Applicability Of An EM-CCD For Spatially Resolved TIR-FCS

Diffusion constants (DC) of surface near fluorescent particles can be measured by total internal reflection fluorescence correlation microscopy (TIR-FCS). The usage of EM-CCDs instead of photo diodes offers a high degree of parallelization and the possibility of extracting additional information by spatial cross-correlation (TIR image correlation spectroscopy, TIR-ICS).Since temporal autocorrelation functions of particle fluorescence critically depend on CCD parameters such as pixel size and geometry, binning, sampling rate, and gain, we explored systematically the performance of an EM-CCD as detector in TIR-ICS. We found that variations in the sample geometry can be well described by a structure term (ST). Whereas in axial direction the ST is described by evanescent field depth, the lateral extension of the detection volume was found to be well approximated by a Gaussian fit to the convolution of the CCD pixel geometry with the measured point-spread-function for single pixel read-out. For higher binning we empirically could show a linear relationship between the Gaussian approximation for the lateral ST and the size of the quadratic ROI on the CCD used for detection (binning), with a correction factor (slope) that is independent of the CCD chip used.To test the performance of CB TIR-ICS we measured diffusion coefficients (DC) and particle numbers (PN) of fluorescent probes of different sizes (Fluorospheres and GFP) at varying viscosities, concentrations, and sampling rates. This allowed calculating the resolution of the method expressed as the minimal relative resolvable difference in PNs or DCs. Distinguishing between different probe concentrations was possible with differences in PN of 30%. In contrast differences as low as 6% in DC could be distinguished at DC-to-sampling-frequency-ratios smaller than 0.5∗10ˆ4nmˆ2.This renders TIR-ICS suitable and ideal for measuring spatially resolved dynamics of proteins in viscous media such as in live cells.