smICA: Open-Source Software for Quantitative, Lifetime-Resolved Mapping of Absolute Fluorophore Concentrations in Living Cells

TL;DR

The presented methodology, along with the software, is a promising tool for quantitative single-cell studies, including, but not limited to, protein expression, degradation of biomolecules (such as proteins and mRNA), and monitoring of enzymatic reactions.

Abstract

Advanced microscopy techniques are essential in biomedical research for visualising and tracking biomolecules within living cells and their compartments. Conventional fluorescence microscopy methods, however, often struggle with accurately measuring the absolute concentrations of fluorescent probes in living cells. To overcome these limitations, we introduce an open-source analysis tool, smICA (Single-Molecule Image to Concentration Analyser). The smICA method offers quantitative mapping of absolute fluorophore concentrations, lifetime-resolved filtering methods of the signal, intensity-based cell segmentation, and requires only a few photons per pixel. Our approach also reduces the time required for the determination of the mean concentration per cell, compared to the standard FCS measurement performed in multiple posts. To highlight the robustness of the method, we validated it against standard fluorescence correlation spectroscopy (FCS) measurements by performing in vitro (aqueous solutions of polymers) and in vivo (polymers and EGFP in living cells) experiments. The presented methodology, along with the software, is a promising tool for quantitative single-cell studies, including, but not limited to, protein expression, degradation of biomolecules (such as proteins and mRNA), and monitoring of enzymatic reactions.

Figures (5)

• Figure 1: Localization of the FCS measurement spots. The LSCM image of the cell with fluorescently-labelled (Cy5) mRNA injected into the cytoplasm marked as hashed region marked in the figure corresponds to the locations where FCS measurements were typically performed. The dashed curve represents the cell boundary. The FCS measurements were chosen to omit the endoplasmic reticulum and nucleus.
• Figure 2: The workflow scheme for the concentration imaging method. First, we calibrate the FCS setup using a dye with a known diffusion coefficient. From calibration measurements, we get $\omega$ - the width of the focal volume. Next, we performed the FCS measurements on the target samples to find the molecular brightness, $B$. Finally, the raster imaging was performed using the laser scanning confocal microscopy equipped with the SPAD detectors. The raw data, in the form of a time trace of photons acquired during scanning, were further filtered to remove unwanted background, afterpulsing, or autofluorescence photons. We used fluorescence lifetime-resolved filters that are typicaly used for the fluorescence lifteime corrlation spectroscopy (FLCS) described in references enderlein2005using and kapusta2007fluorescence. The filtered signal was combined into frames and integrated over all frames. Knowing the $\omega$ and $B$, we calculated the mean concentration of fluorophores in each pixel averaged over the data acquisition time.
• Figure 3: Statistical analysis of the influence of experimental parameters on the determination of concentration. Panels depict heatmaps of the estimated conditional probability $\hat{P}$ that the $C_\mathrm{im}\slash C_\mathrm{FCS}$ takes values between 0.75 and 1.25. $\hat{P}$ was estimated from the experimental data of $C_\mathrm{im}\slash C_\mathrm{FCS}$ for various pairs of parameters: (a) pixel size vs. total expected number of photons per molecule ($\Gamma$), (b) scan velocity $\Delta x\slash\Delta t$vs. molecular brightness, $B$. The $\hat{P}$ was estimated utilising the Nadaraya–Watson methodWatson1964Nadaraya1964 using the isotropic Gaussian kernel with adaptive bandwidth being the $k$-th ($k$=40) nearest neighbour distance.Demir2010 Black dots marks the experimental points, solid lines indicate $\hat{P}$ contour lines. The regions marked with /// depict regions of low trust where the effective local sample size, $n_{\mathrm{eff}}<100$.
• Figure 4: The influence of filtration and Region of interest size on the imaging of concentration. Panel a depicts a plot consisting of the fluorescence decay patterns for the non-filtered signal, the signal filtered by the TCSPC background removal, and the TCSPC pattern filtration methods; all acquired for the TRITC-labelled dextran polymer in water. Both filtering methods were based on the algorithm described in references enderlein2005using, and kapusta2007fluorescence. In the background removal method, we subtracted the constant value of the background (see main text) to find the TCSPC pattern to be filtered. In the TCSPC-pattern method, we applied a previously registered pattern of fluorescence decay acquired in a system with a sufficiently high S/N ratio. Panel b shows box plots representing the dispersion of the $C_\mathrm{im}\slash C_\mathrm{FCS}$ ratio data obtained for different region-of-interest sizes. ROI size $\sqrt{N_\mathrm{px}} = 256$ corresponds to the full frame.
• Figure 5: Application of the TCSPC-based filtration to the imaging of concentration in living cells. Figures a and b depict raw fluorescence decay patterns acquired in the pulse interleaved excitation, PIE, mode in living cells where two fluorophores (EGFP and TRITC-labelled dextran) were present. Figure a shows the data registered in the TRITC-channel ($\lambda_{\mathrm{em}}>594$ nm). Although the photons were registered in the TRITC channel that was hardware-filtered by the beamsplitter, some EGFP molecules emitted the photons of wavelengths exceeding the $\lambda_{\mathrm{em}}>594$ nm, due to the long tail of the emission spectra of the fluorescent protein. By analogy, Figure b depicts the fluorescence decay for the photons registered in the EGFP channel ($\lambda_{\mathrm{em}} = 525 \ pm 25$ nm, the band-pass filter). Here, part of the TRITC molecules were also excited by the 485 nm laser. Figure c shows boxplots comparing the $C_\mathrm{im}\slash C_\mathrm{FCS}$ ratio for the filtered and non-filtered data. The filtering was performed using a simple time-gating method, which involved selecting photons that fell within the specified decay-time range, along with the TCSPC background removal method.