Session 3: G-Proteins

00:00:01.14 Hello, my name is Alfred or Fred Wittinghofer and I'm an emeritus group leader
00:00:07.24 from the Max-Planck Institute for Molecular Physiology in Dortmund, Germany.
00:00:13.06 What I would like to tell you today is about a class of proteins
00:00:18.08 that bind GTP (so GTP-binding proteins) that work as molecular switches.
00:00:23.11 And I will tell you in my first lecture, how they work
00:00:26.14 and in the second lecture, how they lead to a number of different diseases that we have studied.
00:00:32.29 So imagine, for example, that you have a quiescent cell that sits there in the G0 phase.
00:00:38.04 It doesn't grow, doesn't differentiate and it needs a signal from the outside
00:00:43.04 in order to start proliferation right away.
00:00:45.14 And the way it works is that a growth factor hits the cell
00:00:48.22 and induces a series of reactions indicated here by these arrows.
00:00:53.14 And then this series of reactions comes to the cell nucleus
00:00:57.10 where DNA is duplicated and the cell decides to proliferate.
00:01:00.09 And one of the most important elements in the signal transduction chain is Ras
00:01:06.10 a GTP-binding protein, one of the leading molecules in this class
00:01:12.09 and this then regulates cell growth as a molecular switch.
00:01:17.21 And you can imagine, if this regulation does not work,
00:01:21.21 and cell growth is uncontrolled, then you have cancer.
00:01:25.26 And Ras is one important element of cancer formation
00:01:28.16 and I will talk about that in my second seminar.
00:01:33.18 In another example, for example, you have a quiescent cell
00:01:37.04 where you have the actin cytoskeleton marked here in the left part of the pictures
00:01:44.16 where the actin cytoskeleton is very diffuse in blocks and small stripes and so on
00:01:51.16 and then suddenly you hit these cells with
00:01:53.15 a particular class of GTP-binding proteins called Rho
00:01:56.15 and then you see what happens.
00:01:58.05 In one case you get stress fibers,
00:02:00.14 in the other case you get a structure at the cell periphery which are called lamelopodia
00:02:06.26 or here you get structures that are making long extensions which are called philopodia
00:02:13.00 which make the cell move or proliferate or differentiate and so on and so on.
00:02:18.21 And these reactions are also controlled by a GTP-binding protein
00:02:22.27 and they are called Rho, Rac or Cdc42.
00:02:26.13 So the question then boils down to the thing,
00:02:29.28 how do you construct a molecular switch that is reversible,
00:02:33.17 that can be regulated at any level and then does the thing that it's supposed to do?
00:02:40.19 And so nature has devised a very large class of proteins called GTP-binding proteins
00:02:47.03 that comes in two flavors; in the GTP bound state it's on and in the GDP bound state it's off.
00:02:53.01 So the difference between these two states is a single phosphate.
00:02:56.08 And to show that this is a very important class of proteins,
00:03:02.29 we can find more that 38,000 GTP-binding proteins or G proteins, as I would like to call them,
00:03:08.29 in about 1300 genomes by December 2010.
00:03:13.20 So that tells you these are really important molecules found in all kingdoms of life
00:03:18.21 and some of these proteins are the most highly conserved proteins in nature at all.
00:03:24.25 So, let me tell you about how these molecular switches work.
00:03:31.05 And I will talking mostly about Ras-like proteins because these are the ones that we work with
00:03:35.22 and they are sort of the prototype for learning how these are regulated.
00:03:40.01 So you start out with a signal that comes, for example, like a growth factor
00:03:44.29 or whatever, that hits, somehow these G proteins, (and I will be talking to you about that later)
00:03:50.16 and that induces a series of steps that lead to the protein becoming loaded with GTP.
00:03:58.01 And then it has its downstream effect.
00:04:00.10 And in Ras-like proteins it works the following way:
00:04:04.29 these nucleotides are usually bound very tightly (picomolar range)
00:04:09.27 such that GDP never comes off by itself
00:04:12.15 but needs the action of a nucleotide exchange factor which is called GEF (guanine nucleotide exchange factor)
00:04:19.12 which allows GDP to be released much faster
00:04:25.16 and then allows GTP to bind to the protein.
00:04:27.25 And now it is active.
00:04:29.16 And now it can do its effect but you obviously, since it's a molecular switch, you want it to be switched off again.
00:04:36.06 And the way you do that is not the reverse, not the exchange of GTP for GDP,
00:04:42.10 but rather it is the irreversible step, the GTPase hydrolysis.
00:04:48.02 So GTP is hydrolyzed to GDP and Pi and there is another protein that
00:04:53.20 stimulates that reaction because it is intrinsically very slow
00:04:57.27 and becomes stimulated by a protein called GTPase Activating Protein or GAP.
00:05:04.07 And I will be talking about that, obviously, in great detail in my second seminar
00:05:08.07 because that is where you see a lot of diseases being due to inability to hydrolyze GTP.
00:05:15.28 So the downstream effector is then something that is mediating the biological effect
00:05:22.10 and the effector is a molecule that recognizes, specifically, only the active
00:05:26.10 GTP-bound form and not the inactive GDP-bound form.
00:05:31.21 So, just to make you familiar with the way this thing can work,
00:05:36.14 since the cycle of GDP to GTP is regulated by Kdiss or Kd (dissociation) for GDP
00:05:49.19 or is regulated by the GTPase reaction, which is Kcat or Koff,
00:05:55.18 you can see that the signal can either increase Kdiss or it can decrease Koff.
00:06:02.10 In both cases you get an increase in the effect, in the biological effect
00:06:07.18 because the effect is finally determined, really, by the GEF reaction or the GAP reaction.
00:06:16.08 And you can quantify this and say that the amount of biological effect
00:06:21.13 that is coming out of this system is
00:06:23.27 directly proportional to Kon (to the introduction of GTP)
00:06:28.25 or is inversely proportional to Koff (to the GTPase reaction).
00:06:33.26 If you make Kdiss faster, you get more GTP bound to protein
00:06:38.13 or if you make the GAP reaction slower you also have an increase in GTP.
00:06:45.18 So, let's now come to how these proteins look, how you recognize them and so on.
00:06:51.10 So, obviously, are there sequence motifs?
00:06:53.14 Are there structures or biochemistry--are they similar between these proteins?
00:06:57.23 Yes, indeed. What I will show you is that you can identify
00:07:01.17 these proteins very easily from sequence motifs, from structure
00:07:05.10 and also the biochemistry are rather similar between all these different proteins.
00:07:09.25 And you can present some of the general rules for recognizing
00:07:14.13 and working with these proteins which I will do in the next 30 minutes.
00:07:21.01 So there are, obviously, when you look at a new protein you may have sequenced
00:07:27.10 and then you compare the amino acid sequence,
00:07:30.00 if you find these 5 sequence elements, indicated here,
00:07:33.01 called G1, G2, G3, G4 and G5, standing for G binding motifs,
00:07:40.24 then you immediately know that you're dealing with a GTP binding protein or G protein.
00:07:45.22 And these elements are, for example, the first one is the so called P loop
00:07:51.04 is a motif G 4-times x (which means any amino acid there)
00:07:56.28 then another conserved glycine, a conserved lysine, S or T,
00:08:02.25 and this is one of the most frequently occurring sequence motifs in the database
00:08:09.03 because not just G binding proteins but also ATP binding proteins have this sequence motif.
00:08:16.00 The second one is just the conserved threonine
00:08:18.22 and a conserved D x x G as the Switch I and Switch II motifs.
00:08:23.11 I will show you later on what Switch I and Switch II means.
00:08:26.00 And lastly there are two motifs, N K x D (G4) or s A k
00:08:33.20 where only, for example in the last motif, the alanine is totally conserved.
00:08:38.19 These are the motifs that are involved in binding the nucleotide
00:08:41.23 and also are involved in the specificity of the nucleotide.
00:08:47.07 And I should also remind you that there are a few proteins
00:08:51.15 which are also GTP-binding but they have a different fold and a different sequence and so on.
00:08:57.16 These are the most famous examples: tubulin or the bacterial homolog FtsZ
00:09:02.24 which you have heard about in other of these iBio seminars for example from Ron Vale.
00:09:08.14 And there are also a few metabolic enzymes.
00:09:11.23 So it's a strange coincidence that almost all the metabolic enzymes
00:09:17.07 that need energy to catalyze the chemical reaction, that they use ATP and not GTP
00:09:24.15 and only very few examples, indicated here, use GTP.
00:09:28.00 So that seems to be as if nature decided that in order to transmit energy it uses ATP
00:09:34.01 and for the regulation of processes it uses GTP,
00:09:38.14 except for these few examples here.
00:09:44.00 So obviously as biochemists we...or lets say a structural biochemist,
00:09:48.18 that I and my lab is involved in, we would like to
00:09:52.28 understand the system we're working with and the biochemistry and the biology of it
00:09:56.22 by knowing the structure because this is usually giving you the most deep understanding of your favorite system.
00:10:08.05 So, to be brief, for those of you that have never worked with protein structures...
00:10:13.10 and we're using x-ray crystallography to determine the structure
00:10:16.05 although there are other methods that I will not be talking about
00:10:19.01 and there will be actually an iBio seminar series coming up later on.
00:10:24.05 But, I'll give you a brief introduction
00:10:27.10 to give you a feeling for how you get at your particular structure.
00:10:30.21 So you can see that you're not afraid of using the method in your favorite system.
00:10:36.13 So what you start out with is you have to have pure protein
00:10:40.19 sometimes a lot or, let's say lots of milligrams of protein.
00:10:44.23 You crystallize them. Hopefully that works.
00:10:48.07 And once you crystallize them you put them through an x-ray beam.
00:10:51.26 And sorry if this is still a German slide
00:10:56.02 because Mr. Rontgen is the one who discovered x-rays
00:11:00.09 and we still like to call them Rontgen-strahlung which means x-rays.
00:11:05.10 And so you shine these x-rays through a crystal and then these x-rays are diffracted.
00:11:11.22 You sample the diffraction pattern and then you use that to calculate your result.
00:11:19.22 Just to make sure you understand this slide I put the English version of it down there.
00:11:25.09 X-rays shone into a crystal and then being analyzed on a detector.
00:11:30.13 So you end up, by a complicated mathematical calculation,
00:11:35.12 which is standardized by the way so you don't have to really learn all the details about it,
00:11:40.07 but under that method you end up with an electron density map of your favorite protein
00:11:45.09 and now you have to do the really fun part which is build an atomic model of it.
00:11:51.00 And I give you here an example of a particular part of Ras polypeptide. So the chain runs from left to right.
00:11:57.15 You see that, for example, this would be a 5-membered ring that can only be proline.
00:12:06.09 You see, for example, down there, a bifurcated amino acid
00:12:10.00 which can be aspartic or threonine or valine.
00:12:12.07 The same one down here.
00:12:13.18 And there's an aromatic residue up there which has also a little tip on it so that must by a tyrosine.
00:12:19.22 So in other words, you end up
00:12:22.00 (and you probably did it while you were looking at the screen)
00:12:25.21 you end up with your atomic model with the tyrosine, then an aspartic acid,
00:12:31.03 a proline, and a threonine down there
00:12:33.01 and if you look into the sequence of your particular protein
00:12:35.16 you know this must be a certain part of the sequence of your protein.
