Scratching the Cell Surface

TIRF EXPLAINED

Courtesy of © 2004 Elsevier Science

These time-lapse TIRF images illustrate the dynamics of actin organization as a cell moves. Two frames were collected 1 second apart, pseudocolored, in order, green and red, and then superimposed. (A) Long filaments spanning half the length of the cell are evident in the right image. A new protrusion appears in the upper part of the middle image, and another one is expanding on the right (closed arrows). Retracting regions appear green (open arrowheads in the left and right images) and are accompanied by a dense texture of red, i.e., newly formed, filamentous elements at their base. (B) A leading edge of the cell shown in (A) at four stages of protrusion. The red border of the edge represents cell expansion within 1 sec. Scale bars represent 5 microns. (Reprinted with permission from T. Bretschneider, Curr Biol, 14:1–10, 2004.)

As its name suggests, TIRF microscopy is based on the phenomenon of total internal reflection (TIR), the principle underlying fiber optics. Light traveling between two media of different refractive indices bends or refracts at the interface. If the incident light comes in at larger than a certain angle (known as the critical angle) through the medium with a higher refractive index, all of the light will reflect back into the first medium without propagating into the second. The process generates an electromagnetic field (or evanescent wave) in the second medium, which persists to a depth of a few hundred nanometers and decays exponentially. (In TIRF microscopy the two media are the glass cover slip, with a refractive index of 1.5, and the aqueous cell sample, with a refractive index maximum of 1.38.)

"The evanescent wave is a very shallow field of excitation. It can excite fluorescent molecules within a 50 to 80 nanometer depth into the specimen," explains Stephen Ross, senior applications scientist at Melville, NY-based Nikon Instruments. Because the wave excites only fluorophores at or near the interface between the cell membrane and the microscope slide, there effectively is no background signal. "So you can get a much higher signal-to-background [ratio] than almost any other optical technique, and this allows you to do things like image single molecules against a background," Ross adds.

Daniel Axelrod at the University of Michigan, TIRF's inventor, says developing the technique was technologically "trivial" – his first efforts in the 1980s involved sending laser light through a prism at a high incident angle. "That wasn't technically difficult at all, and the equipment cost around $50 just for the right prism. And it wasn't even critical what the angles of the prism were, or the index of refraction, as long as you just introduced the light so that you could get TIR at the sample plane," explains Axelrod.

But this configuration was of limited use for biological applications. Mounting the prism on the microscope's condenser mount limited access to the sample, and the standard setup required an open laser. So scientists today generally use incident light (either from a laser or arc lamp) that passes directly through the objective to the sample.

STUDYING SECRETION

One of TIRF's best-known applications is the study of secretion events, endocytosis and exocytosis, at the plasma membrane. "With TIRF we can selectively illuminate the plasma membrane and the adjacent 50–100 nanometers of cytoplasm, without being confused by the rest of the cell. So it's very easy to see a single secretory vesicle undergo exocytosis, and it's almost as simple to see a single endocytic event," explains Almers, whose group was the first to use TIRF to study endocytosis and exocytosis.¹²

Guy Rutter and colleagues at the University of Bristol, UK, use TIRF microscopy to look at insulin release from individual secretory vesicles at or just beneath the surface of pancreatic beta cells. Insulin exocytosis is a relatively rare event, occurring 10–20 times per minute in an individual beta cell; Rutter explains that because it provides high resolution at the surface, TIRF microscopy enables him to pick up more of these events than confocal microscopy can. "Because the TIRF microscope … allows us to select for activity at the cell surface, we see many more of the secretory events that we want to see … than if we take a slice through the cell with the confocal, because then we just see the rim of the cell as a kind of pencil line," Rutter says.

