Telomeres and Human Disease

Part 1: Telomeres and Human Disease

00:00:15.04 Hello.
00:00:16.22 My name is Titia de Lange
00:00:18.08 and I work at The Rockefeller University in New York,
00:00:22.10 and in this lecture I'm going to
00:00:25.06 discuss how telomeres solve the end-protection problem,
00:00:28.29 how they protect the ends of chromosomes.
00:00:32.12 So, telomeres are made up of
00:00:35.01 a simple DNA sequence --
00:00:37.23 it's the TTAGGG repeats.
00:00:41.06 So, all the telomeres in human cells
00:00:44.03 and in mouse cells have exactly the same sequence,
00:00:47.18 this short repeat, six nucleotide repeat,
00:00:51.00 repeated over and over and over again.
00:00:56.14 Thousands of repeats are present
00:00:58.10 at every chromosome end.
00:01:00.05 In human cells, a chromosome end
00:01:02.08 can be as long as 10 kb (kilobases)
00:01:04.02 of telomeric DNA,
00:01:05.11 in mouse cells as much as 50 kb.
00:01:08.09 And then at the very end
00:01:10.04 there is a special structure to the telomere,
00:01:12.15 where the G-rich strand,
00:01:15.06 which sticks behind me, here,
00:01:18.02 sticks out 3' to the end of the chromosome,
00:01:20.00 and this region is about 50-300 nucleotides (nt),
00:01:24.08 and it's very important for telomere function. 1 This DNA is maintained by the enzyme telomerase,
00:01:31.26 that needs to maintain DNA
00:01:34.17 because telomeric DNA is lost with cell divisions.
00:01:36.26 So, telomerase is a reverse transcriptase
00:01:39.05 that can recognize the end of the chromosome
00:01:42.12 and add, using its template RNA sequences,
00:01:45.18 to the chromosome end.
00:01:46.29 We're not going to talk about telomerase
00:01:48.29 -- that's part of the previous lecture, part 1.
00:01:52.24 Today, we're going to talk about
00:01:54.19 what these telomeric DNA...
00:01:57.01 these telomeres are actually doing.
00:01:59.25 They solve the end-protection problem,
00:02:02.11 which is a problem that originates
00:02:05.03 from the fact that mammalian cells,
00:02:07.22 like all eukaryotic cells,
00:02:09.29 have a very good system to detect DNA breaks in their genome.
00:02:13.24 So, when a DNA break occurs,
00:02:17.03 cells respond to this.
00:02:18.18 They can find the break and activate
00:02:20.19 two powerful transducing kinases,
00:02:22.08 the ATM kinase and the ATR kinase.
00:02:25.23 And these kinases will enforce cell cycle arrest,
00:02:30.00 sometimes even cell death,
00:02:31.26 if the damage is very extensive,
00:02:35.05 and they will orchestrate the repair of the DNA break.
00:02:38.19 And of course the telomere
00:02:40.23 looks just like a broken piece of DNA,
00:02:43.04 and yet these kinases
00:02:45.24 should not become active at the chromosome end,
00:02:48.10 otherwise the cells would never divide,
00:02:50.11 and cells might try to repair their telomeres,
00:02:53.16 which would lead to a mess in the genome.
00:02:57.12 So, telomeres must be ignored
00:02:59.04 by the DNA damage response.
00:03:02.01 How does this work?
00:03:04.06 Well, the DNA is inert -- it doesn't do too much --
00:03:08.01 but it has this one great ability,
00:03:10.11 and that is to recruit
00:03:12.19 a telomere-specific protein complex
00:03:14.09 that we've called shelterin.
00:03:17.00 It's a six subunit complex,
00:03:19.02 of which two subunits
00:03:21.14 bind to the double-stranded part of the telomere,
00:03:22.27 and a third subunit also binds telomeric DNA,
00:03:27.25 but in single-stranded form.
00:03:30.01 We'll go into this in some detail.
00:03:32.02 So, this complex was discovered in my lab,
00:03:35.13 not as a complex
00:03:37.24 -- we just found one telomeric protein, TRF1.
00:03:42.00 We found TRF1 as we were looking for proteins
00:03:45.04 that could bind to telomeric DNA,
00:03:47.10 because we suspected that
00:03:49.27 the end-protection problem was solved by a protein complex.
00:03:53.14 After we had identified TRF1, purified it,
00:03:57.12 sequenced the protein, cloned the gene,
00:04:00.27 and showed it to be at telomeres,
00:04:03.07 we started looking for other telomeric proteins,
00:04:04.25 and we found one protein
00:04:06.24 that's pretty similar to TRF1,
00:04:08.25 TRF2,
00:04:10.17 that we found in a database,
00:04:12.16 and found that it's also at telomeres.
00:04:17.00 And then we found Rap1, this little blue guy, here,
00:04:18.29 as an interacting factor of TRF2.
00:04:22.19 TIN2, the yellow protein in the middle,
00:04:26.01 was discovered by Judy Campisi.
00:04:28.20 We then found TPP1
00:04:30.08 as an interacting factor of TIN2.
00:04:34.01 And finally, POT1,
00:04:37.13 the single-stranded DNA binding protein,
00:04:38.27 was found by Tom Cech's lab,
00:04:40.06 based on homology to a telomeric protein
00:04:42.25 in unicellular organisms,
00:04:46.03 the only protein in shelterin
00:04:48.16 that is conserved enough to be
00:04:50.02 identified by homology.
00:04:52.29 So, let's look at these proteins in a little bit more detail.
00:04:57.03 Let's first look at TRF1 and TRF2,
00:04:58.28 the two proteins that bind to
00:05:01.24 double-stranded DNA.
00:05:02.25 They're very similar and they both
00:05:05.04 bind double-stranded DNA
00:05:06.19 through a C-terminally located
00:05:09.10 Myb domain, a three-helix bundle
00:05:11.24 that recognizes the telomeric sequence.
