How Actin Filaments Doff Their Pi Cap

The cytoskeletal protein actin (or one of its relatives) is found in almost all life forms. Yet, despite its ubiquity, there are still some fundamental things about its structure and function that are not well understood. For example, we know that monomeric actin subunits (called G-actin) polymerize into filaments (F-actin), but it’s still unclear how actin dynamics–filament assembly and disassembly, and turnover of individual actin subunits–are managed within these filaments. As cellular processes like motility are regulated by actin polymerization and depolymerization, understanding actin dynamics will provide important insight into these biological processes and how they in turn are regulated. Toward that end, Antoine Jégou and colleagues examine the details of actin filament polymerization and depolymerization in their paper published this week in PLoS Biology. It’s well known that individual actin molecules bind to the nucleotide ATP, an energy-providing molecule. Actin filaments grow asymmetrically by adding ATP-actin to just one end of the filament (called the ‘‘barbed end’’; the opposite end is called the ‘‘pointed end’’). ATP-actin addition is followed rapidly by ATP hydrolysis, leaving behind ADP plus a phosphate molecule (called Pi) bound to the actin subunit within the filament. These ADP-Pi-actin subunits can then release Pi to generate ADP-actin, and the subunits can eventually disassemble. Thus, a growing actin filament is composed of three different species of actin: ATP-actin (at the very tip), ADP-Pi-actin, and ADPactin. But many proteins that interact with actin have different affinities for these different forms of actin, and actin polymerization can be regulated by proteins binding to the filaments, making it a challenge to understand how these different species contribute to the makeup of actin filaments. Jégou and colleagues were particularly interested to know more about how Pi is released from ADP-Pi-actin within actin filaments, as this destabilizes the actin-actin bonds and thus lowers the filament rigidity. There are two theories that may explain what controls Pi release in actin filaments. The ‘‘vectorial’’ hypothesis holds that Pi release can only happen on ADP-Pi-actin subunits that abut an ADP-actin subunit, while the ‘‘random’’ hypothesis says that Pi release can happen anywhere in the actin filament. Jégou et al. set out to test these two hypotheses. Unfortunately, you can’t tell the nucleotide composition of a given actin subunit just by looking at it. Instead, it’s necessary to infer this information by watching how the filament as a whole behaves–specifically, how it disassembles. Actin filaments are polymers in equilibrium with monomers, and taking the monomers away causes their depolymerization. Adding additional free ATP-actin to the environment causes filaments to elongate at their barbed ends; taking it away makes them shrink from the barbed end on back. But ADP-Pi-actin falls off the barbed end of a filament more slowly than does ADPactin, so, by removing all the free actin and watching the rate at which filaments dissolve, one can infer whether the actin coming off the barbed end is ADP-actin or ADP-Pi-actin. Jégou and colleagues designed a special apparatus to carry out these observations. The new apparatus uses microfluidics to quickly remove all the free actin from a chamber containing actin filaments (that are anchored by their pointed ends to the chamber walls). As individual filaments then shrink in length, their behavior can be monitored with a microscope. Using this setup, the authors saw that filaments start out shrinking rather slowly, but that the rate of shrinkage accelerates in an exponential fashion. The only way to account for this behavior, say Jégou et al., is if Pi release from actin is random, not vectorial. They explain that actin subunits closer to the