Data normalization.

Image-filtering and thresholding steps were applied to DAPI (blue) and mCherry (red) images to generate the masks for nuclear, cytoplasmic, and whole-cell areas. Using these masks, integrated intensities (3 wavelengths) were measured for all regions for each cell (600 to 800 cells/set). The values were imported into R software (http://www.r-project.org) for filtering and analysis. The integrated nuclear DAPI (blue), integrated nuclear mCherry (red), integrated cytoplasmic mCherry (red), and cellular GFP (green) intensities (in arbitrary relative fluorescence units [RFU]) were plotted over time for each cell within each experimental set. Cells with no changes in mCherry or GFP (i.e., not transfected) were discarded. However, as the background GFP intensity was generally low, the filtering criteria were set loosely to avoid falsely omitting cells that might be expressing some 2A^pro. The time tracks were cut short if the cell started exhibiting CPE (>10 h posttransfection). Since the total integrated mCherry intensity of each cell (the sum of nuclear and cytoplasmic integrated mCherry intensities) did not change measurably throughout an experiment (Fig. 5), for mCherry-NLS cells, the integrated nuclear mCherry intensities were normalized by the total integrated mCherry intensities (ratioN) at each time point. Likewise, for the mCherry-NES cells, at each time point, the integrated cytoplasmic mCherry intensities were normalized by the total integrated mCherry intensity (ratioC). The change in the nuclear or cytoplasmic mCherry ratio (ΔratioN or ΔratioC) throughout the course of the experiment with respect to the initial ratio was calculated for each 2A^pro (200 to 250 cells/2A^pro type). The ΔratioN or ΔratioC was normalized for cellular 2A^pro expression levels by dividing the calculated ΔratioN or ΔratioC by the average cellular GFP intensity to compare efficiencies of different 2A^pro to disrupt nuclear transport.