Intracellular actin networks assemble through the addition of ATP-actin subunits at the growing barbed ends of actin filaments. This is followed by \"aging\" of the filament via ATP hydrolysis and subsequent phosphate release. Aged ADP-actin subunits thus \"treadmill\" through the filament before being released back into the cytoplasmic monomer pool as a result of depolymerization at filament pointed ends. The necessity for aging before filament disassembly is reinforced by preferential binding of cofilin to aged ADP-actin subunits over newly-assembled ADP-Pi actin subunits in the filament. Consequently, investigations into how cofilin influences pointed-end depolymerization have, thus far, focused exclusively on aged ADP-actin filaments. Using microfluidics-assisted Total Internal Reflection Fluorescence (mf-TIRF) microscopy, we reveal that, similar to their effects on ADP filaments, cofilin and cyclase-associated protein (CAP) also promote pointed-end depolymerization of ADP-Pi filaments. Interestingly, the maximal rates of ADP-Pi filament depolymerization by CAP and cofilin together remain approximately 20-40 times lower than for ADP filaments. Further, we find that the promotion of ADP-Pi pointed-end depolymerization is conserved for all three mammalian cofilin isoforms. Taken together, the mechanisms presented here open the possibility of newly-assembled actin filaments being directly disassembled from their pointed-ends, thus bypassing the slow step of Pi release in the aging process.

Intracellular actin networks assemble through the addition of ATP-actin subunits at the growing barbed ends of actin filaments. This is followed by \"aging\" of the filament via ATP hydrolysis and subsequent phosphate release. Aged ADP-actin subunits thus \"treadmill\" through the filament before being released back into the cytoplasmic monomer pool as a result of depolymerization at filament pointed ends. The necessity for aging before filament disassembly is reinforced by preferential binding of cofilin to aged ADP-actin subunits over newly-assembled ADP-Pi actin subunits in the filament. Consequently, investigations into how cofilin influences pointed-end depolymerization have, thus far, focused exclusively on aged ADP-actin filaments. Using microfluidics-assisted Total Internal Reflection Fluorescence (mf-TIRF) microscopy, we reveal that, similar to their effects on ADP filaments, cofilin and cyclase-associated protein (CAP) also promote pointed-end depolymerization of ADP-Pi filaments. Interestingly, the maximal rates of ADP-Pi filament depolymerization by CAP and cofilin together remain approximately 20-40 times lower than for ADP filaments. Further, we find that the promotion of ADP-Pi pointed-end depolymerization is conserved for all three mammalian cofilin isoforms. Taken together, the mechanisms presented here open the possibility of newly-assembled actin filaments being directly disassembled from their pointed-ends, thus bypassing the slow step of Pi release in the aging process.

Cellular actin networks display distinct assembly and disassembly dynamics resulting from multicomponent reactions occurring primarily at the two ends and the sides of actin filaments [1-3]. While barbed ends are considered the hotspot of actin assembly [4], disassembly is thought to primarily occur via reactions on filament sides and pointed ends [3, 5-11]. Cyclase-associated protein (CAP) has emerged as the main protagonist of actin disassembly and remodeling - it collaborates with cofilin to increase pointed-end depolymerization by 300-fold [6, 7], promotes filament \"coalescence\" in presence of Abp1 [12], and accelerates nucleotide exchange to regenerate monomers for new rounds of assembly [13-15]. CAP has also been reported to enhance cofilin-mediated severing [16, 17], but these claims have since been challenged [7]. Using microfluidics-assisted three-color single-molecule imaging, we now reveal that CAP also has important functions at filament barbed ends. We reveal that CAP is a processive barbed-end depolymerase capable of tracking both ends of the filament. Each CAP binding event leads to removal of about 5,175 and 620 subunits from the barbed and pointed ends respectively. We find that the WH2 domain is essential, and the CARP domain is dispensable for barbed-end depolymerization. We show that CAP co-localizes with barbed-end bound formin and capping protein, in the process increasing residence time of formin by 10-fold and promoting dissociation of CP by 4-fold. Our barbed-end observations combined with previously reported activities of CAP at pointed ends and sides, firmly establish CAP as a key player in actin dynamics.

As a regulator of actin filament turnover, Arabidopsis thaliana CAP1 plays an important role in plant growth and development. Here, we analyzed the phenotypes of two Arabidopsis cap1 mutants: cap1-1 (a T-DNA insertion mutant) and Cas9-CAP1 (generated by CRISPR-Cas9 gene editing). Phenotypic analysis demonstrated that loss of CAP1 results in defects in seed germination and seedling morphology, with some seedlings exhibiting one or three cotyledons. The cap1-1 mutant took longer than the wild type to complete its life cycle, but its flowering time was normal, indicating that loss of CAP1 prolongs reproductive but not vegetative growth. Moreover, loss of CAP1 severely reduces seed production in self-pollinated plants, due to disruption of pollen tube elongation. RNA-seq and qRT-PCR analyses demonstrated that CAP1 may be involved in osmotic stress responses. Indeed, the cap1-1 mutant showed increased tolerance of salt and mannitol treatment, indicating that CAP1 plays a negative role in osmotic stress tolerance in Arabidopsis. Taken together, our results demonstrate that CAP1 functions not only in plant growth and development, but also in Arabidopsis responses to osmotic stress.