Single-particle plasmon spectroscopy has become a standard technique to detect and quantify the presence of unlabeled macromolecules. Here, we extend this method to determine their exact distance from the plasmon sensors with sub-nanometer resolution by systematically varying the sensing range into the surrounding by adjusting the size of the plasmonic nanoparticles. We improved current single-particle plasmon spectroscopy to record continuously for hours the scattering spectra of thousands of nanoparticles of different sizes simultaneously with 1.8 s time resolution. We apply this technique to study the interaction dynamics of bacterial MM proteins with supported lipid membranes of different composition. Our experiments reveal a surprisingly flexible operating mode of the MM proteins: In the presence of cardiolipin and membrane curvature induced by nanoparticles, the protein oscillation occurs on top of a stationary MinD patch. Our results reveal the need to consider membrane composition and local curvature as important parameters to quantitatively understand the MM protein system and could be extrapolated to other macromolecular systems. Our label-free method is generally easily implementable and well suited to measure distances of interacting biological macromolecules.