00:05:14.11 The Myb domains
00:05:17.14 cannot bind to telomeric DNA by themselves,
00:05:19.18 they can't form a stable complex.
00:05:21.22 They can only bind to telomeres
00:05:23.26 when they're in a dimeric state,
00:05:25.29 and so both TRF1 and TRF2 are dimers,
00:05:30.15 as I tried to draw here,
00:05:33.11 and they're dimerizing through these
00:05:36.07 very large dimerization domains.
00:05:39.14 So, these two proteins
00:05:42.12 are very similar in architecture,
00:05:43.24 but they're in fact very different,
00:05:45.17 they interact with different proteins,
00:05:48.12 and they have very different function at telomeres,
00:05:50.18 as you will see.
00:05:51.28 Now, let's look at the other DNA binding protein
00:05:55.09 in shelterin, POT1,
00:05:57.09 which binds to the single-stranded DNA.
00:06:00.08 POT1 is also a fairly simple protein.
00:06:02.19 It has a C-terminus
00:06:04.16 with which it binds to TPP1,
00:06:06.20 and TPP1 recruits POT1
00:06:10.10 onto the telomere by binding to TIN2,
00:06:12.04 which binds to TRF1 and TRF2.
00:06:15.06 And then POT1 has in its N-terminus,
00:06:19.06 two DNA binding domains
00:06:20.27 that are very good at binding single-stranded DNA.
00:06:23.06 This structure, here, from Tom Cech's lab
00:06:24.28 shows these two OB folds
00:06:27.03 that are protein folds
00:06:30.13 that can bind nucleic acids and oligosaccharides,
00:06:34.01 binding specifically to the TTAGGG repeats.
00:06:38.08 They can bind this DNA both at a 3' end,
00:06:41.24 as drawn here,
00:06:43.27 but also all along the single-stranded telomeric DNA.
00:06:49.00 So, shelterin is very abundant,
00:06:51.03 it's there at telomeres in all cells,
00:06:53.18 it's ubiquitously expressed,
00:06:55.20 and there are so many copies of shelterin
00:06:58.17 per telomere
00:07:00.03 that there is enough TRF1 and TRF2
00:07:02.07 to cover all the telomeric DNA.
00:07:05.06 There's also enough POT1
00:07:07.20 to bind all the single-stranded telomeric DNA at the telomere.
00:07:10.21 In fact, there is ten times more POT1
00:07:16.11 than there is room for POT1 on the single-stranded DNA.
00:07:19.01 But, TPP1 and POT1
00:07:21.12 are less abundant than the other shelterin components,
00:07:24.15 and although I don't draw this in my slides...
00:07:27.14 so, there are shelterins that don't have TPP1 and POT1
00:07:31.01 on them.
00:07:32.16 Shelterin is highly conserved,
00:07:34.21 and because of this we can
00:07:37.07 use mice to study shelterin function,
00:07:41.03 which is a very convenient and good genetic system.
00:07:44.14 And telomeres also contain nucleosomes,
00:07:46.20 that I don't draw,
00:07:48.00 and shelterin-associated proteins.
00:07:51.18 So, this little protein complex, six subunits,
00:07:56.27 has a formidable task.
00:07:58.29 It needs to block two different
00:08:03.01 signal transduction pathways
00:08:04.20 that could become inappropriately activated at telomeres,
00:08:07.25 and it has to block DNA repair.
00:08:11.03 So, let's first look at the ATM kinase pathway.
00:08:14.05 This is a pathway that recognizes DNA ends.
00:08:17.23 It recognizes DNA ends
00:08:19.20 through the agency of
00:08:22.20 a trimeric complex called MRN,
00:08:24.17 which stably binds to the end,
00:08:26.06 recruits the ATM kinase,
00:08:27.29 and the ATM kinase then goes on to
00:08:31.22 phosphorylate proteins in the chromatin nearby,
00:08:34.28 which leads to the accumulation of DNA damage foci.
00:08:38.19 So, the chromatin at a DNA break is altered,
00:08:42.06 over many kbs (kilobases), sometimes hundreds of kbs,
00:08:46.15 and is marked as a site of DNA damage.
00:08:49.11 This will guide repair to that break
00:08:54.24 and control the repair reaction,
00:08:56.16 so that the appropriate repair takes place,
00:08:59.23 depending on the cell cycle.
00:09:01.07 I want to point to one protein, here,
00:09:05.06 53BP1, which is a component of these
00:09:08.12 DNA damage foci
00:09:09.25 that we'll use as a marker
00:09:12.10 for ATM and ATR kinase signaling in this talk.
00:09:15.04 So, the ATM kinase, also,
00:09:17.27 when it becomes active,
00:09:19.10 can enforce cell cycle arrest.
00:09:20.18 It does so through phosphorylation
00:09:23.15 and activation of Chk2,
00:09:26.01 which then activates p53
00:09:28.23 and inactivates Cdc25 phosphatase,
00:09:31.29 making it impossible for cells
00:09:33.26 to transit from G1 into S phase,
00:09:36.28 when DNA replication occurs.
00:09:39.09 So, shelterin blocks this pathway.
00:09:43.01 It also block the ATR kinase pathway,
00:09:45.24 a pathway that is initiated
00:09:48.27 when cells have single-stranded DNA at a break.
00:09:52.23 ATR kinase is activated
00:09:56.20 through the agency of RPA.
00:09:59.12 RPA is a very abundant DNA binding protein
00:10:03.08 that binds to single-stranded DNA,
00:10:06.10 regardless of the sequence.
00:10:08.03 Once RPA is at one of these breaks
00:10:11.02 where a single-stranded DNA is formed,
00:10:13.07 it joins 9-1-1,
00:10:16.02 a ring-shaped molecule that loads onto
00:10:18.27 the double-stranded/single-stranded transition,
00:10:21.05 and these proteins then
00:10:24.06 recruit and activate the ATR kinase,
00:10:26.09 which then does pretty much the same thing
00:10:29.03 that the ATM kinase does.
00:10:30.22 It leads to the formation of DNA damage foci
