A cell membrane is a highly heterogeneous medium containing a wide variety of molecules and structural elements. All of these species may interact with each other, leading to rather complex lateral diffusion behaviors in the lipid bilayer. Depending on the local composition of its membrane neighborhood, a molecule can experience one of four prominent diffusion modes1 (Figure 1). Figure 1. The four principal modes for molecules diffusing in membranes. Molecules are shown as red spots, black lines represent their diffusion paths, and the fluorescence correlation spectroscopy (FCS) observation spot is indicated as a light green circle. (a) Free diffusion is possible when no obstacles are present; (b) Impermeable obstacles (dark gray) restrict particle motions to the free space; (c) Dynamic partition model where molecules can diffuse into and out of permeable domains (light gray), where they can be transiently trapped; (d) Meshwork model: Multiple adjacent barriers (gray solid lines) hinder the diffusion of particles. They can, however, exhibit a hop-like diffusion, jumping from one partition to the next. See P.-F. Lenne and L. Wawrezinieck, et al. (2006). Dynamic molecular confinement in the plasma membrane by microdomains and the cytoskeleton meshwork. EMBO Journal, Vol. 25, pp. 3245-3256. A species can move freely, if no significant interactions between it and other membrane components occur. Should impermeable obstacles be present, diffusion of the molecule will be restricted. Furthermore, the molecule can diffuse into and out of immobilized or slow-moving domains and even be transiently trapped there. If a series of barriers compartmentalize the membrane, such as the cytoskeleton, the molecular species will follow a hop-like diffusion pattern. Each of these models follow a specific diffusion law, which makes it possible to identify them by plotting the experimentally obtained diffusion time against the observation spot area2 (Figure 2). Since cell membranes play an active role in many biological processes and interactions, studying the different motion behaviors of these diffusing species can provide insights into the function, composition and organization of membranes in cell biology. Figure 2. Simulated diffusion time curves as a function of observation spot area obtained for the four diffusion models. Determining the diffusion time for spot sizes below the diffraction limit (green area) requires the use of stimulated emission depletion (STED)-FCS. For spot sizes above that limit, classic spot-variation FCS can be used. See P.-F. Lenne and L. Wawrezinieck, et al. (2006). Dynamic molecular confinement in the plasma membrane by microdomains and the cytoskeleton meshwork, EMBO Journal, Vol. 25, pp. 3245-3256. Fluorescence correlation spectroscopy One of the most commonly employed methods for studying diffusion processes is fluorescence correlation spectroscopy (FCS), a powerful method for determining the average diffusion coefficients of fluorescent molecules in solution or membranes. FCS measurements rely on recording the transition of several thousands of molecules through the focal volume3. The combination of short measurement times along with free positioning or scanning of the observation spot makes FCS an excellent tool for investigating diffusion heterogeneity over time and space. Differentiating between the heterogeneous diffusion modes requires the acquisition of diffusion times from different-sized observation areas. This method is called spot-variation FCS (sv-FCS)4. In recent years, sv-FCS has helped in elucidating the formation and behavior of lipid rafts in cell membranes5 and the dynamic confinement of molecules in membranes by the cytoskeleton and micro domains6. This method has also revealed how the active remodeling of cortical actin regulates the organization of the cell surface7,9. Stimulated emission depletion microscopy Since the observation volume in a confocal microscope is diffraction-limited, probing the membrane’s heterogeneity on the sub-100-nm scale becomes difficult and the diffusion behavior is only inferable by extrapolations. Being able to directly follow the diffusion beyond the diffraction limit would be of great use since many biologically relevant cell membrane structures are smaller than this limit. A method to realize this is stimulated emission depletion (STED) microscopy9, a technique that has become well-established for achieving spatial superresolution below 50 nm. Figure 3. (a) Schematic overview of a confocal microscope equipped with STED add-on. (b) Cross sections of the focal plane with and without STED donut (red), altering the observation area (shown in blue). Diffusion paths are indicated as black lines; the green parts represent path sections contributing to the FCS data. One way to perform FCS measurements with improved lateral resolution is to interface a time-resolved confocal microscope, such as PicoQuant’s MicroTime 200, with a STED add-on (Figure 3). This instrument is equipped with a segmented phase plate. The microscope has an illumination system featuring at least two pulsed diode lasers: one or more for exciting the fluorescent labels and a STED laser to deactivate their emissive state10. After all the laser beams have been spatially overlaid, they pass through the segmented phase plate before being imaged onto the sample by the objective. The segmented phase plate modulates the polarization profile of the STED laser leading to a donut in the focal plane, while the beam profile of the excitation lasers is retained. The STED donut deactivates the excited state of the fluorescent species present in its area by inducing stimulated emission at a longer wavelength than that of the STED laser. Light from the excitation and STED lasers as well as from the stimulated emission are excluded from detection by means of appropriate bandpass filters. Therefore, only the desired emission from the center of the STED donut will be detected, leading to a smaller effective observation area with a diameter below the diffraction limit (Figure 4). Figure 4. Plot of the observation spot size as a function of the STED laser power, as determined