00:12:41.10 So you end up then, in the end, with a ribbon model of your complete protein.
00:12:47.07 And you can see here that this one is composed of alpha helices and beta sheets
00:12:52.04 and that's why I just call it an alpha-beta protein
00:12:55.14 which is, by the way, typical for any nucleotide binding proteins
00:12:59.01 which is an alpha-beta fold.
00:13:01.14 If you go into the database you will see that.
00:13:04.09 So, if you now look at the sequence motif that I have indicated to you a while ago,
00:13:11.17 G1 to G5, you can now see...where can I find them?
00:13:17.00 And they're actually found only in the loops.
00:13:19.14 So, for example, if you see the first beta strands...
00:13:22.10 so this is the N-terminus, the first beta strand, you go through G1 into the other helix down here
00:13:27.28 and then down into the other loops down here
00:13:31.12 and you see that G1, G2, G3, G4 are all in loops
00:13:35.18 which is again very typical for a protein structure--that the actual fold is a very stable entity
00:13:43.01 and the business part of the protein where you see changes,
00:13:46.29 where things bind and are released,
00:13:50.06 where things are hydrolyzed or chemical conversions are happening,
00:13:54.02 they are happening in loops that combine these structural elements.
00:14:00.00 And if you look at the structure you see, actually, that
00:14:02.22 all the conserved elements here are really on just the one part of the structure right here.
00:14:11.07 And the other part, down there, is probably unimportant
00:14:14.19 in terms of, at least, the interactions with other molecules.
00:14:19.05 Just to give you a flavor of some of the motifs that we are dealing with here,
00:14:24.20 this is, for example, the P loop
00:14:26.25 which is a connection between the beta strand down there, it goes though a loop and ends up in the helix.
00:14:33.27 And you see that the first conserved glycine (which has the number in Ras by the way)
00:14:39.24 and going through the loop. Coming to the other conserved glycine on top, there.
00:14:44.02 The conserved lysine is here and then a serenine or threonine.
00:14:47.23 So what you actually see is that the phosphate sits right in the middle of this loop
00:14:52.29 and that's why it is some how neutralized by the charges in the loop
00:14:59.04 and this the most frequent sequence motif in the database.
00:15:05.02 Many ATP-binding and GTP-binding proteins just contain this motif.
00:15:10.00 For example, you heard about kinesin, about myosin in these iBio structure series,
00:15:14.15 and they also contain the same type of motif.
00:15:20.00 This is shown here again in a little more detail.
00:15:23.15 You can see that the beta sheet that sits right in the middle of this P loop
00:15:28.13 makes the main chain interactions and lysine interaction
00:15:33.02 so that the negative charge of phosphate is neutralized by binding into the P loop.
00:15:39.04 It is also called the polyanion hole
00:15:41.10 by Georg Schulz many many years ago
00:15:43.07 when he worked adenylate kinase.
00:15:47.19 So, now we know the structure of Ras and that was the first structure to be solved
00:15:53.29 many years ago by us.
00:15:56.04 So in order to look at different structures
00:15:59.26 let me first introduce you to the Ras super family of GTP binding proteins
00:16:03.08 where each of the sub-families is, first of all, defined by sequence.
00:16:09.14 So proteins within the sub-family are more similar to each other than proteins outside the sub-families.
00:16:14.09 And if you align them by sequence you also align them by function because
00:16:19.10 each sub-family is involved in some kind of different function.
00:16:23.03 For example, the Ras sub-family is involved in general signal transduction reactions.
00:16:28.03 The Rab family is involved in vesicular transport.
00:16:31.26 The Rho protein sub-family, I introduced you to already, regulates the actin cytoskeleton.
00:16:37.11 And I showed you Rho, Rac and Cdc42.
00:16:40.10 And for example, the Ran sub-family is involved in nuclear transport.
00:16:45.04 It's a nuclear version of Ras. That's why it's called Ran.
00:16:47.22 And Rho, for example, is called Rho for Ras Homology.
00:16:51.05 Rab is called Rab for Ras in the brain and so on.
00:16:54.22 So all the names derive, actually, from the grandfather of the family, which is Ras.
00:17:00.00 So we have been working with a number of these proteins
00:17:02.20 and I will show you a few examples of these
00:17:05.03 and just, for example, compare their structure.
00:17:07.25 And we'll also talk about function in my second seminar.
00:17:12.04 So that indicates that a number of structures have been solved
00:17:16.07 in the meantime by us and many other people.
00:17:18.12 And the first correct structure of Ras was in 1989
00:17:23.13 and we now have about 400-500 deposits in the pdb database.
00:17:29.24 And not just the protein itself, but also complexes with effectors;
00:17:34.00 complexes with GEFs and complexes with GAPs.
00:17:37.02 Obviously, these helped us to understand, actually, the complicated regulation of these proteins
00:17:41.14 much better than only biochemistry would have done.
00:17:46.03 So if you look now, let's say, at a few of these structures,
00:17:48.21 on a first view you would say, immediately, "Yeah, they look totally identical."
00:17:53.11 which is obviously true. If you, for example, compare Ras and Rap, they are rather similar
00:17:59.10 and also Rho and Cdc42 and also Rab.
00:18:06.00 So the overall fold is the same but you see small additional elements.
00:18:09.22 For example, in the Rab-Rho family you see this extra helix here and there.
00:18:15.04 And in the Arl family you see an extra N-terminal helix.
00:18:20.00 But other than that, you see that the structure is the same.
00:18:22.11 So you would ask yourself, "Why would these proteins do different things if they have the same structure?"
00:18:26.26 And obviously the answer is very simple.
00:18:30.05 It's not the fold that determines what these proteins are doing in the biological system.
00:18:34.10 It's what they're interacting with
00:18:36.16 and the interaction is determined by the surface of the protein where you have different amino acids.
00:18:41.18 This can be indicated, for example, by just looking at the charges of the surface
00:18:46.15 where red means negative charge and blue means positive charge.
00:18:50.17 And now if you compare, for example, the similar Ras and Rab proteins you see, really,
00:18:55.25 at the surface it's different enough to
00:18:58.11 make sure that the Ras protein interacts only with its downstream effectors
00:19:02.26 and Rap interacts only with its downstream effectors.
00:19:06.05 And the same is true for the Rho protein which is different from Cdc42
00:19:11.12 but is also different from Ras and Rap and so on and also Arl and so on.
00:19:15.25 So the message from all of this is that the overall fold is exactly the same for
00:19:21.05 all of these proteins but they have additional elements and they have a different surface
00:19:25.01 that makes these proteins do its particular biological reaction that they are involved in.
00:19:32.13 So, the next thing that I would like to introduce you to is how the switch works.
00:19:38.22 So. obviously, in our schemes I have also always shown that the GDP-bound form
00:19:43.25 is different from the GTP-bound form by a symbol.
00:19:47.01 But now we will obviously look at the structure and biochemistry
00:19:49.21 and would like to know: How does this structure actually change
00:19:52.05 when the protein goes form the inactive to the active conformation.
00:19:56.02 So the basic element of the switch mechanism that we would like to understand...
00:20:02.12 how does the protein change structure when it goes from the GDP-bound to the GTP-bound state?
00:20:07.17 ...is the following...most of the structure (which is this grey part down here)
00:20:14.25 does not change at all when the protein goes from one state to the other.
00:20:18.18 But there are two elements in the structure called switch I, down there, and switch II, down here
00:20:24.13 which is part of the conserved elements G2 and G3.
00:20:27.20 And they contain two amino acids; threonine-35 is the number in Ras
00:20:35.24 and glycine-60, again the number in Ras
00:20:38.00 and in different proteins the number would be different.
00:20:40.08 They are bound to the gamma phosphate by these two main-chain hydrogen bonds
00:20:44.05 which are indicated here as springs that are loaded by binding to the gamma phosphate.
00:20:49.13 So switch I and switch II are really important elements of the structure change
00:20:55.22 and they are obviously called switch I and switch II because
00:20:58.23 they change their structure when they go from GDP-bound to GTP-bound.
00:21:02.04 These are the important elements for the biological function because
00:21:07.02 that's where they change structure when going from one form to the other.
00:21:12.07 And I will show you a number of examples of how this looks in detail.
00:21:16.21 For example, in the case of Ras, you see that if you overlay
00:21:21.09 the structure of GDP- and GTP-bound conformations they are almost totally identical
00:21:26.18 in most of the secondary structure elements, shown as beta sheets and helices here.
00:21:32.01 And the only changes, really, are happening in the switch regions,
00:21:35.03 which are down here in this colored area here.
00:21:38.15 where, for example, you have a tyrosine-32 that sits inside and goes outside
00:21:44.16 and you have a threonine down here which goes from the outside to the inside.
00:21:48.20 So you see these two different changes in switch I up there (which is the purple color)
00:21:54.27 and the cyan color down here (which is switch II) you see, again, structural changes,
00:21:59.06 a melting of the helix down there.
00:22:01.02 So there's a localized change
00:22:03.06 and this would be the part where, obviously, proteins that
00:22:06.18 are acting downstream of Ras would be recognizing this part of the structure
00:22:12.07 which is shown here again in a movie.
00:22:16.00 You'll see the grey part of the structure doesn't change much.
00:22:18.07 What is happening is down there.
00:22:20.07 You see these colored loops that are changing
00:22:23.16 and this is element where proteins that would recognize this Ras in the GTP-bound conformation
00:22:29.09 would come and recognize this conformation.
00:22:31.18 We can show the same thing by coloring the surface of the protein
00:22:36.12 and you see here in color again the two switches when they go from GDP- to GTP-bound conformations.
00:22:43.17 And this is actually a simulation of the reaction between the two states.
00:22:48.08 But what you see again, is that the surface of the protein
00:22:51.01 where the action happens changes when it goes from one state to the other.
00:22:57.02 And so that's where the business end of the protein is.
00:22:59.15 This is where almost all the factors that recognize Ras would attack the protein.
00:23:04.20 Just to give you a more dramatic example:
00:23:08.11 So the conformation change is canonical. I showed you that in one of the previous pictures.
00:23:13.08 But there are some wonderful dramatic changes that we can observe
00:23:17.14 by looking at different Ras-like proteins,
00:23:19.28 for example, the protein Ran, that I introduced you to briefly before,
00:23:23.23 a protein that regulates nuclear transport
00:23:26.29 and has a the C-terminus an extra element, an extra helix, which we call the C-terminal helix
00:23:33.10 which is bound by its highly negatively charged end to the positively charged protein.
00:23:39.20 So it sits on the surface in the GDP-bound state.
00:23:42.12 Now, see what happens in the GTP-bound state.
00:23:45.06 You see that now you have the canonical triggering of the switch
00:23:50.16 down at the gamma phosphate side which would be somewhere here.
00:23:53.05 The two switches no change their structure and by doing that
00:23:56.28 they kick out the C-terminal end which does something involved in nuclear transport.
00:24:04.15 An even more dramatic example is shown by the protein Arl or Arf.
00:24:10.02 So here you see a protein up in the GDP-bound state.
00:24:14.06 You see there's alpha and beta phosphate bound but the gamma phosphate
00:24:19.01 binding site is very far away from where it should be if its in the canonical structure.
00:24:24.12 So what happens now, if you have a tri-phosphate bound to the protein down there,
00:24:28.25 it moves what is called the inter-switch region which are actually two beta strands.
