YouTube Facebook LinkedIn Google+ Twitter Xingrss  

Myriad advances in light microscopy are increasing resolution, accelerating confocality, improving detection — and toying with the laws of nature

By Rabiya S. Tuma

Sept 16, 2004 | Knowing that the laws of physics limit the resolution of a light microscope, biologists have long relied on different types of microscopes for different tasks. No matter how good the lenses, objectives, mirrors, and other gadgetry, the best resolution with a light microscope until recently was about 200 nm in the X-Y axes, and about 700 nm to 800 nm in the Z-axis with a confocal microscope.

Of course, mathematical deconvolution and other computer wizardry could fudge this limit, but such techniques gained about a twofold increase in resolution. Anything smaller required an electron microscope.

That, however, is no longer the case.

A variety of new light microscopes has increased not only resolution in the X-Y and Z axes, but also the temporal resolution, allowing researchers to look at individual parts of the cell in ways never before possible.

DIFFRACTION ABSTRACTION: Stefan Hell in front of the equation indicating "breaking of the diffraction barrier."
For example, Leica Microsystems (Bannockburn, Ill.) recently announced the availability of its new TCS 4Pi microscope for beta testing at a limited number of institutions. The 4Pi microscope uses two opposing lenses with high numerical aperture to simultaneously illuminate a single focal spot. Because the incident light is in the form of coherent wave fronts, the two opposing beams result in constructive interference at the main incidence point, thereby increasing the resolution in the Z-axis to about 100 nm, says Stefan Hell, director of the Department of NanoBiophotonics at the Max Planck Institute for Biophysical Chemistry in Göttingen, Germany, who initially developed the 4Pi system.

Hell and colleagues ( are now working on a microscope that improves the resolution of fluorescence images to about 10 nm to 30 nm in the X-Y axes — about a 10x increase over the diffraction limit of 200 nm. "We now have evidence of 10nm to 15nm spatial resolution, and we have the possibility of getting to the 3nm to 5nm range," Hell says. That, he says, approximates single-molecule resolution.

In building this new microscope, called stimulated emission depletion (STED), Hell's group has developed a system in which it illuminates a region with intense light at a given wavelength. That incident light excites the fluorescent dye in the whole area, while the team immediately shines a donut-shaped pulse of light covering all but the central spot. The speed and wavelength of this second pulse quenches the excited state of the dye molecules before they have the opportunity to emit light. Thus, the team ends up with just a small area, well below the diffraction limit of light, within which the fluorescent dye is excited and emits light (Dyba et al. Nature Biotechnology 21, 1303; 2003).

EXTREME CLOSEUP: Carl Zeiss' LSM 510 META features a detector that can acquire spectral data for each pixel, voxel, and texel.
Although STED microscopes are not yet available commercially, companies will no doubt be producing microscopes with such resolution power in the future. Hell's team is not the only one pushing the diffraction barrier. David Agard, a Howard Hughes Medical Institute Investigator at the University of California at San Francisco, and his associates are working on a slightly different system aimed at the same target.

"There are pros and cons for each system," Hell says, "and one could argue which will be the most successful in the end, but we have faith in ours. Regardless, it is clear that the diffraction limit has been broken ... Ten years ago, people didn't think this was possible."

The Need for Speed
In addition to significant enhancements in spatial resolution, microscope developers are also improving confocal technology. Developers are working to improve the tradeoff between Z-axis resolution, or confocality, and light throughput, ease of use, and the scanning speed. Prime examples include the Olympus Spinning Disk Confocal, the new Nikon Eclipse C1-Plus, and an upcoming, yet-to-be-named Zeiss confocal, which will collect images in real time.

IT'S ALIVE: This image frame, taken with Zeiss' LSM 510, shows mitochondria expressing green fluorescent protein within a living cell.
University of Kentucky researcher Bill Kinsey has just set up a Nikon Eclipse C1-Plus microscope and is impressed with the power. His research group studies embryonic development in zebrafish. As expected with a confocal, Kinsey has no problems obtaining virtual sections through the 700nm embryo. But more impressive is that the new microscope has a dramatically improved signal with much lower laser output, so the cells aren't harmed during imaging.

"Previously, the beam intensity was so high that it was damaging embryos," Kinsey says. "Now we can do rapid scans, with low-intensity light and high-gain settings, which cuts the damage on embryos. We're seeing things we've never seen before."

Kinsey is using his Nikon C1-Plus microscope at a scanning speed of one image every five seconds, but the microscope is designed to collect images as rapidly as one 512x512-pixel image per second, or one every three seconds if scanning in a bi-directional manner.

The upcoming Zeiss microscope, due to be released in October, is designed to provide real-time confocal imaging using a whole new type of technology, which the company is not yet willing to discuss in detail.

What Caltech's Mary Dickinson, who along with Caltech's Scott Fraser helped Zeiss develop the new microscope, is willing to say is that the confocal will be able to collect 90 frames per second. This would be a dramatic improvement over the 15 frames per second that fast confocal cameras currently provide.

"Many people consider video rate (30 frames per second) to be real time, but that isn't real time for ion fluxes in a cell," Dickinson explains. With 90 frames per second, Dickinson believes the new Zeiss microscope will be able to image a whole new range of molecular interactions that are currently too fast to be caught on camera. Sebastian Tille, product manager at Zeiss, says, "The new fast confocal scanner uses a parallel acquisition approach that boosts speed as well as sensitivity."