from images of immobilized crimson beads with 20-nm diameter. Fluorescence lifetime correlation spectroscopy By increasing the STED laser intensity, one can gradually shrink the observation spot even more (Figure 5). Thus, diffusion times can be obtained for spot sizes ranging from 250 nm down to 40 nm, making it possible to determine the diffusion model mode and to probe the local membrane environment on the same scale. Alternatively, the observation spot size can also be reduced by applying a variable time gate to a recorded FCS data set. The latter method is called gated STED (gSTED) or fluorescence lifetime correlation spectroscopy (FLCS) and has been successful in the investigation of the diffusion behavior of dye-labeled lipid analogues in cell membranes11. Time-gated FCS methods do have a series of limitations, including the inability to record a “reference” data set from the same spot under confocal conditions at the same time as STED-FCS data. Figure 5. Schematic representation of the pulsed interleaved excitation pattern used for quasi-simultaneous FCS probing under STED and confocal conditions. Excitation laser pulse (blue), STED pulse (red), and resulting fluorescence (green). The dashed boxes indicate the time gates applied in the data analysis step for separating STED and confocal contributions. Introducing a sample illumination scheme fully based on pulsed interleaved excitation (PIE) allows the collection of FCS data from the same sample spot nearly simultaneously under both confocal and superresolution conditions12 (Figure 5). In the simplest possible case, employing only one excitation wavelength, the sequence consists first of an excitation laser pulse (blue), followed by a slightly delayed STED laser pulse (red). The non-quenched emission light can then be collected, leading to a fluorescence decay obtained under STED conditions. The second part of the sequence features only the excitation pulse and the resulting decay curve is thus collected under confocal conditions. By applying time gates to the collected photon arrival times, one can restrict the data analysis to separate STED and confocal conditions. For multicolor experiments such as those where several fluorescent label types present in the membrane are to be tracked nearly simultaneously, the PIE pattern can be extended to include additional STED/confocal excitation pairs featuring lasers with different wavelengths. Another advantage of this excitation scheme is the ability to fully exploit photon arrival time information by means of time-correlated, single-photon counting (TCSPC). This makes it possible to accurately separate different molecular species by taking into account subtle differences in their fluorescence characteristics13. Since all data is acquired from an identical area at nearly the same time, the obtained diffusion coefficients for STED and confocal conditions can be directly compared, even for spot sizes far beyond the confocal diffraction limit. Such a comparison is a straightforward way to check whether STED illumination has any influence on the investigated diffusion dynamics. Insights into cell membranes FCS measurements with fully pulsed interleaved excitation patterns are a versatile and powerful tool for studying the movement of fluorescent molecular species in cell membranes. A prominent example is the investigation of cholesterol-mediated lipid interactions, which may lead to the formation of nanodomains, such as lipid rafts, capable of capturing the diffusing species14. Furthermore, following the gradual changes in diffusion coefficient as a function of observation area size can provide insights into the organization and dynamics of the cell membrane beyond the diffraction limit. The method can be expanded for studies involving more than one fluorescent species by, for example, adding more excitation lasers to the PIE pulse pattern. In addition to cell membrane investigations, PIE-STED-FCS can also be applied to any research area where precise observation of the diffusion times of fluorescent species with high lateral resolution is desirable or required, such as membrane permeability or protein mobility studies. Meet the authors André Devaux is a technical writer for PicoQuant GmbH in Berlin; email: devaux@picoquant.com. Andreas Bülter is an authorized director for PicoQuant GmbH in Berlin; email: info@picoquant.com. Felix Koberling is the head of system development for PicoQuant GmbH; email: info@picoquant.com. References 1. P. Schwille and J. Corlach, et al. (1999). Fluorescence correlation spectroscopy with single-molecule sensitivity on cell and model membranes. Cytometry, Vol. 36, pp. 176-182. 2. H.T. He and D. Marguet (2011). Detecting nanodomains in living cell membrane by fluorescence correlation spectroscopy. Annu Rev Phys Chem, Vol. 62, pp. 417-436. 3. M.L. Kraft (2013). Plasma membrane organization and function: Moving past lipid rafts. Mol Biol Cell, Vol. 24, pp. 2765-2768. 4. H.T. He and D. Marguet (2011). 5. P.-F. Lenne and L. Wawrezinieck, et al. (2006). Dynamic molecular confinement in the plasma membrane by microdomains and the cytoskeleton meshwork. EMBO Journal, Vol. 25, pp. 3245-3256. 6. P. Schwille and J. Corlach, et al. (1999). 7. K. Gowrishankar and S. Ghosh, et al. (2012). Active remodeling of cortical actin regulates spatiotemporal organization of cell surface molecules. Cell, Vol. 149, pp. 1353-1367. 8. C. M. Blouin and Y. Hamon, et al. (2016). Glycosylation-dependent IFN-γR partitioning in lipid and actin nanodomains is critical for JAK activation. Cell, Vol. 166, pp. 920-934. 9. M. Koenig and P. Reisch, et al. (2016). ns-time resolution for multispecies STED-FLIM and artifact free STED-FCS. Proc SPIE, Vol. 9712, 97120T. 10. C. Eggeling and C. Ringemann, et al. (2009). Direct observation of the nanoscale dynamics of membrane lipids in a living cell. Nature, Vol. 457, pp. 1159-1162. 11. Ibid. 12. M. Koenig and P. Reisch, et al. (2016). 13. T. Niehörster and A. Löschberger, et al. (2016). Multi-target spectrally resolved fluorescence lifetime imaging microscopy. Nature Methods, Vol. 13, pp. 257-262. 14. P.-F. Lenne and L. Wawrezinieck, et al. (2006).