 So, welcome everyone to the 2014 Steambok Lectures. And before I introduce the speaker, I just wanted to introduce Harry Steambok, whom these lectures are named after. So Harry was a member of the Biochemistry Department, and one of his most notable discoveries was the finding that you could use UV light to irradiate food, to enrich the food in vitamin D. And so this led to the basically cure of a lot of vitamin D-based deficiencies, diseases, including rickets. And as you can see here in this picture, this is Harry carrying out an experiment irradiating mice with a UV light. And this is actually captured on a video that you can find on YouTube. So I encourage you to look up this YouTube video, and you can see Harry carrying out one of his seminal experiments. So vitamin D is referred to as the sunshine vitamin, and this led to Harry getting the nickname the sun trapper. So our Steambok lecturer today is Ron Vale, who's known for his work on molecular motors. So you can think of Ron then being as the motor trapper, who's trapped molecular motors. Excuse me. So Ron has a long-standing interest in cell motility and molecular motors. And this dates back to his time as a graduate student, when he carried out a really seminal series of experiments where he identified the molecular motor kinesin, which locks along microtubules. And shown in the bottom left-hand corner is an image from one of the five cell papers he published in 1985, describing this series of experiments. This work was carried out in collaboration with three other researchers, Mike Sheetz, Tom Reese, and Bruce Schnopp. And the work that they did characterizing and identifying kinesin led to the opening of this whole new field of study of molecular motors. And so since this initial work, a number of other different kinesins had been identified. This is now a huge super family. There are 14 sub-families. And in humans, there are 45 different kinesins that carry out a number of different essential roles within cells. So Ron has won a number of different awards, including most recently the Lasker Award in 2012. He's a member of the Cellular Molecular Pharmacology Group at UCSF, where he's been a professor since 1986. He's an investigator with the Howard Hughes Medical Institute. He's been elected to the National Academy of Sciences. And in addition to all of his scientific achievements, he's also been very involved in science outreach. So he's one of the founders of iBiology, which produces the iBio seminars. So he's very involved not only in research, but also promoting research and science research. So today, Ron is going to talk about his work on motor proteins, and tomorrow he'll talk about the work that his lab has recently done with T-cell receptor signaling, and this has led to the discovery of a very elegant mechanism by which the change in energy that's created when the membrane deforms leads to the activation of the receptor. So this is a really nice mechanism in and of itself, in addition to a new finding about how T-cell receptor signaling occurs. And so with that introduction, I'll now turn the stage over to Ron. Wonderful. Thank you, Jill, for the very nice invitation. I appreciate it. It's a great honor to give the Steenbach Lectures. There are a lot of great faculty here, so I'm having a fun time visiting. And I'm also very excited to come here, because this is probably the only time in my life when I'm going to experience a minus 40 to minus 50 windshield factor. So I'll see what happens with that this evening. Yes, so this lecture today will be on our work with molecular motors, and tomorrow's will be on the T-cell. I mean, they both have a similar theme of, you know, I would say using biochemistry and reconstitution to dissect cellular processes. So it's the one theme in common. I did want to also just start off just by maybe telling you about advertising this other project, which is also, I would say, a great passion of mine, which is a project which has now been called iBiology. Actually, it's been growing over years, but it started maybe in 2006 when I just realized that the way we communicate our work through all communication, like in the form of the seminar, that, you know, there's some places like Madison and San Francisco and so forth where you do have great seminar series, but many other places in the U.S. and around the world really don't have access to seminar speakers and don't have an opportunity to hear the work from leading scientists. So over time that gave rise to this project where we filmed scientists, such as, this is actually a recent one we did with Randy Schekman, who's actually giving a talk here in a new series that we have, which is directed for high school students in beginning college, and really the mission is just to make scientific communication, in terms of oral communication, available to anyone around the world, and we film these speakers in a studio in a green screen, so similar the way the weathermen is being recorded, and they are available on this site, which is called iBiology, which I encourage you to go to. There are about 100 seminars, which are similar in flavor to a seminar like this, although they have a much longer separate introduction, which is geared for non-specialists and students. I think one of the difficulties in giving a seminar like this is we, you know, we have to get to our data. We don't have enough time often to give a really proper introduction, but all these seminars have a really great extensive introduction to them. The iBio Magazine series are 15-minute talks. They're a little bit like what I consider the TED talks of our profession, so they'll describe maybe less than 15 minutes showing how a famous Nobel Laureate made their kind of behind-the-scenes that went into their discovery. They'll be talks on education, talks on careers, talks on what scientists do outside of the lab. It's a whole collection of very interesting talks, and a new product, which we just have come up with, which is called iBio Education, which was launched with our new website, is geared to materials so that educators can use them, potentially if you're giving talks to undergraduates here, these can be really useful supplements to teaching. So we are very eager to actually work with, well, graduate educators too, but undergraduate educators who want to use this material, so I would encourage you to have a look and also feel free to email us and we're happy to work with you. We also have, we're starting to produce courses. We have actually a complete course on light microscopy, an advanced course with about 60 videos on it. It's a really amazing resource. It's done by kind of the leaders of microscopy from all over the world and a beginning course, which is about 13 lectures, more for newcomers to microscopy, for example, an