Top

iBiology

Bringing the World's Best Biology to You

Total Internal Reflection Fluorescence (TIRF) Microscopy

00:00:11.10 We're going to talk about total internal reflection fluorescence
00:00:14.10 microscopy. The purpose of total internal reflection fluorescence
00:00:19.23 microscopy, otherwise known as TIRF, is to selectively illuminate
00:00:26.28 fluorophores that are right near a surface. And not illuminate
00:00:31.16 fluorophores that are further into the solution, above the surface.
00:00:35.14 The main reason for doing that is in two applications. One is
00:00:43.13 in single molecule studies, where you want to see the molecules
00:00:48.23 right on the surface, but you don't want to see a lot of background fluorescence.
00:00:52.12 And the other main use is to look at living cells in culture, where
00:00:59.27 you might be interested in the cell substrate contact region, and
00:01:04.16 many organelles that are near there. But you don't want to see
00:01:08.13 the fluorescence from further in the cell. So in both cases,
00:01:13.00 there's a big background fluorescence reduction. And that's the main
00:01:17.00 purpose of TIRF. There are other purposes, too. Which I will get
00:01:21.21 to. But first, I will give some examples, then after the examples,
00:01:28.16 we'll talk about the fundamental principles of TIRF, how to set it up,
00:01:33.05 what limitations there are, and other useful features other than background
00:01:39.01 reduction. And then there's an important question about what does
00:01:43.17 an objective, particularly a high aperture objective in microscope
00:01:48.19 actually see of the fluorescence that's excited. And then the last topic
00:01:54.25 will be looking at the polarization of the field that's produced by TIRF.
00:02:03.25 Because it allows you to study orientations of fluorophores.
00:02:08.22 So we'll start with the examples.
00:02:12.28 This series of pictures is of cells, in this case, cells that are designed
00:02:25.17 to secrete little packets of chemicals. And these packets in this
00:02:33.22 particular sample are labeled with GFP. So you can see where
00:02:39.14 the packets are. The packets are called granules, these are chromaffin cells.
00:02:44.15 Those granules though, are all through the depth of the cell.
00:02:50.24 So, on the bottom here, if you look at the pictures on the two lower
00:03:00.14 parts. That is regular illumination of this kind of preparation.
00:03:06.06 Regular illumination is called epi, epi illumination. And the problem
00:03:13.27 there is that it's very fuzzy. You see a lot of out of focus fluorescence
00:03:20.01 and it's hard to resolve an individual granule because of the low contrast
00:03:26.10 with the large background. With TIRF, you can see more clearly
00:03:31.08 the individual granules and it's because only the granules that are right near the
00:03:37.08 surface are showing up in the picture. Here's another example, the bottom again
00:03:45.04 is Epi illumination. And these are cells that are labeled with diI, which
00:03:52.09 completely labels the plasma membrane on the bottom, on the
00:03:55.29 top, and even some membranes inside. And so in Epi, you don't see
00:04:00.15 that much detail, you just see where the cells are. And it's very bright
00:04:04.14 with a nice edge to it, but with TIR toward the top, you only see
00:04:10.29 the cell substrate contact regions. Here we're going to talk about
00:04:16.05 the principles that TIRF microscopy works on. What's necessary
00:04:23.29 is to illuminate the sample with a beam at a very oblique angle, and the
00:04:29.17 beam could either be a laser or a conventional source, but usually a laser.
00:04:33.25 The beam is coming in here, coming up through the glass to a
00:04:40.11 glass/water interface. And then if the angle of incidence, this angle
00:04:46.06 theta, is big enough, it will totally internally reflect. All the light will
00:04:53.02 reflect back from the interface. At that angle and larger angles, you don't
00:05:01.14 get light propagating out into the water. But you do get a field right above
00:05:10.16 the surface in the water that's exponentially decaying as you go away
00:05:15.09 from the surface in what we call the z direction. And the depth of that decay
00:05:21.07 is very small, it can be as small as 1/5 or 1/10 of the wavelength of light.
00:05:27.23 Depending on how big the incidence angle theta is. So, what you can see
00:05:39.09 here is that if there's a fluorescent molecule, say out here somewhere,
00:05:44.08 it won't be in that field. If it's closer, it will be in the field but in a dimmer
00:05:49.19 part and it will light up a little bit. And if it's really close to the surface,
00:05:53.23 it will get excited easily and be quite bright. So, this field is called the
00:06:03.02 evanescent field. Evanescent means disappearing. And it's not disappearing
00:06:08.14 in time, it's always there, it's just disappearing in space,