00:10:34.29 that have 53BP1 in them,
00:10:37.28 and it impinges on cell cycle progression
00:10:40.15 through its own effector kinase, Chk1.
00:10:43.15 So, shelterin needs to repress this pathway.
00:10:48.15 And then shelterin needs to block DNA repair.
00:10:52.13 And there are various forms of DNA repair
00:10:54.16 that it has to take care of.
00:10:56.07 It needs to bock non-homologous end joining,
00:10:58.03 a simple ligation reaction
00:11:00.16 that I'll go into in some detail,
00:11:04.11 it needs to block homologous recombination at the telomere,
00:11:07.02 and it also needs to prevent excessive resection
00:11:11.12 of the telomere.
00:11:12.18 There is a 5' end, here, at the telomere,
00:11:15.06 where nucleases that
00:11:19.07 like to resect DNA ends could act,
00:11:21.03 and this could of course take away
00:11:23.01 much of the telomeric DNA.
00:11:25.11 So, how do we figure out
00:11:28.18 how shelterin does this complicated task,
00:11:31.25 taking care of so many different pathways,
00:11:35.10 a simple six subunit complex?
00:11:38.27 Well, we've chosen to study shelterin
00:11:42.13 in mouse embryo fibroblasts,
00:11:45.03 because that allows us to use the genetics of mice
00:11:49.08 to dissect the pathways.
00:11:51.17 Shelterin, in general, is essential,
00:11:53.02 so you can't make a mouse with shelterin
00:11:55.12 -- it's dead --
00:11:56.27 but you can use conditional deletion
00:12:00.01 in mouse embryo fibroblasts
00:12:01.17 to remove components of the complex,
00:12:04.13 and the way we do this is by
00:12:06.29 using the Cre recombinase.
00:12:09.06 The Cre recombinase can
00:12:11.14 cut specific sites and make a deletion in the genome.
00:12:15.22 So, what we do is we take shelterin-encoding,
00:12:19.27 subunit-encoding genes,
00:12:22.03 we put the Cre recombinase recognition sites
00:12:25.24 in the genes,
00:12:27.02 and then we have cells that
00:12:29.20 contain what we call floxed alleles,
00:12:31.28 that contain these Cre sites in the gene,
00:12:34.19 and we can then take these cells
00:12:36.12 and infect them with Cre recombinase,
00:12:38.28 and the deletion will occur.
00:12:41.05 The gene will be removed.
00:12:42.22 Now, when we do these experiments,
00:12:44.29 we always do them in cells that don't have p53.
00:12:49.06 The reason is that
00:12:51.05 when you remove shelterin from telomeres,
00:12:52.19 you get a DNA damage response,
00:12:55.00 and in mouse cells this will lead to arrest or apoptosis.
00:13:01.03 We don't want that
00:13:03.17 because then we can't study what's going on at the telomeres anymore.
00:13:06.11 But in the mouse, if you remove p53,
00:13:08.14 they have no idea that ATM or ATR is active,
00:13:12.06 and they keep dividing,
00:13:14.16 and so this will give us another opportunity
00:13:16.29 to see what the phenotypes are at the telomere.
00:13:20.00 Now, it turns out that nature
00:13:22.04 has designed shelterin for these kinds of experiments,
00:13:25.17 it couldn't be better.
00:13:26.26 Now, the reason is that
00:13:29.24 the complex is extremely compartmentalized,
00:13:31.24 so, as shown here,
00:13:34.25 you can delete TRF2 from the cells
00:13:38.14 and TRF2 then disappears from the telomeres,
00:13:42.06 but miraculously the other proteins stay.
00:13:46.24 So, TRF1, on its own,
00:13:48.26 -- it's very similar to TRF2 --
00:13:50.28 is enough to recruit TIN2, TPP1, and POT1
00:13:54.25 to the telomere.
00:13:56.00 The only protein that falls off the telomere,
00:13:58.07 together with TRF2, is Rap1.
00:14:00.18 So, whenever we see a phenotype with TRF2,
00:14:04.06 we wonder...
00:14:07.08 is this due to the loss of TRF2 or due to the loss of Rap1?
00:14:10.19 And what we do then is we delete Rap1,
00:14:12.20 and if we don't see the phenotype
00:14:14.21 we know it was a function of TRF2,
00:14:16.14 and it turns out that Rap1
00:14:18.07 doesn't do very much at telomeres.
00:14:20.21 So, what happens when you delete TRF2 from cells.
00:14:25.27 Well, there's a really striking phenotype that occurs,
00:14:30.03 that you can see in
00:14:32.06 metaphase [spreads] of chromosomes.
00:14:34.04 So, this is a metaphase
00:14:35.28 of normal, healthy mouse cells
00:14:38.18 in which the telomeres have a fully functional shelterin complex.
00:14:44.00 And what you can see is that
00:14:46.10 these little mouse chromosomes,
00:14:48.07 false-colored in red, here,
00:14:49.17 with a DNA stain,
00:14:51.05 all have little telomeres highlighted by telomeric FISH
00:14:54.17 at their ends.
00:14:56.11 Now, let’s see what happens
00:14:58.20 when we delete TRF2.
00:15:02.05 The consequences of TRF2 [deletion]
00:15:04.25 are quite dramatic.
00:15:06.28 You can see in this slide that
00:15:10.02 most of the telomeres are no longer at the end of a chromosome.
00:15:13.25 They are inside long stretches of DNA.
00:15:18.03 What has happened here is that
00:15:19.29 the cells have mistaken
00:15:21.27 their natural chromosome ends
00:15:24.00 for DNA breaks,
00:15:25.15 gone into a panic and started repairing these breaks,
00:15:29.20 and they do it in the only they can
00:15:34.03 in this situation, that is,
00:15:35.14 by fusing one telomere to another.
00:15:37.13 And this is non-homologous end joining
00:15:39.28 that is taking place.
00:15:42.22 Non-homologous end joining is
