Taking FRET to the single-molecule level wasn't an easy feat. In fact, it was so challenging that Ha, who developed the technique as a graduate student in Weiss's lab, recalls thinking it would be effectively impossible for others to ever repeat.

But the concluding statement in the abstract of his seminal paper, which his graduate research advisors insisted upon, reads: “Monitoring conformational changes, such as rotations and distance changes on a nanometer scale, within single biological macromolecules, may be possible with single-pair FRET.” (1)

“I thought that was [BS],” Ha recalls, “but it turned out they were right. It turned out to be a general mechanism to study biomolecular interactions.”

smFRET has become more than merely reproducible. Thanks to technical advances such as more-sensitive detectors, improved fluorophores, and better passivation chemistry, it's something of an everyman's single-molecule technique these days. It is a method almost anyone with a good total internal reflection fluorescence (TIRF) microscope can use. According to the ISI Web of Science, Ha and Weiss' original paper has been cited nearly 400 times.

Starlight, Starbright

In Starlight's case, the FRET donor is a custom quantum dot nanocrystal attached to the back of a DNA polymerase like a snail's shell; the acceptors are the individual deoxyribonucleotides, each bearing a different dye. To make this work—that is, to get FRET to extend as far as the dNTP gamma phosphate position—the system actually uses a variant FRET scheme called a “three-color cascade,” in which the transferred energy hops from the quantum dot to an intermediately placed dye, and then to the final acceptor. The reactions themselves are imaged via simple TIRF microscopy, which limits fluorescence to a narrow optical slice immediately above the slide surface.

The approach is, according to Beechem, “a great way to molecularly confine the [reaction], but not physically confine it.” In other words, between the TIRF illumination and FRET detection, the system can ignore all the fluorophores in the reaction except those that are actually interacting with the polymerase. By comparison, Pacific Bioscience's third-generation technology relies on “zero-mode waveguides” and a tethered polymerase to restrict fluorescence from labeled nucleotides to the volume immediately around the enzyme.

Beechem says Starlight can read up to 300 bases per minute for several minutes, for each of the approximately 100,000 distinct templates immobilized on a surface. But it's not commercially available yet; Starlight technology is still in the research phase of development, Beechem says.

Multidimensional FRET

For all its power, standard smFRET is limited to measuring one interaction at a time—a single vector in a complex three-dimensional space. To really probe molecular dynamics, researchers need to pinpoint those dye molecules relative to atleast one other point. “In my view,” says Ha, “the future lies in multidimensional single-molecule measurements.”

In 2004, Ha and team member Sungchul Hohng took the first step towards such multidimensional smFRET by using three dyes (Cy3, Cy5, and Cy5.5) to study the conformational acrobatics of a DNA complex called a Holliday junction. More recently, Hohng (now assistant professor of physics at Seoul National University, South Korea) and Ha extended smFRET again. By carefully positioning each of four dyes (Cy2 or Alexa488, Cy3, Cy5, and Cy7), they were able to measure six intramolecular distances concurrently, again in a Holliday junction.

Detection in a flash. (Click to enlarge)

Ha says the trick is to cycle between three different lasers, an approach called alternating laser excitation, or ALEX. ALEX is a technical variant that is used to directly probe the presence of active FRET donors and acceptors, and to untangle the FRET signals arising from the different possible intermolecular events. Here's how it works in the case of four-color smFRET: first, Cy5 is excited and its distance to Cy7 is measured using a 633-nm red laser. Then, Cy3 is excited and its distance to Cy5 and Cy7 measured with a 532-nm green laser. Finally, Cy2 is excited and its position relative to everything else is measured with a 473-nm blue laser. “You can, in these three steps, deconvolve everything,” says Ha.

Unfortunately, that approach really isn't scalable, says Achillefs Kapanidis, a university lecturer in biological physics at the University of Oxford, since the addition of each new color makes data collection and analysis more complex. “Five-color ALEX, for instance, would be horribly complicated,” he says.