Generalized Polarization Index (GP)

Calculation of GP value was carried out as:

where I₄₄₀ and I₄₉₀ correspond to the emission intensities of Laurdan at 440 and 490 nm, respectively, using 380 nm excitation wavelength. Values of GP vary from 1 to −1, where higher numbers reflect lower fluidity or higher lateral lipid order, whereas lower numbers indicate increasing fluidity. The GP measurements on live keratinocyte monolayers at different calcium concentrations (from [Ca2+] = 0.06 to 5 mM) were on one hand performed on a micro plate fluorescence reader (CLARIOstar, BMG LABTECH), with Laurdan fluorescence emission excited at 374 nm and recorded from 405 to 600 nm. The emission intensities at 440 and 490 ± 10 nm were used to obtain the GP values according to the above equation. On the other hand, confocal spectral imaging on live NHEKs at calcium concentrations [Ca2+] = 0.06 and 1.5 mM was performed on a Zeiss LSM 780 confocal microscope equipped with a 32-channel GaAsP detector array. Fluorescence of Laurdan was excited at 405 nm and detected between 415 and 691 nm. The images were then analyzed using a custom plug-in compatible with Fiji/ImageJ, as previously described 2 using a gamma variate fit of the spectra. A frequency histogram of the GP values was generated in Origin Pro (Oregon, United States), which disclosed two populations, one with high GP values representing the plasma membrane fraction and one with low GP values revealing the cytosolic fraction (e.g., from organelles). A fit of a double Gaussian distribution to the distribution allowed determining average GP values as well as standard deviations of both populations. In contrast, the plate reader measurements gave an average value over both populations. For both plate reader and spectral imaging n = 3 biological replicate were performed, while from each keratinocyte monolayer a minimum of 25 plate reader recordings or 15 confocal images (spectral images) were acquired and analyzed.

For the calculation of the GP index employing spectral imaging and rendering a GP map, we used a purposed-coded plug-in published elsewhere (Sezgin et al., 2015) and available at https://github.com/dwaithe/GP-plugin). Briefly, the plugin has in-built the GP general polarization formula and the discrete intensity values around the peak maxima (i.e., 440, and 490 nm) are obtained from a fitting to the spectrum obtained per pixel and extracting the values around these peak maxima from the nearest wavelength intervals. This methodology produces a spatial GP map representing the GP value for each pixel of the image. The outputs from these pixels are then exported into OriginPro (OriginLab, Oregon, United States) to produce the normalized frequency counts histograms per GP value. Every image pseudo-color representation has a corresponding custom look-up-table (LUT) matching the highest and lowest GP value obtained from the data.