Rutter and colleagues used TIRF to study the behavior of a protein residing on the vesicle membrane and another protein located within the vesicle. They found, to their surprise, that vesicle fusion to the cell membrane during exocytosis is transient, a so-called kiss-and-run event.³ In the past, Rutter explains, images from electron microscopy led researchers to suspect that vesicles collapse fully into the membrane and are gathered back into the cell by endocytosis in a different spot. With TIRF, says Rutter, "We see that the vesicle fuses very transiently with the membrane, releases its cargo, and then is gathered together at the same point at which it initially fused, as a sort of intact sphere, and then withdrawn back into the cell."

Likewise, Almers notes that TIRF has been used to measure FRET (fluorescence resonance energy transfer) between molecules involved in exocytosis. Molecules that are genetically modified with FRET tags are synthesized throughout the cell, but surface events can be isolated with TIRF. "If I want to see only the probes that are on the plasma membrane, then that's where TIRF is important," says Almers.

© 2002 Nature Publishing Group

A live 3T3 cell showing clathrin-DsRed fluorescence imaged by epifluorescence (a) and 1 sec. later by evanescent field (EF) illumination (b). (c) Actin fluorescence (green) was imaged under EF illumination simultaneously with clathrin-DsRed fluorescence (b) before the images were superimposed. The objective lens was focused on the interface between the glass coverslip and the adhering cell. (Reprinted with permission from C. Merrifield, Nat Cell Biol, 4:691–698, 2002.)

Almers cites Axelrod's studies of exocytosing vesicles as they move vertically to the plasma membrane as an example of exciting TIRF-based imaging. During endocytosis and exocytosis, ripples or dimples appear on the membrane surface that can be visualized with polarized TIRF microscopy when the membrane is labeled with a particular fluorescent dye (a carbocyanine) that orients itself in the membrane with its transition dipole moment parallel to the membrane. "Any place where the membrane is dimpled up, where the fluorophore then has a transition dipole moment with a component in the z-direction, those places would be excited and they'd look bright," Axelrod explains. "So the dimpled places would appear as bright spots on a dark background."

The changes in brightness resulting from vertical movement may be the result of fusion-mediating proteins called SNARES (N-ethylmaleimide-sensitive factor attachment protein receptors) that pull the vesicles toward the plasma membrane, Almers says, but he adds that further experiments are needed to verify this hypothesis.

TIRF AND THE SINGLE MOLECULE

While the study of exo- and endocytosis is one of its best-known applications, TIRF also has single-molecule applications. Stefan Diez of the Max Planck Institute for Molecular Cell Biology and Genetics in Dresden, Germany, uses TIRF microscopy to look at single biomolecular motor molecules interacting with cytoskeletal filaments. Diez's team immobilizes these filaments on a glass surface and visualizes single motor molecules that bind to the filament close to the slide-solution interface.

TIRF has the advantage of filtering out signals from molecules that diffuse in the solution, Diez says, making it easy to capture single-molecule images. Even the free molecules within the evanescent TIRF field do not pose a problem, he says. "We don't see the diffusing ones because they are so fast that they don't really give an appreciable signal on a single pixel of the camera chip," Diez explains.

Using TIRF, Diez and colleagues can determine the direction and speed at which single kinesin motors move along filaments, as well as whether the observed molecule is a monomer or dimer. Now Diez is using the technique to investigate the activity of kinesin-like motor molecules involved in the depolymerization of micro-tubules, to visualize the actin cytoskeleton in fast-moving cells, and to image the transport of DNA molecules on motor-driven microtubules.⁴

Another application of TIRF involves combining it with a technique called fluorescence speckle microscopy (FSM), which was developed by Clare Waterman-Storer of the Scripps Research Institute, La Jolla, Calif. In FSM, low quantities of green fluorescent protein (GFP)-conjugated proteins or fluorescently labeled proteins are injected into cells so that they coassemble with endogenous, unlabeled proteins to make macromolecular assemblies.

