Building a Better Optical Trap

PHYSICS 301: HOW IT WORKS

Optical trapping is just one of a class of techniques called "external field manipulators," which use different field forces to control biomolecules.³ Sometimes that force is light and sometimes it is a magnetic field, but on at least one memorable occasion, it was gravity. Howard Hughes Medical Institute investigator Carlos Bustamante of the University of California, Berkeley, who in 1992 used "magnetic tweezers" to perform the first direct manipulation of a single molecule when he measured the elasticity of DNA,⁴ recalls that 1991 experiment: "Our instrument was the planet Earth," he says, making a punchline out of a now-classic experiment in which his team attached molecules of DNA to a glass slide on one end, and to microscopic beads on the other. "As we put more and more beads on the DNA end, they sank in the water further, and they extended the DNA, so the first force-extension measurement of the elasticity of a single molecule was done by gravitational force."

Optical tweezers make similar use of the properties of light. Nineteenth-century physicist James Clerk Maxwell, of "Maxwell's equations" fame, predicted more than a century ago that a very small force, called "radiation pressure," should be associated with the interaction of light with a material. Some researchers did explore radiation pressure at the beginning of the 20th century, but it wasn't until the advent of lasers in the 1960s that light sources existed that could be focused tightly enough to manipulate a very small object.

"You can generate significant forces on a polystyrene bead with a 60-Watt lightbulb," observes Mark Williams, assistant professor of physics at Northeastern University in Boston. "It's because that force is exerted on a micron-sized object that it actually does something noticeable."

In other words, optical traps must contend with competing forces. The trapping is caused by the bending of light as it's refracted; the resulting change in the photons' momentum causes the attraction of the bead to the center of the laser point. If the light is reflected off the surface, there's also a change in momentum in other directions, which causes the bead to be propelled from the trap.

A MATTER OF PRACTICALITY

What really makes optical tweezers work as a practical lab tool is the ability to accurately measure position while a force is being exerted. If you can truly measure radiation pressure then you can accurately control the force to manipulate matter, exactly as if you were working with a mechanical tool, like metal tweezers. "The localization and manipulation is the crucial part," says Ashkin.

Optical layout of Steven Block's combined optical trapping and single-molecule fluorescence instrument, showing light pathways for the mercury-arc transillumination (light green), trapping laser (dark red), position detection laser (orange), fluorescence excitation lasers (dark blue, green) and fluorescence emission (red). Photodetectors include a QPD, a video-rate analog camera, a digital EMCCD camera, and a SAPD. Electronic shutters provide automatic control of the trapping beam (S1), fluorescence excitation beam (S2), bright-field illumination (S3) and light entering the EMCCD/SAPD (S4). Multiple optical filters isolate the diode laser emission (F1) and block the trap, detection, and excitation laser wavelengths before fluorescence detection (F2-F5). Flipper mirrors alternate between the video-rate camera and SMF subsystem (FM1) and choose the desired SMF detector (EMCCD camera or SAPD) (FM2). (From M. Lang et al., Nature Methods, 1:133–9, 2004.)

© 2004 Nature Publishing Group

Fundamentally, an optical tweezers experiment involves attaching a biomolecule to a trapped bead and pulling on it. Then it becomes possible to ask, how much force must I exert to dissociate a complex, or denature a nucleic acid? The distance the bead moves is directly proportional to the force applied and is referred to as "trap stiffness." It is a high-tech manifestation of Hooke's Law, F = -kx, where F is force, k is the trap stiffness, and x is displacement.

"A trap is a three-dimensional spring made out of light," explains Block. "If the spring is calibrated for whatever force you put on it, you can use it as a measuring tool. By measuring the motions of the bead with stunning accuracy, we're inferring the motions of the molecule attached to the bead."

Of course, that's merely a simplification. "Ray optics works fine, so long as the bead is bigger than the wavelength of the light," says Williams. "With beads on the order of the wavelength of light, you have to think about gradient force due to an electric field on a dielectric, which leads to more complicated calculations."

There is no such thing as a "generic" optical tweezers set-up, because the leaders of the field are continually refining the method. There are, for instance, single-, dual-, and multi-beam configurations, whose different geometries produce different forces on the trapped beads.⁵⁶ With dual beams, for example, you can work farther away from the lens, and so work deeper in solution.

Although magnetic and optical tweezers each have unique advantages and disadvantages, the gross way to distinguish between them, according to Bustamante, is by which technique works best with which level of force. Magnetic tweezers work best with very low forces, under a piconewton (pN). Optical tweezers are best with forces between 1 and 100 pN. Atomic force microscopy is also used in single molecule experiments; it is best for forces between 100 pN and 1 nanonewton.

Mara Prentiss of Harvard University says her lab concentrates more on magnetic tweezers, which make it much easier to replicate experiments. "Most optical tweezers only trap one particle at a time," she writes via E-mail. "David Grier at [New York University] has made tweezers that trap a few, and we have also trapped several, but it is difficult to do a hundred or a thousand and have a reasonable trap depth."