undergraduate. So they're all on, this is what the, I just took this off of the website right now. This is what the site looks like, so you can navigate in these seminars, these short videos and these educational resources, including the microscopy course. Also just a little bit of another advertisement. We are also trying to make this site more interactive so that you just, you know, in addition to seeing the videos, we're also starting interactive formats with scientists through Google Hangout, which is very easy to do. You know, you just get on the internet, pop open your computer. The links are all present here and on February 7th, you can listen to Bruce Alberts talk about, you know, future challenges of science. So we're having one of these about every month. The next one will be with Keith Yamamoto. So again, I think also for those of you who are, you know, graduate students, post docs, undergraduates, this is a really great venue. They've been quite lively discussions, so I encourage you to have a look at that too. And I'm happy to talk to anyone about this project afterwards or over the next couple of days. And Judith has been a big supporter of us at ASCV, so thank you, Judith, who's been helping us with international outreach for this project. Just another little thing that may be of interest, which is another lab project, the open source project, is software that we develop, mainly very talented person in my group, Nico Sturmon. This is open source software for microscopy. So unlike commercial software, which is very expensive, not open source, doesn't run all the equipment that you necessarily need. This software here will control your cameras, your robotics, everything associated with your microscope. It's completely free. It has a big user community already. And anyway, free, you can't argue with the price tag of that. And anyway, we're also happy to help you set this up in your lab as well. Okay, so now on to the science. I think one of the really fun things that I love working in this field, which is biological motion, is that everyone's familiar with motion as a very inherent property of life. Of course, these are obvious muscle movements here. And then if you look at any pond water, even as a child, you can see all kinds of wonderful creatures swimming around in there. So biology has machines that create this type of movement. And these are some of the well-known machines and how they work. This is muscle myosin. If Ivan Raymond is here, a lot of this is really based upon his work on crystal structure of myosin and the lever arm motion and kinesin working, moving along microtubule tracks here. So what this lecture is going to be about is really focused on very recent work on dining. All the work I'm going to be, this is actually the first time I'm going to be talking about all the work in this seminar. So fairly new, unpublished. And we actually maybe switched more from kinesin to dining about a decade ago because very little is understood about how this motor works. And I'll tell you about that later. But I want to start off to tell you a little bit about what we did learn from kinesin and myosin because I think it's at least our aspiration of where we want to see dining several years from now. And also might help to introduce to you what molecular motors do. So molecular motors are machines that take chemical energy. In this case, from the chemical energy source ATP and their enzyme. So they go through and do a number of chemical steps on this energy source, which is first hydrolyze the gamma phosphate bond and then release the phosphate and then release the ADP and then rebind ATP again. So they go through this cycle, which is called the ATP hydrolysis cycle. And during one of these cycles, the motor produces some kind of motion along a track and force associated with that motion. And these machines do this at much greater efficiency than your automobile. So in many cases, these machines work at extracting about 50 to 90% of the theoretical chemical energy and converting that into work. And this just shows from the good old kinesin days, an example of kinesin bound to a plastic bead undergoing these cycles and moving these plastic beads along a microtubule track here. Now, again, as Jill referred to, it may have been quite an amazing journey because we started in 1985 just observing motion really in axons. And these are little vesicles, and that's a mitochondria over there moving along these microtubule tracks. Really not very clear what was underneath these organelles and what they might be doing to produce this magical motion. But since that time, just through the amazing tools that have come out all around us in biology, everything from better structural approaches, single molecule, genomics, et cetera, kinesin world and with mice. And I think we now have a pretty good idea what these machines look like and what kind of structural changes happen in this protein to produce the motion. So I'll get to in the next slide. Every time there's a chemical change in the motion that's accompanied by a structural change, in this case of kinesin that leads to this little region called the neck linker zippering up along the main motor domain and positioning its partner head to a forward binding site. So inherent in driving that model, which is obviously an animation, what one really is trying to do to understand how these machines work ultimately is, first of all, to have a fairly sophisticated understanding of their structure, either through X-ray or EM, but the challenge then is not just to get one static structure, but to piece together different states of that protein, particularly what those states are doing during this ATPase cycle to drive a model for how this protein could create motion. So I think this is kind of an apt analogy. I mean, this is similar to what people were debating at the end of the 19th century of how a horse actually runs, and it was unknown even whether it had all of its legs on the ground at one time or whether how the horse could actually gallop, and that was done through high-speed photography where they could capture different states of that horse and then derive, if you'd like, a pretty good understanding of the cycle of how these legs are moving to propel the horse. And in some ways, that's kind of what we're trying to do with these motors is also to derive a number of different states and piece together how this chemical cycle gets converted into protein-structural changes that lead to motion. So again, I'll say a few words about Kinesin and Myosin before transiting to dining, but in this case, and there I see Ivan in the audience here, but we actually were helped considerably by his earlier work on Myosin because we had a real surprise, and I would say in this case a helpful surprise, that when