br Introduction Cell motility is fundamentally important in

Introduction Cell motility is fundamentally important in morphogenesis, wound healing, and the immune response. One of the best-studied basic types of cell movement is lamellipodial motility [1, 2], characterized by a thin (∼0.1–0.2 μm), broad (∼10–40 μm) motile appendage containing a dynamic recommended site network enveloped by the plasma membrane [3]. Lamellipodial protrusion is driven by actin polymerization at the leading edge [4] (Figure S1A). As new actin assembles at the front, older filaments are pushed away from the leading edge and eventually disassemble. Despite substantial progress in studying the regulation of lamellipodial actin dynamics [2, 5, 6], its complexity still defies quantitative understanding [7]. Here we focus on actin turnover, which is a fundamental part of lamellipodial motility. Two possible limiting scenarios for actin turnover have been put forward: the network-treadmilling model [8, 9, 10], in which actin assembly occurs primarily at the leading edge with slow network disassembly everywhere (global network turnover; Figure S1Bii), and the nucleation-release model [11, 12], where rapid actin assembly and disassembly take place throughout the lamellipodium (local network turnover; Figure S1Biii). Measurements of the actin turnover rates in the lamellipodium support the nucleation-release model in some cell types [12, 13, 14, 15] and network treadmilling in other cell types [16] (see STAR Methods). Actin turnover and recycling in vivo involve a host of actin-binding proteins that associate with actin monomers and/or filaments. The different actin subpopulations interchange continuously, and this has a substantial effect on the dynamics, as each subpopulation is characterized by different kinetics [2, 4]. The relative size of the actin subpopulations is largely unknown, as direct measurements by techniques such as labeling with actin-binding probes, photoactivation, or cell extraction cannot discriminate between all the various subpopulations [4, 17, 18, 19, 20, 21]. It is also unclear how the presence of monomer-binding proteins including thymosin and profilin influence the effective critical concentration of actin in vivo (see STAR Methods). To add to this complexity, recent research shows that the synergistic action of several proteins, including coronin, cofilin, Aip1, twinfilin, and Srv2/CAP [22, 23, 24, 25], as well as the activity of myosin motors [26, 27] accelerate network disassembly by promoting filament severing and debranching, which leads to the production of actin oligomers. Importantly, these studies imply that the actin network does not disassemble into monomers directly but rather first breaks into short actin filaments, which are disconnected from the network and small enough to diffuse in the cytoplasm. Because a detailed molecular picture of all the reactions involved in actin turnover in motile cells is still beyond reach [6, 28], our aim here is to characterize actin turnover and transport in a coarse-grained yet quantitative manner. We seek to measure the spatial distributions of the main lamellipodial actin subpopulations and the transitions between them. To avoid complications related to fluctuating actin concentrations, heterogeneous lamellipodium-lamellum actin networks [29], or the presence of a cell body with a cortical actin network that obscures measurements in the rear part of the cell, we use lamellipodial fragments from motile keratocytes [30, 31]. These fragments have a simple, persistent geometry and undergo rapid movement like whole cells, while lacking a cell body and any competing actin structures. Using this model system, we characterize the steady-state lamellipodial dynamics with unprecedented detail, combining experimental measurements with mathematical analysis to generate a comprehensive picture of actin turnover. We find that there is roughly 2-fold more actin in the diffusible pool compared to the network, and show that oligomers constitute a sizable fraction of this diffusible pool. The network turns over rapidly and locally, with actin assembly and disassembly occurring within seconds throughout the lamellipodium, while the diffusible actin remains nearly uniformly distributed. The model suggests that even though the actin network turnover is local, monomer transport is global. This global transport is made possible by the vast amount of non-polymerizable actin, which exchanges rapidly with the polymerizable actin pool. Consequently, actin subunits typically diffuse across the lamellipodium before reassembling into the network. This has profound implications for lamellipodial motility, allowing cells to move in a robust manner yet maintain the ability to rapidly adapt to changing conditions.