Product Survey: Super-resolution microscopes
Looking over Yonder Wall
by Harald Zähringer, Labtimes 03/2014
Though Ernst Abbe’s diffraction barrier is not breakable, researchers have found ways to look behind it and visualise objects closer than the diffraction limit of light.
In a scholarly piece covering 56 pages, the German physicist, Ernst Abbe, defined the limits for lateral (x,y-plane) and axial (z-axis) resolution of a light microscope, already in 1873. Abbe described this so called diffraction barrier only with words but it may be boiled down into the following two simple formulas: dxy = λ/2n sinα and dz = 2λ/(n sinα)2 (λ stands for the wavelength of light, n represents the index of refraction of the media between coverslip and objective lens, and α is the half-angle of the maximum cone of light that can enter the lens). Abbe called n sinα the numerical aperture (NA) of the objective lens. Until today, it is one of the most important parameters of a light microscope. Typical NA values of oil immersion objectives are in the range of 1.4 to 1.5, with maximal aperture angles of about 140°.
Substituting Abbe’s resolution formula with 500 nm light and 1.5 NA yields a resolution limit (diffraction barrier) of classical light microscopes of about 200 nm in lateral direction and 500 nm along the optical axis. A brief look at the formula immediately shows, which parameters have to be tweaked to improve the resolution: you may shorten the wave length, increase the refraction index or expand the aperture angle of the objective lens.
Shorter wave lengths are realised in electron microscopes accomplishing resolution limits down to 0.1 nm. But high energy electron waves and the necessary vacuum are not suitable for the investigation of living specimens. Filling the space between cover slip and objective lens with natural or synthetic oil to increase the index of refraction is an old trick, already applied by Abbe and other microscopists in the late 19th century. However, the difference between the refraction indices of air (1.0) and oil (approx. 1.5) is way too small to essentially lower the resolution limit.
The idea to increase the aperture angle of the objective lens led to the construction of Confocal Laser Scanning 4Pi-microscopes in the late 1970s and early 1990s by super-resolution microscopy pioneers, Cristoph Cremer and Stefan Hell, then at the universities of Freiburg and Heidelberg, respectively. The 4Pi-microscopes are equipped with two opposing objectives to double the aperture angle, collecting the light from a laser-illuminated object (that’s the very basic concept; the realisation, however, is much more complicated). The 4Pi-microscopes may reach a three-dimensional optical resolution of approximately 100 nm and were amongst the first super-resolution light microscopes that allowed a look beyond the diffraction barrier.
Since Abbe’s diffraction barrier is fundamental, approaching substantially smaller resolution limits in the low two-digit nanometre range, just by further optimising the optical set-up of conventional microscopes according to Abbe’s formulas, is basically impossible.
Hence, microscopy specialists in the labs of big microscope manufacturers obviously quit the idea of trying to overcome the diffraction barrier – it took out-of-the-box thinking, scientific outsiders to come up with new approaches to circumvent the barrier. They put an additional parameter into play, which had hitherto been ignored by the inside-the-box thinking experts of the microscope manufacturers – the factor ‘time’.
In conventional fluorescence microscopy, all parts of the object are illuminated at once to excite fluorescence molecules. Hence, two simultaneously blinking adjacent fluorescence spots, which are closer together than the resolution limit of the microscope, are not resolved. If, however, the fluorophores are excited with a slight time-shift, the distance between two consecutively-blinking, adjacent fluorescent molecules may be exactly determined, even beyond the diffraction barrier.
This concept has been realised in three basic types of super-resolution microscopes. Stefan Hell, meanwhile director of the NanoBiophotonics department at the Max-Planck-Institute for Biophysical Chemistry in Göttingen, Germany, translated it in the late 1990s into the Stimulated Emission Depletion (STED) microscope.
In STED microscopy, the object is scanned point-for-point with a laser beam that excites the fluorophores inside the probe. The trick is to superimpose a stimulation (excitation) beam with a red-shifted, depletion laser pulse that surrounds the excitation beam as a doughnut-shaped ring. The depletion pulse quenches the fluorescence in the “doughnut ring” and reduces the size of the final fluorescent spot to a diameter that is way beyond the diffraction limit.
Theoretically, it is possible to narrow down the fluorescent spot into the sub-nanometre range by increasing the intensity of the doughnut-shaped STED pulse. In practice, STED microscopes have achieved a spatial resolution of about 20 nm.