00:06:11.27 as you go up in the z-direction a fraction of a wavelength.
00:06:18.08 So, how do you set this up in a microscope? There are various ways.
00:06:23.18 There are two main types of ways, one is with a prism that you
00:06:30.14 buy and install externally. And the other is using the microscope objective
00:06:39.23 itself. So first we'll talk about the prism based method. One way is in an
00:06:45.05 upright microscope, it's very simple. Just put something like a trapezoidal
00:06:50.16 prism on the condenser mount of the microscope, and then the beam comes in
00:06:56.11 through the lower part of the microscope. Reflects up through the base
00:07:02.01 and the totally internally reflects at the upper surface of that trapezoid,
00:07:08.12 which is in optical contact with the bottom of a glass bottom tissue culture
00:07:16.12 dish, where the sample is. And from the top, you look at the sample
00:07:22.25 with an air or water immersion objective. You could do something very
00:07:29.14 similar to this, except instead of a prism, you just use a high aperture
00:07:34.12 condenser. In fact, this system is commercially available.
00:07:38.27 If you use an inverted microscope in the prism based system,
00:07:44.16 the objective is below and the prism is above. And there are various
00:07:49.18 types of prisms you could use. You could even have a system where
00:07:56.20 it's a prism based one in an inverted scope, but the sample is
00:08:01.24 completely accessible by coming in from the bottom, going through
00:08:07.08 a prism, and then bouncing back and forth, back and forth, and
00:08:12.00 totally internally reflecting right over the optical axis of the objective.
00:08:15.26 And there's still plenty of access for micropipette or changes of
00:08:21.07 solution. But actually, the most popular way of doing TIRF nowadays is
00:08:29.10 objective based TIRF, where you don't need an external prism
00:08:33.10 at all. But you do need a very high aperture objective. The idea here
00:08:38.22 is as follows. There is a plane in back of the objective, some plane
00:08:46.04 actually in the objective unit itself, called the back focal plane.
00:08:51.08 If the laser beam passes through that back focal plane
00:08:55.21 off axis, in other words, not coming right down the middle,
00:09:00.03 but instead, coming off to the side. Then when it gets up to the objective,
00:09:06.01 it will emerge from the objective at a steep angle. And if the objective's
00:09:13.07 width or diameter or aperture is big enough, then that angle will be
00:09:21.05 above the critical angle for total internal reflection. And then you just
00:09:25.19 look through a coverslip or a glass bottom tissue culture dish, and you'll get
00:09:31.23 total internal reflection. And the same objective is used for gathering
00:09:35.29 the fluorescence that's emitted. The requirement here is not just
00:09:42.16 that you're passing the beam off axis, so you get the angle, but
00:09:48.24 also that the beam is focused at the back focal plane. Because
00:09:54.10 if it's focused there, then the light that emerges from the objective
00:09:58.05 will come out collimated. In other words, all the rays will be parallel
00:10:01.27 and all going at the same angle. And you want them all at the same angle, because
00:10:06.14 you want them all to be totally internally reflecting. So, there are two requirements:
00:10:12.00 focus off axis far enough with a high aperture objective, and make sure
00:10:18.05 that it is focused at the back focal plane. So, this can be done by a
00:10:25.11 variety of methods. You can use what some microscopes have is a
00:10:30.10 side port, where you come into the turret that mounts the objectives from
00:10:37.02 the side. And just have a lens over here that you adjust its position. And its
00:10:44.14 position back and forth, so that you made sure that your off axis
00:10:50.21 focused at the back focal plane. There's other methods, too. You could come in
00:10:57.13 at a port in the microscope that might be marked aperture plane.
00:11:04.25 And make sure that the beam is focused there. And then if it's focused
00:11:10.15 at the aperture plane, well it turns out that aperture plane and back focal
00:11:13.20 plane are really complementary. So if you're focused at an aperture plane, you're also
00:11:18.02 focused at the back focal plane. There's other ways of doing this, if
00:11:23.11 there is no aperture plane port in the back of the microscope, just
00:11:28.09 set your own lens up here and make sure that you're focused at the back focal plane.
00:11:34.15 Many commercial setups are like this, and instead of using a laser beam,
00:11:40.10 just an open raw laser beam, you actually focus the tip of a fiber optic
00:11:46.08 bundle with a lens, to focus at the back focal plane. And that works
00:11:51.10 quite well, too. You don't need to use a laser, you could use a regular
00:11:57.04 mercury arc. The probably is a mercury arc sends light in all the
