Discovery of Chaperonin-Assisted Protein Folding
Transcript of Part 1: Discovery of Chaperonin-Assisted Protein Folding
00:00:07;07 Hi, I'm Art Horwich, from Yale School of Medicine and HHMI. 00:00:10;13 And I want to tell you a story about a discovery that has led to a really wonderful 00:00:17;29 what's now a fifteen to twenty year experience of experimentation that followed. 00:00:23;19 This discovery was really somewhat serendipitous, 00:00:27;09 and I think there was a lot of luck involved. 00:00:29;09 It was helped enormously by two wonderful collaborators, who have remained very close friends, 00:00:36;09 and constant contacts, Ming Cheng, who's now in Taiwan, 00:00:41;12 and Ulrich Hartl, who's now in Martinsried. 00:00:44;13 And so I'll just tell you the story as it occurred. 00:00:48;09 So, the story is about information transfer in the cell. 00:00:52;16 As we know, DNA is transcribed into RNA and is in turn translated into protein 00:00:57;12 and the step that's involved here is the final step of information transfer, 00:01:03;08 namely that of protein folding. 00:01:06;01 So, we know about machines that carry out transcription and translation, 00:01:10;05 and have very fine molecular details to those reactions for some years now, 00:01:16;22 increasingly fine. 00:01:19;11 But protein folding has remained somewhat mysterious 00:01:23;14 and it's really only during the last decade or two 00:01:25;18 that we've really started to gain some understanding about how this works in cells. 00:01:31;23 So, for those unfamiliar, I'd just point out that proteins are composed of twenty different amino acids, 00:01:38;17 and these strings of amino acids ultimately fold into characteristic three-dimensional structures 00:01:44;18 that carry out biological activities. 00:01:47;03 For example, they can be enzymes, cytoskeleton, hormones, hemoglobin, channels, 00:01:52;18 receptors, what have you. 00:01:54;15 Proteins are major effectors in the cell 00:01:57;21 and so their proper folding is crucial to them being biologically active, 00:02:03;08 because when they're effectively strands of spaghetti, 00:02:07;00 leaving a ribosome, they don't necessarily have any biological activity at that point. 00:02:13;08 They have to be folded correctly in order to be active. 00:02:16;24 And so the granddaddy experiment of protein folding 00:02:20;00 that I heard about when I was an undergraduate at Brown University 00:02:24;04 in 1972 was carried out by Christian Anfinsen, in fact, in the late 50's, 00:02:30;29 he received his Nobel Prize for this work in 1972, 00:02:35;27 a really astonishing experiment where he started with an enzyme, 00:02:39;23 ribonuclease, 100-odd amino acids with 4 disulfide bonds, 00:02:44;05 and unfolded that protein by treating it with denaturant, urea, 00:02:48;20 and reductant, mercaptoethanol. 00:02:50;28 That produced essentially a random coil version of the protein, 00:02:54;26 and he asked whether, when he removed the urea and reductant, 00:03:00;14 whether the protein could find its way back, on its on, in a test tube, 00:03:04;21 to the native active form, and astonishingly, it did so. 00:03:09;01 So, this proved that a polypeptide chain contains all of the information 00:03:14;12 for folding into the unique native structure in its primary amino acid sequence structure. 00:03:20;21 He also presumed that in fact this native state would lie at an energetic minimum. 00:03:27;17 And so, this astonishing experiment led people to believe that perhaps inside cells, 00:03:36;19 proteins also fold spontaneously. 00:03:39;13 But increasingly, there were instances where that didn't really happen, 00:03:44;03 both in test tube, where many proteins could also not be spontaneously refolded, 00:03:50;05 as in the Anfinsen experiment, with ribonuclease, 00:03:53;18 but inside cells when the biotech industry began to express mammalian proteins in E. coli, 00:03:58;28 for example, they also saw these terrible looking aggregates with no biological activity 00:04:04;27 and failed to recover their favorite proteins in a native, active form. 00:04:09;13 And this, once again, one saw, as in the test tube, 00:04:13;26 a misfolded, aggregated, inclusion body, as it were. 00:04:18;21 And so this suggested that something more might be required in the cell 00:04:23;24 to deal with these so-called, what would be kinetic problems of folding, 00:04:30;10 where proteins simply get it wrong and go off pathway 00:04:34;07 and misfold and aggregate. 