Top
SCIENCE. STORIES. IMPACT.

iBiology

Bringing the World's Best Biology to You

Developing Single Molecule Technologies to Study Nanomachines

Part 1: Developing Single Molecule Technologies to Study Nanomachines

00:07.2 Hi.
00:09.1 My name is Taekjip Ha.
00:11.0 I'm a professor of biophysics and biomedical engineering
00:13.2 at Johns Hopkins University.
00:16.1 I am also an investigator with the Howard Hughes Medical Institute.
00:21.3 Today, I'm going to tell you
00:23.2 how single molecule measurement technologies
00:26.1 are changing the way we study
00:28.2 and understand nature's nano-machines.
00:34.2 You may have heard people say that
00:36.2 proteins can be viewed as nano-machines.
00:40.0 Why?
00:41.2 First, because they are unimaginably small.
00:45.2 They are only a few nanometers across.
00:48.1 How small is a nanometer?
00:50.1 Well, the smallest object that human eyes can see
00:54.2 is about 80 [micrometers] across,
00:58.0 which is a human hair.
01:00.0 That means that you can put 20,000 proteins
01:03.2 across a single human hair.
01:05.2 So, it's really, really small and you need special technologies to be able to
01:10.0 view and measure them.
01:12.2 Proteins are also called nano-machines
01:14.2 because they can perform jobs
01:17.1 that human-made machines can do.
01:19.2 For example, they can convert chemical energy
01:22.0 into mechanical energy
01:24.1 to power proteins along DNA or other molecules,
01:30.2 or you can store energy into a protein
01:33.1 just like you can wind up a wind-up toy
01:36.1 to store mechanical energy.
01:41.3 Here is an example:
01:44.2 a virus can package its genomic material,
01:47.2 the DNA,
01:49.1 into a very small volume using
01:51.2 a protein nano-machine
01:54.1 called DNA packaging motor,
01:56.0 and this packaging motor can use the energy coming from ATP consumption
02:00.1 to move the DNA into the capsid
02:03.3 one base pair at a time,
02:06.1 and it can do this until it can build
02:10.1 a pressure that matches the pressure
02:12.2 inside a champagne bottle.
02:15.3 So it's an amazing machine that can,
02:17.2 you know,
02:19.2 cram a large piece of DNA into a very, very small volume.
02:24.0 Here's another example of a nano-machine.
02:26.1 This is a protein called kinesin
02:29.1 carrying a cargo inside a living cell,
02:32.1 and kinesin is one of the molecular motors
02:34.3 that can move on the cellular highways directionally,
02:39.3 using the energy coming from ATP.
02:43.2 Just like cars are moving on the highway
02:46.0 using gasoline as a fuel,
02:48.1 these molecules can move on microtubules
02:51.3 using ATP as the fuel.
02:58.2 So, we wanted to study the mechanism
03:02.2 of the movement of these motor proteins.
03:05.2 One way of achieving this is to
03:08.1 label the motor
03:11.0 using a single fluorescent molecule
03:13.2 and measure the movement of a motor protein
03:16.0 at the single molecule level
03:17.2 using a microscope in a dark room,
03:20.0 as shown here.
03:22.3 The labeling can be done by
03:25.1 attaching a small fluorescent molecule
03:28.1 to the site of your interest.
03:30.1 In this case, let's say a molecular motor
03:33.1 called myosin V
03:34.3 is moving on the actin filament,
03:36.2 and it's known that the center of mass of the motor
03:40.1 moves on the track
03:42.1 with a step size of 37 nanometers.
03:45.1 Here, you can put a fluorescent molecule
03:48.0 on the foot of the motor
03:50.1 and you can image that single molecule
03:52.1 as a function of time in a dark room
03:54.3 using a microscope,
03:56.2 a very sensitive microscope
03:59.1 using CCD-based detection.
04:02.2 Now, if you image a single molecule
04:07.0 using your camera,
04:08.2 you get a pixelated image shown on the left,
04:13.0 and here the pixel size is about 80 nanometers,
04:16.1 but the motor protein itself is
04:19.3 much, much smaller than even a single pixel size,
04:23.1 about 10 times smaller,
04:27.0 but because of a process called light diffraction
04:29.1 #NAME?
04:31.2 based on Heisenberg's Uncertainty Principle --
04:34.2 when you take an image of a single molecule,
04:37.1 you know, using even the best optical microscope,
04:40.2 this molecule appears to be much bigger,
04:42.3 about 250 nanometers across,
04:45.1 which is about the half of the wavelengths of the light
04:50.0 used to image the molecule.
04:51.2 However, if you have a good enough
04:54.0 signal-to-noise ratio in your imaging,
04:56.1 then you can fit this pixelated image
05:00.2 using a 2-dimensional Gaussian function
05:03.0 and then find the center
05:05.0 with a much better precision
05:07.0 than the width of the image itself.
05:10.1 In fact, mathematics tell you that
05:13.3 you can do it down to about 1 nanometer precision,
05:16.2 and in practice it has been shown that,
05:20.1 yes, you can even approach that level
05:22.1 by, you know, going down to
05:24.1 about 1.5 nanometer precision.
05:26.0 So, it's amazing that you
05:28.0 can actually follow the movements of these motor proteins
05:30.0 in the lab,
05:33.3 down to single nanometer precision.
05:40.0 So, here's an example
05:42.1 where we imaged a single myosin V motor protein
05:45.2 moving on the track
05:47.3 and, as you can see,
05:53.1 as time goes on,
05:56.1 this spot, coming from a single molecule,
05:58.2 moves diagonally across to the top left,
06:01.2 and you can visualize individual stepping motions
06:05.1 as discrete movements,
06:07.1 you can find the center,
06:09.1 and then plot that position as a function of time,
06:13.0 and you can see beautiful steps,
06:14.2 and the step size turns out to be 74 nanometers,
06:18.1 which is twice the step size of the center of mass.
06:21.3 And this actually showed, conclusively,
06:24.1 for the first time,
06:25.2 that the step size of...
06:27.2 actually, that the molecule,
06:29.2 instead of crawling on the track,
06:31.3 it walks like a human walker, a bipedal walker,
06:34.2 and it proved a model called the hand-over-hand model
06:38.1 in the most definitive way.
06:42.1 So, to recap what I just told you,
06:44.1 what we did with the myosin V molecule
06:46.3 was just this:
06:49.0 you know, if you want to follow the movement
06:51.1 of your favorite soccer player on the field,
06:53.0 you can put a fluorescent molecule
06:56.0 on a body part of the player
07:00.0 and then follow its movements as a function of time.
07:07.1 But this measurement alone does not tell you why
07:10.2 -- you know, my favorite soccer player, here, Messi --
07:12.2 is such a great player, not a merely average player.
07:16.1 Soccer players, as you can imagine,
07:18.3 have to undergo multiple conformational changes...
07:22.0 actually, conformation is a jargon that
07:24.2 biophysicists use to denote the shape change of a molecule,
07:27.3 so a soccer player has to change his body shape
07:30.2 and posture,
07:32.2 you know, repeatedly
07:35.1 to perform the amazing action
07:37.3 of kicking and moving the ball.
07:39.1 And you are blind to the conformational changes
07:43.3 if you just follow one spot
07:47.1 in a body part.
07:50.1 In addition...
07:52.1 so, that's basically what you will be seeing,
07:55.0 you know, just one spot,
07:57.0 you know, nothing else.
08:00.0 So, we have, actually, another technology
08:01.2 that can be used to address this issue.
08:04.3 It's called FRET for
08:07.2 fluorescence resonance energy transfer.
08:10.0 Here, instead of just using one dye,
08:12.1 we use two dye molecules,
08:13.3 and one is called "donor", in green,
08:16.1 and the other is called the "acceptor", in red.
08:20.1 Now, when the two dyes are very far away from each other,
08:22.2 you use a laser light to excite the green molecule
08:25.3 and then you get emission from the green molecule,
08:28.1 so you get green photons from the molecule.
08:33.0 But if the acceptor moves closer to the donor,
08:37.0 then there is more and more
08:39.3 energy being transferred from the green molecule
08:41.3 to the red molecule.
08:43.1 As a result,
08:44.3 now you get photons coming from the red molecule,
08:47.2 of a red color.
08:49.3 So, by measuring the relative intensities
08:53.1 of the green and red molecules,
08:57.0 you can deduce how far they are away from each other
08:58.3 and if you know where the dyes
09:01.2 are attached to the molecule,
09:03.2 you can, for example,
09:05.2 distinguish between different conformations of a molecule.
09:08.3 Now, this technique
09:11.1 has a very strong distance dependence
09:13.2 and it can go from 100% energy transfer to 0
09:17.3 over a very few nanometer length scale change,
09:21.1 between 3-8 nanometers,
09:23.1 which is about the size of many proteins
09:25.3 and many other important biological molecules.
09:28.1 So it's the perfect technique
09:30.2 to measure the shape changes of
09:34.2 all kinds of important nano-machines inside our body.
09:39.2 Here's an example.
09:41.3 Imagine there's a protein
09:43.2 -- it's a butterfly that I'm using as an example --
09:45.3 you can put two dye molecules, green and red,
09:48.1 on two different locations,
09:50.1 so that in the closed conformation
09:53.3 energy transfer is efficient,
09:55.2 you get many red emissions,
09:57.2 whereas in the open conformation over there,
09:59.2 you get mainly green emission,
10:01.0 because the energy transfer efficiency is not high.
10:08.1 Here's another illustration
10:10.1 using the world famous using the world famous rap singer
10:13.2 Psy's Gangnam Style as an example.
10:15.2 So, he's going between different conformations,
10:17.3 The Sunglasses,
10:19.2 The Cowboy,
10:20.3 The Horseback Riding,
10:22.0 and The Finish,
10:23.1 and if you have dyes on the two hands,
10:25.2 then you can go in between intermediate distance,
10:28.2 a large distance,
10:30.0 and small distance,
10:31.3 and different, you know,
10:33.3 intensities of two different dyes
10:36.0 can be measured as a function of time.
10:39.3 Now, unlike what the cartoon suggests,
10:44.0 in a microscope you cannot see these two dye molecules
10:46.2 as two distinct spots because they are so close to each other,
10:50.0 much closer than the diffraction limit,
10:53.0 you know, 250 nanometer resolution of the microscopy
10:57.1 that we are using,
10:58.2 so they appear as a mixture of green and red,
11:02.2 in yellow, as shown here.
11:04.2 And when you move to the next conformation,
11:07.0 then it appears as green,
11:09.2 and then as red and so on,
11:12.0 so you can go back and forth between two conformations,
11:15.2 and you can read out the movement,
11:17.2 how quickly this movement occurs,
11:19.1 and how big the amplitude of the movement is
11:21.1 by just looking at the color change
11:23.2 as a function of time.
11:24.2 So, this is the idea,
11:26.2 and here's the final reaction.
11:28.2 So, if you measure intensities of green and red fluorescence
11:32.3 as a function of time, as you can see,
11:35.3 you can see beautifully anti-correlated changes
11:39.0 that report on the conformational changes
11:44.1 during the molecular dance.
11:48.1 Here are the results...
11:50.2 in addition to intramolecular conformational changes
11:53.1 that you can measure, as shown here,
11:56.0 you can use single molecule FRET
11:58.3 also to measure intermolecular changes,
12:02.1 interactions between two different molecules.
12:05.3 You can put a green dye on a protein
12:08.3 moving on the DNA lattice,
12:12.2 let's say, powered by ATP,
12:14.2 and then put the red dye at the destination end of the DNA,
12:20.1 then as time goes on,
12:21.3 as the protein approaches the red fluorescent molecule,
12:25.2 FRET increases and you'll see green signal going down
12:29.1 and red signal going up.
12:33.1 So, this is the idea behind using FRET to measure
12:36.1 movements between two molecules.
12:39.0 This animation is not the actual data,
12:41.2 it's what I prepared using a PowerPoint,
12:46.1 but our data is not that bad, it's almost as pretty.
12:49.0 So, here's an example where single molecule data
12:52.2 show that green and red intensities go up and down,
12:55.1 you know, green goes down initially, red goes up initially,
12:58.0 as shown in the cartoon.
13:00.1 The surprising observation here was that,
13:03.1 instead of falling off at the end of the DNA track,
13:06.2 when the molecule reaches the end of the DNA track,
13:09.3 it somehow reappears at the beginning of the track