00:24:33.22 It moves it by two amino acids towards the N-terminus
00:24:36.21 and kicks out the N-terminus and thereby induces a large conformational change
00:24:41.23 which is shown in a movie again, down here.
00:24:45.01 So you see these two beta strands that move by a very large distance
00:24:49.07 and kick out the N-terminal end,
00:24:51.16 which, the N-terminal end is actually interacting now with the plasma membrane
00:24:54.28 and does something that has to do with vesicular transport.
00:24:59.07 So the message from all of these things is that there is a canonical change
00:25:05.21 when the protein goes from GDP-bound to GTP-bound.
00:25:10.19 The trigger is the same but the affects can sometimes be very dramatic.
00:25:15.10 And I showed you another dramatic example for that.
00:25:20.01 And, incidentally, you have heard a lot in different seminar of this series
00:25:28.26 that motor proteins like myosin or kinesin also do a conformational change
00:25:33.06 between the ATP- and ADP-bound state.
00:25:35.20 And it turns out that the mechanism for that structural change is
00:25:40.07 almost totally identical to that of G protein.
00:25:42.20 So you see here, for example, in red on the left side, for the motor protein,
00:25:46.23 the conserved sequence element.
00:25:49.00 And on the right you see the sequence elements that I have introduced you to before
00:25:53.03 which from GTP-binding proteins and they are actually very very similar.
00:25:56.17 And the conformational change is very similar,
00:25:59.19 except that in the case of motor proteins it is transmitted to do work.
00:26:04.04 So, for example, in myosin the conformational change of the switch II
00:26:07.15 makes a lever arm movement that makes the protein walk along the actin filament.
00:26:12.10 And in the case of kinesin it walks along the tubulin filament.
00:26:18.02 And so in Ras this conformational change is obviously converted not into work
00:26:22.23 but is used to make an irreversible change from GTP-bound to GDP-bound state
00:26:28.15 by the GTPase reaction.
00:26:31.16 So let me at the end, or towards the end,
00:26:35.18 show you some of the biochemical properties of the protein,
00:26:39.07 in particular, of small G proteins which we usually work with.
00:26:42.27 So all of them have a high affinity to nucleotides.
00:26:47.26 High affinity means in the pico molar or nano molar range.
00:26:51.01 Which means that (and this is the second point here)...
00:26:55.07 that means that the dissociation of nucleotide is very slow.
00:26:58.17 The half-life of dissociation at 37 degrees for GDP would be on the order of 20 minutes to 30 minutes.
00:27:06.00 Which would be way too long for a signal transduction process
00:27:10.18 where this protein is activated within minutes.
00:27:13.04 And that's why, obviously, you have these nucleotide exchange factors
00:27:16.07 that I showed you before.
00:27:17.17 The third thing that I would like to introduce you to is that the affinity is magnesium dependent.
00:27:23.20 So in the absence of magnesium, for example, the affinity
00:27:27.06 of nucleotide is 1000-fold less which is a very interesting technical thing for us.
00:27:32.25 I will show you that in a minute also...so, why we can work with these proteins.
00:27:37.02 You can reduce the affinity just by removing magnesium.
00:27:40.19 The fourth things that's here on the slide is
00:27:43.23 that there is a very high specificity for guanine nucleotide.
00:27:47.17 For example, I told you the affinity is in the order of pico molar to nano molar.
00:27:52.02 But the affinity to ATP or adenine nucleotide is in the order of millimolar.
00:27:58.21 So, a huge difference in affinity of guanine nucleotide versus adenine nucleotide.
00:28:05.10 And I'll, again, show you that in a minute, how that comes about.
00:28:07.25 Also, I showed you already that the GTPase is very often very slow,
00:28:13.06 again, with a half-life at 37 degrees of 20 to 30 minutes.
00:28:17.27 And again, in a biological process, you want this reaction to be much faster.
00:28:21.27 So, actually, almost all G proteins are very slow enzymes, very slow phospho-transfer enzymes.
00:28:29.12 But, they become, obviously, very good transfer enzymes
00:28:32.25 when in the presence of the GTPase Activating Protein.
00:28:36.08 And finally, again, which is a very universal observation, the GTPase reaction
00:28:45.05 just like the ATPase reactions, are always magnesium dependent.
00:28:49.00 And I'll show you again what that looks like.
00:28:50.20 So, you need a divalent ion first to make it high affinity
00:28:54.29 and for the GTPase reaction.
00:29:00.06 So for example, how does the specificity of guanine nucleotide come about?
00:29:05.02 Here you see the two elements that I introduced at the very beginning,
00:29:08.03 G4 and G5 down there which is N or T KxD which is the G4 motif
00:29:16.15 or the sAk where alanine is probably the only really conserved element.
00:29:22.12 And what you see is the guanine base making a very strong, bifurcated hydrogen bond
00:29:27.21 to this aspartic acid down there which obviously adenine couldn't do.
00:29:32.27 And also the alanine makes a main chain hydrogen bond with the oxygen
00:29:37.24 which would be the amino group in the case of adenine and again that couldn't happen.
00:29:41.21 And you see that the lysine which sits underneath the base
00:29:44.18 and the alanine make additional contacts.
00:29:47.04 So that's the reason for the specificity.
00:29:50.15 And also if you mutate any of the residues around the guanine base
00:29:55.13 binding site you make the binding very much lower affinity.
00:30:01.23 So here you see the essential magnesium ion.
00:30:07.06 It is bound between the beta and the gamma phosphate.
00:30:11.06 These are two of the ligands of magnesium.
00:30:13.17 You see there are two ligands coming from the protein itself; a threonine and a serine.
00:30:18.10 This is always conserved in the GTP-bound state for all GTP-binding protein.
00:30:22.22 And then you have two water molecules that make the coordination field complete.
00:30:30.17 And a general theme of any ATPase and any GTPase is that you have magnesium as
00:30:37.22 a bidentate interaction like this one here where the gamma phosphate is to be transferred
00:30:43.12 but if you have a cleavage at the alpha phosphate here,
00:30:46.14 magnesium is sitting between alpha and beta.
00:30:48.10 So, that's a general thing:
00:30:50.15 that you need a bivalent ion to neutralize the charges of the phosphate.
00:30:57.26 And I will talking about that, obviously, in my second seminar
00:31:00.23 where we'll talk in detail about the GTPase reaction.
00:31:05.01 Another thing that is very important for analyzing and working biochemically with
00:31:11.15 these proteins is the high affinity is such that if you isolate the protein, let's say
00:31:18.03 from E. coli (make it recombinately), that you want to make sure that
00:31:23.03 you know what nucleotide is bound and to what extent.
00:31:28.03 So, if you run, for example the protein, on an HPLC column,
00:31:31.06 you see this peak here at a certain elution volume.
00:31:34.19 If you compare that to a standardized elution diagram
00:31:39.20 where you take a mixture of GMP, GDP and GTP in a control reaction you see that this is GDP.
00:31:47.05 And you can also quantify that because this has been equilibrated
00:31:52.23 with a certain about of nucleotide such that you now know
00:31:55.28 that your protein has a 1:1 complex of protein and GDP,
00:32:00.26 which is important because if you want to analyze the protein
00:32:04.07 in any of the reactions that this protein does
00:32:06.12 you want to make sure that you know which protein is bound to it already and to what extent.
00:32:13.15 People make a lot of mistakes by forgetting that the protein already has nucleotide
00:32:19.24 and that is important if you analyze the following reaction that I will show you.
00:32:25.12 For example, you want to do an exchange.
00:32:29.12 You need to to convert the protein into the form that
00:32:34.03 you would like it to be in in oder to analyze the following reaction.
00:32:37.24 And the way you do that is you use EDTA to make a very quick exchange of nucleotides.
00:32:44.09 So, EDTA picks off the magnesium
00:32:46.04 and without EDTA you see here that the exchange reaction would be very slow.
00:32:49.16 So you're looking at the reaction of Ras-GDP and you want to introduce a GTP version of that,
00:32:55.17 for example, tritium labeled GTP or gamma-P32 labeled GTP.
00:32:59.20 Here, you see that reaction would take a very, very long time in order to get to completion.
00:33:04.19 But in the presence of EDTA, it takes minutes
00:33:06.27 and you have a full conversion of the protein into the GTP-bound form, for example,
00:33:11.16 if you want to, let's say, analyze the GTPase reaction
00:33:14.17 because if you start with the GDP form,
00:33:17.24 the rate-limiting step would be exchange, rather than the GTPase reaction.
00:33:22.15 And we also use, for example, of fluorescent analogs of GDP, which is shown here.
00:33:31.15 So you have on the ribose or the deoxyribose (depending on which you want to use)
00:33:36.06 you have a fluorescent reporter group which is called mant or mGDP or GTP
00:33:42.12 which is a very sensitive probe for the interaction of the protein with
00:33:47.18 other proteins, with nucleotide itself and so on.
00:33:50.14 And we have shown that this fluorescent reporter group does not
00:33:54.05 disturb most of the reactions of the protein that we are working with.
00:34:00.03 For example, here, this shows you what a wonderful fluorescent reporter group this is.
00:34:05.13 So, the lower curve here would be the emission curve of the mant-GDP or -GTP
00:34:13.18 which is a certain amount of emission spectrum you get
00:34:17.16 and if you add Ras you see that you get a huge increase in the fluorescence emission
00:34:25.04 which means, probably, that the probe is now in a completely different environment
00:34:29.00 and that's why you get such a wonderful change
00:34:31.18 which you can use now to do lots and lots of different reactions.
00:34:34.29 Just to again give you an example, you first load the protein with a fluorescent analog
00:34:40.26 (this is Ras bound mGDP) and now you want to analyze the exchange reaction
00:34:45.22 with one of the guanine nucleotide exchange factors that I showed you in the beginning.
00:34:51.28 So, in the absence of an exchange factor, at room temperature,
00:34:57.19 there would be almost no exchange in the time frame that is shown here (600 seconds).
00:35:03.10 But now, if you add one of the exchange factors
00:35:05.22 (which is obviously the catalytic domains of one of the exchange factors),
00:35:08.27 you see you get a very fast release of nucleotide.
00:35:13.05 The way we do this is we take a large excess of unlabeled GDP
00:35:17.10 to replace the fluorescent derivative and thereby, you get a decrease in fluorescence
00:35:25.01 because when the mant-GDP becomes free, fluorescence is about one half of what it was before.
00:35:31.02 And you use a very high excess of GDP in order to
00:35:35.05 make the back reaction spectroscopically silent.
00:35:38.18 In other words, you can now analyze this by a first order reaction analysis
00:35:45.10 and say what is Kcat of the exchange reaction of Ras + Sos.
00:35:51.11 In the very end, let me briefly show you some multi-domain G-proteins
00:35:59.18 because so far I have talked, mostly, about Ras-like proteins which consist mostly, of just the G domain.
00:36:05.13 And obviously there are a larger number of G proteins that consist of
00:36:10.14 many more domains and are much, much larger.
00:36:13.04 So the proteins that we talked about, the Ras-like proteins are in the order of 20 kDa
00:36:17.14 but the proteins that are indicated here are much, much larger.
00:36:21.17 For example, the translation factor...
00:36:24.17 so these are the factors that regulate ribosomal biosynthesis
00:36:28.04 which are called EF-Tu or EF-1, EF-G and so on.
