Mass-sensitive particle tracking to elucidate the membrane-associated MinDE reaction cycle

Abstract

In spite of their great importance in biology, methods providing access to spontaneous molecular interactions with and on biological membranes have been sparse. The recent advent of mass photometry to quantify mass distributions of unlabeled biomolecules landing on surfaces raised hopes that this approach could be transferred to membranes. Here, by introducing a new interferometric scattering (iSCAT) image processing and analysis strategy adapted to diffusing particles, we enable mass-sensitive particle tracking (MSPT) of single unlabeled biomolecules on a supported lipid bilayer. We applied this approach to the highly nonlinear reaction cycles underlying MinDE protein self-organization. MSPT allowed us to determine the stoichiometry and turnover of individual membrane-bound MinD/MinDE protein complexes and to quantify their size-dependent diffusion. This study demonstrates the potential of MSPT to enhance our quantitative understanding of membrane-associated biological systems.

Fig. 1. Principle of MSPT.

a, Schematic displaying the iSCAT-based measurement principle of MSPT. Exemplary structures of three aldolase oligomer states (PDB 4S1F, ref. ) are shown in the top panel, and their respective iSCAT images at the bottom. Scale bar, 1 µm. b, Probability density distributions of standard proteins determined using the conventional mass photometry landing assay (left) or using MSPT (right). All data represent pooled distributions of three independent experiments per condition: alcohol dehydrogenase (ADH) (particle number n = 9,828), BSA (n = 11,408), TEV protease (TEV) (n = 1,705), β-amylase (bAm) (n = 10,043), protein A (prA) (n = 12,720); divalent streptavidin (Strep) (n = 16,699 trajectories), divalent streptavidin with biotinylated aldolase (Strep-ALD) (n = 16,727 trajectories), divalent streptavidin with biotinylated BSA (Strep-BSA) (n = 8,842 trajectories) and divalent streptavidin with biotinylated protein A (Strep-prA) (n = 22,424 trajectories). Dashed lines mark peaks not considered for mass calibration (left). Continuous lines represent oligomer states included in the mass calibration. Two-dimensional plots of mass versus diffusion coefficient for the four proteins measured with MSPT (right) are shown in Supplementary Fig. 4. c, Comparison of the contrast-to-mass calibration for mass photometry and MSPT, derived from peak contrasts in b and their assigned sequence masses (Supplementary Tables 2 and 3). Error bars represent the standard error of the peak locations estimated by bootstrapping. d, Two-dimensional KDE of 1.25 nM tetravalent streptavidin bound to biotinylated lipids on a SLB (n = 73,901 trajectories of three independent replicates; particle density: 0.2 µm⁻²). Marginal probability distributions of the molecular mass (top) and the diffusion coefficient (right) are presented. Source data

Fig. 2. Lateral MinD–MinD interactions lead to self-assembly into large homo-oligomers.

a, Schematic of the canonical membrane binding–unbinding cycle of MinDE. Upon ATP complexation, MinD dimerizes (1) and attaches to the membrane interface (2). In the event of MinE binding (3), MinE stimulates the intrinsic ability of MinD to hydrolyze ATP, which upon inorganic phosphate (P_i) release leads to the dissociation of MinD from the membrane in its monomeric form (4). b, MinD mass distribution in solution (gray line) (n = 16,101 particles) and on attachment to the SLB (blue line) (n = 13,917 trajectories). For solution experiments, 175 nM MinD with 0.5 µM ATP were measured in the conventional mass photometry landing assay. The membrane mass distribution of MinD was determined using MSPT at a particle density of 0.03 µm⁻². c,d, Two-dimensional KDE of membrane-attached MinD (c) and MinD D40A (d) at particle densities of 0.1 µm⁻² (light blue) (n = 117,086 trajectories) and 0.8 µm⁻² (dark blue) (n = 152,685 trajectories) and 0.1 µm⁻² (light green) (n = 7,831 trajectories) and 0.8 µm⁻² (dark green) (n = 3,150 trajectories), respectively. Marginal probability distributions of both molecular mass (top) and diffusion coefficient (right) are presented. e,g, Representative mass distributions (gray) of MinD (e) and MinD D40A (g) and estimation (black line, colored lines highlight underlying components) of its six components (MinD monomer–hexamer, light blue–dark blue; MinD D40A monomer–hexamer, light green–dark green) for three different particle densities (0.1 µm⁻², 0.3/0.5 µm⁻² and 0.8 µm⁻²). f,h, Relative oligomer abundance as a function of particle density: MinD (f) and MinD D40A (h). Error bars are the standard deviation of fitting results from three data subsets. The oligomer analysis is based on a total of n = 1,102,940 trajectories for MinD and n = 194,545 for MinD D40A. Source data

Fig. 3. Subunit (dis-)assembly of MinD particles diffusing on membranes.

