:03:33.16	the A tubule,
00:03:35.23	and then they reach across
00:03:37.03	to the neighboring B tubule
00:03:39.15	and they cause a relative sliding
00:03:41.21	between these adjacent microtubules.
00:03:44.20	Now, this sliding motion
00:03:46.11	of the two microtubules
00:03:49.01	in the cilia
00:03:50.27	gets converted into this sinusoidal
00:03:52.27	beating pattern
00:03:54.14	by a process that's still
00:03:56.10	very poorly understood.
00:03:59.00	Now, in addition to these axonemal dyneins,
00:04:01.12	there are cytoplasmic dyneins
00:04:02.27	that are more cargo-transporting motors.
00:04:06.15	So, one of these
00:04:07.28	is cytoplasmic dynein 2
00:04:09.11	and it's responsible
00:04:10.21	for a particular type of cargo transport
00:04:12.24	that actually occurs inside of cilia and flagella,
00:04:16.15	and it's responsible for transporting,
00:04:19.01	first of all, building blocks
00:04:21.00	up and down
00:04:23.10	for the maintenance of cilia and flagella,
00:04:24.23	and also some kinds of signaling molecules
00:04:26.20	are also present in cilia,
00:04:29.05	and they are trafficked
00:04:31.27	by motor proteins as well.
00:04:34.09	This shows an image of what
00:04:36.13	cargo transport looks like,
00:04:38.29	what intraflagellar transport looks like,
00:04:40.13	by fluorescently tagging
00:04:42.06	one of these cargo subunits.
00:04:43.29	Now, a kinesin motor
00:04:46.15	transports the cargo up to the tip,
00:04:48.17	to the plus end,
00:04:50.00	but dyneins are minus end-directed motors,
00:04:51.19	and so the dynein
00:04:55.08	transports cargo in the opposite direction,
00:04:56.24	from the tip towards the cell body.
00:04:59.02	Now, there's also a cytoplasmic dynein 1,
00:05:01.24	and that's responsible
00:05:03.09	for virtually all of the cargo transport
00:05:05.00	that occurs in the cytoplasm of cells
00:05:08.06	towards the minus end.
00:05:10.01	So kinesins,
00:05:11.23	as you remember from the first lecture,
00:05:13.03	move things out to the plus end,
00:05:14.22	and this one type of cytoplasmic dynein
00:05:16.08	carries out
00:05:17.28	virtually all of the transport in the opposite direction,
00:05:20.18	towards the minus end.
00:05:21.27	So, things that are transported
00:05:23.12	include membranes,
00:05:24.29	mRNAs,
00:05:26.14	protein complexes are transported,
00:05:28.02	as are viruses.
00:05:29.16	Here's just one example, here,
00:05:32.08	which is a melanocyte cell.
00:05:33.27	This is a skin cell
00:05:35.21	that carries pigments
00:05:37.25	which are melanosomes,
00:05:39.27	and in some organisms
00:05:41.22	like amphibians and fish
00:05:43.23	they can change the distribution
00:05:45.07	of their melanosomes,
00:05:47.14	so that when they're distributed outward
00:05:50.11	the skin color is dark,
00:05:52.18	when they're moved inward
00:05:54.07	the skin color turns lighter,
00:05:55.14	and this transport
00:05:57.11	towards the center
00:05:58.20	that you see here
00:05:59.27	is driven by cytoplasmic dynein.
00:06:02.12	Now, in the first iBiology talk,
00:06:04.29	I told you that myosin and kinesin, in fact,
00:06:08.17	are relatively similar to one another.
00:06:11.12	They're similar in structure
00:06:12.27	and they clearly evolved,
00:06:14.11	some time in evolution,
00:06:16.11	from a common ancestor.
00:06:18.28	But even though kinesin and dynein
00:06:22.02	both operate on microtubules,
00:06:24.05	they're not at all related to one another.
00:06:26.03	In fact, dyneins
00:06:27.27	emerged from a completely
00:06:29.29	different evolutionary lineage of ATPases
00:06:32.18	and they belong to a group of ATPases
00:06:34.26	that are called the AAA ATPases.
00:06:38.15	And in fact dynein
00:06:40.11	is a rather unusual member
00:06:41.27	of this AAA ATPase family.
00:06:44.01	Most of them are not traditional motors
00:06:46.00	that you think of in terms
00:06:47.16	of moving along a track.
00:06:49.05	Instead, they use ATP energy
00:06:51.24	to produce mechanical work
00:06:54.03	on molecules like proteins
00:06:56.00	to basically break them apart
00:06:57.27	and unfold them.
00:06:59.13	So, an example
00:07:01.16	that occurs in bacterial and eukaryotic proteolysis
00:07:06.05	is that there's AAA ATPases
00:07:09.07	that sit on top
00:07:11.07	of the proteolytic chamber
00:07:12.15	and their job is to take an incoming polypeptide
00:07:15.19	and basically unravel it
00:07:17.10	and stuff it through this hole
00:07:19.24	into the proteolytic chamber
00:07:21.28	so that the polypeptide can be degraded.
00:07:25.20	So, let me tell you
00:07:27.04	a few things are kind of
00:07:29.00	more universal
00:07:30.13	about these AAA ATPases.
00:07:32.12	First of all, most AAA ATPases
00:07:36.16	encode a relatively small protein
00:07:39.06	that has two domains,
00:07:40.13	a large domain and a small domain.
