Metal-Induced Energy Transfer (MIET) for Live-Cell Imaging with Fluorescent Proteins

Metal-induced energy transfer (MIET) imaging is an easy-to-implement super-resolution modality that achieves nanometer resolution along the optical axis of a microscope. Although its capability in numerous biological and biophysical studies has been demonstrated, its implementation for live-cell imaging with fluorescent proteins is still lacking. Here, we present its applicability and capabilities for live-cell imaging with fluorescent proteins in diverse cell types (adult human stem cells, human osteo-sarcoma cells, and Dictyostelium discoideum cells), and with various fluorescent proteins (GFP, mScarlet, RFP, YPet). We show that MIET imaging achieves nanometer axial mapping of living cellular and subcellular components across multiple time scales, from a few milliseconds to hours, with negligible phototoxic effects.


INTRODUCTION
Fluorescence nanoscopy 1 beyond the classical diffraction limit of optical microscopy has become an indispensable tool for modern life sciences, allowing to discern the spatial organization of biological structures down to molecular length scales. The first successful method of super-resolution microscopy was stimulated emission depletion (STED) microscopy, 2 which was followed by the big family of singlemolecule localization microscopy (SMLM) techniques. The latter comprises photoactivated localization microscopy (PALM), 3 stochastic optical reconstruction microscopy (STORM), 4 fluorescent PALM (fPALM), 5 direct STORM (dSTORM), 6 and point accumulation for imaging in nanoscale topography (PAINT). 7 These methods routinely achieve lateral resolutions down to a few dozen nanometers.
Achieving a comparable resolution along the third dimension (the optical axis) requires additional modifications of these methods: For STED, one has to apply special phase plates to generate an optical bottle along the optical axis, 8 and for SMLM, methods such as biplane imaging, 9 astigmatic imaging, 10 wavefront shaping, 11 or single-molecule selfinterference 12 have been developed for three-dimensional single-molecule localization. However, the axial resolution achieved by all these methods is typically by a factor of 3−5 worse than the lateral resolution, very similar to the situation encountered in classical diffraction-limited confocal microscopy.
To overcome this anisotropic resolution, interferometric methods such as iPALM 13 and isoSTED 14 have been developed. These methods use two opposing objective lenses to image the sample, similar to what is done in diffractionlimited 4π microscopy. 15 While achieving truly isotropic superresolution down to a few nanometers, these methods are technically highly complex which prevented their wide application so far. This is also true for the latest addition to the pool of super-resolution methods, MINFLUX, 16 which uses a complex three-dimensional triangulation method for pinpointing the position of a single molecule with nanometric and even isotropic resolution. 17 For achieving axial super-resolution close to an interface (∼200−300 nm axial distance), a different family of superresolution methods has been developed that rely on electromagnetic near-field effects. The first one is variable-angle total internal reflection fluorescence (vaTIRF) microscopy. 18 It is based on conventional TIRF microscopy but records several consecutive images of the same sample under different incidence angles of the totally reflected excitation light. This incidence angle variation generates evanescent excitation intensities with varying decay length. By applying an appropriate mathematical analysis, the axial distance of fluorescent structures from the glass surface can be recovered. 19 The second of these near-field methods is supercritical angle fluorescence microscopy (SAF microscopy), 20 which uses the fact that the electromagnetic nearfield of an emitting fluorescent molecule, which usually does not take part in its far-field emission (i.e., which does not contribute to observable fluorescence intensity), can couple into propagating and detectable far-field modes in the coverslip glass when the emitter comes sufficiently close to the glass surface (few hundred nanometers). These near-field-generated propagating modes travel at angles above the critical angle of total internal reflection and are thus called supercritical angle fluorescence (SAF). In contrast, the usual far-field emission of the molecule does strictly travel at angles below this critical angle (undercritical angle fluorescence or UAF). While the SAF emission intensity is strongly dependent on the distance of a molecule to the glass surface, the UAF emission intensity is independent of it. By measuring the ratio of both intensities, the axial position of an emitter can be determined with nanometer accuracy. 21−23 Metal-induced energy transfer (MIET) imaging exploits similar physics for achieving axial super-resolution. 24 Here, a glass coverslip covered with a thin metal film is used instead. The near-field modes of the emitting fluorescent molecule, which can be modeled as an oscillating electric dipole, couple to so-called surface plasmons (collective metal electron oscillations) in the thin metal layer (typically 20 nm of gold film). This near-field coupling is similar to Forster Resonance Energy Transfer (FRET) where a fluorescent molecule (the donor) transfers its excited state energy to an acceptor molecule next to it. This energy transfer leads to a strong distance-dependent modulation of fluorescence lifetime and brightness within a distance range of 10 nm. In the case of MIET, the energy from the fluorescent molecule is transferred to metal surface plasmons, that act here as the acceptor, within a distance range similar to vaTIRF or SAF (i.e., up to ∼150− 200 nm). 