The (Never Ending?) Search for Higher-Resolution Microscopy

Lynn Savage, Features Editor, lynn.savage@photonics.com

Can optical microscopy further shred the Abbe diffraction limit? Is it important to even try?

Until about two decades ago, optical microscopists were stuck viewing their samples with a resolution no better than 200 nm. Now, well-funded labs have access to imaging down to 20 nm – and sometimes better still. Super-resolution optical techniques, in which fluorescent particles are manipulated into providing imaging resolutions far less than 200 nm, are primarily used in imaging and photolithography. Any technology that seeks to further improve upon the long-defeated diffraction limit codified by Ernst Abbe over a century ago has much to do.

An alphabet soup of techniques is available to researchers with plenty of spare change: stimulated emission depletion (STED), photoactivated localization microscopy (PALM), stochastic optical reconstruction microscopy (STORM), ground state depletion microscopy-individual molecule return (GSDIM) and many more.

Total internal reflection fluorescence (TIRF) microscopy provided the image on left, whereas using the superresolution technique photoactivated localization microscopy (PALM) produced the image on the right of a tdEos-paxillin fusion protein inside a cultured cell. Images courtesy of Carl Zeiss Microimaging LLC.

These techniques are increasingly being used to examine biological problems from a wide variety of angles. Cellular features such as nuclei and Golgi complexes, larger tissue samples and objects as small as individual proteins and peptides can be elucidated easily. Interactions between individual molecules and cell dynamics can be watched in real time.

Achieving resolution of 20 nm along the X- and Y-axes and 10 nm in the Z-axis is very good, said Prashant Prabhat, an applications specialist at optics manufacturer Semrock Inc. in Rochester, N.Y. He added that the overall resolution could be better, but not enough to be worthwhile in all likelihood.

Diffraction-limited imaging with today’s ultrahigh-numerical-aperture optics can reach resolutions down to ~200 nm, said Stephen T. Ross of Nikon Instruments Inc. in Melville, N.Y. However, it is localization techniques such as STORM that perform at an order of magnitude better, getting down to ~20 nm.

A TIRF image (left) and a direct stochastic optical reconstruction microscopy (dSTORM) image (right) each shows Alexa 647-tagged antibody staining for tubulin in a cultured cell.

“Beyond that, it is really a question of signal to noise,” Ross said. “If you have enough photons from any single molecule, and low enough background, you can reach single-nanometer resolution.

“This is why the use of small molecule dyes have a resolution advantage over probes that photobleach more quickly. I do not believe we will get beyond this resolution with classical optics. To achieve higher resolutions, I believe we need to go to other methods such as electron microscopy.”

Localization precision is primarily dependent upon the number of photons emitted by a single molecule, agreed Duncan McMillan of Carl Zeiss MicroImaging LLC in Thornwood, N.Y.

A wide-field microscopy image (left) and a superresolution structured illumination microscopy (SR-SIM) image (right) each shows actin (green) and tubulin (red) cytoskeleton in a primary chicken fibroblast.

“Resolution depends upon labeling density, which in turn depends upon the size of the molecule; that is, you can localize the position of a 30-nm molecule to a precision of <10 nm if it emits sufficient photons, but even if two adjacent molecules are found as close as possible to each other on a structure, the best resolution you will achieve is >60 nm,” he said.

Prabhat concurred.

“Localization to 1 nm is possible,” he said, “but that’s not the same as resolution. We are already there (with resolution), or at least where we need to be.”

To achieve high localization, he added, one must obtain the most relevant photons, which is the role of high-quality transmission and blocking filters, and other good optics.

What are they looking for?

A high-level goal is to get optical techniques to approximately the same resolution as electron microscopes. “But does that make any sense?” Prabhat wondered. “The strategies are very different.” Lighting a protein with another protein requires a linker, which introduces uncertainty, thus degrading resolution.