00:15:44.11 a very simple pathway
00:15:46.20 that's very active in mammalian cells,
00:15:49.04 and it repairs DNA breaks.
00:15:51.18 It starts out by the loading of a ring-shaped molecule
00:15:56.02 onto the DNA end,
00:15:57.23 the Ku70/80 heterodimer.
00:15:59.17 This is a structure of the Ku70/80 heterodimer on DNA,
00:16:04.21 done by Jonathan Goldberg's lab,
00:16:07.29 and you can see that this molecule
00:16:10.25 has an opening that's
00:16:13.05 just big enough to allow DNA to go through.
00:16:17.00 Ku70/80 can only get on DNA
00:16:20.05 when there is an end,
00:16:22.19 so it normally will not load
00:16:25.22 onto our chromosomal material,
00:16:27.24 only when there is a DNA break can it load,
00:16:30.12 and once it's loaded,
00:16:32.27 it can bring two DNA ends together
00:16:35.24 -- this is called synapsis,
00:16:38.14 it needs a lot of additional factors to do this,
00:16:40.29 which I won't go into --
00:16:43.04 and then DNA ligase IV can come in
00:16:45.00 and ligate the two ends together.
00:16:47.28 Ligase IV also has accessory factors that I won't mention.
00:16:52.10 So, this is the pathway
00:16:54.26 that gives rise to the telomere fusions
00:16:57.05 after deletion of TRF2.
00:16:59.02 How do we know this?
00:17:00.29 Well, what we did is
00:17:02.19 we suspected it was non-homologous end joining,
00:17:05.00 so what we did is we looked at
00:17:07.28 the genetic requirement for Ku70/80 and ligase IV
00:17:11.17 in the formation of these fused telomeres,
00:17:15.20 so we deleted TRF2 from cells
00:17:18.06 that either lacked Ku70/80
00:17:20.28 or lacked ligase IV,
00:17:23.01 and then the fusions didn't occur.
00:17:25.05 So, we know TRF2 itself,
00:17:29.01 not TRF1, TIN2, TPP1, POT1, and also not Rap1,
00:17:34.16 is dedicated to the repression of non-homologous end joining.
00:17:39.22 But it does a second thing.
00:17:41.22 It blocks the activation of the ATM kinase.
00:17:45.27 How do we know this?
00:17:48.06 Well, when we delete TRF2 from cells,
00:17:53.00 we see that now, all of a sudden,
00:17:56.23 rapidly, the telomeres
00:18:00.27 contain the 53BP1 protein.
00:18:03.03 This is the protein that I pointed out
00:18:05.25 we can use as a marker for ATM signaling,
00:18:09.00 and you can see, here, that 53BP1
00:18:12.03 is now at telomeres,
00:18:13.26 colocalizing,
00:18:16.10 the green IF with the purple telomeric FISH,
00:18:20.07 inside nuclei.
00:18:22.21 And this only occurs when TRF2 is deleted with Cre.
00:18:26.22 In a non-Cre situation,
00:18:28.27 when TRF2 is there at the telomere,
00:18:31.09 53BP1 is not found at telomeres.
00:18:35.29 We know this is the ATM kinase pathway,
00:18:38.26 again from genetic experiments,
00:18:41.10 in which we deleted TRF2 from cells
00:18:43.28 that did not have the ATM kinase.
00:18:47.03 So, we used genetics to make such cells,
00:18:49.12 and it we delete TRF2
00:18:52.02 from an ATM kinase-null setting,
00:18:54.07 there is no DNA damage response.
00:18:58.19 So, TRF2, then, does two things.
00:19:02.08 It's dedicated to the repression of two pathways
00:19:05.06 -- on the one hand,
00:19:08.00 the non-homologous end joining pathway,
00:19:09.29 on the other hand,
00:19:11.24 the ATM kinase pathway.
00:19:13.10 And, interestingly, both pathways
00:19:15.12 are initiated by an end-loading factor
00:19:19.11 -- Ku70/80, the ring,
00:19:21.28 in the case of non-homologous end joining,
00:19:23.25 and the Mre11 complex, MRN,
00:19:25.26 in the kinase of ATM signaling.
00:19:28.29 So, how does this work?
00:19:31.04 How does this simple DNA binding protein,
00:19:34.07 TRF2,
00:19:36.02 prevent these end-loading proteins
00:19:38.23 from getting at the telomere?
00:19:40.25 Well, the clue to this came from work
00:19:45.20 that was initially done by Jack Griffith,
00:19:48.09 in collaboration with us,
00:19:50.09 who looked at the structure of telomeric DNA
00:19:53.08 in mammalian cells,
00:19:55.12 and he found that the telomeres are not linear,
00:19:58.25 but can attain a t-loop structure.
00:20:02.07 This is an EM, Kleinschmidt spreading,
00:20:05.20 of purified, protein-free telomeric DNA from HeLa cells,
00:20:11.02 and you can see that this telomere
00:20:13.24 has a very large lariat structure.
00:20:17.16 Here's another EM analysis,
00:20:20.08 in this case done by Woodcock's lab,
00:20:22.02 in which they analyzed chicken telomeric chromatin.
00:20:25.22 They purified the whole chromatin
00:20:29.14 and could see these t-loop structures.
00:20:31.28 And in fact telomeres in protozoa,
00:20:35.05 and in plants, and in many organisms
00:20:38.17 are now known to be in a t-loop structure.
00:20:42.02 So, what is the t-loop, really?
00:20:44.15 Well, we looked into the details of this structure
00:20:47.16 and we surmised that
00:20:51.00 what is really going on here is that
00:20:53.11 the 3' overhang of the telomere strand
00:20:59.10 invades into some part in the double-stranded part
00:21:02.09 of the telomere,
00:21:04.02 forming a displacement loop,
00:21:05.24 single-stranded DNA
00:21:07.20 because the telomere overhang goes in,
00:21:09.10 so the other strand has to come up,
00:21:12.06 and forming this structure.
00:21:16.06 The point of insertion of the telomeric overhang
00:21:19.26 seems to be variable
00:21:21.13 -- sometimes the loops are big,