13:13.3 and goes back to a low FRET value very quickly,
13:17.2 and then it then shows this gradual increase in FRET
13:22.2 and then repeats this process many, many times.
13:27.2 So, we have done a lot of different measurements
13:29.3 to understand the physical basis for this phenomenon
13:34.2 and how the molecule is able to perform
13:38.0 such acrobatic activities,
13:40.2 and also to deduce how this type of interesting,
13:46.0 you know, repetitive movements
13:48.2 can be useful for biological functions of the molecule.
13:52.3 So, after many months of research of this kind,
13:56.1 we eventually asked a magazine
14:01.2 to publish our research,
14:05.3 and after a usual back and forth
14:09.0 they eventually agreed to publish the study.
14:11.0 So... and then we asked an artist to,
14:15.1 you know, make a cover illustration
14:18.2 to suggest how, you know,
14:21.2 this enzyme may be performing the function.
14:23.2 So here's an example:
14:25.2 we proposed that this protein can sit on the DNA
14:28.2 and then reel in the single-stranded DNA
14:32.0 using the ATP-powered movement,
14:34.1 and then remove proteins bound to the DNA
14:36.3 to clear the DNA of unwanted proteins.
14:41.1 Well, we submitted this illustration to the magazine,
14:46.2 hoping that they would show it on the cover,
14:48.3 but unfortunately they just said, "No,"
14:52.1 and instead they showed a picture of a diabetic mouse,
14:55.3 and, well, that's fine...
14:59.1 you win some, but, you know, lose some,
15:02.1 so we just move on.
15:04.0 Here's another illustration that we paid an artist to create...
15:08.1 same idea, now we have a superhero,
15:10.3 my student Jeehae Park, who did the work,
15:14.0 sitting on the DNA and reeling in the DNA,
15:16.2 and removing asteroids bound to the DNA
15:19.0 using a powerful motor function.
15:21.3 Again, the magazine did not pick this cover either,
15:24.2 so I have to, you know,
15:26.2 use these images whenever I give public lectures.
15:32.3 Alright, so let me show you some real data
15:35.3 using DNA repair as an example.
15:42.2 As you know, DNA is an important genetic material
15:46.2 required to make proteins,
15:50.3 and proteins are the actual players
15:53.1 of the living cells in our body,
15:56.3 but DNA is under constant threat of damaging reagents.
16:00.2 If you smoke or if you go out to the sun in California,
16:04.2 then your DNA gets damaged
16:07.1 because of the smoke and also sunlight,
16:10.0 and DNA damage can cause
16:15.3 aging or cancer or other genetic disorders.
16:19.0 I see that I'm getting older every time I look at myself in the mirror,
16:23.2 so this is an important topic for me as well.
16:30.2 Do you know how much DNA you make in your lifetime?
16:35.1 It turns out that,
16:38.1 in your own body at this moment,
16:39.2 you have about 100 trillion cells,
16:46.0 and each cell contains about 2 meters of DNA,
16:50.0 and during your lifetime you,
16:52.3 you know, regenerate the cells
16:55.3 about 100 times on average.
16:57.1 So, if you connect all of the DNA molecules
16:59.2 that you produce back-to-back,
17:02.2 it can actually span a distance of one light year.
17:07.0 One light year is not a unit of time,
17:10.0 it's a unit of distance,
17:11.2 it's how far the light travels in one year,
17:14.2 so it's an enormous, actually, distance
17:17.1 and you have to make a lot of DNA.
17:19.3 And each cell in the body
17:22.3 can produce more than 1000 damages
17:27.2 even in a single day,
17:30.0 so unless you are able to repair the damages efficiently
17:32.1 you can inevitably accumulate
17:36.0 enough DNA damage to cause cancer.
17:38.1 You know, without proper DNA damage repair,
17:41.1 you'll actually get cancer at a much younger age.
17:47.1 So, there are several ways
17:50.0 to repair DNA damage,
17:52.1 and one is called mismatch repair,
17:54.3 another is called nucleotide excision repair.
17:57.2 Today, I'm going to tell you briefly about
18:00.0 a repair process called
18:04.1 homologous recombination.
18:08.1 Homologous recombination is
18:10.3 about as close as you can get to having sex,
18:14.3 if DNA can have sex.
18:18.0 So, there's a DNA and you have,
18:20.1 you know, DNA damage
18:22.2 causing a break in the middle of the DNA,
18:24.0 producing what's called a double-strand break.
18:28.1 Then, the cell has machineries to
18:31.3 digest one of the two strands of DNA
18:34.2 to produce single-stranded DNA,
18:38.1 and that single strand then finds a matching sequence,
18:42.2 you know, another copy of the same DNA inside the cell,
18:45.1 called homologous DNA.
18:47.0 And then once you find the matching copy
18:49.1 then you can use the information stored
18:51.2 in the other copy
18:53.2 to reproduce the original DNA without any mistakes.
18:59.2 This process requires a protein called RecA
19:04.3 to form a filament around single-stranded DNA,
19:10.1 and this is for bacterial cells,
19:13.2 and for humans we have its own counterpart
19:16.1 called RAD51.
19:19.2 Now, this RecA filament formed on a DNA sequence
19:23.0 has to find the matching sequence
19:26.1 in the sea of millions of different base pairs,
19:29.2 even in a small E. coli bacterium,
19:34.1 so, you know, this is equivalent to finding a soulmate,
19:37.1 your future husband,
19:41.2 you know, when you have millions of people
19:44.2 living in New York City, for example.
19:48.1 How do you actually achieve that?
19:50.1 More like finding a needle in a haystack.
19:56.1 One mechanism that has been proposed is called the 3D search.
20:01.0 Here, the filament is diffusing in solution
20:05.0 and then it randomly collides with the DNA
20:08.0 at random locations.
20:10.0 Unless it has a perfect match of sequence,
20:14.1 it dissociates quickly
20:17.1 and then goes to look for another sequence,
20:22.1 and this is more like dating a random person on the street
20:26.2 to find your soulmate,
20:30.1 and if you have a lot of actually
20:33.0 competing potential partners,
20:37.0 this can take a long time.
20:39.0 In this animation, I got tired,
20:40.1 so I just stopped it in the middle, so...
20:45.1 the filament found the correct target,
20:47.2 but you can imagine that if you millions of possibilities
20:50.2 it's going to take a long time.
20:54.2 Another possibility is called 1D sliding,
20:58.2 and this is equivalent to joining a book club.
21:02.1 So, instead of dating a random person on the street,
21:05.1 you join a club to meet people
21:09.0 who share a common interest,
21:10.2 and then you can date maybe 20 people within the club
21:13.2 to see whether you have a good fit,
21:16.0 and if you don't find the perfect match
21:19.2 then you can join a different club,
21:21.3 like a knitting club,
21:23.2 and then do the sampling again.
21:26.0 So, here's an illustration:
21:29.2 the single-stranded filament can bind
21:33.2 and then slide on the DNA, 1D sliding,
21:35.1 to sample, you know, 200 or 300 base pairs
21:38.0 before falling off,
21:39.2 so you can actually sample multiple potential partners
21:42.2 in one binding event,
21:44.1 and this can make the process much, much faster.
21:51.2 So, we perform single-molecule FRET measurements
21:55.0 of the process
21:57.2 by putting a green dye on the filament itself,
22:02.2 and the red dye on the DNA.
22:07.0 In this experiment,
22:10.1 DNA, target DNA,
22:12.0 actually does not have the matching sequence,
22:13.2 so the filament will still bind to the DNA
22:17.3 and then show sliding motion,
22:20.3 and that sliding motion,
22:22.3 if it indeed occurs,
22:25.1 then it will be visualized as fluctuations in,
22:29.1 you know, single-molecule FRET,
22:32.0 where green and red intensities go up and down
22:35.0 in an anti-correlated manner, very rapidly.
22:38.0 And this is the actual signal that we see
22:40.1 from single molecules,
22:41.3 indeed showing that RecA filament
22:46.3 can slide on the DNA,
22:49.0 possibly in search of a target sequence.
22:52.0 So that was really, really exciting,
22:54.1 but does it really
22:59.2 find the matching sequence while sliding?
23:04.0 So, we performed another experiment, shown here.
23:10.1 So, we designed a DNA sequence
23:12.1 that contains a matching sequence
23:14.1 for the filament DNA,
23:20.1 but the sequence is not very long,
23:23.3 it's only seven base pairs long,
23:25.3 so it's not a perfect match,
23:28.2 so it's as if, you know, you...
23:31.2 you find a near match,
23:34.2 but here in this case we put
23:36.3 two copies of the same near match.
23:39.2 So, you know, again,
23:42.1 to use that example of dating,
23:44.2 you may be dating twin brothers, okay?
23:48.0 Well, both are near matches,
23:50.2 but you spend some time with one of the two twin brothers,
23:53.3 but the match is not perfect,
23:56.0 so you slide off to the other twin brother,
24:00.2 and that's why you see
24:03.0 a change in FRET efficiency,
24:04.2 shown on the top,
24:06.0 and then you then leave the other brother
24:08.1 and you go back to the original brother
24:10.1 and so on.
24:11.1 So, this data showed
24:14.0 very clearly
24:16.1 that the sliding on the same DNA
24:18.2 can be used to sample
24:24.0 different sequences in search of
24:26.3 a matching sequence.
24:28.3 So, overall, data of this type
24:31.2 showed us that 1D sliding
24:34.2 can make the target search process,
24:37.0 you know, 250 times faster than 3D search,
24:41.2 and in principle this sliding activity
24:46.0 that we discovered in this study,
24:47.2 you know, can make DNA repair
24:50.3 that much faster.
24:53.1 Okay, so I have told you about how
24:55.3 single molecule fluorescence imaging,
24:57.1 tracking of single fluorescent molecules
24:59.2 and measuring the interaction between
25:02.2 two different molecules
25:04.2 to, you know, to measure
25:07.0 the activities of single protein molecules
25:11.2 in real time.
25:13.3 There is another technology
25:16.3 that is extremely popular in the single molecule biophysics community.
25:19.2 It is called the optical trap
25:22.0 or optical tweezers technology.
25:24.0 If you want, this can be viewed as
25:26.0 chopsticks made of light.
25:28.1 So, you can use a tightly focused laser light
25:33.2 to grab a single particle in solution
25:40.3 and then if you move the laser beam around back and forth
25:44.0 you can move the particle,
25:46.1 and if the other end of the DNA
25:48.2 linking the particle to the surface
25:51.1 is an enzyme that is transcribing the DNA
25:54.2 -- RNA polymerase --
25:56.2 then the distance between the particle and the enzyme
26:01.1 will change, it will become shorter and shorter,
26:03.2 and you can measure this with Angstrom-level precision,
26:07.2 and you can apply forces that are relevant
26:11.0 to physiological conditions
26:13.1 and measure the mechanical response very precisely.
26:16.2 So, my laboratory and many others
26:22.1 have decided to combine these two different technologies,
26:26.3 single-molecule FRET and the optical trap,
26:29.3 so that you can use fluorescence
26:32.1 to measure conformational changes of a single molecule,
26:35.2 but use the optical trap to apply
26:38.2 a defined level of force
26:40.3 and understand how the tension, or force applied,
26:44.1 can change the activities of a single molecule.
26:51.1 And this is equivalent to
26:55.1 combining two different features of these measurements:
26:58.2 in the optical trap measurements,
27:00.2 you are using your hands to
27:03.2 manipulate a single molecule
27:05.2 and measure its response mechanically,
27:08.0 but with your eyes closed;
27:10.2 but in the FRET measurement,
27:13.2 you are using your eyes to make observations,
27:16.1 passive observations,
27:18.1 with your hands tied in the back.
27:20.3 By combining the two,
27:23.1 we can actually hope to sample
27:25.1 the best of both worlds
27:27.1 and this is going to be the topic of my second talk.
27:34.1 Alright, so I'd like to acknowledge
27:37.2 my former and present colleagues
27:40.1 and collaborators.
27:41.2 We collaborated with Yale Goldman's lab
27:43.2 and Paul Selvin's lab
27:45.1 on the myosin V studies.
27:47.1 And Kaushik Ragunathan, Chen Liu,
27:50.1 and Chirlmin Joo, and Jeehae Park,
27:53.0 and Sua Myong
27:54.3 worked on many of the
27:59.0 single molecule FRET studies I told you about.
28:00.1 And Ahmet Yildiz and Sean McKinney
28:03.1 worked on the
28:06.3 tracking of a myosin V molecule.
28:10.3 Thank you very much.

