Surviving ADP-BeFx-Arp2/3 complexes detach slowly from mother filaments. (A) Schematics of the renucleation experiment for unlabeled ADP-BeFx-Arp2/3 complex branch junctions. Branches are first nucleated and converted to BeFx for 6 min as in Fig. 1 C, then exposed to BeFx buffer with no G-actin for 6 (n = 150 branches) or 13 (n = 261 branches) minutes, before being exposed again to a regular buffer solution containing 0.4 µM Alexa Fluor 568 (10%)–G-ATP-actin (cyan) to allow branches to regrow. During the debranching step, the debranching time for each branch is recorded. In case of branch regrowth, the time interval between this debranching event and re-exposure to G-actin is calculated to assess the branch renucleation ratio versus time of ADP-BeFx-Arp2/3 complex branches. (B) Branch renucleation ratio for ADP-BeFx-Arp2/3 complex branch junctions, probed as shown in A, pulling with an average force of 5.2 pN. Each data point represents a branch renucleation ratio computed on at least 40 branches. Error bars are standard deviations. Exponential fit of the data gives a rate of detachment of the surviving ADP-BeFx-Arp2/3 complex from the mother filament of 2.8 (±2) 10⁻⁴ s⁻¹, and a fraction of Arp2/3 complexes remaining bound to mother filaments upon branch dissociation of 0.92 (±0.08). (C) Debranching rates as a function of constant pulling forces (log-linear representation), for ADP (blue)- and ADP-BeFx (green)–Alexa 488–labeled human Arp2/3 complex branch junctions, or in the presence of 50 mM Pi (light green). Error bars are standard deviations for the force, and confidence intervals on the single-exponential fits, obtained by fitting the upper and lower bounds of the 95% confidence intervals of the survival fractions from debranching experiments. Data are obtained from experiments with at least 28 analyzed branches for each condition. As a comparison, dashed lines represent the behavior of the unlabeled sheep Arp2/3 complex, as characterized in Figs. 1, 2, 3, and 4 in the main text. (D) Branch renucleation ratio as a function of the pulling force for Alexa 488–labeled human ADP-BeFx-Arp2/3 complexes (green), probed at 0.5 µM actin. Data points are from single experiments with 30 analyzed branches. The error bar for each individual data point is the standard deviation of the force, and the 95% confidence interval for the branch renucleation ratio. As a comparison, dashed lines represent the estimated branch renucleation ratio for unlabeled sheep Arp2/3 complexes at 0.5 µM actin. (E) Surviving fractions of unphotobleached Alexa 488–labeled Arp2/3 complex branch junctions over time, probed at an average pulling force of 1 pN, with images acquired every 2 s, obtained from 50 analyzed fluo-Arp2/3 complex branch junctions. Single-exponential fit (black curve) yields a photobleaching rate k_photobleach = 1.68 10⁻³ s⁻¹. The shaded area represents the 95% confidence interval. Vertical ticks on the surviving curves are times where censoring events occurred.