00:12:01.19 directions, in quite a width, instead of a nice thin pencil of light.
00:12:06.19 So, to do TIRF, you don't want light that's close to the optical axis.
00:12:12.09 You only want light that's far away. So you have to block the light that's close
00:12:18.25 to the optical axis with an opaque disk of some type. And only allow
00:12:24.28 peripheral light to get through. And then you could do TIRF even
00:12:30.25 with a mercury arc. And there are commercial setups that implement
00:12:35.26 this, and they're called white light TIRF. Not because the whiteness is important,
00:12:40.27 but just because the mercury arc is white colored. This method
00:12:49.27 of doing TIRF with an arc source, rather than a laser, actually works.
00:12:56.11 If you prepare a sample, as you can see up here, just a glass bottomed dish where the
00:13:04.07 glass is labeled with diI, just sticks to the glass and doesn't dissolve in
00:13:11.26 water. So it's a way of labeling a glass surface. And above it, you put a fluorescent
00:13:17.20 solution. So in TIR, you should only see the surface. The diI is orange, fluorescein
00:13:24.04 is green when you excite its fluorescence. So with regular epi, as you can see
00:13:30.01 over here, you see mostly green because the fluorescein in the water is
00:13:35.11 getting excited as well as the diI on the surface. With TIR, you
00:13:41.13 even with mercury arc TIRF, this opaque block, you see just the
00:13:48.27 fluorophore on the surface, the diI. And if you look at something like diI labeled
00:13:53.21 red blood cells, in epi, you just see round circles biconcave disks.
00:13:59.21 But in TIR, you see just the region where a biconcave disk is touching
00:14:06.27 the glass. So it looks sort of like a ring or a crescent. So, which is
00:14:13.01 better? A prism-based or objective-based TIRF? The prism based method
00:14:19.23 is better in that it has less scattering, because the scattering of the laser light if it had
00:14:27.16 to go through an objective. It's much cheaper, it's hundreds of dollars, versus tens of thousands
00:14:32.16 of dollars. It works well for low magnification and for water immersion
00:14:38.21 objectives. It's easiest to do with an open free laser. And you can
00:14:49.14 access a large range of incidence angles, you like to do that because
00:14:54.14 the bigger the incidence angle, the thinner the evanescent field.
00:14:57.12 And you might want to have control over that. The objective based
00:15:01.23 method is very good for high magnification, high aperture studies.
00:15:06.08 Which actually is quite common in cell biology. Very stable, and no
00:15:11.22 separate optical elements hanging on separate posts. It's easy to set up, it's
00:15:18.05 commercially available. And it works with either a free collimated laser or
00:15:22.27 an optical fiber input from a laser, or conventional arc sources.
00:15:28.00 So what are some of the limitations of TIRF? I mean every technique
00:15:35.25 in science has some drawbacks. Well, of course, TIRF is only good for looking at things at
00:15:45.05 surfaces. That's clear. But you also do get scattering of excitation light in the
00:15:52.23 objective when you use objective based. You can get scattering of light
00:16:00.13 because of index of refraction discontinuities on the sample. Some places on the sample
00:16:05.14 are denser than others, and then even the evanescent light gets converted
00:16:12.13 into scattered light at those places. And then the scattered light can
00:16:18.10 continue on and excite more fluorescence. And that makes an impurity.
00:16:22.00 You don't have a pure exponentially decaying field anymore. And also whenever you
00:16:29.11 are using laser excitation, whether it's TIRF or not, you can get interference fringes.
00:16:34.04 So we're going to talk about each of these little problems in order.
00:16:39.13 How can you measure how much scattering there is in the objective if you're
00:16:43.16 doing objective based TIRF? One way is to get a bead that has a pretty
00:16:49.08 low index of refraction and coat it with diI on its surface, so it's
00:16:56.01 basically looks like a hollow sphere of fluorescence. Then if you have
00:17:03.00 that bead sitting on the surface, then the lower part of it, down here will be
00:17:12.16 in the evanescent field and get excited. The deeper the evanescent field
00:17:16.07 is, the further up the side of the sphere you'll get excitation. And then
00:17:21.27 by looking with an objective from the bottom with a microscope, you
00:17:25.18 just look at how wide is the area of illumination from here to here.
00:17:33.01 And you can get a measure of how deep the evanescent field is.
00:17:36.12 And if that's done, you find out that as it should be, at low incidence angles
00:17:45.08 near the critical angle, the evanescent field is quite broad. And at
00:17:49.08 higher incidence angles, it's thinner. But the problem is it's not exactly
00:17:56.18 exponential. The blue part of these graphs here shows an exponential