00:07:43.11	This is the basic ATP binding unit,
00:07:46.12	but this single subunit
00:07:47.25	is not the functional element
00:07:49.17	of how these proteins work.
00:07:51.22	They self-assemble,
00:07:53.23	oligomerize into a hexamer,
00:07:55.10	and it's this hexamer that's the active agent.
00:07:59.06	And in fact
00:08:01.10	adjacent subunits help one another
00:08:03.01	to hydrolyze the ATP,
00:08:05.02	and in the last example thing ring-like structure
00:08:08.09	is what actually unfolds the polypeptide
00:08:11.14	and stuffs the polypeptide
00:08:13.19	into this chamber that you see here.
00:08:15.27	Now, dynein again
00:08:17.11	is unusual in the fact
00:08:19.06	that it makes
00:08:21.15	a ring of AAA ATPases,
00:08:24.05	but it does so by placing
00:08:26.11	all the AAA domains
00:08:28.01	in one very, very long
00:08:29.29	polypeptide chain.
00:08:31.22	And this just shows the motor domain structure
00:08:35.06	of dynein
00:08:36.26	and shows the positions
00:08:38.18	of the six AAA domains.
00:08:41.05	And because they're in one polypeptide,
00:08:44.15	they've each evolved different amino acid sequences
00:08:46.26	over time
00:08:48.10	and have evolved different functions.
00:08:50.00	So AAA1, for example,
00:08:53.05	is the main ATPase site of dynein,
00:08:55.22	so this is what is really responsible
00:08:57.20	for driving motility, as I'll show you later.
00:09:00.08	AAA2 binds ATP,
00:09:02.09	but it doesn't seem to be...
00:09:05.14	bind it in a cyclic or hydrolytic manner.
00:09:08.18	AAA3 also plays an important role
00:09:10.22	that I'll come back to later
00:09:12.08	and it may be a mechanism
00:09:13.26	of regulating dynein.
00:09:15.14	AAA4 also hydrolyzes [ATP],
00:09:18.03	but it seems to play a very minor role
00:09:20.23	in dynein motility
00:09:22.06	and one that we don't completely understand.
00:09:25.07	So, I'd like to address this subject now
00:09:27.05	of how dynein can move along a microtubule,
00:09:30.17	and in addressing this problem
00:09:32.22	one has to tackle it
00:09:34.24	using different kinds of techniques
00:09:36.19	that are complementary.
00:09:38.04	So, one approach is to measure
00:09:40.09	the motility of dynein,
00:09:41.20	particularly at the single molecule level,
00:09:43.24	and this gives you all kinds of information
00:09:45.25	about the dynamics of dynein
00:09:47.10	and how it's stepping along the microtubule track.
00:09:50.13	But it's relatively low resolution information,
00:09:53.26	in other words it can't really see
00:09:55.28	the protein structure
00:09:57.12	and what it's doing.
00:09:58.22	On the other hand,
00:10:00.12	we can do X-ray crystallography
00:10:01.28	or do electron microscopy
00:10:04.27	and these give higher resolution information,
00:10:07.14	even down to atomic detail,
00:10:09.08	but they're static images,
00:10:11.00	so, you know,
00:10:12.18	we see the protein frozen in time
00:10:14.11	and it doesn't provide the information
00:10:16.02	on the dynamics.
00:10:17.13	So, somehow to piece together
00:10:19.21	the answer to this problem,
00:10:20.28	one has to use the information from both of these techniques
00:10:23.03	and try to work out a model
00:10:25.03	of how dynein might work.
00:10:28.05	So, let me tell you first about
00:10:30.04	in vitro motility assays.
00:10:32.02	This shows an in vitro motility assay
00:10:33.29	for yeast cytoplasmic dynein,
00:10:35.26	where we've labeled the dynein
00:10:37.14	with a fluorophore,
00:10:38.29	and you can see
00:10:40.13	these individual dynein molecules
00:10:41.29	moving beautifully
00:10:43.19	along these microtubule tracks.
00:10:45.02	It's processive movement,
00:10:46.29	meaning the dynein can take many steps
00:10:48.23	along the microtubule track
00:10:50.00	without letting go.
00:10:53.03	Now, we can also
00:10:56.05	measure this motility
00:10:57.14	with greater precision
00:10:59.09	if we use a computational approach.
00:11:02.01	Basically, each of these individual spots,
00:11:05.15	fluorescent spots of dynein that you see...
00:11:08.24	as they pass through the microscope,
00:11:10.25	the light spreads out
00:11:13.28	to what's know as a point spread function,
00:11:16.12	so they appear...
00:11:18.12	these single fluorophores
00:11:20.22	appear to have a diameter
00:11:22.05	of about 250 nanometers.
00:11:24.08	But if you collect enough photons,
00:11:26.21	you can describe that fluorescence,
00:11:28.29	that spread out fluorescence intensity profile,
00:11:33.22	and you can fit that intensity profile
00:11:35.22	with a Gaussian curve,
00:11:37.20	and the center of that Gaussian curve
00:11:39.14	defines kind of the midpoint
00:11:41.08	of where that fluorescent spot is.