00:21:23.19 sometimes the loops are small --
00:21:25.27 but what is constant here is
00:21:28.17 this strand invasion event,
00:21:30.13 and it seems, then, that
00:21:33.13 telomeres are trying to hide the chromosome end
00:21:36.09 inside the telomeric DNA.
00:21:40.15 Now, this links to TRF2
00:21:42.25 because we also found, in collaboration with Jack Griffith early on,
00:21:47.00 that TRF2, in vitro, by itself,
00:21:50.08 can form a t-loop,
00:21:52.02 or a t-loop-like structure.
00:21:54.09 So, we incubated baculovirus-derived TRF2
00:21:57.20 with substrates that looked like telomeric DNA
00:22:01.26 and then examined the product by EM,
00:22:04.27 and as you can see here there are little loops
00:22:07.16 with TRF2 accumulated at the base.
00:22:11.29 More recently, Eric Gilson's lab
00:22:15.01 showed that TRF2 has the interesting ability
00:22:18.02 to wrap telomeric DNA
00:22:20.23 around its very large TRFH domain.
00:22:25.02 This wrapping of DNA often
00:22:29.00 creates a topological change
00:22:31.02 that may melt nearby DNA,
00:22:33.09 and this could of course be very helpful
00:22:35.08 for a protein that is trying to
00:22:37.07 strand-invade the 3' overhang.
00:22:39.13 So, from these kinds of data,
00:22:42.17 the model evolved that TRF2
00:22:45.26 represses non-homologous end joining
00:22:48.09 and ATM signaling
00:22:49.22 by making this t-loop structure,
00:22:52.03 because in the t-loop
00:22:54.17 the telomere end would not be accessible
00:22:56.02 to the two initiating factors,
00:22:59.15 Ku70/80 and the MRN complex.
00:23:02.19 So, how would one test this?
00:23:05.00 It's not so easy to test,
00:23:07.12 but the idea would be simple.
00:23:09.21 The prediction of the model is that
00:23:12.08 if you delete TRF2 from mouse embryo fibroblasts,
00:23:15.02 the loop structure should be lost.
00:23:18.10 So, Jack Griffith and I tried to do this experiment
00:23:21.14 for a very long time
00:23:24.04 and failed, because the EM technique
00:23:25.27 is not really well-suited for quantitative analysis.
00:23:31.20 Fortunately, a new technique
00:23:35.04 was developed by Xiaowei Zhuang, at Harvard,
00:23:39.04 that allowed us to image telomeres
00:23:41.25 based on the telomeric DNA.
00:23:44.08 This is super-resolution STORM imaging,
00:23:46.16 there is an iBioSeminar by Xiaowei on this,
00:23:50.22 and I advise you to watch this.
00:23:54.05 This work with did
00:23:56.18 in collaboration with Xiaowei Zhuang,
00:23:57.24 where we used super-resolution STORM imaging
00:24:00.19 to examine the structure of telomeres
00:24:03.04 in cells with and without TRF2.
00:24:06.15 So, Ylli Doksani worked with the Zhuang lab,
00:24:10.10 examined telomere structure by STORM imaging
00:24:14.07 in cells with TRF2,
00:24:15.24 and you can see here some examples
00:24:18.13 of the t-loops he found.
00:24:20.19 Again, there is always a loop and a tail,
00:24:24.00 but the ratio of the DNA in the loop or the tail
00:24:27.18 can be very variable,
00:24:29.00 allowing us to conclude
00:24:31.12 that it's not so much the loop
00:24:33.16 that's important for the protective activity,
00:24:38.08 but it's the strand invasion event.
00:24:41.03 So then, Ylli wanted to image
00:24:43.23 cells that don't have TRF2,
00:24:45.10 but here he had a problem,
00:24:47.05 because if you delete TRF2,
00:24:48.24 of course all the telomeres fuse together,
00:24:50.28 and if you don't see t-loops at fused telomeres,
00:24:54.00 it's not a very informative experiment.
00:24:57.17 So, we used a genetic trick
00:25:01.02 to delete TRF2 from ATM-null cells,
00:25:03.19 in which case there is no DNA damage signaling
00:25:06.19 at the telomeres, as I showed you,
00:25:08.20 and also no fusion.
00:25:10.15 And this is what Ylli saw.
00:25:13.03 The telomeres in these cells
00:25:15.08 that lacked TRF2,
00:25:17.03 but had all the other shelterin components there,
00:25:20.13 had a linear structure
00:25:22.05 -- the t-loop was lost.
00:25:23.18 He then repeated the experiments
00:25:25.16 in TRF1 knockout cells,
00:25:27.14 POT1, etc., etc. knockout cells,
00:25:30.06 and the t-loops always remained.
00:25:33.11 Only when TRF2 is removed do the t-loops resolve.
00:25:37.26 And so this is fitting with the model.
00:25:40.18 The idea is simply that without TRF2
00:25:43.09 the telomere can no longer
00:25:45.03 maintain the t-loop state.
00:25:46.22 Now it becomes a linear molecule
00:25:48.26 that allows Ku70/80 to load onto the end,
00:25:52.10 and you get non-homologous end joining.
00:25:55.26 Furthermore, the MRN complex
00:25:58.06 can no see this telomere
00:26:00.01 for what it really is,
00:26:01.11 a DNA end,
00:26:03.12 load onto the end and activate the ATM kinase,
00:26:06.11 leading to the 53BP1 foci
00:26:09.10 at the telomere.
00:26:11.07 So, this takes care of
00:26:14.26 one major problem in telomeres
00:26:16.16 -- the ATM signaling pathway
00:26:19.03 and non-homologous end joining.
00:26:22.02 What about the ATR kinase?
00:26:24.04 Again, this is a very powerful kinase
00:26:27.06 that could kill the cells if it was activated
00:26:29.01 at the telomeres.
00:26:30.20 It is repressed by another shelterin component
00:26:35.01 -- it's repressed by POT1.
00:26:37.00 So, POT1, the single-stranded DNA binding component
00:26:42.02 in shelterin,