00:18:03.05 contribution to the intensity, but the yellow part at the bottom just depicts
00:18:08.25 the scattering part. And the scattering part doesn't decay as rapidly as
00:18:14.19 the evanescent field that's exponentially decaying does.
00:18:19.11 So, you can see when you're looking at a fluorophore that's right
00:18:23.21 close to the surface, at what you might call z = 0, most of the intensity
00:18:29.10 is evanescent. But a little bit, maybe 10% is scattering. But as you
00:18:35.27 get further out, the fluorophore further out from the surface. Several
00:18:41.08 wavelengths away, most of it is scattering and not evanescent.
00:18:45.02 So you have to keep that in mind. Then the second limitation
00:18:49.27 has to do with non-homogeneities in the sample itself. Let's say
00:18:54.29 here's a typical cell. It has organelles in it, some organelles or some regions
00:19:01.05 might have a lower index of refraction than the average, some might
00:19:04.24 have a higher index than the average. So here comes the laser beam,
00:19:09.27 where there's a higher index of refraction, you won't get total internal
00:19:15.02 reflection. The light will just propagate out and bounce around all over the place.
00:19:18.10 And even in the places where you do get total internal reflection,
00:19:23.13 because of the discontinuity of the index of refraction, you still get scattering.
00:19:28.06 So you can see that the sample is doing its own scattering. This is still a
00:19:34.11 problem that needs to be resolved theoretically. Exactly how much scattering
00:19:39.23 you get. In practice, it's usually not a serious problem. But it is noticeable.
00:19:45.07 Particularly on a more dense cell. So, just for example, if you simulate
00:19:52.18 a cell with let's say, a glass bead, instead of a cell. And illuminate with TIRF.
00:19:59.09 If you use a TIRF at a lower incidence angle, so that the evanescent
00:20:04.18 field is deeper, if these beads are placed in a fluorescein solution, you can see
00:20:12.20 where the scattered light is. And every bead is causing sort of a
00:20:17.22 flare that disrupts the evanescent field. At a higher incidence angle,
00:20:25.09 the evanescent field is thinner. The disruption isn't as bad. So in general,
00:20:30.02 it's good to go to as high an incidence angle as you can get away with
00:20:34.05 to avoid the disruption due to scattering at non-homogeneities.
00:20:40.15 One thing that is worth pointing out here is that cells are usually
00:20:48.08 growing on collagen, even collagen that they make themselves, rather
00:20:53.28 than collagen that you may or may not have put down.
00:20:56.29 And that positions the cell further from the substrate, perhaps
00:21:02.19 in a region where scattering is more important than the evanescent field.
00:21:06.17 So you want to make sure that the collagen layer is not too thick,
00:21:09.18 otherwise you'll be looking at scattering light excitation, rather than
00:21:15.11 evanescent field excitation. The last problem limitation I wanted to talk
00:21:24.08 about in some detail is just illuminating by a laser gives you interference
00:21:31.08 fringes. Because the laser light is coherent and any defect in the optics,
00:21:36.24 any place, any dust, any warping of any surface, scratches, it will make
00:21:43.05 interference fringes on the sample. And the sample itself will produce
00:21:46.27 interference fringes, too. If you look at it, you see a lot of intensity
00:21:54.13 variations that really aren't on the sample, so that's an artifact.
00:21:59.12 If you could come in with a laser light from many directions,
00:22:07.01 from all kinds of different directions while all being TIR,
00:22:11.13 but all around various azimuthal angles, then those fringes, those interference
00:22:19.04 fringes might average out. And in fact, it does work. By putting the right kind of
00:22:23.23 optics, mirrors that make the beam go around in a circle or spinning
00:22:30.26 wedges, you can spin the azimuthal angle while keeping the angle of incidence still the same.
00:22:39.09 And the interference fringes seem to disappear. They're really
00:22:44.19 there, they're just changing so fast that the camera and your eye
00:22:48.14 doesn't see them. Now we're getting to other useful features
00:22:53.15 of TIRF. People often want to know what is the concentration
00:23:01.06 profile of a fluorophore near a surface? I mean, it might not just
00:23:05.08 be a single molecule or a very thin layer at a particular position.
00:23:08.25 It might have more at one location, less at another location, and you would
00:23:16.18 like to discover what is that profile. You can do that by varying the
00:23:21.23 incidence angle, which varies the evanescent field depth.
00:23:25.28 If you arrange it that you get a pretty deep depth, then the field
00:23:31.22 will be deep enough to excite fluorophores that are pretty far
00:23:37.10 from the surface, as well as fluorophores nearby. And you'll get a certain