00:11:44.15	Now, you can take successive
00:11:46.29	snapshots of dynein moving along the microtubule
00:11:49.18	and at each snapshot
00:11:51.10	you can mark the position of that centroid,
00:11:54.19	and that's what all these individual dots
00:11:58.01	are here, data points are,
00:11:59.16	and this is for a kinesin molecule,
00:12:01.16	but you can see for example, here,
00:12:03.19	the motor protein
00:12:05.05	was pretty much stationary on the track
00:12:07.07	and then it took a jump forward,
00:12:09.26	so it took an abrupt step forward
00:12:12.25	along the microtubule track,
00:12:14.21	and this kind of mechanism
00:12:16.04	allows you to get
00:12:18.20	a great deal of information
00:12:20.10	on the stepping behavior of the motor
00:12:22.06	on the microtubule.
00:12:24.08	And I should say that this general method
00:12:25.24	was first developed
00:12:27.28	by Ahmet Yildiz and Paul Selvin.
00:12:31.22	So, let me first describe to you
00:12:33.28	how kinesin steps along the track
00:12:36.18	for comparison with dynein.
00:12:37.25	So, kinesin always walks
00:12:39.16	in this hand-over-hand manner,
00:12:41.25	where the front motor domain...
00:12:44.24	these two motor domains are identical...
00:12:46.24	but the front one undergoes a conformational change
00:12:49.18	and that causes the displacement
00:12:52.10	of the partner head
00:12:54.02	from a rear site to a forward site.
00:12:58.00	And this is how kinesin
00:12:59.21	walks for long distances
00:13:01.12	in this kind of very regular hand-over-hand manner
00:13:06.16	where it's stepping from one tubulin subunit
00:13:08.18	to the next.
00:13:09.26	And you can even see this
00:13:11.15	if we label the two heads
00:13:14.22	with two different fluorescent dyes.
00:13:17.10	So, we marked the two heads
00:13:19.06	by a red color and a blue color
00:13:21.19	and now we plot
00:13:23.28	the position of these heads
00:13:25.07	as they're stepping along,
00:13:26.28	and you can see here, for example,
00:13:28.26	in this frame over here,
00:13:30.24	the red head is in front of the blue head,
00:13:33.00	just like this diagram,
00:13:35.03	but then the blue head
00:13:36.28	leapfrogs past the red head
00:13:38.29	and now the red head
00:13:41.04	leapfrogs past the blue head,
00:13:42.29	etc, etc,
00:13:44.17	and you can see how these two heads
00:13:46.07	are exchanging position
00:13:47.29	in a regular, alternating manner.
00:13:50.21	Now, dynein stepping
00:13:52.10	doesn't look anything like this,
00:13:54.26	so a similar experiment
00:13:57.16	of marking the two dynein heads
00:13:59.09	with two different dye colors
00:14:02.08	was done by Ahmet Yildiz
00:14:04.11	and Sam Reck-Peterson.
00:14:05.23	Both Ahmet and Sam were postdocs in the lab,
00:14:08.03	but the work that I'm showing you here
00:14:09.18	was done in their independent laboratories
00:14:14.11	at Berkeley and Harvard.
00:14:16.17	So, what you can see here
00:14:18.09	is if we look at the position
00:14:20.02	of the blue head, here,
00:14:22.10	in step number 1,
00:14:24.11	it's taken a big step forward,
00:14:27.17	but in step number 2,
00:14:29.27	instead of the partner taking the step,
00:14:32.04	that same head now
00:14:34.04	has taken yet another step
00:14:35.19	along the microtubule.
00:14:37.06	That is step number 2.
00:14:39.08	And now, finally,
00:14:40.20	the rear head, in step number 3,
00:14:42.23	begins to catch up,
00:14:44.11	but it doesn't pass the blue head.
00:14:45.26	And now in step number 4
00:14:47.16	the blue head
00:14:49.02	still takes another step forward.
00:14:51.05	So, what you can see from this
00:14:52.26	is that the dynein
00:14:54.23	is exhibiting an inchworm pattern,
00:14:57.04	where the two heads
00:14:59.01	can maintain their front and rear position
00:15:01.25	and both step forward together.
00:15:04.08	And second of all
00:15:06.19	the two heads are not necessarily
00:15:08.06	exchanging roles in timing of stepping.
00:15:11.25	Here, for example,
00:15:13.06	the blue head took two successive steps
00:15:15.01	before the red head took a step.
00:15:17.14	So, this is just a very
00:15:19.14	different kind of motility,
00:15:20.27	an irregular motility
00:15:22.21	that's not present in kinesin.
00:15:24.22	Also, dynein can take
00:15:27.03	very different sized steps as well,
00:15:29.06	so for example, here,
00:15:30.16	here's a very large step of dynein
00:15:32.17	going forward,
00:15:35.16	but these steps here are smaller,
00:15:37.10	so the step size of dynein
00:15:39.06	is not as regular as kinesin.
00:15:41.17	Furthermore, if you look at this trace,
00:15:43.17	there are many times when dynein
00:15:45.17	is actually taking a step backward
00:15:47.25	before it takes a step forward,
00:15:49.14	and these backward steps
00:15:51.08	are fairly frequent for dynein
00:15:52.27	and very rarely seen for kinesin,
00:15:55.11	especially if kinesin
00:15:57.14	is not trying to work against the load.
00:15:59.29	So, let me just review
00:16:01.14	the things that I just told you.
00:16:02.21	Kinesin has a very regular step size,
00:16:05.10	this is the distance
00:16:06.21	between subunits on the microtubule track,
00:16:09.03	dynein more variable.