00:26:43.09 is the one that represses
00:26:45.26 the ATR signaling pathway.
00:26:47.00 How do we know this?
00:26:48.02 Again, we used genetics in the mouse
00:26:51.00 and looked at,
00:26:52.09 what is going on at telomeres
00:26:54.08 when POT1 is there?
00:26:55.23 These are cells with an intact shelterin complex
00:27:00.17 and there are no 53BP1 foci at telomeres
00:27:03.17 -- there's no DNA damage response.
00:27:06.10 Now we delete POT1
00:27:08.15 and immediately you see that
00:27:11.12 there is a lot of BP1 at...
00:27:14.13 colocalizing with the telomeric DNA.
00:27:17.17 This is the ATR kinase pathway
00:27:20.07 that has become active
00:27:22.22 at these chromosome ends.
00:27:24.10 How do we know that?
00:27:26.05 Well, we use the genetics in mouse, again,
00:27:28.24 and now deleted POT1 from cells
00:27:31.18 that do not have ATR,
00:27:33.17 and if you do that,
00:27:35.05 without the ATR kinase,
00:27:38.11 no 53BP1 foci at the telomere.
00:27:42.22 So, this told us that
00:27:45.13 while TRF2 is dedicated to the ATM pathway,
00:27:48.08 POT1 is dedicated to the ATR pathway.
00:27:52.25 I should say, parenthetically,
00:27:55.10 that all these experiments were done in mouse cells
00:27:58.05 where there are actually two POT1 proteins,
00:28:00.21 but for simplicity I refer to them
00:28:04.29 as POT1.
00:28:08.27 How does POT1 repress the ATR signaling pathway?
00:28:12.10 Well, we got a clue about the mechanism
00:28:14.28 from looking at the RPA protein
00:28:18.10 that I told you is important to activate the ATR kinase.
00:28:24.00 So, when you delete POT1 from telomeres,
00:28:28.21 you all of a sudden see RPA accumulate
00:28:31.17 at the telomeres, right there.
00:28:33.05 Normally, no RPA at telomeres.
00:28:36.06 So, when POT1 is removed,
00:28:38.05 RPA can bind to the single-stranded telomeric DNA,
00:28:41.29 and this made us think that
00:28:45.06 what really is going on here
00:28:47.17 is that POT1,
00:28:49.15 bound to the telomeric DNA,
00:28:50.22 is normally excluding RPA
00:28:53.08 from the telomere,
00:28:54.14 and since RPA can't bind
00:28:58.02 it can't activate ATR.
00:29:01.25 So, the idea is that
00:29:03.19 the whole single-stranded part of the telomere, here,
00:29:06.02 is covered by POT1,
00:29:07.23 and then RPA can't bind.
00:29:10.25 But there's a problem with this model,
00:29:13.18 and that is that RPA is extremely abundant.
00:29:17.26 So, this is a competition model,
00:29:19.21 where POT1 wants to win
00:29:23.15 in this competition from RPA,
00:29:24.24 but it's dealing with a protein that's about
00:29:27.12 200-fold more abundant than itself
00:29:29.21 -- not a good situation.
00:29:32.12 Furthermore, we found that
00:29:34.26 RPA and POT1 have exactly the same affinity
00:29:37.16 for the telomeric DNA,
00:29:39.11 so POT1 doesn't have an advantage at that level either.
00:29:43.20 So, how then can this
00:29:46.12 simple single-stranded DNA binding protein
00:29:49.15 prevent RPA from binding?
00:29:51.01 Well, the trick here is that
00:29:53.25 POT1 is tethered.
00:29:56.06 So, POT1, after all,
00:29:59.18 is tethered onto the telomere
00:30:02.15 by TRF1 and TRF2,
00:30:04.07 so that if one of the POT1s
00:30:07.10 on the single-stranded DNA
00:30:08.25 lets go of the DNA,
00:30:11.24 first of all the DNA is held there
00:30:13.18 by other POT1 proteins,
00:30:15.25 and secondly the POT1 that has released
00:30:19.20 is still sitting there,
00:30:21.20 because it's tethered to TRF1 and TRF2
00:30:24.08 through TPP1 and TIN2.
00:30:26.02 So it can rebind immediately,
00:30:28.23 whereas RPA, if it lets go of the DNA,
00:30:31.04 like any DNA binding protein,
00:30:34.19 will... will diffuse away, and it has to start all over.
00:30:38.10 So, the prediction of this model is that
00:30:41.19 the tethering of POT1 to the rest of shelterin
00:30:45.08 is critical for the ability of POT1
00:30:47.05 to protect the telomere from ATR signaling,
00:30:49.18 and that's a testable prediction.
00:30:52.00 So, what we did is we created situations
00:30:56.02 where not just POT1 is deleted,
00:30:59.08 in which case the ATR kinase becomes active,
00:31:02.07 but we leave POT1 in the cell
00:31:04.15 and ruin its tethering.
00:31:08.02 We can do that by deleting TPP1.
00:31:09.29 We found that ATR becomes active.
00:31:13.21 Or we can do it by deleting TIN2.
00:31:16.09 We found the ATR becomes active.
00:31:19.04 And probably the best experiment of this set
00:31:21.24 was an experiment in which
00:31:25.03 we took the TIN2-knockout cells
00:31:27.14 and put back into them a mutant form of TIN2
00:31:30.17 that's completely normal, can do everything,
00:31:33.17 except for binding to TPP1.
00:31:36.14 So, these cells now have
00:31:39.25 all shelterin components there, and yet, 1because POT1 is not tethered to the telomere,
00:31:45.25 ATR becomes active.
00:31:49.15 So, the tethering of POT1
00:31:53.03 and the repression of ATR signaling
00:31:56.11 is critical for telomeres,
00:31:58.09 otherwise this signaling pathway becomes active,
00:32:00.20 and it's not only important for telomeres
00:32:03.12 that are in an open configuration, as drawn here.
00:32:07.07 Telomeres may be open when the t-loop opens up,
00:32:11.04 they may be in linear structure
00:32:12.25 when the t-loop opens up during DNA replication,
