Performing Fluorescence Correlation Spectroscopy (FCS) measurements in multiple points within a sample is necessary for understanding the dynamics of intrinsically heterogeneous samples, such as live cells or organisms. However, the number of points in which FCS measurements are performed is strongly limited in conventional FCS for at least three reasons: 1) Conventional FCS is conducted in confocal mode and thus at a single spot at a time 2) Each FCS measurement takes at least several seconds; the resulting photo damage restricts the number of repeated measurements that can be performed in each cell. 3) The sequential points acquired in confocal FCS do not represent an accurate image of a cell at a single time point. These limitations of conventional FCS motivated the development of imaging FCS allowing parallel FCS measurements in many points covering a whole region of interest within the sample. This provides access to previously inaccessible structure-dynamic relations.

Conventional FCS instruments use point detectors thus making multiplexing of measurements difficult. We therefore explored the use of electron multiplying charge coupled device (EMCCD) and scientific Complementary Metal-Oxide-Semiconductor (sCMOS) cameras as detectors for imaging FCS [1, 2]. FCS requires a small observation volume to fluorescence intensity fluctuations caused by single molecules. In confocal FCS this is provided by the laser focus and the confocal pinhole. In camera based imaging FCS the limitation of the observation volume is achieved by the finite extent of the pixels (or group of pixels) on the camera chip and a laser illumination that provides intrinsic z-sectioning. Two distinct illumination schemes have been employed in our instruments. The z-sectioning is achieved either by a thin light sheet generated by a Selective Plain Illumination Microscope (SPIM) or the exponentially decaying evanescent wave generated due to Total Internal Reflection (TIR). We call the latter method, Imaging Total Internal Reflection-Fluorescence Correlation Spectroscopy (ITIR-FCS). ITIR-FCS is ideal for mapping concentrations and diffusion coefficients in planar structures such as supported lipid bilayers or plasma membranes of cells and to study peptide-membrane interactions . Our experiments suggest that ITIR-FCS is a calibration free FCS variant, providing, thus, absolute diffusion coefficients of molecules in membranes [3, 4]. In addition, this imaging approach allows as well the calculation of spatial cross-correlations to obtain information about mass transport and membrane heterogeneity [5]. Imaging FCS enables one to observe the temporal evolution of membrane dynamics as a time lapse ‘FCS movie’ which allows one to directly visualize concerted spatial change in membrane diffusion with pixel resolution. ‘FCS movies’, to the best of our knowledge, is the first demonstration of time lapse diffusion imaging (FCS movie).

Modern cameras allow simultaneous FCS data acquisitions in thousands of pixels (up to a million). The large number of FCS correlation curves obtained in each imaging FCS measurement led us to development of software for imaging FCS data analysis and motivated our interest in algorithms for automatic and objective selection of the appropriate model for fitting the FCS curve analysis (Bayes-FCS) in each individual pixel [6].

Here, we show that by using a scientific CMOS (sCMOS) camera, we can record 1,152,000 (1920×600) autocorrelation functions at 25 fps. The autocorrelations were calculated from a stack of 1500 images. The sample being measured here is 0.2 µm sized fluorescent beads.

The red correlation curve in the figure above shows representative correlation function acquired using a camera at a rate of 100,000 frames per sec. The blue curve shows a correlation acquired at a rate of 1,250,000 frames per sec. The circled region in the blue curve shows the diffusion component similar to the one shown in the red curve. We currently attribute the rest of the decays to detector electronics.

The difference in correlation function (ΔCCF) approach calculates the forward and backward correlation function between two pixels. If the movement is isotropic then it is equally likely that particles move from A to B or from B to A. In that case both correlation functions are the same and the difference of their area under the curve, the ΔCCF value, is 0. If there is, for instance, flow from pixel A to pixel B, then the two functions are not symmetric. Here it is more likely that a particle found in A will be later found in B, while a particle in B is unlikely to be later found in A. Thus the DCCF values will not be 0 but have a positive value (if the flow is reversed the ΔCCF value is negative). Note that the width of the DCCF distribution is an indication of the heterogeneity of the sample and the spread of the diffusion coefficients found on a sample.