we got the crystal structure for Kinesin with Robert Flederich, what we found is even though Kinesin is a microtubule motor and Myosin is an actin motor and they're completely different sizes, they actually appear to have evolved from a common ancestor, and what you see in blue is a common and probably very ancient structural core for both of these motors, that actually overlaps very nicely if you superimpose them. And if I have to summarize how these motors work in a nutshell, which is obviously leaving out a lot of details, is what these proteins have evolved and actually still share in common as a similar common strategy for movement is they have this core structure featured in blue, which is performing the chemistry. Performing the chemistry and also acting as a switch to trigger structural changes in the protein. So this is a core that will hydrolyze the ATP and it has some regions of that protein that change state very subtly depending upon whether there's an ATP in the binding site or whether there's an ADP in the binding site. And in fact, those sensors if you like are quite similar between kinesin and myosin. So the common chemistry and sensing of the nucleotide is very similar. And then that has to be that information that's happening right around the nucleotide has to be communicated to other parts of the protein. Again, amazingly, kinesin and myosin use a very similar strategy of a helix here, which is either called a switch to or relay helix that extends from one end where the nucleotide is to another end, which is where a mechanical element is. In the case of kinesin and myosin, there are very different mechanical elements that the two have evolved. But the helix is a common mechanism for communicating information. When something happens here at the nucleotide, this mechanical element also changes. Yeah, so that's our green helix over there. These are our mechanical elements. And these are the state changes that happen in these mechanical elements. Again, work from Ivan and well as many other people have defined this lever arm movement of myosin. And kinesin also undergoes a state change of this lever arm. So you have this amplification mechanism and converts very small changes that are happening in the nucleotide binding site into much bigger scale motions of another domain of the protein. And that's what is causing this state change of this little element here called the necklinker of kinesin, where it zippers up and propels the partner head forward. And of course, this is an animation, but you can actually put fluorescent probes on the motor domain to measure the state of the necklinker. So we think it's not completely made up. But this is actually a kinesin that's walking along a track. And I'm not going to go into the details of that experiment. But this is a FRED experiment. So when you see it being bright over here, this necklinker is in this red confirmation, but then it switches to yellow, where it's in this yellow confirmation. And you can see it altering back and forth as this motor is walking along the track such as here. Okay, the other thing you should also know about motors is that they're not on all the time. And that kind of makes sense. You don't want the motor to be churning up energy all the time. You want it to be harnessing energy when it's doing something productive, like moving a cargo. So for example, kinesin, there's a lot of kinesin in the cytoplasm in cells. Thank goodness, otherwise I never would have had a chance to purify it. In fact, actually, the real story is, you know, we and probably everyone else in the field thought if there was a motor, it was going to be either bound to the organelle, or it might be bound to the microtubule, but no one was expecting to be floating around the cytoplasm. But there's a ton of it in the cytoplasm. And most of it's in this inactive folded state. Cargo comes along and it activates it. And so each motor has some kind of mechanism that can switch its state from one to another. OK, so that's my intro here. Now let me get on to the dyne molecule and tell you something about it. So dyne has lagged behind kinesin and myosin. And maybe this slide explains why. It's very, very large. It's about one and a half megadaltons in terms of the complete holoenzyme. So it's very complex system to study and dissect what everything's doing. But it contains this ring shaped motor domain that will describe more detail. And then it has what's called a tail domain with many associated subunits that dock dyne into different kinds of cargos in the cell. And so dyne is very important. It's what causes your cilia to beat. For example in your bronchial tracts. Or it will cause sperm to swim or cilia in a paramecium to move. It has many roles in just interface cells. It's the major motor protein that moves to the minus end of microtubules. So that's mostly in most cells transport from the cell periphery to the center. And kinesins move in the opposite direction. They move, it moves organelles, protein complexes, mRNAs, just a tremendous amount of different cargos. And it's also involved in building the mitotic spindle and also involved in activities at the kinetic core as well. So pretty important class of motor protein as Jill would agree too. So the two things I'd like to talk to you about today are first of all how does this motor domain work? And second, last part I'll talk about how this motor protein might be regulated. So we began studying maybe around 2000. And dyne was a graveyard for the first couple people that attempted to work on this in my laboratory. But then I was very fortunate to have Sam Rek Peterson come along to set up Ceravisia as a model system where we can modify the gene by homologous recombination. So begin to put probes on it or manipulate the dyne recombinantly, which was critical for all the success we had with kinesin. She also got in vitro motility assays working from this. And shortly after a few other post docs came to the lab and that's a single molecule assay for dyne. Andrew Carter, very talented protein chemist, worked out a whole bunch of protein biochemistry of dyne, including the crystallization that I'll describe to you. Ahmed Yaldiz was a physicist and got high resolution tracking information about how dyne steps along the track. And Arna Generich developed an understanding of how dyne produces force and responds to different force. The reason I want to show this slide is it's also I think a good example, I think for students or post docs. All these folks had to come to the lab and we couldn't get anything to work with dyne in the beginning. So they all had to work together to get these various assays to work. And it was really critical that they work together in the beginning. And sometimes people are worried, oh, if I work with someone else, will