Fluorescence light microscopy techniques offer several advantages when imaging biological samples, including high image contrast, good labeling specificity, multicolor, and three dimensional (3D) imaging, as well as the capability to image living specimens. The primary disadvantage, however, has been the diffraction-limited 200 nm and 500 nm lateral and axial resolutions, respectively, obtainable due to the wave nature of light.
Image resolution is defined as the smallest separation distance between two objects whereby they can be distinguished as being separate. Electron microscopy (EM) has often been used to obtain resolutions beyond that obtained by light microscopy, down to the nanoscale; however, EM of living samples is not typically feasible, and 3D imaging is cumbersome. The advent of super-resolution light microscopy techniques in the last decade has the capability to bridge the gap between traditional light microscopy and EM.
Super-resolution imaging can be broadly divided into three categories: point spread function engineering (used in stimulated emission depletion [STED] microscopy), structured illumination (SIM), and single molecule localization (SML). The principle behind SML microscopy is illustrated in Figure 1. The resolution limit in conventional light microscopy arises due to diffraction, thereby causing the point spread functions of nearby individual molecules to overlap, resulting in a “blurred” image.
SML techniques utilize the principle of “localization” to obtain higher resolution. It has been shown that one can infer the nanoscale position of a single, isolated molecule by computing the centroid of its point spread function. Importantly, in order to obtain the higher resolution offered by localization, it is essential to have only a small subset of molecules visible at a time within the field of view.
To accomplish this, SML microscopy relies on the ability to convert nearly all molecules in the sample to a dark state with only a few molecules visible (active) at a time. The positions of the visible molecules are subsequently determined by localization of the center of the molecules. The activation and localization cycles are then repeated several thousands of times to obtain a higher resolution image of nearly every molecule in the sample.
A number of SML super-resolution microscopy modalities have been demonstrated that utilize various strategies to image single, isolated molecules in a dense sample. Among others, these include fluorescence photoactivation localization microscopy (FPALM), photo-activated localization microscopy (PALM), stochastic optical reconstruction microscopy (STORM), direct STORM (dSTORM), and PALM with independent running acquisition (PALMIRA).
Each of these techniques is based on the principles outlined previously, but varies based on the fluorophores used and the protocol for obtaining sparse subsets of isolated molecules. While SML may not be amenable to every available fluorophore, a variety of genetically encoded and exogenous dyes can be used; these include photoswitchable fluorescent proteins such as Dendra2 and mEos, and organic molecules such as Alexa Fluor and ATTO dyes to list a few.