Part 2: Combining FRET and Optical Trap to Study the Nucleosome

00:07.2 Hi.
00:09.2 I'm Taekjip Ha.
00:11.1 I'm a professor of biophysics and biomedical engineering
00:13.2 at John Hopkins University.
00:16.1 I am also an investigator
00:18.2 with the Howard Hughes Medical Institute.
00:23.0 This is the second of a three talk series
00:24.3 where I discuss how single molecule technologies
00:28.1 are changing the way we study and understand
00:33.1 many of the important biological molecules.
00:37.1 In this part, I will discuss
00:40.2 how the combination of single molecule fluorescence technologies
00:44.1 and single molecule mechanical manipulation technologies
00:47.2 can be used to study the dynamics
00:51.2 of the nucleosome,
00:53.1 which is the compaction unit of our own genome
00:55.3 inside our cells.
01:03.0 We have been interested in developing new probes
01:05.2 to study the relation between structure and function.
01:10.1 In normal biochemical literature,
01:13.2 structure-function relationship usually denotes
01:17.2 the idea that if you make a change
01:20.1 to the primary structure of a protein,
01:22.1 the sequence of the amino acids,
01:24.2 then you change the function,
01:27.1 but the way we look at this question
01:29.2 is a little bit different, here.
01:31.1 So, imagine having a can opener
01:34.2 and if you have never seen a can opener before,
01:39.0 then you may not be able to guess
01:42.2 what the function is just by looking at the structure.
01:45.0 So, that's the structure,
01:47.0 but its function is to open a can.
01:49.3 The best way to learn how the can opener functions
01:53.2 is to watch someone use it in real time,
01:58.1 and as you know the can opener has to change its shape repeatedly
02:02.3 to carry out the function.
02:05.2 And this is also true for many of the biological molecules
02:08.3 such as proteins, and DNA and RNA molecules.
02:13.3 We have wonderful biophysical technologies
02:16.2 such as X-ray crystallography,
02:20.1 electron microscopy,
02:21.2 and NMR and so on,
02:23.1 to determine the [structure] even at the atomic resolution.
02:27.0 We also have wonderful biochemical tools
02:29.1 to measure their function.
02:32.1 But, really, still a challenge and a goal
02:38.1 is to measure the changes in structures
02:41.2 while they are functioning,
02:43.2 and this is what we are going after.
02:47.1 So, we have a technology that we call
02:49.3 single molecule FRET.
02:51.1 If you want to measure shape changes,
02:54.0 or conformational changes of a molecule,
02:56.1 you can put two different dye molecules,
02:59.3 in green and red,
03:02.0 on the known sites of the molecule,
03:04.2 so that you can distinguish between the
03:07.0 closed and open conformations.
03:09.2 In the open conformation,
03:11.0 if you excite the green dye,
03:12.2 then you get mainly green emission,
03:14.1 but in the closed conformation
03:16.0 energy can be transferred from the green to the red
03:19.0 so that you get photons coming out
03:21.3 from the red molecule.
03:23.1 So, by measuring the relative intensities
03:25.3 of the two colors,
03:27.1 you can deduce the shape of the molecule.
03:31.2 Now, if you want to measure how quickly
03:34.2 the butterfly flaps it wings in ensemble,
03:37.0 you must synchronize the reaction
03:39.1 by starting from one conformation.
03:43.1 You can convince the molecules,
03:45.2 the butterflies,
03:47.2 to sit on the leaf in one conformation,
03:50.3 and then you can frighten them into taking off the leaf,
03:54.0 and you perform a relaxation analysis.
03:58.0 But often times, what you measure here
04:01.2 is not the wing flapping frequency,
04:04.0 but how long it takes for the butterflies
04:06.2 to take off from the leaf.
04:09.1 Ideally, you want to look at a single butterfly
04:12.2 or a single enzyme
04:14.1 as a function of time,
04:15.1 so you can measure
04:18.0 anti-correlated changes in green and red intensities
04:20.1 as a function of time,
04:22.1 and this can give you information
04:24.1 on the amplitude of the motion
04:26.2 and also its timescale.
04:29.0 So, this technology has been really, really powerful
04:34.1 in revealing hidden details about biomolecules.
04:37.2 It's being used by hundreds of laboratories
04:40.0 around the world.
04:43.1 So, why study single molecules, generally?
04:47.0 First, you can detect sub-populations and dynamics
04:51.1 in the most direct way.
04:54.2 And you can also detect
04:56.2 non-synchronizable dynamics.
04:59.1 If your molecules can be synchronized,
05:01.1 just like in a marching band,
05:03.2 you don't have to measure single molecules
05:06.0 to study the details of the movements,
05:08.3 but, you know, imagine trying to do that
05:12.1 to college students on campus.
05:14.1 It's impossible to synchronize them;
05:16.0 you then have to then observe single individuals.
05:22.0 We can also reveal that
05:24.2 nominally identical molecules can be heterogeneous.
05:27.1 You know, they can have
05:30.0 different personalities and different moods
05:31.2 that can only be revealed by looking at single molecules.
05:35.2 Finally, when you perform a single molecule measurement,
05:38.2 you can actually make
05:42.3 measurements of multiple observables
05:44.1 from the same molecule
05:45.2 and make correlations between that,
05:47.3 and this is actually a topic of today's talk.
05:56.2 There's another powerful technology called optical trap.
06:00.1 We can call it chopsticks made of light.
06:03.2 You can grab a single particle
06:05.3 using a focused laser beam
06:07.2 and apply very small forces,
06:09.3 down to single picoNewtons of force,
06:14.1 and a picoNewton of force
06:16.2 is about the weight of a single E. coli bacterium,
06:21.1 and it's a very small amount of force
06:24.0 because if you have your iPhone,
06:25.3 the gravitational force on that iPhone is about a Newton,
06:28.2 so it is 10^12-times smaller than the weight of your phone.
06:36.1 If you can grab a single particle
06:39.1 and apply a force this way,
06:41.2 you can then play a tug-of-war
06:44.0 with an enzyme sitting on a glass surface,
06:47.0 you know, through a DNA molecule,
06:49.0 and this method can give you
06:51.2 information on the force being produced by a molecule,
06:54.3 and also you can measure
06:58.1 changes in DNA lengths down to an Angstrom,
07:01.1 or single base pair resolution,
07:03.0 to follow the activities of these molecular machines
07:08.3 with the ultimate resolution of, you know,
07:12.1 single base pair incorporation and so on.
07:16.2 So, my lab and many others,
07:19.0 shown on the bottom of the slide,
07:21.1 have made efforts to combine
07:23.3 these two different technologies
07:26.0 -- one is single molecule fluorescence,
07:27.2 the other is optical trap --
07:30.0 so that you can measure, for example,
07:32.2 conformational changes of a molecule using FRET,
07:37.2 but as a function of applied force.
07:41.0 And the force, here,
07:43.3 actually mimics the physiological conditions
07:45.1 because these molecules are in a crowded environment,
07:47.2 bumping and tugging each other,
07:51.0 and so, you know, the measurements
07:53.2 performed under tension can be
07:56.2 much more relevant to what's really going on inside a cell.
08:01.0 If you want, optical trap measurements allow you to measure
08:04.2 things mechanically, with your hands,
08:07.1 but with your eyes closed,
08:09.0 whereas in single molecule FRET
08:11.1 you can use your eyes to make passive observations,
08:15.3 but with your hands tied in the back.
08:18.0 And by combining the two into one instrument,
08:21.1 you can sample the best of both possibilities
08:25.1 and reveal new things that you couldn't see before.
08:31.1 Okay.
08:32.3 Let's get into the topic of today's presentation:
08:39.2 DNA and the nucleosome.
08:41.3 This is a famous paper
08:43.3 published more than 60 years ago
08:45.3 by Watson and Crick
08:48.0 to propose a double helical structure
08:50.3 of this iconic molecule, DNA,
08:54.0 and this paper ended with this paragraph, saying that,
08:57.2 you know,
08:59.3 "It has not escaped our notice
00:09:01.26 that the specific pairing mechanism...
00:09:03.28 suggests a possible copying mechanism
00:09:05.26 for the genetic material."
09:07.2 And then later on they published another paper
09:10.2 proposing how A's and G's can...
09:13.3 A's can pair with T's,
09:15.2 and G's can pair with C's
09:17.2 to form base pairing as the explicit
09:21.1 mechanism of genetic material copying.
09:24.0 And the rest became history.
09:28.2 In our own body,
09:30.3 in every cell of our body,
09:32.2 you have about 2 meters of DNA,
09:36.1 and this long DNA has to be
09:38.3 somehow packaged inside a chromosome,
09:42.0 shown at the bottom,
09:43.1 and inside a cell nucleus,
09:46.0 and this happens through multiple stages.
09:49.2 At the most elementary stage,
09:52.0 DNA is bundled into a structure called the nucleosome,
09:57.0 shown in this movie.
10:02.0 To look at this structure in slightly more detail,
10:07.1 here... about 140-some base pairs of DNA
10:10.2 is wrapped around a protein core,
10:13.3 almost 2 turns, 1.7 turns.
10:16.2 So I have seen many, many protein structures
10:20.0 in my career,
10:21.2 but this is still my favorite
10:24.1 because it's so beautiful and you just look at it
10:26.2 and you know immediately what it means.
10:29.0 This is how DNA is packaged inside the cell.
10:32.3 The nucleosome has to satisfy
10:35.1 two conflicting requirements.
10:37.1 First, it has to be stable
10:39.2 to serve as the compaction unit of our genome
10:43.1 in the nucleus.
10:45.0 It also has to be dynamic
10:47.0 to allow gene expression.
10:48.2 You know, proteins have to come and bind to the DNA
10:50.2 to, you know, make messenger RNA molecules,
10:53.1 but if the DNA is still bound to the protein core,
10:55.2 you cannot do that.
10:58.1 So, my student Thuy, shown here,
11:01.3 decided to develop single molecule assays
11:03.3 to measure the dynamics of a single nucleosome
11:07.1 under force, or under tension,
11:10.2 at the level that is, you know,
11:13.2 expected to occur all the time
11:16.2 inside a living cell.
11:19.1 Previous studies done by other labs,
11:22.1 shown at the bottom,
11:24.1 using the optical trap technology...
11:28.1 where they pulled on the ends of DNA in a chromatin,
11:32.2 multiple nucleosomes or a single nucleosome,
11:35.1 led to the following model,
11:38.2 where the nucleosome is unraveled under tension
11:41.1 in multiple stages.
11:45.1 At low forces,
11:47.2 the outer turn of DNA in a nucleosome unravels,
11:52.1 and as you increase the force to a higher value
11:56.0 the inner turn unravels,
11:58.0 and finally at even higher forces
12:00.0 the protein core is ejected from the DNA.
12:04.2 Because we have an instrument that combines
12:06.2 single molecule fluorescence, FRET,
12:09.0 with optical trap,
12:10.2 we decided to test our ability
12:12.3 to examine the same process in a different way.
12:16.1 So, in the previous studies,
12:19.0 people actually used the end-to-end distance measurement
12:22.0 of the DNA tether to,
12:24.1 you know, conclude that
12:26.2 the nucleosome unravels in multiple stages, as shown,
12:28.2 in a symmetric manner.
12:31.0 We decided to put red and green dye molecules
12:34.0 to the DNA on the nucleosome,
12:36.1 as shown,
12:37.3 so that, initially,
12:39.2 in the absence of any tension,
12:41.0 the FRET should be high
12:43.2 because of the small separation between the two dye molecules,
12:46.2 high proximity,
12:48.0 so you get mainly emission in the red channel of the detectors,
12:52.2 but as soon as it unravels,
12:55.2 the outer turn of the nucleosome,
12:57.3 at low force,
12:59.2 then FRET should go down,
13:02.0 you know, quite dramatically.
13:03.2 And depending on the exact configuration
13:05.3 of your measurement,
13:07.2 you may see additional FRET decrease
13:10.0 upon inner turn unraveling.
13:13.2 So, that was the idea.
13:15.2 This is the actual DNA that we used,
13:18.3 where we picked a sequence called 601,
13:22.3 discovered by the late John Widom
13:26.1 -- this is a sequence discovered from in vitro selection,
13:28.3 it's an artificial that's known to
13:31.2 position the nucleosome
13:33.2 in a precise location on the DNA,
13:35.1 down to a single base pair of precision --
13:38.3 then we can put the green dye
13:41.2 near the middle of the DNA
13:44.0 -- this is called the dyad axis,
13:45.2 representing the symmetry axis of the nucleo...
13:49.1 the DNA and nucleosome --
13:51.2 and we can put the red dye toward the end of the DNA,
13:55.1 so that in the 3D structure shown here,
13:59.2 the two dye molecules are very close to each other,
14:02.2 sitting at the entry point of the nucleosome, here.
14:06.2 So, this is a cartoon version of that.
14:10.1 So, you should get high FRET
14:12.1 because of high proximity at the beginning,
14:14.1 but as soon as you unravel
14:17.2 that beginning part of the nucleosomal DNA,
14:19.3 then you should see a major reduction in FRET.
14:23.3 And we call this end-dyad labeling scheme,
14:28.3 and call it ED1
14:32.1 because we are labeling the dyad and positions
14:34.2 of the DNA.
14:37.2 We can use a different labeling scheme,
14:40.1 shown here,
14:42.2 that we call internal labeling.
14:45.1 We moved the dyes to different locations
14:48.1 so that now the dyes
14:52.1 show up at the bottom of the nucleosome,
14:54.3 and they are again very close to each other,
14:57.1 as shown in this cartoon,
14:59.2 but you can imagine that you
15:02.2 need to unravel more than 20 bases...
15:05.0 base pairs of DNA before you expect to see
15:08.2 a decrease in FRET,
15:10.1 because you are labeling internal locations
15:12.2 of the nucleosome.
15:14.1 So, this should be sensitive
15:16.2 to the inner turn unraveling
15:18.3 that is expected to occur under high tension.
15:24.2 Then we can tether the DNA molecule
15:26.2 to a surface using a biotin
15:28.2 -- this is a well-known molecular glue --
15:30.1 and then we can pull the other end of the DNA
15:33.1 using, you know,
15:36.1 an optical trap through a long DNA linker,
15:40.1 and measure FRET at the same time.
15:45.2 So, here, I'm showing you the actual data.