00:16:10.24	Kinesin has this hand-over-hand stepping.
00:16:14.28	Dynein can exhibit this as well,
00:16:16.02	but it also exhibits
00:16:17.18	this inchworm pattern.
00:16:20.00	The two heads of kinesin take turns moving;
00:16:21.23	that is not necessarily true with dynein.
00:16:25.03	And while backwards steps are rare for kinesin,
00:16:27.22	as I showed you they're quite frequent
00:16:29.14	for the dynein molecule.
00:16:31.13	So, now I'd like to go on
00:16:33.11	and discuss:
00:16:34.17	How is it that dynein
00:16:36.12	can actually take these steps along the microtubule track?
00:16:38.20	What is the structural basis for this movement?
00:16:42.02	Well, a first big breakthrough
00:16:44.22	in this problem
00:16:46.13	came from pioneering
00:16:48.03	electron microscopy studies
00:16:49.23	by Stan Burgess,
00:16:51.25	and this shows the images
00:16:53.14	that they got of dynein
00:16:55.04	in two different nucleotide states,
00:16:56.19	and from these EMs (electron micrographs)
00:16:58.10	you can see, for example,
00:16:59.23	the ring of these AAA ATPase domains,
00:17:02.10	but you can also see a couple appendages.
00:17:04.17	One is a long stalk
00:17:06.29	that comes out of dynein
00:17:08.11	that leads at the very tip
00:17:09.28	to its microtubule binding domain,
00:17:12.00	and there's another appendage
00:17:13.17	that you can see here as well.
00:17:14.21	This is something that they termed
00:17:16.11	the linker.
00:17:17.20	It's something that sits kind of across the ring
00:17:21.04	and then extends out of the ring.
00:17:23.15	And what they noticed
00:17:24.25	in these two different nucleotide conformations
00:17:27.13	is that the position of the linker
00:17:29.27	relative to the ring and to the stalk
00:17:33.05	can change.
00:17:34.22	So, here it's sitting...
00:17:37.02	it's emerging from the ring
00:17:39.01	far from the stalk
00:17:40.23	and here they're merging close together.
00:17:43.04	And they thought that this motion of the linker
00:17:46.09	may act kind of like a lever arm
00:17:48.11	or a mechanical element
00:17:50.13	similar to the lever arm of myosin,
00:17:53.29	so what they proposed
00:17:55.18	is that the motion of the linker
00:17:58.21	relative to the ring
00:18:00.24	might be able to generate
00:18:03.05	a force upon a microtubule
00:18:05.00	that would cause it to slide,
00:18:07.21	and I'll come back to this later.
00:18:10.06	So, of course
00:18:12.26	we had to get higher resolution information of dynein
00:18:15.27	and that had to be derived from X-ray crystallography,
00:18:19.28	and it was quite a struggle
00:18:22.07	to get a crystal structure of dynein
00:18:25.06	and in fact our lab
00:18:27.00	was able to get the first crystal structure
00:18:29.09	of dynein in a nucleotide-free state in 2011,
00:18:33.27	but shortly thereafter
00:18:35.13	a whole bunch of other
00:18:37.15	nucleotide conformations of dynein
00:18:39.20	were reported.
00:18:41.20	So, the group of Kon and colleagues
00:18:44.00	from Japan
00:18:46.19	reported a very nice structure
00:18:48.00	of Dictostelium cytoplasmic dynein with ADP,
00:18:52.12	and in the last year or two
00:18:55.21	our lab got a structure of dynein
00:18:57.18	with an ATP analogue called AMPPNP,
00:19:03.00	and Andrew Carter's lab
00:19:04.11	was able to get a structure
00:19:06.13	with ADP-vanadate,
00:19:07.14	which may be mimicking an ADP-Pi state.
00:19:10.09	And what we'd like to do
00:19:12.02	is kind of similar to what you see
00:19:14.00	in this image of the horse here,
00:19:16.00	where you could see different snapshots
00:19:17.25	of the horse
00:19:19.06	taken as it's executing a gallop
00:19:21.29	and from these different snapshots
00:19:23.14	you can see the different conformations
00:19:25.12	of the horse
00:19:26.21	and begin to piece together
00:19:27.29	how this horse
00:19:29.28	is able to execute motility,
00:19:33.08	and by the same principle
00:19:34.21	we're trying to use these snapshots of dynein
00:19:36.19	to understand
00:19:38.06	how it changes its conformation
00:19:39.20	in order to execute motion.
00:19:41.19	So, now I'd like to give you
00:19:43.22	kind of a tour of what we learned
00:19:46.12	about the crystal structures,
00:19:47.27	not just from our lab but from all the crystal structures
00:19:50.05	that have emerged from the field.
00:19:52.23	First of all, here's just an image of dynein
00:19:54.24	compared to kinesin,
00:19:56.07	and you can see how much bigger dynein is
00:19:58.19	compared to kinesin
00:20:00.10	and how much more complicated
00:20:02.11	a motor domain it is.
00:20:06.24	And here's the position of the different AAA domains
00:20:10.25	that I showed you before
00:20:12.15	in this linear diagram,
00:20:13.26	but here's how they map out
00:20:15.19	on the dynein motor protein,
00:20:17.11	and they're all color coded in the same way
00:20:19.28	that you see in this linear diagram, here.
00:20:23.28	So, I'll focus on
00:20:25.29	a few important components...
