- Fast three-dimensional tracking
For researchers at Leiden University in the Netherlands, standard single-molecule fluorescence microscopy is not good enough, even though it can locate fluorophores with very high lateral accuracy. They wanted to extend it from two into three dimensions while imaging at high frame rates over a large area. Because existing methods could not do this, they developed their own solution by placing a cylindrical lens into the optical emission path, thereby introducing astigmatism.
To locate fluorophores in three dimensions while imaging at high frame rates over a large area, researchers placed a cylindrical lens into the optical emission path, thereby introducing astigmatism. Images courtesy of L. Holtzer, T. Meckel and T. Schmidt, Leiden University.
Adding imperfection enabled them to achieve 40-nm positional accuracy in a low-light regime as demonstrated using quantum dots, noted research team member and professor of physics of life processes Thomas Schmidt. With a stronger signal, the technique located sources with even better accuracy. In a fixed sample, the researchers went up to a signal of 5000 counts per quantum dot, which resulted in ~8-nm lateral and ~20-nm axial accuracy, he said.
In two dimensions, the positional resolution of single-molecule fluorescence microscopy can be driven far below the diffraction limit. This is done by collecting photons from a fluorophore and then applying statistics to the measured intensity distribution to better locate the molecule.
However, the trick does not work in the Z-direction because the positional error becomes very large at the focal point, and there is no way to distinguish between positive and negative defocus. Thus, when adding the third dimension, investigators have been forced to resort to such techniques as confocal microscopy, which is slow, is not sensitive enough without custom instruments and can follow only a single molecule at a time.
The researchers sought a way around these limitations. It occurred to Schmidt that a weak cylindrical lens placed in the emission path would change the situation. The introduced axial astigmatism would reduce the obtainable accuracy slightly in X and Y but would result in a large increase in the accuracy in Z.
The introduced astigmatism slightly reduces resolution in the X and Y while substantially improving resolution in the Z. (AOTF = acousto-optic tunable filter.)
Simulations indicated that the trade-off would be acceptable, so the researchers built a demonstration system. They started with an existing 2-D setup, using an argon/krypton laser from Spectra-Physics, a Zeiss microscope, a Roper Scientific CCD camera, a shutter, a dichroic mirror and an emission filter. Behind the filter but in front of the detector they added a cylindrical lens with a 10-m focal length that introduced 184 nm of astigmatism in their system.
After calibration to account for camera readout noise, they followed the path of streptavidin-coated quantum dots in a dextran solution for several minutes, using the 514-nm laser line to excite quantum dot emission at 705 nm. After capturing the images 35 times per second, they reconstructed the 3-D path of the particles and found the positional accuracy to be 47 nm in the X and Y and 90 nm in the Z. The diffusion constant that fit their data was in good agreement with that expected for a 22-nm-diameter particle, the size of the quantum dots.
Schmidt noted that these results are encouraging and that the researchers made these measurements as a check that they had done everything correctly. They then tracked the movement of quantum dots in a study of intracellular transport processes, incubating human embryonic kidney cells in a solution containing quantum dots. They found several modes of motion. Initially the particles seemed to randomly diffuse in all directions. Later they moved in a purposeful way as a result of cellular transport. During these times the quantum dots followed a directed motion along one axis while continuing to exhibit diffusional behavior in perpendicular directions.
Human embryonic kidney cells loaded with quantum dots by endocytosis enable the tracking of intracellular transport process. On the left is a transmission image of the cell measuring 55 × 55 μm. On the right is a fluorescence image of the same cell with a quantum dot circled. The dot can be tracked to an accuracy of a few tens of nanometers in all three dimensions, enabling transport processes to be studied.
The work is detailed in the Jan. 29 issue of Applied Physics Letters.
Such complex motion was unexpected, but Schmidt said that it was precisely to uncover and understand such motion that they developed the system. Currently, they are trying to understand transport of vesicles in cells to the level of detail that their system will allow.
Schmidt noted that the technique could be applied to virus tracking and perhaps elsewhere. “I hope we can be the first to follow individual molecules in tissue,” he said. “We are pretty close.
Contact: Thomas Schmidt, Leiden Institute of Physics, University of Leiden; e-mail: firstname.lastname@example.org.
MORE FROM PHOTONICS MEDIA