00:23:41.04 total fluorescence. Whereas, if the field was much thinner with a
00:23:46.11 bigger incidence angle, you'll only excite fluorophores right near
00:23:50.16 the surface. And by varying the incidence angle continuously and the depth
00:23:57.29 continuously, you can deduce what the concentration profile was.
00:24:03.15 Also, when you look at TIRF, the most noticeable thing is that the field
00:24:12.23 seems very flat. You don't see anything out of focus because places
00:24:17.22 that would be out of focus aren't even being excited. So that means that
00:24:22.09 either the sample is in focus or out of focus, it's not a question of
00:24:25.14 where your focused. And it's easy to do the technique to do image
00:24:31.28 deconvolution, where it's basically a sharpening technique. It's very easy
00:24:37.02 to do with TIRF, because there's only one plane to worry about.
00:24:41.15 There's other features of TIRF that are interesting for cell biology,
00:24:47.19 particularly that normally in cell biology, with epi illumination the cells
00:24:56.01 don't like to be illuminated. They tend to die or at least they stop doing
00:25:01.08 what they are normally doing when they're under the light.
00:25:06.18 TIRF only illuminates the very bottom of the cells, so the cells
00:25:12.00 are not nearly as sensitive to the fact that they're being illuminated.
00:25:17.01 And they survive under a time-lapse movie for a much longer
00:25:22.16 time, in a week rather than just hours. Also, this TIRF technique is
00:25:33.17 particularly good for single molecule spectroscopy. Because often you have
00:25:37.16 molecules on the surface, there's other molecules in the bulk that are
00:25:41.03 fluorescence, and you don't want to see the ones in the bulk.
00:25:43.28 You need a lot of contrast, because the total amount of fluorescence is quite
00:25:48.02 low. So it's good for that. Also, if you come in with a TIRF beam this way,
00:25:57.09 and a TIRF beam that way, on the same sample so that they're
00:26:02.13 intercepting, you get very fine stripes. The stripes in fact are so fine that you can't
00:26:08.22 even see them. They're spaced smaller than the resolution of the microscope.
00:26:12.29 And those stripes are useful in a technique called structured illumination.
00:26:20.12 Which actually was invented at UCSF by Mats Gustafsson. And
00:26:28.25 it's a way of illuminating the sample and then shifting the stripes
00:26:34.29 a little bit, taking a picture, shifting them again a little bit, taking
00:26:37.11 a picture. That really enhances the resolution of the image.
00:26:41.22 So it's a super-resolution technique, and it's very effectively implemented
00:26:46.25 with TIRF. Okay, now the next question is, what do we see or what does the
00:26:58.08 objective see from a fluorophore that's near a surface?
00:27:02.21 This isn't a specifically a TIR or TIRF question. Any fluorophore
00:27:08.13 that's excited near a surface will have a pretty interesting pattern of emission.
00:27:13.03 It's not true that fluorophore just emits in all directions, even an
00:27:19.22 isolated fluorophore with no surface nearby does not emit
00:27:23.23 in all directions. But when you put a surface nearby, the anisotropy
00:27:31.04 of the directions of emission becomes really extreme. So let's say the fluorophore
00:27:37.09 is right there at this surface. It emits a lot of light in a hollow cone
00:27:48.11 along here. And this hollow cone goes all the way around. Depending
00:27:57.11 on the orientation of the dipole of the fluorophore on the surface,
00:28:03.08 you get a different pattern. So the pattern of emission depends not
00:28:07.22 only on whether there's a surface nearby, but also on the orientation of the
00:28:12.08 dipole. And this is one of the reasons for using a high aperture
00:28:19.04 objective, because you can see that much of the light that's emitted
00:28:23.08 from a dipole goes into a very high angle. And if the objective has
00:28:31.14 a low aperture, you'll miss that. So it's not like, well 1.45 aperture
00:28:38.14 is a little bit better than 1.3, 1.45 will catch this ring, this hollow cone,
00:28:45.20 whereas 1.3 will entirely miss it. Then, there's something interesting about this, too.
00:28:55.25 There's this line here, I show it as a dashed white line. That separates
00:29:06.23 two regions, the region below that or smaller angles into the glass, is
00:29:15.12 basically light that's just propagating from the fluorophore. The region
00:29:22.01 above that line is what's coming from what's called the fluorophore near
00:29:28.21 field. That's light that doesn't travel from the fluorophore, and normally
00:29:37.28 wouldn't be seen. But if the fluorophore is right near the surface, like
00:29:43.06 it would have to be in TIRF, but when any type of fluorophore is near a surface,
00:29:48.00 that near field can interact with the glass of the surface and get converted