00:20:27.16	so, the first is AAA1.
00:20:29.24	So, this is, again,
00:20:31.14	the main hydrolytic site.
00:20:32.17	If you make a mutation in AAA1,
00:20:34.15	you completely knock out dynein motility,
00:20:36.28	and interestingly this AAA1
00:20:39.08	is actually the region
00:20:41.08	that's farthest away from the microtubule.
00:20:44.08	Now, the other domain that I...
00:20:46.17	AAA subunit that I mentioned
00:20:48.05	that's important is AAA3,
00:20:50.00	and this is its position over here.
00:20:52.22	As I said before, it also hydrolyzes ATP
00:20:55.00	and plays an important role
00:20:56.20	in the mechanism,
00:20:58.06	and I'll explain how it works later in this talk,
00:21:02.19	but if you prevent ATP hydrolysis by AAA3,
00:21:05.23	dynein isn't completely inactive,
00:21:08.19	but the velocity of movement goes
00:21:11.06	way down with a hydrolysis mutant.
00:21:14.26	So, here's now
00:21:16.23	an atomic resolution image of the linker
00:21:18.23	that I described before as a mechanical element,
00:21:21.16	and here it's shown
00:21:23.20	extending across the ring.
00:21:26.14	Here is the microtubule binding domain
00:21:28.29	that's a small domain
00:21:30.23	that interacts with the microtubule,
00:21:32.12	and in between the ring and the microtubule binding domain
00:21:35.15	lie these two coiled-coils.
00:21:38.12	One is called the stalk,
00:21:41.06	but there's a second coiled-coil called the buttress,
00:21:44.14	which in fact extends out of the ring
00:21:46.20	and makes an important interaction with the stalk
00:21:49.24	that I'll describe in a second.
00:21:53.11	So, one of the interesting things
00:21:54.27	that we want to know from this structure
00:21:57.15	is how information,
00:21:59.10	or conformational changes,
00:22:01.06	are propagated
00:22:03.11	to control various aspects of dynein function,
00:22:06.17	and this is a particularly fascinating question for dynein
00:22:10.02	because we know that when ATP binds
00:22:13.01	at the very top of this molecule over here
00:22:15.23	it has to relay a conformational change
00:22:18.12	all the way down to the microtubule binding domain,
00:22:23.02	which in fact causes this microtubule binding domain
00:22:25.26	to release from the microtubule
00:22:27.23	so it can step forward along the track.
00:22:30.14	So, how this propagation occurs
00:22:32.25	is a fascinating question,
00:22:34.11	especially over this long distance
00:22:36.02	of about 25 nanometers.
00:22:38.11	We also know that the ATP binding
00:22:41.10	must be transmitted also
00:22:43.12	to somehow change the conformation
00:22:45.25	of where this end of the linker
00:22:47.29	is going to be positioned on the ring.
00:22:50.25	So, I'd like to now share with you
00:22:53.23	some ideas of how we think
00:22:55.07	this long range conformational change works,
00:22:58.00	based upon this collection of new X-ray structures
00:23:00.23	that were obtained.
00:23:02.20	So first of all,
00:23:04.20	let me just tell you a hint that we had
00:23:06.22	from our first X-ray crystal structure
00:23:08.26	in the nucleotide-free state,
00:23:11.10	and this just shows the AAA ring,
00:23:13.15	just focusing on the large domains.
00:23:17.09	And the one thing that you notice here
00:23:19.08	is that this ring is not symmetric,
00:23:21.01	it's a very asymmetric structure
00:23:23.08	and there are a couple gaps in this ring
00:23:25.05	where the AAA domains are farther apart.
00:23:29.24	And this gap between AAA1 and AAA2
00:23:32.26	was particularly interesting
00:23:34.23	and also surprising,
00:23:36.10	because this is the region
00:23:38.03	where ATP binds
00:23:40.07	and drives motility,
00:23:42.04	but we know from other AAA proteins,
00:23:46.16	ATPases,
00:23:48.17	that for ATP to be hydrolyzed,
00:23:51.06	these two domains, AAA1 and AAA2,
00:23:53.26	have to come closer together
00:23:56.08	because there are residues
00:23:57.29	that contribute to the hydrolysis
00:23:59.12	both from AAA2 and AAA1.
00:24:01.19	So we speculated,
00:24:03.22	although we just had one nucleotide state here,
00:24:06.01	that what may happen in dynein motility
00:24:08.15	is that in the nucleotide free state
00:24:10.09	there's a large gap,
00:24:11.18	but when ATP binds that gap closes,
00:24:14.12	and that closure then propagates
00:24:16.25	a conformational change around the ring
00:24:20.03	that gets transmitted to the microtubule binding domain
00:24:23.04	and also gets propagated to the linker
00:24:28.07	to change the linker conformational,
00:24:31.01	all, though, initiated by the binding of ATP
00:24:35.09	and the closure of this gap.
00:24:37.29	So, I'll show you that these general ideas
00:24:39.26	appear to be true,
00:24:42.09	and what you're seeing here
00:24:44.03	is a morph,
00:24:45.23	so we're going to slowly
00:24:47.25	go from one crystal structure,
00:24:49.24	which is the ADP structure
00:24:52.26	from the Kon et al lab
00:24:55.23	to another crystal structure,
00:24:58.00	which is ADP-vanadate,
00:24:59.24	which is more like an ATP state.