00:32:16.13 but the same repression of ATR signaling
00:32:19.15 is important when telomeres are in the t-loop configuration,
00:32:23.08 because in the t-loop configuration,
00:32:25.15 at the very base of the t-loop,
00:32:28.12 you have a structure that is a perfectly fine substrate
00:32:32.14 for ATR signaling.
00:32:34.19 It's a structure that has
00:32:37.06 single-stranded DNA in a d-loop (displacement-loop),
00:32:38.27 where RPA can load,
00:32:40.24 and it has a 5' end here
00:32:42.21 and a single-stranded/double-stranded junction,
00:32:45.09 which allows 9-1-1 to load.
00:32:47.27 So, everything here is set up
00:32:51.00 to activate ATR and also, therefore,
00:32:53.13 in this setting, POT1 needs to be able to repress the ATR signaling.
00:32:58.14 T-loops, themselves, don't to anything
00:33:02.01 to repress ATR signaling.
00:33:04.16 When we delete POT1,
00:33:06.00 the t-loops are still there,
00:33:07.29 but the absence of this single-stranded DNA binding protein
00:33:11.17 at the base of the t-loop
00:33:13.24 creates the trouble.
00:33:15.21 So, in summary, then,
00:33:18.15 I discussed with you
00:33:20.09 how telomeres solve
00:33:22.29 the basic tenets of the end-protection problem.
00:33:26.17 Telomeres are made up of a long stretch
00:33:28.25 with very boring repeat sequences.
00:33:31.14 This is important because it
00:33:33.10 provides many binding sites for shelterin,
00:33:35.26 this six-subunit complex
00:33:37.23 that's recruited onto the telomeric repeats.
00:33:41.04 And shelterin then has two main factors
00:33:44.19 that do the job.
00:33:46.13 On the one hand, there's TRF2.
00:33:48.11 It creates the t-loop structure
00:33:50.10 and this represses non-homologous end joining
00:33:53.01 and ATM signaling.
00:33:55.01 On the other hand, there is POT1,
00:33:57.04 which, by its ability to associate
00:34:00.09 with the single-stranded DNA,
00:34:02.17 can repress ATR signaling
00:34:05.21 by keeping RPA off the telomere.
00:34:08.07 Now, I've discussed this in some detail,
00:34:11.05 but I must say there is much else about telomeres
00:34:14.04 that I haven't been able to discuss,
00:34:16.04 and I can't leave you with the idea
00:34:18.22 that this is everything,
00:34:20.05 so I'm just quickly going to
00:34:23.00 list the many interesting subjects in telomeres
00:34:25.26 that I haven't had time to go into.
00:34:29.13 I haven't talked to you about TRF1.
00:34:32.02 It's the first telomeric protein we found.
00:34:34.14 It turns out not to be involved
00:34:36.27 in the end-protection problem;
00:34:38.10 instead, it's a protein that helps the DNA replication fork
00:34:42.20 go through the telomeric DNA,
00:34:44.28 which is extremely difficult for the replication fork
00:34:48.02 to negotiate.
00:34:49.21 I haven't told you how shelterin
00:34:51.15 creates the 3' overhang
00:34:53.16 that's so important to make the t-loop structure.
00:34:56.21 I haven't told you how shelterin
00:34:59.27 represses excessive 3'... 5' end resection,
00:35:02.13 which threatens telomeres,
00:35:04.00 or about homologous recombination,
00:35:06.12 or about the real danger of t-loops,
00:35:10.23 which have a structure at their base,
00:35:13.26 at the tail of the loop's junction,
00:35:16.12 that can resolved by Holiday junction resolvases,
00:35:20.01 that could cleave off the whole loop if they act.
00:35:24.25 I haven't told you about alternative non-homologous end joining
00:35:26.25 and PARP activation at telomeres.
00:35:28.22 And, lastly, the whole story I told you about telomeres
00:35:32.19 is a story about DNA and protein,
00:35:35.17 but it leaves out one component.
00:35:37.23 It turns out telomeres are transcribed into
00:35:40.28 a long non-coding RNA, lncRNA,
00:35:43.06 called TERRA, that undoubtedly
00:35:45.25 has an important function at telomeres,
00:35:47.07 we just don't know yet how it works.
00:35:50.04 So, I want to say that
00:35:53.07 I have discussed work from my lab
00:35:56.02 and I haven't had the opportunity to mention
00:36:00.12 everybody that contributed to this work,
00:36:03.07 and so if you're interested in the people from our lab,
00:36:06.17 please go to our website,
00:36:09.07 delangelab at rockefeller.edu,
00:36:13.06 where you can find all that information
00:36:15.20 about the current members of the lab and the alumni,
00:36:18.05 and it's a fantastic group of people
00:36:20.19 that I've been fortunate enough to work with
00:36:22.17 over the years.
00:36:23.23 It also lists our collaborators,
00:36:25.12 it lists our protocols
00:36:27.20 -- if you're interested in doing these kinds of experiments.
00:36:33.18 It lists information on how to get our reagents from the labs,
00:36:36.00 where to get our constructs -- on Addgene --
00:36:39.16 our mice are in Jackson Lab,
00:36:41.21 the MEF cell lines will be available at the ATCC,
00:36:46.05 so you have a lot to find in our website.
00:36:48.13 We do our work at the Rockefeller University,
00:36:51.16 which is a great place to work,
00:36:53.00 and the Upper East Side of Manhattan
00:36:55.06 is a beautiful campus and
00:36:57.01 a wonderful, collegial environment.
00:36:59.19 And we're currently supported by
00:37:02.02 the funding agencies mentioned here.
00:37:05.10 So, thanks very much for listening
00:37:07.28 and it was a pleasure to do this.