15:49.0 We have this ED1 and internal labeling probes,
15:53.1 so, one for outer-turn probes
15:55.3 and the other is for inner-turn probes,
15:59.0 and I'm showing you FRET as a function of force
16:02.1 that we measure from many, many single molecules
16:04.2 of the same kind,
16:06.1 so you see multiple lines here
16:08.0 because these are from different single molecules.
16:11.3 For the outer turn probe, ED1,
16:14.0 we find that FRET starts
16:16.2 at high values at the beginning,
16:18.2 at low forces.
16:20.1 As you increase the force to about 3 picoNewtons of force,
16:23.1 which is a level that is low enough
16:25.0 that you should expect to see
16:26.3 this level of force all the time
16:28.2 inside a living cell,
16:30.0 then FRET drops to the lowest value.
16:33.0 So, a major unraveling occurs at low forces,
16:36.2 as you expect based on prior studies.
16:40.0 Now, if you look at the inner turn probe,
16:42.2 FRET starts high because, again,
16:44.2 the dyes are very close to each other,
16:46.1 but the FRET decrease does not occur
16:48.2 until you go to much higher forces,
16:50.1 let's say 15 picoNewtons.
16:54.2 So, that was great because
16:57.1 this matches exactly what we were hoping to see,
16:59.2 you know, matching the prediction by other studies
17:03.1 -- outer turn comes off at low forces;
17:05.1 inner turn comes off at higher forces.
17:09.1 We were very happy that you could do this measurement
17:11.0 and we should have published that paper
17:13.2 at that moment,
17:16.2 but we made a mistake of doing one more experiment,
17:19.2 and that led to a...
17:23.2 actually, a three-year saga
17:25.3 to understand what is really going on,
17:28.1 and I will tell you later that we have a happy ending here.
17:31.3 So, my student said that,
17:34.0 well, if our model is correct,
17:35.3 then if you move the probes to the other end
17:38.3 or other entry point of the nucleosome
17:40.2 -- we can call that ED2 --
17:42.2 then that should also unravel at lower forces,
17:46.0 as was seen for ED1 over there.
17:51.1 But when she performed the experiment it turned out
17:55.2 that ED2 construct actually
17:58.2 didn't come down in FRET
18:01.1 until you went to even higher forces than the middle one, here.
18:05.2 So, that was really surprising,
18:08.2 and if this is true, what it means is that,
18:11.2 instead of unraveling symmetrically from both ends,
18:14.3 unravelling is asymmetric.
18:17.2 It may even be directional,
18:19.0 starting from the weak side,
18:23.2 propagating toward the middle,
18:25.3 and finally to the strong side.
18:29.2 Okay.
18:31.0 So, that was surprising,
18:32.3 so we had to do a lot of experiments
18:34.1 to make sure that this not,
18:36.1 you know, an artifact of the experimental configurations.
18:41.1 First, we noted that in our experiment
18:46.1 the ED2 side, that is, the stronger side over there,
18:51.1 yeah,
18:52.2 is tethered to the surface of a coverslip
18:54.2 through a short DNA linker,
18:57.1 whereas the weak side, ED1,
18:59.2 is tethered to a much longer DNA
19:02.2 and to the optical trap.
19:05.0 So there was actually asymmetry in the configuration
19:07.2 and so we though this may be the issue,
19:10.1 and this may be the reason
19:12.2 why we see this strong asymmetry.
19:14.1 So, we did a control experiment
19:16.1 where we swapped the experimental configurations
19:19.2 and in this case we put
19:23.2 a short tether to a surface on the weak side
19:26.1 and then a long tether to the strong side,
19:30.2 and then when you performed the measurement
19:32.2 nothing changed, you know,
19:34.0 what was weak was still weak
19:35.1 and what was strong was still strong,
19:37.0 so it's not the surface tethering
19:39.3 and asymmetry in configuration
19:41.2 that was causing this apparent asymmetry.
19:47.1 Now, if this is true,
19:49.2 we thought, well,
19:51.2 let's say this ED1 is indeed strong
19:53.2 and ED2 is indeed...
19:56.3 sorry, ED1 is indeed weak and ED2 is indeed strong,
20:01.2 then if you move the probes a little deeper
20:04.2 from the weak side,
20:05.3 and we call this ED 1.5,
20:07.3 that side should still be weak,
20:10.0 maybe slightly stronger,
20:11.1 and if you move the probes from the strong side
20:13.3 into the nucleosome,
20:15.2 we call that ED 2.5,
20:17.2 then that side should still be strong.
20:20.2 And indeed that was the case.
20:22.2 So, here's the FRET as a function of force.
20:26.0 ED1, the original weak side, shown in black,
20:29.2 unravels at very low forces.
20:32.2 We go a little bit deeper, 1.5, shown in red,
20:36.2 it unravels still at low forces, but slightly higher.
20:40.1 And if you look at 2.5,
20:42.2 going a little bit into the strong side,
20:44.3 that is still strong, shown in the blue curve.
20:48.1 So we began to feel that this must be real
20:52.0 and it's really
20:57.3 caused by the intrinsic asymmetry of the nucleosome.
21:01.2 What in the end clinched this idea
21:04.2 was an additional measurement that my student did.
21:08.2 So, she decided to add
21:11.2 even more configurations to go around the nucleosome,
21:16.0 so I have already shown you
21:18.2 five different sets of data,
21:19.3 two on the weak side
21:21.3 and then two on the strong side,
21:24.0 and this is the internal one,
21:25.2 and she added two more data points
21:28.0 on the weak side.
21:32.0 They still remained weak,
21:33.2 and two other data points on the strong side
21:35.1 and they still remained strong, really...
21:37.2 you know, showing very clearly that,
21:40.0 indeed, the nucleosome is asymmetrically unraveled
21:44.0 if you apply tension to the molecule.
21:47.2 So, if you want,
21:49.2 instead of unraveling like that, shown here,
21:53.1 we may have a system where we...
21:59.1 unraveling occurs in this matter, like that.
22:03.0 Okay.
22:04.2 So, the next question is,
22:06.2 you know, why is it asymmetric?
22:09.1 And the follow-up question would be,
22:10.2 you know, is it important biologically?
22:16.0 So, we then looked at that
22:20.0 601 sequence in more detail.
22:21.3 So this, you know,
22:24.0 has 147 base pairs
22:26.2 and we said that there's a strong side
22:28.3 and there's a weak side on the DNA.
22:32.3 It turns out, if you look at the nucleosome structure itself,
22:35.1 the protein part in the middle is
22:38.1 completely symmetric,
22:39.2 there's no asymmetry there,
22:41.0 so it must be the asymmetry
22:43.1 in the DNA sequence itself,
22:44.2 which, in fact, is not symmetric.
22:47.0 So, the idea that we had is shown here.
22:50.2 We thought, perhaps,
22:53.1 that the strong half is strong mechanically
22:57.1 because the DNA sequence makes it more flexible.
23:01.1 Why?
23:03.2 In the nucleosome,
23:05.2 the DNA is sharply bent around the protein core
23:08.0 and this sharply bent conformation
23:09.3 is not natural for the DNA,
23:12.1 so it wants to become straight.
23:15.3 But if the sequence makes the DNA more bendable
23:19.1 and more flexible,
23:20.1 then this sharply bent conformation
23:22.1 is better tolerated, even under tension,
23:25.3 and that can be responsible
23:28.2 for the stronger mechanical stability
23:30.2 of one sequence over another.
23:34.1 So, that was our prediction,
23:35.3 but how do we test this prediction?
23:38.1 Well, we need to be able to measure
23:41.1 flexibility of a DNA of very, very short lengths,
23:44.2 about half the length of a nucleosome of DNA
23:47.1 -- 70 to 80 base pairs.
23:52.1 We are lucky because my former student,
23:54.2 Reza Vafabakhsh,
23:57.0 had developed an assay that can just do that.
24:01.2 We call this the single molecule looping assay.
24:03.1 Now, here, we start with a short DNA,
24:06.2 about 100 base pairs in length,
24:09.0 immobilize to a surface
24:11.2 using a biotin-avidin interaction,
24:13.1 and then we have sticky ends
24:15.2 on the both ends of the DNA
24:18.1 that are complementary to each other,
24:19.2 so that they can base pair against each other.
24:22.2 If you start with a condition,
24:24.2 a lower concentration of sodium chloride, for example,
24:27.3 that prefers this linear,
24:30.2 unlooped state of the DNA,
24:32.2 and then rapidly change the solution condition
24:35.1 to go to high sodium chloride concentrations,
24:39.1 that favors this looped conformation,
24:42.1 then you can measure, using FRET, for example,
24:46.0 how quickly the loop population
24:49.0 develops as a function of time.
24:51.0 Okay. So, this is actually data,
24:53.2 the fraction of looped molecules versus time,
24:56.1 and it shows that, for this particular sequence,
24:58.3 looping takes about 15 or 20 minutes.
25:02.3 Now, the idea here is that,
25:06.1 if the sequence makes the DNA more flexible,
25:09.1 then looping will occur faster.
25:13.2 Then you can use the speed of looping,
25:16.2 or rate of looping, or looping rate,
25:19.0 as a measure of,
25:20.3 you know, "flexibility", okay?
25:23.2 So, we decided to measure looping
25:27.2 of the sequences found inside a nucleosome.
25:32.2 I told you we have two halves,
25:34.2 strong and weak halves,
25:36.1 so on the bottom I have loop fraction
25:38.2 as a function of time.
25:40.2 For the red curve,
25:42.1 that's coming from the strong side,
25:44.1 that took about 27 minutes,
25:47.1 whereas on the weak side,
25:49.3 the green curve,
25:51.2 it took more than 3 hours.
25:53.3 So, indeed, the data suggests that
25:57.0 the strong half of the nucleosomal DNA
25:58.3 is actually much more flexible,
26:01.2 as we predicted.
26:03.0 By how much?
26:04.3 About 7 times, according to our measure of flexibility.
26:08.2 So, we were very happy that our prediction
26:13.2 was actually supported by the data,
26:16.2 but this also could be a coincidence,
26:18.2 so we decided to look at this in more detail.
26:22.1 Let me just tell you what we did.
26:24.2 So, we then divided the DNA
26:29.1 into four segments, four quarters,
26:32.0 and then first decided to change
26:34.2 the sequences of the DNA
26:36.3 so that we can swap the outer quarters
26:39.2 of the DNA sequences.
26:40.3 And we performed the single molecule looping experiment
26:43.1 and optical trap experiments,
26:48.0 and found that nothing changed
26:50.0 #NAME?
26:52.0 and what used to be more flexible
26:54.1 was still more flexible and so on.
26:56.1 So, it's not the outer quarters of the DNA
26:59.1 that are causing the asymmetry.
27:02.2 If you instead swap the inner quarters,
27:05.0 as shown in the color scheme here,
27:07.1 while keeping the outer quarters the same
27:09.3 as the original DNA,
27:11.2 then everything swapped
27:15.1 #NAME?
27:18.3 and what used to be more flexible became stiffer.
27:23.1 So, we now believe that
27:27.1 it is the sequence of
27:29.2 the inner quarters of the nucleosomal DNA
27:32.1 and the resulting flexibility
27:35.2 is the determining effect...
27:38.2 is the determining effect for the
27:41.1 mechanical stability of the nucleosome
27:43.0 and the asymmetry that we observed.
27:45.2 And this paper was published in Cell in 2015,
27:50.1 and I told you there would be a happy ending here,
27:52.2 so she was very happy.
27:55.2 Then she went on to study the effects of
27:59.0 not only the effect of the DNA sequence
28:00.3 but also of DNA chemical modifications.
28:03.3 As you may know,
28:07.0 cytosine, C, in the DNA...
28:09.1 well, methylation is an important modification
28:12.2 that can determine gene expression
28:14.2 and epigenetic memory.
28:17.2 When she performed the measurement
28:20.2 of DNA looping on highly methylated DNA,
28:24.2 she found that that makes the DNA less flexible,
28:27.3 and then also found that nucleosomes,
28:32.0 as a result, becomes weaker,
28:35.0 mechanically.
28:36.2 And then we she examined
28:39.0 additional cytosine modifications,
28:41.1 shown here,
28:42.3 that are important intermediates
28:45.1 while you are trying to remove the methyl groups
28:47.0 from the cytosine,
28:48.2 this makes the DNA much more flexible,
28:50.2 even with a single modification,
28:52.2 and the nucleosome much, much stronger.
28:55.2 So, even here,
28:59.1 we find that
29:02.0 whenever you make changes to the DNA
29:03.3 to make the DNA more flexible,
29:05.2 the nucleosome becomes stronger
29:08.1 and vice versa.
29:10.1 This correlation holds every time
29:13.1 you make a change to the DNA
29:15.1 and so much so that,
29:17.0 if we find an exception in the future,
29:19.2 it'll be an interesting exception.
29:23.2 So, the outlook here is the following:
29:27.2 you know, we now want to
29:30.2 use a high-throughput
29:34.2 genome-wide technology
29:36.0 such as sequencing
29:37.3 so we can measure flexibility of the DNA
29:39.3 across the entire genome,
29:42.1 and because of the correlation
29:44.1 between the flexibility and nucleosome stability,
29:48.1 we can then use that information to
29:50.2 predict how stable individual nucleosomes are
29:54.3 inside the cell,
29:57.2 and then use this as a proxy
30:01.2 to make a connection
30:04.2 between the physical properties of DNA
30:06.1 and the gene expression
30:08.0 and other processes involving chromatin.
30:11.2 And we are hoping to discover
30:14.2 the signatures of the flexibility,
30:17.0 you know, that are imprinted
30:19.1 in our own genome
30:21.0 throughout the evolutionary process,
30:23.0 if, indeed, DNA flexibility
30:25.2 is important for biological functions.
30:31.1 Thank you very much.
30:33.0 This is just a slide to
30:35.1 advertise the third part of this three-talk series,
30:39.2 where we will now use
30:43.2 an even more advanced instrument that combines
30:45.1 ultra-high resolution optical tweezers
30:47.0 that can measure DNA processes
30:49.2 down to a single base pair resolution
30:51.3 with single molecule FRET
30:54.0 to measure conformational changes
30:56.3 of a single helicase enzyme
30:59.2 and its activity all at the same time,
31:02.0 and make important correlations and discoveries.
31:09.3 Many of the lab members
31:12.1 have contributed to these studies,
31:14.0 especially Thuy Ngo,
31:16.0 who did the bulk of the research.
31:19.0 We also acknowledge collaborators
31:21.0 in the lab of Alek Aksimentiev
31:23.2 and also in the lab of in Chuan He,
31:25.2 at the University of Illinois Urbana-Champaign,
31:27.2 and the University of Chicago.
31:29.2 Thank you.