The Caltech team is already combining the new confocal's high-speed image acquisition with Zeiss' LSM 510 META detection module. This module, which has been on the market for several years, uses a 32-channel multi-spectral photomultiplier (PMT), which can be used to acquire spectral information for each pixel, voxel, and texel — delivering up to five dimensional data sets, including X, Y, and Z axes, plus time and a lambda stack. Thus the module can separate emission spectra for even very similar fluorophores. The combination of the META detector and the new microscope will result in high spatial and temporal resolution, with the sensitivity to detect the intensity of each fluorophore in each pixel, Dickinson says.

Rockefeller University's Sanford Simon has used the META detector for his work with quantum dots (see "Tracking Cancer Cells with Quantum Dots," page 82), and has been impressed with the instrument. The detector is more sensitive than traditional bandpass filters, which commonly show bleed-through between fluorophores with overlapping emission spectra or those with wide emission peaks.

TIME ZONES: Nikon's new C1-Plus system includes improved software for live cell imaging and FRAP or FREP applications. This image reflects the system's capability to photo-bleach multiple zones in one field.

A Friendlier Package 
One of the new trends in microscope design is to build the system as an integrated unit. In the past, end-users often played a large part in deciding which computers and software they would use to drive their confocal microscopes. With both the upcoming Zeiss real-time scope and the Nikon C1-Plus, the microscope, computer, and software come together in a more integrated, user-friendly package.

A Simple Fix Generates New Power 
Joe Gall, a faculty member at the Carnegie Institute of Washington in Baltimore, has been studying chromosomes and nuclear structures in frog oocytes since the 1950s.

Read More 
The user will still have some choices to make — for example, the LSM 510 META detector can be added on to the Zeiss scope or not — but many of the decisions have been made for them.

With the Nikon C1-Plus, the company has spent a lot of time developing a more ergonomic setup that one person can run, without the need for a dedicated technician — a fact that Kinsey says makes using the equipment much less cumbersome.

Additionally, the computer that comes in the C1-Plus package can be loaded with between 700 MB and 2 TB of memory storage for data, and can record continuous files up to 2 TB. The company has also increased the functionality of the software so that users can crop, rotate, and zoom in on images without having to download the data to a different program.

For the upcoming Zeiss, Tille stresses that the company has experience handling large data files, based on its work with the META detection system, but that the new system will require even more data management. "A conservative example of data rate calculation is for 100 MB per second of data for an image that includes 100 frames per second, with 512x512 pixels per frame, 12 bit per pixel, and 2 channels."

Physicists have understood the principles behind total internal reflection fluorescence (TIRF) microscopes since the mid-1800s, but only recently have biologists figured out what to do with the technology, Simon says. "It is a killer technique for studying membrane dynamics and single events that couldn't be resolved before," he says.

With TIRF imaging, the light is directed at the specimen at a flat angle relative to the surface of the specimen. Consequently, the light penetrates only a short distance into the specimen, illuminating a small region of the cell. As a result, scientists can study the dynamics of vesicles and other structures at the cell surface with little or no background fluorescence.

Tracking Cancer Cells with Quantum Dots 
Quantum dots' unique fluorescence features mean they are highly photostable and can be tuned to a variety of emission wavelengths.

Read More 
The technique has revolutionized biologists' understanding of vesicle transport and fusion. Simon's group and others are using TIRF to study vesicle dynamics at the surface of cells. Simon likens the previous understanding to what someone might see if they tried to decipher the function of school buses by looking at them from a spy plane. "Most of what they would see are the large parking lots full of [off-duty] school buses. Every so often they'd see a stray school bus, but they'd think that was noise and, therefore, not relevant to what the school buses were doing."

"That's what we've been doing: looking for the strongest signals in the cell," Simon continues. "But what we're finding is that the minority event is the important one. And it is often transient, so even if all molecules go through it, we aren't likely to capture it." That all changes with TIRF. Researchers can follow individual events in membrane fusion, exocytosis, and endocytosis. And, as Simon says, they are finding that the rare events appear to be the functionally important ones.

Moving into the Future
Clinical researchers are looking for new ways to detect cancer early, when it is most likely to be treatable. While most of these efforts currently employ spectrophotometry or biochemical techniques, Intel and the Fred Hutchinson Cancer Research Center in Seattle are collaborating on the development of a new microscope that is designed to detect early disease states.

The Intel Raman Bioanalyzer System is based on Raman spectroscopy, a technique that Intel uses to determine the chemical composition of silicon computer chips. When a laser beam is shone on the material — silicon chips or biological samples — each molecule present gives off a unique signature. The idea with the new microscope is that researchers will be able to detect a shift in the signatures present between healthy cells and diseased ones.

The initial results from work with the microscope are expected early next year, but until then the company and the researchers are keeping quiet about the details, except to say that they are pleased with the technique so far.

The resulting picture is clear: Microscopes are revolutionizing the field of cell biology. But that can take time, insight, and creativity. As Simon points out, the physical principles on which the TIRF microscope are based were known for a long time before someone realized how to put them to use in biology. Sometimes the fixes are as simple as hooking up a new camera to an old microscope (see "A Simple Fix Generates New Power," page 80).

With more biologists becoming interested in imaging and more technology specialists heading into biology, people are bound to start uniting applications with questions and questions with applications at an unprecedented speed.


For reprints and/or copyright permission, please contact  Jay Mulhern, (781) 972-1359,