00:36:32.12 So all of these use GTP conversion
00:36:36.10 for driving some part of the ribosomal synthesis and these are between 40 and 80 kDa large.
00:36:44.11 And these are the most highly conserved G-protiens in the database.
00:36:48.17 So, they are highly conserved even between bacteria and man.
00:36:54.07 Then there are the heterotrimeric G proteins.
00:36:57.18 These are the proteins that are coupled to G protein coupled receptors
00:37:01.04 which sits in the membrane and transmits a signal to these G proteins
00:37:07.07 which are trimeric proteins consisting of G alpha, G beta and G gamma
00:37:11.29 but where the G alpha protein is the actual G protein and which is about 40kDa.
00:37:16.23 I talked about the Ras superfamily. It's a very large superfamily of proteins, all of 20-25kDa.
00:37:23.16 Then you have the Dynamin superfamily of proteins of about 80kDa more or less.
00:37:29.07 It's very important for certain aspects of vesicle formation.
00:37:35.15 Then you have a very small family which is the Signal Recognition Particle
00:37:41.18 which takes the ribosomal nascent complex and brings it to the membrane.
00:37:48.04 And its receptor is in a very small family but it is also very well known and highly conserved family between bacteria and man.
00:37:56.21 Then you have, for example, the septins which are proteins that form filaments
00:38:03.07 by taking the G-domain and making polymers out of it.
00:38:05.24 There is a large family of that in people, at least.
00:38:09.13 And there are many small subfamilies that were not mentioned in detail.
00:38:14.04 All together, you have a couple of hundred proteins, let's say, in mammalian cells.
00:38:23.10 So, if you overlay the structure of all of these, you see this is a huge amount
00:38:28.26 of colored spaghetti that I am showing you where each color indicates a different protein.
00:38:34.27 And you can analyze them and superimpose them very well by overlaying them on the G domains.
00:38:41.08 So the G domain is totally conserved between these very very different proteins.
00:38:44.16 And, for example, you see down there the green stuff here,
00:38:47.00 that would be Elongation Factor Tu. This protein here, hGBP1 would be a dynamin-like protein.
00:38:53.01 And you have up there the SRP, the Signal Recognition Particle and so on.
00:38:57.23 And to just give you one example, mainly that of EF-Tu.
00:39:03.20 So if you look at the topology...so the green stuff here,
00:39:06.25 all the green elements here would be the normal G domain, the alternating beta strands and helices.
00:39:14.01 Here you have the conserved loop elements, the conserved sequence motifs.
00:39:18.08 But then at the C terminus you also have two extra domains; domain 2 and domain 3.
00:39:23.23 And all together this protein is about 45 kDa large.
00:39:28.18 And again, you see now, if I show you the next movie,
00:39:33.18 you see that now the conformation change that, again, is the canonical way I've introduced you to.
00:39:40.16 Down there are the switches.
00:39:41.19 But, now these switches change the structure of these two extra domains in a significant way
00:39:48.00 and this protein binds aminoacyl-tRNA only in the GTP-bound, compact state (which is this state)
00:39:54.02 and not in the open state that you see after the conformational change.
00:39:57.08 So that indicates to you that the trigger for the conformational change is the same
00:40:02.02 but yet it can lead to very drastic conformational change in, also, other domains.
00:40:10.00 So, let me then come to my conclusions which are the following:
00:40:14.21 G proteins are universal switch molecules
00:40:17.06 and I hope I have convinced you that the principle of how they work is universal.
00:40:24.08 Their G domain has a typical alpha-beta structure.
00:40:28.02 I showed you that as an alternating sequence of beta strands and alpha helices.
00:40:35.05 They work by this canonical switch mechanism which we have called the loaded spring mechanism
00:40:40.12 where the two switch elements are bound to the gamma phosphate by main-chain hydrogen bounds
00:40:45.27 are now released when the protein losses the gamma phosphate by the GTPase reaction.
00:40:50.26 They are very specific (most of them at least) for guanine nucleotide and don't bind adenine nucleotide.
00:40:58.28 This is again, different from adenine nucleotide-binding proteins which are usually not so specific.
00:41:05.11 For example, for us the difference is 10 to the 6-fold at least, if not more,
00:41:10.23 for the difference in affinity for GTP versus ATP.
00:41:15.04 They have a very slow intrinsic nucleotide exchange where the dissociation is very slow
00:41:19.14 and is catalyzed by this factor called GEF.
00:41:22.17 They have a very slow intrinsic GTPase reaction
00:41:25.28 which is, again, catalyzed by the protein GAP (GTPase activating protein).
00:41:30.05 So the whole system of the molecular switch is regulated by these GEFs and GAPs
00:41:38.04 which make the reaction faster and can be regulated in the context of a biological system.
00:41:43.03 And I will show you, obviously, in my next seminar, in detail how these GAPs function.
00:41:49.04

00:00:00.14 Hello. My name is Fred Wittinghofer.
00:00:02.18 I'm an emeritus group leader at the Max Planck Institute for Molecular Physiology in Dortmund, Germany.
00:00:11.08 And this is my second seminar
00:00:14.01 and in the first one I introduced you to the molecular switches called GTP binding proteins or G proteins.
00:00:21.15 And in this second part I will more concentrate on a particular aspect of our research which is
00:00:28.20 dealing with the mechanism of GTPase reactions and how they lead to a number of different diseases.
00:00:34.09 And I will obviously focus mostly on Ras-like proteins that I talked about in my first seminar.
00:00:42.23 So, again, just briefly, the mechanistic cycle for these proteins:
00:00:49.25 They come in two flavors these proteins; the GDP-bound and GTP-bound state.
00:00:55.13 They have nucleotide bound very tightly.
00:00:58.03 They need a guanine nucleotide exchange factor to release GDP for GTP
00:01:03.23 because in the cell, there is more GTP than GDP.
00:01:06.11 That's why the protein become loaded with GTP once GDP is off.
00:01:10.06 They have a downstream affect in the GTP bound form; interacting with some partner proteins.
00:01:15.10 And in order for the switch to be switched off again, you have the GTPase reaction
00:01:20.27 whereby the protein splits GTP into GDP and Pi.
00:01:24.29 And this reaction is very slow and is catalyzed by proteins called GTPase Activating Proteins
00:01:31.06 which will obviously be the major thing that we will be talking about.
00:01:36.24 So what we're talking about, really, is a really basic biochemical reaction;
00:01:43.23 namely, the hydrolysis of phosphoanhydrides and it's similar to the hydrolysis of phosphoesters.
00:01:53.08 For example, when you have phosphorylated protiens which are phosphorylated on the threonine, serine or tyrosine,
00:01:57.24 you have a similar nucleophillic attack on the phosphate by water.
00:02:05.04 And obviously, people have been talking about this reaction a lot because it is a very
00:02:12.15 slow reaction for the reason that the phosphates are highly negatively charged
00:02:19.16 and the approach of a nucleophile, for example a water which is partially negatively charged also,
00:02:25.06 is very, very disfavored. And that's why this reaction is normally very slow.
00:02:30.25 So even though the reaction delivers energy,
00:02:33.17 it is very slow because you have to overcome the very high activation energy
00:02:38.16 which depends on what I just told you about--the negative charges.
00:02:42.04 And the higher the activation energy, you might know from your biophysical courses,
00:02:46.13 that the higher the activation energy, the slower the reaction,
00:02:51.00 because the reaction rate is directly proportional to the activation energy.
00:02:54.18 So, nature, then, uses enzymes to lower the transition state energy
00:03:00.27 and thereby make the reaction faster.
00:03:03.15 And there's an interesting article that I bring to the attention of my students all the time
00:03:07.21 from Francis Westheimer who wrote an article many years ago:
00:03:12.01 Why Nature Chooses Phosphates.
00:03:13.25 Because chemists never use phosphate as a leaving group
00:03:17.13 but in biology it is a very frequent leaving group
00:03:21.16 So, because of just this purpose here, because of just what I showed here
00:03:26.17 the phosphoanhydride bond or the phosphoester bond is kinetically stable
00:03:32.06 in other words you can have your ATP or GTP in water and it stays forever or hydrolyzes very slowly.
00:03:39.22 But thermodynamically it's principally unstable.
00:03:43.00 If you use an enzyme, it lowers the activation energy--you can make this reaction very fast.
00:03:47.14 So that's an interesting article that I would recommend you to read.
00:03:50.27 For example, that's why DNA is stable and that's why ATP is such a wonderful source of energy
00:03:57.12 because it, in principle, delivers energy but is stable in aqueous solution.
00:04:03.28 So let's start then with first of all Ras and its GAP-mediated GTPase reaction because
00:04:10.25 that is really the paradigm for many of the things that have been developed afterwards.
00:04:15.24 So this is the molecular Ras. Its a ribbon representation of the structure
00:04:21.29 of the G domain of Ras.
00:04:24.16 You see its heart shaped because...
00:04:27.17 obviously we like it very much and we solved the structure in Hiedelberg.
00:04:32.00 The city of Hiedelberg's advertisement, Ich hab mein Herz in Hiedelberg verloren
00:04:36.04 which means, I lost my heart in Hiedelberg.
00:04:37.27 But the most important reason it's heart shaped is because
00:04:41.13 people call it the beating heart of signal transduction.
00:04:44.06 And so its one of the most important molecules that regulates
00:04:48.14 important signal transduction processes like growth, differentiation and sometimes even apoptosis.
00:04:54.17 And you probably know a little bit about the signal transduction process because
00:04:59.07 every text book has this version, one of the paradigms of signal transduction chain
00:05:04.01 whereby, for example, a growth factor binds to the cell membrane
00:05:08.27 and binds to its growth factor receptor RTK (receptor tyrosine kinase)
00:05:14.22 whereby this becomes phosphorylated. It recruits the exchange factor SoS
00:05:19.04 which then activates Ras to the GDP bound form.
00:05:22.16 And then, now, Ras interacts with the downstream component which is Raf kinase
00:05:26.15 which is the starting kinase for what is called the MAP kinase module.
00:05:30.24 There, one kinase activates the next downstream kinase which is called MAP kinase kinase kinase.
00:05:36.29 MAP kinase kinase kinase activates MAP kinase kinase and that activate MAP kinase
00:05:40.28 which then goes into the nucleus and activates transcription.
00:05:45.03 So this is a very simplified version of what Ras is actually doing.
00:05:48.16 and this is the first one to be discovered (the first signal transduction).
00:05:54.15 But, now it becomes more and more complicated.
00:05:57.26 This is still a very simplified version but it shows you already
00:06:00.27 the major thing about signal transduction via Ras in that many upstream components come and activate Ras.
00:06:09.18 It's either tyrosine kinase receptors or G-protein coupled receptors
00:06:13.09 or T-cell receptors. All of them can activate Ras.
00:06:16.08 And then Ras can activate, downstream, many components, not just Raf kinase
00:06:20.18 but also a molecule called Ral GEF, PI(3) kinase, PLC epsilon and others.
00:06:26.22 And they mediate a number of signal transduction reactions
00:06:32.12 which somehow are integrated somewhere, let's say in nucleus,
00:06:38.00 by a transcription factor and initiate, when the threshold is right, when the number of reaction is right,
00:06:46.15 it initiates a cellular response which can be proliferation or differentiation or whatever.
00:06:52.03 And I will not deal with any aspect of this because I want to concentrate on the switch-off reaction.
