With traditional light microscopy, the diffraction of light limits imaging resolution to about 250 nanometers. Super-resolution techniques can sometimes improve that by 10 times—or more, far more in one method mentioned below. Today, this technology spreads across a wide variety of approaches, but there are three main ones: single-molecule localization microscopy, including photoactivated localization microscopy (PALM) and stochastic optical reconstruction microscopy (STORM); structured illumination microscopy (SIM), although some experts point out that SIM’s resolution is only comparable to a good confocal microscope; and stimulated emission depletion microscopy (STED).
“Unfortunately, there’s no simple rule for deciding which super-resolution method to use,” says Mathew Stracy, Sir Henry Wellcome Postdoctoral Fellow at the University of Oxford in England. “Each has its own advantages and disadvantages.”
Scientists use various techniques to pick the right method for a specific project. “Find the simplest solution that solves the problem,” says Yoav Shechtman, assistant professor of biomedical engineering at Technion, Israel Institute of Technology in Haifa. “In the biological imaging context, key considerations include: the required spatial and temporal resolution, sensitivity to photodamage, labeling capability, sample thickness—is this a 2D or a 3D problem?—and background fluorescence level, or cellular autofluorescence.”
ABCs of super-res
The forms of super-resolution microscopy work in different ways. For example, with PALM and STORM, only a small fraction of the fluorescent markers on molecules are turned on, or photoactivated, at any given moment, allowing them to be localized independently with high precision. Running this process for all the fluorescent markers builds up a complete super-resolved image. “PALM/STORM systems are relatively easy to build, but more difficult to apply since the fluorophores have to be photoactivatable,” says Stefan Hell, director of the Max Planck Institute for Biophysical Chemistry and one of the winners of the Nobel Prize in Chemistry 2014 for super-resolution imaging. “Their limitation is that they need to detect individual fluorescent molecules amid cellular background.” He adds that these techniques are “less reliable to use than STED.”
With SIM, interference patterns of light create a grid on a sample during imaging. The grid shifts between images and algorithms based on Fourier transforms use the resulting information to locate features.
STED uses a laser pulse to turn fluorophores on and another one to turn them off. The focal point is scanned over the sample to generate an image. “STED has the advantage that it is a push-button technology,” Hell explains. “It can be used virtually like a standard confocal fluorescence microscope.” It can also image living cells with some fluorophores, such as green or yellow fluorescent proteins and the silicon-rhodamine derivative dyes.
In thinking over which kind of super-resolution to use, Neil Anthony, assistant scientist at Emory University, says, “It all has to be considered case by case.” He points out that PALM and STORM are “not so good for live cells,” but that STED can be used for live or fixed cells.
Rating the parameters
Although all super-resolution techniques surpass the resolution of traditional light microscopy, some make more gains than others. SIM roughly doubles the resolution, dropping down to about 100 nanometers. PALM and STORM can resolve features on the order of 15 nanometers. STED, says Hell, “can provide spatial resolution down to 30 nanometers in living cells and 15 nanometers in fixed cells.”
In assessing specific applications, the single-to-noise ratio must also be considered. In some cases, less resolution but a higher single-to-noise ratio produces better images than better resolution but a lower single-to-noise ratio.
The speed of acquiring images matters in some applications, especially ones with live cells. “All the super-resolution techniques are slower than conventional fluorescence imaging,” says Stracy. “PALM/STORM is the slowest, requiring often tens of thousands of frames to get a single image, SIM requires tens of frames per image, and STED is a scanning technique so the acquisition speed depends on the size of the field of view.”
Biocompare’s Microscope Search Tool
Find, compare and review microscopes
from different suppliers Search
Beyond imaging living or fixed cells, some scientists want to know how things move. For example, Stracy is “interested in looking at the dynamics of biological systems in live cells, not just static images, which people often associate with microscopy.” He can analyze dynamics with PALM combined with single-particle tracking in live cells. In this way, Stracy says that he can “directly track the labeled molecules as they perform their function in cells.” STED plus fluorescence correlation spectroscopy can also be used to watch labeled molecules move across a STED imaging volume. “SIM, on the other hand, isn’t well suited to studying these molecular-level dynamic processes, but since it is comparatively fast to acquire it is well suited to look at the dynamics of larger structures in cells, such as entire chromosomes, over many minutes,” he says.