00:29:54.11 into propagating light in the surface. And then that propagating light
00:30:00.10 can be captured. So it really is nice to have a high aperture objective
00:30:06.07 because it's a way of seeing some of the light from a fluorophore that you otherwise
00:30:11.26 wouldn't see. The light that comes from the fluorophore is near field.
00:30:15.22 Okay, now we're going to talk about polarization of the evanescent field.
00:30:25.26 And it actually turns out to be quite interesting and useful. With
00:30:30.14 regular epi illumination, you can have polarization that's either
00:30:35.21 this way or back and forth, you know, in the plane of this screen.
00:30:43.13 And when the epi illumination comes up through the middle of the
00:30:47.29 objective, it forms a polarization that's either in what's called the x direction or
00:30:54.24 y direction in the plane of the sample. And the only polarization that
00:31:01.20 you get is in the plane of the sample. Even if it's what you might call
00:31:06.12 unpolarized light, it's not really unpolarized because there is no polarization
00:31:11.01 and no electric field energy in the z direction at all, in epi illumination.
00:31:18.11 With TIRF, though. Coming up with the same polarized incident light,
00:31:25.05 but of course coming through the periphery of the objective, so you get the
00:31:28.07 big angle of incidence, the polarization turns out to have a large
00:31:34.04 z component. It's the only way you can get a z component on a
00:31:38.02 sample, is with TIRF. So you can either get a z component or this
00:31:46.04 y component here. This y component. And those two components are the
00:31:54.01 polarizations of the evanescent field. The z component you can
00:31:57.14 see is in the plane of incidence. The plane of incidence defined by
00:32:04.05 the light coming in and the light reflecting. So that's a plane, that's called the
00:32:11.17 plane of incidence. And z is in that plane. So, since it's polarization parallel
00:32:18.26 to that plane, it's called p-polarized, parallel polarization. The other
00:32:25.21 polarization, the y polarization that you can get, depending on the
00:32:29.26 polarization of the incident light is perpendicular to the plane of
00:32:33.15 incidence. And in German, that would be senkrecht, or the s-pol.
00:32:38.24 So we get p-pol and s-pol light, now what can you do with it? Well
00:32:45.23 one thing you can do with it is try to label a sample with a fluorophore
00:32:50.20 that's really oriented. One example is diI, when diI labels a membrane
00:32:57.12 it sits in the membrane as part of the lipid bilayer and orients in there.
00:33:03.25 Its transition dipole moment, which is the preferred direction of its
00:33:10.11 absorption and emission of an electric field is indicated here, and
00:33:18.09 so you get a preferred direction of absorption polarization and emission
00:33:23.05 polarization with diI. So what does that do for you? Let's say you label
00:33:28.04 a membrane with diI, in the parts of the membrane that are flat
00:33:34.28 on the substrate, the polarization or the orientation of the diI will
00:33:40.21 be parallel to the surface. In places where there might be an
00:33:45.27 indentation in the membrane, which in a real membrane might be an
00:33:50.14 exocytotic or endocytotic site, for example. Or some other kind of ripple.
00:33:57.24 Then those places, the diI orientation is perpendicular to the surface.
00:34:05.12 Now let's say you excited the whole business with p-polarized
00:34:10.28 evanescent field. That's the polarization that's along the z-axis.
00:34:15.17 As it is here. That kind of polarization will not excite dipoles or
00:34:24.16 transition dipole moments in the diI that are oriented like that, because you
00:34:29.15 don't get excitation when the electric field of the exciting light is perpendicular
00:34:34.03 to the molecule. But in places like this, where the dipole moment
00:34:41.01 is parallel with the excitation light, both in the z direction. You get a lot
00:34:45.06 of excitation. So with this technique, the indentations show up, they get
00:34:51.25 bright, whereas everywhere else is still dark. So this is a method of
00:34:57.01 seeing indentations in a cell membrane that otherwise with regular epi
00:35:05.00 illumination can't be seen, because you can't get a z-polarization,
00:35:08.26 a z-direction polarization from epi illumination. So let's say you take a
00:35:15.25 picture with p-polarized incident light and take a picture with s-polarized
00:35:21.09 incident light, the two possibilities you have. And you add them together
00:35:25.07 in this linear combination. Take the p image and then double the s image and add it
00:35:31.04 to the p image. The result is approximately how much fluorophore there is in
00:35:39.15 the different locations in the image. You do this at every pixel in the image.
00:35:44.06 And what you see is independent of orientation. This is a combination