I ever get a job? So anyway, they did okay in the end. And they're all great scientists and now starting their own labs. So you could also, someone may say, well, of course, you worked on the kinesin mechanism for so long and dyne is just another, it's another microtubule motor, probably worked like kinesin. Wrong. Because even though they work on a microtubule, dyne really has nothing to do with kinesin, evolutionarily or, and there are probably going to be tons of differences in the mechanism. This is just a gross oversimplification, so don't take any of these distances very seriously, just made up in keynote. But as I told you about mice and kinesin evolved from a common ancestor, they're related to small GTPases as well. But dyne has evolved from a, maybe originally from something that had a common motif in many nucleotideases called a pilu, but it has emerged from a whole other group of enzymes that are called AAA ATPases. And in reality, dyne is kind of the weird uncle in this protein family. Because most of them are actually unfoldases. They unfold proteins for degradation, so this is a system in bacteria, this is a AAA protein called ClpA or ClpX, that takes proteins, unfolds them, and stuffs them into a degradation chamber. The top of the proteasome is also made up of AAA ATPases. Jim Rothman also worked on a factor called NSF, which is a AAA ATPase that breaks apart snare complexes. So many of these proteins use ATP energy to produce work, but not to walk along a track. So dyne is the one unusual group in this family that has evolved an ability to walk along a cytoskeletal track. But probably, you know, we learned a lot by comparing kinesin and myosin. I think we're going to learn more about dyne by comparing it to unfoldases than we will by comparing it to kinesin, for example. So let me give you now a little bit of primer on what these AAA ATPases are and what they do. They all have this common structure of two domains. A large domain, which is a beta sheet and surrounding alpha helices and a small helical subdomain. The majority of them are expressed as a single gene product that self-assembles into a hexameric oligomer. So the majority of these are hexamers, but they originate by self-assembly. They also depend upon one another. So the neighbors depend upon one another for hydrolyzing ATP. So for example there is a critical arginine, which is called the arginine finger, from a neighboring subunit that is part of the active site of the ATPase. So that's why you need these oligomers to get efficient ATP hydrolysis because they depend upon their neighbor. So dyne is also the weird uncle in the sense that it has concatenated a hexamer of AAA domains in one polypeptide into one of the biggest polypeptides in the genome. As a result of that, unlike most AAA ATPases which have the same subunit around the ring, each of these AAA domains in dyne is different. They evolve different sequences, so none of these are identical to one another. Four of them, what's called AAA 1, 2, 3, and 4, which has a big insertion into it, which binds to the microtubule that I'll show you, these all bind ATP. The last two do not. They've lost that ability, but they're probably needed to structurally complete the ring. So dyne also has two big features that interact with the ring, which I'll talk about. One is the stock domain and the other is something that's called the linker, which I'll explain later. Now, this also puts in perspective why dyne has not been easy to study because it is so large. This is just the motor domain, and this puts it in relative perspective to kinesin, for example. So it's a real behemoth of a motor, but it also, just from looking at this structure, you could see how fascinating this motor protein is. This is the main nucleotide hydrolysis site called AAA 1, and for a motor to walk, just like I'm walking, I have to get my feet off of the track, so I have to communicate from my brain to my feet to sometimes plant, sometimes release my foot, and the same thing has to happen with the motor. So sometimes they have to be tightly bound so they can produce force, but also this microtubule binding part has to release so it can walk along the track, and that has to be coordinated. So events that are happening at this ATPase site have to be transmitted through this ring all the way down to change the affinity of the microtubule binding site so it can release. So this is about, you know, 35 nanometers or so, so it's a really fascinating question of long distance transport. We actually think we have some idea what's happening with this, I'm not going to talk about that today, but happen to answer any questions. So anyway, fascinating problem of allosteric communication and dining. The other thing that's very interesting about dining, unlike kinesin or myosin, is has several ATP binding sites, and this just shows data from Khan and Suto a number of years ago where they prevented hydrolysis for each of these binding sites, and you can see that this triple A1 is absolutely essential. You block hydrolysis, the motor doesn't move at all, and there's a second site here triple A3 where if you knock it down it also is a severe, very severe defect. These two are less important. In fact, actually we know now that triple A2 doesn't even hydrolyze ATP. So this is going to be a subject of this talk which is what these different roles of these ATPases are in dining, and I don't think we have the entire answer, but I think we made possible some insights into that question. Now the last thing about dining is how does it produce mechanical force? Well here's the ongoing model for that, which is very nice work done by Stan Burgess, Anthony Roberts, Suto collaborators, but in addition to this ring here there's this really big appendage. This is actually a cryoem image over here that's been stylized, but there's a appendage that sits on top of the ring called the linker, and just like you saw the myosin lever arm rotate, this linker is thought to become undocked from and kind of swing out, and when it goes from this state back to this state it's thought to pull the cargo with it. So that seems to be a plausible, although I would say not completely proven, but a pretty reasonable model that I think we would also agree with. So a big challenge really with dining was to get it to the modern age where we can now think of it as a real structure rather than a Macintosh drawing, so that means getting a crystal structure for the motor. So that was done by a very, two very talented people, Andrew Carter, who I mentioned now is at LMB, and Carol Cho is a graduate student, who worked together and crystallized a actually a GST dimer of the yeast motor domain, originally a lower resolution and pushed later by Andrew Carter to higher resolution where the side chains