Breaking the diffraction limit

The period at the end of this sentence is 1 million nanometers wide. With super-resolution microscopy, scientists can see synaptic vesicles as small as 30 nanometers wide. Imagine taking a picture of the continental United States from the stratosphere and being able to distinguish a single strand of hair.

Fluorescent tags make this possible. Molecules and structures of interest are given a fluorescent tag—either a dye or a genetically engineered tag like green fluorescent protein (GFP). A laser beam is directed at the target sample, which makes the tag emit light, and the resulting fluorescence is recorded to create an image.

The first fluorescence microscopes scanned an entire sample at once, which produced a fair amount of out-of-focus fluorescence. Today’s standard fluorescence microscope—the confocal microscope—scans a sample one point at a time, pixel by pixel, and assembles the pixels to create an image. The size of each fluorescent point in these standard microscopes is determined by how much the laser light diffracts and is limited by the diffraction limit to between 200 and 250 nanometers.

Stefan Hell’s innovation, STED, improves resolution by reducing the size of each fluorescent spot. STED targets the light returning from the sample and a second laser blocks out the fluorescence in a donut shape around the center of each fluorescent spot. Each fluorescent point is reduced to the size of the donut hole. These smaller points of light yield a higher-resolution image. STED microscopy can achieve a resolution of 25 to 80 nanometers, small enough to distinguish cellular vesicles and the folds within organelles.

PALM/fPALM/STORM capture just a few scattered molecules at a time so that they are unlikely to overlap and blur together. Using labels that turn on and off, scientists arrange to have only a few molecules fluoresce at one time; then they take a picture. A computer finds the center of each spot, representing a single fluorescent molecule, on the individual photo. This process is repeated thousands of times, and the photos are then combined. The approach is sometimes called pointillist microscopy, after Impressionist Georges Seurat’s painting technique. Pointillist techniques achieve extremely high resolution, about 25 nanometers. However, the technique can also be slow—it requires many photos to generate one image, and it is dependent on high-powered computers to process the data.

TIRF microscopy, developed in the early 1980s, excites fluorescence in a thin layer near the cell surface, which reduces background fluorescence and improves resolution to between 40 and 100 nanometers. TIRF microscopy is faster than pointillist techniques but has lower resolution and can record only the cell surface.

These are only a few of the high-resolution microscopy techniques available today, and Yale is unusual in that it has all these microscopes—STED, PALM/fPALM/STORM, the electron microscope, and others—in one place, said Derek Toomre. Each has its strengths and weaknesses. “If we knew that there was one type that could do everything, we wouldn’t be investing in all of them. … There’s no clear winner. We’ll see; maybe there will be.”