Part 3: Investigating DNA Helicases Using Single Molecule Technologies

00:07.2 Hi.
00:09.0 I'm Taekjip Ha.
00:10.2 I'm a professor of biophysics and biomedical engineering
00:13.0 at Johns Hopkins University.
00:15.1 I'm also an investigator
00:17.0 with the Howard Hughes Medical Institute.
00:19.3 This is part three of a three-talk series
00:22.1 where we discuss how single molecule measurement technologies
00:26.1 are changing the way we
00:28.2 study and understand
00:30.3 nature's amazing nano-machines
00:35.0 -- nucleosomes, helicases,
00:36.2 and many other proteins.
00:40.2 In this part, I will focus on
00:44.2 the studies of DNA helicases.
00:52.1 As you know,
00:53.2 DNA forms a double helix,
00:55.2 and this important paper by Watson and Crick
00:58.0 proposed that structure for the first time.
01:02.0 This paper ended with a paragraph
01:04.2 saying that this helical structure
01:08.1 suggests how DNA information
01:11.1 can be copied for the next cells and so on,
01:17.0 but, you know,
01:19.0 it was written in such a modest way
01:21.1 that the authors actually were worried that
01:25.3 someone may propose the same mechanism
01:28.0 in a more explicit manner
01:29.3 and take the credit for that discovery,
01:31.3 along with the big prize,
01:34.0 so they published another paper in the following months
01:37.2 proposing base pairing between
01:40.3 A and T, and C and G, explicitly,
01:43.3 and the rest became history,
01:46.2 and the big prize, and so on.
01:49.1 The second paper ended with
01:51.2 the following paragraph, shown here.
01:53.2 Even if this is correct,
01:55.3 there is still much to be learned...
01:57.2 for example, if the DNA duplex
02:02.2 is so stably formed,
02:04.2 then what makes the pair of chains
02:07.2 unwind and separate,
02:09.1 because you need to do that
02:11.0 before making a copy of the DNA.
02:14.1 So, it took more than 20 years
02:16.1 before someone discovered that, actually,
02:19.2 there are enzymes that can perform this function
02:21.3 inside a cell,
02:23.2 and these enzymes are called helicases.
02:31.1 As shown in this animation,
02:34.0 helicases can be viewed as
02:37.2 molecular motors that can move on the DNA
02:40.1 in a directional manner,
02:43.1 in the process unwinding the DNA strands
02:46.1 into two single strands,
02:47.3 so we can view it as a DNA unwinder
02:49.2 or DNA unzipper.
02:51.2 If you give a helicase, as shown in the bottom,
02:54.1 a single strand of DNA,
02:56.2 it can also move on the DNA as a motor protein
02:59.2 in a directional manner,
03:01.2 either from the 3' to 5' direction
03:05.1 or 5' to 3' direction...
03:08.0 because single-stranded DNA has a polarity,
03:09.2 one side is called 3', the other side is called 5'.
03:12.2 And this motion had been shown to
03:15.1 require consumption of ATP,
03:18.0 a cellular fuel molecule,
03:20.1 and one base of DNA can be traversed
03:24.1 per one ATP consumed.
03:32.3 There are probably more than
03:35.0 100 different helices in own our genome,
03:37.2 they function in many, many important processes,
03:41.2 shown here.
03:43.1 Essentially, every metabolic process
03:45.2 that requires DNA or RNA
03:47.3 has a requirement for certain helicases.
03:52.2 Helicases can unwind DNA or RNA,
03:57.1 nucleic acids.
03:58.2 They can also remove proteins
04:01.0 bound to DNA or RNA.
04:02.2 They can cause migrations of 4-way junctions of DNA,
04:06.2 they can remodel chromatins,
04:09.2 and they can also remodel RNA-protein complexes.
04:15.1 If these enzymes are defective,
04:17.3 for example, in Werner proteins and Bloom proteins,
04:22.2 then you can get serious human genetic diseases,
04:27.0 such as Werner and Bloom syndromes,
04:29.1 and often times resulting in
04:31.2 genome instability and cancer.
04:35.0 So, it is actually important to
04:37.2 understand how helicases function
04:41.0 and also how their functions are regulated,
04:44.2 because you don't want these enzymes
04:46.3 to bind and unwind every DNA they see.
04:51.2 You need to make sure that
04:53.1 their action is actually controlled,
04:56.1 under control.
04:58.2 You can classify helicases in two different groups.
05:03.2 For example, some helicases form a ring,
05:06.1 forming a hexameric ring,
05:09.1 shown on the left,
05:12.1 and others do not.
05:14.1 Some helicases function on DNA,
05:16.0 others on RNA,
05:18.1 some 5' to 3'
05:20.1 and others 3' to 5'.
05:22.3 What I mean by a 5'-to-3' helicase
05:25.2 is the following:
05:27.2 to see unwinding by such a helicase,
05:30.0 you need to provide
05:32.2 not only the duplexed DNA,
05:34.1 but a single strand DNA tail in the 5' direction.
05:37.2 So, presumably, the enzyme will bind to that tail
05:41.2 and translocate, or motor along that strand,
05:45.1 in the 5'-to-3' direction,
05:47.2 and displacing the top strand,
05:49.2 therefore unwinding the DNA into two strands.
05:53.1 So, today, I'll focus on our studies
05:56.2 performed on non-hexameric DNA helicases
06:00.2 that unwind the DNA
06:02.3 in the 3'-to-5' direction,
06:04.2 so these guys will bind on the 3' tail
06:07.2 and translocate on the 3'-to-5' direction,
06:11.1 causing DNA unwinding.
06:17.1 So, I'm showing you three names
06:19.1 -- Rep, UvrD, and PcrA --
06:21.1 these are enzymes belonging to the same family,
06:26.1 3'-to-5' DNA helicases,
06:28.2 and they are very similar in terms of their
06:31.1 3D atomic structures,
06:33.2 and here is an example of a structure.
06:37.2 These enzymes function in bacterial replication restart,
06:42.1 DNA repair,
06:44.1 and also phage replication.
06:46.1 And on the bottom
06:48.1 I will show you the names of these enzymes
06:50.0 and whenever I'm talking about a specific enzyme
06:52.3 then I'm going to use to green to denote that enzyme.
06:59.1 The questions that we want to address in this talk
07:01.2 are the following.
07:03.2 First, you know,
07:05.1 what is the functional form of the enzyme?
07:07.2 And there's a discussion whether it's a monomer,
07:11.2 so, a single protein,
07:12.3 or two proteins, a dimer,
07:16.0 functioning to unwind the DNA.
07:18.1 The second subquestion is the following:
07:22.2 the enzyme is known to have
07:24.3 two widely different conformations,
07:26.2 as shown in this animation,
07:28.3 one domain can swivel around the vertical axis by 130°
07:32.2 and obtaining and switching
07:35.1 between closed and open conformations...
07:38.2 so, of the two,
07:40.1 which one is functional for the unwinding?
07:42.0 If you know what that is,
07:43.2 then why do you have the other conformation?
07:47.1 The next question is, you know,
07:48.3 how is unwinding regulated?
07:50.2 Again, you don't want these enzymes
07:52.1 to unwind everything that they see,
07:54.0 you want to turn the activities off, mostly,
07:57.3 until you need to activate these enzymes.
08:00.2 So, the question is, you know,
08:03.1 how do tell the enzymes, you know...
08:06.1 yeah, okay, now you have a go signal,
08:07.3 you go and unwind the DNA,
08:09.2 or, you know, no-go signal,
08:11.0 just stop here and wait until the signal comes.
08:18.1 So, we'll be using a technology
08:20.1 called single molecule FRET
08:22.1 for some of the studies I'll describe today.
08:24.2 In FRET, you can label
08:27.1 two different parts of a biomolecule
08:29.0 using green and red dye molecules,
08:32.0 so that, depending on the distance
08:35.2 between the two dye molecules,
08:37.0 you can get different intensities
08:39.3 of the two colors.
08:41.1 When the two dyes are very close to each other,
08:43.2 in this Horseback position,
08:45.3 you get strong FRET and you get mainly red emission,
08:48.2 shown here,
08:50.1 whereas if they are far away from each other,
08:52.0 shown here in the Cowboy position,
08:55.1 then you get mainly green emission,
08:56.2 because when you excited the green molecule
08:58.1 using the laser,
09:00.0 energy does not get transferred efficiently
09:02.1 to the red molecule,
09:04.0 so you get mainly green emission.
09:05.2 So, that is the idea.
09:07.0 You can follow the conformational changes,
09:09.1 or shape changes of a molecule,
09:11.1 as a function of time using single molecule FRET.
09:18.2 For example, you can use FRET
09:20.1 to measure the movement of a helicase
09:23.1 on a single strand of DNA
09:25.0 using ATP consumption as a fuel source.
09:26.3 And, you know, as you move the protein along the DNA
09:31.0 you can expect to see the
09:33.1 green signal going down
09:34.2 and the red signal going up, and so on,
09:35.3 and that is exactly what we see
09:38.0 in the single molecule data
09:39.1 -- you can see the green signal going down,
09:41.1 red is going up gradually,
09:42.2 it's a little bit noisier than the animation,
09:44.2 but still unmistakable signatures of the movements.
09:58.1 So, coming back to the helicases
10:00.1 -- Rep, UvrD, PcrA helicases --
10:04.1 as I told you,
10:06.2 there are two widely different structures.
10:08.1 One is called the open, shown here,
10:11.0 the other is called the closed.
10:13.3 The two structures are identically, essentially,
10:16.1 regarding three of the subdomains
10:21.1 -- green, yellow, and red --
10:24.2 between the two conformations,
10:26.2 but they are different in the orientation
10:28.3 of the blue domain,
10:30.2 called the 2B domain.
10:33.0 So, if you want to visualize it,
10:35.1 you can imagine a vertical axis, here,
10:39.1 and then you can swivel the blue domain
10:42.0 around the vertical axis by 130°,
10:44.2 like that,
10:46.1 to obtain the closed conformation.
10:48.1 This is a big movement,
10:51.1 you know, highly unusual in the biological structural literature.
10:56.1 Some of the distances between two sites
10:58.1 on the same protein
11:00.0 can change by more than 50 Angstroms
11:02.1 -- huge movements --
11:04.1 so there must be a meaning behind these movements,
11:07.2 and we are interested in finding out what that is.
11:13.0 So, in this structure shown on the right,
11:15.2 to the closed form of the helicase
11:20.0 you see DNA bound to it,
11:23.0 in fact, this DNA has a double helical structure,
11:25.0 plus the 3' tail.
11:28.0 That is exactly the DNA that enzyme
11:30.1 is able to unwind.
11:33.0 So, unsurprisingly,
11:35.0 this led to the proposal that