were visible, but even from this first crystal structure we could really understand the domain organization of how all of these different domains worked and interacted in the dining motor. And one of the big surprises from that first crystal structure is that the ring was not a symmetrical ring, like a lot of AAA ATP aces, and in fact it was highly asymmetrical, and there were if we just looked at what I called the large domains, which is the main ATP binding site, if we just strip those out you could see how it's discontinuous. There's a big gap here between AAA1 and 2, and also a gap between 5 and 6. Now this was really surprising because I mentioned that the neighbor has to provide a residue that tickles the other one to hydrolyze ATP, and we knew this is the major hydrolytic site of dining, and there's no way in this crystal structure this is going to hydrolyze ATP. So we reasoned or made a model that what has to happen in the next phase, somewhere in the next phase of the cycle, is that this gap has to close to bring these two domains together, and that might induce a set of conformational changes that would produce that conformational change of the neck linker. So that was something that we got out of the first crystal structure. I should then say then work from Khan and colleagues in Japan got a very nice high-resolution structure of the dictastelium dining motor domain in an 80p pound state. The last data I showed you was a nucleotide-free. So we now had two different crystal states of 80p and nucleotide-free. A little problematic because we're also from two very different species, but it was pretty clear you know we had to keep going, we had to really get more snapshots to try to understand what dining is doing in its cycle. So again a very fortunate, again to have people with the right skills and personalities who are willing to work together, so Ghirababa and Wei-Chun have been working very closely together to try to understand new structural states of dining. They've been great partners and Ghir is focused more on E.M. and Wei-Chun has worked on extra crystallography and I think you'll see how the two have been very complementary to one another. So what Wei-Chun was able to get was a crystal structure now of the dining motor domain with a non-hydrolyzable ATP analog, which is AMP-PMP, bound to the four active sites of dining. So we can see now instead of being nucleotide-free we'd see the structures for AMP-PMP in the four binding pockets. And well gratifyingly there was a structural change if we compare it to our previous crystal structure and here we're looking on the ring and a side on view and there's the linker that's coming across and you can see that there's a really massive change in the ring where upon ATP, AMP-PMP binding the whole ring shifts up one side of the ring to become much more planar than in the nucleotide-free state. Interestingly this conformational change is restricted to one side of the ring so from pretty much from triple A1 through 4. If we flip the ring over now and look on the other side we see that there's virtually no structural change on the other side of the ring. So half the ring is basically coming up like that when ATP binds. So we can now look at and try to analyze how this large conformational change is happening by looking at each of the triple A domains to see where the motions are occurring. Oh okay well anyway so this we speculated before that this gap would close and in triple A1 and indeed gratifyingly it does you could see the the blue triple light blue triple A2 move like a clamshell to close when there's AMP-PMP bound in the active site. I should say right now that this state if you look at in detail is still not competent for catalytic reaction because the arginine is not in the right spot and I'll come back to that later because it's a key point. But if we analyze all the domains separately you can now see gray is the APO structure and the colored one is our new AMP-PMP structure. You can see as I just showed you gray to light blue this clamshell motion. If we look at what's happening in triple A2 there's basically nothing going on here but if we compare we look at triple A3 you could see a clamshell motion also occurring between three and four that's this motion and nothing happening between four and five. So AMP-PMP binding is clearly having actions going on here both at triple A1 and triple A3 and I told you those are the two most important sites for motility in this enzyme. But in a way the structure was disappointing. Although not really very disappointing but we were speculating from what everyone was thinking in the field that when when ATP bound to triple A1 that we would see a large conformational change of this linker and that this ADP vanadate was supposed to be mimicking an ATP like state. But when we put in a non-hydrolyzable analog we did not see that structural change of the linker. I showed you that linker was still in the exact same position and there it is. It's crossing over the ring looking very much like this. It basically did not change state very much at all between APO and AMP-PMP. I'll come back to that why that is. So we then began by EM because we could screen through many more conditions so Gira did this and she now actually created a mutation in triple A1 where instead of putting in AMP-PMP we can now put in real bonafide ATP but we made a mutation a point mutation in the active site so it couldn't hydrolyze it. So we can get ATP in there but it wouldn't be hydrolyzed and when we did this we actually then saw this big linker swing with ATP which we didn't see with AMP-PMP as I'll show you in a second also by EM. I should say when you do single particle reconstructions also a nice thing that you don't get from X-ray is you can form class averages so you can see if there are different structures of the dynein in your population and indeed actually with this ATP state and hydrolysis mutant we see actually different confirmations of the neck linker suggesting it's not in a single state but probably quite flexible and able to probe different states along the ring. So and as I just said if we feed this thing ATP we saw that linker swing if we fed it AMP-PMP we didn't. So there was one way to interpret that that well actually this way we're thinking for a long time that AMP-PMP is just a bad analog for ATP. It's not ATP you know it's has a different chemical structure so it's just not mimicking this. Until we began to think of the other nucleotide binding site which is AAA3 which I also told you was important for movement and we thought that this scenario might be happening that when you when you give the dynein motor AMP-PMP in this case you would have AMP-PMP a non-hydrolyzable analog bound at both sites so you'd have an ATP like molecule in both sites but