00:06:57.28 So the switch-off reaction is again the scheme you have seen, now, many times
00:07:03.25 but what happens in Ras is also an oncogene;
00:07:06.26 an oncogene that has two types of mutations, either the glycine 12 mutated to any other amino acids
00:07:13.10 or glutamine 61 mutated to any other amino acid.
00:07:16.21 And what this does biochemically is, it blocks the GAP-mediated GTPase reaction.
00:07:23.21 So you can imagine what happens, you have blocked the return to the inactive state
00:07:29.13 and that's why you now accumulate Ras in the GTP bound form.
00:07:33.09 You don't need any of the upstream signaling anymore because Ras is already in the GTP bound state.
00:07:39.03 And now it has an effect that is not regulated anymore and that's why it leads to cancer.
00:07:44.14 So these simple mutations...and that's why, obviously, as structural biochemists
00:07:49.23 it is a very interesting project to ask the question:
00:07:52.00 How can it be that a single point mutation has such dramatic consequences?
00:07:58.19 And its also...obviously since Ras is the most frequent oncogene.
00:08:02.28 About 25% of the people that come to the clinic diagnosed with a tumor,
00:08:09.07 they have a Ras mutation, one of the ones I showed you. At least these are the most frequent ones.
00:08:15.00 So obviously every drug company also are working on trying to inhibit the Ras pathway
00:08:20.29 and Ras signaling as a way of treating Ras-mediated cancers.
00:08:25.28 And there are many approaches that one can think of.
00:08:30.17 I just indicated a few here. For example, you could think of Ras...is farnesylated
00:08:36.25 at the C-terminal cysteine and there's an enzyme called farnesyl transferase that mediates that reaction.
00:08:44.18 There are farnesyl transferase inhibitors that are just in the clinic being tested for their efficacy.
00:08:50.00 But you could also think of maybe inhibiting the interaction with downstream effectors.
00:08:54.15 This is, for example, a structure we determined: the complex between Ras and Raf kinase (or part of Raf kinase).
00:09:00.10 Or you can even think of...and that's what we're working on still...
00:09:04.21 Our dream approach is...the basic feature of oncogenic Ras is that is cannot hydrolyze GTP.
00:09:12.09 Can we think about making small molecules that would induce GTP hydrolysis on oncogenic Ras?
00:09:18.21 This is probably a dream project and we're still working on it to make it feasible.
00:09:25.05 But I will show you, in the course of my seminar presentation, that the reason we still think its possible
00:09:32.23 is that the chemistry of it should not be so difficult.
00:09:36.14 And I hope I can convince you and that you will go with me on that point in the end.
00:09:40.26 So, we are talking about, here, a nucleophilic attack of water on the gamma phosphate of GTP
00:09:49.03 mediated by RasGAP and it produces Ras-GDP and Pi. A simple biochemical reaction
00:09:56.09 but if it doesn't work it leads to very drastic consequences, namely cancer.
00:10:00.21 So first of all (And you have seen that picture also a number of times.
00:10:05.24 This is my introductory slide, always)
00:10:08.09 is that the two amino acids that are most frequently mutated
00:10:13.16 (either of them) in oncogenic versions of Ras, they are very close to the active site.
00:10:18.07 So you see the nucleotide here: that is the gamma phosphate
00:10:22.07 And you see this is glutamine 61 and this is glycine 12 are very close to the active site
00:10:28.10 obviously, as you would probably predict or thought of before,
00:10:32.06 is that they are somehow involved in the GTPase reaction.
00:10:35.22 Now, we'll obviously show you what they actually do and why the mutation leads to a block in GTP hydrolysis.
00:10:43.08 And you also can see here from a surface representation of the Ras active site
00:10:49.12 that if this GTP...so it sits...its bound to the surface
00:10:53.26 and you see the gamma phosphate is still approachable from the outside
00:10:57.03 and that will be important in the context of what I will be talking about
00:11:01.12 and you see also, I talked, last time about magnesium being important.
00:11:05.20 If there is no magnesium around you also have no GTP hydrolysis reaction.
00:11:09.24 So, as always in a mammalian genome, there are not just one RasGAP, but rather, 12 or 13.
00:11:19.08 And I indicated here a few of these.
00:11:24.06 And you see they all contain one particular domain of about 330 residues.
00:11:28.09 This is the RasGAP domain which by itself is able to initiate the fast GTP hydrolysis reaction.
00:11:36.17 And you see that all these proteins are composed of different domains.
00:11:43.03 And some are similar, some are vastly different.
00:11:45.20 And they obviously are regulating Ras in different contexts of the cell
00:11:50.04 due to their different domains that they have in addition.
00:11:53.15 And I will be talking about one, the first RasGAP to be discovered by Frank McCormick which is P120GAP.
00:12:02.06 And I will be talking later on about NF1
00:12:05.05 which is another important element for tumor formation because it is a tumor suppressor gene.
00:12:12.04 So first of all, and again to remind you, the intrinsic GTP hydrolysis is very slow.
00:12:17.23 But, if you add the GTPase activating protein, it is fast.
00:12:21.08 So what I have been doing here is I take radioactive GTP (gamma labelled, for example)
00:12:25.27 and then I measure the production of Pi over time.
00:12:29.26 You see, down there, this reaction (this is at room temperature) is almost negligible.
00:12:35.14 There's almost no hydrolysis of normal Ras at room temperature without GAP.
00:12:40.12 And if you now add a particular amount of RasGAP you see that there is a very fast reaction.
00:12:44.23 Much much faster.
00:12:46.04 And under limiting conditions it's actually about 10^5-fold stimulation of that reaction.
00:12:53.10 So that is a way we want to look at it.
00:12:57.05 We want to analyze how GAP mediates this fast GTP hydrolysis reaction.
00:13:03.14 So, initially, obviously, you ask yourself: What could be the mechanism of hydrolysis
00:13:10.11 and what is the step that is catalyzed by GAP?
00:13:14.02 You can think of Ras-GTP being, in principle, a fast GTP hydrolysis enzyme
00:13:20.14 but it needs to come into a GTPase competent state by an isomerization reaction.
00:13:26.15 And that is very slow and is catalyzed by GAP.
00:13:28.29 Or you could think that the actual cleavage reaction, going from GTP to GDP Pi,
00:13:35.02 is very slow and is catalyzed by GAP.
00:13:38.01 Or you could even think that all of that is still very fast
00:13:41.09 but the release of Pi to product is very slow and that is catalyzed by GAP.
00:13:47.08 For example, if you remember having heard about the hydrolysis of ATP on myosin,
00:13:53.26 myosin is, for example, a very fast GTPase but the Pi release is very slow
00:13:59.00 and needs to be catalyzed by actin.
00:14:01.04 So, in other words, what is the actual step that is catalyzed by GAP in this particular case?
00:14:07.10 And I should say that it is the cleavage reaction itself
00:14:11.12 which is the major point of attack by GTPase activating protein
00:14:16.21 and that only in the presence of GAP do the other reactions
00:14:20.18 become at least partially rate limiting, at least the Pi release.
00:14:24.10 And I'll show you that later on, as well.
00:14:27.29 And although I've shown you that before, I will repeat that again.
00:14:32.12 So what we use a lot in our studies where we do biochemistry with fluorescent nucleotide,
00:14:38.20 we use a mant- or mGDP or mGTP analog
00:14:42.29 where you have, on the ribose, a fluorescent reporter group.
00:14:46.24 So this would be either deoxy ribose or ribose and on the ribose you have bound by
00:14:52.17 these ester bonds, the mant group which is very sensitive to the environment that it sits in.
00:14:59.19 And that's why it always give very beautiful structural changes as I will show you in a minute.
00:15:07.10 We used stop flow for measuring fast reactions because, remember, the GTPase reaction,
00:15:13.29 which is slow in the absence of GAP becomes very fast
00:15:17.15 and so in order to analyze it in detail, we need to use stop-flow kinetics.
00:15:21.16 So what you'll, for example, do: you have Ras labeled with
00:15:25.10 mant-GTP (so it's the fluorescent version of GTP)
00:15:28.16 and you have GAP and you shot them together into a fluorescent-detection cuvette
00:15:34.02 and you have the stop flow up here in order to
00:15:37.17 make sure that you only put a certain amount of liquid from these two syringes into your reaction chamber.
00:15:44.07 So when we do that, when we shot these two things together, we see that there is
00:15:50.09 a fluorescent increase if you use Ras-mantGTP and a decrease.
00:15:54.09 The increase, again, is very fast.
00:15:56.25 It means the two proteins make a complex.
00:15:58.27 And then over time, the complex dissociates because after hydrolysis neurofibromine does not bind to Ras anymore.
00:16:07.18 And all of this is over, as you can see, after one second.
00:16:10.09 So, very fast phospho-transfer reaction in the presence of saturating amounts of GAP.
00:16:15.07 And as a control, we use Ras bound to an analog, GppNHp,
00:16:21.25 where you that between the beta and the gamma phosphate you have an NH group
00:16:25.16 which cannot be hydrolyzed anymore and now you have, also, a very fast increase
00:16:30.17 (which is complex formation) but no dissociation because that cannot be hydrolyzed.
00:16:35.05 So that is also the proof that the dissociation is due to GTP hydrolysis.
00:16:40.21 So one may ask, "In this reaction here, what on GAP, which residues on GAP,
00:16:50.01 are involved in mediating this fast GTP hydrolysis reaction, that 10^5 fold stimulation of the reaction?"
00:16:58.17 And obviously, you were thinking of which residues, which amino acids could do the job.
00:17:03.16 Or is it more than one amino acid?
00:17:05.16 And as a good candidate, we were thinking of an arginine
00:17:08.16 because if you look at another G-protein, G-alpha protein (of the heterotrimeric G proteins)
00:17:15.02 it is known that that consists of two domains;
00:17:18.25 so this blue domain is the G-domain and the red stuff here is a helical extra domain.
00:17:24.21 But in that helical domain you have an arginine which sits right, smack in the active site
00:17:29.24 and it has been shown that that arginine is important for the reaction.
00:17:33.29 And that has been shown, in a very old experiment
00:17:37.20 because there is a bacterial pathogen called Vibrio cholera which induces cholera
00:17:45.29 and what is does is it actually modifies this arginine on this G-alpha protein.
00:17:51.16 You see, you have NAD and the cholera toxin transfers ADP-ribose onto
00:17:58.03 an arginine of G-alpha which is shown here. This is actually from the textbook here.
00:18:02.26 So this a very old reaction. It had been analyzed many, many years ago.
00:18:07.10 And it shows, when you do this reaction, when you block
00:18:10.16 the GTPase reaction on G-alpha protein, there is no GTP hydrolysis anymore
00:18:16.10 and the protein is now producing cyclic AMP,
00:18:19.14 it opens up channels (cyclicAMP-gated ion channels) and then you get all these symptoms of cholera.
00:18:26.19 So again the question is: would arginine be a good candidate?
00:18:29.29 So we looked for conserved arginine in all the RasGAPs that I've shown you
00:18:33.24 and indeed, we found several, most of them didn't make any difference.
00:18:38.06 But there was only one arginine that when mutated shows the following pattern in the reaction.
00:18:43.19 So you, again, have in the green version (this is the wild-type version);
00:18:47.23 fast increase due to complex formation, fast decrease due to dissociation after GTP hydrolysis.