00:25:01.18	So, this is the conformational change
00:25:03.05	that presumably happens with ADP
00:25:06.06	is exchanged for ATP
00:25:08.07	in AAA1.
00:25:10.04	So, when ATP binds to AAA1,
00:25:14.17	you'll see a conformational change
00:25:17.07	and in this video I'm going to focus particularly
00:25:19.10	on these coiled-coils,
00:25:21.07	and how the conformational change
00:25:23.06	can be propagated from AAA1
00:25:25.09	all the way down to the microtubule.
00:25:27.16	So, here's the movie.
00:25:29.08	You can see the whole ring kind of distorting in shape,
00:25:32.22	and if you look at what happens here,
00:25:35.16	this orange coiled-coil, the buttress,
00:25:37.12	gets pulled away from the stalk,
00:25:40.21	so that creates tension on the stalk,
00:25:43.00	over here,
00:25:44.22	and that does something interesting
00:25:46.12	to the two helices that make up the stalk.
00:25:49.05	It causes a sliding motion to occur
00:25:53.03	so that the two helices
00:25:55.08	can move a short distance relative to one another,
00:25:59.08	but that sliding motion
00:26:01.12	gets propagated all the way down the coiled-coil,
00:26:04.13	all the way to the microtubule binding domain,
00:26:06.24	and causes a subtle change
00:26:08.22	in the microtubule binding domain structure
00:26:11.03	that changes its affinity for microtubules.
00:26:13.08	And in fact this kind of mechanism
00:26:15.19	was speculated many years ago
00:26:17.11	by Ian... in 2005
00:26:19.23	by Ian Gibbons and colleagues,
00:26:21.22	and now it looks like there's
00:26:23.28	good structural evidence for this
00:26:25.22	as well as other types of evidence
00:26:28.00	that has been obtained by other laboratories,
00:26:31.11	including Kon and Sutoh.
00:26:34.13	So, I now want to also
00:26:37.10	focus this same morph
00:26:39.03	between these two nucleotide states,
00:26:40.17	but with reference to the linker.
00:26:42.07	And you'll see that when ATP binds to AAA1,
00:26:46.14	you'll see the change in the AAA subunits,
00:26:49.27	and now we'll focus on what's happening in this linker,
00:26:52.29	and you can see it undergoes this large conformational change,
00:26:55.29	effectively going from a straight state
00:26:58.15	to this bent conformation.
00:27:00.23	So, earlier in this talk,
00:27:02.08	I described single molecule motility studies
00:27:04.14	that provide information
00:27:06.08	on how the dynein motor steps
00:27:08.00	along the microtubule track
00:27:09.12	and then I described X-ray crystallography
00:27:11.08	and EM studies
00:27:12.25	that provide information
00:27:14.18	on conformational changes
00:27:16.09	that occur in the dynein motor domain,
00:27:18.03	and now what I'd like to do
00:27:19.22	is to synthesize both pieces of information together
00:27:23.05	into a model that describes how dynein
00:27:26.00	is able to move along a microtubule.
00:27:28.28	And this model is presented
00:27:30.20	in the form of an animal
00:27:32.20	that's made by Graham Johnson.
00:27:34.22	Many parts of this animated model
00:27:37.18	are speculative at the present time
00:27:39.22	and no doubt,
00:27:41.20	as we get more information on dynein,
00:27:43.14	this model will change over the years.
00:27:47.06	But for right now it's useful
00:27:49.05	as a way of synthesizing data that's been gathered
00:27:52.26	by many different laboratories on dynein,
00:27:55.16	and also to generate models
00:27:57.13	for dynein motility
00:27:59.04	that can be tested in the future
00:28:00.27	by experimentation.
00:28:03.03	So first of all, let me show you
00:28:04.10	what you're going to see in this movie.
00:28:06.08	This image that you see here
00:28:08.04	of the dynein dimer
00:28:09.12	is derived from X-ray crystallographic data.
00:28:13.29	However, we don't know very much
00:28:15.23	about how the two dynein motor domains
00:28:17.21	are connected to one another
00:28:19.25	or how they're attached
00:28:21.17	onto a membrane cargo, for example.
00:28:23.28	So this part of the dynein molecule
00:28:26.17	is more stylistic and simple
00:28:28.07	because we simply don't have that structural information
00:28:30.19	right now.
00:28:32.12	Now, when I start playing this movie,
00:28:33.27	you can see the dynein jiggling back and forth.
00:28:36.18	This jiggling is due to
00:28:39.13	Brownian motion,
00:28:41.04	which is driven... thermally driven
00:28:43.16	collisions of water molecules with the dynein,
00:28:46.03	in fact this Brownian motion
00:28:47.22	is probably much more vigorous
00:28:49.23	than shown here in this animation.
00:28:51.27	Here are the different parts of the dynein.
00:28:53.19	Here's the ATPase ring,
00:28:55.14	the stalk that connects the ring
00:28:57.10	to the microtubule binding domain.
00:28:59.01	Here, colored in dark blue,
00:29:01.00	this is the strong binding state of dynein.
00:29:03.05	You'll see it transition
00:29:05.07	to a light blue color
00:29:07.23	when it undergoes a transition
00:29:09.18	to a weak binding conformation.
00:29:12.02	You'll see conformational changes
00:29:13.29	occurring in the linker
00:29:15.16	that I already described
00:29:17.04	and that transition will be shown
00:29:18.24	from a change in color
00:29:21.10	from this yellow state to a red state.