when you feed it ATP here yes you have the ATP analog in or real ATP in AAA1 but AAA3 can do chemistry on it and convert it to an ADP state. So maybe what we're seeing the difference here is not really due to you know real ATP versus bad ATP but the ability of AAA3 the second nucleotide binding site to get in a different nucleotide state. And the answer to that is true because Gira then could now make the mutation that blocks nucleotide hydrolysis in AAA3 and indeed this is what happens so this is just the same image I showed you before this is when you have the mutation AAA1 but AAA3 can still hydrolyze nucleotide. Now we make the double mutant where same as this but now we prevent a nucleotide hydrolysis in site 3 and this linker swing doesn't occur. So this is what I'm going to try and get through in a couple minutes because I realize I'm running out of time but this gives us I think a new view of why you have multiple nucleotide binding sites in dynein and that one of the nucleotide binding sites which is AAA3 its job is actually acting as a regulatory gate and I will say in advance that we don't know completely how that gate and how that regulation is used but we think it may be used for turning dynein on and off in important situations for a cell biological function but here's how we think the gate works basically if you load up ATP in both sides here the dynein is frozen it's mechanically inactive but it's almost like spring loaded it's ready to change its confirmation but you have to hydrolyze AAA the nucleotide at AAA3 that gate opens and now that linker swing can occur and so we're beginning to think of a dynein cycle like this if we start off with nucleotide free dining we bind ATP to these two critical sites but the linker and there's this big conformational change that's happened in one half of the ring nothing that's happened on the other side of the ring as I just showed you so we've got half of the conformational change to occur but not the other half and the next step we need that gate to open which is basically AAA3 hydrolyzing the ATP and that allows essentially the conformational change to propagate throughout the entire ring that leads to this bending of the linker and that is the pre-power stroke state and we think that then when this AAA1 hydrolyzes nucleotide releases phosphate then this linker straightens out again and that is the power stroke and then eventually it gets into another state where these two domains can open up again release the ADP and the interesting thing that we don't know the answer to is how the cycle will continue from here I bet it's not going to be necessary for AAA3 to cycle on every mechanical event I think it's going to act more like a gate and the gate's going to be open i.e. it's going to retain ADP in the active site and it's going to go right back to this state over here but if this releases and binds ATP it's going to go into this off state until AAA3 can hydrolyze so anyway and this just shows actually some of the structural changes that we think are happening I'm going to wait for the cycle again what you're going to see it is it going from the APO state to the AMPP bound state to a state from the Japanese group which we think actually is a close mimic of this state so here it is here's our AMPP and peak conformational change nothing's happening on this side of the ring but then when AAA3 hydrolyzes you can see this conformational change propagating right here around the other side of the ring which we think is a prelude to getting the mechanical swing of the motor I'll think I'll just skip that so okay so now I'd like to shift gears the last few minutes to well I told you a story of how we think the part of the nucleotide can actually regulate turning dining on and off but we also know that there are a lot of other factors in the cell whose job is to regulate dining motility and I'll tell you another story from Rick McKenney who's a very talented postdoc who I think is put together a beautiful story on how some other dining regulation can work so it's known through many years of cell biology that in most cells dining does not work alone it works with a very large protein complex called dinactin so if you make a dining knockout or a dinactin knockout you get very similar phenotypes which is loss of dining motility of some particular cargo but really the details of how dinactin works with dining still not very clear we actually did some work on that many years ago too and other groups have and they found like modest effects on how maybe how long dining can walk along a microtubule but nothing of the magnitude that I'm going to show you right now so this story began when Rick actually tried to switch our lab from yeast where we've been very comfortable and happy to making dining from other sources like human dining which he tried to make recombinantly he also tried again in vitro motility assays to work with purified native dining from brain and nothing he could do could get the dining to work very well so I showed you those movies of yeast dining moving so beautifully but he would mostly get dining molecules that would bind and just jiggle around and they wouldn't engage in really good process of motion and you can follow this on what's called a chimograph where you're looking at the position of dining on a microtubule over time so if it's in the same position you get a straight line if it's changing its position like this it's moving along the microtubule and the bottom line is most of the dinings weren't moving and if he got some to move they were moving incredibly slowly so it was really terrible really couldn't do very much with this so Rick began to you know think through that you know maybe there's an activator in the system and in addition to Dynactin is also very well known actually from genetic screens originally done by Wieschausen, Nusselen-Volhard of another protein called Big D that somehow seems to be involved in dining motility and the recent work actually by another group in the Netherlands Splinter et al actually showed that if you add Big D you could form a very stable complex biochemically between dining Dynactin and Big D so it's cemented into a very stable complex and Rick then asked well what's the motility properties of this complex and the gratifying news is dining when it was complexed with both Big D and Dynactin became ultra possessive so you know compare that to the last video that I showed you and in fact really this Dynactin complex is probably the most possessive motor ever I mean that we've ever studied in the lab or maybe more possessive than any motor currently characterized the run lengths these run lengths you can see here extend for very long distances and you can measure the run lengths and in many cases they're limited by the microtubule length