00:18:53.16 And with these mutants here, arginine mutated to either lysine or alanine
00:18:58.11 or anything you'd like to mutate it to,
00:19:00.03 there is again an increase in the reaction but then it stays up there, no hydrolysis whatsoever.
00:19:06.13 Or, at least under those circumstances, up to a second, there is no hydrolysis.
00:19:10.16 Which tells you already, obviously, that this arginine must be essential for the reaction.
00:19:18.02 So far so good. The biochemistry was clear
00:19:21.09 but obviously you don't really know what arginine is doing in the context.
00:19:24.05 You think you have an idea; it may be going into the active site
00:19:28.02 but, again, we have to wait for the structure to tell us really what it does.
00:19:33.21 First of all, though, let me introduce you to another important concept for analyzing phospho-transfer reaction.
00:19:40.12 And that is using aluminum fluoride complexes.
00:19:43.02 You may wonder why such an inorganic molecule would be important
00:19:47.14 but it turns out that if you look at the nature of the transition state,
00:19:51.20 so this would be GTP being attacked by water.
00:19:55.18 You have a transition state where the phosphate now makes a triginal, flat thing
00:20:00.26 where the nucleophillic oxygen and the leaving group oxygen on the other side
00:20:08.22 are the axial ligands of this penta-coordinate phosphate
00:20:12.11 and this transition state can either be very tight (which is then called associative)
00:20:18.16 or very loose depending on the distance here between phosphate and the two axial ligands.
00:20:25.26 And then you end up with, again, tetragonal phosphate (this Pi free phosphate).
00:20:32.08 And it turns out that aluminum fluoride
00:20:36.05 (which has distance between aluminum and fluoride very similar to
00:20:39.23 between phosphate and oxygen and they are also highly electron negative)
00:20:44.26 that this is a very good mimic of the transition state of the phospho-transfer reaction.
00:20:51.02 The only problem was that while, for example kinesin or myosin or
00:20:56.28 many other phospho-transfer enzymes use aluminum fluoride as a mimic of the
00:21:02.07 gamma-phosphate in the transition state,
00:21:04.17 Ras or Rho and all of these Ras-like proteins, never showed that.
00:21:08.25 Here you see that experiment.
00:21:11.19 You take Ras-mantGDP (it has a certain fluorescent emission spectrum with a maximum at around 440)
00:21:19.03 and now you add the GAP and with one there's no change whatsoever to the spectrum.
00:21:24.12 So nothing happens.
00:21:25.11 But if you now add aluminum fluoride, you see that now you get a blue shift first of all
00:21:31.00 and then an increase in the absorption.
00:21:37.11 That means there is a trimeric complex between Ras, NF1 and aluminum fluoride
00:21:42.07 where the aluminum fluoride sits in the gamma phosphate binding site.
00:21:47.10 And that was very instrumental...
00:21:48.25 and by the way, if you now do this in oncogenic mutants
00:21:52.02 which we know does not hydrolyze GTP and do the same experiment,
00:21:55.15 you see, in the presence of aluminum fluoride and Ras, there's no fluorescent change.
00:22:00.06 And now you add NF1 and there's no increase in fluorescence
00:22:04.12 because you have a mutation (Q61L) that does not hydrolyze.
00:22:08.13 You can also take a mutation of NF1, for example with the arginine mutation,
00:22:13.22 and again there would be no change whatsoever.
00:22:16.05 So in other words, what these experiments again tell us is that Ras in an incomplete phospho-transfer enzyme.
00:22:22.22 It needs the presence of a GAP in order to look like a phospho-transfer enzyme
00:22:28.05 and if there's any mutation that blocks the GTPase reaction, either on Ras or on NF1,
00:22:33.22 we also get no aluminum fluoride complex.
00:22:37.09 So obviously, all of this is nice in terms of doing biochemistry
00:22:42.00 but what really mediates the GTPase reaction, the fast one,
00:22:47.22 we think can only be verified by looking at the structure of the Ras and RasGAP complex.
00:22:55.29 This is shown here. So that was the paradigm for analyzing this type of reaction.
00:23:02.06 So you see, for example, in red you see Ras and down in green is the RasGAP domain (P120 GAP).
00:23:09.13 And what you see then, here...if you look very carefully
00:23:12.04 you see that there is a residue from GAP going into the active site.
00:23:17.14 You see that when it comes back...you see that right there.
00:23:21.01 There is an arginine residue pointing into the active site of Ras
00:23:26.04 and what is does is shown in the next slide.
00:23:29.06 First of all, in this slide it shows the aluminum fluoride really is the flat triangle
00:23:34.00 that sits between the leaving group and the nucleophillic water.
00:23:38.07 So there is a mimic of the transition state.
00:23:39.28 And you see then, also, what are the residues that mediate fast GTP hydrolysis.
00:23:45.17 So this would be the phosphate,
00:23:48.14 so we think now that if we take away aluminum fluoride and
00:23:52.06 think of what the real transition state would look like
00:23:55.06 you have the nucleophillic water and you have the transition state
00:23:58.29 where they are bound to the gamma phosphate
00:24:00.25 and the glutamine is fixating the water relative to the phosphate
00:24:05.07 by an acceptor and a hydrogen bond donor interaction.
00:24:09.17 And the arginine (what we call arginine finger) is first of all stabilizing
00:24:13.29 the position of that glutamine and it also delivers this positive charge
00:24:18.03 of the amino group to neutralize charges in the gamma phosphate.
00:24:22.28 So these are the two residues that are really crucial for the reaction;
00:24:27.08 glutamine 61 from Ras and the arginine finger from RasGAP.
00:24:32.05 And this already explains, for example, why mutations of glutamine 61 in oncogenic Ras
00:24:37.19 mess up the hydrolysis reaction because you can imagine if you have, for example,
00:24:42.09 a leucine or whatever here, if cannot do this kind of interactions here.
00:24:47.00 So any mutation of glutamine 61 is oncogenic because it is a direct catalytic residue.
00:24:53.16 So the structure also explains why mutations of glycine 12 are oncogenic.
00:24:58.22 And that also can be easily seen here on this slide.
00:25:02.04 You have glycine 12 which when mutated to any residue makes it an oncogene.
00:25:06.25 And then you see aluminum fluoride being this flat triangle.
00:25:11.08 You see glutamine 61 being there and you see the arginine finger down here.
00:25:16.09 And they're all very very close together indicated by these blue dashed lines
00:25:20.04 which are indicating that these are almost VanDerWaals distance.
00:25:24.04 And now if you mutate, for example, glycine to the smallest possible amino acid,
00:25:28.11 it would be alanine, you see that it would immediately mess up all these residues in the active site.
00:25:34.01 So for steric reasons, there can be no other residue
00:25:36.24 in the active site in the transition state except for glycine.
00:25:43.08 So the message from all of this was, obviously, that there's an arginine
00:25:49.00 that we call the arginine finger that pulls the trigger on the GTPase reaction.
00:25:54.00 Without it, it is very slow and with it, it becomes 10^5 fold faster.
00:26:01.01 We used also...so coming back to the question: what is now rate limiting in the whole process?
00:26:07.00 So the main regulating step of intrinsic GTP hydrolysis is the very slow chemical step, the slow hydrolysis.
00:26:14.29 But using a different technique we now find out that
00:26:17.25 there is another step that become partially rate limiting and I will show you that in a minute.
00:26:23.29 So we use time-resolved Fourier transform infrared spectroscopy.
00:26:28.10 And with that we can do almost atomic resolution.
00:26:33.11 We can look at atoms in the active site on a millisecond timescale.
00:26:37.27 So if you see an infrared spectrum of a protein, which has amide I and amide II bands,
00:26:47.20 which is not very structured information because its a mixture of information on
00:26:54.26 alpha helices, beta sheets and so on and not a great deal of detail can be taken from such a picture.
00:27:02.26 But what we do is we observe, for a reaction...let's say a protein goes from A to B
00:27:09.05 we observe the different spectrum which at least on this absorption scale
00:27:16.19 does not show much of a difference but if you look in more detail...
00:27:20.15 so you see the absorption there is 0.0 and the absorption down here is 0.02,
00:27:25.14 so we see very small but very reproducible changes
00:27:29.21 in the course of the reaction when A goes to B.
00:27:33.02 And we see negative peaks that mean A goes away and we see positive peaks if B comes up.
00:27:40.22 So we can observe things that are lost and things that come up
00:27:43.29 in the course of the reaction and we can follow these by FTIR.
00:27:49.25 And so if you do...so first of all we need to trigger the reaction at a particular time point
00:27:56.25 in order to observe second timescale structural changes.
00:28:03.04 And in order to do that we use, again, a different analog
00:28:07.00 and the different analog is caged GTP which was developed by Roger Goody
00:28:10.22 who is a colleague of mine in the Institute.
00:28:14.05 And this caged GTP is blocked on the gamma phosphate by what is called the cage group.
00:28:20.22 So it doesn't allow hydrolysis but now, with a flash of light, you can cleave that cage group off
00:28:26.21 and now you have Ras-GTP which can then hydrolyze to GDP and Pi.
00:28:31.19 If you do that with let's say Ras without GAP,
00:28:39.11 what you see is that you have a...for example, you see absorbance...
00:28:44.22 so any absorbance that you might know from infrared, shows an atomic vibration in the bonds.
00:28:51.09 For example, you see vibrations for the
00:28:53.15 alpha, the beta, the gamma, which decrease toward the end and become zero after 2 hours, 5 minutes.
00:29:00.29 So the difference spectrum (subtract the end spectrum from the starting spectrum)
00:29:06.16 and then you see only the changes that are occurring during the reaction.
00:29:11.23 And that is quite normal. There is a single exponential decay when Ras hydrolyzes GTP.
00:29:18.21 No big deal.
00:29:20.15 But what is interesting is that when you analyze the GAP mediated reaction
00:29:24.26 because now you don't see a single exponential decay, but you suddenly see intermediates appear.
00:29:30.14 You see peaks that come up and go down and the most important one is
00:29:35.04 indicated by this number 1113. This is the frequency for that particular change.
00:29:39.28 So you obviously have to analyze what is each band doing, what does it belongs to.
00:29:47.06 So what we're doing in order...we have developed these techniques
00:29:52.17 or our colleagues at the University of Bochum with whom we collaborate
00:29:56.03 they have developed techniques to find out what is each bandwidth,
00:30:01.01 what is each frequency...absorbance change...what is it due to.
00:30:03.28 For example, you find for GTP that there is absorbance change coming up with
00:30:10.13 a rate constant k2. So k1 is the photoisomerization, k2 is one reaction and k3 is the next one.
00:30:17.18 And if you do that, you get an increase and a decrease with k3.
00:30:21.12 And then there is an intermediate which is coming up at 1113.
00:30:26.12 It comes up with k2 and decays with k3.
00:30:30.01 There is free Pi coming with k3. From all of that we can obviously conclude
00:30:35.18 that there is a Pi intermediate with this absorption frequency here, 1113 cm^-1
00:30:44.02 which appears with the rate constant k2 and decays with a rate constant k3.
00:30:50.03 So in other words, the release of Pi is now becoming visible.
00:30:54.25 So we see hydrolysis when this Pi peak comes up.
00:30:58.21 And we see decay when it goes away.
00:31:01.01 And you see this here, for example, in real-time.