00:29:26.22	And when the microtubule binding domain
00:29:28.18	is detached
00:29:30.00	you'll see it also jiggling,
00:29:31.22	kind of moving randomly back and forth
00:29:33.10	along the microtubule.
00:29:34.26	That again is due to Brownian motion
00:29:38.07	and it probably helps this microtubule binding domain
00:29:41.06	execute a search for new binding sites
00:29:43.21	along the microtubule lattice.
00:29:46.09	So now, let's start this movie
00:29:48.19	and watch how dynein steps
00:29:50.21	along the microtubule.
00:29:52.05	And I'll show you the first step
00:29:53.20	and then we'll analyze it in greater detail
00:29:55.23	in the second step.
00:29:57.15	So, here this leading head
00:29:59.14	takes a step forward,
00:30:01.06	it's jiggling around
00:30:02.18	and now it redocks onto a microtubule binding site.
00:30:05.24	Now you'll see the rear head take a step.
00:30:08.05	It took a step forward
00:30:09.20	and you can see this linker
00:30:11.11	undergo a conformational change
00:30:13.06	from this yellow state to this red state,
00:30:16.22	and this conformational change
00:30:19.22	is accompanied, we think,
00:30:22.23	potentially, by a rotation of the ring,
00:30:26.11	and this rotation of the ring
00:30:28.19	can change the angle of the stalk,
00:30:31.12	pointing it and the microtubule binding domain
00:30:35.00	forward on the microtubule track,
00:30:37.12	which then allows this microtubule binding domain
00:30:39.26	to reattach to a tubulin subunit
00:30:42.27	farther towards the minus end of the microtubule.
00:30:45.25	So that's what you'll see in this next step.
00:30:48.07	It's going to redock,
00:30:50.05	right there,
00:30:52.12	and once it rebinds
00:30:54.19	that is accompanied by, we believe,
00:30:56.23	hydrolysis of ATP
00:30:58.13	and the release of phosphate from AAA1,
00:31:01.21	and that release of phosphate
00:31:03.24	causes this conformational change,
00:31:05.19	again, of the linker
00:31:07.17	from this bent red state
00:31:10.08	to this straighter yellow state,
00:31:12.06	and this conformational change,
00:31:13.29	we also think,
00:31:15.08	may produce a tug on the cargo
00:31:17.22	that advances the cargo forward along the track.
00:31:20.12	So, now let's see
00:31:23.02	these conformational changes again,
00:31:25.17	in this next sequence,
00:31:27.03	and you'll also see the different types of dynein
00:31:29.14	stepping in this next part of the movie.
00:31:32.23	So here, the leading head steps forward,
00:31:35.18	again, takes a big step forward.
00:31:37.27	It redocks,
00:31:39.13	but now it actually takes a step backward
00:31:42.12	along the microtubule track.
00:31:43.25	Here's the rear head,
00:31:45.23	it actually, by Brownian motion,
00:31:47.13	scoots around the other head
00:31:49.04	in this hand-over-hand motion.
00:31:51.06	It now takes another step forward
00:31:53.17	along the microtubule track,
00:31:55.04	and now its partner head
00:31:57.08	again undergoes a conformational change
00:32:00.15	and takes a step forward
00:32:02.20	along the track.
00:32:04.08	And we think by this
00:32:06.29	kind of process
00:32:08.07	the dynein molecule is able
00:32:10.23	to progressively move along on the track,
00:32:12.19	and now let's have another look
00:32:16.04	at this video
00:32:17.28	and see all these steps in action one more time.
00:33:24.19	Now let me come at the end
00:33:26.13	to this other AAA domain,
00:33:29.13	AAA3,
00:33:30.29	and let me tell you how we think that works.
00:33:32.20	As I said,
00:33:34.04	this also plays an important role in motility,
00:33:37.10	and in particular we know
00:33:38.23	if we block ATP hydrolysis,
00:33:40.17	the motor stops working.
00:33:43.06	And the mystery was why that was true,
00:33:46.07	because we know that
00:33:49.13	hydrolysis in AAA1
00:33:51.15	is sufficient
00:33:53.09	to do all the conformational changes
00:33:54.27	of the linker
00:33:56.13	and for dynein to take a step forward,
00:33:58.18	so the reason why AAA3
00:34:01.23	seemed to be important
00:34:03.08	wasn't that clear.
00:34:04.20	But an answer to this
00:34:06.11	came from structural studies
00:34:09.04	from our lab,
00:34:11.00	Gira Bhabha and Hui-Chun Cheng,
00:34:12.29	where they looked at
00:34:15.21	the conformation and the conformational changes of dynein,
00:34:19.09	not so much when there are different nucleotides
00:34:22.17	in AAA1,
00:34:24.07	but in two different nucleotide states
00:34:26.08	in AAA3.
00:34:27.23	So, in particular,
00:34:29.18	comparing when ADP in bound in AAA3
00:34:32.04	versus when ATP is bound in AAA3.
00:34:35.22	And I'll show you,
00:34:37.21	when AAA3 is in these different nucleotide states,
00:34:40.25	what happens to the conformational change
00:34:44.14	that occurs when ATP binds to AAA1.
00:34:46.29	So, the first is the movie you just saw.