11:36.2 the closed form is active in DNA unwinding,
11:39.0 basically, using the ATP
11:41.2 to pull in the single-stranded DNA tail
11:44.2 one base at a time,
11:46.2 causing the DNA to unwind.
11:49.2 And this is a proposal actually made
11:53.0 by many groups,
11:56.0 as shown in this animation,
11:58.1 combining the snapshots of the crystal structure
12:00.2 of the UvrD helicase
12:03.1 bound to the DNA in different ATP states.
12:06.2 And they led...
12:08.2 this led to the proposal that the closed form
12:11.2 is active for unwinding as a monomer,
12:13.3 and then also,
12:15.3 this 2B domain that swivels around,
12:18.1 contacts the duplex DNA,
12:20.2 and this motion can cause a twisting-opening of the DNA,
12:25.1 one base pair at a time.
12:27.2 And these studies also led to the proposal
12:30.1 that the other form,
12:32.0 the open form,
12:34.0 is important for protein displacement.
12:36.2 But based on biochemical data
12:39.2 and single molecule studies
12:42.2 that my collaborators and we have performed,
12:45.2 we had a different view that actually
12:49.0 the closed form is the inactive form,
12:51.1 it's the off state,
12:52.2 it's the no-go signal
12:54.1 telling the enzyme that now you have hit duplex DNA...
12:58.0 stop going there
13:00.2 until there is a signal to continue to move.
13:04.3 And, you know, some of the reasons
13:07.1 why we thought this is the case
13:08.2 are shown on this slide.
13:12.2 So, we decided to test, you know,
13:17.1 this idea in the most direct way,
13:20.2 using the following approach.
13:24.2 I'm again showing you
13:26.3 the two different structures of these enzymes
13:30.1 -- we are using the helicase as an example here --
13:33.0 closed and open,
13:35.3 and we put two cysteines
13:40.2 into the protein domains,
13:42.1 shown on the [left],
13:44.0 shown in the red residues,
13:46.0 so that in the closed form
13:48.0 they are very close to each other,
13:50.0 only 7 Angstroms apart.
13:52.2 So, you can use a bifunctional crosslinker
13:56.1 to link the two cysteines chemically,
13:59.3 but in the open form
14:03.1 these two cysteines are 42 Angstroms apart,
14:05.2 so you cannot actually obtain the open form
14:08.0 once the closed form is achieved,
14:10.0 using the crosslinking reagents.
14:12.1 So, that construct
14:17.1 will give you a protein that is only and always
14:21.1 in the closed form
14:23.1 and we call this Rep-X
14:25.1 #NAME?
14:28.3 As a control, we made another helicase mutant,
14:32.1 where we put two cysteines in such a way,
14:34.2 that's shown in here,
14:36.1 the orange residues,
14:38.2 that are close to each other in the open form
14:40.2 but they are far apart from each other,
14:43.2 20 Angstroms apart,
14:45.1 in the closed form.
14:47.2 So if you crosslink the protein using the two cysteines,
14:50.0 you can only see the open form
14:52.1 but not the closed form.
14:57.2 Again, our prediction was that
15:05.2 this closed form of the enzyme, Rep-X,
15:07.3 will be inactive in DNA unwinding
15:09.3 and the open form of the enzyme
15:11.3 will be active for DNA unwinding.
15:16.2 And because of my prediction,
15:18.1 I was surprised when my student came to my office,
15:21.2 told me that he saw just the opposite,
15:24.2 okay?
15:26.2 He couldn't see any unwinding by this open form,
15:30.0 crosslinked form,
15:31.2 he saw unwinding from this one.
15:35.2 In fact, unwinding was much, much better
15:38.1 once you crosslink the protein into the closed form,
15:40.2 so Rep-X was a much, much better helicase
15:44.0 than, you know, than the wild type.
15:47.3 With the wild type, we couldn't see any unwinding
15:50.0 by a single monomer of the helicase,
15:51.2 but a monomer of Rep-X
15:53.2 could unwind DNA quite well,
15:55.1 according to my student.
15:57.0 So, I was surprised,
15:59.1 but if the better data says that you're wrong,
16:03.2 that's fine,
16:05.1 so we quickly changed our model
16:08.2 to propose that the closed form is active
16:11.3 and the open form is inactive.
16:15.2 And then...
16:19.1 well, okay, wait, the lesson of the story
16:21.0 is to never bet against a structural biologist.
16:25.1 Then we asked ourselves, okay,
16:27.3 if this enzyme becomes an active enzyme
16:30.0 as a monomer,
16:31.1 which is much better than the wild type,
16:33.1 how much better is this?
16:35.1 So we used the optical trap,
16:38.0 the optical tweezers assay,
16:40.0 to measure the activity of Rep-X.
16:45.1 Here, we have two laser traps,
16:48.1 and one is holding a bead
16:51.0 with a long DNA molecule,
16:53.3 6000 base pair long double-stranded DNA molecule
16:56.0 with a 3' tail
16:59.0 that is needed to start the winding reaction,
17:02.0 and the second bead has an antibody
17:05.0 that is holding our Rep-X protein using,
17:08.2 you know, an epitope,
17:10.1 and then when Rep-X binds to the 3' tail
17:14.2 you can form a tether,
17:16.1 and then you can move the tether
17:18.2 that formed into a chamber
17:21.2 that contains ATP, the fuel molecule,
17:24.1 to start the unwinding reaction.
17:27.3 As the helicase unwinds the DNA,
17:30.1 that helicase will move to the left,
17:33.2 unwinding more and more DNA,
17:36.1 and the bead will follow the helicase,
17:38.1 so the distance between the two beads
17:40.1 will decrease as a function of time,
17:42.2 as the helicase unwinds the DNA,
17:45.3 and this is exactly what we saw.
17:47.2 So, initially, we formed a DNA tether
17:51.1 between the two beads
17:53.0 through a DNA/helicase linkage,
17:55.1 and then you move into the ATP channel
17:58.2 in a microfluidic chamber
18:00.2 to start the unwinding reaction.
18:02.0 So, when you do that,
18:03.3 as shown here from the yellow region
18:05.1 to the blue region,
18:07.0 DNA lengths starts to change,
18:09.0 they start to shorten
18:12.1 as soon as you go into the ATP chamber,
18:14.2 and this unwinding reaction continues
18:16.2 over thousands of base pairs
18:18.1 without falling off.
18:19.2 So, this was a complete surprise
18:21.1 because none of the other enzymes of this type
18:24.2 had such, you know,
18:27.1 amazingly processive activity.
18:30.0 Processivity, here,
18:32.2 is just defined as how many base pairs of DNA
18:35.1 are unwound before the enzyme falls off,
18:37.2 and here it's limited only by
18:39.2 the length of the DNA.
18:41.1 In principle, if you had a much longer DNA,
18:44.0 we suspect the enzyme would be able to unwind
18:46.2 that DNA as well.
18:48.1 So, that was a big surprise
18:51.2 and, you know,
18:53.2 we concluded that the Rep-X protein is a superhelicase.
18:57.2 It can unwind DNA for
18:59.2 many thousands of base pairs without falling off,
19:02.0 and we also showed that it can do this
19:05.0 even against a strong opposing forces.
19:07.1 It's robust against the opposing force.
19:16.1 Okay, so we now have data
19:18.3 that shows that the closed form is active
19:20.2 and, in fact, it becomes a superhelicase,
19:23.1 but why did nature create the open form?
19:26.1 Why do you need to have the other form
19:28.1 that is not active for DNA unwinding?
19:32.1 To address this question,
19:34.0 we turned to a different instrument
19:36.2 that combines ultra-high resolution optical tweezers
19:41.3 that can measure DNA unwinding
19:43.2 at the single base pair resolution
19:46.0 with single molecule FRET.
19:47.1 You can use the two laser traps,
19:50.2 here and here, to, you know,
19:52.3 trap the beads and apply force,
19:55.1 and then use
19:57.3 a laser light in the middle
19:59.1 to illuminate a fluorescent molecule on the helicase
20:02.1 to measure how many copies of the helicase
20:05.1 are sitting on the DNA at a given time,
20:08.1 and also you can use FRET,
20:10.0 if you have two labels of different colors,
20:12.0 to measure the helicase conformation
20:14.0 at the same time as you measure DNA unwinding
20:16.2 using the optical tweezers.
20:21.0 Here's a more detailed schematic.
20:24.1 We call this the hairpin assay.
20:27.1 We have a DNA hairpin that forms
20:29.1 in the middle of the DNA
20:31.1 and there's a bit of DNA single strand region,
20:33.1 right here,
20:35.2 to which the helicase can come and bind,
20:40.2 and then this can cause
20:44.1 unwinding of the hairpin region,
20:47.0 as time goes on,
20:48.3 and this will lengthen the DNA tether
20:51.1 between the two beads,
20:52.2 then you can read the reaction progress
20:54.3 as the length change
20:57.2 as a function of time.
20:59.3 So that's how you read out the helicase activity.
21:01.2 But at the same time you can
21:05.1 focus the fluorescence laser excitation
21:07.3 in the middle of the DNA tether
21:09.2 and then measure fluorescence
21:12.1 coming from the helicase.
21:14.0 If you have a single dye per helicase,
21:16.1 then you can count the number of dyes
21:19.1 in the middle
21:21.1 to count how many proteins are functioning there
21:22.3 and we can show that,
21:24.2 you know, that single protein can unwind
21:26.2 maybe 10-20 base pairs of DNA
21:28.1 when there is no...
21:32.1 when there's external tension applied to the DNA.
21:35.1 You can also use two colors
21:37.2 to measure conformational changes
21:39.2 through single molecule FRET.
21:42.2 So, let me show you the actual data.
21:44.2 Here, we are using the UvrD helicase,
21:48.0 and then we labeled the protein
21:50.2 with two different dye molecules in such a way that,
21:54.0 in the closed conformation shown on the top,
21:56.3 the two dyes are close to each other,
21:58.1 so you get high FRET,
21:59.2 you get mainly red emission,
22:01.3 in the open conformation they are far away from each other,
22:04.1 you get mainly the green emission in the open...
22:08.0 and low FRET efficiency.
22:11.0 On the top, I'm showing you
22:13.3 DNA base pairs unwound as a function of time.
22:16.3 I told you that as the hairpin unwinds
22:19.0 the two beads are becoming more and more separated,
22:21.2 so the distance between the two will increase.
22:25.0 And we found that a single UvrD monomer
22:28.1 can bind and unwind 10-15 base pairs,
22:31.1 but then it comes back
22:34.0 to rezip the DNA and it unwinds the DNA again,
22:36.2 it rezips the DNA,
22:37.3 and this can happen multiple times.
22:42.0 So it seems to sit there,
22:45.0 like, unwinding a little bit and coming back,