but they're often like 15, 20 microns in length which is very many considering the step is 8 nanometers these are many steps so we now start getting incredible run lengths and then Rick did another experiment where he wanted to actually visualize to really make sure that last experiment I showed you what we were actually visualizing was not dining but but the adapter protein Big D so Rick did another experiment where he actually functionally labeled dining Dynactin and Big D with separate fluorophores by expressing a tag subunits in mammalian cells and also in this experiment he could separate also dining Dynactin from the Big D to make sure that the Big D is really essential for motion and joining this complex together and the bottom line is if he added dining Dynactin alone the movement was terrible so this is what he isolated from the mammalian cells the dining would stick but it wouldn't move Dynactin didn't even bind and to the same tube he adds back Dynact Big D and gets these really long process of runs and this is a three color image where I guess you're seeing dining in green, Big D in blue and Dynactin in red they're separated they're not colloquializing because the microscope takes one image changes filters takes the other image changes filters so there's a little time delay so that's why they're not colloquializing but you can see them actually moving beautifully at a single molecule level as is Big D Dynactin triple labeled complex and we can see this complex by EM which Rick did with Gira and it looks like from this EM this little filament here is a short actin filament which is part of Dynactin and you can see that it binds to the dining tail domain over here here you can see the heads there's the Dynactin heads there's the Dynactin it also seems to be very flexible so it appears that this Dynactin can move in many orientations relative to the dining and the working model which Rick is now working on now of course want to find the mechanism and I'll just tell you and so far experiments might be looking promising in this direction is that the Dynactin is providing some kind of tether that's somehow linking dining more strongly to the microtubule and assisting its movement through an attachment to the microtubule binding domain but sorry attachment directly to the microtubule in some complex that may look like this but that's still work in progress so anyway I want to thank all the I'm here giving the talk but it's these people that did all the fantastic work and I've been very fortunate you know they had a privilege of having these extremely talented people in our laboratory Gira Baba and Wade Shand did all the structural work that I told you about Rick McKinney working on the big D regulation and I also want to thank our EM collaborators who have helped Gira many stages of this project Yifan Chang and Malfoo Lau from UCSF and Arna Mola from Scripps and thank all of you for your attention all right so there's one over there but no I didn't talk about that at all yeah because it's almost like a completely separate story but well you know the bottom line is we don't entirely know the answer to that but let's see if I find an appropriate slide well maybe this this is not entirely the answer to that oh no I actually do have one so the question is how does things get communicated to the microtubule binding domain so I mean basically what you can see here is a very long coil coil and I think most of the time people think of a coil coil as a really dumb structure you know it dimerizes things but it doesn't have any intelligent life onto its own but the current model in the field is that these two helices may slide relative to one another and the design of the coil coil is kind of interesting because one of the coil one of the one of the helices coil coil usually has two hydro has a heptad repeat so seven residues one and then two of the residues of the seven are hydrophobic and in the interior of the interacting helices so that forms the core of the coil coil so one of the one of the helices has a very canonical what's called a plus d hydrophobic pattern all the way through it the second one has effectively one good hydrophobic residue in the heptad and the second one is very weak all throughout the entire length so one idea is that in fact these two helices may if there is tension placed on them could potentially slip relative to one another and if they slide by four residues that would be a way of transmitting information all the way down this long structure to change tweak something at the microtubule binding domain the evidence for that is we work with Ian Gibbons many years ago who was the discoverer of dining and with him this largely Ian's work made two different versions like froze the coil coil in these two different registries so it took the microtubule binding domain and the two helices and froze it in like registry one or registry two and the result was that the microtubule binding affinity at the end was altered the other evidence is that same group in Japan put cross linkers in registry so two cysteines that should cross link in registry one or registry two and they could cross link the dining and the result was that also changed the microtubule affinity and the ATPase properties in the intact dining now no one's actually observed this thing sliding in a real dining molecule but that's the idea but also the other interesting thing that came out of even the first crystal structure this thing is now dead but there was that long stalk or they call it the stalk that goes the microtubule binding domain but we found in addition to that there's a second coil coil and that's called the buttress over there and it cements halfway up the other coil coil through a bunch of hydrophobic interactions so without I can't tell you any details of exactly how this works but I think while these AAA proteins are moving around as I showed you they are moving around then they are going to be yanking on this buttress and stalk so if they're moving around on the base and the buttress and stalk are moving but they're connected at one fixed point where they're both interacting then you could possibly pull the helices through the movements of the ring which would get converted through some sliding of these helices that would then get converted to some probably subtle domain changes that are happening in that microtubule binding domain so it's kind of maybe this really complicated Rube Goldberg kind of string of events but anyway we think there's something happening to these helices and that this connection between the buttress and the stalk is a way of communicating things that are happening in the ring to linear motions that are happening with these helices details to be figured out yeah well the concentrations are really very very low so we're