25 Over this distance range, the monotonous fluorescence-lifetime−distance curve (MIET curve) can be calculated accurately using a theoretical model described thoroughly in several previous publications and which can be used to convert the measured lifetime into an emitter's axial position. 24,25,29 Provided the availability of a fluorescence lifetime imaging microscope (FLIM), MIET does not require any modifications of the experimental setup except that it needs glass coverslips covered with a thin metal layer, which can be routinely produced today with chemical vapor deposition. For a MIET measurement, a fluorescent sample . A pulsed laser source is used for excitation. The excitation beam (shown in orange) is guided through a cleanup filter followed by a beam splitter cube (90:10) which reflects 10% of the excitation light to the scan unit. After the scan unit, the beam is focused onto the sample through a high numerical aperture objective lens. Fluorescence light is collected by the same objective and propagates through the scan unit to the beam splitter cube which transmits 90% of emission and reflects 10%. The transmitted light is then passed through an appropriate long-pass filter after which the beam is focused through a pinhole by a lens. Light after the pinhole is refocused onto the active area of a single photon-sensitive avalanche photodiode. Optionally, a band-pass filter can be used before the detector to reject residual scattering light originating from the metal film of the MIET substrate. (B and C) Intensity and fluorescence lifetime images, respectively, of an exemplary D.d. cell. Scale bar: 10 μm. on top of the coverslip is scanned with a focused laser beam for acquiring photons within a thin axial section with a thickness between 500 nm and 1 μm, determined by the numerical aperture of the objective and the pinhole diameter. MIET does not affect the light collection of the confocal microscope but only affects the fluorescence properties of the dyes within its range. Instead of using the brightness of a single emitter or a monolayer of emitters as in vaTIRF or SAF microscopy, MIET exploits the measured fluorescence lifetime to determine axial distances from the metal surface. This makes it more robust against intensity-affecting artifacts that may impact vaTIRF or SAF microscopy results. The localization accuracy depends on the number of detected photons which determines the accuracy of the determined fluorescence lifetime, and with a few thousand of detected photons one typically achieves an axial resolution down to a few nanometers.
Previous applications of MIET imaging include mapping the topography of the basal membrane of living cells, 24 threedimensional reconstruction of focal adhesions and stress fibers, 26 measuring the distance between the inner and outer envelope of the nucleus, 27 visualizing the dynamics of epithelial−mesenchymal transitions (EMT), 28 single-molecule localization and colocalization, 29,30 mapping the basal membrane and lamellipodia of human aortic endothelial cells, 31,32 and, in combination with SMLM, three-dimensional isotropic resolution imaging of microtubules and clathrin pits. 33 Recently, it was shown that substituting the metal layer with a single sheet of graphene (graphene-induced energy transfer or GIET imaging) achieves a ca. 10-fold higher axial resolution than MIET (down to a few Ångstroms), within a distance range of ∼25 nm from the graphene. 34−37 From here on, we refer to the axial distance of molecules or structures from the surface as height for consistency with previous works.
So far, all MIET/GIET applications mentioned above were done exclusively with synthetic organic dyes as fluorescent labels. For applying MIET to live-cell super-resolution microscopy, it is important that MIET also works with fluorescent proteins in living cells. Fluorescent proteins are widely used to label cellular structures of interest and are conveniently expressed within genetically modified living cells. 38 However, applying MIET to live cell imaging with fluorescent proteins is much more challenging than using MIET with synthetic organic dyes. First, one has to ensure that the average fluorescence lifetime of the used proteins is homogeneous throughout the cellular structures (in the absence of any quenching metal layer). Second, one has to cope with the usually non-monoexponential fluorescence decay of fluorescent proteins. 39 In the present paper, we demonstrate that fluorescent proteins (in particular, wild-type GFP and mScarlet) provide indeed fluorescent labeling with sufficiently homogeneous lifetime distributions across living cells, which is a necessary prerequisite for MIET and that the error introduced by the non-monoexponential decay of the proteins' fluorescence lifetime into the lifetime-to-distance conversion still allows for an axial resolution of approximately 10 nm. We demonstrate fluorescent-protein MIET imaging of actin stress fibers in living mammalian cells, and of the basal membrane dynamics of Dictyostelium discoideum (D.d.) cells with a nearly video-rate of imaging. Our work demonstrates the suitability of MIET for super-resolution microscopy of living cells and subcellular components labeled with fluorescent proteins while keeping phototoxicity to a minimum and achieving high temporal resolution. This is of paramount importance for the application of MIET to a wide range of biologically important problems.

