Dual-color live-cell confocal imaging detects a weak mechanical coupling of neighboring RDs. (A) Schematics of the experimental approach to measure the correlation angle α between trajectory pairs. A 2D temporal sequence (2 Hz, 30–300 frames) of dual-color–labeled CTs is acquired (left, green: ATTO 633–labeled RDs, magenta: ATTO 565–labeled RDs). The position of each RD is determined, and RD tracks are generated (middle). The correlation angle α is calculated at every time step from the corresponding displacement vectors (right). Then, for every trajectory pair, the mean α is calculated, and the averaged mean correlation angle <α> is extracted as a function of distance between pairs. To sample a broad distance, interval trajectories from dual-color–labeled RDs pulsed at increasing intervals between labeling pulses as described above, Δt = 0 min, 15 min (n = 3,144 pairs, 26 cells), 30 min (n = 8,256 pairs, 27 cells), 45 min (n = 5,126 pairs, 21 cells), 60 min (n = 5,306 pairs, 15 cells), 90 min (n = 4,103 pairs, 22 cells), and 120 min (n = 4,329 pairs, 22 cells) are acquired. (B) Time-lapse sequence of dual-color–labeled RDs at Δt = 0 min (top) and 60 min (bottom). Bars: 2 µm; (magnified portion) 200 nm. The zoomed-in panels show the correlated movement of RDs for Δt = 0 min and the uncorrelated movement of RDs for Δt = 60 min. (C) Density plot of averaged mean correlation angles for every trajectory pair <α> versus the spatial distance (blue line <α>, SD shaded blue region). Bundled data are from all Δt datasets (n = 30,264 pairs, 133 cells). Two linear-regression models were fitted respectively to data with distances <400 nm and data with distances > nm (red dashed lines). The transition point of correlated to uncorrelated movement was determined by the intersection of both (gray dotted lines).