looking at single molecules I mean it's a good question but the EM shows that we see single particles so they don't look like multimers and also if we measure the fluorescent intensity it's again most consistent with one copy of Big D Dynan Dynactin and they're not forming some kind of multimers so it looks like it's working by you know a single stoichiometric complex as I said the reason in that experiment why they look shifted is that this dynein is actually moving really fast so in that experiment we have full ATP there then every time we switch filters the dynein is already moved along the microtubule yeah that's a good question we've been I'm trying to think we're going to actually answer that definitively we've had trouble actually separating dynein and dynactin with great purity of separation so even in things where we've tried to separate like native dynein to a dynein and a dynactin fraction that looks pretty clean on a gel by the single molecule experiment any small contamination of dynactin gets hooked up to Big D and dynein so we see even a minor minor population of dynactin in a dynein prep getting recruited to dynein so as of yet we haven't been able to get a completely pure dynein prep with absolutely no dynactin in it that we've added Big D to it's one of the things about the single molecule experiment you can observe like a 1% contaminant very very easily in those kinds of experiments so yeah maybe at current stage the answer is no but I strongly suspect that Big D plus dynein alone is not going to be possessive also the work very nice work by Splinter et al again showed that in their biochemistry Big D that the Big D did not interact with dynein very well unless there was dyneactin so they seem to synergize in forming this complex together but it's a good question yeah the case of the human dynein when it's not moving is it still hydrolyzing ATP and just remaining stationary or is it the ATP cycle fixed as well? yeah it looks like it looks like still hydrolyzing ATP yeah I mean one thing about dynein in general it's a very poorly regulated ATPase overall compared to myosin and kinesin so kinesin and myosin have a very high activation other ATPases by the polymers like you know 500 fold or something dynein if you look at it with or without microtubules it's fairly minor I mean usually like eightfold so it's not extremely tightly coupled ATPase looks like it's there to help generate heat in the body during cold days yeah that's a good question I mean part of the problem is it's really hard to develop when you're measuring ATPase you don't know where that ATP is coming from and it's hard to make conclusions completely from mutants like oh I made a mutant here and therefore everything I see left is here because these two sites may be talking to one another so that's something we're interested in doing in the future is making probably some fluorescent probes that could distinguish you know the chemical cycles and the different sites but right now we don't have the answer to that so I think that's a very important thing to do yeah so the question is how does the communication around the ring go well the the arginine finger is probably essential for catalysis we don't know how important it is for the conformational change mechanism but for example that big movement that you saw over there probably not dependent on the arginine finger because it's not really it looks like it's not engaged in anything you know basically I think that the what it looks like is the conformational change starts at AAA1 that's where the biggest movement occurs it kind of uh it brings along with it many of the adjacent subunits and then interestingly the linker spans over the ring so it starts at AAA1 and it docks at AAA5 and in our AMP-PMP structure it's actually making physical contact with the ring on the other side and I think that's important for blocking because you've got now cemented like the ring at one point and you cemented the ring on the other point and that is probably what's blocking the conformational change from propagating around subunits on the far side of the ring so I think what has to happen is that linker has to pop off and then you can get that second phase of movements that I showed you in that morphed diagram so I think what's happening is a AAA3 hydrolysis somehow is popping the linker off of the ring and that allows the dominoes to keep going so I kind of use this you know in a by analogy like imagine you have a string of dominoes right and you start the first domino going and it goes click click click and knocks down you know one two and three but you're holding your finger on the third domino and that's what I think the ATP and AAA3 is doing it's kind of holding the finger down on that domino once that hydrolyzes to phosphate, phosphate comes out it's like that finger releases and the the dominoes can keep propagating around the ring so that's kind of a visual graphic description of how I think you know that conformational change is going and what the role AAA3 is doing and part of that domino effect going around like I said is probably to pop that linker off which is freezing it's kind of connecting the two parts of the ring together so they can't move so unless you get that linker off you know you don't have the freedom of movement anymore yes yeah yeah yeah you're right so in that case one of the tail domain that I showed you that had all those other subunits it's a different dining though this is what is called cytoplasmic dining but the concept is is kind of similar so instead of like normally you attach the tail to a cargo and when dining walks you can carry the cargo along the microtubule in the case of the psyllium the tail domain of the dining is attached to one of the pairs of the microtubule and it's locked in there tightly and the modal domains are attaching to the partner microtubule and what happens with that is one microtubule slides relative to the other and that was actually beautifully shown by Ian Gibbons many many years ago where they could isolate the axonium kind of proteolyze connections that are holding the axonium together and when they started activating the dining instead of getting beating they just got microtubule sliding so one of the dinings would just push out the partner microtubule outside of the axonium the big mystery is how you convert sliding motion which is what dining normally would like to do it's just pushing things into a waveform so that motion between these microtubules that are sliding relative to one another is somehow regulated in a way that dinings on one side of the axonium are turned on the other ones are off and then they go back and forth so there's some really interesting and complicated regulation that's converting a simple linear motion into a beating kind of motility