Near-Video-Rate Monitoring of Changes in Cell− Substrate Adhesion of D.d. Cells.
Fluorescence intensity and lifetime images were recorded with a home-built confocal laser-scanning microscope; see Figure 1A and the "Materials and Methods" section for details. This microscope allows for taking live-cell fluorescence lifetime images with video-rate acquisition speed (fluorescence lifetime imaging microscopy or FLIM). The first biological system that we studied were live cells of the slime mold D.d., a social amoeba growing in soil and a model organism for investigating cell adhesion, motility, chemotaxis, and signal transduction. 40 Binding of cyclic AMP (cAMP) to the cells' cyclic AMP receptor 1 (cAR1) results in amplification of actin polymerization at the leading edge of D.d cells, which leads to the formation of membrane protrusions known as pseudopodia. 41,42 The cell line used in this work was derived from the axenically growing strain D.d. carA-GFP. MIET imaging was done for cells in their vegetative and development stages (see the "Materials and Methods" section for detailed description of cell culture and transfection protocols). Unlike mammalian cells, D.d. cells lack integrin. Since they can bind equally well to hydrophobic and hydrophilic surfaces, adhesion is likely governed by van der Waals interactions of membrane glycoproteins with the substrate. 43 Furthermore, cell−substrate adhesion changes during the developmental time of D.d. cells. 44 The D.d. cells in their developed stage used in this study were starved for 6 h and pulsed with 50 nM cAMP every 6 min over the duration of starvation. The pulsing renders the cells chemotactic, which makes them polarized and let them migrate faster toward the source of cAMP. As a result, these cells show rapid motion and membrane dynamics which makes them ideal systems for checking the live-cell imaging capabilities of MIET. Figure  1B,C shows fluorescence intensity and corresponding lifetime images of a representative D.d. cell where cAR1 in the cell membrane was labeled with green fluorescent protein (GFP).
We fitted recorded time-correlated single-photon counting (TCSPC) histograms from each pixel to obtain fluorescence lifetime values (see the "Materials and Methods" section for details on lifetime fitting). To convert lifetimes into heights, we first measured the (free-space) lifetime of GFP attached to cAR1 for cells on a glass slide without any metal layer. The measured value was τ 0 = 3.1 ns and did not show any systematic variation across cells. Together with the fluorescence quantum yield value of GFP of ϕ = 0.79, 45 we computed the lifetime-on-distance MIET calibration curve for cells on gold-coated glass cover slides (see the "Materials and Methods" section and Supporting Information section S3 for MIET curve calculation, Figure S7 for free-space lifetime measurement on GFP and MIET calibration curve calculation, and Figure S8 for estimated variations in height determination at various theoretical quantum yield values of GFP when attached to cAR1). Figure 2A presents reconstructed height images obtained from FLIM images that were converted using the precalculated MIET lifetime-to-height calibration curve. Shown are cells in their developed (left panel) and vegetative stage (right panel) (see Supporting Information section S3 and Figure S9 for FLIM images and Videos S2, S2e, S3, and S3e). We observe a mean height of 47 ± 8 nm for the cell in its vegetative stage, ACS Nano www.acsnano.org Article while the cell in its developed stage approaches the surface much closer, exhibiting a mean height of 27 ± 3 nm (see Figure 2B).  Figure S9) were recorded at a frame rate of 20 Hz (see the "Materials and Methods" section). However, pertaining to the requirement of a reasonable photon budget for fluorescence lifetime fitting, further analysis was done with a 4× frame binning. Figure  Three-Dimensional Architecture of Actin Stress Fibers in Mammalian Cells. The second biological system studied with FP-based live-cell MIET imaging was the threedimensional architecture of actin stress fibers in mammalian osteosarcoma cells (SAOS-2) and in human mesenchymal stem cells (hMSCs) using electroporation with pLifeact-mScarlet (see the "Materials and Methods" section). The force transmission from contractile acto-myosin stress fibers

ACS Nano