Down to one
In 2017, Hell’s team reported on MINFLUX super-resolution microscopy.1 “This super-resolution method routinely attains—for the first time—spatial resolution on true molecular scale, 1 nanometer,” Hell says. “Moreover, it allows tracking of individual molecules in living cells with at least 100-fold larger speed than any other method before.”
MINFLUX—with 1 nanometer resolution—clearly resolves eight distinct molecules that are about 11 nanometers apart from each other. (Image courtesy of Francisco Balzarotti, Yvan Eilers, Klaus Gwosch, Arvid Gynnå, Volker Westphal, Fernando Stefani, Johan Elf and Stefan Hell.)
Other scientists also proclaimed the value of MINFLUX super-resolution microscopy. According to Shechtman, “There is a constant stream of new applications and method developments, but two exciting advances come to my mind.” One, he says, is MINFLUX. In describing its benefits, he says, “a clever approach is utilized to obtain molecular localization at very high precision and under an extremely limited photon budget.”
As a second exciting recent advance, Shechtman points out that W.E. Moerner—also one of the winners of the Nobel Prize in Chemistry 2014 for super-resolution imaging—and his Stanford University colleagues improved imaging resolution, which can be limited by the anisotropic emission of fluorescent single molecules. To solve this, says Shechtman, the scientists in Moerner’s lab “measure the orientations of the molecules along with their positions, by using different excitation polarizations. Alternatively, they developed sophisticated pupil plane engineering, which removes orientation-induced localization bias altogether.”2-4 Techniques such as these improve the ability to localize structures.
Looking at labels
In many super-resolution applications, the labeling really makes a difference, and some commercial options exist. For example, Germany-based Miltenyi Biotec joined forces with Abberior, a start-up company co-founded by Stefan Hell, to offer custom antibody-conjugation services for super-resolution microscopy dyes.
In many super-resolution applications, the labeling really makes a difference, and some commercial options exist.
Other companies also offer labels that work well with super-resolution microscopy. For example, Christoph Eckert, marketing at ChromoTek, says, “Our Nano-Boosters are very small—about 15 kilodaltons—and highly specific.” These proteins are bound to green and red fluorescent proteins (GFP and RFP, respectively) and vimentin-binding proteins. “The binding proteins are derived from single-domain alpaca antibody fragments, termed VHHs or nanobodies,” Eckert explains. “These VHHs domains are very small, have excellent binding properties and are produced at constant high quality without batch-to-batch variations.”
The size of these labels makes the difference in super-resolution microscopy. Coupled to fluorescent dyes, VHHs are useful tools in super-resolution. As Eckert points out, “A less than 2 nanometer epitope-label displacement minimizes linkage error.” He adds, “The fluorophores of conventional detection systems are more distant, usually 15–30 nanometers.” These labels work well with a range of super-resolution techniques, including SIM, PALM, STORM and STED.
Ai-Hui Tang, assistant professor at the University of Maryland School of Medicine, and colleagues used ChromoTek’s GFP-Booster with STORM to explore information transmission in the nervous system.5 At synapses—places where neurons communicate—the authors found nanoclusters of molecules in the pre- and post-synaptic neurons that make what they described as nanocolumns. The scientists concluded: “This architecture suggests a simple organizational principle of central nervous system synapses to maintain and modulate synaptic efficiency.”
The versions of super-resolution imaging and the increasing range of methods will continue to give scientists closer views of biology. In some cases, biologists can even watch cells in action—all while breaking the diffraction limit of visible light.
References
1 Balzarotti, F, et al. “Nanometer resolution imaging and tracking of fluorescent molecules with minimal photon fluxes,” Science 355:606–612, 2017. [PMID: 28008086]
2 Backer, AS, et al. “Enhanced DNA imaging using super-resolution microscopy and simultaneous single-molecule orientation measurements,” Optica 3:3–6, 2016. [PMID: 27722186]
3 Backlund, MP, et al. “Removing orientation-induced localization biases in single-molecule microscopy using a broadband metasurface mask,” Nature Photonics 10:459–462, 2016. [PMID: 27574529]
4 Lew, MD, Moerner, WE. “Azimuthal polarization filtering for accurate, precise, and robust single-molecule localization microscopy.” Nano Letters 14: 6407–6413, 2014. [PMID: 25272093]
5 Tang, AH, et al. “A trans-synaptic nanocolumn aligns neurotransmitter release to receptors,” Nature 536:210–214, 2016. [PMID: 27462810]
Image: Dreamstime Images