Part 2: How Telomeres Solve the End-protection Problem

Videos in this Talk

Part 1: Telomeres and Human Disease
Audience:
- Student
- Researcher
- Educators of H. School / Intro Undergrad
- Educators of Adv. Undergrad / Grad
Part 2: How Telomeres Solve the End-protection Problem
Audience:
- Student
- Researcher
- Educators of H. School / Intro Undergrad
- Educators of Adv. Undergrad / Grad

Speaker: Titia de Lange

Total Duration: 1:14:51

Recorded: August 2016

All Talks in Genetics and Gene Regulation

Talk Overview

In this seminar, Dr. Titia de Lange gives an overview of telomeres, the protective repeats at the ends of chromosomes. Because telomeres resemble double stranded DNA (dsDNA) breaks, they could be recognized by the DNA damage response (DDR) pathway leading to cell cycle arrest and genome instability. De Lange discusses the proteins that protect telomeres, how telomeres are maintained by telomerase, mechanisms of telomere shortening, and the role of telomere shortening in cancer and several inherited human diseases.

In her second lecture, de Lange describes the function of shelterin, the protein complex that binds to telomeres and prevents the recognition and activation of the DDR pathway. Important to this process are two shelterin proteins, TRF2 and POT1. TRF2 and POT1 block the activation of the DDR-initiating kinases, ATM and ATR, and prevent inappropriate DNA repair. De Lange explains that TRF2 blocks ATM signaling via the formation of an altered telomere structure called the t-loop. She also discusses the mechanism by which POT1 represses the activation of the ATR kinase.

Speaker Bio

Titia de Lange

Dr. Titia de Lange is the Leon Hess Professor, an American Cancer Society Research Professor, and the Director of the Anderson Center for Cancer Research at Rockefeller University. After getting her doctorate from the University of Amsterdam, she pursued her postdoctoral training with Dr. Harold Varmus at UCSF. Here, she isolated human telomeric DNA and… Continue Reading