Researchers looking into superresolution techniques most often want to compare and contrast the multitude of available techniques. The big trade-off people must consider, Ross said, is temporal versus spatial resolution.

“Techniques like structured illumination microscopy, as used in our N-SIM system and others, can achieve temporal dynamics fast enough to study dynamic events in live cells with [a] large field of view in multiple colors, albeit only doubling the standard diffraction-limited resolution (~100-nm spatial resolution),” he said. “Localization techniques have been shown to be used for imaging in ‘live cells’ such as in the case of single-particle-tracking PALM or live-cell STORM, but there are significant limitations, only imaging small fields of view in specimens with relatively low molecular density.”

Another common question from clients that Ross noted is how many colors are viewable. Imaging down to single digits of nanometers is of limited usefulness, he said, if one cannot elucidate the interactions of proteins, which requires multiple colors.

More limitations

Driving future improvements into microscopes are other resolution-draining factors, such as inhomogeneities in optics and in the specimens themselves, as well as optics alignment issues. Developing new fluorophores is a major limiting factor for improved resolution; others include time and the ability to peer deeply into a sample. Also, existing systems are still quite expensive and complicated – out of reach of many labs and requiring highly trained operators.

“Today, many universities are merging multiple disciplines in integrative departments, such as Stanford’s BioX and Harvard’s Systems Biology programs, where the goal is to bring together scientists in areas of cell biology, neuroscience, physics, mathematics, computer science, etc.,” Ross said. “And these types of cross-disciplinary approaches [are] developing and utilizing these types of techniques.”

The plan at Nikon, as well as at other instrument makers, seems to be to ride the wave of customer demand. Deeper probes into biological systems will require expanding the boundaries of what microscopes can achieve.

Ross said that Nikon will continue to strive to push the limits of superresolution imaging as well as to increase temporal resolution, facilitating research into the dynamic events occurring in live cells.

“I believe that in 10 years all core imaging facilities will need superresolution imaging modalities to remain competitive,” he said.

Touring the Biological Pathways:
An Interview with Yuan Wang, University of California, Berkeley

What has led to your working in subdiffraction-limit techniques?

Optical microscopy and lithography are the two major technologies applied in characterizing and manufacturing electronic devices over the past several decades due to their high accuracy and yields. My research goal is to explore nanophotonic systems that can manipulate light and reveal the fundamental mechanisms of light/matter interactions at [the] nanometer scale. This began with the work on near-field microscopy and lithography, and then included the employment of plasmonic nanostructures to concentrate optical energy in deep-subwavelength scale for nanoimaging and nanomanufacturing.

What are you working on currently?

The minimum feature size or the optical resolution of an optical system is limited by the intrinsic diffraction property of light. The emerging nanoscale device commercialization calls for breakthroughs in nanotechnologies that are flexible for frequent design changes and high-throughput biosensing. My research on plasmonics and nanophotonics is to address this concern, via the strong interactions between the light and matter to realize nanoscale optical energy confinement. Ultimately, these techniques will have profound impacts on [the] electronic industry and health care industry.

What do you think the ultimate limit of optical imaging will be?

Subdiffraction-limit imaging has become a very active research area recently. The development of this field is incredible. Many earlier subdiffraction-limit techniques such as near-field scanning optical microscopy and solid immersion lens imaging have demonstrated much higher optical imaging resolution (<20 nm). But, practically, the signal-to-noise ratio of modern imaging systems will become the ultimate limit for improving the imaging resolution as well as the imaging speed.

What biological questions will be answered by nanoimaging?

An imaging tool with nanometer-scale spatial resolution can visualize the biological pathways at the molecular level, to sense the transient responses in real cellular conditions and to ultimately understand the mechanisms of human physiology in real time. Subdiffraction-limit imaging techniques may lead to exciting applications in fundamental molecular and cellular biology fields including cancer research, disease pharmacology, stem cell research and perhaps many other fields yet to be discovered.