00:35:50.17 that is independent of orientation, but it is proportional to the total amount
00:35:55.19 of fluorophore there and of course, the intensity of the illumination.
00:36:01.01 But the key thing is to look at the ratio of the image excited by p-polarized
00:36:08.07 light to the image excited by s-polarized light. You take the ratio,
00:36:13.23 that ratio depends only on orientation, it does not depend on how much
00:36:19.24 of the fluorophore was there or the illumination intensity. And it doesn't depend
00:36:26.03 on how far the fluorophores are from the surface. That P/S is entirely
00:36:30.09 orientation. So you can get basically an image or a reconstructed image
00:36:36.07 of orientation of membrane or other other organelles you might be
00:36:41.10 looking at, just by taking the P/S ratio of polarized TIRF illumination.
00:36:48.08 And it works. You have a red blood cell here, for example, labeled with dI.
00:36:54.16 This is the s-polarized, this is the p-polarized image. You take the ratio.
00:36:59.22 The places that you see that look bright are the edges where the membrane
00:37:05.29 is not parallel to the surface, but in fact, it's oriented with a component
00:37:12.13 perpendicular to the surface. This lower series is a macrophage
00:37:17.24 labeled with dI, and the s-polarized, the p-polarized. If you take the ratio,
00:37:25.12 the places that turn out bright are endocytic sites. That's the
00:37:32.07 job of a macrophage, it's to eat things. It has a lot of indentations on
00:37:37.04 its surface. They show up independently of the local concentration
00:37:42.27 of the fluorophore. It's just a map of orientations. And this can be done
00:37:50.10 also not just with two polarizations, but you might want to introduce
00:37:54.17 a third color so you can see some other process going on in the cell at the same
00:37:59.22 time. Like for example, there might be an indentation in the surface
00:38:05.06 right where an exocytotic event is taking place. You can use the other color
00:38:11.26 illumination to track that. And then watch how the membrane changes during
00:38:19.15 a secretion event. So for example, this is granules, secretory granules
00:38:28.05 excited with one color, in the blue. And where the granule is,
00:38:40.02 you see a granule and then in the next frame, all of a sudden it's gone. That
00:38:45.16 was an exocytotic event. And it's gone. If you look in this double labeled sample,
00:38:54.01 at the P/S ratio to look at orientations, there's not too much distinguishable going on
00:39:03.07 there. But shortly after the secretion event, you often see an indentation.
00:39:12.01 A higher P/S ratio, and in fact, that indentation lasts for quite a long time.
00:39:18.21 It could be for minutes even. And by doing this, you can see where and when
00:39:29.07 the secretion event happens, and then what happens to the orientation of the plasma membrane
00:39:35.03 during that event. So in conclusion about the polarized TIRF business,
00:39:42.01 you can detect possible local regions of altered fluorophore orientations, even
00:39:49.11 if these regions are submicroscopic, you can actually resolve them. You'll still
00:39:54.21 see a point of light in the P/S ratio. You can look at fast processes in cells.
00:40:01.29 And you might not be able to see these processes by any other
00:40:08.00 optical technique. And to summarize the whole view of TIRF,
00:40:14.22 you can use TIRF to look at small random motions of organelles
00:40:19.23 toward or away from the membrane. I didn't talk about that that much, but
00:40:23.06 if an organelle is fluorescently labeled and it moves further away,
00:40:26.22 it gets dimmer. If it gets closer, it gets brighter. And you can actually
00:40:30.14 quantitatively recover how rapidly it moves in the z-direction, even
00:40:35.21 down onto the scale of 10 nm. You can look at submembrane events, like
00:40:43.26 I talked about. You can look at membrane folding and indentations. You can
00:40:49.23 also measure by combining this with various other techniques.
00:40:53.13 If you have a fluorophore on the surface, how long does it stay on the surface
00:40:57.16 before it comes off? Or how long does it take for a new fluorophore to
00:41:01.17 show up? You can combine it with photobleaching fluorescence recovery
00:41:06.20 after photobleaching. Or fluorescence correlation spectroscopy.
00:41:10.13 To get those kinetic rates. It's also useful, as I mentioned, for single
00:41:15.12 molecule fluorescence because of the low background. And for
00:41:20.01 looking at the surface diffusion of molecules moving around on the
00:41:24.01 surface. So for this talk, I wanted to thank Ron Holz, my collaborator
00:41:33.06 at the University of Michigan, who is a specialist in cellular
00:41:39.08 secretion. And former graduate students and postdocs of ours, Susan Sund,
00:41:45.27 Miriam Allersma, Alexa Mattheyses, and Arun Anantharam.
00:41:52.14 And Joel Swanson is a professor at Michigan in Microbiology.
00:41:57.10 Thanks.