22:47.2 and unwinding a little bit and coming back, and so on.
22:50.1 It turns out that the reverse direction
22:52.1 also depends on the ATP concentration
22:54.1 in terms of its speed,
22:55.3 so we think that it's not just
22:58.2 sliding back on the same strand
23:00.0 but it's moving to the other strand
23:02.1 and walking back on the other stand.
23:04.2 Now, if you look at the fluorescence data
23:07.1 in the same experiment,
23:09.2 the middle panel shows you the raw data
23:12.2 of the green and red intensities,
23:14.1 the bottom shows you
23:16.2 the calculated FRET efficiency.
23:18.1 You know, a high FRET value
23:20.2 means that the two dyes are close to each other,
23:22.1 high proximity,
23:23.2 and vice versa.
23:25.1 Now, we can begin to interpret the DNA.
23:27.1 On the bottom slide,
23:29.1 low FRET state,
23:31.0 we can interpret that as
23:33.1 an open conformation of the enzyme,
23:35.0 and the high FRET state, shown here,
23:37.1 you know, we interpret that as
23:39.1 the closed conformation.
23:41.1 So, the molecule is switching back and forth
23:43.1 between the closed and open conformations
23:45.2 as time goes on.
23:48.2 We can begin to make correlations
23:50.2 between FRET and unwinding
23:52.3 by dividing the time trajectories
23:55.0 into two sections.
23:58.1 In the first section,
24:00.2 we have high FRET values, colored in red.
24:02.3 In the second section,
24:05.1 we have low FRET values, colored in green.
24:07.2 Now, if you look at the top graph,
24:10.1 whenever you see a red color
24:12.2 -- high FRET, closed conformation --
24:15.0 the distance tends to increase as a function of time,
24:18.2 so DNA is being unwound by the enzyme,
24:21.2 but whenever you see a green color
24:23.1 the distance tends to decrease
24:26.1 during the...
24:27.2 meaning that DNA is being rezipped.
24:31.3 So, this means that
24:34.1 a high FRET conformation, closed form,
24:35.3 is needed for DNA unwinding
24:38.2 and the open form is needed for DNA rezipping.
24:45.0 You can actually analyze data of this type
24:47.1 for many, many molecules
24:49.2 and plot the two-dimensional histogram.
24:53.1 The x axis here is FRET
24:54.3 and the y axis velocity,
24:57.0 so whenever you have high FRET values
24:59.1 then velocity tends to be positive,
25:01.3 meaning that DNA is being unwound.
25:04.2 Whenever you have low FRET values,
25:06.2 you have negative velocity,
25:08.1 meaning that DNA is being rezipped.
25:11.1 So, we had to revise our model yet again.
25:15.1 I told you that based on,
25:17.2 you know, superhelicase data,
25:20.1 we proposed that this high FRET state,
25:24.2 closed conformation is active for unwinding
25:27.1 and the open state is inactive for unwinding,
25:30.3 it's a no-go signal,
25:32.1 but instead actually
25:34.2 what we have is a U-turn signal.
25:36.2 So, when you take the open conformation,
25:38.1 it is a signal to take a U-turn,
25:41.0 to walk back on the DNA,
25:42.2 to let the DNA rezip.
25:48.1 So, this is the U-turn model,
25:52.2 where this enzyme
25:55.0 can walk from the 3'-to-5' direction
25:58.1 as a motor protein
26:00.1 while it is in the closed form,
26:02.1 causing the DNA to be unwound.
26:04.2 But after unwinding 10-15 base pairs,
26:07.2 it changes to the open conformation,
26:10.1 that we call the U-turn signal,
26:13.2 and then this allows the DNA helicase
26:16.2 to change strand from the 3' strand
26:19.1 to the 5' strand,
26:21.2 and the enzyme then moves on the 5' strand downward,
26:25.1 causing the DNA to rezip.
26:28.3 And then this can switch again
26:30.3 and switch again
26:32.2 and it can go back and forth many, many times.
26:35.1 So, it looks like nature has made
26:38.1 this open conformation
26:40.2 to disallow the protein from...
26:43.3 to go far deep into the DNA duplex
26:46.1 until there is a signal to unwind the DNA.
26:49.3 In fact, we have data that shows that
26:51.2 if you add a partner protein
26:53.2 that is known to make the enzyme processive
26:55.3 and active as a helicase,
26:57.2 then you stabilize the closed conformation,
27:00.1 thereby you prevent the formation
27:02.2 of the open conformation required for switching.
27:05.2 This also explains why
27:10.3 Rep-X crosslinked into the closed form
27:12.1 becomes a superhelicase.
27:14.0 It cannot take a U-turn;
27:15.3 it just keeps going in the same direction.
27:18.1 You know, it's not a crazy...
27:20.1 it's not a Porsche or a Ferrari,
27:22.2 it's probably a Honda,
27:24.2 but it just keeps going without taking a U-turn,
27:30.1 and that persistence makes it a superhelicase.
27:39.2 We are now pursuing
27:42.0 several biotechnological applications
27:45.0 because you have a unique enzyme, it's a monomeric superhelicase.
27:48.0 Monomeric means that it's easy to handle
27:51.1 and easy to design
27:55.3 assays using this enzyme
27:57.2 and it doesn't have any,
27:59.1 you know, associated nuclease activity,
28:01.2 so it doesn't chop the DNA
28:03.1 that you are trying to unwind,
28:05.0 unlike some other enzymes.
28:06.2 So we are now pursuing applications
28:09.2 of using the superhelicase
28:11.1 to sequence single DNA molecules
28:13.2 through a nanopore,
28:15.0 because our enzyme is very robust
28:17.2 against applied force.
28:19.0 You know, the hope is that
28:22.0 this can be used
28:24.1 to sequence thousands of basepairs of DNA
28:26.3 in one shot,
28:29.1 giving you a very good signal-to-noise ratio
28:32.3 when you are...
28:34.3 still able to apply a strong voltage
28:36.1 to increase your current through a nanopore.
28:39.1 We can also use it to amplify DNA
28:41.2 at a constant temperature,
28:43.2 isothermal DNA amplification.
28:46.1 You know, in normal DNA amplification
28:48.1 using PCR machines,
28:50.1 you need to go up and down in temperature 30 times,
28:52.3 because every time you make a copy of the DNA
28:55.2 you have to melt the DNA before you start the next cycle,
28:59.0 to allow the primers to bind to the DNA,
29:01.3 but if you have a superhelicase
29:04.0 that can unwind the DNA for you,
29:05.2 then you don't need to do thermocycling.
29:07.2 You can just keep everything at a constant temperature
29:09.3 and then let the amplification continue.
29:14.2 So, you don't need expensive equipment,
29:16.1 you may not even need electricity,
29:18.3 to detect pathogenic DNA
29:22.1 from viruses and bacteria
29:24.1 in, you know, Africa and so on,
29:26.1 so that's one area
29:28.2 that will be very exciting to look into.
29:35.1 Finally, I just want to mention that,
29:38.0 you know, we don't have to stop here,
29:39.3 we can go into even more complex systems,
29:42.3 that I call multidimensional single molecule measurements.
29:47.2 You know, an example would be DNA replication.
29:49.2 Amazingly rapid,
29:52.2 amazingly accurate,
29:55.1 it has many important parts that are required for function.
29:57.3 You can imagine using
30:01.1 multi-axis optical traps
30:03.2 to apply forces and measure the progress
30:06.1 of the replication machinery
30:08.2 at the single base pair resolution,
30:11.1 and at the same you label different components
30:14.1 of different proteins with different colors
30:16.2 to determine how many copies
30:18.3 of which proteins are present
30:21.1 at every reaction step
30:24.1 and also to measure their conformational changes,
30:26.1 and make correlations between different properties,
30:29.1 all at the same time.
30:31.2 So, we are not there yet,
30:33.1 but we have made important progress in that direction.
30:36.1 I just told you we now have,
30:38.2 you know, single base pair resolution optical tweezers
30:41.2 that can measure single molecule fluorescence at the same time,
30:43.2 and recently my collaborators in Korea
30:49.0 advanced the FRET technology into four colors,
30:53.1 that makes it possible to measure
30:55.1 not only one distance as a function of time,
30:57.0 but up to six distances as a function of time,
31:00.1 all at the same time.
31:03.3 On a, you know,
31:07.0 more global scale,
31:09.1 we can also include
31:12.1 other single molecule methods.
31:14.2 You know, so far I have talked about fluorescence
31:17.1 such as FRET,
31:18.3 and force-based methods such as optical tweezers,
31:22.2 but you can also imagine using
31:24.1 perhaps the first single-molecule measurement technology,
31:26.3 you know, patch clamps and so on
31:29.1 that can measure molecular activities
31:32.0 through electrical measurements,
31:33.3 through channels and pores and so on.
31:36.1 You can also do single molecule measurements
31:38.0 essentially inside a computer, in silico,
31:40.1 by running full-atom simulations
31:42.3 using, you know, computers.
31:48.1 As we become more ambitious
31:50.2 in trying to understand
31:52.3 ever more-complex systems,
31:54.2 we need to actually combine
31:57.1 different measurement methods into one instrument
31:59.3 so that we can maximize the information content of analysis
32:05.2 and, you know,
32:08.1 really give additional confidence
32:10.1 to the data interpretation.
32:11.2 So, I've told you only about
32:13.2 the top arrow, here,
32:15.2 combining fluorescence and force,
32:17.1 but there are many, many other research labs
32:19.2 trying to combine different approaches
32:22.1 in many different combinations,
32:24.1 and making really exciting and important discoveries.
32:31.3 Okay, so that's all.
32:34.3 Thank you very much for listening to my talk.
32:38.2 In this particular talk,
32:41.0 we had strong collaborations
32:43.2 with the labs of Yann Chemla at the University of Illinois Urbana-Champaign,
32:47.2 who specializes in ultra-high resolution optical tweezers
32:50.3 in combination with fluorescence,
32:54.2 and Tim Lohman in Washington University School of Medicine in St. Louis.
32:59.1 He's a long-time collaborator
33:01.1 on the studies of helicases.
33:03.2 Sinan Arslan was the main person
33:06.1 who did the superhelicase work
33:08.1 and Matt Comstock was the main person
33:11.2 who performed combined optical trap
33:16.1 and fluorescence measurements
33:18.1 of the UvrD helicase,
33:19.2 who is now a faculty member of
33:21.3 Michigan State University in the physics department,
33:25.1 and then additional people from the lab
33:27.1 who made contributions in the past.
33:29.0 Thank you very much.

