Introduction to Single Molecule Localisation Microscopy (SMLM)
The resolution of light microscopy is limited by diffraction to approximately half the wavelength of light (i.e., about 250 nm). Objects smaller than this cannot be resolved from one another using conventional light microscope techniques, but the example below shows how adjacent objects smaller than the diffraction limit can be resolved using Single Molecule Localisation Microscopy (SMLM).
Panel A shows a light microscope image of fluorescently labelled DNA origami structures immobilized on glass. Each structure appears as a bright dot whose size is defined by the resolution of the optics and the wavelength of light and is usually referred to as the Point Spread Function (PSF) of the microscope (scale bar 5 µm). Panel B shows a zoomed image of the square outlined in white in panel A. The underlying structure of the DNA origami cannot be resolved because the diffraction-limited PSFs are too large (scale bar 0.3 µm). Panel C shows a single molecule localization microscopy image of the same region. SMLM allows us to see that each spot is actually a DNA origami structure labelled with two fluorescent molecules, which are 94 nm apart (scale bar 0.3 µm).
The principle of SMLM
The problem illustrated above is that the fluorescent signals of adjacent objects overlap if they are closer than the limit set by diffraction, which is why two adjacent fluorophores can seem to form a single spot in panel B. SMLM solves this problem by imaging the fluorophores one at a time and not simultaneously. It does this by ensuring that in each image, light is only emitted from a few sparsely distributed fluorophores, which can then be precisely localised without their emission overlapping with that of a neighbour. The fluorophores only emit for a short time, so in the next image a different set of fluorophores is localised. This process is then repeated, usually for >10000 images and a final super-resolution image is constructed from the full set of localisations.
|
| Blinking timelapse, localisations and reconstructions |
Normally when a fluorescent sample is illuminated using techniques like widefield or confocal microsopy effectively all its fluorophore molecules will emit light simultaneously. For SMLM, a method is required to ensure only a few molecules emit in a single image; that those molecules do not emit in the next image; and that a different set of molecules emits in subsequent images, so that all the molecules can be localised with as little duplication as possible. The on/off switching of fluorophores between frames is often described as 'blinking', as seen in the GIF below.
SMLM Techniques
The principles above are effectively the same for all SMLM techniques. The main practical difference between experimental approaches is in how the blinking is achieved. The most common SMLM techniques are summarised below.
Photo-activated Localization Microscopy (PALM) or Fluorescence Photoactivation Localization Microscopy (FPALM)
PALM (or FPALM) was the first SMLM technique to be described in papers by Betzig et al and Hess et al. The original implementation used a photoactivatable form of GFP (PA-GFP) which is normally in a nonfluorescent state, but which can emit green fluorescence after photoconversion with 405 nm light. Sparse subsets of PA-GFP fluorophores were coverted into their emissive state using 405 nm light. The molecules were imaged until they photobleached and then another subset was photoconverted. This process was then repeated for thousands of images. Photoactivatable forms of fluorescent proteins can be challenging to image because they only become fluorescent after conversion, which makes it difficult to set up experiments. Photoconvertible FPs can also be used. These emit light at different wavelengths depending on whether they have been activated. For example, mEos2 normally emits green light but emits red light after photoconversion. In the case of photoconvertible FPs the blinking is achieved by converting a subset of the fluorophores in each frame. It is also possible to use organic dyes rather than FPs. In this case the dyes would be 'caged' and therefore would be in a non-emissive state that undergoes photoconversion.