Fig 2: CELFIS quantifies CD95 oligomerization.(A) Measurement objectives. (B) Confocal fluorescence image of mEGFP- and mCherry-labeled CD95 in the cell membrane. Cells 1 to 3 are alive. Cell 4 underwent apoptosis. (C) Methodological approach. Left: Distribution of donor fluorescence lifetimes in absence (D0; gray) and presence (DA; yellow) of FRET. FRET-induced donor decay εD(t) with fluorescence fraction (xFRET) in presence of the acceptor. Right: Conversion of 1% xFRET into 2.8% oligomer fraction from theoretical considerations accounting for the probability of mature donor-acceptor pairs, pAD, and the fluorophore cloud correction, ξ. The conversion was also experimentally confirmed. a.u., arbitrary units. (D) xFRET as a function of receptor surface density. About 3 % xFRET values confirm the monomeric character of CD86, and nearly constant ~37 % xFRET values confirm the dimeric nature of CTLA4. xFRET of CD95 alone indicates primarily monomeric (≥96%) and some dimeric (≤4%) receptors. For CD95(ΔDD), ≥88% monomers and 12% dimers are found. After CD95L incubation, ≤21% of CD95 or CD95(ΔDD) receptors form oligomers. N > 108 cells; ≥4 independent experiments per condition. Note the adjusted y axis. (E) Dynamics of oligomerization after CD95L addition. Box plots of oligomer fraction calculated from n(max) cells. The “max” in case of “dead CD95 + L” indicates the initial cell number, which decreases over time. Dashed line indicates 15% oligomer fraction from (G). (F) Exemplary evolution of the oligomer fraction in single cells. Legend as in (E). (G) Oligomer fraction right before apoptosis. (H) Oligomerization rate over ≤3 hours, depending on the apoptosis time point. Legend as in (E). Two-sided Mann-Whitney U test, ***P < 0.001.

Fig 3: Quantitative STED imaging reveals randomly distributed CD95 spots and small oligomer formation.(A) Objective of STED measurements. (B) Schematic representation of CD95-mEGFP with GFP-nanobody Atto647N labeling. (C) Exemplary STED image (left) of HeLa CD95KO membrane transiently transfected with CD95-mEGFP and deconvolved image (right) using Huygens Professional software (see STED imaging and analysis in Materials and Methods). On average, 20 spots/μm2 were detected. Green box indicates threshold-based detected spot analyzed in (D). (D) Gray panel illustrates methodological approach: Spot centers from deconvolved images are registered and superposed on raw data for Gaussian fitting (see Materials and Methods). From the fit, the SD σ, the brightness, and the average number of photons 〈NPh〉 per pixel are derived. (E) 2D probability density representation of 〈NPh〉 per pixel and σ values derived from individual spot analysis. Frequency histograms of 〈NPh〉 per pixel and σ are depicted on the side and top of each graph. Left column: Simulation of monomer receptors up to dimer of trimer receptors per spot including their random distribution on the membrane surface. Simulation parameters (brightness, σ, and crowding factor; for details, see Materials and Methods and table S3) were adjusted to match the measured CD86 and CD86-mEGFP-mEGFP distributions precisely. Parameters of all oligomer simulations were kept constant. Right column: Measured 〈NPh〉 per pixel and σ for monomer, pseudo-dimer control, and CD95 before and after CD95L incubation (for other receptors, see fig. S9). From each simulation, isolines enclosing 95% of data points are calculated and depicted in different panels for data comparison. N > 5000 objects analyzed per sample.

Fig 4: Probing CD95 signal initiation models with receptor variants over a broad range of molecular concentrations and in space and time.(A) Structure and cartoon of CD95 with genetically fused mEGFP and trimeric CD95L. Four-letter abbreviations are protein data bank IDs. For simplicity, only one of up to three CD95 receptors is shown together with CD95L. (B) Scheme of proposed TNFR signal initiation models. Model 1: Monomeric receptors bind trimeric ligands and form up to trimer-trimer receptor-ligand configurations. In the receptor activated state, the intracellular death domain (DD) recruits an adaptor molecule [Fas-associated death domain protein (FADD) in case of CD95]. A cascade of (pro)caspase activation follows (74) along with mitochondrial dysfunction (75) and protein cleavage, resulting in cell apoptosis (76). Model 2: Before activation, TNFRs form inactive dimers, which assemble into a supramolecular hexagonal lattice (units of ~24 nm in diameter depending on TNFR) (12). After ligand binding, the receptor dimers decouple and recruit FADD to the DDs. FADD may cross-link the DDs, from where the (pro)caspase cascade evolves as in model 1. (C) Overview of test strategy using (super-resolution) microscopy and multiparametric fluorescence spectroscopy techniques covering single molecule to cellular scales. (D) Scheme of engineered CD95 variants exhibiting different signaling competencies (I to VIII), monomer controls (IX to XI), and dimer controls (XII and XIII). Bicistronic plasmids are used for CELFIS, and monocistronic plasmids are used with all techniques. Numbers indicate the amino acid, and dashed lines indicate optional linkers. Gray panel indicates a methodological highlight.