Stochastic modeling of photoswitchable fluorophores for quantitative superresolution microscopy

Stochastic modeling of photoswitchable fluorophores for quantitative superresolution microscopy

Frahm, Lars

Doctoral thesis

Date of Examination: 2016-11-23

Date of issue: 2017-04-13

Advisor: Hell, Stefan Prof. Dr.

Referee: Munk, Axel Prof. Dr.

Referee: Grubmüller, Helmut Prof. Dr.

Persistent Address: http://hdl.handle.net/11858/00-1735-0000-0023-3E16-0

Files in this item

Name: lars_frahm_dissertation.pdf

Size: 14,3 MB

Format: PDF

View/Open

Abstract

English

The diffraction limit of optical microscopy can be overcome by switching fluorophores between on- and off-states. In the on-state, the fluorophores emit fluorescence and can be detected. In the off-state, they remain dark. By modeling the stochastic switching of the fluorophores between the on- and the off-state, quantitative insights can be gained into the image formation process of superresolution microscopy. Reversible saturable optical fluorescent transitions (RESOLFT) can be used to drive the fluorophores to a non-fluorescent off-state using light. In a RESOLFT microscope, the size of the region in which there is a high probability that the fluorophores remain in the on-state can be confined to the sub-diffraction scale. To count the number of fluorophores in a RESOLFT image, the photoswitching process of reversibly switchable fluorescent proteins was modeled. Based on this model, a method was developed to calibrate the brightness per fluorescent protein directly from the image data. The result of the analysis is an estimate of the number density of fluorophores in the image. In stochastic optical reconstruction microscopy (STORM), a superresolved image can be reconstructed by switching fluorophores stochastically between the on- and off-state, and localizing the position of single molecule events. In activation-based multicolor STORM, several species of activator-reporter fluorophore pairs can be distinguished by activating them specifically at different points in time. In this method, the crosstalk between the color channels is typically high due to random activation of the reporter fluorophores, which can happen at the same time as the specific activation. We introduce a new approach to avoid this principal source of crosstalk in multicolor STORM, based on estimating the on-switching time of single molecule events with sub-frame precision. Our method enables the assignment of switching events to the correct molecular species with an improved error rate. This significantly reduces the crosstalk in activation-based multicolor STORM, and presents a solution for high resolution STORM imaging of multiple molecular species.

Keywords: Superresolution microscopy; Quantitative modeling; RESOLFT; STORM