00:04:36;00 And so we never thought we would have anything to do with this, 00:04:40;18 but we were busy in the late 80's studying mitochondrial protein import. 00:04:48;19 So, it turns out that most of the proteins of mitochondria are encoded in the nuclear genome, 00:04:54;22 the messages for these proteins are then translated on cytosolic polyribosomes 00:05:02;17 and an N-terminal targeting peptide in these proteins directs them to mitochondria. 00:05:10;06 And they're then imported through the mitochondrial membranes, 00:05:14;16 and the targeting peptide is then clipped and the mature-sized protein 00:05:19;00 then has to fold into its biologically active form. 00:05:22;29 Well, an important experiment was carried out in 1986 by Gottfried Schatz and his coworkers 00:05:28;18 that showed that a protein had to be unfolded in order to transit the mitochondrial membranes. 00:05:35;19 He used dihydrofolate reductase as a substrate in these experiments, 00:05:41;03 showing that if you lock DHFR into its native state, 00:05:44;22 by adding the ligand methotrexate, the protein couldn't go through the membranes, 00:05:49;19 whereas, if you now removed methotrexate or denatured the protein, 00:05:54;06 it would go through the membranes. 00:05:55;25 So, the question that arose to us, as we were studying mitochondrial protein import, 00:06:01;05 was, do proteins, upon reaching the mitochondrial matrix space, 00:06:07;03 refold spontaneously to their active form, as might be suggested by the Anfinsen experiments, 00:06:13;12 or could there be a machinery that was required to assist the protein to reach its native, active form. 00:06:20;27 And so we were in a position to actually test this 00:06:24;11 because we had a bank of temperature sensitive lethal yeast mutants 00:06:28;28 that contained a galactose regulated version of a mitochondrial protein, 00:06:34;20 ornithine transcarbamalase, a human protein that we had shown in yeast already 00:06:39;10 could be expressed and would go faithfully to its mitochondria, have its signal peptide cleaved, 00:06:44;12 and become biologically active and make OTC enzymatic activity. 00:06:48;05 And so our screen was basically to start with temperature sensitive lethal yeast mutants, 00:06:54;11 the idea being from Schatz and others in the field 00:06:57;03 that if you can't make new mitochondrial proteins and new mitochondria, 00:07:02;10 the cell basically stops its growth and ultimately, it's a lethal. 00:07:07;09 So, we start with a bank of lethal mutants, 00:07:10;12 programmed to express mitochondrial OTC behind a gal operon promoter, 00:07:16;03 and we proceeded to test each mutant individually for the ability to make OTC activity 00:07:21;28 after shift to non-permissive temperature and shift into galactose, 00:07:26;05 to turn on OTC expression under non-permissive conditions. 00:07:30;13 We then analyzed OTC activity; if it wasn't present, we would then Western blot 00:07:36;20 the particular mutant after shift to 37 C and induction of OTC to see what the fate of the OTC subunit was. 00:07:44;09 Was it present as a mature size protein, or was it still a precursor 00:07:48;11 that hadn't been recognized by mitochondria and so our idea was to dissect the mitochondrial import pathway. 00:07:54;26 Well, as we were asking this question about refolding inside the mitochondrial matrix, 00:08:01;28 we knew what we would be looking for in the form of a mutant 00:08:05;05 that affects refolding in the mitochondrial matrix space. 00:08:08;20 We would have mature size OTC subunits because their signal peptide would be cleaved 00:08:13;29 and they'd be inside the matrix space, but we would have no enzymatic activity, 00:08:18;26 because they would have failed to correctly fold, and in the case of OTC, 00:08:22;23 trimerize into the active form of the enzyme. 