In STED microscopy whole ensembles of fluorescent molecules are switched on and off by a laser pulse. Eric Betzig and Harald Hess, now at the Janelia Farm campus of the Howard Hughes Medical Institute in Virginia, USA, followed a slightly different strategy when assembling a super-resolution microscope in the living room of Hess’ house in La Jolla in 2005.
Their Photo Activation Localization Microscope (PALM) is based on the precise localisation of single fluorescence molecules which are gently turned on and off with a weak laser pulse. The low laser power ensures that only a very small population of photoactivatable fluorescence proteins is stochastically activated and detected via a CCD camera. Repeated rounds of activation, detection and bleaching deliver a map of the blinking molecules, which is transformed by a computer programme into a super-resolution image.
Similar single molecule techniques lead to the nearly identical, super-resolution microscopy methods Fluorescence Photoactivation Localization Microscopy (FPALM) and Stochastic Optical Reconstruction Microscopy (STORM), which have been implemented in commercial super-resolution microscopes. The resolution of single molecule, super-resolution microscopy techniques, such as N-STORM, may reach 20 nm in the x,y-plane and approximately 50 nm along the z-axis.
Mats Gustafson, who sadly passed away in 2011, gave birth to the third prototypic super-resolution microscope, termed Structured Illumination microscope (SIM), when working at the University of California at the beginning of the new millennium. In SI-microscopy, Moiré patterns, which may be generated by placing a diffraction grid between the excitation laser beam and the object are superimposed on the specimen. The grid pattern is shifted or rotated in tiny steps against the object and an image is captured at each step. The generated image set is subsequently translated by a computer algorithm into a super-resolution image.
Though the resolution of SI-microscopes is ‘only’ about 100 nm in x,y and 250 nm along the z-axis, it offers some attractive features for cell biologists. One of SIM’s major advantages is the increased resolution along the optical axis, even through a 10 µm thick sample, allowing 3D super-resolution images of intracellular components.
STED microscopy and PALM (STORM) rely on the switching of photoactivatable fluorescence proteins between on-and-off states with sophisticated and hard-to-implement laser techniques. However, much simpler super resolution imaging methods are already on the horizon. In a recent Nature Methods paper, Peng Yin’s group from Harvard, Boston, published a very smart variation of the imaging technique PAINT (Point Accumulation for Imaging in Nanoscale Topography), dubbed DNA-PAINT (Jungmann et al., doi:10.1038/nmeth.2835).
The basic idea of DNA-PAINT is pretty simple: fluorescent-labelled DNA molecules interact transiently with the sample and the resulting fluorescence signals are converted into a super-resolution image. DNA-PAINT relies on the interplay between DNA ‘docking’ and DNA ‘imager’ strands.
The docking strands are immobilised on the cellular structure of interest via a streptavidin bridge between a biotinylated antibody against the target and the biotinylated docking strand. Each time a fluorescent-labelled imager strand transiently hybridises with a docking strand, a fluorescent signal is emitted (in the unbound state only background fluorescence is observed). Similarly to PALM, fluorescent signals are collected by a CCD camera and are subsequently translated into a super-resolution image of the cellular structure.
DNA-PAINT is especially suited for multicolour and multiplex imaging. To demonstrate multicolour imaging with DNA-PAINT, Jungmann and his colleagues labelled microtubules with a docking strand hybridising to a Cy3b-labelled imaging strand and mitochondria with one matching an Atto-655-labelled imager strand.
For multiplex super-resolution imaging, the group applied a variation of DNA-PAINT called Exchange-PAINT. In this case, the target structures are initially labelled with orthogonal docking strands, before the first imager strand is added and a picture is taken. The imager strand is subsequently removed by a wash step and a second one is added (harbouring the same fluorescence label) that hybridises to another docking strand and so on.
By assigning a different pseudo-colour to each newly-added imager strand, a multicolour image of the sample is generated (applying the same fluorescence label and the same laser at each imaging step). Hence, super-resolution multiplexing with Exchange PAINT is only restricted by the number of orthogonal DNA docking sequences and not by the number of spectrally distinct dyes as in current multiplexing approaches.
Sounds very promising but only time will tell, whether DNA-PAINT really is a game changer that enables super-resolution microscopy without specialised equipment or just an amendment of existing super-resolution microscopy techniques.
First published in Labtimes 03/2014. We give no guarantee and assume no liability for article and PDF-download.
Table of Products as PDF-download: Formatted for Printout