In 2014, the Nobel Prize in Chemistry was awarded to Eric Betzig and William Moerner who, working separately, laid the foundation for SMLM. In essence, this method relies on the possibility to turn the fluorescence of individual molecules on and off. Scientists image the same area multiple times, allowing only a few interspersed molecules to glow each time. By superimposing these images, a dense super-image can be resolved at the nanolevel. With the development of this technique, Betzig and Moerner were able to overcome Abbe’s diffraction limit, allowing for the production of high resolution images that, before SMLM, had not been possible.
Towards the end of the nineteenth century, Ernst Abbe and Lord Rayleigh formulated what is commonly known as the “diffraction limit” for microscopy. Roughly speaking, this limit states that it is impossible to resolve two elements of a structure that are closer to each other than about half the wavelength (?) in the lateral (x, y) plane and even further apart in the longitudinal (z) plane. Another consequence of the same diffraction limit is that it is not possible to focus a laser beam to a spot of smaller dimension than about ?/2.
In the case of light (optical) microscopy, an important tool for the imaging of biological structures, this means that two objects within a distance between 400/2 = 200 nm (far blue) and 700/2 = 350 nm (far red) cannot be resolved. Although this is no real limitation for electron microscopy, in which the wavelength is orders of magnitude smaller, this method is very difficult to use on living cells. For instance, the length-scale of the E. coli cell is about 1,000 nm (1 ?m) which is larger than, but of similar magnitude, as the diffraction limit.
This explains why, prior to the development of SMLM, it was difficult to image details of the internal structures of living bacteria. Perhaps this may be the reason why bacteria are considered to be “primitive” organisms with little internal structure. With single-molecule localization, more precise structures of bacteria and other small-scale entities, e.g. individual viruses, can be resolved.
In SMLM, the photochemical properties of fluorescent proteins are exploited to induce a weakly emissive or non-emissive “dark” state. From the dark state, very small populations of fluorophores are returned to an emissive state by shining a weak light pulse that activates only a fraction of the fluorophores present.
These fluorophores are excited and detected by glowing until they are bleached, at which point the procedure is repeated on a new subgroup of fluorophores. In order to be identified, however, the emission profile must exhibit minimal overlap in each image. The centroid position of each identified molecule is statistically fitted, often to a Gaussian function, and with a level of precision scaling with the number of detected photons.
By imaging and fitting single emitters to a sub-diffraction limited area over thousands of single images, enough data is generated to create a composite reconstruction of all identified emitters. Single-molecule localization is a broad category consisting of specific techniques, such as STORM, PALM, and GSDIM, that operate using the conceptually similar procedure outlined above. The main difference between these types is the exact fluorophore chemistry used to turn the fluorescence of individual molecules on and off.
The real breakthrough in single-molecule localization occurred in 2006, when Betzig and colleagues coupled fluorescent proteins to the membrane enveloping the lysosome, the cell’s recycling station. By activating only a fraction of the proteins at a time and superimposing the individual images, Betzig ended up with a super-resolution image of the lysosome membrane. Its resolution was far better than Abbe’s diffraction limit of 0.2 ?m, a barrier that previous microscopy techniques could not bypass.
Since the ground-breaking discovery, SMLM has allowed organelles and single molecules to be resolved with an order of magnitude better resolution (with a localization accuracy of about 10 nm), in multiple color channels, and in 2D as well as 3D. Single-molecule microscopy allows quantification of the number of proteins within biological assemblies and characterization of protein spatial distribution, permitting the determination of protein stoichiometry and distribution in signaling complexes.
For instance, for the ?2 adrenergic receptors, SMLM was used to show that the receptors are partially organized in mini-clusters only in cardiomyocytes but not in any other cell lines, and that these oligomers are not lipid raft related but rather depend on actin cytoskeleton integrity. Most importantly, the results of this study were different from those obtained from a similar report which used near-field scanning optical microscopy (NSOM), demonstrating the better precision of SMLM over other techniques.
An additional important aspect of SMLM is that it can be used with other imaging techniques to elucidate receptor complex structures. In one study by Nan et al. (2013), the powerful sensitivity of FRET imaging to detect receptor proximity was combined with the capability of SMLM to obtain direct visualization of receptor oligomers in studying RAF, a strategic protein involved in RAS signaling.
By means of cluster analysis, Nan and colleagues were able to show how RAF exists between an inactive monomeric state in the cytosol and a multimeric condition at the cell membrane when activated. The results from single-molecule localization confirmed the importance of dimer and oligomer formation in RAF signaling, even though the precise biological role of these different multimeric states is yet to be determined.
The better definition of biological structures in the nanometer range as a result of SMLM has had most relevance in the field of neuroscience, where the morphology of neurons composed of dendritic spines and synapses is not suitable for confocal microscopy.
For example, Dani et al. (2010) used single-molecule microscopy to image presynaptic and postsynaptic scaffolding proteins in the glomeruli of the mouse olfactory bulb to show distinct punctate patterns that were not resolved by conventional fluorescence imaging. Lastly, the high resolution of SMLM has enabled a deeper understanding of chromosome organization and genome mapping. Wang et al. (2011) determined the distribution of nucleoid-associated proteins in live E. coli cells, while Baday et al. (2012) were able to label 91 out of a total of 107 reference sites on a 180 kb human BAC gene with a 100 bp resolution. DNA mapping with such resolution offers the potential to uncover genetic variance and to facilitate medical diagnosis in genetic diseases.
Nonetheless, there are a few challenges that come with single-molecule microscopy, namely errors in detection efficiency and localization uncertainty. Since using fluorescent proteins as labels involves the complications associated with protein expression, errors in this step (e.g. misfolding, incomplete maturation, etc.) can lead to the production of label molecules that are not fluorescent. This can directly affect counting studies, as the number of counted molecules can be underestimated.
However, it is possible to use the obtained count (after correcting for blinking artifacts) for the counting. In one study that involved identification of protein complex stoichiometry by counting photobleaching steps, Renz et al. (2012) accounted for errors in detection efficiency using a binomial model, which was found to provide accurate results. Incorporating detection efficiency into a model for the ratio between monomers and dimers can also rectify efficiency errors.
In terms of localization uncertainty, each photon from the emitter molecule provides a sample of the point spread function (PSF) from the molecule. Based on these samples, single molecule localization algorithms provide an estimate for the position of the fluorescent molecule. This estimate is prone to uncertainties, especially due to limited sampling (i.e. the limited number of photons obtained from the molecule). By ensuring that the imaged molecules within a frame are spatially separated enough so that the localization algorithms can correctly identify them, however, it is possible to minimize the effect of localization uncertainty on counting measures.
Despite its potential shortcomings, single-molecule localization enables high resolution imaging on the scale of nanometers, which defies Abbe’s diffraction limit of 0.2 ?m. SMLM has been used to elucidate specific cell structures, as in Betzig’s visualization of the lysosome membrane, and receptor complexes, as in the case of RAF.
The technique has also been used to refute results of similar studies that used different imaging protocols, as shown when determining the specific location of ?2 adrenergic receptors. Overall, SMLM has ushered in a new era of high resolution imaging that not only allows for accurate insight into individual cell and protein structure, but also enables identification of abnormalities in cellular processes that ultimately manifest as genetic diseases.