a, Schematic representation of the time-resolved mass analysis of single MinD trajectories, which reveals attachment (at) and detachment (dt) events along the trajectory as well as a particle’s full membrane release (rl) at the end of its trajectory. b, Representative mass time traces of MinD trajectories (gray line) and corresponding step fits (black line) determined by a step-finding algorithm that locates mass change points within a trajectory. c, Mass step size distribution derived from step fits as depicted in b revealing MinD subunit sizes for at and dt events at particle densities: 0.1 µm⁻², pale blue (n = 20,796 plateaus); 0.3 µm⁻², light blue (n = 26,177 plateaus); 0.6 µm⁻², blue (n = 25,864 plateaus) and 0.8 µm⁻², dark blue (n = 7,506 plateaus). d, Dwell time plots for MinD particles before at events (top plot), representing the lengths of mass plateaus preceding a mass increase and before dt events (bottom plot), representing the lengths of mass plateaus preceding a mass decrease. Dwell times are shown for the MinD dimer (66 kDa) (light blue triangles) and the MinD tetramer state (132 kDa) (dark blue triangles). Plateau numbers for at are dimer, n = 23,782; tetramer, n = 5,088 and for dt are dimer, n = 3,143; tetramer, n = 10,406. Inset shows box plots that indicate second and third quantile (box), median (horizontal line) and 1.5× the interquartile range (whiskers) of bootstrapped mean dwell times (n = 10,000). e, MinD mass distribution for rl events. MinD particle densities: 0.1 µm⁻², pale blue (n = 117,086 plateaus); 0.3 µm⁻², light blue (n = 169,957 plateaus); 0.6 µm⁻², blue (n = 284,916 plateaus) and 0.8 µm⁻², dark blue (n = 120,654 plateaus). f, Plot of the dwell times before rl for the MinD dimer (light blue) and tetramer state (dark blue). Plateau numbers are dimer, n = 562,011; tetramer, n = 73,037. Inset shows box plots, details of which are as described in d. Source data

Fig. 4. MinE interconnects MinD oligomers into large complexes with a prolonged membrane dwell time.

a, Two-dimensional KDE of membrane-attached MinDE complexes at particle densities of 0.1 µm⁻² (pink) (n = 200,436 trajectories) and 0.8 µm⁻² (purple) (n = 30,082 trajectories). Marginal probability distributions of molecular mass (top) and diffusion coefficient (right) are presented. b, Mass step size distribution revealing MinDE subunit turnover (at and dt events) on membrane-bound particles at particle densities of 0.1 µm⁻², pale pink (n = 105,438 plateaus); 0.3 µm⁻², light pink (n = 73,471 plateaus); 0.6 µm⁻², pink (n = 9,247 plateaus) and 0.8 µm⁻², purple (n = 4,040 plateaus). c, MinDE mass distribution for membrane release (rl) at MinDE membrane particle densities of: 0.1 µm⁻², pale pink (n = 200,436 plateaus); 0.3 µm⁻², light pink (n = 158,660 plateaus); 0.6 µm⁻², pink (n = 35,501 plateaus) and 0.8 µm⁻², purple (n = 30,082 plateaus). d, Analysis of oligomer-specific diffusion coefficients for MinD (blue lines) and MinDE complexes (pink lines). Light blue/pink, dimer (MinD/MinDE n = 439,568/206,422 trajectories); dark blue/purple, tetramer (MinD/MinDE n = 118,922/47,136 trajectories). Inset shows box plots that indicate second and third quantiles (box), median (horizontal line) and 1.5× the interquartile range (whiskers) of bootstrapped diffusion coefficient maximum (n = 10,000). e, Dwell time plots for MinD and MinDE attachment (at) events (top) as well as for detachment (dt) events (bottom). Dwell times are shown for the MinD dimer (light blue) and tetramer (dark blue) as well as for their respective MinDE versions (hetero-dimer state, pink; hetero-tetramer state, purple). Plateau numbers for at are dimer MinD/MinDE n = 23,782/37,278; tetramer MinD/MinDE n = 5,088/11,698 and for dt are dimer MinD/MinDE n = 3,143/3,974, tetramer MinD/MinDE n = 10,406/22,501. Inset shows box plot details, as described in d for mean dwell times. f, Plot of the dwell times before rl for the MinDE dimer and tetramer state and the respective MinD versions. Plateau numbers are dimer MinD/MinDE n = 562,011/277,782; tetramer MinD/MinDE n = 73,037/60,114. Inset shows box plot details, as described in e. Source data

Fig. 5. Schematic of the proposed membrane-associated MinDE reaction cycle.

1. On nucleotide exchange from ADP to ATP, membrane-dependent dimerization of MinD is triggered. 2. On the membrane, lateral MinD interactions and recruitment of MinD subunits from solution lead to MinD higher-order structures that assemble through attachment of subunits at a location different from the canonical dimerization site. MinD assemblies then either dissociate from the membrane (3) or encounter MinE (4). MinE promotes the interconnection of very large heteromeric MinDE complexes that, due to their multivalent MTS structure, reside substantially longer on the membrane interface (5). 6. However, MinE also induces nucleotide-conversion of MinD subunits, thereby weakening the overall membrane avidity of the MinDE complex, before its full release from the membrane interface in complexes >350 kDa.