www.acsnano.org Article ness in order to detect fibers near the basal membrane with short acquisition times, (b) highly monoexponential fluorescence decay, and (c) highly uniform expression levels and decay curves between different cells. Promising fluorophores with near monoexponential decay were selected based on past lifetime measurements in the lab. 51,52 Test measurements were done with lifeact-TagRFP, -YPet, -mTurquoise, and -mScarlet (see Figures S10−S12). In our experiments, mScarlet showed the best overall performance. We have performed MIET imaging of stress fibers in live hMSC 24 h and SAOS-2 cells 48 h after transfection. Figure 4 shows an overview of lifetime distributions for both cell lines and Figure 5 shows images of stress fibers emerging from focal adhesions together with calculated height curves. Brightness in these images corresponds to fluorescence intensity, while color encodes the fluorescence lifetime. Height profiles for the three annotated fibers in Figure 5A are shown in Figure 5B−D. This allows to reconstruct the 3D structure of stress fibers, which is important to fully understand mechanical cell−matrix interactions, in particular the direction of generated and transmitted forces. For the three analyzed stress fibers, actin heights above substrate vary from 60 to 100 nm which is in good agreement with fixed-cell measurements using interferometric PALM (iPALM) 50 and MIET imaging. 53 Our results show very shallow inclination angles of stress fibers with respect to the substrate surface (around 1°), which is in good agreement with our recently reported values using fixed cells and synthetic dyes. 26 Restructuring of the cytoskeleton and focal adhesions can occur on time scales ranging from minutes to hours. It is thus imperative to observe these cellular processes quantitatively over long periods of time. Using our MIET imaging setup, we were able to image cytoskeletal changes in living cells over 7 h; see Video S1 and the Supporting Information for details.

CONCLUSION
Our results highlight the feasibility and capabilities of MIET microscopy for axial super-resolution live-cell imaging with fluorescent proteins. We showed this for two different types of mammalian cells (hMSCs and SAOS-2) and for the slime mold D. discoideum, using different fluorescent proteins (mScarlet and GFP). We achieved an axial resolution of a few nanometers at an effective image rate of 5 Hz on two different developmental stages of D. discoideum cells, which is only limited by the brightness and density of the fluorescent proteins but not the MIET imaging microscope itself. Our results suggest that the basal membranes are twice higher from the substrate and more rugged in the vegetative stage than in the developed stage of D. discoideum cells. The ability to image live cells allows us to follow the temporal evolution of membrane protrusions with respect to the substrate with subsecond resolution. Furthermore, we presented long-term measurements of up to 8 h of SAOS-2 cells, demonstrating the low phototoxicity of MIET imaging, which offers the possibility to follow and quantify cellular processes with extreme axial resolution over long time scales. In particular, this will be of considerable interest for the study of the structure and dynamics of the cytoskeleton and of the focal adhesion machinery.
An important result is a demonstration that MIET imaging works well with FP labeling. FPs are on average less bright than organic dyes and do often show a non-monoexponential fluorescence lifetime decay. Nonetheless, our results show that both these limitations still allow for decent live-cell MIET imaging with excellent signal-to-noise ratio and superior axial resolution. Considering different possible origins for the biexponential fluorescence decay we could show that we accurately estimate the radiative rate based on the average fluorescence decay time τ ̅ (see the Supporting Information), even for nearly biexponential decays with similar amplitudes of both decay components. For typical FPs, the error of the determined height based on our calculations in Supporting Information section S1 is considerably less than 2 nm. This makes MIET imaging an ideal microscopy modality for a wide range of applications requiring live-cell imaging with fluorescent proteins. When compared to nonfluorescent interferometric techniques such as quantitative phase imaging 54 that deliver nanometer axial resolution, MIET imaging provides the specificity of fluorescence in addition to a similar axial resolution. Moreover, because MIET imaging is mostly insensitive to refractive index variations, it is not affected by inhomogeneous protein or cell organelle distributions. Taken together, we believe that live-cell MIET imaging with fluorescent proteins can be an important addition to the toolbox of axial super-resolution fluorescence microscopy techniques for cell biology. When the cells deplete their food source and start to starve (in our case, the HL5 medium is removed and replaced by buffer), they enter a developmental cycle. For the preparation of experiments, cells were starved in shaking phosphate buffer at 150 rpm (PB, 2 g of KH 2 PO 4 and 0.36 g of Na 2 HPO 4 ·2H 2 O per 1 L, pH 6) for 5 h at a density of 2 × 10 6 cells/mL. The shaking culture was pulsed with 50 nM cAMP (Sigma) every 6 min over the course of the starvation time. After the corresponding starvation time, the cells were harvested and washed in PB. An aliquot of the cell suspension was applied to the experimental setup and the cells were allowed to spread on the glass substrate for 10 min at room temperature before starting the experiment. For experiments with vegetative D.d. cells, they were detached from the Petri dish bottom, washed twice with PB, and applied directly to the experimental setup without any additional starvation time.