00:31:03.25 So the absorption change with the rate constant k2...you see the scheme here:
00:31:09.05 Ras when its in the on state goes to the Pi state and you see there is a protein bound Pi
00:31:17.13 that comes up and the Pi band goes down in the course of the reaction.
00:31:21.20 And this is repeated a number of times.
00:31:23.07 And you also so that there is an arginine finger that is coming in and out of the reaction chamber.
00:31:30.00 So we can follow not just protein bands,
00:31:32.16 we can follow the phosphate bands at atomic resolution on a milli second timescale.
00:31:40.26 And that tells us now, all together, that is the message from all of this,
00:31:46.01 that while you have a high activation energy for the intrinsic hydrolysis reaction of 92kJ/mol
00:31:54.01 you now seperate the reaction in two partial reactions,
00:31:57.16 which have a lower activation energy and that's why is makes it so much faster.
00:32:01.08 So you have the first activation energy for the cleavage reaction. which leads to Ras-GDP-Pi.
00:32:08.03 And you have the second step where you have release of Pi and now you have the product.
00:32:12.14 So that is, again, a general theme of enzymology;
00:32:15.12 that an enzyme catalyzed reaction lowers the activation energy
00:32:20.00 not just by lowering one reaction but also by subdividing it into partial steps
00:32:24.28 each of which has a different activation energy.
00:32:29.22 And here you see that this activation energy is 59 kcal/mol and 66 which means
00:32:36.04 that this is a bit higher and that's why this is the partially rate limiting step of the overall reaction.
00:32:45.12 So let me now give you...so that was Ras and how it leads to tumor formation
00:32:50.14 and we analyzed in detail how this function even on a biophysical level
00:32:55.02 but now let me come back to why certain GTPase reactions, when the don't work,
00:33:03.00 how they lead to different types of diseases.
00:33:05.20 Neurofibromatosis is one of them.
00:33:07.27 I already showed you that neurofibromine, the gene product of the gene is a Ras GAP.
00:33:15.09 And there is a disease called Type I Neurofibromatosis. Its what people have cafe au lait spots on the skin,
00:33:22.21 sometimes small tumors on the skin which can sometimes be rather large and very disfiguring
00:33:28.16 which are caused by mutation of deletion of neurofibromatosis gene.
00:33:35.21 And for example, when we worked on the mechanism of GTP hydrolysis, by GAP and NF1,
00:33:43.00 a colleague from the Charite Clinic in Berlin came to us and told us that he had
00:33:48.03 a female patient which dies at teh age of 35 and she has three sons indicated up there
00:33:55.20 which also have the disease. And he analyzed the blood of the patient and then the tumor itself.
00:34:04.22 He found out that there is a mutation in the sequence of neurofibrobromine
00:34:11.29 where the arginine is mutated to proline. You see that down there; arginine is mutated to proline.
00:34:17.13 And obviously I wouldn't tell you that, if it wasn't the catalytic arginine.
00:34:21.19 So it turns out they have a mutation in the catalytic arginine
00:34:25.23 and when you now do the GTPase reaction that I've shown you before,
00:34:30.09 you see that you have, with normal NF1 you have the blue curve here,
00:34:37.17 it means an increase in fluorescence and a decrease with hydrolysis
00:34:43.03 and now if you take this mutation, R to P, you have increase which means complex formation
00:34:48.03 but no hydrolysis and there is another patient that we've analyzed in the meantime
00:34:53.16 again arginine to anything else, Q in this particular case,
00:34:58.09 leads to a block of its ability to hydrolyze GTP on Ras.
00:35:02.17 So mutation of the essential arginine leads to the disease neurofibromatosis.
00:35:08.16 Also, I should point out that there are many other mutations in neurofibromine
00:35:13.22 which also cause the disease which phenotypically very different in many different patients.
00:35:20.26 Let me introduce you now to a different system that is
00:35:24.06 interesting both from a biochemical standpoint but also from a standpoint of a different disease
00:35:29.11 that I will come to in due course.
00:35:33.03 So what I will be talking about is a molecule called Rap which, obviously, has a cognate RasGAP.
00:35:40.16 And Rap is close homolog of Ras and the name derives from
00:35:44.05 the fact that it is highly homologous to Ras because Rap stands for Ras Proximate.
00:35:49.07 Even though it was considered to be a close homolog of Ras, it does something completely different.
00:35:55.14 It obviously has the same cycle between GDP and GTP. So it works as a molecular switch.
00:36:02.00 But it does not have to do with proliferation, as Ras, or differentiation, but rather
00:36:06.10 it is involved in integrin activation, or platelet activation or other things.
00:36:11.19 So the biology of Rap is completely different.
00:36:14.17 Why the biochemistry is so interesting is that Rap is the only homolog
00:36:22.12 or the only member of the Ras super family that doesn't have a glutamine in the position
00:36:27.18 where glutamine 61 of Ras is involved in GTP hydrolysis.
00:36:33.01 So, it misses the residue that we thought was absolutely crucial GTP hydrolysis and here, its not there.
00:36:39.18 And, obviously, the question is why that is so and how does RapGAP
00:36:42.25 then work on this system and how does it stimulate the reaction.
00:36:48.29 So first of all, let me introduce you to the RapGAPs.
00:36:52.24 There are indicated here five of them but there are more in the human genome
00:36:57.23 but these five RapGAPs all contain a domain that is highly homologous
00:37:04.02 which is indicated here by the light blue and the dark blue staining
00:37:08.09 where the light blue stuff is somewhat different to the dark blue.
00:37:12.05 And I will explain that when we look at the structure.
00:37:13.21 So again, the domain organization of all these five RapGAPs is somewhat different.
00:37:17.25 That means they're probably doing their job in a different biological context.
00:37:23.04 The reason its also interesting is that the dark blue homology region is
00:37:28.10 also conserved in a protein called Tuberin
00:37:30.17 which stands for a disease that I will be talking about in the end.
00:37:34.03 So that's why it was also interesting to look at that reaction.
00:37:37.06 And the third reason its interesting to look at that reaction is indicated in the next slide.
00:37:42.19 But before doing that, let me show you first of all that we do, again, a stopped-flow fluorescence assay.
00:37:48.23 We've developed a system where we can look at this reaction in a biochemical way
00:37:53.12 and analyze mutations and the speed of reaction and so on.
00:37:58.13 So here again, we have Rap-GTP interacting with RapGAP
00:38:03.20 You get a large, quick fluorescent increase which is due to complex formation
00:38:07.28 and a decrease due to GTP hydrolysis and dissociation of the product.
00:38:12.25 And it is all over, again, after one second.
00:38:16.03 whereas without RapGAP, the whole reaction would take hours.
00:38:19.25 So, in other words, we again have 10^5 stimulation of the reaction.
00:38:23.15 But the reason this is also an interesting system biochemically and mechanistically,
00:38:29.09 is that there are a number of conserved arginines in RapGAPs and obviously
00:38:35.00 we thought that one of them would be involved in providing an arginine finger to the system.
00:38:39.06 In fact, we mutated all of them to alanine
00:38:42.06 and none of them has any dramatic effect on activity as one can see here.
00:38:48.02 So the worst reaction is still close to .5/second.
00:38:52.02 So in other words, there is not a dramatic effect when you mutate an of the arginines
00:38:55.29 which makes it unlikely that there is an arginine finger involved in the reaction.
00:39:00.23 So the two residues, intrinsic glutamine and the arginine finger in trans
00:39:09.04 which is Ras and Rho and others make the important catalysis are not here in this system.
00:39:17.05 That means the whole chemistry must be totally different.
00:39:21.15 So we looked at the FTIR of the system again and
00:39:26.07 it shows basically the same features even thought their structures are somewhat different.
00:39:31.16 The basic features are that there is a Pi intermediate whose decrease is rate limiting
00:39:37.23 for the reaction and although its chemically somewhat different,
00:39:41.27 as the different absorption spectra show, it is, indeed, kinetically a most important intermediate.
00:39:50.11 But what was interesting, and we found in this particular case, in the RapGAP case,
00:39:56.10 and it was also found in other systems in the meantime,
00:39:58.27 is that the GTPase reaction is reversible,
00:40:01.29 which sounds sort of crazy because it's a downhill reaction.
00:40:05.06 If you take GDP and Pi and the whole system you would never ever create GTP.
00:40:09.22 But what happens is that if you analyze the reaction by
00:40:13.05 using, instead of normal water, O18 water, which is indicated by this black dot,
00:40:17.23 you would expect that in the reaction you get hydrolysis and then you get GDP and Pi
00:40:23.27 which has one phosphate labelled as O18 oxygen-rather, one oxygen labeled as O18 oxygen.
00:40:30.12 But instead of getting a Pi with one O18 oxygen, you get Pi with two O18 oxygens,
00:40:36.24 with three O18 oxygens, and with four O18 oxygens as analyzed here by a mass spec.
00:40:42.11 You see where you get all these four products, analyzed by mass spectroscopy.
00:40:49.27 So, how does this happen?
00:40:51.09 So it happens because on the protein you have this long-lived intermediate
00:40:56.00 with GDP and Pi, sitting in the active site before going into product that can also
00:41:03.10 do a back reaction to make GTP with one oxygen now on the gamma phosphate.
00:41:08.15 Now, if it reacts again with O18 water, you get incorporation of a second O18 into the product.
00:41:15.27 And if it happens again, a third one and a fourth one.
00:41:18.16 So although the overall reaction is downhill,
00:41:22.12 on the enzyme, you're getting a backwards reaction.
00:41:24.15 And you can never get it once Pi is released
00:41:28.21 because then the activation energy for the reverse reaction is too high.
00:41:33.18 So we solved the structure of RapGAP also.
00:41:37.06 This is a two domain structure where one domain, which you see now to the left,
00:41:44.07 is the catalytic domain--the dark blue region in the homology diagram that I showed you
00:41:49.01 and the light blue stuff is the dimerization domain which is not important for catalysis
00:41:53.08 but is there to dimerize the protein
00:41:56.18 for whatever reason that we really do not know.
00:42:00.13 So in analyzing the catalytic domain, obviously, you ask yourself:
00:42:03.10 What are the conserved residues and which of those play an important role in catalysis?
00:42:08.04 The purple helix that I've indicated here is the most highly conserved region.
00:42:12.00 So it's likely that residues on this helix are somehow involved in catalysis.
00:42:18.05 And indeed, you can mutate many of these and you see certain effects.
00:42:21.12 But the most dramatic effect happens when you mutate
00:42:25.01 N290, so an asparagine, sitting on this catalytic helix.
00:42:29.08 When you mutate that, you get the following result.
00:42:33.00 So we call it, by the way, the Asn-Thumb to make a difference to the Arginine Finger.
00:42:39.03 So it's an Asn-Thumb and you'll see in a minute why we call it an Asn-Thumb.
00:42:44.26 So, first of all, if you now look at the mutation that I talked about,
00:42:48.27 so if you take the N290A mutation and analyze the reaction by, again,
00:42:55.09 this fluorescent stopped-flow assay, you see that while wild type makes a complex
00:43:00.25 and then decays into product,
00:43:02.26 the mutation here makes a complex, this is even, by the way, tighter than in the wild type case,
00:43:09.15 but there's absolutely no hydrolysis.
00:43:11.07 The reaction goes on and on. It stays up there and never goes down.