DON'T TRY THIS AT HOME

A few companies, such as Cell Robotics of Albuquerque, N.Mex., and P.A.L.M. Microlaser Technologies of Bernried, Germany, both offer prebuilt optical tweezers systems. But, most research labs hack together their own. A single-beam teaching-level laser trapping system can be created for a few thousand dollars. "It's just a microscope objective and a lens," Williams says of the $4,000 system he created from optical parts and a single mode laser diode. "You don't want an eyepiece if you're working with lasers," he adds.

Such a system can trap bacteria and yeast, even single molecules. But if you hope to wow them at the Biophysical Society, you'll need a completely customized combination of optics, piezoelectric stages and computer-controlled lasers, which together cost six figures, and more than likely time as a postdoc with either Block or Bustamante. Jeff Gelles, a Brandeis University professor and collaborator of Block's, says he sends his students to Block's lab at Stanford for experiments that require very precisely calibrated measurements.

In the laboratory equivalent of stripping a Lexus for parts, some labs build their systems by gutting and modifying high-end optical microscopes. Others just build their own systems from the ground up. Matthew Lang, assistant professor of biological engineering at the Massachusetts Institute of Technology, suggests that one advantage to starting with a stock microscope is that it feels familiar to outside collaborators, and so can make basic operations like loading samples easier.

This dual-beam optical tweezers instrument is used for stretching DNA. Two laser beams are focused to the same spot by microscope objectives, trapping a polystyrene sphere. To stretch DNA, another polystyrene sphere is held on the end of a glass micropipette by suction. As the DNA molecule is stretched by moving the micropipette, it exerts a force on the trapped sphere, causing it to move slightly. The distance it moves is proportional to the force exerted by stretching. The resulting measured force reveals information about the molecule's elasticity, stability, and its interaction with the environment.

Courtesy of Mark C. Williams

The research instruments have entire subsystems devoted to position detection (displacement sensing). And it's not impossible to have an instrument with five or even more light sources on it, according to Block. "One to do the trapping, one to do the position detection, one to do fluorescence excitation, and maybe one for another color of fluorescence excitation, plus a mercury arc lamp to provide illumination."

The challenge for Block's single-angstrom resolution experiment was extreme stability for the instrument. "If your system is drifting, you wind up measuring your system, but not the object of interest," says Block. "You measure that something is moving, but not the enzyme.... The real key point is you need a system that is ultrastable, devoid of drift."

He explains the basic physics challenge: "If you propagate a laser beam through air, the laser beam will twinkle a little bit. It will move up and down by tiny, tiny amounts, and it does this for the same reason a star twinkles; because there are density fluctuations in the air, and the density fluctuations give rise to change in refractive index, and these act a bit like a lens to bend the beam of light." Normally, in light microscopy, density fluctuations are too small to be an issue, but that's not the case at the angstrom level. Block says his lab has "gone to extremes over the past several years to build a system that is not subject to drift."

One seemingly obvious solution, to put all the instruments in a vacuum, creates its own complicated, cumbersome challenges, Block says, so his team did the next best thing: They used helium, which has a refractive index much closer to that of a vacuum than air does, so small changes in pressure would not bend the light.

"Using scattered light, we can measure the center of one of the beads to much smaller than the resolution limit of the light microscope, by either interferometry or back focal plane detection," says Block.

Tom Perkins, an associate fellow at JILA, a Boulder, Colo.-based research institute run jointly between the National Institute of Standards and Technology and the University of Colorado, took a different approach to achieving single-angstrom resolution.⁷ Unlike Block's system, which was uncoupled from the surface of the cover slip to reduce the inherent noise, Perkins used a second laser to measure surface noise and then subtracted it out in real-time. "If you can set up an assay where you can pick the whole thing up off the surface, then Steve's technique is a very good choice," says Perkins, "but for assays that are inherently surface-coupled, ours would work better."

FLUORESCENCE AND THE FUTURE

Though much single-molecule biochemistry has been accomplished with optical traps, there are other ways to approach these issues. One involves precise monitoring of fluorescent probes, for instance, on the surface of a molecular motor.

Some researchers are now building composite systems that combine such measurements with optical trapping. Japan's Toshio Yanagida did pioneering work in 1998, using a microfabricated pedestal to separate the optical traps from the location of the fluorescence to avoid photobleaching. More recently, Block and Lang, his former postdoc, demonstrated that it is possible to combine trapping and fluorescence at the same location and on the same molecule.⁸

Commenting on various single molecule manipulation techniques, from optical traps to magnetic tweezers to atomic force microscopy, Julio Fernandez of Columbia University, says, "The instrumentation has a long way to go, but at the end of that road, it's going to revolutionize our understanding of proteins. It's going to change how the books are written."