00:08:26;20 So, we started to look through our bank of mutants 00:08:31;12 and to our astonishment, almost right away, having asked this question, 00:08:35;23 we found a mutant in which exactly that happened. 00:08:38;20 So, OTC was our initial reporter enzyme and what we found was that OTC subunits entered, 00:08:45;20 they reached their mature size, 00:08:48;03 and then there was no biological activity at all. 00:08:50;28 So, this astonished us. 00:08:53;20 It seemed like heresy to have a result like this. 00:08:56;18 So, we though, uh huh, we better start to look at endogenous yeast proteins 00:09:01;27 and see what is happening to them. 00:09:04;04 So, we looked at this particular protein, the F1 beta subunit of ATPase, 00:09:09;05 of the F1 ATPase. 00:09:11;02 And we found that in fact, similarly, this beta subunit had failed to assemble 00:09:17;26 into this stalk like structure and the subunits themselves looked like had misfolded 00:09:24;06 in the mitochondrial matrix space. 00:09:27;11 We were in a bit of a quandary, because, again, it seemed like heresy, 00:09:33;06 and we couldn't really feel comfortable about the data at hand 00:09:38;04 and we were sort of pondering what proteins to look at next 00:09:42;12 when thankfully, the phone rang and it was Ulrich Hartl and Walter Neupert 00:09:47;03 calling from Munich to say, well, we understand that you guys have some mutants of yeast 00:09:52;23 that affect mitochondrial protein import 00:09:55;11 and we would love to get together with you and perhaps help on some of the biochemistry of those mutants, 00:10:01;12 if you'd have some interest. 00:10:03;08 So, at that point in history, my lab had four people in it, 00:10:07;18 and we weren't very well versed at fractionating yeast mitochondria, 00:10:12;29 whereas these were world experts, 00:10:15;09 so we were just delighted to interact and I went to Munich the next week 00:10:20;29 and presented some of the better described mutants that we had 00:10:25;12 that affected the mitochondrial signal peptide processing. 00:10:29;08 But I also, at the end of my talk, described this particular mutant, 00:10:33;13 and Walter and Ulrich were somewhat worried that what was happening in this mutant 00:10:39;07 was that proteins were coming in through this contact site 00:10:43;08 but weren't getting entirely into the mitochondrial matrix, 00:10:47;01 that they were poking their N-terminal signal peptides in and the signals were being cleaved, 00:10:52;04 but then the proteins weren't entirely entering 00:10:54;18 and of course, they wouldn't fold if they couldn't enter through the membranes. 00:10:58;12 And so, I said, well I'd be absolutely delighted to send you the mutant 00:11:02;29 and you can make mitochondria, I am sure, from this mutant, very effectively, 00:11:08;06 and analyze to see what's going on and in fact, as soon as I got back, I sent the mutant to Munich, 00:11:16;10 and two weeks later, Ulrich called up to say, excitedly, you're right, 00:11:22;05 the proteins are entirely in the mitochondrial matrix, I can add exogenous protease, 00:11:27;17 and it doesn't touch them, so it seems like they have gone into the matrix space 00:11:32;10 and they have failed to fold correctly. 00:11:35;06 So, this was a really wonderful time where there was a lot of back and forth, 00:11:39;16 with Ulrich and I crossing the Atlantic to visit each other and discuss the phenotype of this mutant. 00:11:47;07 This is us together at his parents' house, 00:11:51;03 here's his lovely wife Manajit, who is a biochemist herself 00:11:55;13 and has worked some on the GroEL system as well, in recent times, 00:11:59;13 a chaperonin of bacteria. 00:12:02;03 And maybe this is her telling us how this is all going to work out. 00:12:06;00 But in any case, we carried out together a set of further experiments 00:12:13;15 on a number of mitochondrial matrix proteins 00:12:16;12 to see what was happening to them. 00:12:18;18 And one of the most exciting results came with this particular protein, 00:12:22;12 the Rieske iron sulfur protein. 00:12:24;29 So, this is a protein that is imported to the matrix space and has to be cleaved twice 00:12:30;24 as a monomer, before it enters the inner mitochondrial membrane. 