The low fraction of fluorescently labeled subunits produces an image with variations in fluorescent intensity (which look like speckles) along the macromolecular structure. "Those intensity variations serve as markers that can be tracked by computer vision tools to monitor the motion of that macromolecular assembly, and the appearance and disappearance of the speckles tell you about the assembly and disassembly of the structure," Waterman-Storer explains.

© 2003 American Chemical Society

Two rhodamine-labeled micro-tubules (red) transporting five individual YOYO-labeled DNA molecules (green, white arrows), visualized by TIRF microscopy. The dotted white line indicates the direction of transport. (Reprinted with permission from S. Diez, Nano Lett, 3:1251–4, 2003.)

Her group first used FSM to study the dynamics of microtubules and actin filaments. The researchers then combined FSM with TIRF to look at interactions between the actin cytoskeleton and focal adhesion complexes, sites on the bottom of cells that adhere to the coverslip.⁵Interactions between focal adhesion proteins and actin molecules are responsible for generating force in the cell. "By being able to visualize and quantify the dynamics of the proteins that make up the focal adhesion, and simultaneously the actin cytoskeleton, we are able to get a grasp on the nature of this molecular interface between the traction and the force production," says Waterman-Storer.

The technical merger yielded dividends: Waterman-Storer's group identified a protein that may serve as a "clutch" between the actin cytoskeleton and focal adhesion complexes. "It was so clear that the method could reveal stuff about how the cells crawl that nobody else had ever seen, that we had to do it," she adds.

Paul Forscher at Yale uses multimode FSM to look at protein dynamics involved in growth cone motility – specifically, micro-tubule and actin assembly and disassembly. With traditional epi-fluorescence illumination, FSM works well only for very thin portions of the cell, he explains, which in his case restricts imaging to the peripheral domains of growth cones. "It turned out that when we imaged axons and growth cones with TIRF microscopy, we could see a lot of interesting cytoskeletal protein dynamics within the evanescent wave zone," says Forscher. TIRF can provide better time resolution and spatial resolution than confocal microscopy, he adds. "Confocal microscopy is a scanning technique, so if you want a fast image-capture rate you typically have to reduce the number of pixels that you're scanning. In contrast, with TIRF we acquire a full high-resolution CCD [charge-coupled device] image every 50 to 200 milliseconds," says Forscher.

Forscher also uses TIRF microscopy to study growth cone turning events. His lab has developed an assay in which cell adhesion molecules are attached to a glass bead tethered to a micropipette. This construct acts as an artificial cue that instructs a growth cone to turn toward the bead. "What we've been able to do with TIRF microscopy is actually watch the behavior of the actin and microtubule cytoskeletal machinery while growth cones are making turns," says Forscher. "This has given us novel insights into mechanisms that we would never have anticipated without seeing them in TIRF."

TIRF does have its drawbacks; the technique can generate artifacts, for example. "Any time a structure goes out of the evanescent excitation field it disappears," Forscher notes. To address this issue his group is modifying its hardware so that the microscope can switch rapidly between TIRF and epifluorescence illumination, the latter of which provides thicker optical sections. "This will give us fuzzy but greater depth-of-field images to compare with what we're seeing in TIRF, and help us sort out TIRF artifacts from events that might be biologically interesting," explains Forscher.

TIRF's main strength is also its primary limitation. "By the TIRF effect, you can take a slice parallel to the surface of a thickness that you could only do with an electron microscope otherwise," says Almers. "The flip side is that if you want to look for something that's deep inside the cell, you're out of luck." Taking optical sections beyond the surface requires confocal microscopy, but it is technologically difficult to combine the two techniques. Confocal microscopy uses light from a wide range of angles, and TIRF works only with rays of light above a critical angle. "It's hard to see how you can combine a technique that demands a wide range of angles with a technique that demands a very narrow range of angles," says Axelrod. Nevertheless, scientists such as Almers are building microscopes that can rapidly switch between the two.

Aileen Constans (aconstans@the-scientist.com)