00:43:14.21 Whereas if you make a mutation like H287 to alanine,
00:43:20.22 you see that there is no complex formation because the fluorescence stays down here.
00:43:24.04 So that reaction is dead because it cannot bind
00:43:27.08 but the red reaction is deal because we think that this is the important catalytic residue.
00:43:32.12 We think that from looking just at the biochemistry.
00:43:35.15 Obviously, to know what it does, we need again to solve the structure,
00:43:38.26 which we did. This is the Rap-RapGAP complex.
00:43:43.15 And you see here the red and the green stuff is RapGAP and the blue stuff is Rap.
00:43:49.20 And you see GTP. But what you also see if you look at it in detail,
00:43:54.02 is that there is, again, something pointing from the red catalytic domain of RapGAP
00:43:59.06 into the active site and that is an asparagine.
00:44:01.17 Obviously, it's the asparagine that I've talked about.
00:44:04.20 which pokes into the active site of Rap. That's why we call it the Asn-Thumb
00:44:08.28 in relation to the Arginine Finger in the other systems.
00:44:12.27 And if you look in detail at what happens at the active site,
00:44:17.02 and if you compare it to three other structures of Ras protein and their cognate GAPs,
00:44:26.19 taking Ran and RanGAP, Ras and RasGAP, and Rho and RhoGAP,
00:44:31.18 where Rho and RhoGAP are from Rittinger et al. and the other structures are from us,
00:44:35.14 you see that the other three, these three structures here,
00:44:40.12 have a glutamine pointing towards the catalytic water,
00:44:44.29 which attacks the gamma phosphate indicated here,
00:44:47.09 which is the aluminum fluoride,
00:44:50.01 a transition state which mimics the gamma phosphate that is transferred.
00:44:54.08 And the red structure here has, instead of the glutamine 61, has a threonine
00:45:00.19 which is pointing aways from the active site; has nothing to do with catalysis.
00:45:04.01 Instead, what you see is that the purple helix here inserts this asparagine
00:45:09.19 into the active site just exactly at the position where the others have the glutamine.
00:45:14.23 So the differences here: the three other structures have a glutamine, which is in cis
00:45:19.29 and Rap and RapGAP have an asparagine in trans
00:45:23.17 which does the same thing; namely, stabilizing the catalytic water.
00:45:30.04 And if you look at the surface of the protein,
00:45:34.01 this is Rap surface shown as a surface representation,
00:45:39.07 and added to it is just the helix from RapGAP.
00:45:43.09 It sits on the surface and inserts this asparagine into the active site.
00:45:48.13 So the gamma phosphate peaks out of that hole.
00:45:50.08 And all RapGAP really does is it has on the helix this Asn, it puts it into the active site,
00:45:57.17 and that alone seems to be, at least in a chemical sense,
00:46:03.03 responsible for stimulating the GTPase reaction by 10^5 fold
00:46:07.10 because if you mutate that thing, it still binds alright, but there's absolutely no hydrolysis.
00:46:11.22 So that makes us think about, again, the future of designing anti-Ras drugs.
00:46:18.02 If all it takes is to insert such a residue into the active site,
00:46:22.11 from a chemistry point of view it should be doable,
00:46:24.24 but we obviously have to develop molecules that bind in a correct way
00:46:28.12 onto the surface, which is not so easy and we are still thinking that we might be able to do it.
00:46:35.15 This is, again, coming back to that.
00:46:38.00 So as an approach to anti-cancer drug target,
00:46:42.03 let's find molecules that induce hydrolysis of oncogenic Ras.
00:46:46.20 And from what we have observed with RapGAP we think, we have hope that it can be done.
00:46:54.26 And the third reason why working with the Rap-RapGAP system is that
00:47:03.05 it's connected to a disease called Tuberous Sclerosis, a benign tumor.
00:47:06.19 People come in with hamartomas in many organs.
00:47:09.16 But the most obvious feature and what the name comes from is that
00:47:14.07 people have, when they do an NMR of the brain, they have
00:47:16.29 these sclerotic, tuberous sort of things in the brain that you see in the NMR of the brain.
00:47:24.17 The people have mutation in two proteins called Tuberin and Hamartin.
00:47:30.22 And Tuberin, as I showed you before, has high homology
00:47:34.22 to RapGAP in this dark blue region, this means the catalytic region.
00:47:39.07 And if you now look, for example, where patient mutations in Tuberous Sclerosis,
00:47:44.23 in Tuberin, where they are occurring.
00:47:47.00 And this is an alignment of different RapGAP sequences and they align with Tuberin sequence.
00:47:53.27 You see that most highly conserved region is, again,
00:47:57.09 this thing here (the red stuff) where you have the catalytic helix.
00:48:01.08 And one of the mutations in a patient with Tuberous Sclerosis
00:48:04.22 (and many actually have that mutation) is on
00:48:07.06 this asparagine that we know is the important catalytic residue.
00:48:11.08 So this is another version of what we have seen before;
00:48:14.21 an important catalytic residue is mutated in the disease Tuberous Sclerosis.
00:48:21.19 So let me, in the last five minutes, just give you another disease,
00:48:25.22 just to make sure that it's not just the Ras and Rap system,
00:48:28.28 that there are many diseases where the inability to hydrolyze GTP leads to very many different diseases.
00:48:35.07 So this is Retinitis pigmentosa.
00:48:38.28 People have pigments on the retina. That's what the name comes from.
00:48:42.28 And they lose vision, they lose peripheral vision.
00:48:47.00 So this is a normal person seeing that building and this is a patient with the disease
00:48:52.15 that loses more and more, progressively, his peripheral vision.
00:48:56.00 And then loses complete vision after the disease develops fully.
00:49:01.19 And there are a number of genes mutated in Retinitis pigmentosa
00:49:06.12 that are called RP12x and whatever.
00:49:09.28 And there is one form that is called RP2.
00:49:13.03 It's an X-linked disease where people have mutations in...a lot of point mutations in a protein.
00:49:21.16 We determined the structure of the protein, which looks like this.
00:49:24.24 This is a beta helix domain, and this is another domain.
00:49:27.20 And you see many of the mutations that you then analyze.
00:49:31.06 You find out that what you see is that most of the mutations
00:49:37.16 determine or mess up, probably, the structure because they are inside
00:49:41.24 the hydrophobic core of the protein and you would imagine that
00:49:45.04 they don't do anything to the catalysis or the interaction of this protein.
00:49:50.19 But there are a few mutations, indicated here, for example,
00:49:54.01 E138G or R118 to H, K, or C. So these residues point into solution.
00:50:03.27 So you would think that they do something important for interaction of the protein
00:50:08.20 or in something--that they are involved in what these proteins are normally doing.
00:50:14.20 So we solved, again, the structure of the complex between Arl3.
00:50:19.13 So we knew that RP2 interacts with Arl3, another Ras-like protein.
00:50:24.26 It's called Arl because it's an Arf Related Protein.
00:50:28.04 And we solved the structure of that. So you see this stuff here would be the RP2 in purple.
00:50:37.25 And you see in green would be Arl3 in a GTP bound form.
00:50:41.25 And what you see then, if you, again, look very closely,
00:50:44.26 you see that there is a residue from RP2, right here, which points into the active site of the Arl.
00:50:52.10 So one didn't know what RP2 was doing, but when we solved the structure,
00:50:55.17 we could immediately see that it smells like a GAP because
00:50:59.18 it puts an Arginine Finger into the active site of the other protein.
00:51:05.08 And if you look at the active site in detail you see that this would be GDP, again,
00:51:10.28 this is aluminum fluoride, this is glutamine from Arl3,
00:51:15.14 and these are three residues from RP2--Q116, R118, E138.
00:51:21.26 And, obviously, all of these residues are mutated in retinitis pigmentosa.
00:51:27.15 And you can see that the arginine is doing the same thing that you've seen now many times before.
00:51:32.22 And it is stabilized by these other residues and all three residues, when mutated,
00:51:38.10 mess up the GAP activity of RP2.
00:51:41.26 So here, the structure really told us what the protein is doing.
00:51:44.19 There was no idea about the function and the structure told us, exactly, that this is a GAP for Arl3.
00:51:53.17 So let me now come to the conclusions from what I have been telling you about.
00:52:00.12 First of all, maybe, conclusions about Ras itself because it's the most important oncogene.
00:52:06.14 It's an incomplete enzyme, it cannot hydrolyze GTP very fast.
00:52:11.09 But then comes RasGAP which stabilizes switch II and the glutamine 61,
00:52:16.10 which is the important structural catalytic element and it also supplies an arginine finger into the active site.
00:52:23.04 You have Gln1 mutation--any mutation of glutamine 61 is an oncogene
00:52:28.29 because it misses, then, the catalytic residue (the system).
00:52:33.00 You have glycine mutants which are sterically compromised to do GTP hydrolysis.
00:52:38.21 There is no way that the arginine finger
00:52:42.06 can go into its proper position when there is a mutation of glycine 12.
00:52:46.04 I also showed you that there is a strongly bound GDP-Pi intermediate.
00:52:50.02 And that Pi release in the system becomes rate-limiting.
00:52:54.28 Whereas without GAP, the chemical cleavage reaction is the rate-limiting step.
00:53:00.23 The second conclusion--more general to the Ras superfamily.
00:53:05.01 Ras proteins are all incomplete enzymes.
00:53:07.20 They all hydrolyze GTP very, very slowly.
00:53:09.27 They have cognate GAPs.
00:53:12.02 So each of the sub-family proteins and sometimes even proteins within the sub-family,
00:53:16.24 have a specific cognate GAP that are required for fast GTP hydrolysis, for catalysis.
00:53:22.17 Some GAPs supply an arginine finger
00:53:25.02 and I have shown you now many different examples--Ras and Ran and Rho.
00:53:30.15 RabGAPs deliver an agrinine and a Gln.
00:53:34.23 Some GAPs supply an Asn thumb.
00:53:37.00 I showed you the example of RapGAP and Tuberin probably does the same thing.
00:53:41.03 Pi release is very often rate-limiting.
00:53:44.10 And what is even more important, and that is my message for the whole talk,
00:53:47.19 is that the perturbed GTPase reaction is involved in a number of diseases:
00:53:53.24 cancer, neurofibromatosis, tuberous sclerosis,
00:53:58.03 retinitis pigmentosa and many more that I haven't talked about.
00:54:01.07 Thank you for your attention. But I would...
00:54:03.00 Before I stop, let me first thank the people that have done the work.
00:54:07.18 The oldest story that I have told you about is that of Ras and RasGAP,
00:54:11.22 which was done by three post-docs in the lab, Reza Ahmadian, Klaus Scheffzek and Robert Mittal.
00:54:17.23 The RapGAP story was done by the students Oli Daumke, Astrid Kramer,
00:54:23.14 Partha Chakrabarti and Andrea Scrima.
00:54:26.18 And the RP2 story was done by Stefan Veltel and Karin Kuhnel.
00:54:32.04 And a lot of the movies that I have been showing you are doen by Ingrid Vetter
00:54:36.20 who also runs my crystallography lab.
00:54:38.15 And she was very, very helpful in almost all the projects.
00:54:42.13 And we have...on the FTIR we have collaborations with
00:54:46.03 our colleagues at the University of Bochum which is Klaus Gerwert and Carsten Kotting.
00:54:51.03 Thank you for your attention.
00:54:52.10