This Talk
Speaker: Dan Axelrod
Audience:
  • Researcher
Recorded: May 2012
More Talks in Microscopy Series
All Talks in Microscopy Series

Talk Overview

Total Internal Reflection Fluorescence (TIRF) Microscopy is a technique that only illuminates dye molecules near a surface. In this video, the pioneer of TIRF Microscopy describes what this technique is used for, explains the principles of the evanescent wave, gives many examples of different microscope configurations used in TIRF, and shows how polarized light TIRF can be used to image membrane orientation.

Questions

  1. What are the main uses of TIRF microscopy?
    1. Imaging single molecules and cell-substrate contact areas
    2. Obtaining optical slices throughout living cells
    3. Imaging cell nuclei
    4. Imaging developing embryos
  2. TIRF microscopy reduces background by
    1. Rejecting fluorescence from out-of-focus areas in the detection path
    2. Illuminating only the area right near the surface
    3. Deconvolution
    4. Only collecting near-field emission
  3. The evanescent field
    1. Forms all through the sample
    2. Can only be explained by quantum mechanical theories
    3. Decays exponentially and is ⅕ to 1/10th of the wavelength of the excitation light
    4. Can only be formed by transmitting light through an optical fiber
  4. TIRF illumination can only be achieved through the objective, True or False.
  5. TIRF illumination can only be achieved using laser illumination, True or False.
  6. In comparison to epi-illumination, TIRF
    1. decreases cell survivability and is unsuited for single molecule spectroscopy
    2. decreases cell survivability and is well-suited for single molecule spectroscopy
    3. increases cell survivability and is unsuited for single molecule spectroscopy
    4. increases cell survivability and is well-suited for single molecule spectroscopy
  7. What advantages does a high numerical objective have for imaging dye molecules near a surface?
    1. Allows the capture of more emitted light, including some of the near-field emission
    2. Has greater magnification
    3. None
    4. Captures higher wavelengths better
  8. DiI embedded in membranes will be excited by polarized TIRF with
    1. p-polarized light in flat areas and s-polarized light in indentations in the membrane
    2. p-polarized light in flat areas and p-polarized light in indentations in the membrane
    3. s-polarized light in flat areas and s-polarized light in indentations in the membrane
    4. s-polarized light in flat areas and p-polarized light in indentations in the membrane

Answers

View Answers

Speaker Bio

Dan Axelrod

Dan Axelrod

Dan Axelrod specializes in developing fluorescence microscopy techniques that are useful to cell biologists including total internal reflection fluorescence (TIRF) microscopy, a popular technique for viewing single molecules at surfaces. Axelrod is Professor Emeritus in the Department of Physics and Biophysics at the University of Michigan. Continue Reading

Playlist: Microscopy Series

Reader Interactions

Leave a Reply

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