00:34:49.28	That is the conformational change
00:34:52.27	that I showed you
00:34:54.29	where the linker undergoes
00:34:56.20	this large conformational change
00:34:58.08	and the whole ring changes its structure.
00:35:00.24	But, if we now
00:35:03.22	load AAA1 with ATP,
00:35:05.29	but now there's ATP in AAA3 as well,
00:35:09.24	what you'll see is a very different picture.
00:35:13.21	The conformation of this side of the ring
00:35:16.13	changes,
00:35:17.26	you can see a dramatic conformational change there,
00:35:19.20	but the conformational change
00:35:21.22	stops at about AAA4
00:35:23.29	and doesn't get propagated
00:35:25.20	around the rest of the ring,
00:35:27.08	and never causes a conformational change
00:35:29.04	in the linker
00:35:31.00	or in the stalk domain.
00:35:32.24	So AAA3
00:35:35.01	is in effect blocking the conformational change
00:35:37.25	and preventing it from propagating
00:35:40.02	throughout the ring.
00:35:41.13	So the way I like to think about this
00:35:43.11	is that AAA3
00:35:45.20	seems to be like a gate
00:35:47.15	that controls the propagation
00:35:49.08	of conformational change
00:35:51.11	throughout the dynein ring.
00:35:53.03	AAA1 is the trigger,
00:35:55.03	so in this image of dominos here,
00:35:57.26	it's what initially kicks off the chain reaction
00:36:01.08	that moves from one AAA domain
00:36:03.11	to the next,
00:36:04.25	and eventually can move all the way down
00:36:06.26	through the ring to AAA6
00:36:09.00	and cause this massive conformational change.
00:36:12.18	But if AAA3
00:36:14.22	has ATP in this site,
00:36:17.12	it actually acts
00:36:19.23	to block the propagation.
00:36:21.18	It's almost as if I have
00:36:23.29	a finger holding this domino down
00:36:26.15	and preventing the propagation
00:36:28.19	from going any further.
00:36:30.24	So the blocking of the conformational change,
00:36:33.17	or the release to allow it,
00:36:35.29	seems to be the primary activity
00:36:38.25	of AAA3.
00:36:41.09	So, that gives an update
00:36:43.04	of what we've learned about dynein
00:36:45.04	in the last few years,
00:36:46.28	but I must say we're still
00:36:48.19	very much at the beginning
00:36:50.04	and there are a tremendous number of unknown questions.
00:36:52.20	So, I illustrated some atomic structures
00:36:55.25	and conformational changes that occur,
00:36:58.22	but we don't really know
00:37:01.10	how those structural changes relate
00:37:03.12	to the stepping of dynein on the microtubule.
00:37:06.22	What would be particularly nice
00:37:08.08	is instead of just getting static images of dynein,
00:37:11.10	we can actually monitor and measure
00:37:14.09	dynein structural changes
00:37:16.01	while it's in the act of motility.
00:37:17.25	And there are ways of doing this,
00:37:19.28	for example techniques such as single molecule FRET,
00:37:22.21	which act as probes
00:37:24.17	to measure certain conformational changes
00:37:26.13	that occur in a protein,
00:37:28.02	and perhaps those kind of techniques
00:37:29.29	can be applied to dynein
00:37:31.17	so we can actually see
00:37:33.08	dynein stepping
00:37:35.03	and simultaneously measure
00:37:36.27	conformational changes.
00:37:38.06	I also gave you structural information
00:37:40.10	on the role of AAA3,
00:37:43.00	showing that it can block
00:37:44.26	a conformational change of dynein
00:37:47.12	and thereby prevent its motility.
00:37:50.06	But we don't really understand
00:37:52.00	how and why
00:37:54.10	AAA3 does this.
00:37:56.21	How does the cell
00:37:58.20	use this control mechanism
00:38:00.03	to regulate dynein motility?
00:38:01.21	How does it actually control
00:38:03.22	whether AAA3
00:38:05.13	has an ATP or an ADP in the active site?
00:38:08.17	So, we have no idea
00:38:10.20	on this issue right now,
00:38:13.28	and this is obviously
00:38:16.04	going to be important for understanding
00:38:17.23	what the real purpose of AAA3 is
00:38:20.22	in dynein cell biology.
00:38:24.00	So, with that,
00:38:25.16	I'd like to thank the many people
00:38:26.27	that contributed to this work.
00:38:28.10	First of all,
00:38:29.20	people that were in the lab previously,
00:38:32.25	a fantastic group of individuals
00:38:35.19	that helped launch the dynein project
00:38:37.18	in the lab
00:38:39.05	-- Sam, Ahmet, Andrew, and Arne --
00:38:43.16	now have all gone off to their own labs
00:38:45.11	and are very successful,
00:38:47.17	and I've discussed a lot of their work
00:38:49.17	from their independent labs in this talk.
00:38:51.09	And Carol Cho was a graduate student
00:38:54.06	who has now gone on to Korea.
00:38:55.28	And the more recent work
00:38:58.02	is the work of Gira Bhabha
00:39:00.24	and Hui-Chun Cheng.
00:39:02.28	Gira is still in the lab
00:39:04.11	and Hui-Chun has moved
00:39:06.12	to her own lab in Taiwan.
00:39:07.29	And with that I'd like to thank you for your attention,
00:39:11.18	and in my third iBiology talk
00:39:13.20	I will discuss the regulation
00:39:16.03	of mammalian cytoplasmic dynein.