00:12:35;18 And when we looked at this iron sulfur protein, 00:12:38;19 we found it in either a completely uncleaved state 00:12:42;24 or in a once-cleaved state, as if the protein had not folded correctly 00:12:47;00 and had not undergone the proper cleavage that enables its correct biogenesis. 00:12:52;01 So, this was the first of a monomeric protein where we saw effects in this particular mutant. 00:12:58;02 We, at this point, we rescued the mutant with a yeast library. 00:13:04;12 So, we put on a bank of yeast clones and the gene that rescued was sequenced 00:13:12;07 by Ming Cheng during, I believe it was summer 1988. 00:13:17;06 And at this point, we knew that there was a modestly heat inducible protein 00:13:23;02 inside mitochondria that had been identified the year before by Richard Hallberg, 00:13:27;26 working in Iowa, in Tetrahymena. 00:13:30;11 And we had a feeling that Hallberg was analyzing the sequence of the corresponding homologue in yeast 00:13:38;06 with a view to doing further experiments in that system. 00:13:41;16 So, on a Saturday afternoon in that summer, Ming Cheng called up Richard Hallberg 00:13:48;12 with, to see whether her sequence of this gene would match the sequence that he had 00:13:54;18 of the homologue he was studying. 00:13:56;22 Well, I had gone to JFK airport to pick up Ulrich, 00:13:59;28 because the two of us were going to a Gordon conference on mitochondria the next day, 00:14:03;28 and my son came with us, it was many hours, but we got back to the, to our little beach house 00:14:10;18 at around 6 o'clock in the evening and there's Ming Cheng and the rest of the lab, 00:14:15;01 standing in our front yard and Ming's just jumping up and down holding a sequencing gel, 00:14:20;00 saying, "it's a dead match, it's a dead match!" and we said, well, what's the dead match? 00:14:25;07 She said, oh, I called up Hallberg, our protein corresponds to this ring assembly that he's studying! 00:14:31;10 And so, two and two came together in a really amazing moment and we had a wonderful evening of celebration, 00:14:38;00 but basically, what it amounted to was what we then called heat shock protein 60, 00:14:44;05 collectively, with Richard and Ulrich and Walter and everybody. 00:14:48;26 This modestly heat inducible protein 00:14:52;29 was actually assisting the folding of all these proteins being imported in the mitochondrial matrix. 00:14:59;16 And in fact, however, Hsp60 is essential at all temperatures, 00:15:06;08 so its assistance to polypeptide chain folding is required under normal conditions, 00:15:12;01 which embody, often, relatively high temperatures, 00:15:15;25 and very high solute concentrations in the mitochondrial matrix, 00:15:19;13 ergo this machine is essential under all conditions. 00:15:22;27 Now, shortly after this, Ulrich carried out a beautiful experiment with Joachim Ostermann 00:15:28;03 where they analyzed another monomeric protein, dihydrofolate reductase. 00:15:33;14 And what they could see was that when they put a signal peptide on DHFR to make it go into mitochondria, 00:15:39;04 much as in the Schatz experiments, 00:15:41;13 letting it go into mitochondria, it became associated with this complex, 00:15:46;24 with this ring assembly that Hallberg had first identified, the Hsp60 complex, 00:15:52;12 as long as ATP was relatively deprived from mitochondria. 00:15:56;15 If they added ATP back to this system, they now saw DHFR depart from this complex, 00:16:02;13 and appear as a native folded protein. 00:16:05;02 While the protein, I should say, was associated with Hsp60, 00:16:09;12 it seemed to be non-native, very protease susceptible, 00:16:12;21 recognizable by antibodies that specifically saw non-native forms of DHFR, 00:16:17;25 and so these experiments, with Rieske iron sulfur protein, with DHFR, 00:16:23;01 convinced us that Hsp60 was some sort of a folding machine in the mitochondrial matrix. 00:16:28;21 And of course, there's some generality to this because when we saw the sequence of Hsp60, 00:16:36;13 we knew already that it was going to be homologous, and is homologous, 00:16:41;14 to the GroEL protein in the bacterial cytoplasm, essential in that context, 00:16:47;05 and had been shown to be involved, or had been implicated 00:16:50;23 early as being involved in lambda phage head assembly, 00:16:53;29 now this idea had to move to lambda phage protein folding. 