Videos in this Talk
  • Part 1: Developing Single Molecule Technologies to Study Nanomachines
    Part 1: Developing Single Molecule Technologies to Study Nanomachines
    Audience:
    • Researcher
  • Part 2: Combining FRET and Optical Trap to Study the Nucleosome
    Part 2: Combining FRET and Optical Trap to Study the Nucleosome
    Audience:
    • Student
    • Researcher
  • Part 3: Investigating DNA Helicases Using Single Molecule Technologies
    Part 3: Investigating DNA Helicases Using Single Molecule Technologies
    Audience:
    • Researcher
    • Educators of Adv. Undergrad / Grad
Speaker: Taekjip Ha
Total Duration: 1:34:12
Recorded: May 2016
All Talks in Biophysics

Talk Overview

Recent improvements in single molecule technologies have enhanced our understanding of nanomachines; the cells’ tiny engines. Dr. Taekjip Ha explains how scientists have used fluorescence microscopy, in combination with other tools, to study nanomachines. For example, by tagging the protein myosin with a fluorescent dye, his laboratory was able to provide conclusive evidence that this protein does not “crawl”, but “walks” along actin. Ha explains the basis of Fluorescence Resonance Energy Transfer (FRET) and how it can be used to measure conformational changes within a protein (intramolecular) or movement between two molecules (intermolecular).

In his second lecture, Ha shows the value of combining FRET with optical trapping. An optical trap uses focused light to “grab” a particle and “pull” allowing scientists to measure the force required to move a molecule. Ha’s lab used this combination of technologies to study the dynamics of nucleosome binding to DNA. His lab was able to show that nucleosome “unwrapping” from the DNA is asymmetrical, unidirectional, and sequence dependent. Combining FRET and optical trapping technologies with other tools, like whole genome sequencing, scientists can predict novel areas of high transcriptional activity.

In Part 3, Ha explains how his laboratory used optical tweezers (a variation on an optical trap) to study helicases and their regulation. Their data confirmed that the closed conformation of the 3’-5’ DNA helicase is the active form of the enzyme. They combined optical tweezers with FRET to further understand the role of the open conformation of the helicase. Ha encourages scientists to continue combining these single molecule technologies to better understand nature’s complex nanomachines.

Speaker Bio

Taekjip Ha

Taekjip Ha

When Taekjip Ha moved from Korea to the United States to begin graduate school, he also transitioned from being a theoretical physicist to being an experimental biophysicist.  During his last few months of graduate school, he developed fluorescence resonance energy transfer, or FRET, a technique he has since used to investigate many questions in cancer… Continue Reading

Playlist: Imaging in Biology

Related Resources

Part 1:
Ha, T.  Enderle , D.F. Ogletree, D.S. Chemla, P.R. Selvin, S. Weiss. Probing the interaction between two single molecules: fluorescence resonance energy transfer between a single donor and a single acceptor. Proceedings of the National Academy of Sciences of the United States of America 93, 6264-6268 (1996).

Yildiz, J.N. Forkey, S.A. McKinney, T. Ha, Y.E. Goldman, P.R. Selvin. Myosin V Walks Hand-Over-Hand: Single Fluorophore Imaging with 1.5-nm Localization. Science 300, 2061-2065 (2003).    

Joo, H. Balci, Y. Ishitsuka, C. Buranachai, T. Ha.  Advances in Single-Molecule Fluorescence Methods for Molecular Biology. Annual Review of Biochemistry 77, 51-76 (2008).    

Part 2:
C. Neuman, S. M. Block. Optical trapping. Review of Scientific Instruments 75, 2787-2809 (2004).

T. M. Ngo, Q. Zhang, R. Zhou, J. G. Yodh, T. Ha.  Asymmetric Unwrapping of Nucleosomes under Tension Directed by DNA Local Flexibility. Cell 160, 1135-1144 (2015).

Vafabakhsh, T. Ha, Extreme bendability of sub-100 bp long DNA revealed by single molecule cyclization. Science 337, 1097-1101 (2012).

Part 3:
M.R. Singleton, M.S. Dillingham, D. B. Wigley. Structure and Mechanism of Helicases and Nucleic Acid Translocases. Annual Review of Biochemistry 76, 23-50 (2007).

S. Arslan, R. Khafizov, C. D. Thomas, Y. R. Chemla, T. Ha, Engineering of a superhelicase through conformational control. Science 348, 344-347 (2015).

Reader Interactions

Leave a Reply

Your email address will not be published. Required fields are marked *

iBiology and iBiology Courses are part of the Science Communication Lab (SCL). Our mission remains the same, to connect people to science. However, our focus has shifted to producing and evaluating cinematic films for education and the public, which you can find on the Science Communication Lab website. For more information, please see this blog post!