The principle of ΔCCF is shown in graphs A-C. In graph D the distribution of ΔCCF value over a whole image is shown for diffusion and flow. (Refer fig. 5 in [7] and fig. 3 in [8])

The FCS diffusion law (see Fig. below) was introduced by Wawrezinieck et al. in 2005 [9] and is a further development of the idea that the dependence of the diffusion coefficient on the area of observation contains information of the structure of the medium in which diffusion takes place, as earlier applied already in FRAP. Therefore, the FCS diffusion law can provide information of possible domains and mesh works, which influence the diffusive process, even below the diffraction limit. Imaging FCS is ideal for the implementation of the diffusion law as observation areas of different size can be created by binning the pixels on a detector. Thus in imaging FCS one has to take a single measurement only, and can then in the processing step create areas of different size.

FCS diffusion law plots for various diffusion modes in the membrane: free diffusion (black dotted line), hindered diffusion in microdomain confinement (red line) and hop diffusion in meshwork compartmentalization (light blue line). Range of spatial resolutions of the observation areas (in scale) for each experimental variation: STED-FCS (dark blue circles), confocal FCS (green circles), z-scan FCS and Imaging FCS (orange squares) are shown. (This figure is adapted from the one in Ref. [10]). Our recent publication has characterized the FCS diffusion law in the presence of multiple diffusive modes [11].

[1] B. Kannan, J.Y. Har, P. Liu, I. Maruyama, J.L. Ding, T. Wohland, EMCCD-camera based fluorescence correlation spectroscopy, Anal. Chem., 78(10): 3444-51 (2006).

[2] Singh, A.P.; Krieger, J.W.; Buchholz, J.; Charbon, E.; Langowski, J.; Wohland, T. The performance of 2D array detectors for light sheet based fluorescence correlation spectroscopy, Optics Express, 21(7): 8652-8668 (2013).

[3] Bag, N.; Sankaran, J.; Paul, A.; Kraut, R.; Wohland, T. Calibration and Limits of Camera-Based Fluorescence Correlation Spectroscopy: A Supported Lipid Bilayer Study, Chemphyschem 13(11): 2784-2794 (2012).

[4] Sankaran, J.; Bag, N.; Kraut, R.S.; Wohland, T. Accuracy and precision in camera-based fluorescence correlation spectroscopy measurements, Anal. Chem. 85 (8): 3948-54 (2013).

[5] Sankaran, J.; Manna, M.; Guo, L.; Kraut, R.; Wohland, T. Diffusion, transport, and cell membrane organization investigated by imaging fluorescence cross-correlation spectroscopy, Biophys J, 97: 2630-2639 (2009).

[6] Guo, S.M.; Bag, N.; Mishra, A.;Wohland, T.; Bathe, M. Bayesian Total Internal Reflection Fluorescence Correlation Spectroscopy Reveals hIAPP-Induced Plasma Membrane Domain Organization in Live Cells, Biophy. J. 106: 190–200 (2014).

[7] Kraut, R.; Bag, N.; Wohland, T. Fluorescence Correlation Methods for Imaging Cellular Behavior of Sphingolipid-Interacting Probes, Methods in Cell Biology 108 (2012) 395-427.

[8] Bag, N.; Wohland, T. Imaging Fluorescence Fluctuation Spectroscopy: New Tools for Quantitative Bioimaging. Annu. Rev. Phys. Chem. 65 (2014) 225–48.

[9] Wawrezinieck, L., Rigneault, H., Marguet, D., & Lenne, P.-F. (2005). Fluorescence Correlation Spectroscopy Diffusion Laws to Probe the Submicron Cell Membrane Organization. *Biophysical Journal*, *89*(6), 4029–4042.

[10] Ng, X.W.; Bag, N.; Wohland, T. Characterization of Lipid and Cell Membrane Organization by the Fluorescence Correlation Spectroscopy Diffusion Law, CHIMIA International Journal for Chemistry 69 (3), 112-119.

[11] Veerapathiran S, Wohland T. The imaging FCS diffusion law in the presence of multiple diffusive modes. Methods. 2017 Dec 5. pii: S1046-2023(17)30229-3.