00:16:58;15 There was also a protein inside of chloroplasts called the rubisco binding protein, 00:17:04;04 identified by John Ellis as being required for assembly of rubisco, 00:17:08;17 which is a CO2 fixing enzyme, in fact, the most abundant protein in the biosphere. 00:17:14;01 This now, had to go to the level of thinking that rubisco binding protein 00:17:19;07 was actually mediating protein folding of rubisco subunits. 00:17:23;29 And finally, shortly thereafter, working together, Ulrich and I, with Jonathan Trent, 00:17:29;18 identified a heat shot protein in thermophilic archaebacteria 00:17:34;02 that we called thermophilic factor 55, another double ring assembly, 00:17:38;24 that is highly induced when you take these normally heat loving bacteria 00:17:43;12 and take them to an even higher temperature, 00:17:45;21 that turns out to look just like a chaperonin, and it also could turn over ATP 00:17:51;21 and behave just like these other ring assemblies. 00:17:54;24 And, in fact, when we analyze this subunit, we found a homology 00:17:59;08 to a protein in the eukaryotic cytosol called TCP1, or CCT for short, 00:18:05;00 this being a member of the eukaryotic cytosolic chaperonin 1 subunit of the 8 subunit rings 00:18:12;07 that comprise that chaperonin, essential for folding actin, tubulin, beta propeller proteins, 00:18:18;02 and many other proteins. 00:18:19;25 So, the generality of these ring assemblies in mediating protein folding to the native state 00:18:25;00 seemed to be getting established. 00:18:27;17 And so finally, mechanistically, 00:18:30;08 20 years later, we understand at least how the bacterial chaperonin system works 00:18:35;23 in pretty good detail now, with much biochemistry, X-ray crystallography, 00:18:40;28 NMR studies and a lot of EM conducted on this particular machinery, the GroEL machine 00:18:47;19 and its cooperating component, GroES, and the bottom line in terms of mechanism is 00:18:53;25 that an open ring of GroEL with all this hydrophobic surface specifically binds non-native proteins 00:19:00;03 through their exposure of hydrophobic surface that will be buried to the interior 00:19:04;22 in the native state, thus GroEL specifically recognizes non-native proteins 00:19:10;01 and once it's bound to protein, when GroES now encapsulates the same ring 00:19:15;23 that's bound by polypeptide, the polypeptide is released into this hydrophilic folding chamber, 00:19:22;18 all this blue here, signifies electrostatic residues, 00:19:26;21 and the polypeptide now folds in isolation, in a sequestered environment, 00:19:31;17 where it simply can't aggregate with any other protein and it basically uses Anfinson's principles, 00:19:38;14 its primary structure, to find its way to the native state without the fear of aggregation or going off-pathway 00:19:46;19 and being lost. So, in general, molecular chaperones, 00:19:51;23 if one wants to talk about all of the classes of these components, 00:19:55;27 they basically recognize exposed hydrophobic surfaces 00:20:01;18 that are proffered when a protein misfolds, or during biogenesis, 00:20:06;12 and the idea of the chaperone is to bind these surfaces and prevent them 00:20:12;03 from self-associating with each other in multi-molecular aggregates. 00:20:16;13 The idea of all of the chaperones is to prevent aggregation 00:20:20;19 by binding these surfaces in various geometric contexts. 00:20:24;05 The remarkable thing about the chaperonins, however, 00:20:27;00 that's really unique, is that they use nucleotide to release the polypeptide 00:20:32;13 into a sequestered folding chamber where the polypeptide really has no chance of aggregation 00:20:38;08 and has a chance to fold using its own primary structure 00:20:42;11 to the native state, one of nature's really most beautiful machines, 00:20:46;22 used for this final step of information transfer. 00:20:49;23 So, thanks very much for listening to this lecture, I hope you enjoyed it. 00:20:54;05 Thank you.