Date of Award

Fall 10-1-2021

Document Type


Degree Name

Doctor of Philosophy (PhD)


Molecular Biophysics and Biochemistry

First Advisor

Pollard, Thomas


Photoconvertible fluorescent proteins (PCFPs) are widely used in super-resolution microscopy, and studies of cellular dynamics. Their photoconversion properties have enabled single-molecule localization microscopy (SMLM) by temporally separating closely-spaced molecules. However, our understanding of their photophysics is still limited, hampering their quantitative application. For example, counting fluorescently-tagged fusion proteins from the discrete localizations of individual molecules is still difficult. The red-to-green photoconvertible fluorescent protein mEos3.2 is favored by many due to its monomeric property, high brightness, photostability, compatibility with live cells, and 1:1 labeling stoichiometry. The fluorescent protein mEos3.2 is fused to the coding sequence of a protein of interest in the genome for endogenous expression or expressed exogenously and transiently in cells. Irradiation at 405 nm photoconverts mEos3.2 molecules from their native green state with an emission peak at 516 nm to their red state with an emission peak at 580 nm. Sparsely distributed photoconverted red mEos3.2 are excited at 561 nm and then localized for SMLM imaging. Understanding the factors that affect mEos3.2 photophysics can greatly strengthen its applications in imaging and quantitative measurements. However, we still do not know 1) how the behavior of mEos3.2 in live cells compares with fixed cells, and how the imaging buffer influences mEos3.2 photophysics in fixed cells, 2) how different imaging methods and laser intensities affect the behavior of mEos3.2, and 3) if there are unknown dark states of mEos3.2 that can further complicate imaging and quantitative applications of mEos3.2. In this body of work, I first reviewed the usage of photoconvertible fluorescent proteins in SMLM with a focus on its quantitative application. I discussed the significance, advantages, and challenges of counting molecules of interest tagged with mEos3.2 by SMLM. I highlighted how our limited understanding of mEos3.2 photophysics hampers its application in quantitative SMLM, thus requiring further investigation. Parts of this chapter are taken from Sun et al., 2021. In Chapter 2, I combined quantitative fluorescence microscopy and mathematical modeling to estimate the photophysical parameters of mEos3.2 in fission yeast cells. I measured the time-integrated fluorescence signal per cell, and rate constants for photoconversion and photobleaching by fitting a 3-state model of photoconversion and photobleaching to the time courses of the mEos3.2 fluorescence signal per cell measured by quantitative fluorescence microscopy. My method can be applied to study the photophysical properties of other photoactivatable fluorescent proteins and photoconvertible fluorescent proteins quantitatively, an approach complementary to conventional single-molecule experiments. This chapter is taken from Sun et al., 2021. In Chapter 3, I investigated how fixation affects the photophysical properties of mEos3.2, so that I could compare experiments conducted in live and fixed yeast cells with mEos3.2. Light fixation has been used to preserve cellular structures and eliminate movements of proteins to simplify the imaging and quantification process of quantitative SMLM. I discovered that formaldehyde fixation permeabilizes the S. pombe cells for small molecules, making the photophysical properties of mEos3.2 sensitive to the extracellular buffer conditions. To find conditions where the photophysical parameters of mEos3.2 are comparable in live and fixed yeast cells, I investigated how the pH and reducing agent in the imaging buffer affect the mEos3.2 photophysics in fixed cells. I discovered that using a buffer at pH 8.5 with 1 mM DTT to image mEos3.2 in fixed cells gave similar photophysical parameters to live cells. My results strongly suggested that formaldehyde fixation did not destroy mEos3.2 molecules but partially permeabilized the yeast cell membrane to small molecules. This chapter is taken from Sun et al., 2021. In Chapter 4, I investigated the effects of fixation and imaging buffer on mEos3.2 photophysics over a wide range of laser intensities by point-scanning and widefield microscopy, and also by SMLM. This chapter is taken from Sun et al., 2021. In Chapter 5, I alternated illumination at 405- and 561-nm to investigate the effects of 405- and 561-nm illumination separately. I discovered that 405-nm irradiation drove some of the red-state mEos3.2 molecules to enter an intermediate dark state, which can be converted back to the red fluorescent state by 561-nm illumination. I established the “positive” switching behavior (off-switching by 405-nm and on-switching by 561-nm illumination) of red mEos3.2 in addition to the previously reported “negative” switching behavior (switching off by 561-nm and switching on by 405-nm illumination), which could potentially affect counting the number of localizations of red mEos3.2 by quantitative SMLM. This chapter is taken from Sun et al., 2021. In Chapter 6, I described my ongoing progress towards developing a method to count molecules with SMLM using internal standards tagged with mEos3.2. I summarized the preliminary data on the internal calibration standards that I have tried. Further work is needed to optimize the standards and test the robustness and the reproducibility of the standards. Ultimately, this work can be applied to count the number of molecules in diffraction-limited subcellular structures with SMLM by converting the number of localizations to the number of molecules.