ATP-dependent actin barbed-end fluctuations at steady state

Abstract

Actin networks in cells are dynamic and constantly turning over as actin subunits exchange between monomer and polymer pools. Understanding these dynamics in vivo requires a detailed understanding of pure actin behavior in vitro. The prevailing model is treadmilling, which predicts continuous growth of filament barbed ends balanced by continuous shrinkage of pointed ends, while filament length remains roughly constant. While treadmilling has been observed directly at the level of single filaments, length fluctuations—expressed as “length diffusivity”—are much greater than expected, and the origin of these high diffusivity values could not be identified experimentally. By imaging single filaments tethered to glass via α-actinin, we observed frequent, low-amplitude barbed-end fluctuations of ±2–6 subunits per event and rarer, high-amplitude fluctuations of ±50–150 subunits per event. Barbed-end fluctuations depended on adenosine triphosphate (ATP) hydrolysis and release of inorganic phosphate (Pi) and were blocked by phalloidin and barbed-end capping agents. Barbed-end fluctuations consumed an estimated one-third of the ATP at steady state, while two-thirds is still consumed by treadmilling. Fluctuations result in diffusive spreading of filament lengths that exceeds predictions from pure treadmilling yet is lower than previously reported values, suggesting that earlier measurements captured additional sources of variability. Our results provide direct experimental evidence for kinetic models that proposed a dampened form of dynamic instability in pure actin where transient loss of the ATP/ADP•Pi barbed-end cap leads to rare, large excursions superimposed on more frequent, small fluctuations.