MATERIALS AND METHODS
Transfection. Cells were transfected with the plasmid pLi-feAct_mScarlet-i_N1, a gift from Dorus Gadella (Addgene plasmid #85056, RRID:Addgene_85056), using the Lonza 4D Nucleofector (Lonza, Cologne, Germany, AAF-1002B & AAF-1002X). hMSCs were transfected using the P1 buffer (Lonza, Cologne, Ger, V4XP-1012) and pulse code FF-104. SAOS-2 cells were transfected using ACS Nano www.acsnano.org Article the SF buffer (Lonza, Cologne, Ger, V4XC-2012) and pulse code DS-150. Subsequent to transfection cells were seeded onto prepared dishes at 80.000 cells per well. Dead cells were removed by rinsing with medium 12 h after transfection. The cell line carA-GFP were derived from the axenically growing strain D.d. AX3. They were kindly provided by G. Gerisch (Max Planck Institute for Biochemistry, Martinsried, Germany). Sample Preparation. For SAOS-2 and hMSC measurements, glass coverslips of 25 mm diameter (VWR, Darmstadt, Ger, ECN631−1584) were covered with 2 nm of Ti, 15 nm of Au, 1 nm of Ti, 20 nm of SiO 2 . Both, gold-coated and regular glass coverslips were cleaned and functionalized with collagen-I using heterobifunctional cross-linking chemistry as described earlier. 55 In brief, all slides were plasma cleaned for 15 min, sonicated first using 99% EtOH and second using 2% APTES (Sigma-Aldrich Co., St. Louis, MO, 440140) in EtOH and treated with a 0.5% glutaraldehyde solution (Sigma-Aldrich Co., St. Louis, MO, C7651). Glasses were rinsed twice with PBS and once with HEPES buffer. Sulfo-SANPAH (Thermo Fisher Scientific Inc., Waltham, MA, 22589) was added to the surface and activated under UV for 10 min. A solution of 0.2 mg/ mL collagen I (Rat tail collagen I, Corning Inc., New York, NY, 354236) in PBS is added and incubated overnight at 4°C. Glasses are rinsed twice with PBS and glued into bottomless ibidi dishes (ibidi, Grafeling, Ger, DIO01110) using UV-curable glue (NOA68, Norland products inc., Cranbury, NJ, 6801). Prepared dishes are UV sterilized for 2 h and rinsed three times with PBS before seeding of cells. For MIET measurements on D.d. cells, glass coverslips were first sonicated with 1 M KOH and plasma cleaned for 15 min. These cleaned coverslips were then coated with a 2 nm titanium layer (for better sticking of gold on glass) followed by evaporation of 15 nm of gold, 1 nm of titanium, and 10 nm of SiO 2 . Gold-coated coverslips were then glued into bottomless ibidi dishes (ibidi, Grafeling, Ger, DIO01110) using UV-curable glue (NOA68, Norland Products Inc., Cranbury, NJ, 6801). Prepared dishes are UV sterilized for 2 h and rinsed three times with PBS before seeding of cells.
Setup. Fluorescence lifetime images were acquired on a custombuilt confocal setup. The excitation light was generated by an 80 MHz pulsed white-light laser (SuperK Power, Koheras) and the wavelength was selected by an AOTF (SpectraK Dual, Koheras). The beam was coupled into a single-mode fiber (PMC-460Si-3,0-NA012-3APC-150-P and fiber coupler 60SMS-1-4-RGBV-11−47, both Schafter+Kirchhoff GmbH, Germany) and after the fiber recollimated by an objective (UPlanSApo 10×, 0.40 N.A., Olympus). After passing a cleanup filter (F37−563 and F49−488 (for D.d. measurements), AHF), a 90/10 beam splitter was used to reflect the excitation light into the microscope and separate it from the emission. The reflected beam was directed into a laser scanning system (FLIMbee, PicoQuant) and then into a custom side port of the microscope (Olympus IX73). The three galvo mirrors in the scanning system are imaged onto the backfocal plane of the objective (UApo N, 100×, 1.49 N.A. oil, Olympus) with 180 mm and 90 mm achromatic lenses. The sample could be moved by a manual xy-stage (Olympus) and a z-piezo stage (Nano-ZL100, MadCityLabs). Fluorescence in the sample was collected by the same objective and descanned in the scanning system. The fluorescence light that passed the 90/10 beam splitter was then focused onto a pinhole (100 μm, Thorlabs) with a 180 mm achromatic lens. Data Analysis. Data analysis was carried out using custom routines written in Matlab (MathWorks). First, scan lines were aligned using a linear shift. Then, TCSPC histograms of individual pixels were generated and corrected for dead-time artifacts. 56 A monoexponential fit including a measured instrument response function (IRF) was used to estimate the fluorescence lifetime. The TCSPC curve was fitted with the function where χ(t) denotes the instrumental response function of the TCSCP system, τ denotes the fluorescence lifetime, a denotes the fit amplitude, b denotes the background, a s denotes the scattering amplitude, and s denotes the color shift of the IRF. The scattering amplitude accounts for the luminescence of the gold, and the shift s was introduced as the IRF depends on the count rate. We found for our detectors that linear shifts was sufficient to account for this effect. The lifetime was converted into a height using the appropriate MIET curve for the sample. Images and movies were created with the lifetime or height information in color, using the intensity (rescaled to 0−1) as the transparency of the image. MIET Curve Calculation. The theoretical curve for converting measured lifetimes into heights were computed based on the model by Chance, Prock, and Silbey. 25,57,58 For the substrate, we assumed a layered system of (from bottom to top): glass, 2 nm of titanium, 15 nm of gold, 1 nm of titanium, and 20 nm of silica. The medium of the fluorophores and above was assumed to have the refractive index of water, refractive indices of the metals were taken from literature. 59 The theoretical MIET curve for GFP in connection to D.d. measurements were calculated similarly where we assumed a layered system (from bottom to top) glass, 2 nm of titanium, 15 nm of gold, 1 nm of titanium, and 10 nm of silica. Please see Supporting Information section S3 for the theoretical MIET curve.
Impact of Multiexponential Fluorescence Decays. Like other methods that take advantage of fluorescence quenching (Forster resonance energy transfer, Stern−Volmer quenching) the analysis is based on the fluorescence decay time, being the inverse of the rates depopulating the excited state. These are the emission rate k fl and nonradiative processes that are summarized into k nr . The quenching adds a further rate to the total and the analysis is all about quantifying this additional contribution. For this, k fl and k nr are assumed to be constant and can be obtained from the fluorescence decay time and the quantum yield of the dye in "free-space" conditions. All these processes lead to a strictly monoexponential decay of the decay allowing seamless analysis of the data. In reality, however, one usually observes fluorescence decays that deviate from the expected behavior. Most common are biexponential decays. These are usually explained as two populations of slightly different forms of the fluorophore (e.g., in a different local environment, charges near the chromophoric part of the molecule etc.).
Accurate determination of two lifetimes and their respective contribution to the decay is surprisingly hard and requires collecting many photons. Therefore, we estimate the amount of quenching based on the average fluorescence decay time, which can be ACS Nano www.acsnano.org Article determined quite simply and reliably. For checking how this simplified analysis impacts the obtained height values, we simulated biexponential decays based on the rate constants of lifeact-mScarlet and the MIET curve (applied to each monoexponential decay component). The average lifetime of these decays was determined and used to calculate a height values using again the MIET curve. The results prove that the analysis based on the average fluorescence decay time adds no significant error to the determined height values of the fluorophores (see Supporting Information section S1). Protein Expressions for Quantum Yield Measurement. lifeact-mScarlet from SAOS-2. For quantum yield measurements, SAOS-2 cells (passage 13) were cultured and transfected as previously mentioned. The fusion protein (lifeact-mScarlet) was extracted using a RFP-Trap (Chromotek, Planegg−Martinsried, Germany, rta-20). Cells were harvested, lysed, and protein was extracted as advised by vendor. For elution, we bound proteins by adding 50 μL of 0.2 M glycine (pH 2.5) under 30 s of constant mixing. After 2 min of centrifugation at 2.500 g, the supernatant was transferred to a new tube and pH neutralized by the addition of 5 μL of 1 M Tris base (pH 10.4).